DGCR8 regulates multiple processes of transcription coupled nucleotide excision repair

Abstract

Ultraviolet (UV) radiation is a major environmental factor that induces DNA lesions. Cells have evolved repair pathways, in which the transcription-coupled nucleotide excision repair (TC-NER) has a central role in removing the lesions. Here we demonstrate that DGCR8, known as a crucial component in microRNA biogenesis, coordinates the UV-induced formation of the TC-NER complex by interacting with TC-NER factors. These interactions could depend on the phosphorylation of Serine 153 of DGCR8, potentially serving as a functional switch from miRNA biogenesis to the TC-NER process. Interestingly, DGCR8 is also involved in recruiting chromatin remodelers, SPT16 and SMARCA5, for the TC-NER initiation, regulating UV-induced DNA/RNA hybrids (R-loops), and modulating DNA replication through the ATR-CHK1 checkpoint pathway. These findings reveal a novel essential regulator of TC-NER independently of miRNA processing and provide new insights into the relevant biological processes and pathological mechanisms.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-38338-5.

Introduction

Ultraviolet (UV) irradiation is a potent environmental mutagen that generates DNA lesions, primarily cyclobutane pyrimidine dimers (CPDs) and 6 − 4 photoproducts (6-4PPs)^1,2. These lesions disrupt essential cellular processes and can lead to cell cycle arrest, mutagenesis, apoptosis, inflammation, photoaging, immunosuppression, and tumorigenesis^1,2. In mammalian cells, UV-induced DNA damage is predominantly repaired by the nucleotide excision repair (NER) pathway, which plays a critical role in maintaining genomic integrity³. Defects in NER underlie several genetic disorders, including xeroderma pigmentosum (XP), Cockayne syndrome (CS), and UV-sensitive syndrome (UVSS)². NER operates through two sub-pathways: global genome NER (GG-NER) and transcription-coupled NER (TC-NER). While GG-NER detects lesions throughout the entire genome via sensors such as XPC and UV-DDB, TC-NER specifically resolves transcription-blocking lesions, with RNA polymerase II (RNAPII) serving as a primary damage sensor^2–5.

We previously identified an unexpected link between TC-NER and DGCR8, a protein best known for its role in microRNA (miRNA) biogenesis⁶. DGCR8, together with Drosha, processes primary miRNAs (pri-miRNAs) into precursor miRNAs (pre-miRNAs)⁷. Mature miRNAs regulate gene expression post-transcriptionally and influence diverse cellular processes, including DNA damage responses^8–10 and the pathogenesis of numerous human diseases^11–15. Notably, both miRNAs and their miRNA-processing machinery are implicated in the UV damage response: UV exposure alters miRNA expression profiles, and cells deficient in miRNA-processing proteins exhibit increased sensitivity to UV^10,16–19. Beyond its canonical role in miRNA processing, DGCR8 also contributes to UV resistance through phosphorylation at Ser153. The phosphorylation does not impact miRNA biogenesis, strongly suggesting that DGCR8 has distinct, non-canonical functions outside of miRNA regulation. However, the molecular basis of its involvement in TC-NER remains poorly defined.

Among DNA repair pathways, NER is particularly versatile, capable of removing a wide range of structurally diverse lesions, including intra-strand crosslinks, oxidative DNA adducts, and transcriptional byproducts such as R-loops, in addition to CPDs and 6 − 4 PPs⁵. R-loops are RNA: DNA hybrid structures formed during transcription, DNA replication, and DNA repair^20–26. While R-loops serve important physiological roles in transcription termination, gene regulation, and DNA repair, excessive R-loop accumulation can interfere with replication fork progression, promoting genomic instability and contributing to diseases^22,24,27,28, including cancers and neurological disorders, such as amyotrophic lateral sclerosis type 4 (ALS4), ataxia with oculomotor apraxia type 2 (AOA2), Aicardi-Goutières syndrome (AGS), and fragile X syndrome (FXS). To regulate R-loops, cells employ a variety of mechanisms^20–25. These include mRNA processing and export factors (e.g., THO, XAB2)²¹; topoisomerases; RNA methyltransferases (e.g., METTL3, TRDMT1)^29,30; RNase H1/H2, which specifically degrade RNA within RNA: DNA hybrids, helicases that dissolve RNA: DNA hybrids (e.g., SETX, DDX5, DDX17, DHX9), and chromatin remodelers and DNA repair nucleases²¹.

TC-NER has also been implicated in R-loop processing, although this process can result in DNA double-strand breaks, thereby contributing to genome instability. UV-induced R-loops are associated with alternative splicing changes³¹, but broader molecular mechanisms linking UV-induced lesions, R-loop dynamics, TC-NER regulation, DNA replication stress, and tumorigenesis remain poorly understood. Both UV-induced DNA damage and R-loop accumulation activate key DNA damage response (DDR) pathways mediated by the ATR and ATM kinases³². ATR is recruited to RPA-coated single-stranded DNA formed under replication stress or UV damage. Once activated via autophosphorylation at Thr1989, ATR phosphorylates CHK1 to stabilize replication forks and suppress origin firing. ATM responds to double-strand breaks via recruitment by the MRE11–RAD50–NBS1 (MRN) complex and activates downstream effectors such as CHK2, p53, 53BP1, BRCA1, and MDC1 to coordinate DNA repair, cell cycle arrest, senescence, or apoptosis. Crosstalk between these pathways ensures integrated responses to genotoxic stress. The intricate interplay among UV-induced DNA damage, subsequent regulation, DNA replication stress, and ATR/ATM checkpoint pathways requires further clarification.

In this study, we uncover multiple roles for DGCR8 in the context of TC-NER. Specifically, DGCR8 is required for UV-induced assembly of the TC-NER complexes, recruitment of chromatin remodelers to sites of damage, suppression of de novo R-loop formation, and potential involvement in the ATR-CHK1 checkpoint pathway. These findings reveal that DGCR8, beyond its canonical role in miRNA biogenesis, also serves as a key coordinator of transcription-coupled DNA repair and highlight its broader importance in maintaining genomic stability under UV-induced stress.

Results

UV irradiation induces interactions between S153-phosphorylated DGCR8 and TC-NER factors

Our previous work revealed that DGCR8 interacts with the TC-NER factors, RNA polymerase II (RNAPII) and CSB; however, whether these interactions are UV-dependent remained unresolved. To investigate this, we employed proximity ligation assays (PLA), which enable sensitive detection of endogenous protein–protein interactions (Fig. 1a). We first examined the UV dependency of DGCR8-TC-NER factor interactions. The PLA revealed UV-dependent interactions between DGCR8 and both RNAPII and CSB, core TC-NER components (Fig. 1b, c, left). These interactions were confirmed by immunoprecipitation in U2-OS or human keratinocyte HaCaT cells (Fig. S1a), although the sensitivity of immunoprecipitation is much lower than that of PLA. To test whether Ser153 phosphorylation of DGCR8 mediates these interactions, we generated genome-edited U2-OS cells expressing an S153A mutant (serine replaced with alanine). The S153A mutation abolished the UV-induced band shift and PLA signals specific to phosphorylated DGCR8, despite comparable protein levels to wild-type DGCR8 (Fig. S1b-d). In these S153A clones, UV-induced DGCR8-RNAPII and DGCR8-CSB interactions were markedly reduced (Fig. 1b, c, right; Fig. S1e, f). In contrast, we further generated a phosphorylation-mimetic S153D mutant, undetectable with the anti–phospho-S153 antibody, which enhanced the DGCR8–CSB interaction under non-irradiated conditions (Fig. S1g). This result suggests the importance of S153 phosphorylation in the DGCR8-associated interactions. The UV-induced PLA signals between DGCR8 and other TC-NER proteins, CSA and UVSSA, were also diminished in S153A cells (Fig. 1d, e), while the DGCR8–USP7 interaction was less affected (Fig. 1f). These interactions were absent in DGCR8-knockout (KO) cells, confirming their specificity (Fig. S1h–l). Collectively, these results indicate that DGCR8 forms UV-induced, phosphorylation-dependent interactions with early-stage TC-NER components containing RNAPII, CSA, and CSB.

Fig. 1.

UV-induced interactions between DGCR8 and TC-NER factors require S153-phosphorylation. (a) Schematic overview of the experimental design and proximity ligation assay (PLA). (b–f) Quantification and representative PLA images showing interactions between DGCR8 and the specified TC-NER proteins. PLA signals (red) were detected in nuclei counterstained with DAPI (blue). U2-OS cells expressing either wild-type DGCR8 (left) or the S153A mutant (right) were irradiated with 20 J/m² UV-C and allowed to recover for 1, 2, or 4 h (purple dots); non-irradiated controls are shown in gray (UV-). Horizontal black bars indicate the median of each group. Asterisks indicate statistically significant differences relative to non-irradiated controls or between genotypes (*p < 0.05, **p < 0.01, ***p < 0.001). White scale bars, 10 μm.

S153 phosphorylation of DGCR8 is essential for TC-NER complex formation

We next examined the role of S153 phosphorylation in assembling the TC-NER complex. The S153A mutation disrupted key interactions that occur early in TC-NER, including RNAPII–CSB and CSB–CSA (Fig. 2a, b). The disruption of the RNAPII–CSB interaction was also confirmed in two additional S153A clones (Fig. S2a). The effects of S153A mutation on these interactions were confirmed by immunoprecipitation using DGCR8-/- MEFs with transduction of wild-type or mutant DGCR8 (Fig. S2b). In contrast, the phosphorylation-mimetic S153D mutant enhanced the RNAPII–CSB interaction under non-irradiated conditions, suggesting the importance of S153 phosphorylation in regulating TC-NER complex formation. (Fig. S2c). Other interactions, such as CSB–UVSSA and UVSSA–USP7, were moderately impaired (Fig. 2c–e), consistent with the observation that DGCR8–USP7 binding is less dependent on S153 phosphorylation. Interestingly, while RNAPII–USP7 interaction was independent of UV exposure, it still required S153 phosphorylation of DGCR8 (Fig. 2f). These findings support a model in which S153-phosphorylated DGCR8 promotes the formation of a core TC-NER complex comprising RNAPII, CSB, and CSA, necessary for subsequent recruitment of additional factors (Fig. 2g). Loss of these interactions in DGCR8-KO cells further highlights its essential role in TC-NER (Fig. S2d–i).

Fig. 2.

UV-induced TC-NER complex formation requires S153-phosphorylation. (a–f) Quantification and representative PLA images showing interactions between the specified TC-NER proteins. PLA signals (red) detected in nuclei counterstained with DAPI (blue). U2-OS cells expressing either wild-type DGCR8 (left) or the S153A mutant (right) were irradiated with 20 J/m² UV-C and allowed to recover for 1, 2, or 4 h (purple dots); non-irradiated controls are shown in gray (UV-). Horizontal black bars indicate the median of each group. Asterisks indicate statistically significant differences relative to non-irradiated controls or between genotypes (*p < 0.05, **p < 0.01, ***p < 0.001). White scale bars, 10 μm. (g) Schematic model illustrating the proposed DGCR8-centered protein interaction network in response to UV.

UV-induced DGCR8–TC-NER interactions are conserved across cell types

To assess whether UV-induced DGCR8 interactions are conserved, we generated S153A-mutant mouse embryonic fibroblasts (MEFs) by genome editing. These cells exhibited significantly reduced PLA signals for DGCR8–CSB and CSB–CSA interactions following UV exposure, despite normal DGCR8 expression (Fig. 3a, b; Fig. S3a). Similar UV-dependent interactions were observed in wild-type DGCR8-transduced MEFs and in human keratinocyte HaCaT cells (Fig. S3b, c). Notably, these interactions were impaired in both homozygous and heterozygous S153A MEFs, suggesting potential haploinsufficiency or dominant-negative effect (Fig. 3a, b). These findings indicate that pS153-dependent DGCR8–TC-NER interactions are conserved across species and cell types.

Fig. 3.

S153A mutation disrupts UV-induced interactions between DGCR8 and TC-NER factors in MEFs. (a, b) Quantification and representative PLA images showing CSB–CSA (a) and DGCR8–CSB (b) interactions in MEFs. Wild-type, heterozygous (wt/S153A), and homozygous (S153A/S153A) MEFs were irradiated with 10 J/m² UV-C (purple dots) or left untreated (gray dots), followed by 2 hours of recovery. PLA signals (red) were detected in nuclei counterstained with DAPI (blue). Horizontal black bars indicate the median of each group. Asterisks denote significant differences relative to non-irradiated or between genotypes (***p < 0.001). White scale bars, 10 μm.

DGCR8 interacts with chromatin remodelers in response to UV irradiation and May act as a molecular switch for TC-NER initiation

We next explored whether DGCR8 contributes to chromatin remodeling at TC-NER initiation sites. The PLA analyses revealed UV-dependent interactions between DGCR8 and SPT16/SMARCA5, chromatin remodelers involved in TC-NER initiation^4,33–36, implying that DGCR8 may recruit chromatin remodelers to TC-NER sites (Fig. 4a, c). To test whether Ser153 phosphorylation mediates a functional shift in DGCR8, we performed PLA using a pS153-specific antibody. The PLA analyses revealed UV-induced interaction between phosphorylated DGCR8 and CSA but not with Drosha, its canonical microprocessor partner (Fig. 4b, c). CSA knockdown confirmed the specificity of this interaction (Fig. S4a). In contrast, UV exposure modestly reduced Drosha–DGCR8 interactions (Fig. S4b), suggesting that Ser153 phosphorylation may shift DGCR8 from RNA processing to DNA repair, thereby acting as a molecular switch for TC-NER initiation.

Fig. 4.

DGCR8 interacts with chromatin remodelers and may act as a molecular switch in response to UV irradiation. (a, b) Representative PLA images showing interactions between DGCR8 and SPT16 or SMARCA5 (a) and interactions between phosphorylated S153-DGCR8 (pS153) and Drosha or CSA (b). U2-OS cells were irradiated with 20 J/m² UV-C (purple) or left untreated (gray), followed by 1, 2, or 4 h of recovery. White scale bars, 10 μm. (c) Quantification of PLA signals shown in (a) and (b). PLA signals (red) were detected and quantified in nuclei counterstained with DAPI (blue). Horizontal black bars indicate the median of each group. Asterisks denote significant differences relative to untreated cells (***p < 0.001).

DGCR8 regulates UV-induced R-loop formation

TC-NER has been implicated in regulating R-loops, DNA:RNA hybrids formed during transcription. Although UV-induced R-loop accumulation has been reported^31,37, its connection to TC-NER is not well defined. Using the S9.6 antibody to detect DNA: RNA hybrids, we investigated the role of DGCR8 in UV-induced R-loop dynamics. S9.6 immunostaining revealed R-loop accumulation following UV exposure in Dgcr8⁻/⁻ MEFs (Fig. 5a), U2-OS, and HaCaT cells (Fig. S5a). Wild-type DGCR8 reduced R-loops in Dgcr8⁻/⁻ cells, whereas S153A DGCR8 was less effective (Fig. 5a). Consistently, S153A U2-OS cells also exhibited elevated UV-induced R-loop accumulation across the nucleoplasm (Fig. 5b).

Fig. 5.

DGCR8-pS153 contributes to R-loop regulation. (a) R-loop detection by S9.6 immunostaining in Dgcr8^−/− MEFs transduced with wild-type human DGCR8, the S153A mutant, or an empty vector. Note that the vertical axis scale differs due to microscope camera settings (see Methods). White scale bars, 10 μm. (b) S9.6 immunostaining in wild-type and S153A U2-OS cells. (c) Schematic of the R-loop probe: a catalytically inactive RNase H1 mutant (D210N) fused to EGFP (mRNH) was used to visualize R-loops in live or fixed cells. (d) Colocalization of R-loops with nucleoli in non-irradiated U2-OS cells, with nucleoli labeled using Nucleolus Bright Red (Dojin). (e, f) PLA analyses of interactions between DGCR8 and R-loops (e) and between R-loops and UV-induced cyclobutane pyrimidine dimers (CPDs) (f). Cells were irradiated with UV-C (purple) at 10 J/m² (a) or 20 J/m² (b, e, f) or left untreated (gray). Red fluorescence indicates PLA or immunostaining signals; nuclei were counterstained with DAPI (blue). In (d) and (e), green fluorescence marks R-loops labeled by the mRNH probe. Horizontal black bars indicate the median of each group. Asterisks indicate significant differences relative to untreated cells or between genotypes (*p < 0.05, ***p < 0.001). White scale bars, 10 μm. (g) Schematic model of the proposed DGCR8-mediated mechanism of R-loop regulation.

To monitor R-loops in live cells, we used a catalytically inactive RNase H1 mutant (D210N) fused to enhanced green fluorescent protein (EGFP), termed mRNH (Fig. 5c), which binds but does not resolve RNA: DNA hybrids. Under non-irradiated conditions, mRNH predominantly localized to nucleoli, as confirmed by Nucleolus Bright Red staining (Fig. 5d). Following UV irradiation, PLA using anti-GFP and anti-DGCR8 antibodies revealed additional DGCR8–R-loop interaction signals that appeared across the nucleoplasm beyond nucleolar regions (Fig. 5e). In Fig. 5f, a subset of the UV-induced R-loop signals was observed in proximity to CPDs, primarily repaired by TC-NER involving DGCR8, and increased in S153A cells. These observations support the idea that S153-phosphorylated DGCR8 contributes to restricting UV-induced aberrant R-loop formation at transcriptionally active sites. The DGCR8–R-loop interactions were also confirmed in HaCaT cells via S9.6-PLA (Fig. S5b). Together, these findings indicate a critical role for DGCR8-pS153 in regulating UV-induced R-loop dynamics (Fig. 5g). We also examined whether R-loops interact with another UV-induced photolesion, 6-4PP, by using PLA (Fig. S5c). The UV-induced interaction in mRNH-expressing U2-OS cells was reduced compared to that observed for CPDs (Fig. 5f) and the interaction in the S153A mutant appeared to decline to the baseline level. These results may be explained by the rapid repair kinetics and the distinct repair pathway of 6-4PPs, as well as their relatively low abundance compared with CPDs. These aspects are further discussed below.

The S153A mutation alters ATR–CHK1 signaling in response to UV

Given the known effects of R-loops on DNA replication, we investigated whether DGCR8 influences replication dynamics using DNA fiber assays. Cells were sequentially labeled with IdU and CldU, with UV irradiation administered after IdU labeling (Fig. 6a). U2-OS cells expressing wild-type DGCR8 exhibited UV-induced replication slowdown, while S153A and DGCR8-KO cells did not (Fig. 6b; Fig. S6a). Despite expected accumulation of CPDs and R-loops, DNA replication proceeded normally in S153A cells, suggesting impaired checkpoint activation.

Fig. 6.

The S153 mutation disrupts UV-induced control of DNA replication and CHK1 activation. (a) Schematic of DNA fiber assay. U2-OS and S153A cells were sequentially labeled with IdU (red) and CldU (green) for 20 min each. UV-C irradiation (20 J/m²) was applied after IdU labeling. (b) Quantification of CldU (green) to IdU (red) track length ratios. Representative DNA fiber images are shown. Horizontal black bars indicate the median of each group. Asterisks indicate significant differences relative to untreated cells or between genotypes (***p < 0.001). (c, d) Representative images (left) and quantification (right) of ATR (c, phosphorylated T1989) and CHK1 (d, phosphorylated S345) activation in U2-OS and S153A cells, with or without 20 J/m² UV-C treatment, assessed at 1, 2, and 4 h post-irradiation. Immunofluorescence (green) was measured in nuclei counterstained with DAPI. Horizontal black bars indicate the median of each group. Asterisks indicate significant differences relative to untreated cells or between genotypes (***p < 0.001). White scale bars, 50 μm.

Immunofluorescence analysis showed comparable ATR activation (phospho-Thr1989) in wild-type and S153A cells after UV irradiation (Fig. 6c). In contrast, CHK1 activation (phospho-Ser345) was significantly impaired in S153A cells (Fig. 6d). Moderate activation of ATM together with γH2AX in S153A cells (Fig. S6b, c) suggests that unresolved replication stress may give rise to single-strand and/or double-strand DNA breaks. These data support a role for DGCR8-pS153 in coupling UV-induced DNA damage to CHK1-mediated replication checkpoint signaling.

Discussion

In this study, we demonstrated that S153-phosphorylated DGCR8 interacts with multiple TC-NER components in a UV-dependent manner and facilitates the assembly of the TC-NER core complex. We also observed UV-induced interactions between DGCR8 and chromatin remodelers. These findings support the idea that S153 phosphorylation functions as a molecular switch, redirecting DGCR8 from its canonical role in miRNA biogenesis to participation in TC-NER. Additionally, S153-phosphorylated DGCR8 regulates de novo R-loop formation and contributes to CHK1 checkpoint activation in response to UV-induced replication stress.

DGCR8 is a novel TC-NER factor with multiple interaction partners

TC-NER is initiated when elongating RNAPII stalls at DNA lesions and recruits CSB. In the absence of DNA damage, CSB transiently associates with RNAPII and promotes transcription elongation; this interaction becomes stabilized in response to DNA damage^2,4,5. The WW domain of DGCR8 constitutively binds to the phosphorylated C-terminal domain (CTD) of elongating RNAPII³⁸. Our PLA data showed enhanced DGCR8–RNAPII interactions following UV exposure in a manner dependent on Ser153 phosphorylation (Fig. 1b), suggesting a conformational change of RNAPII that may facilitate stabilization of the RNAPII–CSB complex.

The RNAPII–CSB complex then recruits CSA, which forms an E3 ubiquitin ligase complex with CUL4A, RBX1, and DDB1^2,4,5. CSA recruitment to CSB requires a 13-amino acid motif within CSB³⁹ and is regulated through multiple layers, including RPB1 ubiquitylation, association with the COP9 signalosome (CSN), and neddylation of CUL4A^40,41. Although the precise role of DGCR8 in this cascade remains to be defined, our results suggest it may modulate these events.

UVSSA and USP7 stabilize CSB by inhibiting its polyubiquitylation^42,43. It is reported that UVSSA is constitutively bound to CSA and USP7 independently of DNA damage and is recruited to the RNAPII–CSB complex through CSA^42–44. In our study, however, we observed several distinct features. First, UVSSA–USP7 and USP7–CSA interactions were only moderately UV-dependent but were impaired by the S153A mutation (Fig. 2d, e). Since these proteins interacted with S153-phosphorylated DGCR8 (Fig. 1d, e,f), DGCR8 appears to contribute to the UV-dependent assembly of the CSA–UVSSA–USP7 complex. Second, USP7 interacted with RNAPII independently of UV irradiation (Fig. 2f), suggesting a potential scaffolding role for USP7 in TC-NER complex formation. The reduction of this interaction in S153A cells implies that DGCR8 stabilizes USP7–RNAPII interactions through an undefined mechanism.

DGCR8 May coordinate chromatin remodeling during TC-NER

Chromatin remodeling is essential for TC-NER progression. SMARCA5 and NAP1L1 modulate CSB binding and function^33,36, while SPT16, a subunit of the FACT complex, facilitates H2A/H2B turnover at transcription-blocking lesions^34,35. Our results showed that DGCR8 forms UV-induced interactions with both SMARCA5 and SPT16 (Fig. 4a, c), suggesting a role in recruiting these chromatin remodelers during TC-NER initiation. These data support a broader function for DGCR8 as a dynamic scaffold that helps organize TC-NER components at sites of UV-induced DNA damage.

S153 phosphorylation acts as a molecular switch from MiRNA biogenesis to TC-NER

MEFs heterozygous or homozygous for the S153A mutation lacked DGCR8–TC-NER factor interactions (Fig. 3), raising the possibility of a dominant-negative effect. DGCR8 contains a heme-binding domain (Rhed; residues 276–498) that promotes dimerization and pri-miRNA recognition ), which forms a dimer that recognizes the structure of pri-miRNA hairpins^45,46. Although S153 lies outside the Rhed domain, the S153A mutation may impact DGCR8 dimerization, monomer–dimer transitions, or heme binding. Our data suggest that S153 phosphorylation triggers dissociation from Drosha and facilitates association with CSA (Fig. 4b, c). Drosha is phosphorylated and degraded upon cellular stress, including heat shock, oxidative stress, and serum starvation^47,48, and is translocated to the cytoplasm upon UV exposure (unpublished data). Together, these observations suggest that S153 phosphorylation of DGCR8 acts as a functional switch that redirects it from miRNA processing to DNA repair.

DGCR8 mediates regulation of UV-induced R-loops

NER endonucleases XPF and XPG can resolve R-loops in cells lacking RNA/DNA helicases or treated with camptothecin, an inhibitor of topoisomerase I⁴⁹. R-loops are also implicated in UV-induced alternative splicing³¹, yet their regulation by TC-NER remains unclear. In this study, we showed that both DGCR8 and its phosphorylation at Ser153 are required to suppress UV-induced R-loop accumulation, linking DGCR8 function to TC-NER in R-loop regulation. Notably, constitutive R-loops were found in nucleoli, while UV-induced, de novo R-loops interacting with DGCR8 localized to the nucleoplasm and frequently colocalized with CPDs. This observation suggests that pS153-DGCR8 participates in resolving R-loops at active transcription sites through TC-NER.

We also examined the potential interaction between R-loops and another UV-induced photolesion, 6-4PP. Several factors may account for the difference between CPDs and 6-4PPs in their interaction with R-loops. CPDs, predominantly repaired through TC-NER, can persist for over 24 h, whereas most 6-4PPs are repaired rapidly via global genome NER (GG-NER) within approximately three hours^50,51. Moreover, 6-4PPs represent only ~ 10–20% of the total UV-induced photolesions^52,53, further reducing the detectable R-loop–6-4PP interactions. R-loop formation also requires time for collision between transcription and DNA damage. Therefore, in many cases, 6-4PPs are often repaired before R-loops form. In S153A mutant cells, defective DGCR8-mediated TC-NER likely causes preferential accumulation of CPDs, increasing the CPD:6-4PP ratio. Consequently, collisions between transcription and 6-4PPs may be reduced, which could account for the reduced R-loop–6-4PP interaction observed in the mutant upon UV irradiation.

DGCR8 is known to interact with RNA helicases such as DDX5, DDX17, and DHX9⁵⁴, which are also involved in R-loop resolution. These interactions may underlie DGCR8-associated R-loop regulation independent of TC-NER. Importantly, mutations in these helicases are implicated in neurodegenerative diseases and cancers. For example, SETX mutations cause ataxia with oculomotor apraxia type 2 (AOA2)⁵⁵ and amyotrophic lateral sclerosis (ALS4)⁵⁶, while DDX41 mutations are associated with myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML)⁵⁶. Future studies should determine whether DGCR8–helicase interactions are regulated by UV and Ser153 phosphorylation.

An additional layer of regulation may involve N⁶-methyladenosine (m⁶A) RNA modification, the most abundant modification in eukaryotic mRNAs and non-coding RNAs. The m⁶A modifications regulate RNA metabolism and have roles in splicing, secondary structure formation, maturation, nuclear export, translation, and degradation. DGCR8 has been shown to recognize m⁶A-modified transcripts as miRNA precursors⁵³, and UV-induced m⁶A accumulation has been proposed as a marker of R-loop formation^30,37,57,58. Thus, DGCR8 may regulate UV-induced R-loops partly through m⁶A recognition.

DGCR8-pS153 contributes to CHK1 activation during UV-induced replication stress

UV-induced stalling of RNAPII and accumulation of R-loops generates RPA-coated single-stranded DNA (ssDNA), which recruits and activates ATR. RPA-coated ssDNA is also generated as an intermediate in TC-NER or results from replication fork stalling². ATR then phosphorylates CHK1, resulting in S/G2-phase arrest and suppression of CDK2 activity, thereby slowing DNA replication⁵⁹. The ATR–CHK1 axis also promotes NER activation, replication fork stabilization, and suppression of dormant replication origin firing⁵⁹.

In this study, CHK1 activation (via Ser345 phosphorylation) was impaired in both S153A and DGCR8-KO cells, despite intact ATR phosphorylation, indicating that DGCR8-pS153 may be required for effective CHK1 signaling. Although UV irradiation predominantly activates ATR, the ATM pathway can be engaged in non-cycling cells via the UV-induced displacement of spliceosomes, which facilitates R-loop formation³¹. We observed mild ATM activation in S153A cells (Fig. S6b), but whether this results from unresolved R-loops or replication fork collapse remains to be determined.

Notably, somatic DGCR8 mutations affecting miRNA processing have been identified in Wilms tumors⁶⁰. The E518K mutation is associated with familial multinodular goiter⁶¹. S153 is conserved across placental mammals. While phosphorylation at this site is unlikely to affect miRNA processing, mutations near S153, such as L152F and S156N in melanomas and S156C in skin cancers, are listed in the cBioPortal for Cancer Genomics (http://www.cbioportal.org/). These variants may disrupt the role of DGCR8 in TC-NER and R-loop regulation. Elucidating this newly identified function of DGCR8 may yield insights into mechanisms underlying cancer and neurodegenerative diseases.

Methods

Cell lines and genome editing

U2-OS cells were obtained from ATCC. Immortalized human keratinocytes (HaCaT) were kindly provided by Paul Nghiem (University of Washington). Dgcr8-knockout MEFs were purchased from Novus Biologicals and cultured according to the manufacturer’s instructions. All cells were maintained at 37 °C in a humidified 5% CO₂ incubator and cultured in DMEM supplemented with 10% fetal bovine serum (FBS). Genome editing was performed using S.p. HiFi Cas9 Nuclease V3 and tracrRNA (IDT). Single-stranded oligonucleotide donors (ssODNs) and crRNAs were designed using Benchling and synthesized by IDT. For electroporation, 2 × 10⁵ cells were transfected using the Neon Transfection System (Thermo Fisher) with 18 pmol Cas9, 22 pmol gRNA duplex, and 200 pmol ssODN in a 10 µL Neon tip under the following conditions: 1300 V, 15 ms, 1 pulse. Three days post-transfection, cells were trypsinized and clonally expanded by limiting dilution. Genomic DNA was screened by PCR-restriction fragment length polymorphism (PCR-RFLP), based on loss of the DdeI site at Ser153 of human DGCR8, and confirmed by Sanger sequencing.

Guide RNA, ssODN, and primers

Sequences of crRNAs, ssODNs, and primers for PCR-RFLP are listed in Supplementary Table 1.

Mouse embryonic fibroblasts (MEFs)

Genome editing in mice was carried out using the improved genome editing via oviductal nucleic acid delivery (i-GONAD) method^62,63. For the i-GONAD procedure, pregnant female mice were anesthetized with inhaled isoflurane delivered via a calibrated vaporizer and continuously monitored to ensure adequate depth of anesthesia. After recovery from anesthesia, the animals were returned to their home cages. Embryos or pups were screened by PCR-RFLP, based on the creation of a PvuII site at the S153A mutation and verified by Sanger sequencing. Allele-specific in silico analysis was conducted using Benchling. At the experimental endpoint, when required, all mice were humanely euthanized by cervical dislocation, which ensured rapid loss of consciousness without prolonged suffering. In the case of pregnant mice for the generation of MEFs, all fetuses were harvested from the treated animals. E12.5 embryos were obtained by crossing heterozygous male and female wt/S153A mice, washed twice in PBS with 1% penicillin-streptomycin-amphotericin B (Nacalai), minced in 0.05% Trypsin-EDTA (Gibco), and incubated at 37 °C for 20 min. The dissociated cells were neutralized in DMEM with 10% FBS and 1% penicillin-streptomycin-amphotericin B, plated in T75 flasks, and cultured for 2 days. Only cells at passage 3 or earlier were used in the experiments. All animal experiments were approved by the Institutional Animal Care and Use Committee of Tokai University (approval numbers: 191016, 202006, 213021, 224003, 235007, and 241043). All methods were carried out in accordance with the relevant institutional and national guidelines and regulations for the care and use of laboratory animals. This study was reported in accordance with the ARRIVE guidelines.

Antibodies

Rabbit polyclonal antibodies against phospho-S153 DGCR8 (human: #5626; mouse: #8889) were generated by PhosphoSolutions, as previously described⁶. The following commercial antibodies were also used: DGCR8 (10996-1-AP, Proteintech; sc-377249, Santa Cruz), RNA polymerase II (CDT-pS5, sc-55492[F-12], Santa Cruz), CSB (sc-25370, Santa Cruz), CSA (sc-376981, Santa Cruz), USP7 (300 − 033, Bethyl), UVSSA (GT816, GeneTex), CHK1-pS345 (2341 L, CST), ATM-pS1981 (4526 S, CST), SPT16 (sc-165987, Santa Cruz), SMARCA5 (sc-365727, Santa Cruz), CPD (CAC-NM-DND-001, Cosmo Bio), 6-4PP (CAC-NM-DND-002, Cosmo Bio), and GFP (50430-2-AP, Proteintech; AE012, Abclonal).

Immunofluorescence and proximity ligation assay (PLA)

Immunofluorescence was performed essentially as previously described⁶⁴. Cells were grown on ibidi µ-slides to semi-confluency, exposed to 20 J/m² UV-C, and allowed to recover for 1–4 h. Cells were fixed in 4% paraformaldehyde (Wako) and permeabilized with 0.5% Triton X-100 (Wako) for 20 min at room temperature, then blocked with 1% BSA for 30 min. After overnight incubation at 4 °C with primary antibodies, cells were washed three times with 0.05% Tween20 in PBS and labeled with Alexa Fluor 488-conjugated goat anti-rabbit IgG and Alexa Fluor 594-conjugated goat anti-mouse IgG (Thermo Fisher) for 2 h.

For proximity ligation assays (PLA), fixed cells were incubated with Duolink In Situ PLA Probes (anti-rabbit MINUS, DUO92005; anti-mouse PLUS, DUO92001; Sigma Aldrich) for 2 h at 37 °C, followed by ligation and amplification using the Duolink In Situ Detection system (DUO92008). Nuclei were counterstained and mounted with SlowFade Gold Antifade Mountant with DAPI (S36938, Thermo Fisher). Images were acquired using fluorescent microscopes, Ti2E (Nikon; Figs. 1, 2, 3, 4, 5b and d–f and 6; S1h–l; S2d–i; S4a; S5c; S6) or Axio Imager M2 (Zeiss; Figs. 5a, S1d–g; S2a, c; S3; S4b; S5a, b). The Ti2E was equipped with DS-Qi2 camera, CFI Plan Apo Lambda 40x/0.95 or 60x/0.95 objectives, DAPI/FITC/mCherry fluorescence filter sets, and a D-LEDI light source for fluorescence (Nikon). NIS-Elements AR software (Nikon) was used to create a GA3-based automated workflow for nuclear immunofluorescence or PLA signal quantification. Macro keyboards, Stream Decks, were used to facilitate semi-automated image acquisition and analysis⁶⁵. Axio Imager M2 images were analyzed using ZEN microscopy software (Zeiss) and CellProfiler cell image analysis software (Broad Institute).

Western blot analysis

Cells were washed in 0.15 M NaCl and fixed in 10% trichloroacetic acid (TCA; Nacalai) in 0.15 M NaCl on ice for 30 min. After centrifugation at 13,000 rpm for 3 min at 4 °C, pellets were resuspended in PBS, homogenized with a sonicator, Bioruptor (BM Bio), mixed 1:1 with EasyApply sample buffer (Atto), and sonicated until homogeneous. Protein concentrations were measured using Qubit Protein Assay and Qubit 3 Fluorometer (Thermo). Samples were boiled at 95 °C for 5 min, separated by SDS-PAGE on 4–20% gradient gels (UH-R420, Atto), and transferred to 0.2 μm nitrocellulose membranes (170–4158, Bio-Rad) by semi-dry blotting for 15 min (WSE-4125, Atto). Membranes were stained with 5% Ponceau S in 0.5% acetic acid for loading control, blocked with 5% skim milk in 0.1% Tween 20 in Tris-buffered saline (TBS-T) for 30 min at room temperature, and incubated overnight with primary antibodies at 4 °C. After three TBS-T washes, membranes were incubated with HRP-conjugated secondary antibodies (anti-rabbit or anti-mouse IgG, 1:25,000) for 2 h at room temperature. Chemiluminescent signals were visualized using ECL Prime and Amersham Imager 680 (Cytiva).

Immunoprecipitation

Cells in 10-cm dishes were fixed with 1% paraformaldehyde (Wako) for 10 min at room temperature. Residual paraformaldehyde was quenched with 125 mM glycine (Wako) for 5 min at room temperature. The cells were washed twice with ice-cold PBS and lysed in 1 mL of lysis solution containing 0.2 µL TurboNuclease (N0103P, Accelagen), 10 µL protease inhibitor cocktail (161-26023, Wako) and 10 µL phosphatase inhibitor cocktail (07574-61, Nacalai). The lysates were incubated for 10–20 min at 4 °C with shaking. For pre-clearing, 8 µL of Dynabeads Protein G (10004D, Thermo) was added to the lysate and rotated at 4 °C for 1–3 h. For antibody-bead coupling, 24 µL of Dynabeads Protein G was incubated with 1–2 µg of the primary antibodies or normal IgG in 0.5 mL of TBST (0.1% Tween 20 in TBS) and rotated at 4 °C for 1–3 h. After removing the TBST, the pre-cleared lysate was transferred to the coupled antibody–bead and rotated at 4 °C for 2–5 h for immunoprecipitation. To obtain whole cell extract (input), 24 µL of each lysate was mixed with 24 µL of 2× EasyApply sample buffer (AE-1430, Atto). Beads were washed three times with 1 mL of ice-cold TBST for 5 min each, resuspended in 32 µL of 1× EasyApply buffer, and stored at − 20 °C until analysis.

R-loop detection using catalytically inactive RNaseH1

A DNA construct encoding Kozak sequence, SV40 nuclear localization signal (NLS), catalytically inactive human RNaseH1 (D210N) fused to enhanced green fluorescent protein (EGFP), and SV40 poly-A signal was synthesized by GenScript and inserted into the BamHI site of pLenti-X1-PGK to generate pLenti-mutRNaseH1-EGFP. Lentiviral particles were produced in Lenti-X 293T cells (Takara) co-transfected with the construct and Lentiviral High Titer Packaging Mix (Takara) using TransIT-293 (Mirus). Media were collected after 3 days, filtered through a 0.45 μm membrane, quantified using Lenti-X GoStix Plus (Takara), and stored at -70 °C. U2-OS cells were transduced with virus and 8 µg/mL Polybrene for 6 h at 37 °C, then selected with 1 µg/mL puromycin after 3 days.

DNA fiber analysis

Cells were sequentially pulse-labeled with 20 µM 5-iodo-2′-deoxyuridine (IdU, Sigma-Aldrich) for 20 min, washed with PBS, exposed (or not) to 20 J/m² UV-C, and then pulse-labeled with 200 µM 5-chloro-2′-deoxyuridine (CldU, Sigma-Aldrich) for 20 min. Cells were resuspended in cold PBS (5 × 10⁵ cells/mL), and 2 µL were applied to APS-coated slides, air-dried, and lysed with 7 µL lysis buffer (200 mM Tris-HCl pH 7.5, 50 mM EDTA, 0.5% SDS) for 10 min. DNA was stretched by tilting the slide (15°–45°), air-dried, and fixed with methanol/acetic acid (3:1) for 10 min. DNA was denatured with 2.5 M HCl for 60 min, neutralized with PBS, and blocked with 0.1% Tween-20 in 5% BSA/PBS for 30 min. The IdU and CldU tracks were visualized using anti-BrdU antibodies (mouse B44 for IdU; rat Bu1/75 for CldU) for 3 h, followed by Alexa Fluor 488-conjugated anti-rat and Alexa Fluor 594-conjugated anti-mouse secondary antibodies for 1 h. Images were acquired using a Nikon Ti2E microscope and a 60x/0.95 objective (CFI Plan Apo Lamda; Nikon) and analyzed with NIS-Elements AR (Nikon). At least 50 replication tracks were measured per condition.

Statistical analysis

Data comparisons between two groups were performed using the Mann-Whitney U tests. For comparisons involving three or more groups, Kruskal-Wallis tests followed by Bonferroni correction were applied where appropriate. Statistical analyses were conducted using SPSS statistical software (IBM). P values < 0.05 were considered statistically significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

T.W. designed and performed experiments, analyzed data, and wrote the manuscript. D.Y. performed experiments and analyzed data. Y.T. performed experiments and prepared reagents. H.Y. performed experiments. M.O. provided technical supports. T.T. directed the study and wrote the manuscript. All authors reviewed the manuscript.

Funding

This work was supported in part by JSPS KAKENHI [15K21771 (Fund for the Promotion of Joint International Research (Home-Returning Researcher Development Research)) to T.T., 23K21303 to T.T and T.W., 19K06493, 24K15292 to T.W.], the Naito Foundation to T.T., the Uehara Memorial Foundation to T.T., Tokai University Tokuda Memorial Cancer/Genome Basic Research Grant to T.T. and T.W. and Tokai University School of Medicine Research Aid (2018, 2019, 2021 and 2023 to T.W.; 2022 and 2023 to D.Y.), 2019–2020 Tokai University School of Medicine Project Research to T.T.

Data availability

The datasets generated and/or analyzed during the current study are available in the DDBJ database under accession numbers LC909907, LC909908, LC909909, LC909910, LC909911, and LC909912.

Declarations

Competing interests

The authors declare no competing interests.