Long non-coding RNA PARTICLE bridges histone and DNA methylation

PARTICLE (Gene PARTICL- ‘Promoter of MAT2A-Antisense RadiaTion Induced Circulating LncRNA) expression is transiently elevated following low dose irradiation typically encountered in the workplace and from natural sources. This long non-coding RNA recruits epigenetic silencers for cis-acting repression of its neighbouring Methionine adenosyltransferase 2A gene. It now emerges that PARTICLE operates as a trans-acting mediator of DNA and histone lysine methylation. Chromatin immunoprecipitation sequencing (ChIP-seq) and immunological evidence established elevated PARTICLE expression linked to increased histone 3 lysine 27 trimethylation. Live-imaging of dbroccoli-PARTICLE revealing its dynamic association with DNA methyltransferase 1 was confirmed by flow cytometry, immunoprecipitation and direct competitive binding interaction through electrophoretic mobility shift assay. Acting as a regulatory docking platform, the long non-coding RNA PARTICLE serves to interlink epigenetic modification machineries and represents a compelling innovative component necessary for gene silencing on a global scale.

The majority of ribonucleic acids synthesized from the human genome represent long non-coding (lnc) transcripts greater than 200 base pairs. Such lncRNAs have lower expression and greater tissue-specificity compared to messenger RNAs, suggestive of their putative regulatory function 1 . LncRNA plasticity mainly contributes to their capability to interact with diverse biomolecules (DNA, RNA or protein) 2 . This poses a challenge for deciphering the protagonists implicated in lncRNA activities essential for cellular complexity and phenotypic determination.
The lncRNA PARTICLE operates an active feedback silencing mechanism upon the putative tumor suppressor MAT2A to limit its expression rapidly once MAT2A is up-regulated in response to low dose radiation 3 . PARTICLE (1432 bp) is transcribed in the antisense direction from within the promoter of MAT2A, the product of which encodes the catalytic subunit of methionine adenosyltransferase 4 . PARTICLE triplex formation has been demonstrated in vitro in the 'shore' region of a MAT2A promoter CpG island, with evidence found that this lncRNA leads to increased DNA methylation and binds to the Polycomb Repressive Complex 2 (PRC2) subunit Suppressor of Zeste 12 (SUZ12) 3 .
It has emerged that SUZ12 is key for locating the PRC2 catalytic subunit responsible for trimethylation (me3) of histone 3 at lysine 27 (H3K27) during heterochromatin formation 5 . PRC2 also harbors a control module preventing deposition of H3K27me3 on transcriptionally active genes 5 . It has been suggested that focused activity of epigenetic modifiers such as PRC2 and the histone code influence the propensity of an individual gene to become hyper-methylated in malignant tissue, contributing to inactivation of tumor suppressor genes 6,7 . Histone point mutations can hamper H3K27me3 deposition leading to adverse events such as those implicated during aberrant differentiation of mesenchymal stem cell to skeletal tumorigenesis 8 . Unable to target genomic regions Results PARTICLE and low dose irradiation act synergistically to enhance the H3K27me3 modification. A histone 3 lysine 27 trimethylation (H3K27me3) ELISA revealed augmentation of the H3K27me3 repressive mark within 2 hr post transfection for PARTICLE over-expression compared to lipofectamine only controls (LF) (11.9 fold increase; p = 0.0074, Fig. 1A). The H3K27me3 modification level was further substantially augmented in OE versus LF (133.3 fold increase; p = 0.0079) by 24 hr with reduction in this modification at the 48 hr time point following 0.025 Gy irradiation exposure (Fig. 1A). Of note, upregulation of endogenous PARTICLE does not occur post 2 hr to 48 hr after such irradiation dosage (manuscript under consideration). This enabled the independent effects of 0.025 Gy irradiation on H3K27me3 to be assessed. Of interest, when PARTICLE over-expressing cells were irradiated, H3K27me3 profiles were even more elevated when compared to irradiated LF at 2 hr (23.0 fold increase; p = 0.0003) and 24 hr (239.9 fold increase; p = 0.0024) (Fig. 1A). Western blotting and immunofluorescence analysis revealed a global increase in the H3K27me3 heterochromatin repressive modification in PARTICLE over-expressing (OE) cells relative to controls (Fig. 1B,C). A synergistic escalation in this histone modification was apparent in OE 24 hr post very low dose irradiation (0.025 Gy) (Fig. 1B,C). PARTICLE has been found to increase expression of EZH2 (Enhancer of Zeste homolog 2), the PRC2 component which catalyzes the addition of methyl groups to histone H3 at lysine 27 ( Fig. S4) 15 .
Based on the observation that histone modifications tend to cluster to form domains, a spatial clustering method for the identification of ChIP-enriched regions (SICER) 16 was utilized to identify signals unlikely to appear by chance. This approach was adopted for the identification of H3K27me3 enriched domains within ChIP-seq datasets. MDA-MB-361 WT (lipofectamine transfected) and PARTICLE OE cells 24 hr post exposure to very low irradiation (0.025 Gy) were directly compared. Having passed the quality and purity filter screen (Fig. S1) clear differences emerged. This revealed 24,946 genomic regions with significantly increased H3K27me3 modification in irradiated OE versus WT (Fig. 1D,E). An assessment of average read per million (RPM) signal values was undertaken as quantification of H3K27me3 enriched regions does not typically deliver bell-shaped, symmetrical peaks 17 . The genome-wide distribution patterns of H3K27me3 in irradiated OE versus WT differed considerably with notable increased or decreased presence of this modification in distal intergenic regions or 5′ untranslated regions respectively of the human genome following PARTICLE overexpression (Fig. 1F).

PARTICLE alters both the local and global distribution of H3K27me3 in the human
genome. An integrative genomics viewer 18 enabled virtual H3K27me3 positioning across the human genome to be visualized. ChIP-seq tracking information for irradiated WT and OE revealed an enhancement of this modification in the latter within all autosomal chromosomes. Interestingly, the X-chromosome appeared to have further enrichment of the H3K27me3 modification upon the overexpression of PARTICLE ( Fig. 2A).
Analysis identified that MAT2A and WWOX genomic loci (as well as others) were subjected to H3K27me3 repression (Fig. 2B). PARTICLE has been reported to increase the methylation of the CpG island 108368 (Chr. 2: 85765695-85766983; NCBI Homo sapien build number 37 version 2) that surrounds the MAT2A transcription start site (chromosome 2: 85766100) 3 . ChIP-seq findings revealed a considerable shift in the position and intensity of the H3K27me3 signal upstream of this CpG island (a region of established PARTICLE triplex formation 3 ) and the MAT2A promoter (Fig. 2B). H3K27me3 enrichment along a 1.1 Mb stretch was also found spanning the majority of the WWOX locus at two consensus motifs on chromosome 16 in irradiated PARTICLE overexpressing (OE) cells versus wild type lipofectamine -only controls ( Fig. 2B-E). These results reveal that PARTICLE over-expression enhances the histone repressive modification mark across the human genome and specifically within MAT2A and WWOX tumor suppressor genes.
A predominance of PARTICLE triplex binding sites throughout WWOX has been identified (manuscript under consideration). Enriched H3K27me3 clustering domains from OE samples were merged with Triplex Domain Finder (TDF) in silico data for predicted PARTICLE triplex sites within the human genome. This revealed that the PARTICLE 627-646 bp domain had significantly higher potential to bind the target H3K27me3 modified domains than randomly chosen similar sized regions from the human genome (806 regions versus 380 regions respectively, p = 0.00001). This would suggest that H3K27me3 modifying enzymes might be guided to specific PARTICLE triplex sites to exert their function. INGENUITY integration of ChIP-seq H3K27me3 data with TDF evidence for PARTICLE triplex binding sites revealed significant associations with molecular functions eg amino acid metabolism (−log(p-value) = 3.6) and diseases eg cancer (−log (p-value) = 5.2) (Fig. S2).
Analysis of the dynamics of DNMT_RFP showed an exponential decline from 160.35 ± 10.47 to 70.87 ± 10.0 arbitrary units (AU) in the nuclear signal (R 2 = 0.91) over the 2 hr recording period (up to 60 min shown in Fig. 3F). Lower DNMT_RFP signal intensity was evident in the cytosolic/extranuclear compartment, and this also diminished during this time period (10.14 ± 2.0 to 8 ± 0.9 AU, R 2 = 0.7) (Fig. 3F)  DFHBI-1T addition, dbPARTICLE signal intensity increased in the nucleus from 35.86 ± 9.0 to 80.2 ± 19 AU with a later reduction to 60.15 ± 4.5 AU until 120 min. Under the same conditions, dbPARTICLE gradually increased from an intensity level of 19.4 ± 0.5 to 40 ± 11.2 arbitrary units (2 fold increase) in the cytosol within 120 min (up to 60 min shown in Fig. 3G). This finding demonstrates transport of dbPARTICLE from the nucleus to the cytosol during this time period. The dbPARTICLE was co-localized with DNMT_RFP in the nucleus within 10 min (χ 2 = 0.87 ± 0.03; Fig. 3C and E). At this time point DNMT_RFP was not solely associated with dbPARTICLE (χ 2 = 0.66 ± 0.02; Fig. 3C and E) as noted by the predominant red signal intensity profiling in the nucleus. By 30 min DNMT_RFP and dbPARTICLE showed a strong co-localization in the nucleus (approx. χ 2 = 0.8; Fig. 3C and E). Independent dbPARTICLE in the nucleus became evident after 60 min with diminishing co-localisation signal with time ( Fig. 3C and E).
PARTICLE is implicated in global methylome enhancement, WWOX CpG island methylation and enzyme activity with DNMT1 interaction. Global methylome measurements were quantified in MDA-MB-361 over-expressing PARTICLE in comparison to lipofectamine only or negative control (NC1) transfected cells (6.1 ± 1 fold increase in the percentage of 5-methylcytosine (% 5-mC), p < 0.05) (Fig. 4A). Exposure of these cells to 0.025 Gy 24 hr previously, demonstrated an even higher degree of global 5-methylcytosine reaching 1 ± 0.02% (p = 0.01) with a synergistic enhancement from irradiation under the same conditions, especially in the presence of PARTICLE overexpression (16.29 fold increase compared to LF or negative control, p < 0.05) (Fig. 3A). When compared to controls no significant alteration was found in 5-methylcytosine levels upon PARTICLE knockdown or in combination with previous exposure 24 hr earlier to a very low radiation level (Fig. 3A).

CpG island
CpG island   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21   PARTICLE was found to influence the methylation status of a WWOX promoter CpG island (annotated CpG105476) of 990 bp located on chromosome 16: 78133076-78134066 (NCBI homo sapiens build number 37/hg19). The transcription initiation site for WWOX resides within this region at position chromosome 16: 78133327 orientated in a forward direction (NCBI gene id. 51741). The CpG105476 was hyper-methylated by 13.3 ± 0.9% with the remainder being predominantly unmethylated in sham irradiated lipofectamine transfected MDA-MB-361 cells (LF) (Fig. 3B). Following exposure to a very low irradiation level (0.025 Gy) 24 hr previously, the level of hyper-methylation of CpG105476 increased 1.38 fold to 31.83 ± 3% reflecting methylation events independent of endogenous PARTICLE which is not activated after this dosage (Fig. 3B). Nevertheless, in MDA-MB-361 with PARTICLE knockdown, the extent of basal CpG10576 hyper-methylation is reduced to 6.25 ± 4% yet increased to 18.19 ± 2% after 0.025 Gy exposure, perhaps reflecting an endogenous influence of this lncRNA on the methylation status of this WWOX CpG island. Of interest, over-expression of PARTICLE augmented the basal hyper-methylation status of CpG105476 by 4 fold (66.67 ± 7%). Exposure to a very low level of radiation resulted in a further escalation of 5 fold (80.11 ± 3%).
DNMTs transfer methyl groups from s-adenosylmethionine to cytosine to methylate DNA substrate enabling the methylated DNA to be recognized with a 5-methylcytosine antibody. The quantity of methylated DNA which is proportional to enzyme activity was measured thorough ELISA with the activity of the DNMTs proportional to the optical density intensity with time. The significant influence of PARTICLE on DNA methyltransferase activity was subsequently revealed providing evidence for the relevance of their interaction. DNMT activity did not differ between controls (p > 0.05) or with PARTICLE knockdown (p = 0.19) (Fig. 3C). A significant 1.5 ± 0.3 fold increase in DNMT activity was noted when PARTICLE (p = 0.011) or dbroccoli-PARTICLE (p = 0.013) were over-expressed in comparison to cells transfected with the negative controls (NC1 or lipofectamine only) (Fig. 3C). Augmentation of DNMT activity was noted in MDA-MB-361 exposed 24 hr previously to a very low irradiation dose (0.8 ± 0.1 fold increase, p = 0.035) (Fig. 3C). Combining PARTICLE or dbroccoli-PARTICLE overexpression with such irradiation exposure (0.025 Gy) further significantly increased DNMT activity (2 ± 0.4 fold increase, p = 0.022 and p = 0.034 respectively) relative to negative controls (Fig. 3C). Flow cytometric analysis of MDA-MB-361 transfected with dbPARTICLE and DNMT_RFP also demonstrated their association within 20 min of DFHBI-1T addition with independent dbPARTICLE similarly evident by the 2 hr time point (Fig. 3D). These findings were supported by the association of dbPARTICLE with DNMT1 by competitive interaction and gel retardation. An electrophoretic mobility shift of the recombinant human Dnmt1 protein was identified in the presence of biotinylated dbPARTICLE plus PARTICLE and absent with negative control transcript (Fig. 3E,F). Binding specificity was demonstrated by displacement of the amount of Dnmt1-bound biotinylated dbPARTICLE with increasing the concentration of PARTICLE (up to 1000 fold molar excess) (Fig. 3F). These data were further supported by immunoprecipitation of DNMT1 with PARTICLE in crosslinked cells proving direct interaction of this lncRNA and this DNA methytransferase (Fig. 3G).
Overall, evidence is provided of direct PARTICLE DNMT1 interaction that boosts methyltransferase activity levels. Such interaction potentially unveils part of the underlying mechanism by which this lncRNA impacts the global methylome with its notable influence on the CpG island methylation status of the tumor suppressor WWOX.

Discussion
Most recently discovered lncRNAs prevail in the nucleus and tend to be involved in epigenetic regulation 1 . This milieu of lncRNAs either associate with histone or DNA methylation mechanisms 20 . Such double pronged silencing approaches have not been found for a single lncRNA until now. Herein, evidence is provided that PARTICLE represents the missing link that unites both histone and DNA methylation. PARTICLE is sufficient to affect histone H3K27me3 (via influencing EZH2 expression) throughout the human genome and to enhance this repressive modification mark within the neighbourhood of both the MAT2A and WWOX tumor suppressor genes. Our findings establish direct interaction between PARTICLE and the maintenance DNA methyltransferase DNMT1 coinciding with increased enzyme activity, a global shift in the methylome and an upsurge in WWOX CpG island methylation.
The unique ability of lncRNAs to adopt complex secondary and tertiary structures contributes to their greater functional complexity compared to mRNA 21,22 . While proven for proteins, it is now becoming evident that structure-functional relationships provide important information on lncRNA characteristics for mediating biomolecule interaction in such diverse processes as chromatin organization and transcriptional regulation 22 . PARTICLE binds to both the lysine methytransferase G9a and to the SUZ12 component of the polycomb repressive complex 2 (PRC2) 3 . Following low dose irradiation, it could be proposed that PARTICLE provides a functional targeting platform enabling specific targeting of otherwise promiscuous repressive modifiers such as PRC2 to chromatin. In keeping with the recognized role of lncRNAs in genomic architectural regulation 23 and given the interaction between PARTICLE and SUZ12 3 , it is tempting to speculate that PARTICLE also acts as an epigenetic modifying platform, in this case for PRC2 recruitment to target sites to modulate chromatin structure. Other lncRNAs such as Xist and Firre (functional, intergenic, repeating RNA element) depend on increased H3K27me3 levels for inactivating and maintaining X chromosome repression 24,25 . ChIP-seq evidence reported in this study shows H3K27me3 to be generally enhanced throughout the human genome and specifically increased throughout the WWOX locus when PARTICLE is over-expressed. This data also revealed a shift in the positioning of this modification over the MAT2A CpG island, the methylation and expression of which is influenced by PARTICLE (Fig. S3) 3 .
Temporary H3K27me3 marks associated with CpG rich genomic regions become replaced by DNA methylation representing a more permanent means of transcriptional repression 26 . Here we show the direct interaction of PARTICLE and DNMT1. RNA dependent DNA methylation suggests recognition by DNMT and perhaps recruitment to RNA: DNA: DNA triplexes 27 such as that formed by PARTICLE genome wide (manuscript under consideration). Evidence is presented for enhanced WWOX CpG island hyper-methylation with PARTICLE over-expression along with synergistic augmentation of this gene repressive modification by very low radiation exposure. Diminished hyper-methylation of this CpG island upon PARTICLE knockdown may reflect its inhibitory influence on DNMT1 action with evidence provided here that this lncRNA directly effects the activity of DNA methyltransferases.
Epigenetic mechanisms that incorporate histone modifications, DNA methylation alterations and non-coding RNA expression have been identified as prominent hallmarks for distinguishing physiological from pathological cellular conditions, including tumor suppressor inactivation 28,29 . Thus, PARTICLE may constitute an important bond in the internal crosstalk of the broad language of epigenetics, orchestrating transcriptional silencing of genes including tumor suppressors with wider implications for eliciting carcinogenesis and progression.
PARTICLE, a long non-coding RNA, is transcribed in response to irradiation and enables histone modification and DNA methylation to be interwoven. These mechanisms were considered to be relatively independent until now. Operating a double pronged approach enabling methylome repression, PARTICLE serves to interlink epigenetic modification mechanisms and represents a compelling innovative component necessary for quelling gene transcription.

Materials and Methods
Propagation and maintenance of cell lines. MDA-MB-361 (American Type Culture Collection (ATCC)) was cultivated as previously described 30 . U2OS (ATCC) were grown under similar conditions except Roswell Park Memorial Institute (RPMI) 1640 media (GibcoTM cat # 21875-034) and FBS (10%) were utilized. The identity of all cell lines was verified by microsatellite analysis (Eurofin Medigenomix, Forensik GmbH, Germany). All cultures were routinely checked for mycoplasma contamination using a MycoAlert Mycoplasma detection kit (Lonza, cat. # LT07-218). In general, cells were grown to 80% confluency prior to removal from the dish using trypsin (0.25%)/EDTA (0.02%) and sub-culturing or harvesting.

PARTICLE overexpression. PARTICLE was cloned into the pGEM ® -T vector (p.PART) (GenScript)
and transformed into Top10 bacteria. A colony was grown in ampicillin (100 μg/ml) overnight and plasmid midiprep (Promega) performed. Plasmid concentration and purity was assessed (NanoDrop 1000, Thermo Fisher Scientific) with A260/280 ratio determination with automated sequence validation (GenScript). Plasmid linearization was carried out using 1 μg plasmid DNA and SacI overnight digestion at 37 °C. PARTICLE (1432 bp) was in vitro transcribed from a pGEM ® -T vector (GenScript) using the TranscriptAid T7 High Yield transcription kit (Thermo Scientific, cat # K0441). Transcripts were treated with RNase-free DNase 1 (Thermo Scientific) and purified using an RNeasy mini-elute cleanup kit (Qiagen, cat # 74204) and verified by TBE-agarose (1.8%) electrophoresis. Prior (24 hr) to transfection, MDA-MB-361 or U2OS were seeded (10 5 cells/35 mm dish) in growth media (described above) in the absence of antibiotic/anti-mycotic to ~70% confluence at the time of transfection. The control template included in the Transcript T7 High Yield Transcription kit (Thermo Fisher Scientific, cat # K0441) as utilised for the production of a 2223 bp 'run off ' transcript serving as a negative control (NC1) for over-expression studies. Cells were transfected with lipofectamine and PARTICLE (4 μg) or negative control (4 μg) as per standard conditions with incubation for 72 hr prior to irradiation exposure. Histone 3 lysine 27 trimethylation (me3) immunofluorescence. MDA-MB-361 were cultivated as previously described 3 on glass coverslips. Having reached ~60% confluence, the media was removed and cells washed two times for 5 min with 1× PBS. Cells were fixed upon exposure to 4% paraformaldehyde for 1 hr and washed for 5 min with 1x PBS. Cells were permeabilized in 1x PBST (1x PBS including 0.5% Triton ™ X-100 (Sigma-Aldrich ® , cat. # X100-5ML) for 30 min. Following one wash for 5 min in 1x PBS, cells were placed in blocking solution (1x PBS containing 2% goat serum, 5% bovine serum albumin and 0.5% Triton ™ X-100) for 1 hr at room temperature. Cells were then exposed to antibody representing rabbit anti-tri-methyl-histone 3 (Lys27) (Thermo Fisher Scientific cat. # PA5-31817, 1: 200 in blocking solution) with o/n incubation at 4 °C.
Cells were washed three times for 15 min in 1x PBS and incubated in Alexa fluor ® 488 goat anti-rabbit IgG (H + L) (1:500; in blocking solution) for 1 hr at room temperature in the dark. Cells were washed three times for 15 min with 1x PBS and air dried in the dark. To prepare for microscopy, cells on coverslips were mounted in VECTASHIELD ™ HardSet ™ (Vector; cat. # H1500) containing DAPI, and placed on a glass slide. Results were visualized using an epifluorescence microscope (Zeiss AxioVision).
Nuclear Isolation. Nuclei were isolated from cell lines (U2OS and MDA-MB-361) using a nuclear extraction kit (Millipore cat # 2900). In brief, cells were grown to 70-90% confluency and removed by trypsinization following standard protocols. Cell pellets (2 × 10 6 cells) were resuspended in cytoplasmic lysis buffer (500 μl) containing 0.5 mM DTT and protease inhibitor cocktail (1 in 1000 dilution) with incubation on ice for 15 min. Following centrifugation at 250 g for 5 min at 4 °C, the supernatant was discarded and cell pellet resuspended again in cytoplasmic lysis buffer (200 μl). Cell lysis was performed by drawing the cell suspension through a 27-gauge needle. Following centrifugation at 8,000 g for 20 min at 4 °C, the nuclear pellet was resuspended in nuclear extraction buffer (70 μl) containing 0.5 mM DTT and protease inhibitor cocktail (1 in 1000 dilution). Nuclei were disrupted via passage through a 27-gauge needle and incubation for 60 min at 4 °C. Following centrifugation at 16,000 g for 5 min at 4 °C, the supernatant representing the nuclear extract was obtained.
ELISA H3K27me3 quantification. This procedure utilized the histone 3 (tri-methyl K27) quantification fluorometric kit (abcam ® , cat. # ab115073). In brief, a standard curve was established using a kit included standard control (H3K27me3, 100 μg/mL) diluted with antibody buffer to provide a concentration range from 1.5-100 ng/μl. Histone extract (100 ng) was deemed optimal for determining modification levels within the standard curve range. Antibody buffer (50 μl) was added to standards and samples in a 96 well transparent dish (Thermo Fisher Scientific ™ , cat. # 168136). The plate was covered with parafilm, mixed very gently and incubated for 2 hr at room temperature. Wells were aspirated and washed three times with 1x wash Buffer (150 μl). Detection antibody (1:1000, 50 μl) was added to each well and incubated for 60 min at room temperature on an orbital shaker at 0.7 × g. Wells were aspirated and washed with 1x wash buffer (150 μl, X 6). Away from light, pre-mixed fluoro-development solution (50 μl) was added into each well and incubated for 5 min at room temperature. Fluorescence signal was measured at an excitation wavelength of 530 nm and an emission wavelength of 590 nm using a Tecan Infinite ® M200.

Chromatin immunoprecipitation. MDA-MB-361 cells treated with lipofectamine (WT) and/or
PARTICLE over-expression were exposed to 0.025 Gy (as described above). After 24 hr 1 × 10 8 cells were formaldehyde cross linked according to the published protocol 31 . Cells were sonicated using the following conditions: sonication level = 3, duration time = 5 min with 30 seconds on/off mode, without probe contact with the 5 ml eppendorf tube (Sonfier B-12; Branson sonic power company). Genomic DNA fractionation was assessed by 1.5% TBE/agarose gel electrophoresis. Dynabeads sheep anti-rabbit IgG (Life Technologies, cat # 11203D) were placed in blocking solution (1x PBS including bovine serum albumin (0.5%)) at 4 °C and collected on a magnetic rack. Following washes (X 20) in blocking solution, beads were re-suspended in blocking solution (250 μl) containing rabbit anti-tri-methyl-histone 3 (Lys27) (anti-H3K27me3 (10 μg), Thermo Fisher Scientific cat. # PA5-31817) with o/n incubation on a rotator at 4 °C. Anti-H3K27me3/dynabead mix was added to the sonicated cell lysate with o/n incubation at 4 °C. Further washing, elution and crosslinking reversal were performed as previously described 31 . DNA purification was carried out using a Maxwell ® 16 LEV Blood DNA kit (Promega cat # AS1290) and Maxwell ® 16 machine. DNA quantification was determined using a Qubit ® dsDNA high sensitivity assay kit (Invitrogen) with integrity (quality and fragment distribution) tested on a 2100 Bioanalyzer (Agilent Technologies). Illumina TruSeq ChIP library preparation and data analysis was outsourced to IMGM laboratories (Munich, Germany) and Active Motif (CA, USA) respectively. ChIP-seq track comparison was performed using the Integrative Genomics Viewer online software from the Broad Institute 18 .
Analysis of DNA methylation status of the CpG island in the WWOX promoter. Genomic DNA was isolated from MDA-MB-361 (conditions: LF, OE, NC1, KD, NC2) 24 hr post irradiation (0.025 Gy) or sham-irradiation. Genomic DNA was digested using methylation-dependent and sensitive restriction enzymes using an EpiTect Methyl II DNA restriction kit (Qiagen, cat. # 335452) as per manufacturers' instructions and assessed by SYBR green real-time PCR detection using the EpiTect Methyl II PCR primer assay for WWOX (CpG island 105476): (Qiagen, cat. # EPHS105476-1A). To accurately measure the relative percentage of unmethylated and methylated DNA within the CpG island region an online analysis tool (http://www.sabiosciences.com/ dna_methylation_data_analysis.php) was utilized.