Introduction

DNA damage induced by exogenous or endogenous factors is repaired by various DNA repair pathways, which are also components of the cellular DNA damage response (DDR) network. If damaged DNA cannot be repaired efficiently and accurately, it may cause mutations in the genetic material and lead to genomic instability. Such changes may cause alterations to the normal functioning of the cell by disrupting cell cycle, or the induction of apoptosis, and may result in the initiation and promotion of cancer development [1, 2]. Long noncoding RNAs (lncRNAs) have been known for many years, but their potential involvement in the DDR, and more specifically in DNA repair, is only now emerging. It has been reported that a certain number of lncRNAs play important roles in diverse biological pathways [3, 4]. In particular, it has been suggested that lncRNAs are engaged in the cellular DDR across a range of different signaling pathways it encompasses. For instance, lincRNA-p21 [5, 6], LincRNA-RoR [7], PANDA [8], DINO [9], and Meg3 [10] are all induced by p53 upon DNA damage and interact with various different proteins to mediate cell cycle arrest and apoptosis; two important end points of the DDR.

Since DNA double-strand breaks (DSBs) are the most lethal forms of genetic damage in cells [11, 12], the repair of DSBs has been deeply investigated and broadly evaluated. Multiple proteins are engaged in two major repair pathways to restore damaged DNA sites, thus preventing chromosomal aberrations and cell death. The nonhomologous end-joining (NHEJ) pathway operates throughout the cell cycle, but is crucial within the G0 and G1 cell phases when DSBs are recognized and repaired by NHEJ.

During DSB repair by NHEJ, Ku70 plays a key role in conjunction with Ku80 to form the Ku70–Ku80 heterodimer. This complex has a preformed ring-like conformation that senses and binds the double-stranded ends of DSBs [13, 14]. Repair is initiated following the recruitment of the DNA-dependent protein kinase catalytic subunit (DNA-PKcs), which binds Ku separately. DNA-PKcs recruitment stabilizes Ku complex binding at the DSB and is the protein kinase responsible for signaling the presence of the DSB, by phosphorylating the histone variant H2AX to form gamma H2AX. This process marks the presence of the DSB and serves to recruit other DNA repair factors such as XRCC4-like factor (XLF), X-ray cross-complementing protein 4 (XRCC4), which together manifest as the repair foci [11, 15].

It was recently reported that in triple-negative breast cancer (TNBC) cell lines, lncRNA LINP1 is involved in DSB repair. They demonstrated that this lncRNA promotes efficient DSB repair possibly via interacting with Ku and DNA-PKcs where it serves as a scaffold [16]. Overexpression of LINP1 in these cells leads to significantly enhanced rates of NHEJ that causes the drug resistance to chemo and radiotherapy in TNBC cells. It is established that during the recognition of DSBs, the Ku heterodimer is able to bind different types of DSBs, including those with overhangs or blunt ends [17]. It is also known that the Ku heterodimer can bind with the RNA component of telomerase and assists in recruiting telomerase to telomeres [18,19,20]. Finally, we noted that prolonged binding of DNA repair factors at the sites of DSB’s can trigger the DDR. Conceivably, cells coordinately regulate the transcription of specific lncRNAs, which mediate genome stability via the repair of DSBs by the NHEJ pathway, in response to DSB-inducing agent treatment.

Here, we report that repair of DSBs is regulated and enhanced by the damage-induced expression of a novel lncRNA we refer to as lncRNA interacting with Ku (LRIK, NONHSAG044790.2). Using a custom-designed microarray comprised of around 2000 uncharacterized lncRNAs, we screened for the presence of differentially expressed lncRNAs in response to cells treated with the DSB-inducing agent methyl methanesulfonate (MMS). LRIK interacts with the DSB recognition factor Ku heterodimer and enhances cellular resistance against DSB-inducing agents such as MMS, X-rays, and phleomycin. We also demonstrate that knockdown of LRIK significantly reduces the occupancy of Ku and its downstream factors on damaged chromatin. Depletion of LRIK reduces the efficient formation of DSB repair foci in a similar fashion to knockdown of Ku70 and Ku80, without affecting the induction of DSBs. Failure to sense the DSBs results in the impaired formation of γH2AX foci, reducing efficient DSB repair. Our data demonstrate that the interaction of LRIK with Ku enhances the binding of DSBs, promoting the efficient repair of damage by the NHEJ pathway.

Material and methods

Cell culture and transfections

HeLa, A549, and 293T cells were cultured in Dulbecco’s Modified Eagle’s Medium (Gibco) supplemented with 10% (v/v) Fetal Bovine Serum (Gibco) and 1% (v/v) Penicillin/Streptomycin (Gibco). MDA-MB-436 cells were cultured in RPMI 1640 medium (Gibco) supplemented with 10% (v/v) Fetal Bovine Serum (Gibco) and 1% (v/v) Penicillin/Streptomycin (Gibco). EJ5-GFP-U2OS and DR-GFP-U2OS cells were kind gifts of Prof. Xianlu Zeng (Northeast Normal University, China). All cells were tested for mycoplasma contamination by PCR. Human cells were transfected with reconstructed or unmodified plasmids by using Lipofectamine 2000 (Life Technologies) according to the manufacturer’s instructions. Cells with a stable expression plasmid were selected by culture media with 1 µg/ml puromycin (Life Technologies).

Microarray samples and data analysis

Total RNA samples were extracted from untreated HeLa cells or treated with 0.005% MMS followed by 12 h of recovery. Microarray probes were designed against 1958 noncoding RNAs (ncRNAs). The custom-designed microarray platform was manufactured by Agilent with ID GPL23062 (Agilent-026619 Human ncRNA and mRNA array), consisting of probes for 1860 protein coding genes and 1958 lncRNA transcripts. LncRNA transcripts were collected from a number of different sources including our previous findings [21], unpublished data, as well as our database NONCODE [22] and so on. At least one probe was designed for each lncRNA transcript. Microarray sample preparation, hybridization, and data analysis were conducted as described previously [21]. The array data were subject to quality control and background subtraction by the GeneSpring software. Quantile normalization was modulated by limma package. The log2 fold change threshold, which defines significant transcription alteration in tested samples, was set to 1.5. MA plot was drawn with the R software.

Accession number

The microarray data related to this study have been deposited to the NCBI Gene Expression Omnibus with the accession code GSE94868.

RNA extraction, reverse transcription, and quantitative RT-PCR

Total RNA was extracted using TRIzol agent (Invitrogen) following the manufacturer’s instructions. As for cytoplasmic and nuclear RNA extraction, cells were fractioned with the NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific) to separate each fraction followed by isolating RNA with TRIzol agent. RNA was subjected to synthesize the first strand of cDNA with SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen) according to the manufacturer’s procedure. Quantitative real-time PCR (qPCR) was conducted with UltraSYBR Mixture (Cwbio). qRT-PCR was performed with TransScript II Green One-Step qRT-PCR SuperMix (TransGen) according to the manufacturer’s protocol when using RNA as the template. The primers used in qPCR are listed in Supplementary Table S1.

Rapid amplification of cDNA ends (RACE)

5′ RACE and 3′ RACE were conducted as manufacturer’s instructions using FirstChoice RLM-RACE kit (Life Technologies). Primers used in RACE are listed in Supplementary Table S1. Briefly, for 5′ RACE, total RNA was treated with Calf Intestine Alkaline Phosphatase and Tobacco Acid Pyrophosphatase before an RNA Adapter (supplied by the manufacturer) was ligated to RNA by using T4 RNA ligase. Random-primed reverse transcription reaction was performed with SuperScript III First-Strand Synthesis System (Invitrogen). 5′ end of LRIK was amplified through nested PCR. For 3′ RACE, first-strand cDNA was generated from total RNA using the 3′ RACE Adapter (supplied by the manufacturer). Nested PCR was conducted to get the 3′ end of LRIK. All PCR products of RACE results were separated on 1.5% agarose gels. The bands were purified with the Gel and PCR Clean-Up System (A9282, Promega). The purified DNA was then cloned into the pGEM-T Vector Systems I (A3600, Promega) to retrieve the sequence. At least five colonies were sequenced for every RACE PCR.

Northern blot

Northern blot was performed using DIG Northern Starter Kit (Roche) following the manufacturer’s protocol. DIG-labeled RNA probes were generated by in vitro transcription using T7 RNA polymerase (Roche) and digoxigenin-11-UTP (Roche). Briefly, 30 μg total RNA was applied to 1.2% (w/v) agarose electrophoresis in MOPS buffer. After separation, RNA transferred onto a positively charged nylon membrane (GE Life Sciences, RPN303B) through capillary blotting followed by cross-linking with UV. The membrane was pre-hybridized with an appropriate volume of DIG Easy Hyb at 68 °C for 30–60 min before conducting the hybridization with DIG-labeled RNA probes overnight at 68 °C. After washing with 2× SSC, 0.1% (w/v) SDS for 5 min twice at room temperature and 0.1× SSC, 0.1% (w/v) SDS for 15 min twice at 68 °C, CPD-Star detection reagent was applied to the membrane. The images were acquired by exposing the membrane to standard X-ray film.

RNA fluorescence in situ hybridization

Appropriate amounts of cells were seeded on glass coverslips the day before experiments. Tested samples were applied with 30 μg/ml phleomycin (ab143437, Abcam) or 5 μm doxorubicin for 2 h and recovery for 2 h. Then, samples were fixed with 4% (w/v) paraformaldehyde in PBS at the indicated times after the treatment. After samples were washed three times with PBS, they were incubated with pre-cold permeabilization buffer (1× PBS, 0.5% (v/v) Triton X-100) for 5 min at 4 °C. LRIK was detected with in vitro transcribed DIG-labeled RNA probe with a length of ~200 nt. Samples were incubated in hybridization buffer (50% (v/v) formamide, 5× SSC, 500 μg/μl yeast tRNA, 1× Denhardt’s solution, 500 μg/ml herring sperm DNA, 50 μg/ml Heparin, 2.5 mM EDTA, 0.1% (v/v) Tween-20, 0.25% (w/v) CHAPS) at 65 °C for 20 min, and they were hybridized with DIG-labeled RNA probes in the same buffer 1 h at the same temperature. Coverslips with samples were washed five times with washing buffer (4× SSC, 0.1% (v/v) Tween-20) for 5 min. Then, all samples were washed with 2× SSC buffer once and 1× SSC buffer once. PBS buffer containing 3% (v/v) H2O2 was applied on washed coverslips with samples for 20 min followed by three times washing with TN buffer (0.1 M Tris-HCl pH7.5, 0.15 M NaCl). The samples were incubated for 30 min at room temperature in TNB blocking buffer (0.1 M Tris-HCl pH7.5, 0.15 M NaCl, 0.5% (w/v) blocking reagent), and then they were incubated for 30 min again after anti-DIG antibody (Abcam, ab51949) was added to the buffer. After samples were washed with TNT buffer, TSA solution (PerkinElmer Life and Analytical Sciences) was applied onto the samples and incubated for 5–10 min at room temperature. Protein detection and nuclear staining were performed as the protocol described. After samples were washed with TNT buffer (0.1 M Tris-HCl pH7.5, 0.15 M NaCl, 0.2% (v/v) Triton X-100) and PBS, they were blocked with 2% (w/v) BSA solution for 30 min at room temperature. The samples were incubated with primary antibodies overnight at 4 °C followed by incubations with appropriate secondary antibodies and NucBlue Live ReadyProbes Reagent (Life Technologies, R37605) for 1 h at room temperature. Samples were mounted onto clean glass slides with antifading agents (ProLong® Gold Antifade Reagent, Life P36930). The coverslips containing samples were sealed with nail polish. Images were acquired using Olympus FluoView 1200 confocal microscope. The intensity and co-localization were calculated using ImageJ software.

Immunofluorescence

Appropriate amounts of cells were grown on glass coverslips the day before experiments. Cells were treated with 2 Gy of X-ray irradiation (MBR-1520R-3, X-ray-irradiation equipment, Hitachi) or 20 µg/ml phleomycin for 2 h, and were harvested at the indicated times. All collected samples were fixed with methanol at 4 °C for 1 h. Specimens were blocked in 2% (w/v) BSA for 60 min after they were gently washed three times in 1× PBS for 5 min. Samples were incubated with primary antibody anti-γH2AX (9718, Cell Signaling Technology, 1/400) overnight at 4 °C followed the incubation with fluorochrome-conjugated secondary antibodies (Goat anti-Rabbit IgG (H+L) Secondary Antibody, Alexa Fluor 488 conjugate, Life Technologies) and NucBlue Live ReadyProbes Reagent (Life Technologies) for 1 h at room temperature. Images were obtained by using Olympus FluoView 1200 confocal microscope.

Plasmid construction

Plasmids stably expressing shRNA were constructed based on commercial vector, pSliencer 2.0-U6 (AM5762, Ambion), according to the manufacturer’s protocol. LRIK overexpression and LRIK antisense expression vector were constructed by inserting full-length cDNA of LRIK or antisense of LRIK into pCDH-CMV plasmid, respectively. The cDNA fragment of X-ray repair cross-complementing protein 6 (XRCC6) and XRCC5 was cloned into p3×Flag-CMV-10 vector (Sigma-Aldrich) following the manufacturer’s protocol. pBABE HA-AsiSI-ER plasmid was provided generously by Prof. Xianlu Zeng (Northeast Normal University, China). pEJ5-GFP, pDR-GFP, and I-SceI plasmids were kindly given by Prof. Xingzhi Xu (Shenzhen University Health Science Center). Primers and shRNA used were listed in Supplementary Table S1.

Colony-forming assay

Prior to cell survival determination by colony-forming ability, plating efficiency of all cell lines had been measured. The appropriate number of cells from each cell line (500 LRIK knockdown HeLa cells; 500 LRIK rescued HeLa cells; 500 antisense of LRIK expression cells; 800 LRIK and BRCA1 knockdown HeLa cells; 1000 MDA-MB-436 cells) was added to petri dishes (10 cm), such that in each case ~200 colonies would form. Appropriate cells of samples were seeded for overnight and specimens were treated with X-ray-irradiation equipment (MBR-1520R-3, Hitachi) of a 2, 4, 6, 8 Gy dose, respectively. For phleomycin (ab143437, Abcam) or doxorubicin treatment, cells were treated at indicated concentrations for 2 h, respectively. Then, all samples were cultured in a CO2 incubator. When colonies were formed after 10–14 days culture, cells were fixed using cold methanol and stained with 0.1% (w/v) crystal violet in 50% (v/v) methanol. The colonies were counted and normalized to plating efficiencies of untreated samples.

Neutral comet assay

Comet assays were conducted as previously described [23]. In brief, cells were treated with 20 µg/ml phleomycin for 2 h and were harvested either just after this treatment or after 2 h of recovery. Olympus FluoView 500 was used to observe acquired samples. The results were analyzed by CASP software and at least 100 cells were examined per condition for statistical analysis. The efficiency of DSB repair was calculated through the tail moment ratio, the mean tail length ratio and the Olive tail moment ratio between DNA from recovered (R) and phleomycin treated samples (D). R represents the samples treated with 20 µg/ml phleomycin for 2 h and recovered for 2 h. D represents the samples that treated with 20 µg/ml phleomycin for 2 h.

RNA pull-down assay

Biotinylated LRIK, antisense of LRIK, and truncated LRIK were generated by in vitro transcription with the Biotin RNA Labeling Mix (Roche) and T7 RNA polymerase (Roche) followed by RNase-free DNase I (Ambion) treatment and purification with NucleoSpin RNA Clean-up XS (Macherey Nagel). RNA pull-down assay was conducted according to the methods described in the literature [24, 25] without any modification. The proteins associated with RNA were subjected to 4–12% NuPAGE Bis-Tris gels (Life Technologies) followed by silver staining with SilverQuest Silver Staining Kit (Life Technologies) for further mass spectrometric analysis.

UV cross-linking RNA immunoprecipitation (UV-RIP)

The UV-RIP was carried out according to the protocol described in the publication [26]. CL1000 Ultraviolet Crosslinker (UVP) was used to conduct the cross-linking. The antibodies used in UV-RIP were anti-Flag (f3165, Sigma-Aldrich), anti-Ku70 (PLA0263, Sigma-Aldrich), anti-Ku80 (PLA0264, Sigma-Aldrich), anti-Ku80 (ab236277, Abcam), normal mouse IgG (sc-2025, Santa Cruz Biotechnology), and normal rabbit IgG (sc-2027, Santa Cruz Biotechnology). The retained samples were removed by proteinase K (Promega), and RNA of samples was extracted with TRIzol reagent for further investigation. The samples were treated with DNase I and were subjected to qRT-PCR to detect the enrichment of LRIK.

Chromatin fraction preparation

Cells were treated with 150 µm phleomycin or 5 µm doxorubicin for 2 h. Samples without recovery time were harvested immediately after the treatment. For recovery samples, cells were cultured for additional 1 h in normal media after PBS washes. Harvested cells were washed with pre-cold PBS twice and samples were extracted twice in CSK buffer (100 mM NaCl, 300 mM sucrose, 3 mM MgCl2, 0.7% (v/v) Triton X-100, 10 mM PIPES, pH7.0, 0.3 mg/ml RNase A (R4875, Sigma-Aldrich)) at 4 °C for 3 min. Samples extracted were washed with ice-cold PBS three times and stored in SDS-loading buffer for further analysis.

Western blot and antibodies

At indicated times, cells were washed with PBS. Whole cell extracts were acquired by homogenizing washed cells in 2× SDS-loading buffer. Extracts were boiled for 5 min. In brief, samples (20–80 µg) were resolved by 8% SDS-polyacrylamide gel (15% for histone H3 and γH2AX detection) electrophoresis and transferred to PVDF membranes (Millipore). The bands were quantified with ImageJ software. Primary antibodies used in this study were anti-β-actin antibody (60008-1-Ig, Proteintech, 1/5000), anti-H3 antibody (sc10809, Santa Cruz Biotechnology, 1/200), anti-phospho-histone H2AX (Ser139) antibody (9718, Cell Signaling Technology, 1/1000), anti-Ku70 antibody (PLA0263, Sigma-Aldrich, 1/1000), anti-Ku80 antibody (PLA0263, Sigma-Aldrich, 1/2000), anti-Ku80 (ab236277, Abcam, 1/1000), anti-DNA-PKcs antibody (sc5282, Santa Cruz Biotechnology, 1/200), anti-XRCC4 antibody (15817-1-AP, Proteintech, 1/1000), anti-ATM antibody (2873S, Cell Signaling Technology, 1/2000), anti-ATR antibody (2790S, Cell Signaling Technology, 1/2000), and anti-Flag antibody (f3165, Sigma-Aldrich, 1/5000). HRP-conjugated anti-rabbit IgG (7074, Cell Signaling Technology, 1/2000) and anti-mouse IgG (7076, Cell Signaling Technology, 1/2000) were the secondary antibodies.

Chromatin immunoprecipitation (ChIP) assay

ChIP assays were carried out as described previously [27] with slight modifications. The AsiSI-HeLa cells were treated with 300 nM 4-OHT for 4 h to induce DSBs. Formaldehyde was added to the medium at a final concentration of 1% and incubated for 10 min at room temperature. Glycine was added to a final concentration of 0.125 M to stop the reaction. Then cells were washed with cold PBS and harvested by scraping. Pelleted cells were incubated in ChIP lysis buffer (50 mM Tris pH8.1, 10 mM EDTA, 1% SDS) for 10 min before sonicated eight times for 10 s, so as to shear DNA to an average fragment size of about 500–1000 bp. Samples were diluted ten times in dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris pH8.1, 167 mM NaCl) and precleared with previously blocked Protein A/G beads (Life Technologies) for 3 h at 4 °C. The beads were blocked with BSA and herring sperm DNA. Precleared samples were incubated overnight with 2-µg-specific antibodies at 4 °C. Add 20 µl Protein A/G beads and incubated for another 4 h at 4 °C following with washing beads in dialysis buffer (2 mM EDTA, 50 mM Tris pH8.0, 0.2% SDS) once and in wash buffer (100 mM Tris pH8.8, 500 mM LiCl, 1% NP-40, 1% NaDoc) for four times. Add 100 µl of elution buffer and incubation for 15 min. Repeat the elution process once and then adding NaCl and incubating overnight at 65 °C to reverse the cross-link. After RNase A and proteinase K treatment, DNA was purified with phenol/chloroform and analyzed by qPCR. The primers used located in proximity to (200 bp) and distal (2 Mb) from the AsiSI site on chromosome 22 at position 20850307 (GRCh37/hg19). The antibodies used are anti-Ku70 antibody (ab83501, Abcam), anti-Ku80 antibody (2753, Cell Signaling Technology), anti-DNA-PKcs antibody (ab168854, Abcam), and anti-XRCC4 antibody (15817-1-AP, Proteintech).

DSB repair assays

To qualify the DSB repair by NHEJ and HR after LRIK knockdown or overexpression, U2OS EJ5-GFP and DR-GFP cells were transfected with indicating shRNA plasmids or overexpression plasmids, then cells were transfected with I-SceI plasmid after 24 h. For HeLa cells, shNeg, shLRIK-2, and shLRIK-3 cells or LRIK overexpression cells were transfected with pEJ5-GFP or pDR-GFP plasmids, and 24 h later cells were transfected with I-SceI plasmid. Cells were suspended in PBS buffer after 48 h and the NHEJ or HR efficiencies were determined by flow cytometry (FACSCalibur or LSRFortessa, Becton Dickinson, USA). Twenty thousand cells of each sample were tested and data were analyzed with FlowJo software.

Protein purification and electrophoretic mobility shift assay (EMSA)

The protein used in the EMSA was expressed in E. coli DL21 (DE3) by using pET-28a expression vector. After transfection, the bacteria were incubated until reaching 0.6 at OD600. Then, the recombinant protein was induced by adding 1 mM IPTG for overnight at 16 °C. Next, the protein was purified with His-tag protein purification kit (Beyotime, P2226) following the manufacturer’s protocol. In brief, E. coli lysate containing overexpressed Ku70 was incubated with BeyoGoldTM His-tag Purification Resin for 30 min. The resin was then eluted eight times with elution buffer after washing four times with the washing buffer.

Full length of LRIK was labeled with biotin through in vitro transcription assays. The EMSA was conducted as pervious report [28]. 0.7 pmol of biotinylated LRIK was used for each reaction. The secondary structure of LRIK was formed by heating at 95 °C for 5 min following with slowly cooled down to room temperature. The indicated amount of purified protein and biotinylated LRIK was incubated in binding buffer (100 mM HEPES pH7.5, 200 mM KCl, 10 mM MgCl2, 10 mM DTT) for 30 min at room temperature. The mixtures were subjected to electrophoresis on a 6% nondenaturing polyacrylamide 0.5x TBE gel. The gel was then transferred to a nylon membrane and imaged by incubation with Chemiluminescent Nucleic Acid Detection Module Kit (Thermo #89880).

Chromatin isolation by RNA purification (ChIRP)

The ChIRP assays were performed as previous described [29]. Briefly, the ChIRP probes were designed at Biosearch Technologies. Cells were transfected with AsiSI plasmid before cross-linking with 0.5 mg/ml AMT [30, 31]. Then cells were exposed at 350 nm UV for 10 min on ice followed by cell lysis (lysis buffer: 50 mM Tris-Cl pH7.0, 10 mM EDTA, 1% SDS) and sonication. Samples were sonicated in a 4 °C water bath with 10 s ON, 10 s OFF pulse intervals and sonication time was set for 15 min. Sonicated specimens were hybridized with 100 pmol biotinylated probes in hybridization buffer (750 mM NaCl, 1% SDS, 50 mM Tris-Cl pH7.0, 1 mM EDTA, 15% formamide, PMSF, P.I. and RNase inhibitor) at 37 °C for 4 h and C-1 magnetic beads were added. After incubation at 37 °C for 30 min, beads were collected and washed with wash buffer (2×SSC, 0.5% SDS, PMSF) for five times. RNA contained in washed samples was isolated with TRIzol and the enrichment of LRIK was examined. DNA contained in washed samples was eluted and purified with phenol/chloroform. The enrichments of the specific DNA fragments were detected with RT-qPCR by using the primers located in proximity to (60 bp) and distal (2 Mbp) from the AsiSI site on chromosome 22 at position 20850307 (GRCh37/hg19). The probes used in ChIRP are listed in Supplementary Table S1.

Statistics

The precise number of samples in each experiment was indicated in figure legends. Except for specific indications, data are expressed as mean ± standard deviation (SD) and the statistical significance was determined by two-tailed Student’s t-test. Significance was referred as *P < 0.05, **P < 0.01, and ***P < 0.001.

Results

Identification of upregulated lncRNAs following induction of DSBs

Since the transcription of many ncRNAs is often associated with their cellular roles [4], we reasoned that ncRNAs involved in the DDR may also be transcribed following exposure of cells to DNA damaging agents. To identify candidate ncRNAs expressed in response to DNA damage, custom-designed ncRNA microarrays with collections of ncRNAs including self-discovered ncRNAs, which have been deposited into the NONCODE database [22], were applied for the detection of damage-induced ncRNAs in HeLa cells, following their treatment with MMS, a DSB-inducing chemical (Fig. 1a). Following the experimental validation of the expression changes upon MMS treatment using qPCR (Fig. 1b), a novel lncRNA (NONCODE Gene ID: NONHSAG044790.2, Transcript ID: NONHSAT114847.2), which we named as LRIK, was selected as a candidate for further investigation based on its relatively higher expression level of induction following exposure to MMS. This chemical can induce various types of DNA damage, including DNA DSBs. Therefore, to test whether enhanced transcription of LRIK is primarily induced by DSBs, we repeated these experiments using other DSB-inducing genotoxic agents, including X-rays and phleomycin [32]. In addition, we repeated these experiments treating a variety of different cell lines, to confirm that the observations are not unique to HeLa cells (Fig. 1c–f).

Fig. 1: Transcription induction of LRIK in response to DNA damage.
figure 1

a MA plot of long noncoding RNAs tested in custom-designed microarrays. Black dots: non-differentially expressed genes (non-DE genes); red dots: upregulated differentially expressed genes (up DE genes) identified in this study; blue dots: downregulated differentially expressed genes (down DE genes) identified in this study. The arrow indicates LRIK. b Validation of transcription induction of lncRNAs listed by using qRT-PCR. RNAs were extracted from HeLa cells treated with MMS. The error bars indicate SD (n = 3 independent experiments). c qRT-PCR analyses in HeLa cells show LRIK was upregulated under 4 or 8 Gy of X-ray irradiation. The error bars indicate SD (n = 3 independent experiments). d Elevated transcriptions of LRIK were observed in HeLa cells treated with phleomycin. e Transcription induction of LRIK in response to phleomycin induced DNA damage stress in 293T cells. f Transcription induction of LRIK in response to phleomycin induced DNA damage stress in A549 cells. TP53 and XRCC6 were used as positive and negative control, respectively, in (df). The error bars indicate SD (d: n = 3, e: n = 3, f: n = 3, independent experiments). g FISH results showed that LRIK was upregulated in phleomycin treated cells compared with untreated HeLa cells. Cells were fixed after the treatment of 30 µg/ml phleomycin for 2 h and recovery for 2 h, then samples were immunostained with LRIK probe (red) and DAPI (blue). The intensity of LRIK is shown below (n = 3 independent experiments, the precise number of cells analyzed in total per condition was indicated). Scale bar, 20 µm. *P < 0.05, **P < 0.01, ***P < 0.001, by two-tailed Student’s t test.

Increased expression of LRIK following induction of DSBs was also examined cytologically by single-molecular RNA fluorescence in situ hybridization (RNA FISH) assay in HeLa cells, following their exposure to 30 µg/ml of phleomycin (Fig. 1g) or doxorubicin (Supplementary Fig. S1). Collectively, these results show that the transcription of LRIK is significantly elevated in response to DSB-induced stress in human cells.

Characterization of LRIK and identification of its interacting partners

In human cells, LRIK is a transcript located on chromosome 6 (127296102–127296760; GRCh38/hg38) (Fig. 2a). Northern blotting analysis revealed a transcript with a length about 600 nucleotides (Fig. 2b), which coincides well with the results of RACE assays (Supplementary Fig. S2a). LRIK has no coding capacity, as predicted by CPAT [33] and Coding-Noncoding Index (CNCI) [34] (CNCI score: −0.1500472) (Supplementary Fig. S2b).

Fig. 2: LRIK interacts with the DSB repair component Ku70.
figure 2

a The transcription region of LRIK. b The representative image of Northern blotting assay. LRIK was detected with DIG-labeled RNA probes and 30 µg total RNA sample was loaded. c Cellular distribution of LRIK. Listed RNAs were measured by qRT-PCR after cell fractionation. The error bars indicate SD (for U1, n = 4, for the other conditions, n = 5 independent experiments). d Top, pull-down assays were performed by using biotin labeled LRIK (7 pmol) incubated with HeLa nuclear extracts (1 mg) to enrich proteins binding with LRIK. The retrieved proteins were subjected to NuPAGE Bis-Tris gels followed by silver stain and mass spectrometric identification. The arrow indicates Ku70 as the binding candidate interacting with LRIK in vitro. Bottom, western blotting analyses following pull-down assays showed the specific interaction between LRIK and Ku70. AS indicates antisense of LRIK. e UV cross-linking RIP results show that LRIK was enriched by Ku70-Flag in 293T cells with overexpression of a Flag epitope-tagged Ku70. RNA was immunoprecipitated with anti-Flag antibody followed by detecting LRIK and U1. The relative enrichment was calculated as RNA associated with Flag IP relative to the U1 RNA. IgG was used as a negative control. The error bars indicate SD (n = 3 independent experiments). f UV cross-linking RIP with anti-Ku70 or anti-Ku80 antibodies (Sigma-Aldrich) confirmed that LRIK interacts with Ku heterodimer in vivo. Co-immunoprecipitation ncRNAs with anti-Ku70 or anti-Ku80 antibodies, respectively, in HeLa cells were detected with qRT-PCR. The error bars indicate SD (n = 3 independent experiments). *P < 0.05, **P < 0.01, by two-tailed Student’s t test. g The results of UV cross-linking RNA IP by anti-Ku80 (Abcam) in HeLa cells. RNA was immunoprecipitated with indicated antibodies followed by detecting LRIK. The relative enrichment was calculated as RNA associated with Ku70/Ku80 IP relative to the input. There is no significant difference between Ku70 IP and Ku80 IP results (P = 0.413). The error bars indicate SEM.

Quantitative RT-PCR assay on both the cytosolic and nuclear fractions revealed the nuclear location of LRIK. In contrast to GAPDH and ACTB but similar to U1, LRIK is mainly located in the cell nucleus (Fig. 2c). To further validate the subcellular localization of LRIK, FISH assay was employed. It was observed that the vast majority of LRIK is located in the nucleus (Supplementary Fig. S2c). Collectively, these results are consistent with LRIK playing a role in the cellular response to DNA damage.

Previous reports have shown that lncRNAs can bind with a protein partner(s) functioning coordinately. Therefore, we tested whether LRIK binds to protein(s) involved in the DDR. To do this, RNA pull-down assays were conducted to discover potential binding protein(s). Biotinylated sense or antisense transcripts of LRIK were incubated with nuclear extracts to enrich for LRIK associated protein(s), whose identities were subsequently characterized by mass spectroscopy (Supplementary Table S2). Among those, Ku70 (XRCC6) was hence identified as a potential binding partner of LRIK (Fig. 2d, top). Western blot analysis following RNA pull-down confirmed the specific interaction of LRIK with Ku70 as compared with the one observed using the antisense transcript of LRIK (Fig. 2d, bottom). To further confirm the interaction between LRIK and Ku70, cell lysates from Flag-Ku70 overexpressing 293T cells were examined using the UV-RIP assay. The results clearly confirmed the interaction between Ku70 and LRIK (Fig. 2e and Supplementary Fig. 2d). Besides, to confirm the direct interaction between LRIK and Ku70, we performed EMSA by using purified Ku70 and biotinylated RNA (Supplementary Fig. S2e, f). Our data showed free LRIK RNA bands decreasing and Ku-RNA shift bands increasing depend on the increased amount of Ku70 protein mixed with LRIK samples. Collectively, these results demonstrate that LRIK interacts specifically with Ku70.

It is known that Ku70 interacts with Ku80 to form the Ku heterodimer [11], which functions as a complex in the NHEJ pathway. To verify the interaction between LRIK and the Ku complex in vivo, the UV-RIP assays were performed with anti-Ku70 and anti-Ku80 antibody, respectively. The results showed that LRIK enrichments obtained by these antibodies were significantly higher than that by IgG (Figs. 2f and  S2g). To further confirm it, we repeated the UV-RIP assays with anti-Ku70 and anti-Ku80 antibody produced by different company (Abcam) and obtained similar results as showed in Fig. 2g (Supplementary Fig. S2h). Together, these observations demonstrate the association between LRIK and the Ku70–Ku80 heterodimer.

Expression of LRIK regulates the efficiency of DSB repair

Since LRIK binds to Ku heterodimer, which is responsible for the sensing of DNA DSBs during NHEJ, we next tested whether LRIK affects the repair of DSBs. Two LRIK knockdown HeLa cell lines were generated by using pSilencer plasmids (Fig. 3a) via shRNA strategy. The neutral comet assay was used to measure the efficiency of DSB repair in control (shNeg) and LRIK knockdown cells. Following phleomycin treatment, long comet tails indicated that DNA damages were generated in control (shNeg) cells and in LRIK knockdown cells (Fig. 3b). Notably, quantifying these data demonstrated that initial levels of DSB induction were not affected by LRIK depletion (Fig. 3c). After 2 h of recovery following phleomycin treatment, the rate of DSB repair as measured by tail moment (Fig. 3d), mean tail length, and Olive tail moment (Supplementary Fig. S3a, b) demonstrated that the DSB repair efficiency in LRIK knockdown cells was lower than that of control (shNeg) cells. Our data demonstrate that LRIK interacts with Ku70 during the DSB repair process and influences the efficiency of repairing DSBs.

Fig. 3: LRIK is essential for efficient repair of DSBs in human cells.
figure 3

a The RNA level of LRIK in LRIK knockdown cells measured by qRT-PCR assay. The error bars indicate SD (n = 6 independent experiments). b Representative images of comet assays displayed reduced DSB repair efficiency in LRIK knockdown cells compared with control cells (n = 3 independent experiments, the precise number of cells analyzed in total per condition was indicated). c Tail moment of cells that treated by phleomycin without recovery was shown in column scatter plot with a line at geometric mean. ns not significant (pooled data from n = 3 independent experiments are shown). d Quantified tail moment of tested samples with CASP software showed significantly impaired DSB repair efficiency in LRIK knockdown cells. D represents samples that were treated with 20 µg/ml phleomycin for 2 h. R represents samples that were treated with 20 µg/ml phleomycin for 2 h and recovered for another 2 h. The error bars indicate SD (n = 3 independent experiments, the precise number of cells analyzed in total per condition was indicated in (b)). e Colony-forming assays showed that LRIK knockdown cells display increasing radiosensitivity compared with control shRNA (shNeg) cells. The error bars indicate SD (n = 3 independent experiments). f Colony-forming assays showed that knockdown of LRIK repressed the survival abilities following their treatment with phleomycin for 2 h at indicated concentrations and that can be rescued by overexpression of LRIK. The error bars indicate SD (n = 3 independent experiments). g The RNA level of LRIK after overexpressing LRIK in knockdown cells. The error bars indicate SD (n = 4 independent experiments). h NHEJ-mediate repair was analyzed in HeLa LRIK knockdown cells or U2OS LRIK knockdown cells by an EJ5-GFP reporter (mean ± SD, n = 4 independent experiments). i HR-mediate repair was analyzed in HeLa LRIK knockdown cells or U2OS LRIK knockdown cells by a DR-GFP reporter (mean ± SEM, n = 3 independent experiments). j NHEJ-mediate repair was analyzed in HeLa LRIK overexpression cells or U2OS LRIK overexpression cells by an EJ5-GFP reporter (mean ± SEM, n = 3 independent experiments for HeLa cells and n = 4 independent experiments for U2OS cells). Vector means overexpression of empty pCDH vector in cells, and oeLRIK means overexpression of pCDH-LRIK in cells. *P < 0.05, **P < 0.01, ***P < 0.001, by two-tailed Student’s t test.

Since inefficient repair of DSB lesions results in cell death, the cellular survival rates of cell lines were determined using the colony-forming assay to confirm the effect of LRIK on DSB repair efficiency. Compared with the control (shNeg) cell line, the LRIK knockdown cell lines exhibited significantly increased sensitivity to X-ray irradiation (Fig. 3e). Similarly, we also observed significantly reduced survival of LRIK knockdown cells following their treatment with phleomycin (Fig. 3f) or doxorubicin (Supplementary Fig. S3c). Overexpression of LRIK in LRIK knockdown cells rescued their sensitivities to phleomycin (Fig. 3f, g), demonstrating that efficient DSB repair requires sufficient levels of LRIK. To further investigate the role of LRIK in DSB repair, the dependence on LRIK for DSB repair was examined by analyzing the sensitivity of cells deficient in other HR factors. Knockdown of BRCA1 results in enhanced sensitivity of LRIK knockdown cells (Supplementary Fig. S3c, d). Moreover, after knocking down LRIK in BRCA1 defective cells (MDA-MB-436), increased sensitivity to phleomycin (Supplementary Fig. S3e, f) was conferred. This result indicates that LRIK functions in the Ku70 initiated DSB recognition process and is required for efficient DSB repair.

In order to explore the functional significance of LRIK in DSB repair, we examined the effect of LRIK on the repair efficiency of two major DSB repair pathways, NHEJ and HR, by using EJ5-GFP and DR-GFP systems, respectively [35, 36]. Cells containing the EJ5-GFP reporter can only repair DSB in the reporter plasmid by NHEJ and display GFP fluorescence following repair of DSB. Similarly, cells infected with the DR-GFP reporter only exhibit GFP fluorescence after DSB induced in the plasmid has been repaired by HR. Our data showed that knockdown of LRIK in EJ5-GFP reporter containing cells caused a significant reduction of the GFP-positive cells compared with the control sample (Fig. 3h), whereas knockdown of LRIK exhibited little effect on HR repair when evaluated using the DR-GFP system (Fig. 3i). Moreover, to address whether overexpression of LRIK could enhance the repair efficiency of NHEJ, we employed EJ5-GFP reporter assays to measure NHEJ efficiency in both LRIK overexpressed and normal cells (Fig. 3j). It indicates that LRIK overexpressed cells display a significantly increased NHEJ efficiency compared with the control cells. Together, these results indicate that LRIK is required for efficient repair of DSBs specifically through the NHEJ pathway.

LRIK enhances the binding of the Ku complex on damaged chromatin

Next, we examined the mechanism by which the interaction of LRIK with the Ku heterodimer enhances DNA DSB repair efficiency. It has been suggested that the early and prolonged binding of repair factors is crucial for the initiation and signaling of DNA damage and its repair [37]. The repair factors, which include DNA-PKcs, XRCC4, DNA Ligase IV, and XLF, bind to DSBs following the sensing of the damage by the Ku70–Ku80 heterodimer [17]. Therefore, we investigated whether LRIK influences the binding of the Ku70–Ku80 heterodimer and its binding partners on damaged chromatin following the induction of DSBs.

We extracted the chromatin fractions from the control (shNeg) cells and LRIK knockdown cells to examine the occupancy of tightly associated proteins with chromatin after exposing them to phleomycin or doxorubicin. Following the treatment with DSB-inducing agents, significantly increased chromatin occupancies of these factors were observed in control (shNeg) cells immediately after the treatments and following recovery for 1 h (lanes 1–3, Figs. 4a and  S4a). However, compared with control (shNeg) cells, the increase in occupancy of Ku70, Ku80, DNA-PKcs, and XRCC4 on chromatin in the two LRIK knockdown cells was significantly reduced both immediately after phleomycin treatment (Fig. 4a compares the lanes 5 and 8 with lane 2; Fig. 4c shows histograms for the relative band intensities) and following 1 h of recovery (Fig. 4a compares the lanes 6 and 9 with lane 3; Fig. 4c shows histograms for the relative band intensities). Similarly, the reduced occupancy of Ku70, Ku80, DNA-PKcs, and XRCC4 on chromatin was also revealed in LRIK knockdown cells after doxorubicin treatment (Supplementary Fig. S4a, c). Importantly, the total protein levels of Ku70, Ku80, DNA-PKcs, and XRCC4 remained stable in control (shNeg) cells and LRIK knockdown cells following treatment with DSB-inducing agents (phleomycin and doxorubicin) either immediately after or following 1 h of repair (Figs. 4b and S4b). This demonstrates that knockdown of LRIK does not inhibit the expression of Ku70, Ku80, DNA-PKcs, and XRCC4, suggesting that reduced binding of these factors is a direct result of reduced levels of cellular LRIK levels (Figs. 4 and S4).

Fig. 4: LRIK enhances the occupancy of repair factors, including Ku70, Ku80, DNA-PKcs, and XRCC4, on damaged chromatin.
figure 4

a The representative image of western blotting analyses of chromatin fractions. The samples were extracted from LRIK knockdown and control shRNA (shNeg) cells that were treated with 150 µm phleomycin for 2 h with or without recovery for 1 h after treatments. b The representative image of western blotting analyses of whole cell extractions. Whole cell extractions were extracted from cell lines with the same treatments described in (a). c The relative band intensity histograms of chromatin fractions showed that knockdown of LRIK impaired the DSB repair factors, including Ku70, Ku80, DNA-PKcs, and XRCC4 binding with damaged chromatin. The results are shown as the ratio of target protein/H3. The error bars indicate SD (for Ku80 and XRCC4, n = 4, for DNA-PKcs and Ku70, n = 3 independent experiments). Accumulations of Ku70 (d), Ku80 (e), DNA-PKcs (f), and XRCC4 (g) at the indicated locus by ChIP assays performed in AsiSI cut cells (+AsiSI) and no AsiSI cut cells (−AsiSI). The bar plots show the enrichment relative to the input DNA, as detected with primers in the proximity or distal of the AsiSI-induced DSB. The error bars indicate SD (n = 3 independent experiments). ns not significant. *P < 0.05, **P < 0.01, ***P < 0.001, by two-tailed Student’s t test.

To test whether knockdown LRIK influences the occupancy of Ku70, Ku80, XRCC4, and DNA-PKcs at DSB sites, ChIP assays were performed in control cells and LRIK knockdown cells with site-directed breaks introduced by AsiSI (Fig. 4d–g). It revealed that Ku70/Ku80 were notably enriched at DSB sites in control cells following DSB induction (Fig. 4d, e), while the reduced associations of Ku70/Ku80 were observed in LRIK knockdown cells (Fig. 4d, e). Correspondingly, the downstream factors, DNA-PKcs and XRCC4, also exhibited significant enrichment in control cells at an endogenous locus cut by AsiSI (Fig. 4f, g). However, impaired accumulation of DNA-PKcs and XRCC4 in cells was observed in LRIK knockdown cells (Fig. 4f, g). Hence, our results illustrate that the interaction of LRIK with the Ku heterodimer enhances repair complex formation at sites of DSBs in chromatin.

LRIK is required for efficient sensing and signaling of DSBs to promote the formation of γH2AX foci

Next, we investigated whether LRIK interacts with damaged DNA in DSB repair. After a 1-h recovery period following the phleomycin or doxorubicin treatments, the level of LRIK showed a significantly increase in the chromatin-associated fraction (Fig. 5a), suggesting that the majority of LRIK were fixed with chromatin in response to the DSB stress. To further test whether LRIK binds to DNA at the DNA damage sites during NHEJ repair, psoralen (AMT) cross-linked ChIRP (chromatin isolation and RNA purification) assays were performed. After introducing site-directed breaks by AsiSI, AMT was used to cross-link the RNA-bound DNA in vivo since it specifically cross-links direct RNA–DNA and RNA–RNA interactions. Chromatin was sonicated, and nucleosomes associated with LRIK were separated for further analysis. By examining the enrichment of two different DNA fragments, which are 2 Mbp and 60 bp away from the AsiSI-induced DSB site, respectively, the results revealed that LRIK interacts directly with DNA at the DSB damaged sites over non-DSB lesions (Fig. 5b, c).

Fig. 5: LRIK enhances the sensing and signaling of DNA double-strand breaks.
figure 5

a Cellular distribution of LRIK measured by qRT-PCR in untreated cells or after recovery for 1 h following DSB-inducing agents treatments. Cytoplasm: cytoplasmic extracts, Nucleus: nuclear extracts, Chromatins: chromatin-associate components. The error bars indicate SD (n = 3 independent experiments). *P < 0.05, **P < 0.01, ***P < 0.001, by two-tailed Student’s t test. b ChIRP-qPCR of LRIK in AMT cross-linked samples (+AMT) and no AMT cross-linked samples (−AMT). GAPDH served as a negative control. c Enrichment of LRIK binding on DNA double-strand breaks verified by ChIRP-qPCR detected with primers in the proximity or distal of the AsiSI-induced DSB. For (b, c), mean + SD, n = 3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, by two-way ANOVA. d Western blotting analyses of chromatin fractions showed that knockdown of LRIK impaired the level of γH2AX. The samples were extracted from cell lines with the same treatments described in Fig. 4a. e FISH images indicated that LRIK foci overlapped with γH2AX. HeLa cells were fixed after doxorubicin treatments for 2 h, and samples were immunostained with LRIK (red), γH2AX (green), and DAPI (blue). The percentage of LRIK co-localizing with γH2AX signal indicated was measured 275 cells from n = 3 independent experiments. Scale bar, 20 µm. f Immunofluorescence analyses of γH2AX foci indicated that LRIK knockdown cells exhibited impaired γH2AX foci formation. Quantitative measurements of γH2AX foci were performed following 2 Gy of X-ray irradiation and fixed at indicated times. Cells with >10 γH2AX foci were scored as positive. The error bars indicate SD (n = 3 independent experiments, the precise number of cells analyzed in total per condition was indicated). g The representative images of γH2AX foci in (f) are shown. Scale bar, 20 µm. h Immunofluorescence analyses of γH2AX foci indicated that LRIK knockdown cells exhibited impairment of γH2AX foci formation following their treatment of phleomycin. γH2AX foci were enumerated in tested samples. Cells with >10 γH2AX foci were scored as positive. The error bars indicate SD (n = 3 independent experiments, the precise number of cells analyzed in total per condition was indicated). i The representative images of γH2AX foci in (h) are shown. Scale bar, 20 µm. *P < 0.05, **P < 0.01, by two-tailed Student’s t test.

Since LRIK is required for recruitment of DNA repair factors to DSBs, we speculated that failure to recruit DNA-PKcs in LRIK depleted cells may inhibit the formation of gamma H2AX foci. As expect, the levels of phosphorylated histone H2AX were impaired in LRIK knockdown cells compared with control (shNeg) cells following phleomycin treatment (Fig. 5d). Since phosphorylated histone H2AX (γH2AX) [38] is a critical marker in response to DSB formation [39], we further examined the localization of LRIK and γH2AX to check the localization of LRIK at DSB repair sites. By employing the FISH assay, we observed that the location of LRIK overlapped well with γH2AX in the nucleus after cells were treated with doxorubicin (Figs. 5e and S4d), suggesting that LRIK may participate in the signaling of DSB lesions.

To further test our hypothesis, we investigated whether the efficiency of DSB sensing and signaling is affected by LRIK through monitoring the formation of γH2AX foci. We observed that LRIK knockdown cells exhibited fewer γH2AX foci than control samples during a 2-h period following irradiation, particularly at 0.5 h (Fig. 5f, g). Consistently, significantly reduced γH2AX foci were observed in LRIK deficient cells compared with control cells (shNeg) at the indicated time points using the X-ray mimic drug phleomycin (Fig. 5h, i). This demonstrates that significant reduction of the cellular levels of LRIK impairs the formation of γH2AX foci, demonstrating that LRIK affects the sensing and signaling of DSB and the subsequent formation of γH2AX foci.

Phosphorylation of H2AX is indispensable for the formation of γH2AX foci, and this occurs during the DSB sensing and signaling event [38, 40, 41]. It has been suggested that H2AX can be phosphorylated by ATR, ATM, and DNA-PKcs in response to DNA damage, while ATM and DNA-PKcs contribute to H2AX phosphorylation in response to DSBs [42, 43]. To investigate the reasons for reduced H2AX phosphorylation in LRIK knockdown cells, we detected the mRNA levels and expression levels of ATM and ATR in these cells. We observed that knockdown of LRIK did not affect the transcription or expression of ATM and ATR (Supplementary Fig. S5b–d). Furthermore, western blotting analyses following the LRIK pull-down assays showed that LRIK does not interact with ATR or ATM in vitro (Supplementary Fig. S5e, f). In addition, LRIK deficiency did not change the protein level of DNA-PKcs in whole cell extracts (Supplementary Fig. S5a). Our data indicate that the impaired formation of γH2AX in LRIK knockdown cells is mainly caused by reduced recruitment of DNA-PKcs at DSB sites.

To further confirm that the interaction of LRIK and Ku heterodimer is crucial for the DSB sensing and signaling event, we examined whether knockdown of Ku70 and Ku80 recreates similar phenotypes to that observed in LRIK knockdown cells. Since the γH2AX foci formation is a consequence of efficient DSB recognition, it can be used as an indicator of the completed DSB sensing and signaling. Depletion of Ku70 or Ku80 impaired the efficiency of γH2AX foci formation after 2 h of recovery, following their treatment with phleomycin (Supplementary Fig. S6). We conclude that LRIK functions in the NHEJ repair pathway via its interaction with Ku heterodimer.

Efficient DSB repair requires the interaction of LRIK and Ku heterodimer

To further confirm the coordinated function of LRIK and Ku heterodimer in DSBs repair particularly in the NHEJ repair pathway, a HeLa cell line expressing antisense of LRIK was constructed by transfecting a plasmid (pCDH-CMV-AS) in HeLa cells (Fig. 6a). It is exhibited that the γH2AX foci formation was impaired in LRIK antisense expression cells instead of control cells after recovery for 2 h following the phleomycin treatment (Fig. 6b, c). The DSB repair efficiency was reduced after hindering the interaction of LRIK and Ku70 through expressing antisense of LRIK measured with the neutral comet assays (Figs. 6d, e and S7a–c). Consistently, expression of antisense of LRIK caused significantly increased sensitivity to phleomycin (Fig. 6f), illustrating that LRIK and Ku70 function in NHEJ repair coordinately.

Fig. 6: Impeding the interaction of LRIK and Ku70 influences the efficient repair of DSB.
figure 6

a The expression level of antisense of LRIK determined by qRT-PCR. The error bars indicate SD (n = 4 independent experiments). b Immunofluorescence analyses of γH2AX foci indicated that expression of LRIK antisense impaired the γH2AX foci formation after recovery for 2 h following phleomycin treated. γH2AX foci were calculated in tested samples. Cells with >10 γH2AX foci were scored as positive. The error bars indicate SD (n = 3 independent experiments, the precise number of cells analyzed in total per condition was indicated). c Representative images of γH2AX foci. Scale bar 20 µm. d Typical images of comet assay revealed impaired DSB repair efficiency in LRIK antisense expression cells compared with control cells (the precise number of cells analyzed in total per condition was indicated). e Quantified tail moment of tested samples with CASP software showed significantly impaired DSB repair efficiency in LRIK antisense expression cells. Samples were treated with 20 µg/ml phleomycin for 2 h with or without recovering for another 2 h. The error bars indicate SD (n = 3 independent experiments, the precise number of cells analyzed in total per condition was indicated in (d)). f Colony-forming assays showed that cells with expression of LRIK antisense exhibited reduced survival rate after phleomycin treated. The error bars indicate SD (n = 3 independent experiments). *P < 0.05, **P < 0.01, by two-tailed Student’s t test. g The sketch map of truncated LRIK used in pull-down assays. Bottom, western blot analyses of pull-down assays with truncated LRIK fragments indicated that the 5′ region of LRIK (1–151 nucleotides) mainly interacts with Ku. h Colony survival rates of overexpressing truncated LRIK in LRIK knockdown cells. L1: 1–333 nt of LRIK, L2: 333–659 nt of LRIK. The error bars indicate SD (n = 3 independent experiments).

Finally, we sought to characterize the interacting regions of LRIK that associate with the Ku heterodimer. The Ku heterodimer binding capacity of different regions of LRIK in vitro demonstrated that the Ku heterodimer primarily binds with the 1–151 nucleotides region of LRIK (Fig. 6g). Moreover, the sensitivities of cells to the DSB-inducing agent were examined following truncated LRIK vectors overexpressed in LRIK knockdown cells. Interestingly, neither the 1–333 nt nor 333–659 nt regions of LRIK can rescue the sensitivity of LRIK knockdown cells (Fig. 6h). This demonstrates that the efficient repair of DSB requires intact LRIK.

Discussion

Despite the discovery of the versatile functions of lncRNAs having been reported [44], only a few lncRNAs have been implicated in the repair of DNA [5,6,7,8,9, 45]. Here, we identified and characterized a novel lncRNA LRIK, which functions during NHEJ repair and participates in repairing DSBs in human cells. Interacting with the Ku heterodimer, LRIK promotes the binding of the Ku heterodimer at DSB sites and facilitates the assemblies of the downstream repair factors that are required for the efficient sensing and repair DSBs (Fig. 7).

Fig. 7: Model of LRIK functioning in the NHEJ repair pathway.
figure 7

LRIK interacts with the Ku heterodimer thus promoting the occupancy of Ku with damaged chromatin, resulting in facilitating the assemblies of other downstream repair factors and promoting γH2AX foci formation, hence influences the efficient repair of DSBs.

Publications have demonstrated that Ku70 exhibits a higher affinity binding to hairpin RNAs with certain structure than linear ssRNAs, dsRNAs, or other types of hairpin RNAs [46], for example, the bulge motif is essential for Ku binding RNAs like TLC1 [47]. Publications suggest that specific sequences of RNAs display high affinities with Ku70 [48], and these sequences (AATG, CATGA) have been found in LRIK. Besides, the other proteins identified by mass spectrometry showed no specific interaction with LRIK compared with Ku70 (Fig. 2d). Hence, it demonstrates that Ku70 selectively interacts with LRIK during the NHEJ repair.

Since the recognition of DNA lesions requires the prolonged association of DNA repair factors with DNA damaged sites [37], we hypothesized that LRIK assists Ku heterodimer during the DSB sensing and signaling event. Consistent with this notion, we observed that the occupancy of Ku70, Ku80, XRCC4, and DNA-PKcs with chromatin was reduced in LRIK deficient cells after they were treated with DSB-inducing agents (Figs. 4 and S4), implying that LRIK may reinforce Ku70 binding with damaged chromatin. Our data show that LRIK interaction with Ku heterodimer results in prolonged association with DSB sites and the recruitment of XRCC4 and DNA-PKcs.

In summary, the formation of DSBs in cells induces LRIK, which together with the Ku heterodimer functions coordinately to enhance the signaling of DSBs by regulating the assembly of downstream repair factors on chromatin, thus enhancing the efficiency of DSB repair (Fig. 7). This study deepens our understanding of both lncRNA functioning and the sensing and signaling event of the NHEJ repair pathway that may pave the way for application in drug designation and cancer treatments.