Interaction of CSN4 to DDB1 regulates its stability and function in DNA damage signaling

Abstract

The COP9 signalosome (CSN) regulates Cullin–RING E3 ligases (CRLs) through deneddylation and complex stabilization. However, the specific molecular functions of individual CSN subunits remain unclear. In this study, we identified CSN4 as a critical regulator of DDB1 stability and CRL4 function during the cellular response to DNA damage. CSN4 binds directly to DDB1, and its loss leads to the destabilization of DDB1, reduced CRL4 assembly, and impaired DNA damage-induced ubiquitination. The DDB1 K1131R mutant did not exhibit downregulated DDB1 expression, owing to CSN4 depletion. CSN4-resistant property showed different results in the DNA damage response, indicating that this protein is essential for the function of DDR-associated CRL. Our results also showed that This CSN4 deficiency induces autophagy, which is utilized for the degradation of DDB1. These findings provide new insights into the coordination between the ubiquitin–proteasome system and autophagy pathways in the DNA damage response.

Keywords : DDB1, CSN4, Autophagy-mediated degradation, Ubiquitin-proteasome system, COP9 signalosome, CRL4 ubiquitin ligase

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-025-06052-6.

Introduction

Cells respond to various stresses through the ubiquitination of stress-related proteins; these ubiquitin-mediated modifications by cullin-RING E3 ligases (CRLs) are the most abundant among the ubiquitin ligase system complexes in mammalian cells. Within the ubiquitin–proteasome system (UPS), ubiquitin-mediated protein modification by CRLs plays a pivotal role in regulating processes involved in cell cycle control and DNA repair [1–4]. Aberrant regulation of this process has been closely associated with cancer initiation and progression [5–7]. Given that CRLs govern the stability of a broad spectrum of substrates, they represent a central hub in maintaining cellular homeostasis. The CRL complex is composed of multiple subunits, with a central cullin serving as the scaffold. The cullin core is linked to a substrate receptor that binds to specific substrates via an adaptor molecule. Additionally, the complex includes a RING protein that facilitates interaction with the E2 conjugating enzyme that functions as a carrier that delivers activated ubiquitin to substrates in cooperation with E3 ligases [8]. Nedd8 modification of the cullin component is essential for E3 ligase activity because it induces conformational changes in the CRL architecture. These structural rearrangements enable efficient ubiquitin ligation of specific substrates. CRLs are regulated by reversible cycles of neddylation and deneddylation. Neddylation is catalyzed by Nedd8-conjugation enzymes, whereas deneddylation is cleaved Nedd8 from CRLs by the COP9 signalosome (CSN). This enzymatic activity regulates the ability of CRLs to ubiquitinate protein substrates that undergo proteasome-mediated degradation [9, 10]. Substrate receptor auto-ubiquitination occurs when deneddylation by CSN is inhibited, resulting in receptor degradation [9, 11].

CSN regulates CRL activity through coordinated structural rearrangements among its eight subunits. Six PCI-domain-containing subunits form a horseshoe-like structure, while CSN5 and CSN6, carrying MPN domains, form a heterodimer that sits atop the helical bundle [12–14]. However, their deneddylase activity remains latent until the CSN holocomplex engages a CRL substrate. Upon CRL binding, conformational changes initiated by CSN4 and other PCI-domain–containing subunits reposition the complex and activate the CSN5/6 catalytic module. In particular, CSN4 undergoes a shift toward CSN2 to form a clamp-like architecture that stabilizes CSN–CRL association and enables proper alignment of the catalytic site [15]. Through this regulated activation mechanism, CSN maintains ubiquitin–proteasome–dependent protein homeostasis and participates in diverse intracellular processes. In addition to the full eight-subunit complex, CSN has also been reported to form smaller “mini-CSN” assemblies or subcomplexes in vitro, although their physiological relevance remains under investigation [16]. Autoubiquitination of the substrate receptors occurs as neddylation continues. For example, autoubiquitination of F-box proteins has been detected in a Cul1-Skp1-F-box receptor E3 ligase model [9]. CSN inhibition promotes SCF self-ubiquitination and degradation [9]. Similarly, RNAi experiments revealed the degradation of receptor molecules via ubiquitination in several CSN subunits [17]. In addition, an inhibitor of CSN5, which possesses catalytic activity as a metalloproteinase, has been shown to promote the degradation of substrate receptors. CSN5i-3, a selective small-molecule inhibitor of CSN5, leads to the degradation of the substrate recognition module of the E3 ligase complex with irreversible neddylation. However, the adaptor molecule Skp1 was unaffected [18]. Furthermore, CSN6 expression stabilizes COP1, a single-subunit E3 ligase, by reducing its ubiquitination [19].

DDB1 is a component of the E3 ligase complex that carries Cul4 as a scaffold protein and various interacting molecules; in other words, DDB1 cullin-associated factors. DDB1-Cul4 E3 ligase systems are widely involved in many cellular processes, including transcription, cell cycle progression, development, and various signaling pathways [20–22]. Abnormal regulation of these multifarious regulatory factors increases genetic instability and cancer progression [6, 23, 24]. When DNA damage occurs due to UV irradiation, the DDB1-DDB2 complex removes DNA lesions through global genome repair, whereas the DDB1-CSA complex removes DNA lesions through transcription-coupled nuclear excision repair [20, 25, 26].

In this study, we identified DDB1, the adaptor component of the Cul4 E3 ligase complex, as a direct binding partner of CSN4. Loss of CSN4 not only altered the turnover of DDB1-associated substrate receptors but also reduced the stability of DDB1 itself, indicating that CSN4 contributes to maintaining the integrity of the Cul4–DDB1 ligase module. Given the central role of the Cul4–DDB1 complex in regulating key cellular pathways, including DNA damage responses, repair mechanisms, and proteostasis, we focused on defining the specific contribution of the CSN4 subunit to the control of this E3 ligase system.

Results

CSN4 binds to DDB1 independent of DNA damage

The involvement of CRLs in the response to DNA damage, including ionizing radiation (IR), has been widely studied [4, 27]. In our previous study, we identified the E3 ligase FBXW7 as a factor that interacts with the MDC1 PST motif [28]. We further detected CSN4, a subunit of CSN, which is one of the major regulators of the CRL. In the present study, DDB1 coprecipitated with CSN4 independently of DNA damage induced by UV radiation or ionizing radiation (Fig. 1A and Fig S1).

Fig. 1.

CSN4 binds to DDB1 directly. (A) HEK293T (human embryonic kidney 293 T) cells were transfected with HA-tagged CSN4 plasmid or empty plasmid for 48 h and subjected to ultraviolet (UV) irradiation (100 J/m²), followed by immunoprecipitation. (B) A schematic representation of wild-type (WT) and deletion mutants of DDB1 used in this study is shown (upper panel). DDB1 derivative proteins purified from baculovirus-infected insect cells were subjected to a pull-down assay using HA-tagged CSN4-transfected HEK293T cell lysate (lower panel). (C) HeLa cells were immunostained with the indicated primary antibodies, and secondary antibodies were labeled with oligonucleotides and subjected to proximity ligation assay (PLA). ns, not significant; Scale bar = 20 μm. (D) HeLa cells were transfected with each siRNA for 48 h, replated onto six-well plates, and after 24 h, treated with hydroxyurea (HU, 2 mM) for 2 h, followed by replacement with fresh medium

To identify which regions of DDB1 mediate interaction with CSN4, we generated a series of domain-deletion mutants (Δ1, BPA; Δ2, BPB; Δ3, BPC; and Δ4, CTD) corresponding to deletions of the first, second, and third β-propeller domains and the C-terminal domain, respectively (Fig. 1B), and purified known domain-deleted DDB1 proteins using a baculovirus expression vector system to produce several recombinant DDB1s in insect cells. Unlike full-length DDB1, no specific domain-deficient mutants are efficiently bound to CSN4. The β-propeller C domain interacted more potently with CSN4 (Fig. 1B). This interaction was further validated by an in vitro direct binding assay, which reproduced the same result (Fig. S2). We performed a PLA experiment to confirm whether the interaction between CSN4 and DDB1 depends on DNA damage. Our analyses revealed that binding occurred regardless of IR (Fig. 1C). Both DDB1 and CSN4 are required for maintaining normal cellular proliferation. Notably, the depletion of DDB1 resulted in a more substantial decrease in plating efficiency compared to the knockdown of CSN4 or MDC1, a key mediator of the DNA damage response pathway. This growth defect was further exacerbated under replication stress induced by hydroxyurea (HU), underscoring the critical role of DDB1 in sustaining cell viability under genotoxic conditions. Although HU treatment further reduced viability, the extent of this reduction was similar in DDB1- and CSN4-deficient cells (Fig. 1D). These experiments displayed that CSN4 and DDB1 interact physically under both basal and DNA-damage conditions.

CSN4 stabilizes DDB1 levels

We used cycloheximide (CHX) chase assays to evaluate DDB1 levels under CSN4 depletion. CSN4-knockdown cells exhibited reduced DDB1 expression following exposure to ionizing radiation (IR, 7.5 Gy). This downregulation was reversed by treatment with MG132 without CSN4 knockdown. This indicated the proteasome-mediated degradation of DDB1. In contrast, cells treated with CSN4 siRNA and MG132 showed lower levels of DDB1 than those treated with siCSN4 only (Fig. 2A). CSN4 depletion resulted in decreased DDB1 expression and checkpoint activation due to UV irradiation or replication stress driven by HU treatment (Fig. 2B, C). However, DDB1 depletion did not affect CSN4 expression (Fig. 2D). Although DNA damage triggered an increase in DDB1 stability, CSN4 knockdown did not result in the same increase in DDB1 stability following DNA damage (Fig. 2).

Fig. 2.

CSN4 is required for DDB1 stability. (A) HeLa cells transfected with siRNA were pretreated with MG132 (5 µM) for 2 h, then irradiated with 7.5 Gy of ionizing radiation (IR), followed by cycloheximide (CHX, 50 µg/mL) treatment. Cells were lysed at the indicated time points, and DDB1 levels were examined by immunoblotting (left panel). The right panel shows normalized and quantified DDB1 levels from the left immunoblot data. (B, C) HeLa cells were lysed at the indicated time points after UV (b) or HU (c) treatment 72 h after CSN4 knockdown. (D) HeLa cells were transfected with siRNA for DDB1 (30 nM) for 72 h and subjected to IR irradiation followed by CHX chase assay. Data represents SD from three independent experiments. Statistical significance was determined using two-way ANOVA. p < 0.001

DDB1 stability depends on CSN4, not CSN5 or CSN6

To determine whether the decrease in DDB1 expression was specific to CSN4 (Fig. 3A), we performed additional experiments to reduce the expression of other CSN subunits, specifically CSN5 and CSN6. Figure 3 depicts the temporal changes in protein levels following IR. The amount of DDB1 remained unchanged despite CSN5 knockdown (Fig. 3A). Additionally, CSN6 knockdown also did not alter DDB1 stability, and DDB1 protein level did not decrease but maintained. (Fig. 3B). These data indicate that DDB1 stability is regulated by CSN4 but may be through an alternative mechanism that may bypass CSN5 or CSN6. CSN5 or CSN6 did not reduce DDB1 protein, underscoring the specificity of CSN4 in regulating DDB1 stability (Fig. S3). Thus, we hypothesized that CSN4 acts as a DDB1-specific regulatory factor in the CSN complex.

Fig. 3.

DDB1 stability depends on CSN4. For IR challenge, cells transfected with siCSN5 (A) or siCSN6 (B) were lysed at the indicated time points after irradiation. HeLa cells were transfected with siCSN5 or siCSN6 for 72 h and lysates were prepared for immunoblotting. Data represents SD from three independent experiments. p < 0.001

CSN4 is required for the stability of E3 ligase components

The inhibition of CSN5, execution of the catalytic process of COP9 deneddylation, and CSN subunit knockdown resulted in the degradation of its substrate receptor [17, 18, 29]. Therefore, we investigated whether CSN4 depletion altered the stability of the E3 ligase components. Neddylated Cul4a levels were considerably higher than those in control siRNA-treated cells. Moreover, the stability of both DDB1 and neddylated Cul4a was markedly reduced in CSN4-knockdown cells compared to control cells (Fig. 4A). CSN4 depletion led to hyperactivation of CRLs, likely due to impaired deneddylation by the COP9 signalosome. In contrast, treatment with MLN4924, an inhibitor of the NEDD8-activating enzyme [30, 31], caused accumulation of deneddylated Cul4a, consistent with effective inhibition of cullin neddylation (Fig. 4B). Because the downregulation of CSN4 expression affects the neddylation and stability of Cul4a through COP9 dysfunction, we evaluated the effects of the downregulation of CSN expression on the protein levels of the substrate receptor module composed of adaptor and receptor molecules. Different results were obtained for different test molecules. CSN4 deficiency decreased adaptor molecules. Adaptor molecules Skp1 and DDB1, which associate with Cul1 and Cul4 E3 ligases, respectively, were both downregulated (Fig. 4C), although the level of Skp1 remained unchanged by CSN5 inhibitor [18].

Fig. 4.

CSN4 is required for DDB1 function and stability of CRL E3 ligase components. (A) HeLa cells were transfected with control or CSN4 siRNA for 72 h, exposed to IR, and subjected to CHX chase assay for the indicated times. Lysates were analyzed by immunoblotting. (B–E) HeLa cells were treated with 30 nM siCSN4 for 48 h or MLN4924 (1 µM) for 24 h and analyzed by immunoblotting. (B) Cullin 4A (Cul4a). (C) Adaptor proteins SOCS2 and Skp1. (D) Substrate recognition subunits DDB2 and SOCS2. (E) The CRL substrate cyclin E. Data represent mean ± SD of at least three independent biological replicates

CSN4 depletion decreased the expression levels of DDB2, the substrate receptor of the E3 ligase complex (Fig. S4). CSN4 knockdown also negatively activates AMBRA1 (Fig. S5), another DCAF [32], while SOCS2 remains unaffected. SOCS2 is a substrate receptor of the CRL5-mediated E3 ligase [33], which has Elongin B/C as its adaptor protein [34, 35]. In contrast to these findings, our results showed no change in cyclin E levels following CSN4 knockdown, despite observing a reduction in the substrate receptor Skp2 (Fig. 4E).

CSN4 regulates DDB1 stability through ubiquitination

DDB1 accumulated in the presence of the proteasome inhibitor MG132 in siControl-treated cells, but DDB1 levels were reduced in siCSN4-treated cells despite the presence of MG132 (Fig. 2A) (Fig. S6). We investigated the involvement of ubiquitination in the downregulation of DDB1 expression following CSN4 knockdown. As shown in Fig. 5A, the CSN4 deletion-induced decrease in abundance was due to ubiquitination. We prepared domain-deletion constructs of DDB1 and performed CHX chase experiments on each mutant to confirm the CSN4-induced ubiquitin modification. Except for the C-terminal-deleted DDB1 (ΔCTD), DDB1 levels were not only reduced by CSN4 knockdown but also showed an increased rate of reduction over time (Fig. 5B). In the case of ΔCTD, CSN4 knockdown did not reduce DDB1 levels; additionally, the amount that decreased over time was significantly reduced compared to that in other mutations (Fig. 5B). DDB1 (ΔCTD) did not show decreased levels in CSN4-knockdown cells, and it also exhibited a detectable ubiquitination signal. However, in contrast to the reduced ubiquitination observed for wild-type DDB1 under CSN4 depletion, the ΔCTD mutant did not exhibit a comparable decrease, even though its ubiquitination signal was still detectable. The reduced sensitivity to CSN4-mediated ubiquitination explains why the CTD-deficient DDB1 protein is not reduced by CSN4 knockdown (Fig. 5C). Furthermore, we generated a mutant in which the 1131 lysine residue in the CTD domain was mutated to arginine and examined the effect of CSN4 knockdown using this mutant. Unlike wild-type DDB1, the K1131R mutant escaped CSN4-mediated degradation, suggesting reduced ubiquitination compared to wild-type DDB1 (Fig. 5D, E).

Fig. 5.

CSN4 downregulation promotes DDB1 ubiquitination. (A) HeLa cells were co-transfected with CSN4-specific siRNA and HA-tagged ubiquitin plasmid. MG132 (10 µM) was added for 5 h before immunoprecipitation (IP) with anti-HA beads. (B) HeLa cells were co-transfected with siRNA for CSN4 and plasmids encoding WT or deletion mutants of FLAG-tagged DDB1. Cells were collected at the indicated time points after CHX treatment and subjected to Western blot analyses. Normalized DDB1 levels were determined. Data represents SD from three independent experiments. (C) HeLa cells were co-transfected with CSN4 siRNA and FLAG-tagged WT or ΔCTD DDB1 mutant. After 72 h, cells were treated with HU (10 mM) and MG132 (10 µM) for 5 h, followed by IP using anti-FLAG beads. (D) HeLa cells were transfected with FLAG-tagged WT or K1131R mutant DDB1 and CSN4 siRNA for 72 h. (E) HEK293T cells were transfected with HA-tagged ubiquitin and WT or K1131R mutant FLAG-tagged DDB1. IP reactions were performed using anti-HA beads and subjected to immunoblotting. Data represents SD from three independent experiments. Statistical significance was determined using two-way ANOVA. p < 0.001

Downregulation of CSN4 expression triggers an autophagic degradation of DDB1

Ubiquitination is associated with the ubiquitin–proteasome system, a major protein degradation system [34]. To determine the mechanism underlying DDB1 reduction in CSN4-depleted cells, we first examined its transcriptional status. qRT-PCR analysis confirmed effective CSN4 knockdown and demonstrated that DDB1 mRNA levels remained unchanged, indicating that the observed reduction in DDB1 protein upon CSN4 depletion occurs at the post-transcriptional level (Fig. 6A). Because CSN4 is a subunit of the COP9 signalosome that regulates CRL activity, we initially hypothesized that DDB1 degradation might occur through the UPS. However, proteasomal inhibition using MG132 failed to restore DDB1 protein levels (Fig. 6B), indicating that DDB1 turnover under CSN4 loss is not primarily mediated by the proteasome. Moreover, MG132 treatment markedly increased LC3-II levels, consistent with compensatory autophagy activation caused by proteasome blockade. Another major protein degradation system is lysosome-mediated degrading system, autophagy. It processes ubiquitinated proteins through autophagy-dependent degradation [35, 36]. Protein homeostasis is maintained through crosstalk between the two major protein degradation systems: the UPS and autophagy [10, 37, 38]. In addition, MG132 elevated activated LC3-II in a time-dependent manner (Fig. S6). LC3 puncta were also observed more intensely in cells that received both siCSN4 and MG132 than in those that were treated with MG132 alone (Fig. 6C). To further assess the contribution of lysosomal degradation, we treated cells with chloroquine (CQ), a lysosomal inhibitor. CQ partially rescued DDB1 levels specifically in CSN4-depleted cells (Fig. 6D). Immunocytochemistry also demonstrated that inhibition of autophagy by treating CQ partially restored DDB1 protein levels (Fig. 6E). DDB1 signals detected by FITC fluorescence were barely detectable after CSN4 knockdown and became evident upon CQ treatment. Each DDB1 fluorescence (green) and LC3 fluorescence (red) intensities from all captured fields were quantified and are presented (Fig. 6E). The quantitative analysis (Fig. 6E, right) confirmed that CQ treatment partially restored DDB1 in CSN4-depleted cells. Unselected ICC fields for all conditions are provided in Figure S7. These findings indicate that lysosomal processing contributes to DDB1 degradation when CSN4 is downregulated.

Fig. 6.

Autophagy is involved in DDB1 degradation by CSN4 depletion. (A) HeLa cells were transfected with control or CSN4-specific siRNA for 48 h, and mRNA levels were measured by qRT-PCR. CSN4 transcripts were efficiently reduced, whereas DDB1 mRNA levels remained unchanged. (B) HeLa cells were transfected with control or siCSN4 for 72 h and treated with MG132 (5 µM) for the indicated times. Immunoblotting showed that proteasomal inhibition did not rescue DDB1 protein levels, while LC3-II accumulation increased in response to MG132. (C) HeLa cells were treated as in (B) and subjected to immunofluorescence staining using anti-LC3 antibody. siCSN4 treatment increased LC3 puncta formation. Scale bar = 20 μm. (D) HeLa cells were transfected with siControl or siCSN4 for 48 h and then treated for 24 h with DMSO, MG132 (2.5 µM), or chloroquine (CQ, 30 µM). Immunoblotting revealed that CQ partially restored DDB1 protein abundance under CSN4 depletion. (E) HeLa cells were transfected with siControl or siCSN4 (30 nM) for 40 h and subsequently treated with MG132 (5 µM) or CQ (30 µM) for 8 h. Cells were stained with anti-DDB1 and anti-LC3 antibodies, and fluorescence intensities were quantified using ImageJ from five images per experiment across three biological replicates (15 images per condition). Scale bar = 20 μm. (F) HeLa cells were transfected with siControl or siCSN4 (30 nM) for 48 h and treated overnight with DMSO, MG132 (2.5 µM), CQ (30 µM), or MG132 + CQ. Immunoblotting demonstrated that combined MG132 and CQ treatment showed additive effects on LC3-II accumulation while modulating DDB1 levels. Protein levels were normalized to tubulin. Data are presented as mean ± SD from three independent experiments. Statistical significance was determined using two-way ANOVA (p < 0.001)

Depletion of DDB1 by siCSN4 was not rescued by MG132, which indicates that proteasomal inhibition alone is insufficient to stabilize DDB1. We next evaluated autophagy-related markers to determine whether CSN4 loss activates canonical autophagy pathways. Furthermore, it could occur a shift toward autophagy–lysosome–mediated degradation when CSN4 is downregulated. Consistent with this, LC3-II levels were elevated while p62 levels decreased, confirming enhanced autophagic flux (Fig. S8). CSN4 knockdown did not alter the level of canonical autophagy regulators such as mTOR and ATG5 (Fig. 6F). These results show that DDB1 degradation occurred without gross alteration of upstream signaling of autophagy but lysosomal-mediated events. Combined inhibition of UPS and autophagy further increased LC3-II accumulation, and CQ robustly elevated p62 levels under both siControl and siCSN4 conditions, confirming that autophagy–lysosome inhibition was effective (Fig. 6F). In addition, CHOP protein, which is typically stabilized by proteasome inhibition, showed attenuated induction under CSN4 knockdown (Fig. 6F). CHOP was undetectable under basal conditions and was strongly induced only by MG132, but the induction was attenuated in siCSN4 + MG132 compared to siControl + MG132, while CQ had no effect on CHOP (Fig. 6F). These data together support a model in which CSN4 depletion differentially impacts proteostasis: the balance of protein degradation shifts from the UPS toward the lysosomal pathway.

Regulation of DDB1 by CSN4 is required for normal cellular response to DNA damage

CSN4 depletion increased the phosphorylation of the histone H2A variant (H2AX), the histone most directly linked to DNA damage. Notably, upon DNA double-strand breaks (DSBs), H2AX is rapidly phosphorylated at serine 139 to generate γH2AX, which serves as a well-established and widely used molecular marker of DNA damage signaling [39]. This indicated that normal cellular response to DNA damage requires COP9 function (Fig. 7A). DDB1 has been implicated in DNA damage repair and regulation of various factors involved in DNA metabolism. To determine the biological significance of the K1131R mutant, we investigated whether the loss of DDB1 inhibited cell cycle progression through DNA damage. Both DDB1-depleted and DDB1 K1131R cells exhibited no cell cycle phase distribution compared to wild-type cells under normal conditions. However, DDB1-depleted cells and DDB1 K1131R mutant cells showed an increased proportion of cells in G1 phase and a decreased proportion in S phase compared with cells reconstituted with wild-type DDB1 under UV-treated conditions (Fig. 7B). HU is a widely used replication-blocking agent that induces replication stress by depleting the nucleotide pool, causing replication fork stalling and, if prolonged, fork collapse that leads to DNA double-strand breaks (DSBs) [40]. DDB1 depletion resulted in DNA damage following replication stress (Fig. 7C). The mutant DDB1—K1131R—which was resistant to CSN4 knockdown, resulted in DNA damage; consequently, the cells remained unrepaired (Fig. 7C). The irregularity of mutant DDB1 hinders cell cycle progression and DNA repair following DNA damage. Taken together, our data indicates that CSN4 supports normal DDB1 function and contributes to efficient recovery from DNA damage.

Fig. 7.

Regulation of DDB1 by CSN4 is required for normal cellular response to DNA damage. (A) HeLa cells were transfected with control or CSN4 siRNA for 72 h, exposed to IR, and subjected to CHX chase assay. Lysates were analyzed at the indicated time points by immunoblotting. (B) HeLa cells were co-transfected with siCSN4 and siRNA-resistant WT or K1131R mutant DDB1 constructs. The next day, cells were selected with puromycin for 48 h, UV-irradiated, and collected after overnight incubation for cell cycle analysis by flow cytometry. (C) HeLa cells were co-transfected as in (B), selected with puromycin for 48 h, treated with HU (2 mM) for 2 h, and harvested for comet assay. Representative comet images are shown (upper panel). Scale bar = 50 µm. Tail moments were quantified from >50 cells per condition. Data represent mean ± SD. p < 0.001.

Discussion

The interaction between the CSN and CRL-NEDD8 is mediated by structural rearrangements that enable the complex to clamp onto CRLs, particularly through CSN2 and CSN4. Conformational changes initiated at CSN4 promote repositioning of the CSN5/6 catalytic module and facilitate activation of CSN5 upon binding to NEDD8 [41]. These dynamic transitions highlight the essential contribution of CSN4 to both the structural and functional regulation of CRL complexes. In CRL2, CSN2 and CSN4 act as clamps, with the C-terminus of CSN2 interacting with the complex. Deneddylation requires a structural shift in CSN5/6, highlighting the dynamic nature of the CSN during CRL regulation.

In this study, we found that CSN4 not only clamps the cullin C-terminus but also directly interacts with the adaptor protein DDB1. This constitutive interaction is critical for maintaining DDB1 stability. CSN4 depletion markedly reduced DDB1 protein levels without affecting its transcription and increased ubiquitination of wild-type DDB1, whereas DDB1 lacking the CTD or harboring the K1131R mutation was largely refractory to CSN4 loss. These findings indicate that CSN4-dependent ubiquitination of the CTD region is required for proper DDB1 turnover. Because DDB1 recruits diverse substrate receptors for CRL4 [20, 22, 32], its destabilization by CSN4 loss resulted in compromised checkpoint activation and impaired repair in response to IR, UV exposure, and replication stress. DDB1 mutants resistant to CSN4 regulation likewise failed to support an adequate DNA damage response. Thus, although the CSN4–DDB1 interaction remains intact irrespective of DNA damage, its functional importance becomes particularly evident under genotoxic stress.

The CSN is composed of eight subunits, and subassemblies such as the mini-CSN complex have been reconstituted in vitro with partial deneddylase activity [12, 13]. CSN4 is an essential component of these assemblies, suggesting that its loss may disrupt both mini-complex function and CSN–CRL super-complex organization. While canonical CSN regulation involves CSN4-dependent activation of CSN5/6-mediated deneddylation [41], our data indicate that DDB1 stability is regulated by CSN4 through a mechanism that may operate independently of CSN5 or CSN6 (Fig. 3 and Fig. S3), revealing a unique CSN4-specific role in controlling CRL4 integrity.

The proteasome and autophagy systems are the two major cellular protein degradation systems. The COP9 complex is remarkably similar to the 19S lid component of the proteasome [42–44], and the possibility that CSN can replace the lid of the 26S proteasome has been investigated [45, 46]. CSN4 depletion increased ubiquitination and degradation of DDB1 and Skp1, suggesting selective disruption of adaptor stability within E3 ligase modules. However, blocking the proteasome with MG132 failed to restore DDB1 levels, whereas inhibition of the autophagy–lysosome system by chloroquine partially rescued DDB1. CSN4 knockdown also elevated LC3-II and reduced p62, consistent with enhanced autophagic flux. These results indicate that although DDB1 degradation is ubiquitination-dependent, its downstream processing shifts from the proteasome to the autophagy–lysosome pathway under CSN4-deficient conditions. This redistribution was further supported by attenuated CHOP induction upon proteasome inhibition, suggesting reduced UPS dependency and compensatory lysosomal engagement.

Although DNA damage is known to induce autophagy, the autophagy-associated changes observed in CSN4-depleted cells occurred independently of DNA damage signaling. CSN4 knockdown did not activate canonical autophagy initiation markers such as mTOR or ATG5 but altered the proteolytic balance such that autophagy became the dominant route for DDB1 clearance. This shift reflects proteostasis rerouting rather than autophagy induction per se. Recent evidence reinforces this interpretation. A newly reported study demonstrated that CSN2 coordinates ubiquitin-linked substrate tagging and autophagic flux, providing strong support for the notion that COP9 signalosome subunits function at the interface between the UPS and autophagy [47].

Collectively, our data show that CSN4 is essential for maintaining DDB1 stability, preserving CRL4 function, and enabling appropriate DNA damage signaling. CSN4 loss disrupts this balance and shifts DDB1 degradation toward the autophagy–lysosome system. Elucidating how the CSN–CRL super-complex coordinates or redirects protein degradation between UPS and autophagy will be critical for understanding genome maintenance mechanisms and may provide insights for therapeutic strategies targeting cancers driven by genomic instability. Future studies dissecting how CSN4 directs this routing decision will be essential for establishing the mechanistic basis of this regulatory shift, as the precise mechanism remains to be fully defined.

Materials and methods

Cell culture and chemicals

HEK293T and HeLa cells were obtained from ATCC and cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin at 37 °C in a 5% CO₂ incubator. Panc-1 cells were obtained from ATCC and cultured in RPMI-1640 supplemented with 10% FBS and 1% penicillin/streptomycin at 37 °C in a 5% CO₂ incubator. MG132 (Z-Leu-Leu-Leu-al, proteasome inhibitor) [48], chloroquine (CQ), cycloheximide (CHX), and hydroxyurea (HU) were purchased from Sigma-Aldrich (USA).

RNA interference and quantitative RT-PCR

Cells were transfected with the indicated siRNAs using RNAiMAX (Invitrogen, USA) or TransIT-X2 (Mirus Bio) according to the manufacturer’s instructions. siRNAs against human DDB1, CSN4, CSN5, CSN6, or negative control were purchased in smartpool format from Dharmacon (USA). For qRT-PCR, total RNA was extracted using HiGene Total RNA prep kit (Biosesang, Seoul, Korea) and reverse-transcribed with RT kit (One Step SYBR PrimeScript RT-PCR kit, TaKaRa-bio, Japan). qRT-PCR was performed with SYBR Green master mix on a QS3 (Applied Bioscience, USA) using primers specific for CSN4, DDB1, and beta-actin as an internal control. Relative expression was calculated by the ΔΔCt method and presented relative quantity (RQ) values. Primer sequences are listed in Supplementary Table 1.

Clonogenic assay

HeLa cells were seeded at 200–500 cells per well in 6-well plates and treated with 2 mM hydroxyurea for 2 h. After treatment, the medium was replaced with a fresh medium, and cells were incubated for 10 days. Colonies containing more than 50 cells were counted. Colony formation efficiency was calculated as the percentage of colonies formed relative to the number of initially plated cells. Each experiment was performed in triplicate and independently repeated twice. Data are presented as mean ± standard deviation.

Protein purification, pull-down assay, and Immunoprecipitation

N-terminally Strep-tagged DDB1 WT and mutant proteins were expressed in Sf9 insect cells using recombinant baculovirus and purified using Strep-Tactin affinity chromatography. Briefly, Sf9 cells were infected with the appropriate baculovirus constructs and harvested 48–72 h post-infection. Cell pellets were lysed in ice-cold lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% NP-40, and protease inhibitor cocktail), and cleared lysates were incubated with Strep-Tactin Sepharose beads for 2 h at 4 °C. Beads were washed extensively with wash buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl) to remove non-specifically bound proteins.

For the pull-down assay, equal amounts of bead-bound Strep-tagged WT or mutant DDB1 proteins were incubated with lysates from HEK293T cells transiently transfected with HA-CSN4. Binding reactions were carried out for 2 h at 4 °C with gentle rotation. After incubation, beads were washed three times with wash buffer, and bound proteins were eluted by boiling in SDS sample buffer. Eluted samples were subjected to SDS-PAGE and immunoblotting with anti-HA and anti-DDB1 antibodies to assess the interaction between DDB1 and CSN4.

To immunoprecipitate proteins, cells were lysed in lysis buffer (50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, and 1 mM EGTA) supplemented with protease and phosphatase inhibitors. Lysates were incubated with anti-FLAG M2 affinity gel (Sigma-Aldrich, USA) or anti-HA magnetic beads for at least 3 h at 4 °C. The precipitates were washed four times with lysis buffer and finally eluted by boiling in SDS sample buffer for immunoblot analysis.

Proximity ligation assay (PLA)

PLA was performed using the Duolink^® In Situ Red Starter Kit (Sigma-Aldrich, USA). Cells grown on coverslips were fixed, permeabilized, and blocked prior to incubation with primary antibodies against MDC1 (Abcam, USA), p-H2AX (Millipore, USA), CSN4 (Proteintech, USA), and DDB1 (Santa Cruz, USA). The PLA’s probing, ligation, and amplification steps followed a previously described protocol [28]. Red fluorescent signals were visualized by fluorescence microscopy and quantified using ImageJ.

Immunoblotting and immunocytochemistry

The following primary antibodies were used for immunoblotting: anti-DDB1, anti-FBXW7, anti-RFP, and anti-Ambra1 (Proteintech, USA); anti–β-actin, anti-Skp2, anti-Skp1, anti–phospho-Chk1, and anti-FLAG (Cell Signaling Technology, USA); anti-HA (Roche, Switzerland); anti–α-tubulin, anti-Chk1, anti-CSN6, and anti-ubiquitin (Santa Cruz Biotechnology, USA); anti-CSN4, anti-Cul4a, and anti-CSN5 (Bethyl Laboratories, USA); anti-DDB2 (Bosterbio, USA); anti-CSN6 (Santa Cruz Biotechnology, USA); and anti-SOCS2 (Abcam, UK). HRP-conjugated secondary antibodies (Bio-Rad, USA) were used, and signals were detected using enhanced chemiluminescence (ECL) reagents. Signals were visualized either by film exposure or using a chemiluminescence imaging system (e.g., Bio-Rad ChemiDoc MP) and visualized on LAS 4000 (Fujifilm, Japan).

For immunofluorescence staining, cells were fixed with 4% paraformaldehyde, permeabilized with 0.25% Triton X-100, and blocked with 1% BSA in PBS. Primary antibodies used were anti-CSN4 (Proteintech), anti-DDB1 (Santa Cruz Biotechnology, USA), and anti-LC3 (Cell Signaling Technology, USA). After washing three times with PBS with 0.03% tween20, cells were incubated for 1 h at room temperature with Cy3- or FITC-conjugated secondary antibodies (Jackson ImmunoResearch, USA). Cells were mounted with Vectashield containing 4′,6-diamidino-2-phenylindole (Vector Laboratories, USA). Images were obtained using a Zeiss LSM 710 confocal microscope (Carl Zeiss, Germany). Fluorescence quantification for immunocytochemistry (ICC) was performed using ImageJ. For each condition, five images were collected from three independent experimental sets (5 images per experimental set). DDB1 and LC3 fluorescence intensities were measured using identical exposure and threshold settings across all images. The resulting values were averaged, and the data are presented as mean ± standard deviation. In the quantification graphs, green bars represent DDB1 fluorescence intensity and red bars represent LC3 fluorescence intensity.

In vivo ubiquitination assay

Cells were transfected as described above. MG132 (10 µM) was added 5 h before cell lysis. Cells were lysed in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with a protease inhibitor cocktail, N-ethylmaleimide (20 µM), and iodoacetamide (5 mM final concentration), followed by incubation on ice for 30 min with occasional gentle mixing. Lysates were clarified by centrifugation at 16,000 × g for 15 min at 4 °C. Supernatants were incubated with anti-HA magnetic beads (Thermo Fisher Scientific, USA) at 4 °C for 3 h with rotation. Beads were washed three times with cold lysis buffer, and immunoprecipitated proteins were eluted with 2× SDS sample buffer, followed by SDS-PAGE and Western blotting.

Comet assay

Comet assay was performed as described in the manufacturer’s instructions (Cell Biolabs, San Diego, CA, USA). Briefly, endogenous DDB1 was depleted with siRNA and rescued with WT DDB1 or DDB1 K1131R. We further treated puromycin to exclude the untransfected cells and subjected them to DNA damage induction. The images were obtained using a fluorescence microscope and analyzed by calculating the length and fraction of DNA within the “tail” of a comet by multiplying the tail length by the percentage of DNA in the tail.

Cell cycle analysis

Cells were harvested using trypsin-EDTA and resuspended in PBS containing 0.1% BSA. Aliquots (500 µl at 5 × 10⁶ cells/ml) were transferred to conical tubes, and 5 ml of cold 70% ethanol was added dropwise with gentle vortexing. Cells were fixed overnight at 4 °C, washed twice with PBS, and resuspended in 1 ml of propidium iodide solution (20 µg/ml; Molecular Probes, USA) containing RNase A (0.5 µg/ml; Molecular Probes, USA). Samples were incubated for 4 h at 4 °C and analyzed for DNA content using a FACSymphony A5 SE flow cytometer (BD Biosciences, USA).

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contributions

Seung Ho Choi: Conceptualization, Methodology, Visualization, Writing – original draft, Validation, Investigation, Writing – review & editing, Funding acquisition. Kyoungjoo Cho: Conceptualization,

Methodology, Visualization, Validation, Investigation, Writing – review & editing, Resources. Eun Seon Kim: Validation. Hae Yong Yoo: Supervision.

Funding

This work was supported by the Basic Science Research Program through National Research.

Foundation of Korea (NRF) of Korea funded by the Ministry of Education; NRF-2021R1I1A1A01057608.

Data availability

All data and materials that support the findings of this study are available.

Declarations

Ethics approval

Not applicable.

Consent to participate

Not applicable.

Consent to publish

Not applicable.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.