Abstract
Coordination between transcription and replication is crucial in the maintenance of genome integrity. Disturbance of these processes leads to accumulation of aberrant DNA:RNA hybrids (R‐loops) that, if unresolved, generate DNA damage and genomic instability. Here we report a novel, unexpected role for the nucleopore‐associated mRNA export factor Ddx19 in removing nuclear R‐loops formed upon replication stress or DNA damage. We show, in live cells, that Ddx19 transiently relocalizes from the nucleopore to the nucleus upon DNA damage, in an ATR/Chk1‐dependent manner, and that Ddx19 nuclear relocalization is required to clear R‐loops. Ddx19 depletion induces R‐loop accumulation, proliferation‐dependent DNA damage and defects in replication fork progression. Further, we show that Ddx19 resolves R‐loops in vitro via its helicase activity. Furthermore, mutation of a residue phosphorylated by Chk1 in Ddx19 disrupts its interaction with Nup214 and allows its nuclear relocalization. Finally, we show that Ddx19 operates in resolving R‐loops independently of the RNA helicase senataxin. Altogether these observations put forward a novel, ATR‐dependent function for Ddx19 in R‐loop metabolism to preserve genome integrity in mammalian cells.
Synopsis
The helicase Ddx19, which functions as a nucleopore‐associated mRNA export factor, has an unexpected additional nuclear role in resolving R‐loops and maintaining genome stability upon DNA damage or replication stress.
Ddx19 translocates into the nucleus upon DNA damage.
Ddx19 nuclear translocation is mediated by the ATR‐Chk1 signaling pathway.
Recombinant Ddx19 displays RNA:DNA helicase and R‐loop‐resolving activity in vitro.
Ddx19 downregulation generates nuclear R‐loops and DNA double‐strand breaks.
Ddx19 function in R‐loop metabolism is independent of the RNA helicase senataxin.
Introduction
The DNA damage response (DDR) is a surveillance pathway that slows down cell proliferation to facilitate DNA repair and recovery from DNA damage (Ciccia & Elledge, 2010). At the apex of the DDR are protein kinases of the PI3K family, among which are ATM and ATR. DNA damage activates ATM and ATR that, in turn, phosphorylate a large number of substrates thus mediating the cellular response to DNA damage (Matsuoka et al, 2007). ATM orchestrates the response to double‐strand breaks (DSBs), resulting in the phosphorylation of both chromatin‐bound and nuclear‐soluble substrates such as the histone variant H2AX (known as γH2AX) and the Chk2 protein kinase. A large number of proteins are then recruited to DSBs, including 53BP1, to form macroscopic nuclear foci that sustain DNA repair (Lukas et al, 2011). ATR is primarily activated upon replication fork slowdown or arrest, a process also known as “replication stress” by certain types of DNA lesions (induced by UV light for example) that lead to functional uncoupling of DNA polymerases and DNA helicase activities, resulting in γH2AX accumulation and phosphorylation of the Chk1 protein kinase (Byun et al, 2005; Recolin et al, 2014). Replication stress, induced by DNA lesions that do not uncouple replication forks, as those induced by radiomimetic drugs (doxorubicin, bleomycin), as well as by difficult to replicate DNA and/or chromatin structures, induces ATM activation, also leading to γH2AX accumulation (Recolin et al, 2014, for review).
Interference between DNA transcription and DNA replication was first analyzed in bacteria (French, 1992) and originally reported to occur in eukaryotic genomes as a source of instability (Prado & Aguilera, 2005). This phenomenon which was also revealed by loss of function of the DNA topoisomerase I gene (Tuduri et al, 2009; El Hage et al, 2010) represents an important source of replication stress (Wellinger et al, 2006; Gan et al, 2011; Hamperl & Cimprich, 2016). Interference between these two processes generates DNA damage in part through accumulation of DNA:RNA hybrids known as R‐loops. These aberrant structures form transiently during transcription (Huertas & Aguilera, 2003) in particular at discrete CpG island‐containing promoters (Ginno et al, 2012). This thermodynamically stable structure is very toxic for the cell if left unresolved, since it can promote formation of DSBs and recombination, thus fueling replication stress‐induced genomic instability (Wellinger et al, 2006; Gan et al, 2011; Skourti‐Stathaki & Proudfoot, 2014; Sollier et al, 2014; Costantino & Koshland, 2015). In addition, mutations in genes controlling RNA biogenesis have been reported to induce R‐loops and associated genomic instability (Huertas & Aguilera, 2003; Li & Manley, 2005; Paulsen et al, 2009; Wahba et al, 2011). The mechanism of action is thought to occur by interference with ongoing transcription leading to accumulation of R‐loops and generation of replication stress (Aguilera & Garcia‐Muse, 2012). In yeast, inhibition of mRNA export through the nuclear pore (gene gating) has also been shown to induce replication stress and activation of the DNA damage checkpoint (Bermejo et al, 2011). Altogether these observations suggest that both sequestration of the mRNA into a ribonucleoparticle (mRNP) and its nuclear export contribute to the exclusion of nascent RNA from the transcription unit thus limiting the occurrence of R‐loops (Gomez‐Gonzalez et al, 2011). In eukaryotic cells, at least two enzymes have been shown to clear R‐loops, RNase H1, and senataxin (Skourti‐Stathaki & Proudfoot, 2014, for a review). RNase H1 is an RNA endonuclease that cleaves the RNA moiety of the R‐loop. Senataxin is a nuclear protein that binds nascent transcripts and interacts with components of the transcription machinery. Senataxin facilitates R‐loop resolution during transcription termination (Mischo et al, 2011; Skourti‐Stathaki et al, 2011) and during interference between transcription and replication in yeast and mammalian cells (Alzu et al, 2012). Recent data show that in mammalian cells, senataxin is targeted to a subset of R‐loops by interaction with the tumor suppressor BRCA1 (Hill et al, 2014; Hatchi et al, 2015). More recently, the RNA helicase Aquarius (Sollier et al, 2014) and the DNA helicase FANCM (Garcia‐Rubio et al, 2015; Schwab et al, 2015) have also been implicated in R‐loop resolution in mammalian cells. Whether there are other enzymes implicated and how R‐loop resolution is linked to the cell cycle is currently unknown.
The identification of mutations in DDR genes observed in tumors at late stages of malignancy (Bartkova et al, 2005) has fueled the search for new DDR genes. In such a search, using an in vitro screen involving cell‐free Xenopus laevis egg extracts, we identified the highly conserved Ddx19/Dbp5/Rat8 RNA helicase as a DDR‐responsive factor. Ddx19 is a superfamily‐2 DEAD box helicase having both ATPase and RNA unwinding activity (Tseng et al, 1998; Schmitt et al, 1999; Alcazar‐Roman et al, 2006). Ddx19 functions in mRNA nuclear export through mRNP remodeling on the cytoplasmic face of the nuclear pore (Snay‐Hodge et al, 1998; Schmitt et al, 1999; Zolotukhin et al, 2009), where it interacts with the nucleoporin Nup159/CAN/Nup214 (Hodge et al, 1999; Schmitt et al, 1999; Weirich et al, 2004; Napetschnig et al, 2009). In mammalian cells, the RBM15 protein bridges the interaction between the mRNP, bound to the essential nuclear export factor Nxf1/TAP1/Mex67 in the interchromatin space, and Ddx19 (Zolotukhin et al, 2009). At the nuclear pore, Ddx19 interacts with two other proteins, IP6 and Gle1. Gle1 activates Ddx19 thus facilitating unidirectional mRNA export (Alcazar‐Roman et al, 2006; Weirich et al, 2006; Montpetit et al, 2011). Here we report a novel, unexpected role for Ddx19 in resolving R‐loops arising from conflicts between transcription and replication, and after DNA damage, and show that this function depends upon the ATR‐Chk1 pathway that mediates its nuclear re‐localization. Moreover, we show that this regulation is important for maintenance of genome integrity by removal of these aberrant structures.
Results
Identification of XDdx19 as DNA damage‐responsive factor
With the aim of identifying new DNA damage‐responsive factors, we screened for proteins that accumulate in UV‐irradiated nuclei reconstituted in vitro using Xenopus egg extracts (see Materials and Methods). This experimental system is naturally transcriptionally silent and supports very efficiently both chromatin assembly and nuclear formation in vitro upon addition of exogenous DNA, as well as the activation of the DDR signaling. Upon screening over ten thousand in vitro translated proteins obtained from independent cDNA pools, we isolated one cDNA, among others, coding for a protein of 65 kDa (Fig 1A) that through database searching, we identified as being a Xenopus ortholog of the mammalian Ddx19 RNA helicase (XDdx19, 85% of identity, Fig EV1A and B). Nuclear accumulation of XDdx19 was also observed after inducing replication stress with aphidicolin, an inhibitor of replicative DNA polymerases that activates ATR (Fig 1B). Analysis of the dynamics of nuclear retention shows that upon UV irradiation XDdx19 accumulates after nuclear membrane formation and not before (Fig 1C). Ddx19 did not accumulate in UV‐damaged nuclei formed in the presence of geminin, a strong inhibitor of DNA synthesis that also precludes ATR activation (Fig 1D, lane 3). Since in this transcription‐free system both aphidicolin and UV light induce replication stress and consequent replication‐dependent Chk1 phosphorylation (Byun et al, 2005; Fig EV1C), altogether these observations show that XDdx19 nuclear retention is both DDR‐ and replication stress‐dependent. Of note, removal of XDdx19 from egg extracts by immunodepletion did not significantly affect global DNA replication (Fig EV1D and E).
Diagram depicting the general principles and workflow of the screen to identify new DDR‐responsive factors. In vitro translated proteins transcribed from individual cDNA pools were incubated in Xenopus egg extracts supplemented with UV‐irradiated or undamaged sperm nuclei. Upon incubation at room temperature for 1 h, nuclei were recovered as described in Materials and Methods and proteins were eluted with Laemmli buffer followed by SDS–PAGE and autoradiography. Right: Autoradiography of proteins translated from a cDNA pool (71.1) and eluted from nuclei formed in the absence (−) or presence (+) of UV irradiation (UV‐C, 300 J/m2). The arrow indicates a 65 kDa polypeptide that specifically accumulates into nuclei after UV irradiation (UV‐C). kDa indicates molecular mass of standard protein markers.
Nuclear accumulation of XDdx19 upon UV‐C irradiation (300 J/m2) or inhibition of DNA synthesis with 100 μg/ml of aphidicolin.
Dynamics of XDdx19 nuclear accumulation during a time course upon UV‐C irradiation (time post‐UV). A sample of the reaction incubated at room temperature for 120 min in the absence of UV irradiation (−) was also included. The time of nuclear membrane formation coincides with the onset of DNA synthesis in this system, and is indicated by an arrow.
Nuclear accumulation of XDdx19 before (−) or after (+) UV‐C irradiation in the presence (+) or absence (−) of the inhibitor of replication fork formation geminin (100 nM).
Source data are available online for this figure.
Source Data for Figure 1 [embj201695131-sup-0018-SDataFig1.pptx]
Alignment of Xenopus and human Ddx19 proteins obtained with Clustal W software.
Detection of XDdx19 with a specific human Ddx19 antibody (HDdx19). Left panel: Xenopus egg extracts or human total cell lysate from HEK293T cells were fractionated by SDS–PAGE and analyzed by Western blot with an anti‐HDdx19 antibody. In parallel, the XDdx19 cDNA isolated from the screen was in vitro transcribed either with T3 (sense) or T7 (antisense) RNA polymerase and translated (IVT), followed by Western blot with the same antibody (middle panel). An aliquot of the in vitro translated reactions was also analyzed by autoradiography (right panel).
Induction of Chk1 phosphorylation upon UV irradiation of Xenopus sperm chromatin incubated in Xenopus egg extracts. Nuclei were isolated as described in Materials and Methods and analyzed by Western blot with the indicated antibodies.
Western blot of Xenopus egg extracts after depletion with control (Mock) or XDdx19 antibodies as described in Materials and Methods.
Kinetics of DNA synthesis of Xenopus egg extracts upon depletion with control antibodies (Mock) or XDdx19 antibodies followed by incorporation of a radioactive nucleotide precursor.
Ddx19 relocalizes to the nucleus in an ATR‐dependent manner upon DNA damage in mammalian cells
To confirm a general involvement of Ddx19 in DNA damage, we analyzed Ddx19 expression in mammalian cells by Western blotting upon exposure to a variety of DNA damaging agents. Ddx19 accumulates in mammalian cells treated with UV light (Fig 2A), but equally with camptothecin (an inhibitor of topoisomerase I), aphidicolin, and the alkylating agent MMS (Fig EV2A–C). However, no changes were observed upon treatment with the radiomimetic drug doxorubicin, which by inducing DSBs activates ATM, or with the interstrand cross‐linking agent mitomycin C (Fig EV2D and E), suggesting that Ddx19 responds to specific types of DNA damage. Curiously, we observed increased Ddx19 abundance only in a short time window of about 30 min post‐irradiation, while no changes were observed upon longer times of recovery (Fig 2B lanes 1–3, and Fig EV2F). Since the abundance of Ddx19 mRNA did not change after UV irradiation (Fig 2C), it suggests stabilization by post‐translational modifications. Consistent with this possibility, inhibition of the proteasome after UV irradiation resulted in Ddx19 stabilization after an even longer recovery time (Fig 2B, lanes 4–6). Interestingly, the inhibition of ATR, and not ATM, impaired Ddx19 stabilization only after UV damage, suggesting a dependency on ATR activity (Fig 2D, lanes 5–6). Observation of eGFP‐tagged Ddx19 by live‐cell imaging (Fig 2E and Movies EV1 and EV2) and quantification in fixed cells (Fig EV2G) shows that in the absence of UV irradiation (− UV) Ddx19 localizes in the cytoplasm and around the nuclear periphery, as expected (Schmitt et al, 1999). Interestingly, upon UV irradiation (+ UV) nuclear relocalization of eGFP‐Ddx19 was observed. Relocalization was transient and occurred during a 30‐min time window (Fig 2E), consistent with its transient stabilization observed by Western blot (Fig 2A and B), after which Ddx19 shuttles back to the nuclear periphery (Movies EV1 and EV2). No such relocalization was observed for GFP‐tagged tubulin (Movies EV3 and EV4), nor with CFP‐tagged lamin A, a component of the nuclear envelope after UV irradiation (Movies EV5 and EV6). Of note, no relocalization of the nucleoporin Nup153 was previously observed after DNA damage (Wan et al, 2013). We also observed UV‐dependent accumulation of endogenous Ddx19 on chromatin (Fig EV2H). Interestingly, UV‐dependent relocalization of Ddx19 was not observed upon inhibition of ATR, while ATM inhibition had no effect (Fig 2E and Movies EV7, EV8, EV9 and EV10). Altogether these results show that in mammalian cells, upon UV damage Ddx19 global levels increase and Ddx19 transiently relocalizes from the nuclear periphery to the nucleus in an ATR‐dependent manner.
Western blot of HeLa cells irradiated with increasing doses of UV‐C. Total extracts were prepared 30 min later as described in Materials and Methods and analyzed with the indicated antibodies. Relative abundance of HDdx19 was quantified with ImageJ software, normalized to the loading control (β‐actin) and indicated as fold increase in respect to the untreated condition (first lane).
Western blot of HeLa cell total protein extracts treated with DMSO (−) or 20 μM MG132 for 3 h before UV‐C irradiation and collected at indicated times after irradiation. The replication licensing factor Cdt1, which is rapidly degraded after UV irradiation (Tsanov et al, 2014), was included as a control. Signals were quantified using ImageJ software and normalized to the loading control MCM2. Fold In indicates the fold change of Ddx19 levels relative to the untreated condition (first lane: time 0, – MG132).
Relative mRNA levels of Ddx19 measured by qRT–PCR in HeLa cells either non‐irradiated or irradiated with increasing UV‐C doses after 30 min recovery. Transcript levels were normalized to four internal control genes: GAPDH, HPRT, SDHA, and HMBS. Means and standard deviation (SD) values for three independent experiments are shown.
Abundance of Ddx19 in cells treated with UV‐C light with or without ATM (KU5593) or ATR (VE‐821) inhibitors. Western blot of HeLa cell protein extracts treated or not with the indicated inhibitors 2 h before UV irradiation (20 J/m2) and added to the culture medium during recovery. Samples were collected after 30 min following UV‐C exposure.
Left: Time frames of HeLa cells transfected with eGFP‐tagged Ddx19 in the absence (control) or presence of either ATR or ATM inhibitors. The same cells were filmed over a time‐lapse of 1 h before and 1 h after UV‐C irradiation at 20 J/m2. T indicates time in minutes (min). Scale bar: 5 μm. See also Movies EV1, EV2 and Movies EV7, EV8, EV9 and EV10. Right: Quantification of the experiment shown in left panel. Fifty transfected nuclei were scored for each condition. The graph represents the mean and standard deviation values from three independent experiments.
A–F Cellular levels of HDdx19 after exposure to the indicated sources of DNA damage. Total extracts were analyzed with the indicated antibodies at the indicated times (hours).
G. Left: Subcellular localization of eGFP‐Ddx19 before and after UV‐C irradiation, as visualized by indirect immunofluorescence microscopy. γH2AX staining serves as a positive control for DNA damage and nuclear marker in addition to DAPI to visualize nuclei. Scale bar: 10 μm. Right: Quantification of transfected cells showing nuclear localization of GFP‐Ddx19. Mean values and standard deviations from three independent experiments are shown. One hundred nuclei were scored for each condition. ***P < 0.001 (Student's t‐test). See also Movies EV1 and EV2.
H. Chromatin binding of endogenous HDdx19 to chromatin. Western blot of HeLa chromatin fractions prepared 30 min after irradiation or not with UV‐C. Extracts were probed with the indicated antibodies. Absence of GAPDH signal demonstrates the purity of the fractionation procedure.
Ddx19 depletion induces DNA damage throughout the nucleus
We next depleted Ddx19 by RNAi in mammalian HeLa cells and analyzed the occurrence of global DNA damage and replication stress with specific DNA damage markers. Inhibition of Ddx19 expression with two different siRNAs resulted in accumulation of γH2AX, a widely used marker of both DSBs and replication stress (Figs 3A and EV3A). Importantly, γH2AX accumulation was suppressed by expression of an siRNA‐resistant Xenopus Ddx19 cDNA (Fig EV3B, lanes 7–10), demonstrating the specificity of the observed phenotype and thus excluding off‐target effects. This result suggests that loss of Ddx19 leads to increased genomic instability, consistent with a previous report (Paulsen et al, 2009), and similar to knockdown of other factors involved in mRNA metabolism. Analysis of additional DNA damage markers showed the presence of phosphorylated ATM, p53, and Chk2, but not Chk1, suggesting activation of the ATM pathway that monitors the presence of DSBs (Fig 3A). Observation of Ddx19‐depleted cells by indirect immunofluorescence microscopy revealed the presence of a high level of γH2AX staining not restricted to the nuclear envelope as expected, but largely present throughout the nucleus, as revealed by co‐staining with lamin (Fig 3B). Analysis of 53BP1 localization, a more specific marker of DSBs than γH2AX (Schultz et al, 2000; Bouquet et al, 2006), showed the presence of a large number of 53BP1 foci in cells depleted of Ddx19, scattered throughout the nucleus (Fig 3C). Analysis of genomic DNA by pulse‐field gel electrophoresis (PFGE) confirmed the presence of DSBs upon Ddx19 downregulation (Fig 3D). Consistent with this observation, γH2AX accumulation upon Ddx19 downregulation was sensitive to ATM and not ATR inhibition (Fig EV3C). Downregulation of mRNA export factors has been reported to also induce DNA damage indirectly by interfering with mRNA synthesis and release. Hence, we analyzed the phenotype of cells depleted of the major mRNA export factor Nxf1 and compared it to that generated by Ddx19 depletion. Nxf1 downregulation strongly inhibited nuclear mRNA export as expected, and resulted in γH2AX induction (Fig EV3D and E). Interestingly, however, the pattern of 53BP1 accumulation observed was different than that observed upon Ddx19 knockdown, in that accumulation of 53BP1 foci was observed mostly close to the nuclear periphery (Fig 3C). Taken together, these observations suggest that Ddx19 depletion in mammalian cells induces DNA damage, including DSBs, scattered throughout the nucleus, while downregulation of the mRNA export factor Nxf1 generates DNA damage preferentially around the nuclear periphery, suggesting a possible additional role of Ddx19 in nuclear mRNA metabolism.
Western blot of total cell extracts prepared from HeLa cells transfected with either a control siRNA (siLuc) or a Ddx19‐specific siRNA (si_b) and probed with the indicated antibodies.
Immunofluorescence microscopy fields showing the pattern of γH2AX and lamin A staining after knockdown of Ddx19 for 48 h in HeLa cells. Scale bar: 20 μm. Nuclei were visualized by DAPI staining. Arrows indicate mitotic cells with pan γH2AX staining.
Left: HeLa cells were transfected with non‐specific siRNA (Luc), or siRNA specific for Ddx19 or Nxf1, and 53BP1 foci formation was analyzed by indirect immunofluorescence. Scale bar: 10 μm. Right: Cells displaying 53BP1 foci were quantified and expressed in the graph as percent of 53BP1‐positive cells. Cells in which the staining was located around the nuclear periphery or throughout the nucleus were also scored. One hundred nuclei were scored for each condition. The graph represents the mean and standard deviation values from three independent experiments.
Induction of DSBs upon Ddx19 downregulation. HeLa cells transfected with siDdx19 or negative control (siLuc) for 72 h were collected and subjected to PFGE analysis. The image is representative of n = 3 experiments. Arrows indicate the occurrence of DNA breaks. A sample of cells treated with the radiomimetic drug bleomycin (5 mU/ml) for 6 h (lane 2) is included as a positive control.
Western blot of total cell extracts prepared from cells transfected with either a control siRNA (si‐Luc) or two independent Ddx19‐specific siRNAs (b, d). Extracts were probed with the indicated antibodies.
Rescue of the Ddx19 siRNA‐dependent phenotype by expression of the siRNA‐resistant Xenopus XDdx19. HeLa cells were co‐transfected with either control siRNA (siLuc) or four different siDdx19 (a‐d), and an empty vector (EV) or XDdx19 cDNA. Total cell lysates were collected after 72 h and analyzed by Western blot with the indicated antibodies.
Phosphorylation of H2AX (γH2AX) upon Ddx19 knockdown is sensitive to ATM inhibition. Western blot of total cell extracts prepared from HeLa cells transfected with either a control siRNA (siLuc) or a Ddx19‐specific siRNA for 48 h, then treated for 6 h with mock, ATR‐ or ATM‐specific inhibitors prior to analysis.
Western blot of total cell extracts prepared from cells transfected with either a control siRNA (si‐Luc), Ddx19 siRNA, or Nxf1 siRNA and probed with the indicated antibodies.
Poly(A)+ RNA localization by RNA FISH examined in control (siLuc) or Nxf1‐depleted HeLa cells 72 h post‐transfection. Nuclei are indicated by DAPI staining. Scale bar: 5 μm.
Ddx19 relieves conflicts between replication and transcription
We also observed that mammalian cells treated with Ddx19 siRNA accumulated in the S and G2/M phases of the cell cycle, as assessed by measurement of the DNA content by FACS analysis (Fig 4A upper panel). Pulse labeling with the nucleotide analogue BrdU to monitor DNA synthesis revealed a strong reduction in the number of actively replicating cells (Fig 4A, lower panel), suggesting that most of the Ddx19‐depleted cells had reduced DNA synthesis. To obtain further insights into the contribution of DNA replication in the generation of DNA damage upon Ddx19 depletion, we monitored γH2AX levels in non‐proliferating, serum‐starved, primary human cells (RPE‐1) and found no difference compared to cells treated with a control siRNA (Fig 4B). In contrast, strong γH2AX accumulation was detectable in proliferating RPE‐1 cells depleted of Ddx19 (Fig 4C), as a result of checkpoint activation, confirming results in HeLa cells. Importantly, we did not observe γH2AX accumulation in replicating nuclei isolated from transcription‐incompetent Xenopus extracts depleted of XDdx19, but rather only in extracts set for transcription (Prioleau et al, 1994; Amodeo et al, 2015) lacking Ddx19 (Fig 4D). Taken together these results suggest that Ddx19 may be involved in genome maintenance in proliferating cells in resolving conflicts arising between transcription and replication.
A. FACS analysis of propidium iodide (upper panel) staining and incorporation of BrdU (lower panel) in HeLa cells treated with either control siRNA (siLuc) or Ddx19‐specific siRNA (siDdx19). Ddx19 knockdown efficiency was > 90%.
B, C Left: Western blot of total RPE‐1 cell extracts upon serum starvation (B) or asynchronously growing (C) after knockdown of Ddx19. Cell cycle distribution was verified by FACS analysis (right panel).
D. Induction of γH2AX in egg extracts incompetent or competent for transcription upon Ddx19 depletion. Upper panel: Transcription competence of extracts supplemented with either 1,000 or 10,000 nuclei/μl of egg extract was determined by incorporation of the ribonucleotide analogue EU. Nuclei were detected by DAPI staining. Lower panel: Induction of γH2AX in egg extracts incompetent or competent for transcription upon Ddx19 depletion. Egg extracts depleted with either control antibodies (Mock) or XDdx19 antibodies (Ddx19) were supplemented with either 1,000 or 10,000 nuclei/μl of egg extract and incubated at room temperature for 90 min after which nuclei were recovered as described in Materials and Methods. Eluted nuclear proteins were analyzed by Western blotting with the indicated antibodies. Scale bar: 5 μm.
E. Diagram explaining the experimental setup for sequential pulse labeling. Cells were treated by siRNA for 72 h. RNase H1 overexpression was achieved by transient transfection 48 h after siRNA treatment. Cordycepin was added at 50 μM for 90 min prior to the first pulse labeling. Camptothecin was added at 1 μM where indicated at the time of the second CIdU pulse.
F. CldU (red) tracks length quantification of the experiment shown in panel (E). For each condition the lengths of the CIdU tracks of n = 100 fibers were quantified using the ImageJ software. Fibers were labeled with anti‐DNA antibody, and only intact fibers were scored. No differences in the fibers length were observed between the different conditions and only red tracks preceded by a green tract were measured to assure that incorporation had already been initiated. A representative plot of three independent experiments was generated using the GraphPad Prism software. The bar dissecting the data points represents the median of tract length. ns: not significant, **P < 0.005, ***P < 0.001, ****P < 0.0001 (Mann–Whitney U‐test).
We then analyzed the dynamics of DNA replication in cells treated with Ddx19 siRNA by fiber stretching, a technique that allows analysis of ongoing DNA synthesis (Jackson & Pombo, 1998). To this end, cells were first pulse‐labeled with the nucleotide analogue IdU (green) to identify newly replicated DNA regions, and then CIdU (red) as depicted in Fig 4E, to visualize the progression of replication forks during the second pulse. Quantitative analysis of track length shows that, in cells treated with Ddx19 siRNA, both IdU‐ (Fig EV4A) and CldU‐labeled tracks (Fig 4F, lanes 1–2) were significantly shorter than those observed in control cells, suggesting a slowdown of ongoing fork progression. In contrast, inhibition of Nxf1 expression had only a mild, non‐significant effect on ongoing DNA replication (Fig 4F, lane 3). Since data shown in Fig 4B–D suggest that endogenous DNA damage arising upon Ddx19 knockdown is dependent upon both replication and transcription, we investigated the possibility that shorter CldU tracks in Ddx19 depleted cells may be due to conflicts between these two processes, leading to R‐loop formation. To this end, cells transfected with Ddx19 siRNA were also transiently transfected with RNase H1 to specifically degrade the RNA moiety in RNA:DNA hybrids, thus removing the R‐loop. RNase H1 treatment alone did not significantly affect track lengths in control cells, but restored normal track lengths in Ddx19‐depleted cells (Fig 4F, lanes 4–5). To further assess the occurrence of transcription‐dependent replication stress in Ddx19‐depleted cells, we inhibited transcription elongation using the RNA‐specific chain terminator 3′‐deoxyadenosine cordycepin (COR) by adding it 90 min prior to pulse labeling (Fig EV4B). Quantification reveals that cordycepin significantly restored the reduced CldU track lengths observed in Ddx19‐silenced cells to a degree comparable to that obtained by RNase H1 overexpression (Fig 4F, compare lanes 2, 5–7), while it did not have any effect on control cells. These results prompted us to determine whether Ddx19 may be required for replication fork recovery in the presence of DNA damage. To this end, cells were treated with camptothecin, which by inhibiting topoisomerase I, leads to replication fork stalling and generates DNA damage such as DSBs and single‐stranded DNA gaps (Sakasai & Iwabuchi, 2016). Cells were treated with camptothecin for a short period during the second, 30‐min pulse labeling to challenge ongoing DNA synthesis with DNA damage. As can be seen (Fig 4F, lanes 8–9), upon camptothecin addition, red tracks were much shorter in cells treated with Ddx19 siRNA compared to those observed in control cells. Because camptothecin also contributes to stabilization of R‐loops, this result suggests that Ddx19 also functions in recovery of replication forks challenged by DNA damage and/or R‐loop. In summary, these observations are consistent with Ddx19 playing a role in facilitating DNA synthesis progression and its recovery, and suggest that this might depend on R‐loop metabolism.
Quantification of ldU (green) tracks length of the experiment shown in Fig 4F. For each condition the length of the IdU tracks of n = 100 fibers were quantified using the ImageJ software. Only intact fibers labeled with blue anti‐DNA were analyzed, and only green tracks that are followed by a red tract. A representative scatter blot of three independent experiments was generated using the Graphpad Prism software. The bar dissecting the data points represents the median of tract length. ****P < 0.0001 (Mann–Whitney U‐test).
Inhibition of transcription by cordycepin. Quantification of EU incorporation into nuclei upon treatment with cordycepin. One hundred nuclei were scored for each condition. The graph represents the mean and standard deviation values from three independent experiments.
ATPase assay of recombinant Ddx19WT or the Ddx19E243E helicase‐DEAD mutant.
Quantification of mcherry‐RNaseHD10R‐E48R mutant or mcherry‐RNaseHWT foci per nucleus in cells expressing a stable shRNA targeting RNase H1, and treated with either Ddx19 siRNA or siLuc (control). One hundred nuclei were scored per condition and represented as a plot using GraphPad Prism. The bar dissecting the data points represents the median. ***P < 0.005 (Mann–Whitney U‐test). ns: non‐significant.
Western blot of total cell extracts prepared from cells transfected with either a control siRNA (si‐Luc), Ddx19 siRNA, or Gle1 siRNA and probed with the indicated antibodies.
Cell viability of HeLa cells treated with control (siLuc) or Ddx19‐specific siRNA for 72 h, after which either DMSO or camptothecin (CPT) was added to the cells for an additional 24 h before assessing viability. In one situation cells were also transfected with RNase H1 plasmid 48 h after RNAi treatment.
HeLa cells were transfected with the indicated expression vectors. Total cell extracts were made 24 h post‐transfection and analyzed by Western blot with the indicated antibodies.
Recombinant Ddx19 resolves RNA:DNA hybrids in vitro
To explore a possible function of Ddx19 in R‐loop metabolism, we expressed and purified recombinant Ddx19 in bacteria, and assayed for its ability to unwind RNA:DNA hybrids in vitro. We observed that Ddx19WT can efficiently unwind RNA:DNA hybrids and that unwinding is dependent upon its ATPase activity (Figs 5A and EV4C). Kinetic analysis shows that recombinant Ddx19WT can unwind ~50% of an RNA:DNA hybrid within 20 min in vitro (Fig 5B and C). In addition, we tested the ability of recombinant Ddx19WT to resolve R‐loops formed in vitro on a plasmid transcribing the CpG island mAirn DNA (Fig 5C) previously described to form an R‐loop (Ginno et al, 2012). In this system, R‐loop formation causes a reduced mobility of a supercoiled plasmid in agarose gels compared to non‐transcribed or RNase H1‐treated samples in which the shift is not detectable (Powell et al, 2013). In line with results shown in Fig 5A and B, Ddx19WT, but not helicase‐dead mutant (Ddx19E243Q), eliminated the mobility shift of a transcribing plasmid, similar to RNase H1 treatment, which demonstrates the ability of Ddx19 to unwind an R‐loop structure in vitro (Fig 5D).
Unwinding reactions using the indicated substrate (depicted on the right hand side of the panel) were set up in the presence of increasing amounts of recombinant HDdx19 wild‐type (WT, lanes 4–5) or Ddx19DEAD mutant (E243Q, lane 6). Unwinding activity of Ddx19WT was also assayed with no ATP (lane 7), or in the presence of non‐hydrolysable ATP (ATP‐γS, lane 8). The presence of free radiolabeled RNA (star, cartoon on the right) is indicative of the unwinding activity observed.
Time course of Ddx19 wild‐type unwinding activity of the indicated substrate (depicted on the right hand side of the panel).
Quantification of radioactive signals shown in (B). Signals were quantified using ImageJ software, and percentage of unwinding was calculated relative to the input signal of the hybrid product at t = 0 min.
In vitro R‐loop unwinding assay. pFC53 plasmid (lane 1) was transcribed in vitro (lane 2), and samples were treated with RNase A only, or in addition to either RNase H1 (lane 3), Ddx19WT (lane 4), or Ddx19DEAD mutant (lane 5). Samples were fractionated by agarose gel electrophoresis and visualized by ethidium bromide staining.
Source data are available online for this figure.
Source Data for Figure 5 [embj201695131-sup-0019-SDataFig5.pptx]
Ddx19 functions in R‐loop metabolism in vivo
We next probed cells depleted of Ddx19 with the anti‐RNA:DNA S9.6 antibody to detect formation of R‐loops in cells by indirect immunofluorescence microscopy. We observed accumulation of RNase H1‐sensitive R‐loops in cells transfected with Ddx19 siRNA (Fig 6A, upper panel). We also detected accumulation of nuclear R‐loops using a different approach. This consists in expressing the catalytically inactive RNase HD10R E48R mutant that remains bound to the R‐loop (Britton et al, 2014). Using this system, we observed that nuclear foci of this mutant and not wild‐type RNase H significantly accumulated in cells depleted of Ddx19 compared to control cells (Fig EV4D). R‐loops were observed throughout the nucleus, highly enriched in the nucleolar compartment, similar to the pattern previously observed in cells depleted of senataxin (Yeo et al, 2014), and consistent with the pattern of γH2AX staining and 53BP1 foci (Fig 3C). Of note, in cells depleted of Nxf1, R‐loops accumulated mainly close to the nuclear periphery (Fig 6A, upper panel), again consistent with the observed pattern of 53BP1 foci upon Nxf1 downregulation (Fig 3B and C), and in line with a function of Nxf1 at the nuclear face of the nuclear pore. In addition, cells depleted of Ddx19 displayed a statistically higher number of these hybrid structures than Nxf1‐depleted cells (Fig 6A, lower panel). Importantly, significant accumulation of R‐loops was not observed upon downregulation of the essential Ddx19 cofactor for mRNA export at the nuclear pore Gle1 (Figs 6A and EV4E). This result strongly argues that accumulation of R‐loops is specific to inhibition of Ddx19 expression. We also confirmed R‐loop accumulation by DNA:RNA hybrids immunoprecipitation (DRIP assay) using the S9.6 antibody (Fig 6B) at various actively transcribed genes previously reported to naturally form these structures (Ginno et al, 2012). We observed that Ddx19 knockdown induced R‐loop accumulation at the analyzed gene loci (BTBD19, RPL13A, and EGR1), to a similar level than cells depleted of senataxin. Importantly, this was not the case upon downregulation of Gle1, the Ddx19 cofactor in mRNA export. Moreover, Ddx19‐depleted cells showed a significantly reduced viability compared to control cells, which was severely exacerbated by low doses of camptothecin, and rescued by RNase H1 treatment (Fig EV4F), suggesting a dependency upon R‐loop accumulation. Taken together, these observations suggest a possible additional role for Ddx19 in R‐loop metabolism, alongside its mRNA export function at the nucleopore.
Upper panel: Immunostaining with S9.6 (green) and nucleolin (red) antibodies in HeLa cells transfected with the indicated siRNAs for 72 h. In conditions involving RNase H1 treatment, cells were transfected also with RNase H1 plasmid 48 h after siRNA transfection. A merge of the two channels is shown, with the nucleus (stained with DAPI) outlined. Scale bar represents 10 μm. The signal levels of all panels were adjusted equally with ImageJ software. Lower panel: Quantification of RNA:DNA hybrids per nucleus for the experiment described in the upper panel. The DAPI signal was used to create a mask of the nucleus. The nuclear S9.6 signal intensity was then determined by subtracting the nucleolin signal and analyzing the remaining S9.6 signal. One hundred nuclei were scored per condition and represented as a plot using GraphPad Prism. The bar dissecting the data points represents the median. **P < 0.005, ****P < 0.0001 (Mann–Whitney U‐test).
DRIP‐qPCR using the S9.6 antibody at the genes BTBD19, EGR1, RPL13A (positive loci) and SNPRN (negative locus), in HeLa cells upon 72 h transfection with the indicated siRNAs. Genomic DNA extracted from the samples was treated with RNase H1 as a negative control. The relative abundance of RNA:DNA hybrids immunoprecipitated is represented as percentage of the input material. The graph shows mean and SD values from three experiments. *P < 0.05, **P < 0.005, ***P < 0.0005 (Mann–Whitney U‐test).
We then set off to pinpoint the requirement of specific Ddx19 domains in R‐loop resolution, by assaying the ability of either Ddx19WT or of several Ddx19 mutants previously described (Hodge et al, 2011; Schmitt et al, 1999 and Fig EV4G) to remove R‐loops induced by inhibiting topoisomerase I with camptothecin (Tuduri et al, 2009; El Hage et al, 2010; Marinello et al, 2013). Nuclear relocalization of GFP‐Ddx19 was observed in live cells upon camptothecin treatment (Movie EV11 and Fig EV5A). While R‐loops accumulated in cells transfected with empty vector, they were strongly reduced by expression of Ddx19WT, similar to treatment with RNase H1 or by senataxin expression (Fig 7A). In contrast, overexpression of Ddx19 helicase‐dead mutant (Ddx19E243Q) or a mutant in the RNA‐binding domain (RMR, Ddx19R372G) did not suppress R‐loop formation, despite these mutants being nuclear localized in the presence of camptothecin (Fig EV5A). Interestingly, expression of these mutants induced an increase in R‐loops even in the absence of camptothecin, suggesting a dominant‐negative effect. Importantly, overexpression of Ddx19 mutated in the nucleopore binding site, that is defective in nuclear mRNA export (Ddx19S138R, Rajakyla et al, 2015; Schmitt et al, 1999), suppressed R‐loops to a similar extent as either Ddx19WT or senataxin (Fig 7A) and localized into the nucleus upon camptothecin treatment (Fig EV5A), making unlikely the possibility that R‐loop clearance induced by ectopic expression of Ddx19WT is solely due to increased nuclear export. Consistent with this possibility, ectopic expression of Nxf1 was not effective in resolving R‐loops (Fig 7A), although Nxf1 was localized into the nucleus after camptothecin treatment (Fig EV5B).
A, B HeLa cells were transfected with the indicated constructs and were untreated (− CPT) or treated with 250 nM of camptothecin (+ CPT) for 18 h prior fixation. R‐loops were scored by indirect immunofluorescence using the S9.6 antibody and eGFP‐Ddx19, eGFP‐Nxf1 or eGFP‐senataxin (SETX) subcellular localization was observed by eGFP fluorescence. Scale bar: 10 μm.
C. Quantification of the Western blot shown in Fig 7D. Signals were quantified using ImageJ software and normalized to the loading control MCM2 (ratio of GFP‐Ddx19/MCM2).
Quantification of transfected cells showing nuclear R‐loops staining. HeLa cells were transfected with the indicated constructs and treated with 250 nM of camptothecin (CPT) for 18 h prior to fixation. Formation of R‐loops was determined by indirect immunofluorescence using the S9.6 antibody. Transfected cells showing nuclear S9.6 signal (excluding nucleolar signals) were scored positive. Scoring was performed for 100 nuclei per condition, and the graph shows the mean and SD values for three independent experiments.
Subcellular localization of eGFP‐Ddx19 mutants in the Chk1 phosphorylation site and their effect on R‐loop resolution. All GFP‐positive cells displayed the subcellular localization shown in the picture. Scale bar: 10 μm.
Ddx19S93E does not interact with Nup214, nor with senataxin. Left panel: HeLa cells expressing the indicated constructs minus (lanes 1, 3, 5, 7) or plus (2, 4, 6, 8) camptothecin (250 nM), were lysed and GFP‐tagged proteins were recovered from total cell extracts using GFP‐trap beads. Right panel: HeLa cells expressing GFP‐DdX19WT were lysed and analyzed by immunoprecipitation to check for Nup214 interaction. Samples were loaded onto a SDS–PAGE and analyzed by Western blot with anti‐GFP, anti‐Nup214 or senataxin (SETX) antibodies.
Suppression of Chk2 phosphorylation induced upon Ddx19 knockdown by expression of DdX19WT or the indicated Ddx19 mutants. HeLa cells were co‐transfected with either control (siLuc) and empty vector (EV) or Ddx19‐specific siRNA and the indicated eGFP‐Ddx19 mutants or myc‐tagged RNase H1. Total extracts were obtained and analyzed by Western blot with the indicated antibodies. A star indicates a non‐specific cross‐reacting polypeptide.
R‐loop suppression by Ddx19 requires Chk1‐dependent nuclear localization
Following the observation that Ddx19 nuclear localization induced by UV irradiation is dependent on ATR but not ATM, we reasoned that the Ddx19 function in R‐loop suppression might depend upon the ATR pathway. Ddx19 was previously identified in a phosphoproteomic screen as a substrate of the Chk1 kinase, a key ATR downstream target. The site phosphorylated by Chk1 in Ddx19 was mapped to serine 93 (Blasius et al, 2011). To determine the significance of Ddx19 phosphorylation by Chk1, we mutated Ddx19Ser93 to either a non‐phosphorylatable form (Ddx19S93A) or to a phosphomimetic form (Ddx19S93E). We then analyzed both nuclear localization and the ability of these mutants to clear R‐loops. In accordance with the impaired nuclear relocalization of Ddx19WT observed upon ATR inhibition (Movies EV7 and EV8), Ddx19S93A did not localize to the nucleus after UV irradiation and did not clear R‐loops (Fig 7A and B, and Movies EV12 and EV13). In contrast, the Ddx19S93E phosphomimetic mutant was constitutively nuclear and was very effective in suppressing R‐loops (Fig 7A and B, and Movies EV14 and EV15). Moreover, by immunoprecipitation, we did not observe complex formation between Ddx19S93E and the Nup214 nucleoporin compared to Ddx19WT (Fig 7C), suggesting that Ddx19 phosphorylation by Chk1 regulates Ddx19 localization at the nuclear pore.
We also determined the ability of the above‐described Ddx19 mutants to suppress spontaneous DNA damage generated in cells depleted of Ddx19 by RNAi. While Ddx19WT completely suppressed Chk2 phosphorylation observed upon Ddx19 downregulation, the Ddx19E243Q helicase‐dead or the Ddx19S93A mutant did not (Fig 7D, lanes 3–5 and Fig EV5C), suggesting that both Ddx19 helicase activity and nuclear localization are indispensable for suppressing genomic instability. Interestingly, Chk2 phosphorylation generated by Ddx19 knockdown was only partially suppressed by expression of nuclear export‐defective Ddx19S138R mutant, suggesting that not all DNA damage generated by Ddx19 knockdown depends upon inhibition of mRNA export. Consistent with this interpretation, expression of either RNase H1 or the phosphomimetic Ddx19S93E also partially reduced Chk2 phosphorylation (Fig 7D, compare lanes 6–8). This result argues for an R‐loop resolving function of Ddx19 independent of its mRNA export function, but rather dependent upon Chk1. Altogether these observations suggest that R‐loop resolving activity, suppression of endogenous DNA damage, and nuclear localization of Ddx19 are ATR/Chk1‐pathway dependent.
Ddx19 and senataxin function independently of each other
Finally, we analyzed senataxin expression and localization in cells depleted of Ddx19 and exposed to camptothecin. Downregulation of either Ddx19 or senataxin did not significantly alter the expression level of either proteins (Fig EV6A). Further, senataxin focus formation induced by camptothecin treatment, as previously reported (Yeo et al, 2014), was not impaired in cells depleted of Ddx19 (Fig EV6B and C). Interestingly, quantification shows the presence of a significant number of senataxin foci in cells depleted of Ddx19 compared to control cells, also in the absence of camptothecin (Fig EV6C), confirming spontaneous R‐loop formation in Ddx19‐depleted cells. Finally, no physical interaction between Ddx19 and senataxin could be observed by immunoprecipitation (Fig 7C). Altogether these observations suggest that Ddx19 downregulation does not alter senataxin levels and localization, and because R‐loops persist in cells depleted of Ddx19, they also suggest a non‐redundant function for Ddx19 in resolving R‐loops.
Ddx19 knockdown does not affect senataxin stability and vice versa. Western blot of total cell extracts upon transfection with the indicated siRNAs.
Cells were transfected with the indicated siRNA for 38 h, then treated with 250 nM of camptothecin for 18 h prior fixation. Endogenous senataxin (SETX) was visualized by indirect immunofluorescence with a senataxin‐specific antibody (Yuce & West, 2013), and γH2AX staining was used as a control for endogenous DNA damage. Scale bar: 10 μm.
Quantification of endogenous senataxin foci observed in panel (B) was obtained by scoring one hundred nuclei per condition. Only nuclei showing more than ten foci per cell were scored as positive. The graph represents mean and SD values from three independent experiments. *P < 0.05 (Student's t‐test).
Discussion
In this work, we propose a novel role for Ddx19 in R‐loop metabolism in mammalian cells, aside from its well‐known function in mRNA nuclear export at the nuclear pore. This conclusion is supported by several independent observations. First, Ddx19 downregulation induces both R‐loop accumulation and spontaneous DNA damage throughout the nucleus. R‐loops could be immunoprecipitated at genomic loci in cells depleted of Ddx19, while this was not the case upon downregulation of Gle1, the Ddx19 mRNA export cofactor that also localizes at the nuclear pore. Further, Ddx19 overexpression and not that of the mRNA export factor Nxf1 suppressed R‐loop formation, which makes unlikely the possibility that R‐loop clearance mediated by Ddx19 is a secondary effect due to increased mRNA export. Furthermore, the Ddx19S138R mRNA export‐defective mutant (Schmitt et al, 1999; Rajakyla et al, 2015) suppresses R‐loop formation, in contrast to the Ddx19 helicase‐dead or Ddx19 mutant in the RNA‐binding domain that do not. Ddx19 depletion also produces a phenotype distinct from that observed by depletion of Nxf1, or the TREX‐2 complex involved in mRNP biogenesis that also interacts with the nucleopore (Umlauf et al, 2013), which results in loose mRNA tethering to the nuclear pore (Bermejo et al, 2011), and only slightly induces formation of R‐loops in mammalian cells (Bhatia et al, 2014). Finally, we have demonstrated both RNA:DNA hybrids and R‐loop unwinding activity in vitro for recombinant Ddx19WT and not helicase‐dead Ddx19. Collectively these observations strongly suggest that R‐loop accumulation in the absence of Ddx19 is a consequence of a direct role in R‐loop metabolism although a contribution of impaired mRNA export cannot be ruled out. A nuclear role for Ddx19 in transcription was previously proposed in yeast by interaction with the TFIIH subunit of RNA polymerase II (Estruch & Cole, 2003). Our results suggest that in mammalian cells, this role is likely to be in removing R‐loops.
In addition, we have shown that the R‐loop resolving activity of Ddx19 is dependent upon the DDR, and in particular upon ATR. Our data propose a model in which activation of the ATR‐Chk1 pathway, by replication stress or by certain types of DNA damage that primarily activate ATR, triggers Ddx19 relocalization to the nucleus to clear R‐loops generated in proliferating cells (Fig 8). Because in the absence of Ddx19, ongoing DNA synthesis is slowed down, and replication forks fail to overcome DNA damage induced by camptothecin, it suggests that Ddx19 may be implicated in removing R‐loops generated upon interference between replication and transcription. This possibility is supported by the absence of spontaneous DNA damage observed in serum‐starved, transcriptionally active, non‐proliferating cells depleted of Ddx19. This is also consistent with the presence of a very low level of R‐loops observed in post‐mitotic cells (Yeo et al, 2014), and the absence of both spontaneous DNA damage and replication defects upon Ddx19 removal from transcription‐incompetent Xenopus egg extracts. Resolution of R‐loops (Britton et al, 2014) and formation of senataxin foci (Yuce & West, 2013) have been recently shown to be dependent upon the DDR, although the specific molecular mechanism implicated is currently unknown. However, unlike senataxin, whose focus formation was shown to be affected by inhibiting different branches of the DDR, including ATM and DNA‐PK (Yuce & West, 2013), we have shown that Ddx19 is specifically regulated by the ATR‐Chk1 pathway, and not by ATM, suggesting a dependency upon generation of replication stress. This interpretation may explain why Ddx19 also localizes into transcription‐incompetent Xenopus nuclei after UV damage, probably as a default ATR‐dependent relocalization triggered by activation of the replication checkpoint, although an implication in resolving other aberrant structures generated at arrested replication forks cannot be completely excluded. In yeast, it has been proposed that torsional stress, generated when a replication fork encounters a transcription unit, slows down DNA synthesis and activates the DDR that in turn phosphorylates nucleoporins and possibly factors involved in mRNA metabolism to induce the transient release of the mRNA from the nuclear pore, so to relieve torsional stress and facilitate fork progression (Bermejo et al, 2011). The results presented in this work propose Ddx19 as a novel important mediator of this regulation linking control of mRNA nuclear export with R‐loop clearance through Chk1 phosphorylation in mammalian cells. We speculate that once R‐loops are removed, the DDR signal is quenched and Ddx19 may be dephosphorylated or degraded, consistent with its transient proteasome‐dependent stabilization observed upon DNA damage. A similar scenario could also apply in the presence of DNA damage that activates the DDR in a replication‐dependent manner resulting in transient Ddx19 relocalization.
Activation of ATR by replication stress as a result of conflicts between replication and transcription, or induced by DNA damage (camptothecin, UV‐C, red stars) triggers Ddx19 phosphorylation by Chk1 and weakens its interaction with Nup214. This allows nuclear localization of Ddx19 and R‐loop resolution.
Efficient Ddx19 function is indeed essential for proper maintenance of genome stability as also previously proposed (Paulsen et al, 2009), although the specific molecular mechanism involved was unknown. In light of Ddx19's function in reducing R‐loop formation and facilitating interaction between transcription and replication, DSBs accumulation and subsequent activation of the ATM‐Chk2 pathway when Ddx19 is not present in the cell are very likely the consequence of persistent unresolved R‐loops (Sordet et al, 2009; Sollier et al, 2014), but also the result of cells entering mitosis with under‐replicated chromosomes. Finally, the observed reduced viability of cells depleted of Ddx19 in the presence of camptothecin puts forward Ddx19 as a novel target to sensitize cancer cells to treatment with this drug.
Materials and Methods
Screening for DDR‐responsive factors in Xenopus egg extracts
This screen was previously described (Lustig et al, 1997). Briefly a Xenopus eggs cDNA library constructed in the pRN3 vector (a kind gift from A. Zorn, Cincinnati, Ohio, USA) was diluted so to prepare small cDNA pools containing around 100 cDNAs per pool. The cDNA of each pool was in vitro transcribed and translated in the presence of 35S methionine in a rabbit reticulocyte lysate (TNT, Promega). Cytoplasmic Xenopus extracts (low speed egg extracts) were prepared as previously described (Murray, 1991). Xenopus egg extracts were then supplemented with radiolabeled in vitro translated proteins at a ratio of 1:4, 20 min after addition of sperm nuclei (2,000 nuclei/μl) in a final volume of 12 μl in 96‐well plates. The reaction was further incubated at room temperature for 90 min. Cytoplasmic and nuclear fractions obtained by subcellular fractionation as previously described (Recolin et al, 2012) were eluted with Laemmli buffer and analyzed by 12% SDS–PAGE. Proteins were blotted to a nitrocellulose membrane (Hybond C+), dried, and exposed to a PhosphorImager screen (Molecular Dynamics). Autoradiographic signals were revealed by phosphorimager scan (Molecular Dynamics).
DNA replication and transcription assays in egg extracts
Egg extracts were supplemented with α‐[32P]dATP (3,000 Ci/mmol, Perkin Elmer). At the indicated time points, samples were neutralized in 10 mM EDTA, 0.5% SDS, 200 μg/ml proteinase K (Sigma) and incubated at 37°C overnight. Incorporation of radioactive label was determined by TCA‐precipitation on GF/C glass fiber filters (Whatman) followed by scintillation counting. For detection of EU incorporation in nuclei assembled in egg extracts, 5‐EU (5 mM final concentration) was added starting at the time of sperm DNA addition. Extracts were incubated for 90 min at room temperature and nuclei were then fixed onto coverslips as previously described (Recolin et al, 2012). EU detection was done with the Click‐iT RNA Alexa Fluor 594 Imaging Kit (Invitrogen, Molecular Probes) according to the manufacturer's protocol.
Cell culture
Cells were grown in Dulbecco's modified Eagle's medium (DMEM) for HeLa cells or DMEM‐F12 for RPE‐1 cells, supplemented with 10% fetal bovine serum and 2 mM glutamine at 37°C under a humidified atmosphere with 5% CO2.
Preparation of cell lysates and immunoblotting
Cells were harvested post‐treatment, washed once in PBS, and incubated with ice‐cold lysis buffer (50 mM Tris–HCl, pH 7.4, 100 mM NaCl, 50 mM NaF, 5 mM EDTA, 40 mM β‐glycerophosphate, 1% Triton X‐100 and protease inhibitors) for 30 min on ice. Whole‐cell lysates were clarified by high‐speed centrifugation at 16,000 g for 10 min. 25 μg of proteins was loaded per lane and separated on 10% SDS–PAGE and then transferred onto nitrocellulose membranes (Amersham, GE Healthcare). Membranes were blocked with 5% milk in TBST wash buffer for 1 h then incubated with the appropriate primary antibody. Secondary antibodies were peroxidase‐conjugated anti‐mouse and anti‐rabbit antisera (Sigma‐Aldrich, Saint‐Louis, MO, USA).
Plasmids
Human eGFP‐Ddx19 wild‐type and mutant expression vectors were previously described (Schmitt et al, 1999). Human eGFP‐Ddx19 mutants in the Chk1 phosphorylation site were obtained by site‐directed mutagenesis of wild‐type eGFP‐Ddx19. The GFP‐tagged Nxf1 was a gift from E. Bertrand. The GFP‐tagged senataxin construct was previously described (Yuce & West, 2013).
Recombinant proteins
Recombinant XDdx19 was obtained by expression in E. coli (BL21 codon Plus) upon subcloning the Xenopus Ddx19 cDNA from pRN3XDdx19 into pET28a(+) from Novagen. Induction was obtained upon addition of 0.7 mM IPTG for 3 h at 37°C. Recombinant XDdx19 was purified to homogeneity on a Nickel column (Qiagen). The recombinant protein was used to immunize rabbits to obtain Xenopus‐specific Ddx19 antibodies. Recombinant HDdx19 used in the ATPase and RNA:DNA hybrids unwinding assay was expressed and purified in bacteria as for XDdx19 except that cells grown at 37°C to an A600 of 0.4 were shifted to 18°C and induced with 0.25 mM IPTG for 20 h. Eluted proteins were dialyzed through a Microcon membrane (8 kDa MW cutoff, Amicon) into 30 mM Tris pH 7.5, 150 mM NaCl, 2 mM MgCl2, 1 mM DTT, and 5.0% glycerol.
Antibodies
The XDdx19 antibody was raised in rabbits against recombinant 6HisXDdx19. The following antibodies were purchased from the indicated companies: Ddx19 (A300‐546, Bethyl laboratories), PCNA (PC10), and β‐actin (AC‐15) from Sigma‐Aldrich; GAPDH (ab9484), MCM2 (ab4461), histone H3 (ab1791), H2AX (ab11175), GFP (ab290), Nup214 (ab70497), Gle1 (ab96007), and lamin B1 from Abcam; γH2AX (2577), P‐Chk1Ser345 (2341), and P‐Chk2Thr68 (2668) from Cell Signaling Technologies; P‐ATMSer1981 (10 h11.E12), P‐p53Ser15, and 53BP1 (MAB3802) were from Millipore; Chk1 (sc‐8408) from Santa Cruz. RPA, ORC, MCM7, and MCM3 were previously described (Recolin et al, 2012). NXF1 antibody was a kind gift from E. Bertrand. The previously described S9.6 antibody was a kind gift from the laboratory of P. Pasero. The rabbit polyclonal senataxin antibody OY7 was previously described (Yuce & West, 2013).
Drugs and irradiation
Aphidicolin, mitomycin C, methyl methanesulfonate (MMS), doxorubicin, camptothecin, bleomycin, cycloheximide, and MG132 were purchased from Sigma‐Aldrich. Concentrations and incubation times are indicated in figure legends. In all experiments, UV‐C irradiation at 254 nm was performed with microprocessor‐controlled crosslinker (BIO‐LINK ®) at the indicated doses. The ATR inhibitor VE‐821 was purchased from TINIB tools. The ATM inhibitor KU5933 was purchased from Tocris Bioscience. Both inhibitors were used at 10 μM for the indicated times.
Transfections and RNAi treatments
Two siRNA duplexes targeting different reading frames of Ddx19 were retained from a SMART pool (Dharmacon) and used for RNAi‐mediated knockdown:
si_b (5′‐GCUCCAAGCUCAAGUUCAUdTdT‐3′)
si_d (5′‐GGACGGGAAUCCUGACAAUdTdT‐3′)
The Ddx19 si_b was used for subsequent knockdowns and referred to as siDdx19. Nxf1 and Gle1 downregulation was achieved using the following siRNAs, respectively: (5′‐UUGCUCUGAAUCAUGCUCAdTdT‐3′) and (5′‐CAGCGCGUGAAGCAAGCAGAAdTdT‐3′). Senataxin knockdown was achieved using a previously published siRNA sequence (Skourti‐Stathaki et al, 2011).
An siRNA against luciferase (Luc) was used as a control with the following sequence (5′‐CACGUACGCGGAAUACUUCGAdTdT‐3′).
The siRNA transfections were done using INTERFERin (Polyplus‐transfection) at 10 nM. Depletion was verified by preparing total cell lysates 48 and 72 h post‐RNAi treatment and analyzing by Western blots. For overexpression studies, cells were seeded 24 h before transfection and were transfected using jetPEI reagent (Polyplus‐transfection) with 3 μg of DNA. Cells were analyzed 24 h after transfection by immunofluorescence. For co‐transfection experiments, cells were first transfected with siRNA using INTERFERIN for 24 h, followed by transfection with the DNA constructs using jetPEI. Twenty‐four hours later, cells were harvested and analyzed by Western blot.
Immunofluorescence and microscopy
Cells were grown on coverslips overnight prior transfection, washed once with phosphate‐buffered saline (PBS), and fixed with ice‐cold methanol for 10 min. Samples were then washed three times with PBS containing 0.1% Tween‐20 (PBST) and blocked with 3% bovine serum albumin in PBST for 1 h at room temperature. Primary antibodies were diluted in the blocking buffer (1:200 for the S9.6 antibody and 1:500 for all other antibodies used in immunofluorescence) and incubated for 2 h at room temperature. After three 5‐min washes in PBST, samples were incubated with secondary antibodies for 1 h at room temperature. Coverslips were washed three times for 5 min each in PBST and mounted using ProLong Gold antifade reagent (Invitrogen) containing DAPI (4′,6‐diamidino‐2‐phenylindole). Immunofluorescence signals were analyzed with Leica DM6000 epifluorescence microscope (RIO imaging facility). Images were acquired using a Coolsnap HQ CCD camera (Photometrics) and the Metamorph software (Molecular Devices).
DRIP assay
DRIP was performed mainly as previously described (Ginno et al, 2012). In brief, HeLa cells transfected for 72 h with the indicated siRNAs were collected by trypsinization, washed with PBS, resuspended in 1.6 ml of TE pH 8.0, and treated overnight with 41.5 μl of 20% SDS and 5 μl of proteinase K (800 U/ml) at 37°C. The next day, DNA was gently extracted by phenol:chloroform:isoamylalcohol (25:24:1), and ethanol‐precipitated. Precipitated DNA was spooled onto a glass rod, washed two times with 70% ethanol, dissolved gently in TE, and digested overnight with a cocktail of restriction enzymes (EcoRI, HindIII, XbaI, SspI, BsrGI) at 37°C in presence of 1 mM spermidine. After digestion, DNA was extracted again with phenol/chloroform followed by ethanol precipitation. For the negative control, 8 μg of digested DNA was treated overnight with 3 μl of RNase H (M0297S, New England Biolabs) at 37°C. Four micrograms of the digested DNA (treated or non treated with RNase H1) was bound to 10 μl of S9.6 antibody (1 mg/ml) in 500 μl of binding buffer (10 mM NaPO4 pH 7.0, 140 mM NaCl, 0.05% Triton X‐100) in TE, overnight at 4°C on a rotating wheel. DNA‐antibody complexes were immunoprecipitated using Sepharose rProtein A (GE Healthcare) during 2 h at 4°C and washed three times with binding buffer. DNA was eluted with 50 mM Tris pH 8.0, 10 mM EDTA, 0.5% SDS, treated 45 min with 7 μl of proteinase K (800 U/ml) at 55°C and cleaned with phenol/chloroform. qPCR was performed to analyze enrichment of DNA:RNA hybrids at actively transcribed genes. Primer sequences used are shown below:
RPL13A Forward: AATGTGGCATTTCCTTCTCG
Reverse: CCAATTCGGCCAAGACTCTA
BTBD19 Forward: GGCTGCTCAGGAGAGCTAGA
Reverse: ACCAGACTGTGACCCCAAAG
EGR1 Forward: GCCAAGTCCTCCCTCTCTACTG
Reverse: GGAAGTGGGCAGAAAGGATTG
SNRPN (Negative) Forward: TGCCAGGAAGCCAAATGAGT
Reverse: TCCCTCTTGGCAACATCCA
qRT–PCR
Total RNA was isolated with TRIzol reagent (Invitrogen) according to the instructions of the manufacturer. Reverse transcription was carried out with random hexanucleotides (Sigma) and Superscript II First‐Strand cDNA synthesis kit (Invitrogen). Quantitative PCRs were performed using Lightcycler SYBR Green I Master mix (Roche) on Lightcycler apparatus (Roche) in 96‐well plate format. All primers used were intronspanning. The relative amount of target Ddx19 cDNA was obtained by normalization using geometric averaging of multiple internal control reference genes (HPRT, HMBS, GAPDH, and SDHA). Bars represent the mean standard deviation SD of multiple observations (n = 3). The sequences of the primer used were as follows: Ddx19 forward 5′ CGTCCATCCAAGATACAAGAGA 3′, reverse 5′ TTGGGCAATTAAGTTCTGTGG 3′; GAPDH forward 5′ CCTCCTCCTAAGATGGTGTCTG 3′, reverse 5′ GACCGATGCGTCCAAATC 3′; HPRT1 forward 5′ TGACCTTGATTTATTTTGCATACC 3′, reverse 5′ CGAGCAAGACGTTCAGTCCT 3′; SDHA forward 5′ TCCACTACATGACGGAGCAG 3′, reverse 5′ CCATCTTCAGTTCTGCTAAACG 3′; HMBS forward 5′ CTGAAAGGGCCTTCCTGAG 3′, reverse 5′ CAGACTCCTCCAGTCAGGTACA 3′.
FACS analysis
Cells were harvested, washed twice in PBS, and then fixed in ice‐cold 70% ethanol at −20°C overnight. Cells were then washed twice in PBS and incubated with 50 μg/ml RNase A at 37°C for 1 h. Denaturation was done with 2 N HCl for 30 min at room temperature, followed by incubation with 0.1 M borate for 2.5 min. Cells were then centrifuged, dissolved in PBS 0.5% NP‐40, 0.1% BSA, and incubated with primary anti‐BrdU (BD347580) for 1 h at room temperature followed by incubation with FITC‐coupled mouse secondary antibody. Staining with propidium iodide (P4864 Sigma) at 25 μg/ml was done for 10 min in the dark before analysis. Cells were analyzed with a FACSCalibur flow cytometer using CellQuestPro software.
Fiber stretching
HeLa cells were first pulse labeled for 10 min with 25 μM of the nucleotide analogue IdU and then washed twice with warm media. The second pulse labeling was done with 50 μM CIdU for 30 min under the conditions specified in figure legends. Cells were scraped into ice‐cold PBS and fibers were stretched on silanized coverslips after mixing with spreading buffer (200 mM Tris–HCl pH 5.0, 50 mM EDTA, 0.5% SDS). After fixation with methanol/acetic acid (3:1), denaturation with 2.5 M HCl for 1 h was done followed by blocking for 1 h with 1% BSA in PBST. The CldU and IdU were detected with the rat anti‐CIdU antibody C117‐7513 and anti‐BrdU (347580BD, Becton Dickinson), respectively. The whole fibers were marked with the single‐stranded DNA MAB3034 antibody from Millipore. DNA fibers were analyzed on a Leica DM6000B microscope equipped with a CoolSNAP HQ CCD camera (Roper Scientifics, Sarasota, Florida, USA) and a 40× objective. Image acquisition was performed with MetaMorph software. Representative images of DNA fibers were assembled from different fields of view and processed as described with ImageJ software.
Pulse‐field gel electrophoresis (PFGE)
For direct detection of cellular DNA breaks, 106 cells were collected after the indicated treatment and casted into agarose plugs before incubation with proteinase K buffer (0.1 M EDTA, 1% N‐laurylsarcosyl, 0.2% N‐deoxycholate, 1 mg/ml proteinase K) at 37°C for 48 h. Plugs were washed four times in wash buffer (20 mM Tris pH 8; 50 mM EDTA), loaded onto a 0.9% agarose gel and separated by PFGE Rotaphor 6.0® apparatus with the following parameters: 23 h; interval 30‐5 s log, angle 120‐110 linear; voltage 180‐120 V log, 13°C. DNA was visualized by subsequent staining with ethidium bromide.
ATPase and RNA:DNA unwinding assays
ATPase assay was performed as previously described (Schmitt et al, 1999). RNA:DNA unwinding assay was performed on RNA:DNA hybrids hybridized in vitro. The sequence of the top RNA strand was as follows: 5′‐AGCACCGUAAAGA‐3′, and the sequence of the bottom DNA strand was 5′‐TAAAACAAAACAAAACAAAACAAAATCTTTACGGTGCT‐3′ to obtain a hybrid with 5′ssDNA overhang. The top (short) RNA strand was labeled with γ‐32P‐ATP using polynucleotide kinase (New England Biolabs) according to standard protocols and cleaned up with Bio‐Spin® 6 columns (Biorad). Top RNA and bottom DNA strand oligos were mixed at equimolar ratios and annealed by initial incubation at 95°C for 5 min, then slow cooling from 95 to 25°C in 10 mM MgSO4, 10 mM Tris pH 7.5, and aliquots were stored at −20°C.
Unwinding reactions were carried out at 30°C for 20 min in a mixture (20 μl) containing 30 mM Tris–HCl (pH 7.5), 50 mM NaCl, 5 mM MgCl2, 2 mM DTT, 1 unit/μl RNaseOUT (Invitrogen), 0.01% NP‐40, 0.1 mg/ml BSA, 4 mM ATP, 0.5 nM 32P‐labeled RNA:DNA substrate, and 600 nM of recombinant proteins. The reaction was stopped with a buffer containing 6% Ficoll, 13 mM EDTA pH 8.0, 10% SDS and proteinase K (800 U/ml), 0.05% bromophenol blue, and 0.05% xylene cyanol. Reactions were analyzed by 15% non‐denaturing PAGE. Gels were exposed to a PhosphoImager screen overnight.
In vitro R‐loop formation and unwinding
The pFC53 plasmid containing the mouse Airn CpG island under control of the phage T3 promoter was previously described (Powell et al, 2013). This plasmid was transcribed in vitro with T3 RNA polymerase (Promega) for 30 min at 37°C and heat inactivated at 65°C for 10 min. Samples were then treated with 0.5 μl of 1 mg/ml RNase A only, or in addition to either 3 μl RNase H (New England Biolabs), Ddx19WT, or Ddx19E243Q‐DEAD mutant for 30 min at 37°C. Proteinase K was added and incubated for an additional 30 min at 37°C. Samples were then loaded onto 0.9% agarose gel and post‐stained with ethidium bromide. R‐loop formation upon in vitro transcription was detected as an RNase H1‐sensitive mobility shift of the plasmid from the supercoiled state to the relaxed state.
Immunoprecipitation by GFP‐Trap
HeLa cells were transfected with either GFP or GFP‐tagged Ddx19 (WT or S93E), and treated or not with 250 nM of camptothecin overnight. Cell were lysed in ice‐cold lysis buffer (50 mM Tris–HCl, pH 7.4, 100 mM NaCl, 50 mM NaF, 5 mM EDTA, 40 mM β‐glycero‐phosphate, 1% Triton X‐100 and protease inhibitors). Lysates were cleared by centrifugation and incubated with GFP‐Trap®_A beads (Chromotek) equilibrated in lysis buffer on a rotating wheel for 2 h at 4°C. Beads were washed three times with ice‐cold lysis buffer, and bound proteins were eluted in 2× Laemmli buffer and boiling prior to SDS–PAGE.
Additional/extended Materials and Methods information is available in the Appendix.
Author contributions
DH performed all the experiments in mammalian cells. BR performed the screen in Xenopus and isolated XDdx19. NT helped in the isolation of XDdx19. SM analyzed XDdx19 nuclear translocation in Xenopus. KS performed experiments described in Fig EV4D and reproduced experiments shown in Fig 7A. RAM contributed to results shown in Figs 2 and EV2, and provided financial support to DH. DM designed the experiments and wrote the manuscript.
Conflict of interest
The authors declare that they have no conflict of interest.
Expanded View
Expanded View Figures PDF [embj201695131-sup-0002-EVFigs.pdf]
Movie EV1 [embj201695131-sup-0003-MovieEV1.zip]
Movie EV2 [embj201695131-sup-0004-MovieEV2.zip]
Movie EV3 [embj201695131-sup-0005-MovieEV3.zip]
Movie EV4 [embj201695131-sup-0006-MovieEV4.zip]
Movie EV5 [embj201695131-sup-0007-MovieEV5.zip]
Movie EV6 [embj201695131-sup-0008-MovieEV6.zip]
Movie EV7 [embj201695131-sup-0009-MovieEV7.zip]
Movie EV8 [embj201695131-sup-0010-MovieEV8.zip]
Movie EV9 [embj201695131-sup-0011-MovieEV9.zip]
Movie EV10 [embj201695131-sup-0012-MovieEV10.zip]
Movie EV11 [embj201695131-sup-0013-MovieEV11.zip]
Movie EV12 [embj201695131-sup-0014-MovieEV12.zip]
Movie EV13 [embj201695131-sup-0015-MovieEV13.zip]
Movie EV14 [embj201695131-sup-0016-MovieEV14.zip]
Movie EV15 [embj201695131-sup-0017-MovieEV15.zip]
Acknowledgements
We thank E. Izaurralde, S. West, O. Becherel, P. Pasero, E. Bertrand, F. Chédin, and M. Malumbres for reagents. We thank J. Basbous, M. Larroque, C. Ribeyre, and C. Kermi for technical advices. We also thank E. Bertrand, E. Schwob, and C. Dargemont for useful discussion and J. Hutchins for critical reading of the manuscript. This work was supported by grants from “Fondation ARC”, FRM, INSERM, Ligue contre le Cancer and Campus France (PHC‐CEDRE) in D.M. laboratory. D.M. is supported by INSERM. A 3‐year scholarship to D.H. was funded from Lebanese University‐NCSRL.
Funding
Fondation ARC pour la Recherche sur le Cancer (ARC)http://dx.doi.org/10.13039/501100004097 3615References
- © 2017 The Authors
