Actinomycin D Induces Histone (cid:1) -H2AX Foci and Complex Formation of (cid:1) -H2AX with Ku70 and Nuclear DNA Helicase II*

Formation of (cid:1) -H2AX foci is a P. O.cellular response to genotoxic stress, such as DNA double strand breaks or stalled replication forks. Here we show that (cid:1) -H2AX foci were also formed when cells were incubated with 0.5 (cid:2) g/ml DNA intercalating agent actinomycin D. In untreated cells, (cid:1) -H2AX co-immunoprecipitated with Ku70, a subunit of DNA-dependent protein kinase, as well as with nuclear DNA helicase II (NDH II), a DEXH family helicase also known as RNA helicase A or DHX9. This association was increased manifold after actinomycin D treatment. DNA degradation diminished the amount of Ku70 associated with (cid:1) -H2AX but not that of NDH II. In vitro binding studies with recombinant NDH II and H2AX phosphorylated by DNA-dependent protein kinase confirmed a direct physical interaction between NDH II and (cid:1) -H2AX. Thereby, the NDH II DEXH domain alone, i.e. its catalytic core, was able to support binding to (cid:1) -H2AX. Congruently, after actinomycin D treatment, NDH II accumulated in RNA-containing nuclear bodies that predominantly co-localized with (cid:1) -H2AX foci. Taken together, these results suggest that histone (cid:1)

The histone H2A family contains three members named H2A1-H2A2, H2AZ, and H2AX (for reviews, see Refs. 1 and 2). In mammals, 75-98% of H2A is represented by H2A1 and H2A2, whereas only 2-25% of all nucleosomes contain H2AX. Although the situation is different in yeast, where the two H2AX orthologs HTA1 and HTA2 account for 95% of the H2A complement, in both, H2AX plays an important role in DNA repair, particularly after inducing DNA double strand breaks (DSBs) 1 (3). H2AX is highly conserved in eukaryotes and contains a distinguishing C-terminal extension carrying a SQ(E/ D)(I/L/F/Y) consensus motif. This motif is recognized and phosphorylated by members of the family of phosphatidylinositol 3-phosphate kinase-related protein kinases, ataxia telangiectasia mutated (ATM), ATM and Rad 3-related (ATR), and DNAdependent protein kinase (DNA-PK) (1,2). Within the SQ(E/ D)(I/L/F/Y) phosphorylation motif of H2AX, serine 139, the fourth amino acid from the C-terminal end, is preferably sub-jected to phosphorylation. Acknowledging the physiological importance of this phosphorylation, the phosphorylated H2AX is designated ␥-H2AX. The level of ␥-H2AX is normally low in mammals, unless a DSB is introduced, e.g. by ionizing radiation (4), caspase-activated DNases (5), meiotic DNA recombination (6), or V(D)J recombination during lymphocyte development (7).
Although ATM, ATR, and DNA-PK are all able to phosphorylate H2AX, the three kinases seem to signal in different cellular responses. In this respect, ATM has been found to be responsible for phosphorylation of H2AX in response to DSBs caused by ionizing irradiation (15), whereas ATR seems to be required for phosphorylation of H2AX on a stalled replication fork, induced by e.g. hydroxyurea or UV irradiation (16). Further results suggest that ATR may be activated by singlestranded DNA coated with replication protein A. The presence of a single strand of DNA of sufficient length seems to be a prerequisite for the recruitment and activation of ATR (17), which leads to phosphorylation of Chk1 in a checkpoint pathway that depends on Rad17 and the proliferating cell nuclear antigen-like Rad9-Hus1-Rad1 complex (18).
Actinomycin D is an anti-tumor drug that particularly intercalates into transcriptionally active regions of chromosomal DNA and thereby abrogates RNA synthesis (19). Using actinomycin D at a concentration to inhibit transcription, i.e. 0.5 g/ml, we observed induction of ␥-H2AX foci formation. The presence of ␥-H2AX in transcriptionally active chromatin was substantiated by an association with RNA polymerase II, as demonstrated by co-immunoprecipitation. Nevertheless, whereas a prolonged actinomycin D treatment led to degradation of RNA polymerase II, it also enhanced the association of the newly formed ␥-H2AX to Ku70 and to NDH II. We found that DNA mediated Ku70 association to ␥-H2AX, whereas NDH II was directly bound to ␥-H2AX via protein-protein interactions. The presence of NDH II in these complexes is not surprising, as NDH II has been shown to be a substrate of DNA-PK (20) and is hypothesized to play a functional role in transcription. This is inferred by analogy from its Drosophila homolog MLE, which plays an essential role in X chromosome dosage compensation (21)(22)(23). If low doses of actinomycin D would solely distort DNA structures at transcriptionally active sites, there may exist a cellular response to disarranged transcriptionally active DNA. This response may invoke DNA-PK phosphorylation of H2AX and may promote recruitment of NDH II to ␥-H2AX foci.

EXPERIMENTAL PROCEDURES
Antibodies-A mouse monoclonal antibody against ␥-H2AX (JBW301) was purchased from Upstate Biotechnology (Charlottesville, Virginia). The mouse monoclonal antibody against the catalytic subunit of DNA-PK (DNA-PKcs)  was from NeoMarkers (Fremont, CA). A goat polyclonal antibody against ATR (N-19), a rabbit polyclonal antibody against the N terminus of the large subunit of RNA polymerase II (N-20), and mouse monoclonal antibodies against Ku70 (A-9) or p53 (D0 -1) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). A mouse monoclonal antibody (8WG16) against RNA polymerase II was from Covance (Princeton, NJ). Rabbit antiserum against NDH II has been described previously (22).
Proteins-Recombinant human H2AX was obtained from Upstate Biotechnology (Charlottesville, Virginia). DNA-PK purified from human cells was purchased from Promega (Madison, WI). Full-length human NDH II (amino acids 1-1269) and its deletion versions comprising amino acids 1-952, 313-1269, and 313-952 were generated in our laboratory and have been described previously (24). All of these NDH II recombinant proteins carried His 6 tags at their N termini and were expressed from recombinant baculoviruses in Sf9 insect cells. glutathione S-transferase fusion proteins containing the N terminus (amino acids 1-312) and C terminus (amino acids 953-1269) of NDH II were as described previously (24).
Cell Culture-Human HeLa and HEp2 cells were maintained in Dulbecco's modified Eagle's medium (C.C.Pro GmbH, Neustadt/Weinstrasse, Germany) supplemented with 10% fetal bovine serum (Invitrogen). Human MCF-7 cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum. All cell lines were cultured in an incubator with 10% CO 2 at 37°C. Actinomycin D, when applied, was added to the cell medium to a concentration of 0.5 g/ml, followed by further incubation for 3.5 h, if not indicated otherwise.
Immunofluorescence-Cells attached to a coverslip were washed one time with cold PBS (10 mM sodium phosphate, pH 7.4, 140 mM NaCl, and 3 mM KCl) and then treated with Nonidet P-40 (0.5% in PBS) for 15 min on ice, followed by fixation with 4% formaldehyde in PBS for 15 min. Alternatively, the cells were directly fixed without pretreatment by Nonidet P-40. The fixed cells were permeabilized with 0.5% Triton X-100 in PBS for 15 min and then blocked with 5% bovine serum albumin (fraction V, Sigma) in PBS for 30 min at room temperature. Subsequent steps were as described previously (25). The mouse monoclonal antibody against ␥-H2AX (JBW301) was diluted 1:100 and the rabbit antiserum against NDH II, 1:1000. Immunofluorescence was observed at a 100-fold magnification with a Zeiss Axiovert 135 microscope in connection with a Sony charge-coupled device color video camera. For confocal microscopy, we used a Zeiss LSM 510 laser scanning confocal microscope at 63ϫ magnification. All microscopic data were processed by Adobe Photoshop and Canvas.
Western Blot-After electrophoresis through an SDS-polyacrylamide gel, the proteins were transferred to a Hybond-C extra nitrocellulose membrane (Amersham Biosciences) using a semidry electroblotter. For immunodetection, a dilution of 1:1000 was used for the mouse monoclonal antibodies against ␥-H2AX (JBW301) and Ku70 (A-9) or the rabbit polyclonal antibodies against RNA polymerase II (N-20) and nuclear DNA helicase II. A dilution of 1:500 was used for the mouse monoclonal antibody against RNA polymerase II (8WG16) and 1:200 for the mouse monoclonal antibodies against DNA-PKcs (25-4) and p53 (DO-1), or the goat polyclonal antibody against ATR (N-19). Secondary antibodies coupled to avidin or a streptavidin-horseradish peroxidase complex, both diluted at 1:5000, were applied for enhanced luminescence immunodetection (Amersham Biosciences).
Immunoprecipitation-HeLa cells, untreated or treated with actinomycin D, were pelleted by centrifugation at 320 ϫ g for 10 min. After washing with PBS and recentrifugation, the harvested cells (ϳ0.5 ml/ pellet) were disrupted by four passages through a 25-gauge needle in 1 ml of 20 mM Tris/HCl, pH 7.2, 15 mM KCl, 2.5 mM MgCl 2 , 0.05% Nonidet P-40, and the proteinase inhibitors aprotinin, leupeptin, and pepstatin, each at 5 g/ml. The cell suspensions were then centrifuged at 3000 ϫ g for 5 min at 4°C. After removal of the supernatant (i.e. cytosol), the nuclear pellets were resuspended in 1 ml of 50 mM Tris/HCl, pH 7.8, 150 mM KCl, 5 mM MgCl 2 , 250 mM sucrose, and the proteinase inhibitors, followed by six strokes of sonication with a Branson sonicator, each at an output setting of 20 for 4 s. The nuclear extracts were then placed on top of a cushion containing 0.88 M sucrose, prepared in the same buffer as above for resuspending the nuclei, followed by centrifugation at 13,000 ϫ g for 15 min. Supernatants containing the nucleoplasmic components were then divided into two equal parts. The first half was mixed with the mouse antibody against ␥-H2AX (JBW301, ϳ15 g) or the rabbit antiserum against NDH II (10 l), and the second half incubated to the same amount of mouse or rabbit control IgG, respectively. The samples were incubated under rotation for 1 h at 4°C. A 50-l bed volume of protein A-agarose (Oncogene, Darmstadt, Germany) was added to each mixture, and the incubation was continued for another hour at 4°C. Finally, the agarose beads were collected by centrifugation at 10,000 ϫ g for 1 min and washed three times each with a 10-fold bed volume of IP-40 buffer (50 mM Tris/HCl, pH 8.0, 25 mM NaCl, 0.1 mM EDTA, 0.2% Nonidet P-40 plus proteinase inhibitors, as described above). After removal of the remaining supernatant, the agarose beads were mixed each with an equal volume (50 l) of SDS-PAGE loading buffer and heated for 5 min at 95°C. Co-immunoprecipitated proteins were analyzed by SDS-PAGE and Western blotting.
In Vitro Phosphorylation by DNA-PK-The indicated amount of H2AX was phosphorylated by DNA-PK (25 units/l) in a reaction buffer consisting of 50 mM HEPES, pH 7.5, 100 mM KCl, 10 mM MgCl 2 , 0.2 mM EGTA, 0.1 mM EDTA, 1 mM dithiothreitol, 0.2 mM ATP, 0.07 g/l activated calf thymus DNA (Sigma) and 80 g/ml bovine serum albumin. The reaction mixture was incubated for 15 min at 30°C. H2AX phosphorylated at serine 139 was distinguished from nonphosphorylated H2AX by Western blotting with the mouse antibody against ␥-H2AX (JBW301).
Phosphorylation of NDH II, as well as its truncated mutants, by DNA-PK was performed as described above but with the addition of 2 Ci of [␥-32 P]ATP. After SDS-PAGE, the radioactive phosphorylation signals were detected by autoradiography.
In Vitro Studies of Protein-Protein Interactions-1 g of H2AX was first phosphorylated by DNA-PK as described above in a volume of 50 l. After phosphorylation, 300 units/ml of micrococcal nuclease (MNase, Amersham Biosciences) were added to the reaction mixture, which was also supplemented with 1 mM CaCl 2 . Incubation for 15 min at room temperature sufficed to degrade DNA. The mixture containing ␥-H2AX was then mixed with recombinant NDH II (0.5 g) and incubated for another 15 min at room temperature. An equal volume of this protein solution was mixed with the mouse antibody against ␥-H2AX (JBW301, 3.5 g) and as control, the same amount of mouse IgG. After preincubation for 1 h at 4°C, protein A-agarose (25-l bed volume) was added, and incubation was continued for another hour at 4°C. Immunoprecipitated proteins were harvested from the washed agarose beads with SDS-PAGE loading buffer, as described above for immunoprecipitation from nuclear extracts. The proteins, interacting with ␥-H2AX, were examined by Western blotting. To probe the interaction with H2AX or ␥-H2AX, NDH II was translated in vitro from its full-length cDNA subcloned into a pBluescript plasmid (24) using the TNT T7coupled reticulocyte lysate system (Promega, Mannheim, Germany) and [ 35 S]methionine for the radioactive labeling of NDH II.

Actinomycin D Induced the Formation of Nuclear ␥-H2AX
Foci-Actinomycin D inhibits transcription by preferentially intercalating into GpC-rich double-stranded DNA and, as a consequence, abrogates the elongation phase of RNA synthesis in interfering with the extension of the transcription bubble (19). Surprisingly, we found that 0.5 g/ml of this drug led to the appearance of nuclear ␥-H2AX foci that so far have been taken as an indicator for the presence of DSBs (1, 2) ( Fig. 1, A-D). To observe chromatin-bound proteins, we first extracted the cells with 0.5% Nonidet P-40 and then fixed the chromatinbound fraction with formaldehyde (26). After pre-extraction, ␥-H2AX foci were also found in untreated cells, although at a very weak intensity (Fig. 1, E and F). The ␥-H2AX fraction of normal nuclei was widely distributed as small dots, whereas the nuclear foci of actinomycin D-treated cells appeared to be larger and more intensely stained (Fig. 1, G and H).
To address the possibility that ␥-H2AX foci, present in cells treated with 0.5 g/ml actinomycin D, were caused by DSBs, we extracted DNA from these cells and examined the integrity of chromosomal DNA by agarose gel electrophoresis. The results showed that some DNA strand breaks were also induced by treatment with 0.5 g/ml actinomycin D. This was confirmed by DNA end labeling with terminal deoxynucleotidyl transferase and a fluorescein-labeled deoxynucleotide (Fig.  2B). On the other hand, degradation of chromosomal DNA was much more striking when the concentration of actinomycin D was raised to 10 g/ml ( Fig. 2A).
Despite this, after treatment with low doses of actinomycin D, a timely increase of ␥-H2AX could be demonstrated by Western blotting of whole cell lysates (Fig. 2C). This was accompanied by a rapid degradation of RNA polymerase II and a subsequent induction of the tumor suppressor protein p53 (Fig. 2C).
H2AX Was Associated with NDH II and Ku70 -To look for proteins associated with ␥-H2AX induced by actinomycin D, nuclear extracts of actinomycin-treated or untreated cells were prepared for immunoprecipitation with a mouse monoclonal antibody against ␥-H2AX (JBW301). The co-immunoprecipitated proteins were examined by Western blotting. Under this condition, RNA polymerase II, NDH II, and Ku70 were found to co-precipitate with ␥-H2AX (Fig. 3A). Although ␥-H2AX was not detected in nuclear extracts of untreated cells (Fig. 3A, input fraction), it became visible in the corresponding immunoprecipitates (Fig. 3A, ␥-H2AX immunoprecipitated fraction). This is another indication that ␥-H2AX, indeed, was present in untreated cells, although at a very low level. In contrast, in actinomycin D-treated cells, ␥-H2AX was easily detectable in nuclear extracts and could be further enriched by immunoprecipitation (Fig. 3A, second panel from top). The amount of NDH II and Ku70 co-immunoprecipitating with ␥-H2AX was substantially higher in actinomycin D-treated cells than in control cells. Unexpectedly, RNA polymerase II only seemed to coimmunoprecipitate with ␥-H2AX in untreated cells. We suggest that protein degradation decreased the amount of RNA polymerase II and made it undetectable in the ␥-H2AX immunoprecipitates. The potential degradation of RNA polymerase II oc-curred earlier, i.e. 30 min, after actinomycin D treatment. From this time point on, a diminishing amount of RNA polymerase II in association with ␥-H2AX was observed (Fig. 3B).
NDH II Directly Bound to ␥-H2AX, whereas Association of Ku70 with ␥-H2AX Was Mediated by Chromosomal DNA-Coimmunoprecipitation of ␥-H2AX, Ku70, and NDH II led us to examine whether these proteins interact directly with each other or whether chromosomal DNA mediates an indirect interaction. To distinguish between these two alternatives, we immunoprecipitated ␥-H2AX from actinomycin D-treated cell extracts and treated the washed precipitates with MNase. The results showed that once DNA was degraded, the presence of the otherwise co-immunoprecipitated Ku70 antigen decreased to an undetectable level, whereas the same treatment hardly affected the amount of NDH II associated with ␥-H2AX (Fig.  4A). These findings strongly suggest that NDH II interacted with ␥-H2AX by forming a direct protein-protein contact, whereas the interaction between Ku70 and ␥-H2AX was mediated by DNA. Here, the association of ␥-H2AX with Ku70, which is disrupted by degrading the DNA, was also observed in immunoprecipitates with a mouse antibody against Ku70 (Fig. 4B).
To demonstrate binding of ␥-H2AX to NDH II in vitro, recombinant H2AX was phosphorylated by purified DNA-PK holoenzyme. An antibody against ␥-H2AX revealed a signal on Western blots only when both DNA and ATP were added to the reaction mixture (Fig. 5A). This result corroborated the essential need of both ATP and DNA for the reaction and underscored the specificity of the ␥-H2AX antibody used for detecting the phosphorylated (but not the unphosphorylated) form of H2AX. To examine the binding of NDH II to ␥-H2AX, the already described phosphorylation mixture was complemented with recombinant full-length NDH II after DNA degradation with MNase, followed by immunoprecipitation of ␥-H2AX. Fig.  5B evidences that NDH II appears in the immunoprecipitate of ␥-H2AX but not in the immunoprecipitates of unphosphorylated H2AX or the mouse Ig control. Again, Ku70 was not detectable in the immunoprecipitate of ␥-H2AX. This confirms the in vivo immunoprecipitation experiment as presented in Fig. 4 and supports the conclusion that Ku70 does not directly interact with ␥-H2AX or NDH II.
To further distinguish whether NDH II prefers binding to either the phosphorylated or the unphosphorylated form of H2AX, this histone variant was incubated with DNA-PK in the absence or presence of ATP and spotted onto a nitrocellulose membrane. The membrane was then probed with in vitrotranslated and [ 35 S]methionine-labeled NDH II. This experiment confirmed that NDH II only bound to ␥-H2AX, i.e. H2AX phosphorylated by DNA-PK, and not to unphosphorylated H2AX (Fig. 5C).
Aiming at identifying a region of NDH II responsible for binding to ␥-H2AX, several truncated forms of NDH II were generated and probed for an interaction with ␥-H2AX. These NDH II variants contained either the N-terminal amino acids 1-952, with a deletion of the C terminus carrying the RGG box, or the amino acids 313-1269 with a deletion of the N terminus, including the two double-stranded RNA binding domains or the amino acids 313-952 in the middle, bearing the helicase catalytic motif. All truncated constructs of NDH II displayed binding to ␥-H2AX (Fig. 5D). On the other hand, pull-down experiments with glutathione S-transferase fusion proteins containing only the N-terminal amino acids 1-313 or the Cterminal amino acids 953-1269 of NDH II had no affinity to ␥-H2AX (Fig. 5E). These results led to the conclusion that the central part of NDH II, i.e. the region with the helicase signature motifs, was responsible for binding to ␥-H2AX.

␥-H2AX Suppressed the DNA-PK-mediated Phosphorylation of Full-length NDH II and That of Its DEXH Catalytic Core-
Recently, we have shown that DNA-PK phosphorylated NDH II in vitro with either DNA or RNA as cofactor (20). Therefore, we assumed that chromosome-associated complexes consisting of DNA-PK and ␥-H2AX are also capable of phosphorylating NDH II. Co-immunoprecipitation experiments with an antibody against NDH II revealed that the helicase co-immunoprecipitated with DNA-PKcs and ␥-H2AX from nuclear extracts of actinomycin D-treated cells (Fig. 6A). However, these interactions were not strong enough to maintain a stable complex that could be detected by Sephacryl S-200 gel filtration (Fig. 6B). Nevertheless, NDH II was phosphorylated by DNA-PK in a time-dependent manner when ATP and DNA were present (Fig. 6C, panel a). When this experiment was repeated in the presence of ␥-H2AX, the phosphorylation of NDH II was inhibited (Fig. 6D, panel a). We further found that phosphorylation of NDH II by DNA-PK occurred within the catalytic DEXH domain, whereas the N-terminal part of NDH II was not phos-phorylated (Fig. 6C, panel b). Similar to full-length NDH II, a phosphorylation of the DEXH core by DNA-PK was also inhibited by ␥-H2AX (Fig. 6D, panel b). We further realized that the DNA-PK-dependent phosphorylation of NDH II led to a stimulation of the DNA-dependent ATPase activity of NDH II, which was suppressed when H2AX was present in the reaction mixture (Fig. 6E). Thus, the activity-stimulating phosphorylation of NDH II by DNA-PK could be counteracted by histone ␥-H2AX.
After Actinomycin D Treatment, NDH II Accumulated in RNA-containing Nuclear Bodies That Were Partially Adjacent to the ␥-H2AX Foci-Because actinomycin D is capable of in-  5. ␥-H2AX interacted directly with NDH II in vitro. A, phosphorylation of H2AX by DNA-PK in vitro. Recombinant H2AX was incubated with DNA-PK under the conditions indicated. The H2AX input was visualized as silver-stained protein after SDS-PAGE. The phosphorylated H2AX (␥-H2AX) was detected with the monoclonal antibody JBW301. B, direct interaction of ␥-H2AX with NDH II. One g of ducing nuclear ␥-H2AX foci and NDH II displays direct interactions with ␥-H2AX, it was of interest to study whether NDH II co-localizes with ␥-H2AX in vivo. To answer this question, immunofluorescence studies were performed as described above by pre-treating cells with 0.5% Nonidet P-40. In untreated cells, NDH II displayed a nuclear distribution similar to that of ␥-H2AX, although the signals of the latter were comparatively weak (Fig. 7, A-C). Strikingly, actinomycin D treatment led to an accumulation of NDH II into nuclear bodies that, in part, co-localized with the induced ␥-H2AX foci (Fig. 7,  D--F). These NDH II-containing nuclear bodies were apparently sustained by RNA, as RNase treatment caused their fragmentation into smaller particles that became distributed over the entire nucleus (Fig. 7, G-I). Possibly, these NDH II-containing structures arose as a stress response to actinomycin D, which in turn led to the termination of transcription and probably to an increased accumulation of NDH II into ribonucleoprotein complexes that contained aberrant RNA products (27). On the other hand, RNase digestion of actinomycin D-treated cells alone did not induce the co-localization of NDH II with ␥-H2AX (Fig. 7, J-L).

DISCUSSION
␥-H2AX knock-out mice display DNA repair deficiencies that prevent survival after irradiation and provoke an increased frequency of chromosomal aberrations (28,29). Moreover, these mice display deficiencies in meiosis and immunoglobulin classswitch recombination (29). Even the deletion of a single H2AX allele in mice led to an increased susceptibility to developing tumors, which synergistically requires the absence of the tumor suppressor protein p53 (30,31). Recently, ␥-H2AX has been proposed to be a useful indicator in predicting the effect of radiotherapy for cancer treatment, because the persistence of ␥-H2AX foci indicates sensitivity of cancer cells to ionizing irradiation (32).
Actinomycin D is a cyclic pentapeptide that forms a stable complex with DNA by intercalating its phenoxazone chromophore between GpC base pairs (19). Previously, treatment with actinomycin D has been shown to induce a p53 response that eventually induces apoptotic cell death (33). This effect of actinomycin D at 0.5 g/ml is usually considered to be unrelated to DNA damage and to be caused by the inhibition of elongating RNA polymerase(s) (34). When an RNA polymerase progresses through a transcription bubble, the generated topological stress must be released by the DNA nicking and closing activities of a DNA topoisomerase (35,36). It should be noted that even low concentrations of actinomycin D were shown to prevent the religation step of topoisomerase I. This in itself might provoke the production of single-stranded DNA breaks, particularly at transcriptionally active sites (37), or the production of DSBs, when a replication fork collides with the covalently bound topoisomerase (38). Here, we describe that a low concentration (0.5 g/ml or 0.4 M) of actinomycin D induced the formation of nuclear ␥-H2AX foci. At present, we cannot clearly state whether focus formation might be due to the induction of DSB or whether the halt of the elongating polymerase per se is sufficient for this effect. The little DNA degradation observed after treatment with low doses of actinomycin D seems to support the latter view.
As observed earlier (8), ␥-H2AX was also detected to a low level in untreated cells. A basal level of ␥-H2AX can apparently be caused by many physiological processes and furthermore seems to depend on cell cycle phases. Additionally, in view of our data, we propose that the presence of ␥-H2AX may represent a necessary precaution in supervising the transcriptionassociated opening of DNA strands and minimizing the genotoxic effect of a possible DSB. The functional involvement of ␥-H2AX in the transcriptional process is corroborated by our result that RNA polymerase II co-immunoprecipitates with ␥-H2AX in untreated cells.
Tentative unpublished results hint at a role of DNA-PK for phosphorylating H2AX in vivo after treatment with actinomycin D. If these findings can be further corroborated, they may reveal a novel checkpoint, invoking DNA-PK. In contrast, ATM and ATR signal DNA damage in response to ionizing irradiation and the halt of replication forks, respectively (1, 2). DNA-PK consists of three components: the two heterodimerforming subunits Ku70 and Ku86 and the catalytic subunit DNA-PKcs (39). Although DNA-PKcs, itself, has some weak affinity for DNA, it requires Ku antigen for stable DNA binding and the promotion of its kinase activity (40 -42). This property of DNA-PK distinguishes it from ATM or ATR, because the latter two are able to directly bind to DNA (43). Despite this, ATM requires the MRN complex, which consists of Mre11, Rad50, and Nbs1, for efficient activity (44). This led to the suggestion that the MRN complex, rather than ATM, acts as the primary sensor of DSBs and that DSBs direct the activity of ATM to its substrates, such as Chk2, p53, and H2AX (45,46). In a similar manner, ATR requires the single-strand DNA binding protein RPA for specific association to a stalled replication fork (47), as well as the ATR-interacting protein ATRIP for the activation of the kinase activity (48). Therefore, the requirement of loading factors seems to be a common principle among ATM, ATR, and DNA-PKcs and most likely delivers an explanation for differential activation of these kinases, depending on the activated checkpoint. Besides its role in nonhomologous end-joining repair, Ku70/86 has been suggested to be involved in transcription (40,41). It is therefore conceivable that Ku70/86 can be responsible for recruiting DNA-PKcs to transcriptionally active sites or to sites of stalled transcription bubbles. This, in turn, may trigger a checkpoint, monitoring the genomic integrity. Such a scenario is supported by the finding that ␥-H2AX forms a DNA-mediated complex with Ku70, even in untreated cells, and that this type of complex formation is enhanced after the addition of actinomycin D.
The correlation of ␥-H2AX formation with transcriptional activities is further corroborated by the direct interaction obrecombinant H2AX was phosphorylated by DNA-PK as indicated, and the corresponding reaction mixtures were subjected to MNase digestion (see Fig. 4). Thereafter, 0.5 g of full-length recombinant NDH II was added. After incubation for 15 min at room temperature, the mixtures were immunoprecipitated with the mouse antibody against ␥-H2AX (JBW301) or an equal amount of mouse Ig as a control. Immunoprecipitates were examined by Western blotting for the presence of ␥-H2AX, NDH II, and Ku70. H2AX inputs were shown by silver staining after SDS-PAGE. C, specific interaction of NDH II with phosphorylated (but not with unphosphorylated) H2AX. The indicated amounts of H2AX were incubated with DNA-PK without ATP (Ϫ) or with ATP (ϩ) in the phosphorylation mixture. Thereafter, the samples were treated with MNase (see Fig. 5) and dot-blotted onto a nitrocellulose membrane. The membrane was incubated with [ 35 S]methionine-labeled full-length NDH II in phosphorylation buffer containing 0.5% blocking reagent for 1 h at room temperature. The membrane was washed three times with TBS (25 mM Tris, 140 mM NaCl, and 3 mM KCl) for 10 min each and exposed to an x-ray film to observe the radioactive signal. D, interaction of ␥-H2AX with the catalytic core of NDH II. Recombinant proteins comprising the indicated amino acids of NDH II were utilized to examine a physical interaction with ␥-H2AX as described above for B. Rabbit antiserum against NDH II was utilized for detecting its deletion products. E, pull-down of ␥-H2AX (1-g input, see B) with NDH II glutathione S-transferase fusion proteins as indicated (ϳ1 g of each). After washing the glutathione-Sepharose with buffer, as used for immunoprecipitation, the samples were lysed in SDS-PAGE loading buffer and analyzed by Western blotting for the presence of ␥-H2AX.
FIG. 6. Phosphorylation of NDH II by DNA-PK was inhibited by ␥-H2AX. A, NDH II co-immunoprecipitated with DNA-PKcs and ␥-H2AX from nuclear extracts of HeLa cells. After treatment of the cells with 0.5 g/ml actinomycin D for 3.5 h, immunoprecipitation was performed with 10 l of rabbit antiserum against NDH II and a rabbit control serum, respectively. On the Western blots, DNA-PKcs was detected by the mouse monoclonal antibody 25-4. Other antibodies were as described in the legend to Fig. 3. B, NDH II did not interact with DNA-PKcs in vitro.
served here between ␥-H2AX and NDH II. NDH II is a DNAand RNA-dependent helicase that so far has been identified in mammals, Caenorhabditis elegans, and Drosophila (25). The Drosophila NDH II homolog, MLE (maleless), has a well established role in X chromosome dosage compensation, i.e. in increasing the transcription level of the single male to that of the two female X chromosomes (23). On the male X chromosome, MLE has been found associated with many other proteins, including Msl-1, -2, -3, Mof (males absent on the first, a histone acetyltransferase activity responsible for the acetylation of lysine 16 of histone H4), and two noncoding chromosomal RNAs termed Rox1 and -2 (for reviews, see Refs. 49 -51). Moreover, a tandem histone kinase named JIL-1 is involved in the phosphorylation of histone H3 on the male X chromosome (52). Interestingly, all these components co-localize at thousands of discrete sites on the male X chromosome that then seem to assemble into global complexes dependent on the presence of RNA (53). In mammals, NDH II is essential for early embryonic development (54) and physically interacts with RNA polymerase II, the histone acetylase CBP/p300 (55), BRCA1 (56), and several other factors that are necessary for transcriptional activation (57)(58)(59).
Our in vitro results suggest that ␥-H2AX may be capable of distracting NDH II from transcriptionally active chromatin. This may be necessary to slow down RNA synthesis by interfering with co-transcriptional RNA processing on singlestranded RNA in case of genotoxic stress. After removal of the transcription block, the sequestered NDH II may be added back to the transcribed DNA and/or the resulting RNA products. The displacement observed here of NDH II from ␥-H2AX by doublestranded RNA supports the latter view. It is equally noteworthy that the DNA-PK-dependent phosphorylation of NDH II and the ATPase activity stimulated thereby were both suppressed by ␥-H2AX. This suggests a role of ␥-H2AX in physical Interaction between NDH II and DNA-PKcs was examined by gel filtration chromatography on a Sephacryl S-200 column (1 ϫ 26 cm) in a buffer of 20 mM Tris/HCl, pH 8.0, 150 mM KCl, 1 mM EDTA, and 1 mM dithiothreitol. Fractions (500 l of each) were collected and proteins precipitated with 10% trichloroacetic acid followed by SDS-PAGE and silver staining to reveal DNA-PKcs and NDH II. The positions of the void volume and of bovine serum albumin are indicated. C, DNA-PK phosphorylated NDH II in the presence of DNA. a, full-length NDH II (0.3 g) was incubated with DNA-PK in the presence of [␥-32 P]ATP for the indicated time points. Phosphorylated NDH II was observed by exposure of the protein gel to an x-ray film and autoradiography. The NDH II inputs are shown by Coomassie staining. b, the DEIH core (0.5 g) and a glutathione S-transferase fusion protein (1 g) containing the N-terminal double-stranded RNA binding domains of NDH II were incubated with DNA-PK as above in a. Protein inputs and phosphorylation are shown by Coomassie staining and autoradiography, respectively. D, ␥-H2AX inhibited the phosphorylation of full-length NDH II (a) and its DEIH core (b) by DNA-PK. ␥-H2AX was derived from immunoprecipitation by the mouse JBW301 antibody from the DNA-PK phosphorylation mixtures containing increasing amounts of H2AX input. The phosphorylation of NDH II and of its DEIH core domain were performed in a reaction volume of 10 l that was mixed with immunoprecipitates containing increasing amounts of ␥-H2AX. Phosphorylated NDH II and its DEIH core are shown by autoradiography and ␥-H2AX, by silver staining after SDS-PAGE. E, ␥-H2AX suppressed the DNA-PK-stimulated ATPase activity of NDH II. ATPase assays of full-length NDH II (0.3 g) were performed in a 10-l volume of DNA-PK reaction buffer. Activated charcoal (Sigma) was used for separating adsorbed [␥-32 P]ATP versus the released 32 P i . Presented are the ATPase activities of NDH II (pmol of 32 P i released/min) correlated to the nucleotide concentration of calf thymus DNA in the absence or presence of DNA-PK and/or ␥-H2AX (1 g).

FIG. 7. Actinomycin D (AcD) led to an accumulation of NDH II into RNA-containing nuclear bodies that co-localized with ␥-H2AX.
Human HeLa cells were untreated (A-C and J-L) or treated (D-I) with actinomycin D as described in the legend to Fig. 1, followed by immunofluorescence after pre-treatment of the cells with 0.5% Nonidet P-40 without (A-F, J-L) or with RNase A (0.1 mg/ml) (G-I) for 15 min at room temperature. Images were taken by confocal laser scanning microscopy. NDH II was monitored by its rabbit antiserum, and a secondary antibody conjugated with Cy3 fluorophore (A, D, G, and J). ␥-H2AX was detected by the mouse JBW301 antibody and a fluorescein isothiocyanateconjugated secondary antibody (B, E, H, and K). C, F, I, and L display the merged signals of these two proteins. Fluorescence measurements along the indicated lines, as shown in CЈ, FЈ, IЈ, and LЈ, are presented for demonstrating the co-localization of these proteins. sequestration of NDH II and not in its functional activation. It has been described that ␥-H2AX promotes an enrichment of DNA repair proteins at damaged DNA sites rather than being involved in DNA repair by itself (60). Despite this and in addition to a passive role in protein sequestration, ␥-H2AX may actively assemble functional protein complexes for DNA repair or other chromosomal processes. After treating cells with actinomycin D, the so-induced formation of ␥-H2AX apparently accumulates NDH II at sites of stalled transcription bubbles.
Here, NDH II either may facilitate the structural reorganization of the bubble and/or the processing of abnormally synthesized RNAs, both of which may be critical for cellular recovery.
Interestingly, a previous report (61) shows that expression of a truncated form of NDH II exhibited dominant negative effects on the wild-type functions of BRCA1. These led to a series of abnormalities such as an impaired recruitment of BRCA1 to sites of DNA repair, delayed cell division, and defective centrosome replication (61). BRCA1 physically interacts with NDH II (56) and is important for chromatin remodeling, transcription, homologous DNA recombination, DNA repair, and apoptosis (62). This may be critical for supervising the genomic integrity at transcriptionally active sites that are more susceptible to the attack of DNA-damaging agents. According to the findings reported here, NDH II may share some functions with BRCA1 and may be involved in similar pathways as this tumor suppressor protein. Further investigations on the suspected role of NDH II in DNA repair, particularly based on its association with ␥-H2AX, may broaden our view on the mechanisms that are involved in DNA repair and the surveillance of genomic integrity.