Search

Figure S4 shows the calibration of the initial level of TOP2αcc after cell treatment with etoposide or F14512.

Supplementary Table S1

Centromeres are defined by chromatin containing the histone H3 variant CENP-A assembled onto repetitive α-satellite sequences, which are actively transcribed throughout the cell cycle. Centromeres play an essential role in chromosome inheritance and genome stability through coordinating kinetochores assembly during mitosis. Structural and functional alterations of the centromeres cause aneuploidy and chromosome aberrations which can induce cell death. In human cells, the tumor suppressor BRCA1 associates with centromeric chromatin in the absence of exogenous damage. While we previously reported that BRCA1 contributes to proper centromere homeostasis, the mechanism underlying its centromeric function and recruitment was not fully understood. Here, we show that BRCA1 association with centromeric chromatin depends on the presence of R-loops, which are non-canonical three-stranded structures harboring a DNA:RNA hybrid and are frequently formed during transcription. Subsequently, BRCA1 counteracts the accumulation of R-loops at centromeric α-satellite repeats. Strikingly, BRCA1-deficient cells show impaired localization of CENP-A, higher transcription of centromeric RNA, increased breakage at centromeres and formation of acentric micronuclei, all these features being R-loop-dependent. Finally, BRCA1 depletion reveals a Rad52-dependent hyper-recombination process between centromeric satellite repeats, associated with centromere instability and missegregation. Altogether, our findings provide molecular insights into the key function of BRCA1 in maintaining centromere stability and identity.

In contrast to its well-known role in the DNA damage response during interphase, the function of BRCA1 in the maintenance of chromosomal stability during mitosis remains to be defined. In this study, we uncover a novel role of BRCA1 in preserving centromere integrity in mitotic human cells. Using immunofluorescence and chromatin immunoprecipitation approaches, we report BRCA1 association with centromeric chromatin during mitosis. BRCA1 depletion impairs centromeric cohesion, leading to an increase in interkinetochore distance and in unpaired sister-chromatids frequency during prometaphase. Moreover, BRCA1 loss partially decreased accumulation of the Aurora B kinase at the centromere. We found that proper recruitment of the DNMT3b DNA methyltransferase to satellite sequences is BRCA1-dependent during mitosis, suggesting that DNA hypomethylation contributes to Aurora B mislocalization. BRCA1-deficient cells exhibited decreased ability to correct improper Aurora B-dependent chromosome-spindle attachments and to align chromosomes at metaphase. Finally, we show that BRCA1 disruption promotes merotelic kinetochore attachments that represent a major mechanism of aneuploidy in human cells. In summary, we report here a novel function of BRCA1 in maintaining chromosomal stability through its contribution to the mitotic centromere integrity necessary for faithful segregation of sister-chromatids during cell division.

Poisons of topoisomerase II (TOP2) kill cancer cells by preventing religation of intermediate DNA breaks during the enzymatic process and thus by accumulating enzyme–drug–DNA complexes called TOP2 cleavage-complex (TOP2cc). F14512 is a highly cytotoxic polyamine-vectorized TOP2 inhibitor derived from etoposide and currently in clinical trials. It was shown in vitro that F14512 has acquired DNA-binding properties and that the stability of TOP2cc was strongly increased. Paradoxically, at equitoxic concentrations in cells, F14512 induced less DNA breaks than etoposide. Here, we directly compared etoposide and F14512 for their rates of TOP2cc production and resolution in human cells. We report that targeting of TOP2α and not TOP2β impacts cell killing by F14512, contrary to etoposide that kills cells through targeting both isoforms. Then, we show that despite being more cytotoxic, F14512 is less efficient than etoposide at producing TOP2α cleavage-complex (TOP2αcc) in cells. Finally, we report that compared with TOP2αcc mediated by etoposide, those generated by F14512 persist longer in the genome, are not dependent on TDP2 for cleaning break ends from TOP2α, are channeled to a larger extent to resection-based repair processes relying on CtIP and BRCA1 and promote RAD51 recruitment to damaged chromatin. In addition to the addressing of F14512 to the polyamine transport system, the properties uncovered here would be particularly valuable for a therapeutic usage of this new anticancer compound. More generally, the concept of increasing drug cytotoxicity by switching the repair mode of the induced DNA lesions via addition of a DNA-binding moiety deserves further developments. Mol Cancer Ther; 16(10); 2166–77. ©2017 AACR.

Our results reveal a network of new potential Bcl-2 partners identified through the Bcl-2 immunocapture and mass spectrometry approach and analyzed by gene ontology mining. Importantly, we report for the first time the identification of galectin-7, a member of a family of β-galactoside-binding lectins, as a new mitochondrial Bcl-2 interacting partner., Although the anti-apoptotic activity of Bcl-2 has been extensively studied, its mode of action remains incompletely understood. Deciphering the network of Bcl-2 interacting factors is necessary to better understand the key function of Bcl-2 in apoptosis initiation. To identify novel Bcl-2 mitochondrial partners, we have combined a Bcl-2 immunocapture with a mass spectrometry analysis using highly pure mitochondrial fractions isolated from human cancer cells. We identified at high confidence 127 potential Bcl-2–interacting proteins. Gene ontology mining reveals enrichment for mitochondrial proteins, endoplasmic reticulum–associated proteins, and cytoskeleton-associated proteins. Importantly, we report the identification of galectin-7 (Gal7), a member of a family of β-galactoside–binding lectins that was already known to exhibit a pro-apoptotic function, as a new mitochondrial Bcl-2 interacting partner. Our data further show that endogenous Bcl-2 coimmunoprecipitates with Gal7 and that recombinant Gal7 directly interacts with recombinant Bcl-2. A fraction of Gal7 is constitutively localized at mitochondria in a Bcl-2–dependent manner and sensitizes the mitochondria to the apoptotic signal. In addition, we show that the Bcl-2/Gal7 interaction is abolished following genotoxic stress. Taken together, our findings suggest that the binding of Gal7 to Bcl-2 may constitute a new target for enhancing the intrinsic apoptosis pathway.

Inherited mutations of the breast cancer susceptibility gene 1 (BRCA1) confer an increased risk for breast, ovarian and prostate cancer. BRCA1 has been involved in regulation of cell cycle progression, DNA damage signaling and repair, maintenance of genome integrity, ubiquitination and regulation of transcription. Aside from its essential functions in the DNA damage response BRCA1 has been also involved in the cellular response to microtubule damage. Emerging evidence indicates that BRCA1 regulates the duplication and the function of centrosomes, participates in mitotic spindle assembly and is required in the spindle checkpoint. Given BRCA1 distinct functions in microtubule-dependent pathways, we hypothesized that BRCA1 might be regulated following microtubule damage. In the present study, we report the novel finding that BRCA1 is phosphorylated by the checkpoint kinase Chk2 on the previously identified site Ser988 following anti-mitotic treatment in human cancer cells. Ser988-phosphorylated BRCA1 accumulates at centrosomes in response to microtubule damage but Ser988 is not essential for BRCA1 localization at the microtubule-organizing centers. We further demonstrate that the Ser988 phosphorylation is important for the inhibiting microtubule nucleation activity of BRCA1 and for BRCA1 function in cell survival following microtubule damage. These findings reveal a striking outcome of BRCA1 phosphorylation by Chk2 on its role in microtubule-dependent pathways and suggest a fine cross-talk between DNA damage and spindle damage responses.

Accurate repair of DNA double-strand breaks (DSB) caused during DNA replication and by exogenous stresses is critical for the maintenance of genomic integrity. There is growing evidence that the Polo-like kinase 1 (Plk1) that plays a number of pivotal roles in cell proliferation can directly participate in regulation of DSB repair. In this study, we show that Plk1 regulates BRCA1, a key mediator protein required to efficiently repair DSB through homologous recombination (HR). Following induction of DSB, BRCA1 concentrates in distinctive large nuclear foci at damage sites where multiple DNA repair factors accumulate. First, we found that inhibition of Plk1 shortly before DNA damage sensitizes cells to ionizing radiation and reduces DSB repair by HR. Second, we provide evidence that BRCA1 foci formation induced by DSB is reduced when Plk1 is inhibited or depleted. Third, we identified BRCA1 as a novel Plk1 substrate and determined that Ser1164 is the major phosphorylation site for Plk1 in vitro. In cells, mutation of Plk1 sites on BRCA1 significantly delays BRCA1 foci formation following DSB, recapitulating the phenotype observed upon Plk1 inhibition. Our data then assign a key function to Plk1 in BRCA1 foci formation at DSB, emphasizing Plk1 importance in the HR repair of human cells.

The BRCA2 tumor suppressor has been implicated in the maintenance of genomic integrity through a function in cellular responses to DNA damage. The BRCA2 protein directly associates with Rad51, that is essential for repair of double-strand breaks (DSBs) by homologous recombination (HR). In this report, we study the BRCA2-defective Chinese hamster cell mutant V-C8 for its ability to perform homology-directed repair (HDR) between repeated sequences. V-C8 cells were recently shown to be defective in Rad51 foci formation in response to DNA damage. Strikingly, we find that these BRCA2 mutant cells exhibit a strong stimulation of HDR activity compared to the V79 parental cells, which harbor a wild-type BRCA2. Furthermore, molecular characterization of the HDR products shows that loss of BRCA2 in V-C8 cells leads to significant reduction in Rad51-dependent gene conversion but strong enhancement of Rad51-independent single-strand annealing (SSA) events frequency. These data imply that, when HDR by conservative gene conversion is impaired, DSBs usually repaired by this pathway are instead resolved by other non-conservative HDR subpathways. Therefore, high chromosomal instability in BRCA2-deficient cells presumably results from enhancement of error-prone repair mechanisms, such as SSA.

In mammalian cells, the rate of nucleotide excision repair of UV dimers is heterogeneous throughout the genome, with repair occurring more rapidly in the transcribed strand of active genes than in the genome overall. This repair pathway is termed transcription-coupled repair (TCR) and is thought to permit the rapid resumption of RNA synthesis following UV irradiation. To evaluate the inducibility of the TCR process, we examined the repair of UV-induced cyclobutane pyrimidine dimers (CPDs) at the level of the gene following exposure of hamster cells to a sub-lethal UV fluence, 3 h prior to a higher dose. Repair was detected by a well-established technique allowing quantification of CPDs at the level of a specific strand by Southern blot hybridization. Here, we show that prior low-dose irradiation clearly enhanced the early rate of CPD removal in the transcribed strand of the active DHFR gene. Furthermore, the RNA synthesis recovery following UV exposure was stimulated by the priming UV dose. Thus, we provide evidence for an inducible TCR response to CPDs in hamster cells. This pathway is independent of the p53 activation, since the hamster cell line that we used expresses high levels of mutant p53 protein.

This study examines the capacity of a mammalian cell to repair, at the gene level, DNA base lesions generated by photoactivation of acridine orange. Chinese hamster ovary fibroblasts were exposed to acridine orange and visible light, and gene-specific DNA repair was measured in the dihydrofolate reductase (DHFR) gene and in the mitochondrial genome. DNA lesions were recognized by Escherichia coli formamidepyrimidine-DNA glycosylase (FPG) which removes predominantly 8-oxodG and the corresponding formamidopyrimidine ring opened bases, and subsequently cleaves the DNA at the resulting apurinic site. FPG-recognized DNA lesions increased linearly with increasing photo-activation of AO, while cell survival was not affected by light alone and was negligibly affected by preincubation with AO in the dark. The frequency of induction of FPG-sensitive DNA damage by photoactivation of AO was similar in the transcribed and non-transcribed nuclear DNA as well as in the mitochondrial DNA. FPG-sensitive sites in the DHFR gene were repaired quickly, with 84% of adducts repaired within 4 h. The lesion frequency, kinetics and percent of repair of non-transcribed genomic DNA did not differ significantly from repair in the active DHFR gene up to 1 h postexposure. At late time points, transcribed DNA was repaired faster than the non-transcribed DNA. Mitochondrial DNA was efficiently repaired, at a rate similar to that in the active nuclear DNA.

Escherichia coli RecA protein plays a central role both in DNA repair and in recombination. We report biochemical properties of three new RecA proteins mutated at positions 199 (RecA694), 207 (RecA659), and 211 (RecA611) in the putative DNA binding site. RecA694 had a wild-type phenotype, whereas RecA611 and RecA659 were deficient in promoting both the self-cleavage of LexA repressor and the DNA-strand exchange reaction. In order to determine the origin of this inhibition, we examined the capacity of wild-type and mutant proteins to bind to single-stranded DNA (with and without single-stranded binding protein, SSB), double-stranded DNA, and ATP. DNA strand exchange defects were correlated with the inability of mutant proteins to displace SSB from DNA. For the recA659 mutation this inhibition was reversed by equimolar wild-type protein. In contrast, mixtures of either wild-type/RecA659 or wild-type/RecA611 proteins remained deficient in LexA cleavage, suggesting that the dominant negative phenotype of the mutant proteins may be a consequence of the formation heterologous RecA complexes. Various mutations in the putative DNA binding site of RecA protein altered ATP binding, ATPase activity, displacement of SSB from single-stranded DNA, and protein-protein interactions. These results are consistent with the hypothesis that DNA binding to this site of RecA relays allosteric effects to several functional domains throughout the protein.

The human excision repair gene ERCC-1 gene restores normal resistance to UV and mitomycin C in excision repair deficient chinese hamster ovary cells of complementation group 1. To investigate the involvement of the ERCC-1 gene in gene-specific repair of bulky lesions, we have studied the removal of damage induced by the antitumor agent cisplatin in CHO mutant 43-3B cells of group 1, with or without transfection with the ERCC-1 gene. Firstly, we determined the contribution of the ERCC-1 gene to the repair of interstrand crosslinks (ICL) induced by cisplatin and found efficient removal of ICLs from the dihydrofolate reductase (DHFR) gene in the ERCC-1 transfected 43-3B cells. We then assessed the contribution of ERCC-1 to the repair of intrastrand adducts (IA) induced by cisplatin. Compared to the wild-type parental cell line, the ERCC-1 transfected 43-3B cells repaired the IAs in the DHFR gene inefficiently. Thus, our data suggest that the ERCC-1 gene is more involved in the repair of interstrand crosslinks than in the removal of intrastrand adducts.

We have measured the DNA damage formation and repair in the ribosomal and the dihydrofolate reductase (DHFR) genes after treatment of hamster cells with different types of DNA damaging agents. In mammalian cells, the ribosomal DNA (rDNA) is transcribed by RNA polymerase I, whereas the DHFR is transcribed by RNA polymerase II, whereas the DHFR is transcribed by RNA polymerase II. Cells were treated with agents that induce different types of lesions, and that are known to be repaired via different pathways. We used UV (254 nm) irradiation, treatment with cisplatin and treatment with the alkylating agents nitrogen mustard (HN2) and methyl methanesulphonate (MMS). UV induced pyrimidine dimers were detected with the enzyme T4 endonuclease V, which creates nicks at the dimer sites; the breaks are then resolved and identified by denaturing electrophoresis and Southern blot. Intrastrand adducts formed by the alkylating agents HN2 and MMS were quantitated by generating strand breaks at abasic sites after neutral depurination. Interstrand crosslinks (ICL) formed by HN2 and cisplatin were detected by a denaturation-reannealing reaction before neutral agarose gel-electrophoresis. We find that the repair of the pyrimidine dimers is significantly less efficient in the RNA polymerase I transcribed rDNA genes than in RNA polymerase II transcribed DHFR gene at 8 and 24 h after irradiation. ICL and intrastrand adducts induced by HN2 are also removed more slowly from the rDNA than from the DHFR gene. In contrast, MMS induced intrastrand adducts and cisplatin induced ICL are repaired equally efficiently in the RNA polymerase I and RNA polymerase II transcribed genes. We conclude that for some types of DNA damage, there is less repair in the ribosomal genes than in the DHFR; but for other DNA lesions there is no difference. The difference in repair efficiency between the rDNA and the DHFR genes may reflect the different RNA polymerase involved in their transcription. It may, however, alternatively, reflect the different nuclear localization of these genes.

We previously demonstrated that phospho-Thr56 Bcl-2 colocalizes with Ki-67 and nucleolin in nuclear structures in prophase cells and is detected on mitotic chromosomes in later mitotic phases. To gain insight into the fine localization of Bcl-2 on mitotic chromosomes, we further investigated Bcl-2 localization by immunostaining of Bcl-2 with known components of metaphase chromosomes and electron microscopic immunocytochemistry. Immunofluorescence analysis on HeLa mitotic cells together with chromatin immunoprecipitation assays showed that Bcl-2 is associated with the condensed chromatin. Co-immunostaining experiments performed on mitotic chromosome spreads demonstrated that Bcl-2 is not localized on the longitudinal axis of chromatids with the condensin complex, but partially colocalizes with histone H3 on some regions of the mitotic chromosome. Finally, most of the Bcl-2 staining overlaps with Ki-67 staining at the chromosome periphery. Bcl-2 localization at the periphery and over the mitotic chromosome was confirmed by immunoelectron microscopy on mitotic cells. Our results indicate that Bcl-2 is an integral component of the mitotic chromosome.

International audience; OPA1, an intra-mitochondrial dynamin GTPase, is a key actor of outer and inner mitochondrial membrane dynamic. OPA1 amino-terminal cleavage by PARL and m-AAA proteases was recently proposed to participate to the mitochondrial network dynamic in a DeltaPsi(m)-dependent way, and to apoptosis. Here, by an in vitro approach combining the use of purified mitochondrial fractions and mitochondrial targeting drugs, we intended to identify the central stimulus responsible for OPA1 cleavage. We confirm that apoptosis induction and PTPore opening, as well as DeltaPsi(m) dissipation induce OPA1 cleavage. Nevertheless, our experiments evidenced that decreased mitochondrial ATP levels, either generated by apoptosis induction, DeltaPsi(m) dissipation or inhibition of ATP synthase, is the common and crucial stimulus that controls OPA1 processing. In addition, we report that ectopic iron addition activates OPA1 cleavage, whereas zinc inhibits this process. These results suggest that the ATP-dependent OPA1 processing plays a central role in correlating the energetic metabolism to mitochondrial dynamic and might be involved in the pathophysiology of diseases associated to excess of iron or depletion of zinc and ATP.

Zeocin is a member of the bleomycin/phleomycin family of antibiotics, known to bind and cleave DNA. We established human SK-OV-3 cells that stably express the Zeocin resistance gene (Sh ble) using an ecdysone-inducible mammalian expression system. Surprisingly, our results demonstrated that Zeocin, added in the culture medium to maintain the expression of the ecdysone receptor, was responsible for the formation of DNA strand breaks in the recombinant cells. This suggests that the Zeocin is not completely detoxified and is still able to cleave DNA, despite the stable expression of the Sh ble gene in the recombinant clones. Our study indicates that one needs to be very cautious in the interpretation of data involving stable cell lines selected with Zeocin.

DNA interstrand cross-links (ICL)-inducing agents such as cisplatin, mitomycin C (MMC) and nitrogen mustards are widely used as potent antitumor drugs. Although ICL repair mechanism is not yet well characterized in mammalian cells, this pathway is thought to involve a sequential action of nucleotide excision repair (NER) and homologous recombination (HR). The importance of unraveling ICL repair pathways is highlighted by the hypersensitivity to ICL-inducing agents in cells of patients with the genetic disease Fanconi anemia (FA) and in cells mutated in the Breast Cancer susceptibility genes BRCA1 and BRCA2. To better characterize the involvement of HR in the sensitivity to ICL-inducing agents, we examined spontaneous and ICL-induced HR in rodent FA-like V-H4 cells. In this report, we show that MMC-hypersensitive V-H4 cells exhibit an increased spontaneous homology-directed repair (HDR) activity compared to the resistant V79 parental cells. Elevated HDR activity results mainly in increased conservative Rad51-dependent recombination, without affecting non-conservative single-strand annealing process (SSA). We also show that HDR activity is enhanced following MMC treatment in parental cells, but not in rodent FA-like V-H4 cells. Moreover, our data indicate that Rad51 foci formation is significantly delayed in these FA-like cells in response to crosslinking agent. These findings provide evidence for an impairment of HR control in V-H4 cells and emphasize the involvement of the FA pathway in HR-mediated repair.

DNA cross-linking agents such as mitomycin C (MMC) and cisplatin are used as chemotherapeutic agents in cancer treatment. However, the molecular mechanism underlying their antitumor activity is not entirely clear. Critical steps in cytotoxicity toward cross-linking agents can involve DNA repair efficiency, inhibition of replication, cell-cycle checkpoints, regulation, and induction of apoptosis. The complexity of the mechanisms of the mammalian cell defense against cross-linking agents is reflected by the existence of many complementation groups identified in rodent cells that are specifically sensitive to MMC. We recently showed that increased induction of apoptosis contributes to the MMC sensitivity of the group represented by the V-H4 hamster mutant cell line. In this study, through the analyses of a substractive library, we discovered that sensitive V-H4 cells display a 40-fold increase of steady-state expression of metallothionein II (MT-II) mRNA compared with resistant parental V79 cells. Down-regulation of MT-II by antisense oligonucleotides partially restores MMC resistance in V-H4 cells, indicating that MT-II overexpression is directly involved in MMC hypersensitivity of these cells. MTs have been reported to regulate the activation of NF-kappaB, one of the key proteins that modulates the apoptotic response. Here we found that NF-kappaB activation by MMC is impaired in V-H4 cells and is partially restored following down-regulation of MT-II by antisense oligonucleotides. All these data suggest that the overexpression of MT-II in V-H4 cells impairs NF-kappaB activation by MMC, resulting in decreased cell survival and enhanced induction of apoptosis.

To elucidate the mechanisms of the mammalian cell defense against cross-linking agents, we studied previously cellular responses to mitomycin C (MMC) treatment in two MMC-hypersensitive hamster cell mutants' V-H4 and V-C8, as well as their parental cell line V79. In the present report, we investigated whether alterations in cell cycle checkpoints and induction of apoptosis could be responsible for the MMC hypersensitivity of the V-H4 and V-C8 mutant cell lines. First, we found that parental and mutant cells exhibited similar cell cycle responses to MMC concentrations of equivalent cytotoxicity, arguing against a defective cell cycle checkpoint in hypersensitive cell lines. In contrast, we showed that mutant cells underwent greater levels of apoptosis following MMC treatment than parental cells. These findings indicate that increased induction of apoptosis contributes to the hypersensitivity of V-H4 and V-C8 cells to the growth inhibitory effect of MMC. This differential apoptotic response was observed with both equimolar and equitoxic MMC doses and was specific to the cross-linking agent MMC, suggesting that control of the apoptotic process is altered in both MMC-hypersensitive mutants. The defective genes in V-H4 and V-C8 cells would then function in the regulation of an apoptotic pathway triggered by MMC-induced damage and independent of p53-mediated transcription.

A major factor in the progression of senescence and of age-associated diseases is the accumulation of DNA damage. Several studies have documented an age-associated increase in oxidative DNA damage, notably 7,8 dihydro-8-oxodeoxyguanosine (8-oxodG). Although not all studies have agreed on damage accumulation in the nuclear DNA, there seems to be a consensus that the level of 8- oxodG increases with aging in mitochondrial DNA (Richter et al., 1988). Oxidative DNA damage may (1) induce genes that are involved in the regulation of cellular proliferation (an example of this would be the activation of the p53 tumor suppressor gene) or could (2) simply inactive specific genes and thereby provoke progression to senescence or malignancies. If this DNA damage is a key element in senescence, then a major goal in aging research should be to clarify the mechanism involved. One likely mechanism is that the DNA repair processes which normally remove these DNA damages, deteriorate with age and thereby permit the damage accumulation. We therefore need to investigate the change in DNA repair processes and their efficiency with increasing age.

The ability to repair DNA interstrand cross-links may be an important factor contributing to mitomycin C (MMC) and cisplatin cytotoxicities. We have assessed the repair of interstrand cross-links induced by MMC in two MMC-hypersensitive hamster cell mutants and their resistant parental cell line. Using a gene-specific repair assay, we found no evidence for repair of MMC cross-links in either parental or mutant cells, suggesting that persistence of DNA interstrand cross-links is not responsible for the differential toxicity of MMC towards hypersensitive cells. Repair of cisplatin-induced interstrand cross-links was efficient in resistant as well as in mutant cells. Therefore we concluded that a defect in excision repair of interstrand cross-links was not responsible for the cytotoxic effects of MMC and cisplatin in these hypersensitive mutants.