|Home | About | Journals | Submit | Contact Us | Français|
The product of the breast and ovarian cancer susceptibility gene BRCA1 has been implicated in several aspects of the DNA damage response but its biochemical function in these processes has remained elusive. In order to probe BRCA1 function we conducted a yeast two-hybrid screening to identify interacting partners to a conserved motif (Motif 6) in the central region of BRCA1. Here we report the identification of the actin-binding protein Filamin A (FLNA) as a BRCA1 partner and demonstrate that FLNA is required for efficient regulation of early stages of DNA repair processes. Cells lacking FLNA display a diminished BRCA1 IR-induced focus formation and a delayed kinetics of Rad51 focus formation. In addition, our data also demonstrate that FLNA is required to stabilize the interaction between components of the DNA-PK holoenzyme, DNA-PKcs and Ku86 in a BRCA1-independent fashion. Our data is consistent with a model in which absence of FLNA compromises homologous recombination and non-homologous end joining. Our findings have implications for the response to radiation-induced DNA damage.
The presence of DNA damage triggers a series of events collectively known as the DNA Damage Response (DDR) pathway.1,2 Its biological role is to promote efficient DNA repair and to coordinate this activity with cell cycle progression. Injuries to DNA primarily activate three PI3K related kinases ATM, ATR and DNA-PK which are recruited to DNA breaks by the Mre11/Rad50/NBS1 complex, ATRIP and Ku86, respectively.3 This process is guided by different DNA structures: ATM and DNA-PK are activated by double stranded breaks (DBS) and ATR is activated by Replication Protein A (RPA) coated single stranded DNA (ssDNA).1
Two main mechanisms exist to promote DSB repair. The error-prone non-homologous end joining (NHEJ) is functional in all phases of the cell cycle, while the error-free homologous recombination (HR) only functions in S and G2 phases.1,2 HR is initiated by ssDNA resulting from resection of DNA ends at the DSB.1 RPA then binds ssDNA with very high affinity and is visualized as nuclear foci detected by immunofluorescence.4 The RPA-coated ssDNA is the substrate for RAD51 recombinase, which is loaded by BRCA2 and mediates the DNA pairing during HR.1,5 RAD51 co-localizes with the tumor suppressors BRCA1 and BRCA2 in radiation-induced nuclear foci.6,7 BRCA1 and BRCA2 are part of a complex that controls RAD51 function and the efficiency of HR.5,8
Germline mutations in BRCA1 lead to increased predisposition to breast and ovarian cancer.9,10 Cloning of BRCA1 in 1994 made possible the genetic testing of individuals with strong family history of breast cancer and set the stage for an intensive effort to understand its biological functions and the nature of its tumor suppressive activities.11,12 However, 15 years later it is still not clear which of its many activities can be related directly to its role as tumor suppressor.
BRCA1 has been implicated in several aspects of the DNA damage response (DDR). Its role in the DDR seems to span a wide range of activities from damage signaling to participation in repair and the coordination of cell cycle checkpoints.1,13-15 In particular, BRCA1 has been implicated in HR,16 microhomology-mediated17 and NHEJ DNA repair.18,19
Our laboratory has focused on a systematic analysis of domains and motifs in BRCA1 as a means to understand its biochemical functions.20 Here we analyzed a conserved region, called Motif 6, spanning amino acids 845–869 coded by BRCA1's large exon 11.21,22 Using a yeast two hybrid screen we identified the actin-binding protein Filamin A (FLNA) as an interacting partner of BRCA1. Interestingly, FLNA has been shown to interact with BRCA2 and to participate in the DDR.23-25 Cells lacking FLNA exhibit prolonged checkpoint activation leading to accumulation of cells in G2/M after ionizing radiation.23
We show that BRCA1 and FLNA interact in mammalian cells and this interaction is mediated by Motif 6 and by another uncharacterized region in BRCA1 N-terminus called Motif 2.21 Binding to BRCA1 is mediated by the C-terminus of FLNA, a region that includes its dimerization domain. Introduction of a BRCA1 missense variant found in individuals with family history of breast cancer abrogates the interaction. Lack of FLNA leads to a broad defect in DNA repair with accumulation of ssDNA combined with the hyperactivation of ATM and ATR-mediated signaling. We show that this phenotype is due to a combined failure of Ku86 and DNA-PKcs to form stable complexes, and to defects in BRCA1 and Rad51 focus formation implicating FLNA in the control of DNA repair.
In order to identify interactors to the conserved Motif 6 of BRCA1 spanning amino acid residues 845–869 (Suppl. Fig. 1A) we performed a yeast two-hybrid screening against a human mammary gland cDNA library. Two overlapping clones coding for human Filamin A (FLNA; OMIM # 300017), spanning amino acid residues 2443-2647 and 2477–2647 (Suppl. Fig. 1B), were identified. This region includes repeat 23, the hinge region, and repeat 24 in the C-terminus FLNA (Suppl. Fig. 1B).26 We mapped the minimal region of FLNA that interacts with BRCA1 Motif 6 by testing binding of a series of FLNA deletion mutants (Suppl. Fig. 1B). Only the fragment aa 2477–2647 was able to bind BRCA1 Motif 6 (Suppl. Fig. 1B).
Next, we tested whether endogenous FLNA interacted with endogenous BRCA1 in mammalian cells. Immunoprecipitation using a specific monoclonal antibody against BRCA1 pulled down FLNA in HeLa and HCT116 cells (Fig. 1A). In addition, immunoprecipitation using an antibody against FLNA was able to pull down BRCA1 (Fig. 1A). Thus, BRCA1 and FLNA interact in vivo and the interaction is mediated by the C-terminus of FLNA.
Because FLNA and BRCA1 have been demonstrated to be primarily cytoplasmic and nuclear, respectively, we biochemically fractionated HCT116 cells to determine in which subcellular compartment the interaction occurs (Suppl. Fig. 1C). We found that FLNA is expressed in the nucleus and cytoplasm and BRCA1 can be co-immunoprecipitated by FLNA in the nuclear fraction (Suppl. Fig. 1C). We also determined that the interaction is direct as bacterially expressed GST-tagged BRCA1 (aa 141–302) can pull down bacterially expressed His-tagged FLNA C-terminus (Suppl. Fig. 1D).
In the course of our yeast experiments we noted that the interaction between Motif 6 and FLNA was relatively weak (data not shown). Thus, we hypothesized that other regions in BRCA1 might contribute to binding. We co-expressed in-frame fusions of GST to deletion fragments of BRCA1 and a FLAG-tagged FLNA fragment (aa 2477–2647) in 293FT cells (Fig. 1B) to assess each region's contribution to binding.
We immunoprecipitated FLAG-FLNA using α-FLAG agarose beads and the eluate with FLAG-peptide was separated by SDS-PAGE. Western blot against FLNA and BRCA1 confirmed that FLAG-FLNA was properly folded and interacted with endogenous BRCA1, respectively (Fig. 1C, left). Western blot against GST revealed interaction of FLNA with different fragments of BRCA1 under low stringency (Fig. 1C, left). Interaction with fragments 1 (aa 1–324), 3 (aa 502–802) and 4 (aa 758–1064) was detected even under high stringency conditions (Fig. 1C, right). Reverse pull-downs of endogenous FLNA using GT-beads confirmed that the interaction is mediated by BRCA1 fragments 1, 3 and 4 (Fig. 1D). In both experiments, BRCA1 fragment 1 showed the strongest interaction (Fig. 1C and D).
Fragment 1 (aa 1–324) includes the RING finger (aa 1–101)11 and nuclear export signals (aa 22–30 and aa 81–99).27,28 To determine whether the interaction was mediated by these motifs we used deletion mutants of BRCA1 fragment 1 (Fig. 2A). Initially, we identified BRCA1 residues 141–240 as the interacting region to FLNA (aa 2477–2647) (Fig. 2A). Further mapping identified residues 160–190 as the minimal region required for binding (Fig. 2B). This region, called Motif 2, had been previously identified as conserved motif in BRCA1 orthologs.21,22
To assess whether BRCA1 and FLNA interaction might contribute to breast cancer we searched the Breast Cancer Information Core database (research.nhgri.nih.gov/bic/) for variants in this region. Variant Y179C is a frequent missense change recorded in the database (BIC Database). Introduction of BRCA1 Y179C mutant significantly reduced BRCA1 interaction to FLAG-FLNA aa 2477–2647 and to endogenous FLNA (Fig. 2C) further demonstrating the specificity of the interaction. Because other regions in BRCA1 besides Motif 2 also contributed to the binding we investigated whether the Y179C mutation would disrupt binding to FLNA in the context of full length BRCA1. Introduction of the Y179C mutation significantly reduced the interaction in the full length context as compared to wild type BRCA1 (Fig. 2C, right). In summary, these experiments demonstrate that Motif 2 primarily mediates the interaction to FLNA. Taken together these data raised the possibility that lack of FLNA might impair BRCA1 foci formation after DNA damage. Thus, the following experiments were directed at assessing the role of the interaction in the DNA damage response.
To further characterize the functional significance of FLNA/BRCA1 interaction we obtained the M2 melanoma cell line which lacks FLNA and its counterpart A7 which was obtained by reconstituting M2 cells with full length FLNA cDNA.29 First, we assessed the kinetics of double strand break (DSB) repair after ionizing radiation (IR). We irradiated or mock treated the FLNA- and FLNA+ cell lines and collected cells at several time points after IR. We monitored the presence of DSB with an antibody against histone H2AX phosphorylated at Serine 139 (called γ-H2AX), a marker for DSBs.30 Whereas the FLNA+ cell line efficiently repaired DSBs and by 8 h after IR there was no detectable γ-H2AX (Fig. 3A), FLNA- cells had a sustained high level of γ-H2AX for up to 32 h after IR. We confirmed this observation using Comet assays (Fig. 3B).
Next, we assessed whether cells lacking FLNA had a compromised DNA damage signaling. Thus, we tested whether ATM and ATR were properly activated upon DNA damage. Phosphorylation of ATM S1981 was not compromised in FLNA- cells (Fig. 3C, Top). Likewise, phosphorylation of CHK2 T68 and CHK1 S317, markers of ATM and ATR activation, respectively, did not show a defect (Fig. 3D). Intriguingly, we consistently observed higher levels of phosphorylation of ATM, CHK2 and CHK1 in cells lacking FLNA (Fig. 3C and D) indicating an upregulation of ATM and ATR signaling. These results confirmed previous data by Meng et al.23 showing a sustained activation of CHK2 and CHK1 in FLNA-deficient cells following damage.
To determine whether the deficiency in repair was due to defective recruitment of factors required for the DDR we performed immunofluorescence analysis in non-irradiated or irradiated cells at 1 and 24 h after IR. Accumulation of γ-H2AX and pS343-NBS1, early markers of DNA damage, was comparable in both cell lines at 1 h (Fig. 4A). In order to determine if there were small differences we quantified foci-positive cells (Fig. 4A, lower). Results were comparable in both cell lines at 0 and 1 hr after IR, but FLNA-negative cells showed increased number of foci-positive cells after 24 hr. Likewise, recruitment of mediator proteins MDC1, 53BP1, was also comparable at 1 h (Fig. 4B). Finally, repair factor RPA did not show any difference between the cell lines at 1 h (Fig. 4C). Consistent with our western blot results (Fig. 3) where we detected abnormally high levels of γ-H2AX, pT68-CHK2 and pS317-CHK1 at 24 h, we detected persistent foci of γ-H2AX, pS343-NBS1 and RPA at 24 h after irradiation only in FLNA-deficient cells (Fig. 4A–C). Thus, the repair defect in FLNA-deficient cells was not due to a failure to initiate the DNA damage response.
Next, we investigated the ability of BRCA1 and Rad51 to form IR-induced foci. A detailed analysis showed that FLNA-deficient cells are unable to efficiently form BRCA1 IR-induced foci as compared to FLNA-proficient cells (Fig. 4B and bottom). Although Rad51 displayed a comparable initial response at 3 h after IR, it failed to mount a response comparable to FLNA-proficient cells at 6 h after IR. Rad51 presented a delayed kinetics of foci formation with a peak at 15 h in FLNA-deficient cells (Fig. 4C and bottom). Taken together these data suggest that the compromised repair capacity in FLNA-deficient cells may be, at least partially, mechanistically tied to inefficient HR.
During our analysis we noted that RPA foci in FLNA-deficient cells were not only persistent 24 h after damage but were also significantly larger (Fig. 4D). To determine whether those foci were associated with chromatin we pre-extracted cells with Triton X100 before fixation. This method has been successfully used to detect only the fraction of RPA tightly bound to chromatin.31 Interestingly, FLNA-deficient cells accumulate large chromatin-bound RPA foci whereas FLNA+ cells present fewer and smaller chromatin-bound RPA foci at 24 h after IR (Fig. 4D). Whereas most FLNA+ cells have recovered from G2/M arrest and represent an asynchronous population at 24 h, most FLNA- cells remain arrested in G2/M at 24 h after IR.23 Thus, these large tracts of ssDNA found in FLNA- cells are unlikely to be due to replication foci.
To gain more insight of the mechanism by which FLNA participates in DNA repair we transfected FLNA+ and FLNA- cell lines with flag-tagged Filamin A aa 2477–2647 construct (BRCA1-interacting fragment). For the sake of simplicity we will refer to this FLNA BRCA1-interacting fragment as FLNA-Bf. At 24 h post transfection we irradiated cells with 8 Gy IR and collected samples at different time points. Transfection of FLNA-Bf did not lead to checkpoint recovery in FLNA- cells as measured by phosphorylation of CDC2 Y15 (Fig. 5A). Interestingly, transfection of the same fragment in FLNA+ cells led to a similar pheno-type as that found for FLNA- cells as shown by phosphorylation of CDC2 Y15 and H2AX S139 (Fig. 5B). We also confirmed that expression of FLNA-Bf acts in a dominant negative fashion in a stable transfection context (Fig. 5C). We generated HCT116 cells stably expressing GFP-FLNA-Bf or GFP alone that were mock-treated and irradiated. Cells expressing GFP-FLNA-Bf retained high levels of phosphorylated H2AX up to 32 h after damage while cells expressing GFP alone showed levels returning to unirradiated levels at 8 h after damage (Fig. 5C).
Next we asked whether expression of a GST-tagged BRCA1 FLNA-interacting fragment (BRCA1-Ff) could also lead to a dominant negative phenotype (Fig. 5D). In order to verify the specificity of the interaction we transfected a mutated BRCA1-Ff carrying the Y179C mutation and determined whether it lead to a dominant negative phenotype. Introduction of the Y179C mutation (Fig. 2C) significantly reduced the BRCA1-FLNA interaction. The wild type BRCA1-FLNA-Ff led to increased and sustained phosphorylation of CDC2 Y15 and H2AX (Fig. 5D) while the BRCA1-Ff Y179C (Fig. 5D) displayed a dominant negative effect that is intermediate between vector control (Fig. 5B) and the wild type construct (BRCA1-Ff; Fig. 5D). This intermediate effect could be due to the residual binding of BRCA1-Y179C mutant to FLNA. Alternatively, this could also be due to the inability of the mutant to disrupt the binding of FLNA to other regions of endogenous BRCA1 that participate in the interaction (Fig. 1C and D).
Our previous experiments demonstrated that FLNA- cells displayed defective DNA repair, showed signs of compromised HR, and accumulated large tracts of ssDNA. Because mammalian cells also repair DSBs using non homologous end joining (NHEJ) we hypothesized that lack of FLNA also had an impact on the NHEJ pathway.
First, we tested whether FLAG-FLNA aa 2477–2647 interacted with NHEJ factors. FLAG-FLNA aa 2477–2647 immunoprecipitated DNA-PKcs in 293FT cells independent of DNA damage (Fig. 6A). To determine whether FLNA was required for the stability of the Ku86/DNA-PKcs complex we performed immunoprecipitation experiments in FLNA+ and FLNA- cell lines in the presence or absence of irradiation (Fig. 6B). Interestingly, in FLNA- cells Ku86 and DNA-PKcs complex formation was compromised in IR-treated and untreated cells (Fig. 6B).
Next we tested whether FLNA was required for Ku86 loading onto chromatin after DNA damage. Ku86 was efficiently recruited to chromatin upon DNA damage in the presence and absence of FLNA (Fig. 6C) while we detected DNA-PKcs in chromatin only in the presence of FLNA (Fig. 3B). Interestingly, loading of Ku86 onto chromatin persisted longer and with consistently higher levels in FLNA- than in FLNA+ cells (Fig. 6C).
Finally, we tested whether BRCA1 was required to stabilize the interaction between DNA-PKcs and Ku86. We examined BRCA1-deficient HCC1937 cell line32 and a HCC1937 derivative reconstituted with full length BRCA1 (gift from Junjie Chen). Complex formation between Ku86 and DNA-PKcs was not dependent on BRCA1 under IR-treated or untreated conditions (Suppl. Fig. 2B).
In summary, our results indicate that cells lacking FLNA have a defect in the two principal mechanisms for double strand break repair. Mechanistically, FLNA impacts on HR by contributing to efficient recruitment of BRCA1 and Rad51 to IR-induced foci, and on NHEJ by promoting the stability of the DNA-PKcs and Ku86 complex.
In this paper we shed light on the mechanism by which Filamin A (FLNA) is required for efficient DNA repair. Our data indicates that lack of FLNA impacts on HR and NHEJ. FLNA is an actin-binding protein and its inactivation leads to an array of disorders such as otopalatodigital spectrum disorder, Melnick-Needles syndrome and periventricular heterotopia.26 Although of unclear significance, at least two families carrying germline mutations in BRCA1 have been shown to manifest ventricular heterotopia.33,34 FLNA interacts with a variety of proteins, including BRCA2,25 and deficiency in FLNA leads to sensitivity to DNA damage and a defect in the recovery from G2 arrest.23 Thus, we investigated further its role in the DNA damage response.
FLNA binds BRCA1 using its extreme C-terminus which contains its dimerization domain. BRCA1 interaction with FLNA is mediated by a 30 amino acid region in the N-terminus of BRCA1 which contains a conserved domain called Motif 2.21 Introduction of the Y179C mutation in Motif 2 significantly decreases the interaction. Analyses by the Align GV-GD method or by a yeast-based recombination assay suggest that Y179C may act as a deleterious mutation.35,36 On the other hand, this variant has been found co-occurring in trans with a known deleterious mutation, which indicates that it is unlikely to have severe effects.37 Thus, the Y179C may constitute a hypomorphic mutation with moderate effects on breast cancer predisposition. Of note, Motif 2 is close to the region that has been implicated in binding of BRCA1 to Ku86.38
In order to dissect the molecular role of FLNA in the DDR we took advantage of a well-characterized genetically-defined system. A melanoma cell line lacking FLNA was isolated and subsequently reconstituted with FLNA yielding a pair of cell lines in which the only difference is the presence or absence of FLNA.29 When we irradiated FLNA- and FLNA+ cells, we noticed that FLNA- took much longer to resolve DSBs (Fig. 3A and B). To elucidate the mechanism underlying the repair defect we systematically investigated the proficiency of damage signaling in FLNA- cells.
Initially we investigated the recruitment and activation kinetics of the upstream kinases, as well as their downstream substrates after DNA damage. We found that FLNA deficiency led to the hyperactivation of ATM as judged by phosphorylation of ATM S1981 and CHK2 T68, surrogate markers of ATM activation.39-41 Similarly, lack of FLNA also led to a hyperactivation of ATR, as measured by CHK1 S317 phosphorylation, a marker for ATR activation.42 Moreover, we also found sustained levels of phosphorylation of NBS1 S343 to be higher in FLNA- cells. Although the role of NBS1 phosphorylation in the DNA damage signaling is poorly understood, it is generally thought to reflect ATM and ATR activation.43,44 We also determined that major mediator proteins BRCA1, MDC1 and 53BP1 formed IR-induced foci irrespective of FLNA status. However, BRCA1 foci formation was significantly impaired in FLNA-deficient cells. In addition, Rad51 foci formation displayed a delayed kinetics in cells lacking FLNA. These data indicate that FLNA-deficient cells have impaired homologous recombination. Indeed, during the preparation of this manuscript Yue et al. showed that FLNA-deficient cells have a reduced ability to repair I-SceI-induced DSBs.45
During the course of our experiments we noticed a consistent increase in the number of FLNA- cells displaying IR-induced RPA foci. These foci progressively increased in size at later time points after IR. RPA is a ssDNA binding protein and participates in DNA metabolism processes where there is generation of ssDNA such as replication, repair and recombination.46,47 Phosphorylation leads to inability of RPA to associate with the replication centers and leads to the association with DNA damage-induced foci instead.48 Interestingly, lack of NHEJ proteins DNA-PKcs and Ku86, which together with Ku70 form the active DNA-PK complex, leads to accumulation of ssDNA in S phase.49 Thus, we further investigated how the lack of FLNA impacted on DNA-PK complex formation.
Remarkably, Ku86 failed to interact with DNA-PKcs in the absence of FLNA. The reduced stability of the interaction is not due to Ku86 failure to load onto chromatin, as FLNA- cells displayed sustained higher levels of chromatin-bound Ku86 than FLNA+ cells after damage. Ku86 is one of the first molecules to bind DNA ends after DSBs50 and recruits DNA-PKcs via its C-terminus.3 Taken together these results establish that lack of FLNA results in an unstable association of Ku86 and DNA-PKcs impairing the function of the complex. This impaired DNA-PK activity leads to a continuous build up of ssDNA and Ku86 on chromatin.
Over 16 phosphorylation sites have been identified in DNA-PKcs although their role is still poorly understood. Nevertheless, DNA-PKcs phosphorylation status is thought to influence its activity.51 DNA-PKcs interacts with Ku86 and free ends of DNA in an unphosphorylated form,52 and autophosphorylation is required for NHEJ progression.53 Thus, we investigated the status of the two major phosphorylation clusters in DNA-PKcs, namely the 2056 and 2609 clusters. Clusters 2056 and 2609 were consistently phosphorylated upon treatment with IR irrespective of FLNA status. The fact that DNA-PKcs is phosphorylated upon damage in the absence of FLNA suggests that DNA-PKcs is interacting with the Ku86/DNA complex albeit transiently. Alternatively, it is possible that phosphorylation of DNA-PKcs is not mediated by autophosphorylation at the synaptic complex but rather via hyperactive ATM and ATR in FLNA-deficient cells.
We showed that FLNA and BRCA1 interact and that FLNA deficiency leads to a marked decrease in BRCA1 foci formation after damage. To investigate further the role of BRCA1 we tested whether expression of the BRCA1 FLNA-interacting fragment in FLNA-proficient cells could also act in a dominant negative fashion leading to a phenotype similar to FLNA-deficient cells. Strikingly, expression of the BRCA1-Ff lead to a defect in DNA repair as judged by CDC2 pY15 and γ-H2AX markers. This effect is specific because expression of BRCA1-Ff containing a mutation that disrupts FLNA/BRCA1 interaction does not lead to the same phenotype. Taken together, these data establish that BRCA1 participates in the FLNA-dependent regulation of the DNA damage response.
Our data shows that absence of FLNA leads to defective DSB repair. The defect is a combined result of compromised HR and NHEJ processes. At this stage we cannot distinguish whether FLNA-deficiency leads to a defective step that is common to both pathways or, alternatively, it impacts different steps in these pathways. In fact, the interplay between these two arms of the DNA repair process is not fully understood,54 in particular after IR, which generates an array of different DNA modifications. The observed phenotype is consistent with a model in which Ku86 recognizes and binds free ends of DNA, but in the absence of FLNA, fails to make a stable complex with DNA-PKcs. We propose that unstable Ku86/DNA-PKcs interaction results in impaired end processing, accumulation of ssDNA, and hyperactivation of DNA damage signaling.
In addition, in FLNA-deficient cells BRCA1 displays impaired foci formation suggesting that FLNA also plays a role in stabilizing BRCA1 at the DSBs. BRCA1 colocalizes with Rad50/Mre11/NBS1 complex at IR-induced foci55,56 and inhibits Mre11 exonuclease activity.57 Thus, the diminished amounts of BRCA1 at IR-foci may lead to an unregulated Mre11 exonuclease activity with formation of the observed extended tracts of RPA-coated ssDNA in FLNA-deficient cells (Fig. 4C and D). BRCA1 has also been implicated in the regulation of Rad51,7,58 although the mechanism by which it happens is obscure.59 The kinetics of Rad51 foci formation in FLNA-deficient cells suggests that there is no problem in the initial recruitment to foci (see Fig. 4C, bottom, 3 h time point). The extended plateau observed in Rad51 foci (from 3 to 12 h after IR) may indicate an accumulation of DSBs that do not fulfill the end processing requirements for efficient Rad51 loading. Although further research will be needed to test this proposed model, it provides a tractable system to dissect the interplay between different processes involved in DNA repair.
It is possible that FLNA provides a framework for the assembly of factors in the synaptic complex. While unrepaired DNA in yeast (which lacks recognizable DNA-PKcs and FLNA orthologs) migrates to so-called DNA repair centers,60 the picture is different in mammalian cells where broken chromosome ends are essentially immobile.61,62 It will be interesting to determine whether lack of FLNA affects the mobility of broken ends.
GST-fusion fragments of BRCA1 in the mammalian expression vector pEBG BF 1-6 were a gift from Toru Ouchi. BRCA1 fragments BF1A (aa 1–70), BF1B (aa 71–140), BF1C (aa 1–101), BF1D (aa 141–240), BF1D1 (aa 160–190), BF1D2 (aa 190–210), BF1D3 (aa 160–210), BF1E (aa 241–324) and BF1F (aa 1–302) were obtained by PCR using pEBG BF1 as template (primer sequences are available upon request). The PCR products were digested and cloned into pEBG vector63 and sequenced. Construct BF1D Y179C was obtained by site directed mutagenesis using BF1D as template for the PCR reaction. FLAG FLNA-Bf was obtained by cloning a PCR fragment of FLNA (aa 2477–2647) in frame to FLAG in pCMV2-FLAG vector.
The FLNA-deficient M2 melanoma cell line and its isogenic cell line, A7, reconstituted with full length FLNA cDNA29 (gift from Thomas Stossel) was grown in MEM (Sigma) with 8% newborn calf serum (Sigma) and 2% fetal bovine serum (FBS; SAFC Biosciences, Lenexa, KS). A7 cells were grown in the presence of 0.2 mg/ml G418 (Fisher). HeLa (ATCC, Manassas, VA) was grown in DMEM with 5% FBS (Sigma). HCT116 (ATCC) was grown in McCoy's with 10% FBS. 293FT (InVitrogen) cells were grown in DMEM media (Sigma) with 10% FBS. Tissue culture media was supplemented with penicillin and streptomycin. Transfections were performed using Fugene 6 (Roche) according to the manufacturer's instructions.
The following antibodies, peptides and beads were used: α-BRCA1 mouse monoclonal antibody MS110 (Ab-1; Calbiochem; San Diego, CA) and SG11 (gift from Livingston D); α-Filamin A mouse monoclonal antibody PM6/317 (Chemicon International); α-FLAG M2 mouse monoclonal antibody (Sigma); 3xFLAG-peptide (Sigma); α-GST goat polyclonal antibody (Pharmacia Biotech); GT-sepharose 4B beads (GE Healthcare); α-Ku86 monoclonal antibody B-1 and α-Rad51 rabbit polyclonal antibody H-92 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA); α-TP53BP1 mouse monoclonal antibody and α-phosphoserine 343-NBS1 mouse monoclonal antibody (Upstate Biotechnology); α-p34 RPA mouse monoclonal antibody Ab-1 and α-DNA-PKcs mouse monoclonal antibodies Ab-2 (Neomarkers, Freemont, CA); α-phosphoserine 2056 DNA-PKcs rabbit polyclonal antibody (Abcam, Cambridge, MA); α-phosphoserine 2609 DNA-PKcs rabbit polyclonal antibody (Novus Biologicals); α-MDC1 (SIGMA); γ-H2AX rabbit polyclonal;α-H2AX; α-phosphoserine 1981 ATM; α-ATM; α-ATR; α-phosphothreonine 68 CHK2; α-phosphoserine 317 CHK1 (Cell Signaling); α-β actin (Sigma). Conjugates for immunofluorescence were Alexa fluor 488 or 555 Molecular Probes.
Whole cell extracts were prepared by lyzing cells in a mild RIPA buffer (120 mM NaCl, 50 mM Tris pH 7.4, 1% NP40, 1 mM EDTA, protease inhibitors, 4 mM PMSF) lacking harsher SDS, sodium deoxycholate, and Triton X-100 detergents. The same buffer was used for immunoprecipitation. For high stringency immunoprecipitations the RIPA buffer was supplemented with 0.5% SDS. Antibodies (1 μg) were pre-incubated with protein A/G agarose beads (Santa Cruz Biotechnology, Inc.,), washed twice in RIPA buffer and incubated with the cell extracts overnight at 4°C. After incubation, the slurry were pelleted by centrifugation (2,000 rpm) and washed twice by removing the supernatant. Sample buffer was added to the beads and boiled for 10 min. For GST-pull downs, cell extracts were incubated with GT-beads, washed in RIPA buffer, and boiled.
Samples for western blot analysis were separated by SDS-PAGE and gels were electroblotted on a wet apparatus to a PVDF membrane. The PVDF membrane was blocked with 5% milk in TBS buffer containing 0.1% Tween (TBS-Tween) for 1 h. The membrane was washed three times in TBS-Tween and the antibody was added in 0.5% milk in TBS-Tween. The membrane was washed three times in TBS-Tween and incubated with the appropriate conjugate. After final washes Blots were incubated with ECL (Millipore, Billerica, MA).
Chromatin fractions were obtained by lyzing the cells with mild RIPA buffer and centrifuging at 14,000 rpm for 5 min. The pellet was then washed twice in mild RIPA and extracted with acid extraction buffer (0.5 M HCl, 10% Glycerol, 100 mM BME) and subsequently neutralized using 40 mM Tris pH 7.4 with protease inhibitors and NaOH.
Western blot data was quantified by densitometry using AlphaEaseFC v 3.1.2. Each lane was normalized using the corresponding loading controls and then expressed as a fold change relative to the untreated FLNA+ cells in each blot.
For BRCA1 analysis cells were fixed with 4% formaldehyde for 5 min followed by 5 min incubation with 100% ethanol. Cells were permeabilized with 0.25% Triton X-100 in PBS for 10 min, washed with PBS, and then blocked for 30 min with 5% BSA in PBS at room temperature (RT). After blocking, BRCA1 monoclonal antibody (SG11; kind gift from David Livingston) was added to 1% BSA in PBS for 1 h RT. Cells were washed and goat α-mouse Alexa Fluor 488 (Molecular Probes) was added for an additional 1 h RT.
For all other antibodies, cells were plated onto chamber slides and after 24 h they were washed with cytoskeleton buffer (10 mM HEPES/KOH pH 7.4, 300 mM sucrose, 100 mM NaCl, 3 mM MgCl2) and fixed with 4% formaldehyde for 30 min RT. For analysis of chromatin bound RPA cells were pre-extracted for 2 min on ice with cytoskeleton buffersupplemented with 0.5% Triton X-100 before fixation.64 After fixation cells were permeabilized with 0.25% Triton X-100 in PBS for 5 min RT and then washed and blocked with 5% BSA in PBS for 30 min RT. Primary and secondary antibodies in 1% BSA in PBS were added for 1 h each.
Cells were washed and mounted with Prolong Gold medium (Molecular Probes). Images were taken on a Leica Confocal Microscope. For quantification of BRCA1 and Rad51 immunofluorescence foci approximately 100 cells were scored per each time point. Cells were scored as foci-positive if they presented with more than 10 foci per cell (an example can be found in Supp. Fig. 2A). For γ-H2AX and NBS-P-343 at least 50 cells per time point were counted for each condition. Cells were scored as foci-positive if they presented with more than 20 foci per cell. A threshold of 20 foci was chosen based on the number of foci found in unirradiated samples using the described antibodies. Determination of foci number per cell was done using Definiens Developer XD 1.1 (Definiens AD, Germany). A rule set was developed to segment nuclei based on the DAPI stain and then segment foci within the nucleus based on an intensity threshold. Representative results from at least two independent experiments are shown instead of statistical data on a small number of measurements with variability as recently recommended.65
Comet assays were performed in neutral conditions using a comet assay kit (Trevigen, Gaithersburg, MD) according to manufacturer's instructions. Briefly, cells were collected at the indicated time points, combined in low melting agarose (Trevigen, Gaithersburg, MD), spread over the comet slide area and allowed to set. Then, slides were immersed in lysis buffer for 30 min at 4°C. Electrophoresis was run in TBE buffer for 20 min at 1 V/cm voltage. Image analysis was done with Comet Analysis System 2.3.3 software (Loats Associates Inc., Westminster, MD).
The authors are indebted to members of the Monteiro Lab for helpful discussions and to Marcus Smolka and Ed Seto for a critical reading of the manuscript, to David Livingston and Arcangela De Nicolo for a generous gift of SG-11 antibody; Thomas Stossel for the M2 and A7 cell lines; Junjie Chen for the HCC1937-BRCA1wt cells; and Toru Ouchi for GST-BRCA1 constructs. This paper is dedicated to the memory of Hidesaburo Hanafusa, Professor Emeritus of the Rockefeller University, who passed away on March 15, 2009.
Financial disclosure: This work was supported by a predoctoral fellowship [BC083181] to A.V. from the US Department of Defense Breast Cancer Research Program; a National Institutes of Health grant [CA116167]; a grant from the Florida Breast Cancer Coalition Foundation to A.M.; and supported in part by the Analytic Microscopy and the Molecular Biology cores at the H.L. Moffitt Cancer Center & Research Institute.
Note: Supplementary materials can be found at: www.landesbioscience.com/supplement/VelkovaCC9-7-Sup.pdf