RNF4 and USP7 coordinate spatial regulation of SLX4 stability within the PML nuclear bodies

Abstract

To protect the genome from the formation of DNA breaks by nucleases involved in DNA repair, cells have evolved multiple levels of regulatory strategies. One key regulator of nuclease activity is the scaffold protein SLX4, which plays important roles in repairing DNA damage induced by mitomycin C (MMC) and camptothecin (CPT) as well as in the resolution of stalled replication forks. Since SLX4 regulates the activity of nucleases such as SLX1, MUS81, and XPF, whose uncontrolled activity could jeopardize genome integrity, the protein level and localization of SLX4 must be tightly regulated. Here, we show that the ubiquitin E3 ligase RNF4 is associated with SLX4 and is responsible for the ubiquitin-dependent proteasomal degradation of excessive SLX4 under normal conditions. Conversely, promyelocytic leukemia nuclear bodies (PML NBs) promote SLX4 stability. In PML NBs, the stability of SLX4 is maintained by the deubiquitinase USP7, managing the amount of SLX4 necessary for a rapid response to DNA damage. These findings suggest that SLX4 and its associate nucleases are confined within PML NBs and that the optimal protein level of SLX4 is maintained by the coordinated activities of RNF4 and USP7. Our findings provide insight into how cells effectively control the potentially harmful activities of nucleases in the absence of DNA damage by a spatial regulatory mechanism.

Graphical Abstract

Graphical Abstract.

Introduction

The DNA damage response (DDR) and DNA repair pathways maintain genome integrity, which plays an important role in preventing tumorigenesis [1]. In response to DNA damage, factors involved in genome maintenance sense DNA damage and subsequently mediate signal transduction or facilitate the assembly of protein complexes for DNA repair. Upon clearance of DNA damage, DDR needs to be deactivated, and DNA repair components dissociated from undamaged DNA so that cells can regain a normal functional state [2, 3]. Similarly, under normal condition, DDR should be suppressed, and the DNA repair machinery should be kept away from intact DNA. Although investigations of genome maintenance have thus far focused on the activation of the DDR and DNA repair pathways, the molecular mechanisms underlying the controlled inactivation of these pathways in the absence of DNA lesions are less well studied.

A class of enzymes particularly threatening to intact DNA are endonucleases, which mediate nucleolytic DNA lesions in various DNA repair pathways, including nucleotide excision repair (NER), interstrand DNA crosslink (ICL) repair, and resolution of stalled replication forks. For example, within the NER pathway, XPF-ERCC1 heterodimer and XPG mediate incision 5′ and 3′ to lesions, respectively, and their activity is regulated by additional proteins in the NER complex such as XPA and replication protein A [4]. For ICL repair and resolution of stalled replication forks, the scaffold protein SLX4 is a key regulator of DNA cleavage activity, as it interacts with SLX1, MUS81-EME1, and ERCC1-XPF endonucleases [5–10]. However, given their DNA-cutting activities, precautions must be taken to prevent mislocalization of intact DNA or overexpression of endonucleases to prevent the generation of harmful DNA breaks under normal conditions. Indeed, it was reported that overexpression of flap endonuclease 1 and SLX4 induces spontaneous DNA damage, resulting in cell cycle arrest and poor cell survival [11, 12]. Therefore, a delicate regulation of the protein levels or positional sequestering of those endonucleases must exist in cells under normal condition.

Biallelic mutations in SLX4/FANCP have been identified in individuals diagnosed with Fanconi anemia [6, 13], a rare recessive disorder characterized by enhanced chromosomal instability, progressive bone marrow failure, and increased susceptibility to cancers [14]. Associated with multiple endonucleases, including XPF-ERCC1, MUS81-EME1, and SLX1, SLX4 plays a crucial role in the ICL repair pathway [5, 7–9, 15]. XPF-ERCC1 within the SLX4 complex is responsible for unhooking by creating a nucleolytic incision in ICL repair substrates [16], whereas MUS81-EME1 and SLX1 are involved in Holliday junction resolution in the last step of the repair of double-stranded breaks, including those formed at ICLs [17–19]. In addition to the role in ICL repair, the SLX4 complex has been implicated in DNA double-strand break repair, resolution of stalled replication [20, 21], and telomere maintenance [22]. Therefore, SLX4, a scaffold harboring multiple domains, has been proposed to modulate the delivery of its binding partners to different types of DNA lesions to trigger DNA cleavage and repair [10, 23]. The UBZ domain of SLX4 is responsible for recruiting the SLX4 complex to ICLs [24–26], while the SUMO-interacting motif (SIM) facilitates the recruitment of the SLX4 complex to stalled replication forks through interaction with SUMOylated targets [27]. It is worth noting that immunofluorescence imaging analysis has revealed the presence of distinct SLX4 foci even in the absence of DNA damage [6, 25, 28], suggesting that SLX4 might exist in distinct locations in the presence or absence of DNA damage. However, the functional significance of these observations is yet to be determined. Thus, understanding the molecular mechanisms governing the spatial regulation of SLX4 may provide insights into how cells manage DNA nuclease activity both in the presence and absence of DNA damage.

Promyelocytic leukemia nuclear bodies (PML NBs) are structures without a membrane that have been found to exist as biomolecular condensates by phase separation [29]. The PML proteins serve as core components and building blocks of these bodies. PML contains a SIM and multiple SUMO-modified sites that mediate interactions between several PML proteins and facilitate the formation of PML NBs [30–32]. More than 150 nuclear proteins are associated with PML NBs through SUMO-SIM interactions with PML core sequences [30, 33]. These PML-associated proteins vary in their association dynamics and mediate various cellular processes, including DNA repair [34], telomere elongation and stability [35], apoptosis [36], nuclear proteolysis [37], cell cycle, and cell survival [38]. It has been reported that SLX4 colocalizes with PML NBs in the absence of DNA damage [39]. However, the functional relevance of SLX4 localization within PML NBs has not yet been determined. Post-translational modifications play a crucial role in maintaining genomic integrity by controlling protein stability, localization, and activity through various signaling pathways [40, 41]. One such modification is ubiquitination, which adjusts protein stability by serving as a signal for proteasomal degradation [42]. Degradation is often counteracted by the reversible removal or modification of ubiquitin by deubiquitinating enzymes (DUBs). Over 100 DUBs are encoded by seven protein families [43]. One of these DUBs is USP7, and it is a PML-associated protein with the potential to regulate the stability of other PML NB-associated proteins [44].

In the present study, we demonstrate that excessive SLX4 is subjected to ubiquitin-dependent proteasomal degradation. SLX4 interacts with RNF4, which serves as an E3 ubiquitin ligase responsible for SLX4 ubiquitination. Intriguingly, we found that the lysine residues crucial for SLX4 ubiquitination were located within a segment comprising the SIMs of SLX4. Conversely, PML NBs promoted SLX4 stability. We showed that, under normal conditions, SLX4 localized predominantly within PML NBs, where it is associated with USP7, thereby increasing SLX4 stability through deubiquitination. Collectively, these results imply that SLX4 is sequestered within PML NBs and that the protein level of SLX4 is finely tuned through the coordinated actions of RNF4 and USP7, which are important for the suppression of harmful DNA breaks by DNA repair endonucleases in the absence of DNA damage.

Materials and methods

Cell lines and culture

Patient fibroblasts are immortalized using a catalytic subunit of telomerase (hTERT) and/or are transformed using HPV E6 and E7 proteins (in the text, the corresponding cell line is indicated as “EhT”). Fibroblasts were grown in Dulbecco’s modified Eagle medium supplemented with 15% (v/v) fetal bovine serum (FBS; Gibco, Franklin Lakes, NJ), 1% MEM non-essential amino acids (Gibco), 1% GlutaMAX (Gibco), and 1% antibiotic/antimycotic (Gibco) in 3% oxygen. To make the stable cell lines, cells are transfected with plasmid HA-Flag tagged versions of SLX4 WT, Δ3SIM, ΔMLR, ΔSAP, ΔUBZ, ΔTBM, and Δ3SIM-ΔTBM and selected to grow in media that additionally contained 1 μg/ml puromycin. U2OS, HeLa, and HEK293T were precultured in Dulbecco’s modified Eagle medium supplemented with 10% (v/v) FBS (Gibco, Franklin Lakes, NJ) and 1% antibiotic (Gibco) in a 5% CO₂ atmosphere at 37°C in an incubator containing 100 units of penicillin per ml.

Small interfering RNAs

All small interfering RNA (siRNA) duplexes were purchased from Integrated DNA Technologies Inc. (Coralville, IA, USA). siRNAs (20 nM) were transfected into cells using the Lipofectamine RNAiMAX reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Experiments were conducted 48–72 h post-transfection. The siRNA sequences used in this study are listed in Supplementary Table S1.

Antibodies

The antibodies used in this study are listed below, along with their respective working concentrations. For western blot analysis, rabbit anti-XPF (A301-315A, 1:1000), rabbit anti-USP7 (A300-033A, 1:1000) from Bethyl; mouse anti-GFP (sc-9996, 1:1000), mouse anti-MUS81 (sc-5332, 1:1000), mouse anti-Ubiquitin (P4D1) (sc-8017, 1:1000), rabbit anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; sc-25778, 1:2000), mouse anti-GAPDH (sc-47724, 1:2000), mouse anti-actinin (sc-17829, 1:2000), mouse anti-c-Myc (sc-40, 1:1000), mouse anti-EFP(TRIM25)(E-4) (sc-166926, 1:1000), mouse anti-ERCC1 (sc-17809, 1:1000) from Santa Cruz Biotechnology; mouse anti-HA (MMS-101R, 1:1000) from BioLegend; mouse anti-SLX4 (ab169114, 1:500) from Abcam; mouse anti-RNF4 (H00006047-A01, 1:1000) from Abnova; mouse anti-phospho-Histone H2A.X(ser139) (JBW301, 1:500) from EMD Millipore; rabbit anti-BTBD12 (NBP-2879, 1:500) from Novus; mouse anti-IgG (H + L) (115-035-003, 1:2500), rabbit anti-IgG (H + L) (115-035-003, 1:2500), mouse anti-IgG Light chain (115-035-174, 1:2500), and rabbit anti-IgG Light chain (211-032-171, 1:2500) from Jackson ImmunoResearch.

Antibodies for immunofluorescence analysis, rabbit anti-HA (C29F4, 1:200) from Cell signaling; rabbit anti-ERCC1 (FL-297) (sc-10785, 1:200), mouse anti-ERCC1(FL-297) (sc-17809, 1:200), mouse anti-PML(PG-M3) (sc-966, 1:500), and mouse anti-USP7 (sc-137008, 1:500) from Santa Cruz Biotechnology; goat anti-HA (NB-600-362, 1:500) from Novus; rabbit anti-HAUSP(USP7) (ab190183, 1:500), Donkey Anti-Goat IgG H&L (Alexa Fluor® 647) (ab150131, 1:2500), Donkey Anti-rabbit IgG H&L (Alexa Fluor® 488) (ab150061, 1:2500), Donkey Anti-rabbit IgG H&L (Alexa Fluor® 594) (ab150064, 1:2500), Donkey Anti-mouse IgG H&L (Alexa Fluor® 488) (ab150109, 1:2500), and Donkey Anti-mouse IgG H&L (Alexa Fluor® 594) (ab150112, 1:2500) from Abcam.

Mutagenesis and cloning

The multi-point mutated SLX4 vector was generated using the Q5® Site-Directed Mutagenesis Kit (NEB #E0554) using the pDONR223-SLX4 template. The SLX4-K1 (K1081R, K1083R) site mutations were generated using the QuickChange II XL Site-Directed Mutagenesis Kit (Agilent Genomic), according to the manufacturer’s instructions. The other K-site mutation (SLX4-Δ3SIM-Δ13KR and SLX4-mutant vector) was introduced sequentially using a mutagenesis kit (NEB #E0554). Gateway Technology was used to clone SLX4 and SLX4 mutant vectors, which were then cloned into pDONR223 (Invitrogen). Each of the pDONR223 derivatives was inserted into the Gateway-compatible destination vectors HA-Flag tag, GFP-tag, or Myc tag for protein expression using LR clonase. The primer sequences used for mutagenesis are listed in Supplementary Table S2.

RNA isolation and quantitative real-time PCR

Real-time polymerase chain reaction (RT-PCR) was performed using the QuantStudio 6 Pro (Thermo Fisher Scientific) at Gyeongnam Bio and Anti-aging Core Facility Center. Total cell RNA was extracted using RNeasy Mini Kit and QIAshredder (QIAGEN) and synthesized into complementary DNA through the SuperScript III First-Strand Synthesis System (Invitrogen) as suggested by the manufacturer. Messenger RNA (mRNA) level was measured using 2 × qPCRBIO SyGreen Blue Mix Lo-ROX (PCRBiosystems, Wayne, PA, USA) according to the manufacturer’s instruction, and relative mRNA levels of target genes were quantified using the ΔCt method after normalization against GAPDH [45]. The primer sequences used for qRT-PCR are listed in Supplementary Table S3.

Immunoprecipitation

Cells were lysed in MCLB buffer [50 mM Tris–HCl (pH 8.0), 150 mM NaCl, and 0.5% NP-40] supplemented with a protease inhibitor cocktail (Roche,11836145001). Cells were sonicated, and crude lysates were cleared by centrifugation at 13 000 rpm at 4°C for 20 min, and supernatants were incubated with 4 μg of antibody overnight followed by incubation with protein A-agarose beads (Cytiva) for 2 h at room temperature [46]. The immunocomplexes were washed five times with MCLB buffer and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). western blotting was performed using the antibodies indicated in the figure legends.

Ubiquitin assay for denaturing condition

For HA- or Myc-tagged protein denaturing conditions, cells were lysed by adding 5–10 cell volumes of 2% SDS-containing Tris buffer and vortexing vigorously. The mixture was incubated at 95°C for 10 min. DNA was sheared mechanically by passing the homogenate through a 19-gauge needle 20 times using a 1 ml syringe. The homogenate was centrifuged at 10 000 × g for 10 min to remove the DNA, and the supernatant was transferred to a new tube and diluted 20-fold in MLCB buffer. The mixture was then centrifuged at 10 000 × g for 10 min, and the supernatant was used directly for immunoprecipitation [47, 48].

Alternatively, for His-tagged protein purification under denaturing conditions, cells were resuspended in Buffer A (6 M guanidinium-HCl, 10 mM Tris, and 100 mM sodium phosphate [Na₂HPO₄ and NaH₂PO₄], pH 8.0) supplemented with 10 mM imidazole. The cells were then briefly sonicated, and the lysates were centrifuged at 15 000 rpm for 20 min. The supernatant was incubated with 50 μl of packed Ni²⁺-NTA agarose beads (QIAGEN) at room temperature for 2 h. The beads were washed twice with Buffer A, followed by two washes with A/T1 buffer (a 1:3 mixture of Buffer A and T1 buffer; T1 buffer contains 25 mM Tris-Cl and 20 mM imidazole, pH 6.8). The beads were then washed once with T1 buffer. Proteins were eluted using 2× protein sample buffer containing 200 mM imidazole.

Immunofluorescence and image quantification

The cells were seeded in six-well plates (SPL). Before fixing the cells, 1 μM camptothecin (CPT) or mitomycin C (MMC) was treated for 3 h to induce DNA damage. Cells were treated with 0.1%–0.2% Triton X-100 to remove cytosol and fixed with 3.7% paraformaldehyde. The fixed cells were washed with phosphate-buffered saline (PBS), permeabilized in 0.5% NP-40, and stained for indicated primary antibodies at room temperature for 2 h or overnight at 4°C [49]. After incubation with the primary antibody, the coverslips were washed three times with ice-cold PBS and further incubated with goat-anti-IgG conjugated with fluorescence at room temperature for 1 h in the dark. The cells were washed three times with PBS and stained with a vector shield. Images were obtained using a confocal microscope (Nikon A1, Tokyo, Japan, and OLYMPUS FV1000 at Core Facility for Supporting Analysis and Imaging of Biomedical Materials at Wonkwang University) using a 60× oil objective. All foci count quantification was performed using ImageJ. For measuring the foci in each nucleus, each cell was counted through the area of DAPI, and isolated foci intensity in the DAPI area was measured through “Find Maxima,” which is an ImageJ tool. Then, the value of the foci was calculated by dividing it by 250 pixels. The intensity profile of the foci was measured using NIS elements C software (Nikon).

Proximity ligation assay

A Duolink kit (Sigma–Aldrich) was used to conduct the proximity ligation assay (PLA) according to the manufacturer’s instructions. Cells were grown in a 12-well dish with coverslips, fixed, and permeabilized following an immunofluorescence protocol (previously described). The primary antibodies, PML antibody (PG-M3) (sc-966, 1:500) and HA antibody (C29F4) (CST#3724, 1:200), were added to the cells in PBS and incubated for 2 h at room temperature. The cells were then washed twice with buffer A. The minus and plus probes for the PLA were diluted 1:5 in the provided dilution buffer and incubated with the cells at 37°C for 20 min. The ligation reaction was performed at 37°C for 30 min. Following the reaction, cells were washed twice with buffer A. The amplification reaction was then conducted at 37°C for 100 min. The reaction was stopped by washing twice with buffer B. Finally, the coverslips were placed on slides with a DAPI-containing mounting medium.

Laser microirradiation and live cell imaging

U2OS cells were seeded in a glass bottom confocal dish, and cells were transfected with GFP-tagged SLX4 or PML IV expression vectors. After cells were transfected for 48 h, cells were incubated with 10 μM 5-brome-2′-deoxyuridine for 24 h before laser irradiation. Single- or double-strand breaks in the DNA were induced using a Nikon A1 laser microdissection system (Nikon). Cells were subjected to a 405 nm laser with the indicated settings [3 s/64 lines] using a 60× oil objective. After irradiation, cells were incubated in 5% CO₂, 37°C incubator for 2 h and then fixed using the methods described earlier.

Statistical analysis

Immunofluorescence images were initially processed using ImageJ or NIS software, and the resulting data were subjected to statistical analysis using GraphPad Prism 10. Statistical analyses were performed using two-tailed t-tests (unpaired) using GraphPad Prism 10. Data are expressed as median or mean ± standard deviation (ns = nonsignificant, > .05, *P < .05, **P < .01, ***P < .001, ****P < .0001).

Results

PML nuclear bodies enhance the stability of SLX4

While most proteins engaged in DDR typically form foci in response to DNA damage, there are instances in which DDR proteins form foci in the absence of DNA damage. One such protein is SLX4, which is observed to form foci under normal conditions [6, 25, 28]. Importantly, it has been reported that SLX4 relocates to sites of DNA damage [12, 24, 25, 27], suggesting that SLX4 foci may exist in distinct locations under normal conditions and in response to DNA damage. In order to more precisely determine the subcellular localization of wild-type SLX4 within the nucleus, we established immortalized SLX4-null patient-derived dermal fibroblasts (RA3331/EhT; primary RA3331 dermal fibroblasts transformed by E6E7 and immortalized by hTERT) [13, 24], stably expressing HA-tagged wild-type SLX4 (HA-WT-SLX4), and performed immunostaining of the SLX4 using an anti-HA antibody. In addition, we employed endogenous ERCC1 foci as a surrogate marker to track endogenous SLX4 because the formation of ERCC1 foci relies on and aligns with SLX4 localization [6, 25] (Supplementary Fig. S1A and B). In line with previous findings indicating the colocalization of SLX4 with PML NBs [39, 50], we employed an immunofluorescence assay to examine the colocalization of SLX4 and PML NBs. As expected, we observed that the ∼80% of SLX4 foci were localized within the PML NBs under normal condition in RA3331/EhT and HeLa cells (Fig. 1A and B and Supplementary Fig. S2). Next, we investigated whether there was a mutual influence between PML NBs and SLX4 on foci formation. To address this, we performed an immunofluorescence analysis of SLX4 foci in the presence and absence of PML NBs. Surprisingly, we found that depletion of PML led to abrogated SLX4 foci formation (Fig. 1C and D). By contrast, the depletion of SLX4 did not affect the formation of PML foci (Fig. 1E and F), implying that PML NBs may serve as upstream regulators of SLX4 localization. Next, we asked whether the loss of SLX4 foci in the absence of PML NBs could be attributed to the reduced stability of SLX4. To this end, we performed a western blot analysis of SLX4 in the presence or absence of PML NBs and found that the stability of SLX4 significantly reduced in the absence of PML NBs, suggesting that PML NBs play a positive regulatory role in maintaining the stability of SLX4 (Fig. 1G and H). Since siRNA-mediated PML depletion did not affect the transcription level of SLX4 (Supplementary Fig. S3), we hypothesized that SLX4 might be a target for ubiquitin-dependent proteasomal degradation. To test this hypothesis, we performed a western blot analysis of SLX4 in cells in which the PML protein was depleted while also treating the cells with the proteasome inhibitor MG132. Interestingly, the results revealed that treatment of MG132 enhances the stability of SLX4 in RA3331/EhT cells expressing exogenous HA-SLX4, indicating that post-translational modifications play a role in SLX4 homeostasis (Fig. 1I). This observation was further confirmed in HeLa cells where SLX4 protein expression was regulated by an endogenous promoter (Fig. 1J). We next tested whether the stability of binding partners of SLX4 is regulated by PML NBs. We found that the depletion of PML NBs did not affect the stability of SLX4-associated proteins such as XPF and MUS81 (Fig. 1G–J and Supplementary Fig. S4A–D). We noticed that PML protein levels were reduced in RA3331 cells expressing HA-SLX4-WT compared to empty vector (EV) controls (Fig. 1G and I). However, as shown in the Supplementary Fig. S4E, transient expression of SLX4 in HEK293T, HeLa, and U2OS cells did not affect PML levels (Supplementary Fig. S4E), suggesting that the effect may be specific to RA3331 cells. We speculate that chronic loss of SLX4 in these cells may have led to adaptive changes that influence PML regulation via unknown mechanisms. Taken together, our findings demonstrate that SLX4 remains associated with PML NBs under normal condition to promote its stability.

Figure 1.

PML NBs enhance the stability of SLX4. (A, B) Immunofluorescence images showing the colocalization between SLX4 and PML foci (left panel), with the corresponding intensity profile of each image indicated (right panel). RA3331 patient-derived SLX4-null cells and HeLa cells stably expressing the HA-SLX4-WT plasmid were subjected to immunofluorescence analysis with antibodies against HA, PML, or ERCC1. Scale bar, 5 μm. (C, D) Representative immunofluorescence images showing HA-SLX4 (ERCC1) and PML foci formation and quantification. RA3331 SLX4-null fibroblasts and HeLa cells stably expressing HA-SLX4-WT were transfected with control siRNA or siPML (small interfering RNAs targeting PML transcripts), followed by staining with antibodies against HA, ERCC1, and PML. More than 100 nuclei were analyzed per condition. Each dot represents the number of foci per nucleus, and the red lines indicate mean values. Scale bar, 5 μm (****P-value < .0001). (E, F) Representative immunofluorescence images showing HA-SLX4 (ERCC1) and PML foci formation and quantification (left panel). RA3331 SLX4-null fibroblasts and HeLa cells stably expressing HA-SLX4-WT were transfected with either control siRNA or siSLX4, followed by staining with antibodies against HA, ERCC1, and PML. More than 100 nuclei were analyzed per condition. Each dot represents the number of foci per nucleus, and the red lines indicate mean values (right panel). Scale bar, 5 μm (****P-value < .0001). (G) Immunoblot analysis of HA-SLX4-WT in RA3331 patient-derived SLX4-null cells stably expressing EV or HA-SLX4-WT plasmid transfected with either control siRNA or siPML. Whole-cell lysates were probed with the indicated antibodies. (H) Immunoblot analysis of SLX4 in HeLa cells transfected with either control siRNA or siPML. Whole-cell lysates were probed with the indicated antibodies. (I) Immunoblot analysis of HA-SLX4 in RA3331 SLX4-null cells expressing HA-SLX4-WT and transfected with either control siRNA or siPML. Cells were treated with 10 μM MG132 for 4 h prior to harvesting. Whole-cell lysates were probed with the indicated antibodies. (J) Immunoblot analysis of SLX4 in HeLa cells transfected with either control siRNA or siPML. Before harvesting, the cells were treated with 10 μM MG132 for 4 h. The lysates were then analyzed with the indicated antibodies.

The SUMO-interacting motifs of SLX4 are crucial for maintaining the stability of SLX4

Next, we explored the molecular mechanism underlying how PML NBs enhances SLX4 stability. Specifically, we aimed to identify the critical determinants affecting SLX4 stability. To address this, we developed a cell-based model system using immortalized SLX4-null cells RA3331/EhT. SLX4 contains multiple annotated domains, including the UBZ, MLR, BTB, TBM, SIM, and SAP, and the functional relevance of these domains has been previously reported [6, 8, 12, 22]. Notably, two different sets of SIMs have been proposed in the literature [12, 27, 39]. To distinguish between them, we refer to the SIMs identified by Ouyang et al. as SIM^ILTL, and those reported by González-Prieto et al. and Guervilly et al. as SIM^VVEV. We generated SLX4 with mutations in various domains (Fig. 2A and Supplementary Fig. S5) and established RA3331 cell lines that stably expressing individual SLX4 mutant by lentiviral transduction. We found that the majority of SLX4 deletion mutants formed foci that were confined within the PML NBs like wild-type SLX4, suggesting that the association of SLX4 with ubiquitin (UBZ), ERCC1-XPF (MLR), TRF2 (TBM), or MUS81-EME1 (SAP) is not required for its localization to the PML NBs. However, neither HA-SLX4-Δ3SIM^ILTL nor HA-SLX4-Δ3SIM^ILTL-ΔTBM formed foci or exhibited this localization pattern (Fig. 2B). This outcome could potentially be attributed to the reduced protein stability of HA-SLX4-Δ3SIM^ILTL. As expected, we observed a significant reduction in the stability and half-life of HA-SLX4-Δ3SIM^ILTL when compared with the wild-type HA-SLX4-WT (Fig. 2C and Supplementary Fig. S6). As we showed that SLX4 is a target for proteasomal degradation, we treated RA3331/EhT cells expressing HA-SLX4-Δ3SIM^ILTL with MG132. Interestingly, our findings revealed a significant increase in the stability of HA-SLX4-Δ3SIM^ILTL (Fig. 2D and see “Discussion” section). However, unexpectedly, transient overexpression of HA-SLX4-Δ3SIM^ILTL in HeLa cells showed normal foci formation (Supplementary Fig. S7A). We hypothesized that this effect might be due to the transient overexpression and, therefore, decided to examine the protein level of HA-SLX4-Δ3SIM^ILTL over a more extended period. To compare the protein stability of HA-SLX4-Δ3SIM^ILTL with wild-type HA-SLX4-WT, we conducted a transient expression experiment in HeLa cells and quantified SLX4 foci over a 4-day period. We observed significantly lower intensities of HA-SLX4-Δ3SIM^ILTL foci compared with HA-SLX4-WT foci, indicating that HA-SLX4-Δ3SIM^ILTL is indeed less stable than HA-SLX4-WT (Supplementary Fig. S7B). We further performed immunofluorescent analysis to test if the inhibition of the proteasomal-dependent proteolysis rescued the reduced foci formation of HA-SLX4-Δ3SIM^ILTL. MG132 treatment restored the formation of HA-SLX4-Δ3SIM^ILTL foci, which largely colocalized with PML NBs (Fig. 2E). Consistently, proximal ligation assays showed that while HA-SLX4-WT was found to be associated with PML NBs, HA-SLX4-Δ3SIM^ILTL was not. However, treatment with MG132 recovered the foci formation, indicating that when stabilized, the HA-SLX4-Δ3SIM^ILTL is localized within PML NBs (Fig. 2F). As mentioned previously, the three SIMs in SLX4 have been defined in prior studies (Supplementary Fig. S5). To address this discrepancy, we obtained the Δ3SIM^VVEV DNA construct [12], subcloned the HA-SLX4-Δ3SIM^VVEV and generated RA3331/EHT expressing the HA-SLX4-Δ3SIM^VVEV mutant. As shown in the Supplementary Fig. S8, HA-SLX4-Δ3SIM^VVEV showed phenotypes identical to HA-SLX4-Δ3SIM^ILTL in the Fig. 2, including reduced protein levels, shortened half-life, and MG132-rescuable foci formation within PML NBs. Transiently expressed Myc-SLX4-Δ3SIM^ILTL, as well as Myc-SLX4-WT, were co-immunoprecipitated with HA-PML IV, suggesting that SUMOylation of SLX4 is not a prerequisite for its confinement within the PML NBs (Fig. 2G). Taken together, our findings emphasize the crucial role of SUMOylation in maintaining SLX4 stability.

Figure 2.

The SIMs of SLX4 are crucial for maintaining the stability of SLX4. (A) Schematic illustrations of HA-tagged SLX4 wild-type (WT) and mutant constructs. Constructs were stably expressed in RA3331 patient-derived SLX4-null cells. EV denotes empty vector. (B) Representative immunofluorescence images showing nuclear foci of HA-SLX4 and PML (upper panel). RA3331 cells expressing the indicated constructs were stained with anti-HA and anti-PML antibodies. The number of HA-SLX4 nuclear foci per cell is quantified in the bottom panel; each dot represents an individual nucleus, and the mean is indicated by a red line. Scale bar, 5 μm (****P-value < .0001). (C) Cycloheximide (CHX) chase assay in RA3331 cells stably expressing HA-SLX4-WT or HA-SLX4-Δ3SIM^ILTL. Cells were treated with 50 μg/ml of CHX and harvested at the indicated time points. Lysates were analyzed by immunoblotting with the indicated antibodies. Quantification of SLX4 WT and Δ3SIM^ILTL mutant protein levels during cycloheximide (CHX) chase assay (right panel). Protein band intensities were normalized to actinin and presented as a percentage relative to the 0-hour time point (mean ± SEM; n = 3 biologically independent replicates). (D) Immunoblot analysis of HA-SLX4 expression in RA3331 cells stably expressing the indicated constructs (EV: empty vector). Quantification of HA-SLX4 protein levels is shown in the right panel (mean ± SEM; n = 3 biologically independent replicates). (E) Representative immunofluorescence images showing nuclear foci of HA-SLX4 and PML (left panel). RA3331 patient-derived SLX4-null cells stably expressing an EV, HA-SLX4-WT, and HA-SLX4-Δ3SIM^ILTL were treated with 10 μM MG132 for 4 h, and the cells were stained with HA and PML antibodies. Each dot in the graph represents the number of foci in the nucleus, and the mean value is indicated by a red line (right panel). Scale bar 5 μm (**** P-value < .0001). (F) Representative PLA images showing the in vivo interaction between SLX4 and PML in RA3331 cells expressing HA-SLX4-WT or HA-SLX4-Δ3SIM^ILTL (left). Cells were treated with MG132 (10 μM, 4 h) and subjected to PLA using anti-PML and anti-HA antibodies, followed by ligation and amplification. PLA signals were quantified (right); each dot represents one nucleus. Scale bar, 5 μm (****P < .0001; ***P < .001; ns, not significant, P > .05). (G) Co-immunoprecipitation of Myc-tagged SLX4-WT or SLX4-Δ3SIM^ILTL with HA-tagged PML-IV in HEK293T cells. Cell lysates were subjected to immunoprecipitation using anti-Myc antibody and immunoblotted with the indicated antibodies.

SLX4 is a target for ubiquitin-dependent proteasomal degradation

Based on the findings earlier, we hypothesized that SLX4 may be subject to degradation through the ubiquitin-dependent proteasomal pathway. To test this possibility, we performed immunoprecipitation of SLX4 under both non-denaturing conditions (conventional IP, Fig. 3A) and denaturing conditions (2% SDS, 95°C, 10 min; Fig. 3B). In addition, we carried out Ni-NTA pull-down assays under guanidine-HCl denaturing conditions (Supplementary Fig. S9A and B), which together provided conclusive evidence of SLX4 ubiquitination. We found that the concurrent presence of Myc-tagged SLX4 with elevated levels of K48-only ubiquitin molecules consistently amplified the ubiquitination of SLX4 (Fig. 3C), which implies that SLX4 undergoes K48-linked ubiquitination. To further understand the molecular basis of the ubiquitin-dependent proteasomal degradation of SLX4, we investigated which lysine residues might be ubiquitinated in SLX4. To this end, we took advantage of proteomic database to determine potential lysine residues that are ubiquitinated [51]. We noticed a lysine-rich region in SLX4 (K1081, K1083, K1093, K1099, K1101 K1112, K1120, K1140, K1144, K1167, K1169, K1179, and K1180) located in the vicinity of the three SIMs of SLX4 (Fig. 3D). Four (K1093, K1120, K1169, and K1179) of these 13 lysine residues are predicted to be ubiquitinated [51] (iPTMnet and Supplementary Fig. S10A). To test if those lysine residues were responsible for SLX4 ubiquitination, we introduced 13 K-to-R mutations into both HA-SLX4-WT and HA-SLX4-Δ3SIM^ILTL constructs. These mutated constructs were designated as HA-SLX4-Δ13KR and HA-SLX4-Δ3SIM^ILTL-Δ13KR, respectively (Fig. 3D). To compare the protein levels of those HA-SLX4-WT and HA-SLX4-Δ13KR mutants, we established RA3331 cell lines stably expressing HA-SLX4-WT, HA-SLX4-Δ3SIM^ILTL, HA-SLX4-Δ13KR, and HA-SLX4-Δ3SIM^ILTL-Δ13KR and performed a western blot analysis. As shown in the Fig. 3E, the introduction of these 13 K-to-R mutations resulted in the restoration of the protein level of HA-SLX4-Δ3SIM^ILTL, recovering it back to the level comparable to that of HA-SLX4-WT, demonstrating that the one or more of the 13 lysine residues are required for SLX4 ubiquitination. Similarly, compared with HA-SLX4-Δ3SIM^ILTL, relatively higher protein stability of HA-SLX4-Δ3SIM^ILTL-Δ13KR was observed in the transient overexpression setting (Supplementary Fig. S10B) and stably expressed cell line (Fig. 3E). Ubiquitination assays showed that Myc-SLX4-WT was ubiquitinated, whereas Myc-SLX4-Δ13KR exhibited only residual ubiquitination, significantly reduced compared to wild-type SLX4 (Fig. 3F and Supplementary Fig. S9C), indicating that ubiquitination of SLX4 largely depends on lysine residues within the cluster of 13 sites, although the specific lysines directly involved remain to be determined. Consistently, we found that the abrogated foci formation of SLX4-Δ3SIM^ILTL was fully recovered by the introduction of 13 K-to-R mutations (Fig. 3G). We also noticed that both HA-SLX4-Δ13KR and HA-SLX4-Δ3SIM^ILTL-Δ13KR foci were found outside of the PML NBs (Fig. 3G and Supplementary Fig. S10C), which implies that ubiquitination-resistant SLX4 (Δ13KR) remains stable even outside of PML NBs. To quantify the SLX4 foci located outside the PML, we measured the degree of overlap between the PML and SLX4 foci using Pearson’s correlation coefficient. As expected, compared with HA-SLX4-WT, the degree of overlap of PML and SLX4 foci in the 13KR mutant decreased (Supplementary Fig. S10D), possibly due to an increase in the number of SLX4 foci outside the PML. We next asked if the mislocalized HA-SLX4-Δ13KR might cause spontaneous DNA damage. As shown in the Supplementary Fig. S11A, transient expression of HA-SLX4-Δ13KR led to and activation of the DDR determined by phosphorylation of CHK1, CHK2, and γH2AX. Similarly, stable expression of HA-SLX4-Δ13KR in RA3331 cells resulted in a persistent activation of the DDR and slow cell proliferation (Supplementary Figs. S11B and C). Although it was reported that SLX4 is SUMOylated, the lysine residues responsible for SLX4 SUMOylation have not been identified. Therefore, we investigated whether the same lysine residues were responsible for both SUMOylation and ubiquitination. However, Myc-SLX4-Δ13KR was SUMOylated to a degree comparable as the wild-type SLX4 (Fig. 3H and Supplementary Fig. S9E), indicating that the lysine residues for SUMOylation are distinct from those involved in ubiquitination. Taken together, these data suggest that SLX4 is a target for ubiquitin-dependent proteasomal degradation and that the responsible lysine residues are located in the vicinity of the three SIMs of SLX4.

Figure 3.

SLX4 is a target for ubiquitin-dependent proteasomal degradation. (A) HEK293T cells were transfected with Myc-SLX4-WT and/or HA-ubiquitin (HA-Ub) and treated with 10 μM MG132 for 6 h prior to harvest. Myc-SLX4-WT was immunoprecipitated from cell extracts using anti-Myc antibody, followed by SDS–PAGE and immunoblotting with the indicated antibodies. (B) HEK293T cells were transfected with Myc-SLX4-WT and/or HA-Ub and treated with 10 μM MG132 for 6 h prior to harvest. Cell lysates were prepared in 2% SDS buffer, incubated at 95°C for 10 min (denaturing condition), and then subjected to immunoprecipitation with anti-Myc antibody. Samples were analyzed by SDS–PAGE followed by immunoblotting with the indicated antibodies to detect ubiquitinated proteins. In parallel, whole-cell lysates were analyzed separately by anti-Myc immunoblotting (lower panel) to confirm Myc-SLX4-WT expression levels. (C) HEK293T cells were transfected with Myc-SLX4-WT and increasing amounts of HA-K48-only ubiquitin mutant and treated with 10 μM MG132 for 6 h prior to harvest. Samples were processed as in Fig. 3B. (D) Schematic representation of the SLX4 lysine mutant constructs. Thirteen lysine (K) residues between amino acids 1078 and 1204 were mutated to arginine (R) in the HA-SLX4-Δ13KR and HA-SLX4-Δ3SIM^ILTL-Δ13KR SLX4 mutants. Red and yellow marks indicate lysine residues and SIMs, respectively. (E) Immunoblot analysis of SLX4 mutants was performed in RA3331 patient-derived SLX4 null cells stably expressing the indicated plasmid (EV: empty vector) (upper panel). Protein level was quantified by densitometry with mean ± SD (n = 3 biologically independent samples) (bottom panel). (F) HEK293T cells were transfected with Myc-SLX4-WT or Myc-SLX4-Δ13KR together with HA-Ub and treated with 10 μM MG132 for 6 h prior to harvest. Cell lysates were prepared as in panel (B). (G) RA3331 SLX4-null cells stably expressing the indicated HA-SLX4 constructs were analyzed by immunofluorescence staining for HA-SLX4 and PML (upper panel). Representative images are shown. The graph (lower panel) quantifies the number of HA-SLX4 foci per nucleus. Each dot represents an individual cell, and red lines indicate the mean. Scale bar, 5 μm (****P < .0001). (H) In vivo SUMOylation of SLX4 was assessed in HEK293T cells expressing Myc-SLX4-WT, Myc-SLX4-Δ3SIM^ILTL, or Myc-SLX4-Δ13KR followed by cell lysis and immunoprecipitation to detect ubiquitinated proteins. Samples were processed as in Fig. 3B.

RNF4 functions as the E3 ligase responsible for ubiquitination of SLX4

The next logical step in this line of research involves the identification of the ubiquitin E3 ligase for SLX4. To screen the E3 ligase, we initially took advantage of reported proteomic analyses, displaying potential interacting partners with SLX4 [52]. It was reported that the depletion of RNF4 enhanced SLX4 stability [52]. Another study showed that siRNA-mediated RNF4 silencing significantly enhances SLX4 foci formation [28]. Therefore, we tested whether SLX4 was associated with RNF4. As shown in the Fig. 4A, immunoprecipitation assays showed that exogenously expressed HA-SLX4 does indeed interact with Myc-RNF4. We confirmed the interaction between SLX4 and RNF4 using endogenous proteins (Fig. 4B), suggesting that RNF4 may function as the ubiquitin E3 ligase responsible for SLX4. Depletion of RNF4 resulted in a significant increase in endogenous SLX4 protein level in HeLa cells (Fig. 4C). Additionally, the protein level of HA-SLX4-Δ3SIM^ILTL in RA3331/EhT/HA-SLX4-Δ3SIM^ILTL cells was also significantly enhanced upon depletion of RNF4 (Fig. 4D), demonstrating that RNF4 functions as a ubiquitin E3 ligase for SLX4. Consistently, as shown in the Fig. 4E, abrogated SLX4-Δ3SIM^ILTL foci formation was fully recovered in cells depleted of RNF4. Similarly, foci formation of wild-type SLX4 was significantly enhanced in cells depleted of RNF4 (Fig. 4E). It is worth noting that, under the same experimental conditions, depletion of RNF4 led to a significant increase in the number of PML NBs, which is consistent with previous reports [53–55]. In vivo ubiquitination analysis showed that the expression of exogenous GFP-RNF4 significantly enhances the ubiquitination of both wild-type Myc-SLX4-WT and Myc-SLX4-Δ3SIM^ILTL (Fig. 4F and Supplementary Fig. S9F). We noticed that RNF4 is associated with both wild-type Myc-SLX4-WT and Myc-SLX4-Δ3SIM^ILTL as well (Fig. 4G), suggesting that SUMOylation of SLX4 is not a prerequisite for its association with RNF4. Taken together, these findings demonstrate that RNF4 functions as the E3 ubiquitin ligase responsible for SLX4 ubiquitination.

Figure 4.

RNF4 functions as the E3 ligase responsible for ubiquitination of SLX4. (A) SLX4 interacts with RNF4. HEK293T cells were co-transfected with HA-SLX4-WT and Myc-RNF4, and lysates were subjected to immunoprecipitation (IP) with anti-Myc followed by immunoblot analysis with the indicated antibodies. (B) Endogenous protein interaction between SLX4 and RNF4. HeLa cell lysates were subjected to IP using anti-RNF4 or control IgG, followed by immunoblot analysis for endogenous SLX4 and RNF4. (C) Depletion of RNF4 increases endogenous SLX4 protein levels. HeLa cells were transfected with control or RNF4-targeting siRNAs (#1, #2, #3, or pooled). Protein levels of SLX4 and RNF4 were assessed by immunoblotting; tubulin served as a loading control. (D) RNF4 knockdown stabilizes SLX4-Δ3SIM^ILTL. RA3331 SLX4-null cells stably expressing HA-SLX4-Δ3SIM^ILTL were transfected with siControl or siRNF4. Whole-cell lysates were analyzed by immunoblotting for HA and RNF4; actinin served as a loading control. (E) RNF4 depletion increases SLX4-Δ3SIM^ILTL foci formation. Representative immunofluorescence images of RA3331 SLX4-null cells stably expressing HA-SLX4-WT or HA-SLX4-Δ3SIM^ILTL, transfected with siControl or siRNF4. Cells were stained with anti-HA and anti-PML antibodies (left panel). Right panel, quantification of HA-SLX4 foci per nucleus (each dot represents one nucleus; red lines indicate the mean). Scale bar: 5 μm (****P-value < .0001, ***P < .001). (F) HEK293T cells were transfected with Myc-SLX4-WT or Myc-SLX4-Δ3SIM^ILTL together with HA-Ub in the presence or absence of GFP-RNF4 and treated with 10 μM MG132 for 6 h prior to harvest. Samples were processed as in Fig. 3B. (G) Interaction between RNF4 and SLX4-Δ3SIM^ILTL mutant. Myc-SLX4 WT (or Myc-SLX4-Δ3SIM^ILTL) and GFP-RNF4 were transfected into HEK293T cells. Cell lysates were immunoprecipitated with anti-Myc, followed by immunoblotting with anti-GFP and anti-Myc.

USP7 augments the stability of SLX4

So far, we showed that SLX4 protein levels are minimally regulated by post-translational modification. However, we observed that foci formation of SLX4 within the PML NBs remains relatively stable unchallenged conditions (Fig. 1). Therefore, we wanted to understand the molecular basis of how the stability of SLX4 is maintained within the PML NBs. To get a clue as to how PML NBs regulate the SLX4 stability, we first inspected the annotated domains of SLX4. We found that SLX4 contains seven binding motifs of the deubiquitinase USP7 (Supplementary Fig. S12) [56]. It has been reported that USP7 is located in the PML NBs [44], raising the possibility that SLX4 might be associated with USP7. To test the possible association between SLX4 and USP7, we performed immunoprecipitation assays using cells transfected with HA-SLX4 and Myc-USP7 expression vectors and found that SLX4 interacted with USP7 (Fig. 5A and B). We further confirmed that endogenous SLX4 was associated with endogenous USP7 (Fig. 5C). Immunofluorescence analysis revealed that SLX4 foci exhibited complete colocalization with USP7 and PML NBs (Fig. 5D). Next, we investigated whether USP7 plays a role in regulating SLX4 stability. To this end, we analyzed SLX4 protein level using western blotting in cells in which USP7 had been depleted. We noticed that SLX4 protein levels were significantly diminished in cells subjected to USP7 depletion using different siRNAs (Fig. 5E), demonstrating that USP7 positively regulates the stability of SLX4. Consistently, the foci formation of ERCC1, the surrogate marker for the SLX4, was significantly reduced in cells depleted with USP7 (Fig. 5F). Together, our findings suggest that USP7 might deubiquitinate SLX4, thereby enhancing its stability. To test the possibility, we performed ubiquitination assays using HEK293T cells expressing Myc-SLX4 and increasing amounts of USP7. We observed a significant reduction in SLX4 ubiquitination with increasing levels of USP7 (Fig. 5G). In addition, while the expression of wild-type USP7 led to a reduction in SLX4 ubiquitination, the expression of the catalytically inactive form of USP7 (USP7 C223S) did not, providing evidence that USP7 promotes SLX4 stability through deubiquitination (Fig. 5H and Supplementary Fig. S9G). Collectively, these findings demonstrate that USP7, located within the PML NBs, promotes SLX4 stability through deubiquitination.

Figure 5.

USP7 augments the stability of SLX4. (A, B) SLX4 interacts with USP7. HEK293T cells were co-transfected with HA-SLX4-WT and Myc-USP7. Lysates were subjected to immunoprecipitation (IP) with anti-Myc or anti-HA antibodies and analyzed by immunoblotting using the indicated antibodies. (C) Endogenous protein interaction between SLX4 and USP7. HeLa cell extracts were subjected to IP using anti-USP7 or control IgG, followed by immunoblotting. (D) SLX4 colocalizes with USP7 and PML in nuclear foci. RA3331 SLX4-null cells stably expressing HA-SLX4-WT were fixed and stained with anti-HA, anti-USP7, and anti-PML antibodies (left panel). Line-scan intensity profiles across representative foci (right panel). Scale bar, 5 μm. (E) USP7 knockdown decreases endogenous SLX4 protein levels. HeLa cells transfected with siControl or siUSP7 were lysed and analyzed by immunoblotting with the indicated antibodies. (F) Representative immunofluorescence images of ERCC1 foci formation (left panel). HeLa cells were transfected with either control siRNA or siUSP7 and subjected to staining with ERCC1 antibodies. More than 100 nuclei were counted. Each dot in the graph represents the number of foci in the nucleus (right panel). The mean value is indicated by a red line (****P-value < .0001, ***P < .001). (G) HEK293T cells were transfected with Myc-SLX4-WT together with HA-K48-only ubiquitin mutant along with increasing amounts of Flag-USP7-WT and treated with 10 μM MG132 for 6 h prior to harvest. Samples were processed as in Fig. 3B. (H) HEK293T cells were transfected with Myc-SLX4-WT together with HA-Ub in the presence of Flag-USP7 WT or Flag-USP7-C223S (catalytic dead) mutant and treated with 10 μM MG132 for 6 h prior to harvest. Samples were processed as in Fig. 3B.

The stability of SLX4 is preserved within the PML nuclear bodies

SLX4 plays critical roles in the repairing ICLs, which commonly occur outside of the PML NBs. To achieve the functional roles of SLX4, it is essential for SLX4 to move to sites of DNA damage. Indeed, there is accumulating evidence that SLX4 moves to the sites of DNA damage [12, 24–25, 27]. Therefore, we asked whether SLX4 is capable of translocating out of PML NBs to DNA lesions in response to DNA damage. We used laser strip micro-irradiation to produce a limited amount of DNA damage. As shown in the Fig. 6A, we confirmed that SLX4 moved to the laser stripes, while PML NBs failed to do so, consistent with the hypothesis that SLX4 is separated from the PML NBs in response to DNA damage. Similarly, we found that ERCC1 focus, a surrogate marker for endogenous SLX4, was colocalized with the Fok1-induced DNA double-strand break [57], while PML NBs focus was not found in the Fok1-induced DNA lesion (Fig. 6B), again consistent with the observation that SLX4 can escape from the PML NBs and localize to the sites of DNA damage. Consistently, we found that treatment of CPT increased the number of SLX4 foci, but we did not find a significant increment in the number of PML NB foci (Fig. 6C), indicating that a portion of SLX4 remaining in the PML NBs moves to the other locations, supposedly DNA lesions. In response to DNA damage, damage-induced SLX4 foci were found outside of the PML NBs. However, at the same time, we noticed that a large portion of SLX4 foci was still overlapped with PML NBs in response to DNA damage (Supplementary Fig. S13), implying that only some portion of SLX4 complex moves to the sites of DNA damage from PML NBs. We showed that SLX4 stability is maintained in the PML NBs by USP7-mediated deubiquitination. We investigated whether damage-induced SLX4 escaping from the PML NBs could be the target of ubiquitination. As shown in Fig. 6D, treatment with MMC or CPT enhanced SLX4 ubiquitination (Fig. 6D), suggesting that SLX4 outside the PML NBs might be a target of ubiquitin-mediated proteasomal degradation. Collectively, these findings suggest that in response to DNA damage, SLX4 relocates from PML NBs to sites of DNA damage, where it is subject to ubiquitin-mediated degradation once its function in DNA repair is complete.

Figure 6.

The stability of SLX4 is preserved within the PML NBs. (A) SLX4, but no PML, is recruited on the laser stripes induced by laser microirradiation. U2OS cells transfected with GFP-SLX4-WT or GFP-PML IV were subjected to laser microirradiation, fixed after 1 h, and stained with anti-γH2AX antibodies. (B) Representative immunofluorescence image of Fok1 foci or ERCC1 and PML foci formation (left panel). U2OS 2-6-3 cells were treated with Shield1 and 4-OHT to induce Fok1 nuclease. The cells were fixed and stained with ERCC1 or PML antibodies. The quantification of the Fok1 overlap with ERCC1 or PML in cells was performed by measuring Pearson’s correlation. (C) Representative immunofluorescence image of damage-induced γH2AX, HA-SLX4, and PML foci formation. RA3331 patient-derived SLX4-null cells were stably expressing HA-SLX4-WT were treated with CPT for 3 h and fixed for immunofluorescence. More than 100 nuclei were counted, and each dot in the graph represents the number of foci in the nucleus. The mean value is indicated by a red line (****P-value < .0001, ***P < .001, ns = not significant, P > .05). (D) DNA damage alters SLX4 ubiquitination. HEK293T cells were transfected with Myc-SLX4-WT and/or HA-Ub and treated with CPT or MMC for 3 h prior to harvest. Samples were processed as in Fig. 3B.

Discussion

The SLX4 is a scaffold protein associated with multiple endonucleases, which play important roles in diverse DNA repair pathways by inducing nucleolytic cleavage of DNA substrates at the sites of DNA damage [5–9, 13, 14]. The physiological importance of SLX4 has been revealed by the identification of SLX4 mutation in the human rare genetic disorder, Fanconi anemia [6, 13]. However, in the absence of DNA damage, excess or misplaced SLX4 activity might lead to harmful DNA break formation and genomic instability [12]. In this study, we addressed the question of how the activity and stability of SLX4 are spatially regulated. We investigated the molecular basis of positional sequestering and regulation of protein levels of SLX4 in cells. We propose that in the absence of DNA damage, SLX4 is stored within PML NBs to prevent degradation by RNF4-mediated ubiquitination. These findings suggest a role for PML NBs as storage compartment where of proteins remain under unchallenged conditions. The spatial regulation of SLX4 is important for genome stability to prevent the inadvertent cleavage of DNA in the absence of DNA damage. However, a basal level of SLX4 needs to be maintained for a rapid response to DNA damage and DNA repair.

It was reported that SUMO-interacting motifs of SLX4 are implicated in SUMOylation of SLX4 and its binding partner, XPF, and that this is important for clearing DNA damage and also for the formation of condensates through phase separation [28]. However, the functional role of the SLX4 SUMOylation has been largely elusive. To investigate the role of SIM in SLX4, we utilized SLX4-null patient-derived dermal fibroblasts, RA3331/EhT cells, stably expressing the wild-type SLX4 or the SLX4 mutant construct [6]. We noticed that the protein level of HA-SLX4-Δ3SIM was significantly reduced compared with that of wild-type SLX4, in contrast to transient overexpression of HA-SLX4-Δ3SIM showing the protein level comparable to the wild-type SLX4 (Supplementary Fig. S7). This difference might be because the method of creating cells that stably express proteins through the lentiviral transduction process involves adapting the mutated proteins to the cells during the selection process, which allows for a more accurate understanding of the function of the mutated proteins. It was reported that the 3SIM is associated with SUMOylated UBC9 and thus important for SLX4 SUMOylation [12]. Our findings revealed that SLX4-Δ3SIM exhibited lower protein levels, which was restored by inhibition of proteasomal degradation, suggesting that SUMOylation promotes SLX4 stability. Consistently, as shown in the Fig. 2E, HA-SLX4-Δ3SIM forms nuclear foci upon proteasome inhibition with MG132, suggesting that the SIMs in SLX4 are important for preventing its proteasomal degradation. This interpretation is further supported by the data in the Fig. 3G, showing that the stabilized HA-SLX4-Δ3SIM-Δ13KR mutant forms nuclear foci both within and outside of PML NBs. Together, these findings indicate that the primary role of the SIMs in SLX4 may lie in promoting protein stability rather than mediating its confinement to PML NBs. We found that the lysine residues responsible for SLX4 ubiquitination were located within the 3SIM region, implying that coordination of SUMOylation and ubiquitination regulates the minimal SLX4 protein levels in the absence of DNA damage. Therefore, we propose a model whereby SLX4 is SUMOylated by the interaction between SUMOylated UBC9 and 3SIM of SLX4, and a fraction of SLX4 moves to a limited number of PML NBs, while the remainder of SLX4 stays outside of the PML NBs. When the UBC9 dissociates from SLX4, the protein is a target for RNF4-mediated ubiquitination, while USP7 mediated deubiquitination in the PML NBs preserves the stability of SLX4. Another possibility is that there could be competition between occupation of SUMOylated UBC9 and ubiquitination within the 3SIM. UBC9-mediated SLX4 SUMOylation could inhibit SLX4 ubiquitination. We showed that lysine residues responsible for SLX4 ubiquitination is dispensable for SLX4 SUMOylation, suggesting that the lysine resides for SLX4 SUMOylation would exist outside of 3SIMs. To understand the function of 3SIMs, it is critical to elucidate the lysine residues responsible for SLX4 SUMOylation.

The HA-SLX4-Δ3SIM-Δ13KR mutant displayed significantly enhanced protein stability compared to HA-SLX4-Δ3SIM (Fig. 3E), suggesting that removal of both the SIMs and ubiquitination sites synergistically stabilizes SLX4. However, introducing the same Δ13KR mutations into wild-type SLX4 (HA-SLX4-Δ13KR) resulted in only a marginal increase in protein level compared to HA-SLX4-WT. The molecular basis for this difference remains unclear. It is possible that ubiquitination is only one of several post-translational mechanisms governing SLX4 stability. SLX4 is also subject to SUMOylation and phosphorylation, both of which may influence its degradation. Furthermore, alternative mechanisms such as autophagy might also contribute. Additional studies will be required to fully elucidate the molecular determinants of SLX4 stability.

PML NBs sequester factors important for the DDR and DNA repair and release them as soon as insults come to the cells [34]. MDM2 is one of the factors that is critical for DDR. It was reported that in response to DNA damage, nuclear positioning of MDM2 is regulated by PML NBs to enhances p53 stability [58]. Functions of both p53 and SLX4 are important for overcoming cellular stress, particularly in the response to DNA damage. However, given that constitutive activation of both p53 and SLX4 results in poorer cell survival, these proteins need to remain inactive under unchallenged condition [59, 60]. PML NBs thus might function as a site of first response, where proteins are sequestered and can be released quickly in response to DNA damage. In the current study, we found an important function of USP7 within PML NBs for regulation of SLX4 stability. There are multiple factors related to SUMOylation and ubiquitination in the PML NBs, and thus post-translational modification within the PML NBs shed light on the understanding biological functions of PML NBs.

RNF4, also known as the SUMO-targeted ubiquitin ligase, prefers to ubiquitinate SUMOylated proteins [52, 61] or proteins containing SUMO-like domains (SLDs) [62] through the association with SIM and the short arginine-rich motifs in RNF4. We noticed that RNF4 is the E3 ubiquitin ligase responsible for the ubiquitination of not only wild-type SLX4 but also SLX4-SIM, which is not SUMOylated. While the majority of RNF4 targets are recognized to be SUMOylated, there are instances in which RNF4 targets are not subjected to SUMOylation. It was reported that RNF4 recognizes phosphorylated oncogenes and ubiquitinates them to enhance the protein stability [63]. Another study showed that the yeast ortholog of RNF4, Slx5-Slx8, facilitates the ubiquitination of MATα2, a protein that does not undergo any post-translational modification [64]. These observations raise the possibility that RNF4-mediated SLX4 ubiquitination may be independent of SUMOylation. Another possible explanation would be the existence of unidentified SLDs in SLX4 that might serve as a docking sites for RNF4 interaction. Further studies will be required for understanding the molecular basis of RNF4-mediated SLX4 ubiquitination.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

The authors declare no conflicts of interest.

Funding

This work was supported by the National Research Foundation of Korea (NRF) grants funded by the Korean Government (MSIP) (NRF-2023R1A2C3007266, NRF-2021R1A6A1A03038890, and RS-2025-02273042 to Y.K.). This research was partially supported by and funding to pay the Open Access publication charges for this article was provided by the Korea Basic Science Institute (National Research Facilities and Equipment Center) grant funded by the Ministry of Education (No. 2021R1A6C101A564 and RS-2024-00436674), Industrial Technology Innovation Program (RS-2024-00403190) funded by the Ministry of Trade, Industry & Energy of the Republic of Korea, and Korea Drug Development Fund (RS-2024-00463605).

Data availability

All the original data, reagents, and protocols are available upon request. Correspondence and requests for materials should be addressed to Yonghwan Kim.