Dear Editor,

Mammalian cells are constantly experiencing DNA damage through exogenous and endogenous factors that cause genomic instability and subsequently human diseases such as cancers. Thanks to the exquisite DNA repair systems that keep the integrity and stability of our genome, human cells can survive and evolve properly. For DNA double-strand breaks (DSBs), there are two major repair pathways: one is homologous recombination (HR) that uses an intact DNA template to recombine the broken DNA, and the other is non-homologous end-joining (NHEJ) that directly joins two broken ends (Liu et al., 2017). Both pathways consist of a variety of repair factors to complete the repair progress. For example, tumor suppressor BRCA1 which is associated with hereditary breast and ovarian cancers, is a key player in HR repair. BRCA1 was accumulated at DNA damage sites and associated with a variety of proteins. The N terminus of BRCA1 contains RING domain that binds BARD1 and forms BRCA1/BARD1 complex. Interestingly, during HR, BRCA1/BARD1 complex can promote DNA resection to generate ssDNA overhangs as well as the loading of RAD51 onto the ssDNA. Since loss of BRCA1 would severely abrogates HR, BRCA1-deficient cells are sensitive to PARP inhibitors (Farmer et al., 2005). For the NHEJ pathway, as a member of the phosphatidylinositol-3-kinase like kinase (PIKK) family, the catalytic subunit of DNA-dependent protein kinase (DNA-PKcs) is the core factor that mediates end processing during the repair. DNA-PKcs are recruited to the damage sites by Ku70/80 complex and form the DNA-PK holoenzyme to fully function during the repair. Of note, interfering the DNA repair factor in company with radiotherapy and chemotherapy is an attractive therapeutic strategy for cancers. As a matter of fact, several DNA-PKcs inhibitors have shown potential therapeutic effects for various cancers (Liang et al., 2022).

LncRNAs that do not code for proteins have emerged as regulators of numerous biological processes including maintaining chromosome homeostasis. The human genome contains large number of lncRNAs and several studies showed abnormal expression of lncRNAs in human cancer. For instance, lncRNA MALAT1 level is upregulated in a broad type of tumors (Statello et al., 2021). LncRNAs could be induced by DNA damage and several reports showed that lncRNAs participated in the DNA damage responses by different mechanisms including recruiting or stabilizing DNA repair factors (Hu et al., 2020) or preventing ATR mediated RPA2 phosphorylation (Oo et al., 2022). DNA damage-induced lncRNAs (dilncRNAs) were transcribed around DSBs sites and subsequently digested by the endoribonucleases Drosha and Dicer into small DNA damage response RNAs (DDRNAs) that promoted the recruitment of repair proteins (Michelini et al., 2017). Specifically, lncRNA BGL3 was recruited to damage sites by PARP1 and then mediated the interaction between BARD1/BRCA1 complex and HP1γ and RAD51, resulting the retention of the BARD1/ BRCA1 at damage sites. So that targeting BGL3 in combination with PARP inhibition might have effects in killing tumor cells (Hu et al., 2020). Moreover, lncRNA SNHG12 interacted with DNA-PKcs and facilitated its binding with Ku70/80. Consequently, knockdown of SNHG12 levels by treatment with SNHG12-gapmeR leads to increased gH2AX in both ECs cells and in mouse cells (Haemmig et al., 2020).

Proximity labeling is a powerful approach to characterize molecular interactomes. In principle, promiscuous enzymes (such as BioID2, TurboID or APEX2) biotinylates molecules in its proximity (within a range of 1–20 nm, depends on different enzymes), and the labeled molecules could be isolated by streptavidin beads and analyzed by either mass spectrometry (M.S.) or nucleic acid sequencing (Qin et al., 2021). Comparing with traditional co-purification strategy, such as co-immunoprecipitation (Co-IP) or chromatin immunoprecipitation (ChIP), biotin-proximity labeling can detect relatively weak or transient interactions, without crosslinking. In addition, Li et al. developed a method (namely AMAPEX) that used antibody-mediated protein A-APEX2 that could identify molecules surrounding post-translationally modified proteins, such as methylated histones (Li et al., 2022). Since phosphorylated H2AX at Ser-139 is an early marker of DNA damage (Mah et al., 2010), we applied the AMAPEX method against gH2AX to identify new players in DNA damage repair.

The key component of the AMAPEX strategy is a protein A-APEX2 fusion protein (pA-APEX2). The decoy (either post-translationally modified or not) is recognized and bound in situ by specific antibodies, which then tether the pA-APEX2. Activation of APEX2 by H2O2 treatment results in biotinylation of surrounding biomolecules in range of ~ 20 nm (Fig. 1A). In our cases, briefly, we generated DNA damage by treating HeLa cells with 10 Gray (Gy) irradiation and incubated the cells with antibodies against gH2AX 5 min after irradiation. After biotinylation, we used streptavidin beads to purify the labeled molecules, which contained both protein and nucleic acids. Since we only focused on RNA in this project, we reverse transcribed the purified RNAs and sent the samples for lncRNA sequencing (Fig. 1B). To validate the enzymatic activity of the pA-APEX2, we subjected HeLa cells to UV laser micro-irradiation and fixed the cells 5 min later and subsequently incubated the cells with gH2AX antibody, followed by pA-APEX2 treatment. As expected, cells that were micro-irradiated showed clear DNA damage stripes. After washing out unbound pA-APEX2, we treated the cells with biotin-phenol (BP) and H2O2 to introduce biotinylation. The co-localization of biotin and gH2AX suggested that pA-APEX2 could be recruited to DNA damage sites and be activated in situ (Fig. 1C). As a negative control, cells that were not treated with H2O2 did not have biotin stripes. As previously described in Fig. 1B, we enriched the biotinylated molecules with the streptavidin beads and reverse transcribed the RNA into cDNAs. The samples were sent for sequencing and the results reproduced some known lncRNA such as BGL3 that was reported to play important roles in DNA repair. SNHG12 was another lncRNA detected in our assay (Fig. 2A). To validate the sequencing results, we designed primers and performed PCR using the elution from the streptavidin beads. LINP1 was set as positive control (Wang et al., 2018) while GAPDH was a negative control. As we could see, similar to LINP1, SNHG12 level was significantly increased after IR treatment. And the enrichment was caused by the addition of pA-APEX2, since we did not see any signals without the addition of H2O2 or antibody (Fig. 2B). Altogether, our results demonstrated that AMAPEX could be used to identify the proximal RNAs at DNA damage site. As previously reported, SNHG12 interacted with DNA-PKcs and regulated DNA damage in the vessel wall and played important roles in vascular senescence (Haemmig et al., 2020). However, whether SNHG12 play any roles in human cancer cells are not yet known. To explore the potential function of SNHG12 upon DNA damage response, we ordered two independent siRNA to knock down SNHG12 and transfected the siRNA into HeLa and HCT116 human cancer cell lines. The cells were then treated with irradiation at different doses ranging from 0 to 8 Gy, cell number was count 6 days after the irradiation. Downregulation of SNHG12 significantly resulted in cellular hypersensitivity to IR (Fig. 2C).

Fig. 1

Schematic model and verification of AMAPEX labeling. A Activation of APEX2 results in biotinylation of surrounding biomolecules in range of ~ 20 nm. B Process of proximity RNA labeling around DNA damage sites. C Staining γH2AX and biotin using corresponding antibodies after UV mediated DNA damage of HeLa cells (scale bar, 10 μm). pA protein A, APEX2 ascorbate peroxidase 2, decoy target protein, B biotinylated biomolecules, IR ionizing radiation, dsDNA double-stranded DNA, BP biotin-phenol

Fig. 2

Proximity RNA labeling reveals functions of lncRNA in DNA damage response. A Volcano plot to show the sequencing results of biotinylated RNA. HeLa cells with or without irradiation were sent for RNA sequencing. Genes in upper left and right quadrants are significantly differentially expressed. The red dots represent significantly up-regulated genes, the blue dots represent significantly down-regulated genes. B Identification of SNHG12 around DNA damage sites. C Knock down of SNHG12-sensitized HeLa and HCT116 cells to irradiation

In conclusion, we used a powerful proximity RNA labeling method AMAPEX to look for lncRNAs that played roles in DNA damage repair. This method, for the first time, enabled us to map biomolecules that are proximal to gH2AX. We verified some lncRNAs detected in our assay and the results were consistent with previous publications. In addition, we proved that knockdown of lncRNA SNHG12 sensitized tumor cells to irradiation, shedding lights on the cancer therapeutic strategies by inhibiting lncRNA. What’s more, in our assay, we also detected some lncRNAs that were not reported before and these could be of great interests to the field and we would also further explore the effects of these lncRNAs in DNA damage repair (Fig. 2A).