Repairing damaged DNA and removing all physical connections between sister chromosomes prior to anaphase is important for proper chromosome separation to ensure genomic stability. DNA interstrand cross-links (ICLs) induce the stalling of replication forks and DNA breaks (Bredberg et al., 1982; Dardalhon and Averbeck, 1995; De Silva et al., 2000; McHugh et al., 2000), and ICLs repair proceeds through the formation of double-strand breaks (DSBs) (Bessho, 2003; De Silva et al., 2000; McHugh et al., 2000). Immunodetection of phosphorylated histone H2AX (γ-H2AX) at Serine 139 is commonly used to visualize ICL processing and DSB induction in the cell nucleus (Niedernhofer et al., 2004; Redon et al., 2002; Rothfuss and Grompe, 2004; Thiriet and Hayes, 2005).

Homologous recombination (HR) is essential for the correct repair of ICLs and DSBs. HR causes the formation of four-way intermediate DNA structures, called Holliday junctions (HJs), that must be resolved to allow for chromosomal separation (Holliday, 2007; Li and Heyer, 2008). HJ resolvases are needed to resolve recombination intermediates to ensure proper chromosome segregation at mitosis. The BTR complex (BLM-TopoIIIa-RMI1/2), SLX–MUS complex (SLX1-SLX4 and MUS81-EME1), and GEN1 are three pathways that facilitate the dissolution of HJs in human cells (Sarbajna et al., 2014). The nucleolytic resolution of the main HJs is regulated by the SLX4 complex (SLX4-SLX1/MUS81-EME1 and XPF-ERCC1) (Chen et al., 2001; Ciccia et al., 2003; Fekairi et al., 2009) and GEN1 (Ip et al., 2008), which determine whether crossover or non-crossover occurs. Yen1 is the yeast ortholog of human GEN1. The phosphorylation of S679 in Yen1 overlaps the nuclear localization signal, which inhibits nuclear import (Blanco et al., 2014; Eissler et al., 2014; Kosugi et al., 2009). Thereafter, Yen1 is activated by Cdc14 phosphatase, resulting in nuclear entry and enzymatic activation. Resolvase activity resolves the remaining HJs and ensures accurate DNA isolation (Blanco et al., 2014; Eissler et al., 2014). Although Mus81 and Yen1 in yeasts are sequentially activated by phosphorylation events, their activities do not overlap (Matos et al., 2011). In humans, GEN1 also reportedly reaches the chromatin, following nuclear envelope breakdown (Hustedt and Durocher, 2016) at the last stage of mitosis. However, unlike the mechanism that relies on Yen1 phosphorylation in yeast, the translocation of GEN1 is exclusively regulated by nuclear export signals (Chan and West, 2014). GEN1-deficient cells show increased chromosome missegregation, micronucleus formation, and genome instability, suggesting that GEN1 is essential for genome maintenance and proper chromosome segregation (Sarbajna et al., 2014; Wechsler et al., 2011).

Proper chromatin fiber condensation into the highly ordered structure of the mitotic chromosomes is crucial for achieving the equal segregation of sister chromatids until final physical cell division. During the course of the cell cycle, condensin I and II regulate mitotic chromosome compaction and resolution in different ways (Ono et al., 2003). The majority of higher eukaryotes, including humans, have two condensins, condensin I and condensin II. These two condensins share the same SMC2 and SMC4, and also contain non-SMC components with other kleisin subunit and HEAT repeat subunit pairs. Condensin I consists of CAP-H, -D2, and -G and condensin II consists of CAP-H2, -D3, and -G2. NCAPH is one of the non-SMC subunits of condensin I (Hirano, 2012; 2016; Thadani et al., 2012), which acts on chromosomes during mitosis, mostly in the cytoplasm during interphase, and occasionally during nuclear localization (Ono et al., 2003). Further, condensin I plays numerous other roles such as in DNA repair and in the epigenetic regulation of transcription (Chen et al., 2004; Dej et al., 2004). In particular, condensin I was reported to repair DNA single-strand breaks by binding to the PARP-1-XRCC-1 complex (Heale et al., 2006). During meiosis in yeasts, condensin is required for resolving recombination-dependent chromosome linkages (Yu and Koshland, 2003). Similarly, condensin is not associated with the export of ribosomal DNA breaks in the nucleus but rather facilitates the timely repair of meiotic DSBs (Li et al., 2014). Additionally, previous studies on NCAPH have shown that NCAPH-knockdown cells are arrested in the S and G2/M-phase; chromosomal aberrations and DNA damage are caused by the Chk1/Chk2 signaling pathways and play an essential role in the complete condensation of the chromosomes (Kim et al., 2019).

Here, we found that NCAPH regulates recovery from DNA damage-induced cell cycle arrest in both pancreatic cancer MIA PaCa-2 cells and human epithelial kidney HEK293 cells and restricts the formation of ultra-fine DNA bridge (UFB)-associated entanglement structures in sister chromosomes, thereby ensuring accurate chromosome separation. DNA-ICL induction, G2/M-phase arrest, UFB formation, chromosome segregation errors, and cytokinesis failure were observed in NCAPH-depleted cells. Interestingly, NCAPH interacted with GEN1, and the NCAPH–GEN1 interaction was enhanced following the induction of DNA-ICLs. Collectively, it was observed that NCAPH is closely related to GEN1 and is critical for the localization of GEN1 at sites of DNA damage. Therefore, NCAPH plays an important role in the recruitment and stabilization of GEN1 in chromatin to repair DNA damage, such as DNA-ICL, and maintain chromosomal stability.

Cell culture, cell cycle synchronization

MIA PaCa-2 (CRL-1420) human pancreatic cancer cell lines and HEK293 (CRL-1573) human kidney cell lines were obtained from the American Type Culture Collection (ATCC, USA). Cells were grown in high-glucose Dulbecco’s modified Eagle’s (DMEM) medium. Cell culture mediums were used including 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% fetal bovine serum (Gibco Life Technologies, USA). To synchronize cells at late G1-phase (near the G1-S boundary), a double thymidine arrest/release experiments was performed (Whitfield et al., 2000). For a double thymidine block experiments, cells were synchronized with 2 mM thymidine (1‐[2‐deoxy‐β‐d‐ribofuranosyl]‐5‐methyluracil) (Sigma, USA). After the second thymidine treatment, fresh medium was replaced and samples were taken at each time point (Chen and Deng, 2018).

Small interfering RNA (siRNA) or plasmid DNA transfection

Cells were transfected with NCAPH siRNA using the Lipofectamine RNAiMAX transfection reagent (Invitrogen, USA) according to manufacturer’s instructions. NCAPH siRNA was synthesized as 5’-AAU AGU UCU GUG UAA GUG C-3’ and 5’-GCA CUU ACA CAG AAC UAU U-3’. NCAPH siRNA (Georges et al., 2008) and control siRNA were provided by Cosmo Bio (Cosmo Genetech, Korea). For DNA transfection, the cells were transfected with Flag vector, Flag-tagged full length NCAPH, three domains of NCAPH, GST vector and GST-tagged full length GEN1 using the Lipofectamine 3000 transfection reagent (Invitrogen) according to manufacturer’s instructions.

Fluorescence-assisted cell sorting (FACS) analysis

For measurements of cell cycle and DNA content, cells were collected, and washed in 1× phosphate-buffered saline (PBS) and fixed with 70% ethanol at 4°C overnight before being stained with PI (propidium iodide; Sigma) 50 µg/ml and 100 U RNase (ribonuclease A from bovine pancreas; Sigma). The analysis rate was 300-500 cells/s. Clones were assessed in three independent runs. Analysis was performed using FACS Calibur (BD Biosciences, USA) according to standard protocols.

Cytoplasmic, soluble nuclear fraction, and chromatin fractions

Cytoplasmic, soluble nuclear, and chromatin proteins were divided with the NE-PER Nuclear and Cytoplasmic Extraction Reagents (#78835; Thermo Fisher Scientific, USA). Briefly, cell pellets were lysed with CERI buffer, and forceful vortex was applied for 15 s and incubated on ice for 10 min. Next, CER II buffer was added to the samples was shaken strenuously for 5 s and incubated for 1 min on ice (cytoplasmic fraction). After centrifugation at 16,000 × g for 5 min, the pellet was re-suspended with nuclear extraction reagent and incubated on ice for 40 min with intermittent vortex. Nuclear fractions were collected after centrifugation at 16,000 × g for 10 min (soluble nuclear fraction). The pellet was washed two times with 1× PBS. Then, the cell pellet was re-suspended in radio immunoprecipitation assay (RIPA) lysis buffer (Beyotime Biotech, China) and incubated on ice for 40 min and then sonicated for 30 s, it was turned on for 2 s and then turned off for 3 s repeatedly (chromatin fraction). All buffers except CERII buffer were supplemented with PMSF, protease inhibitor, and phosphatase inhibitor cocktails. Fractions were supplemented with 4× (for western blot) or 2× (for immunoprecipitation) laemmli sample buffer (contains 10% or 5% 2-mercaptoethanol; Bio-Rad, USA) and boiled for 5 min at 95°C, and were examined using western blotting analysis or immunoprecipitation.

Western blot and antibodies

Whole cells were washed once with cold 1× PBS and then collected. Cells were homogenized with 1× RIPA lysis buffer, and protein concentration was determined using BCA Protein Assay Kit (Pierce, USA). Protein extracts were re-suspended in 4× sample buffer containing 10% of 2-mercaptoethanol, boiled for 5 min, and subjected to 8%-15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electrotransferred to a Trans-blot nitrocellulose membrane (Whatman International, UK). Transferred membranes were blocked in 5% skim milk in TBST buffer (10 mM Tris [pH 8.0], 150 mM NaCl, 0.05% Tween 20) and incubated with primary antibodies at 4°C overnight. The primary antibodies were purchased from commercial sours: anti-NCAPH (Novus Biological, USA), anti-NCAPG (Bethyl Laboratories, USA), anti-NCAPH2 (Bethyl Laboratories), anti-NCAPD3 (Proteintech, China), anti-GEN1 (Novus Biological), anti-MUS81 (Novus Biological), anti-Cyclin A2 (Cell Signaling Technology, USA), anti-Histone H3 (Abcam, USA), anti-phosphorylated Histone H3 at pS10 (pS10 histone H3) (Abcam), anti-phosphorylated H2AX at S139 (p-H2AX) (Millipore, USA), anti-GAPDH (Thermo Fisher Scientific), anti-LMNA (Bethyl Laboratories), anti-alpaha Tubulin (Abcam), and anti-β-actin (Sigma). Protein expressions analyzed chemiluminescent signals activated by SuperSignal West Pico Chemiluminescent Substrate (Pierce) reacted with horse radish peroxidase tagged secondary antibodies (Jackson ImmunoResearch, USA).

Immunoprecipitation

Flag-tagged full-length NCAPH and GEN1 proteins were expressed in HEK293 cells and purified with FLAG M2 monoclonal antibody affinity gel (Sigma) according to manufacturer’s instructions. GST-tagged full-length GEN1 protein was expressed in HEK293 cells and purified with Glutathione-Sepharose 4B (Thermo Fisher Scientific) according to manufacturer's instructions. HEK293 cells transfected with three truncation domain constructs Flag-tagged-NCAPH (N-NCAPH [1–250], M-NCAPH [251–500], and C-NCAPH [501-741]) or control vector and then purified with Glutathione-Sepharose 4B or FLAG M2 monoclonal antibody affinity gel (Sigma). The primary antibodies were acquired from commercial sours: anti-Flag (Sigma), anti-GST (Cell Signaling Technology), anti-NCAPH (Novus Biological), anti-NCAPD2 (Bethyl Laboratories), anti-GAPDH (Thermo Fisher Scientific), anti-LMNA (Bethyl Laboratories). The ATM-kinase inhibitor (KU-55933) was acquired from Sigma.

Immunofluorescence

Cells were grown on Thermo Fisher Scientific Nunc Lab-Tek II Chamber Slides & trade, fixed with 4% paraformaldehyde for 15 min or PTEM buffer containing 20 mM PIPES (pH 6.8), 0.2%. Triton X-100, 10 mM EGTA, and 4% paraformaldehyde for 10 min, and permeabilized with 0.5% Triton X-100 for 5 min. The fixed cells were incubated for 1 h with % bovine serum albumin and then incubated overnight at 4°C with primary antibody, then the next day add the fluorophore-conjugated secondary antibody for 2 h. Samples were mounted in Vectashield Mounting Medium with DAPI (Vector laboratories, USA), and the results were observed under a confocal microscope (Zeiss LSM 710 or LSM 800, software ZEN; Zeiss, Germany). The primary antibodies were acquired from commercial sours: anti-NCAPH (Novus Biological), anti-GEN1 (Novus Biological), anti-p-H2AX (Millipore), anti-CREST (ImmunoVision, USA), anti-GFP (Invitrogen), and anti-alpha Tubulin (Abcam).

Lentivirus expression of NCAPH and site-directed mutagenesis

The NCAPH expression vector was constructed as a pLenti6/V5-DEST lentiviral vector and was purchased through Addgene (#31214). To produce siRNA-resistant mutant NCAPH (siR-NCAPH), the mutants (siR-NCAPH forward primer: 5’-CAA AGA AGA AGC ACTT GCA TAG GAC TAT TGA GCA GAA C-3’, siR-NCAPH reverse primer: 5’-GTT CTG CTC AAT AGT CCT ATG CAA GTG CTT CTT CTT TG-3’) were generated using QuikChange II Site-Directed Mutagenesis Kit (Stratagene, USA). To produce lentivirus from a NCAPH lentiviral vector, 293FT cells cultured in a 100 mm plate were co-transfected with 3 µg of NCAPH ranging from 4 µg using 3 µg of the packaging vectors (pRsv-REV, pMDLg/pRRE and pMD2.G). Six hours after transfection, cells were changed with MEM medium supplemented with 10% fetal bovine serum but not antibiotics. Another 72 h later, the medium was centrifuged at 500 × g for 10 min and decanted into another tube. One volume of Lenti-X Concentrator solution (Clontech, USA) with tree volumes of the collected-medium were mixed and incubated at 4°C for 40 min. Next, mixture centrifuged 1,500 × g for 45 min and then removes the supernatant. The pellet was re-suspended in DMEM with 10% fetal bovine serum, and virus was aliquoted and stored at –80°C.

Statistical analysis

All data will be represented central tendency obtained in form two or three independent replicate experiments. The data of the experiments were statistically analyzed using GraphPad Prism 5 (GraphPad Software, USA) and presented as mean ± SEM. Statistical analysis was performed using unpaired t-tests, one-way or two-way ANOVA followed by Bonferroni’s multiple comparison analysis tests.

Data availability statement

The authors declare that all relevant data supporting the findings of the study are available within the manuscript, figure, and supplementary information, and are available from the respective authors upon reasonable request.

NCAPH response to DNA-ICLs-induced DSB formation

First, we determined whether NCAPH localization/enrichment in the cytoplasm, nuclear and chromatin fractions is altered during cell cycle progression under DNA-ICL induction with mitomycin C (MMC) and cisplatin, which are DNA damage-inducing agents (Deans and West, 2011). Cell cycle progression proceeded with a double thymidine block. As the second thymidine block treatment was performed, DNA damage was induced by MMC, and time-dependent harvesting was performed for flow cytometry and western blotting analyses (Fig. 1A). Dimethyl sulfoxide (DMSO)-treated cells stopped at the G1/S-phase boundary 0 h, the S-phase population increased at 4 h, and the G/2M-phase population increased at 8 h after release from the second thymidine treatment (Supplementary Figs. S1A and S1B). However, the S-phase population of MMC-treated cells was significantly increased after release from the second thymidine treatment (Supplementary Figs. S1A and S1B). To confirm the occurrence of DSBs during DNA-ICL repair, the formation of γH2AX (p-H2AX), the phosphorylated Ser-139 residue of H2AX, was monitored. In the absence of MMC, NCAPH was weakly expressed in chromatin+ in the late G1-phase at 0 h after release from the second thymidine treatment. However, as the cell cycle progressed into the G2/M-phase, NCAPH localization/enrichment in the chromatin+ increased (Fig. 1B, –MMC). Low expression levels of p-H2AX were detected throughout cell cycle progression. However, p-H2AX expression and NCAPH localization/enrichment in the chromatin+ were increased significantly by MMC-induced DNA-ICLs from 0 to 8 h after release from the second thymidine treatment (Fig. 1B, +MMC). Immunofluorescence specifically confirmed that NCAPH expression in MMC-induced DNA-ICLs occurred in parallel with p-H2AX accumulation (Fig. 1C). Intranuclear expression levels of NCAPH were also increased by MMC-induced DNA-ICLs (Fig. 1D). Thus, it was confirmed that NCAPH localization/enrichment in the chromatin fraction was increased by DSBs caused by DNA-ICLs.

The effects of cisplatin- and MMC-induced DNA-ICLs and NCAPH depletion on the cell cycle were analyzed based on flow cytometry and western blotting analyses. Cells were transfected with control and NCAPH siRNAs and exposed to cisplatin or MMC for 24 h. The flow cytometry results indicated that the G2/M-phase population increased in NCAPH-depleted cells, and the G2/M-phase population increased more in cisplatin-treated NCAPH-depleted cells (Figs. 1E and 1F). Upon MMC treatment, control cells increased in the S-phase population, whereas NCAPH-depleted cells increased in the G2/M-phase population (Figs. 1E and 1F). When cisplatin or MMC was used to induce DNA-ICLs in NCAPH-depleted cells, the expression levels of cyclin A2 and p-H2AX were increased compared with those in the control cells (Fig. 1G). These results suggest that NCAPH plays an essential role in the repair of cisplatin- or MMC-induced DNA-ICLs to ensure proper mitosis progression.

Abnormal phenomena caused by UFB induction in NCAPH-depleted cells

To explore the effect of unresolved recombinant intermediates caused by DNA-ICLs on mitosis, the cells were transfected with control and NCAPH siRNAs and exposed to cisplatin for 1 h. After 24 h, immunofluorescence was performed using anti-PLK1 interacting checkpoint helicase (PICH). The PICH, an SNF2-family DNA translocase, is localized in UFBs and is required for chromosome organization and proper chromosome segregation. PICH is essential for the progression of UFBs via the recruitment of UFB-associated proteins and the mediation of UFB resolution during mitosis (Chan et al., 2009; Chanboonyasitt and Chan, 2021; Wang et al., 2010; 2008). Following the induction of DNA-ICLs by cisplatin, approximately 16% of the NCAPH-depleted cells were observed to be in late anaphase‒telophase, whereas only about 4% of the control cells were observed in this phase (Fig. 2A, left graph). The frequency of UFBs in late anaphase-telophase NCAPH-depleted cells increased significantly to about 66% compared to only 6.6% in the control group (Fig. 2A, middle graph). The number of UFBs observed in NCAPH-depleted cells was approximately 3.3, whereas UFBs were observed only rarely in the control cells (Fig. 2A, right graph).

Immunofluorescence analyses were performed using DNA, CREST (a centromere marker), and α-tubulin to investigate defective abnormal chromosome separation. Following DNA-ICL induction by cisplatin, approximately 75% of the NCAPH-depleted cells in the late anaphase‒telophase stage showed abnormal chromosome separation, such as chromosome bridges and chromosome lagging (Fig. 2B). In contrast, only 6.6% of control cells in this stage showed abnormal chromosome separation. We further examined the effect of NCAPH depletion and cisplatin on cytokinesis during the cell division process. In the absence of cisplatin treatment, 31.9% of the NCAPH-depleted cells showed cytokinesis failure, as opposed to only 5% of the control group. In the presence of cisplatin, 58.8% cytokinesis failure was observed in NCAPH-depleted cells compared with only 10.1% observed in the control group. Thus, there was a noteworthy difference between the extent of cytokinesis failure among NCAPH-depleted cells depending on the presence or absence of cisplatin (Fig. 2C). These results indicate that defects in NCAPH exacerbate chromosome segregation errors and cytokinesis failure owing to sister-chromatid intertwinement, a distinct form of UFB produced by cisplatin-induced DNA-ICLs.

NCAPH interacts with GEN1 via DNA damage response

HR repairs DNA damage prompted by cisplatin-induced DNA-ICLs. To understand the mechanism underlying NCAPH regulation during DNA damage repair progression, we explored its relationship with MUS81 and GEN1 resolving enzymes (Bittmann et al., 2020; García-Luis et al., 2014; Grigaitis et al., 2020; Machín, 2020; Wyatt et al., 2013). To gain a better understanding of the relationship between NCAPH and HJ resolvases, we tested the hypothesis that NCAPH acts as a binding partner of GEN1 by investigating whether NCAPH interacts with GEN1. Binding among the Flag-tagged full-length NCAPH protein, another condensin I component, non-SMC condensin I complex subunit D2 (NCAPD2), endogenous GEN1, and MUS81 was analyzed via immunoprecipitation analysis. Exogenously expressed Flag-tagged full-length NCAPH interacted with both endogenous GEN1 and NCAPD2 but not with MUS81 (Fig. 3A). Immunoprecipitation confirmed the interaction between the Flag-tagged full-length GEN1 protein and endogenous NCAPH (Fig. 3B). Additionally, we analyzed the interaction between NCAPH and GEN1 by producing Flag-tagged full-length NCAPH (Flag-NCAPH) and glutathione S-transferase (GST)-tagged full-length GEN1 (GST-GEN1) and performing immunoprecipitation experiments. These results showed that exogenously expressed Flag-tagged full-length NCAPH interacted with GST-tagged full-length GEN1 (Fig. 3C). Furthermore, a double thymidine block followed by time-dependent harvesting was performed to analyze the expression patterns of NCAPH and the HJ resolvases with western blotting (Supplementary Fig. S2A). The expression levels of NCAPH and GEN1 in the chromatin+ increased during the G2/M-phase 8 h after release from the second thymidine treatment. In contrast, MUS81 expression in the chromatin+ increased only 4 h after release from the second thymidine treatment (Supplementary Fig. S2B). Immunofluorescence indicated that overexpressed GFP-tagged NCAPH localized with GEN1, confirming their presence in the same location in the nucleus (Supplementary Fig. S2C). In addition, immunofluorescence confirmed that GEN1 was co-expressed with p-H2AX in cells displaying MMC-induced DNA damage in a manner comparable to that of NCAPH, thereby demonstrating that GEN1 expression in the nucleus was increased by MMC-induced DNA damage (Supplementary Fig. S2D).

To analyze the signaling pathway involved in the interaction between NCAPH and GEN1, immunoprecipitation experiments of Flag-NCAPH and GST-GEN1 were performed using Ataxia telangiectasia mutated (ATM)-kinase inhibitor treatment, and the resulting fractions were divided into those with and without chromatin. The interaction of NCAPH and GEN1 in chromatin partially decreased under ATM-kinase inhibitor treatment (Fig. 3D). Immunoprecipitation assays using Flag-tagged-NCAPH (N-NCAPH [1-250], M-NCAPH [251-500], and C-NCAPH [501-741]) showed that the N-NCAPH bound strongly to GEN1 and that NCAPD2 bound strongly to the N-NCAPH (Figs. 3E and 3F). These results suggest that NCAPH and GEN1 form a close relationship, that N-NCAPH is important for binding GEN1 with other condensin components and that the ATM signaling pathway is partially involved in the NCAPH–GEN1 interaction.

Finally, we performed flow cytometry and immunoprecipitation experiments to test whether the interaction between NCAPH and GEN1 was altered by MMC-induced DNA damage. Cell cycle analysis revealed that DMSO-treated cells stopped at the G1/S-phase boundary at 0 h, the S-phase population increased after 4 h, and the G2/M-phase population increased after 8 h after release from the second thymidine treatment (Supplementary Figs. S1C and S1D). However, in MMC-treated cells the S-phase population increased 0-8 h after release from the second thymidine treatment (Supplementary Figs. S1C and S1D). In addition, the binding between NCAPH and GEN1 in the chromatin+ was significantly increased owing to MMC-induced DNA damage (Fig. 3G). Consequently, NCAPH interacts with GEN1 via a DNA damage response.

NCAPH is essential for GEN1 recruitment and stabilization in chromatin

To determine the effect of NCAPH on the chromatin localization of GEN1, NCAPH-depleted cells and control cells were double-blocked with thymidine (Fig. 4A). Preferentially, the cell cycle was analyzed 0-12 h after release from the second thymidine treatment (Fig. 4B). The S-phase population and cyclin A2 expression were more increased 4 h after release from the second thymidine treatment in NCAPH-depleted cells than in control cells. After 8 h, the G2/M-phase population and cyclin A2 and phospho-histone H3 (pS10 histone H3) expression were more increased in NCAPH-depleted cells than in control cells (Fig. 4C). Next, western blotting was performed after separation into the cytoplasmic, nuclear, and chromatin components. Although the expression of NCAPH and other condensin I components (NCAPG) in the cytoplasm, nucleus, and chromatin fractions were affected by NCAPH depletion, the components of condensin II (NCAPH2) were not (Fig. 4D). GEN1 expression in chromatin fraction was higher at 8 h than at 0 h after release from the second thymidine treatment (late G1-phase). However, GEN1 expression in the chromatin fraction was significantly reduced 8 h after release from the second thymidine treatment (G2/M-phase) owing to NCAPH depletion (Fig. 4D, Chr). Furthermore, GEN1 expression did not change significantly in the cytoplasm or nucleus (Fig. 4D, Cyt and Nuc). Following the induction of DNA-ICLs via MMC in NCAPH-depleted cells, p-H2AX expression was markedly increased compared with that in control cells. Although there was no significant alteration in GEN1 expression, colocalization between GEN1 and p-H2AX was significantly inhibited (Fig. 4E).

To clarify the involvement of NCAPH in the recruitment and stabilization of GEN1 in chromatin fraction, we conducted reconstitution experiments after NCAPH siRNA transfection. First, we produced lentiviral vector-expressing siRNA-resistant NCAPH (Lenti-siR-NCAPH) and performed reconstitution experiments (Fig. 5A). Next, the cells were separated into cytoplasm, nucleus, and chromatin fractions, and changes in the recruitment and stabilization of GEN1 to DNA damage sites were predicted by reconstruction experiments using western blotting analyses. The decrease in chromatin-specific GEN1 localization in NCAPH-depleted cells was restored by Lenti-siR-NCAPH treatment (Fig. 5B). In addition, the localization of the relative fluorescence signal of GEN1, compared with that of p-H2AX, was decreased in NCAPH-depleted cells, whereas it was restored by Lenti-siR-NCAPH treatment (Fig. 5C). To determine how NCAPH regulates the localization of GEN1 to facilitate DNA damage repair, the degree of colocalization between the fluorophores of GEN1 and p-H2AX was measured using Pearson’s correlation coefficient and Mander’s overlap coefficient analyses. Although the immunofluorescence experiments showed that the colocalization of GEN1 with p-H2AX significantly decreased in NCAPH-depleted cells, both measurement methods confirmed that this decrease was recovered by Lenti-siR-NCAPH treatment (Fig. 5D). Thus, NCAPH acts as a regulator to recruit GEN1 to the chromatin and stabilize it to repair damage induced by DNA-ICLs.

This study adopted two approaches to better understand the function of condensin I, which maintains normal cell division and genomic stability by repairing DNA damage induced by DNA-ICLs in humans. One approach focused on analyzing expression patterns during cell cycle progression, with respect to the function of NCAPH in chromosome segregation in cisplatin- or MMC-induced DNA-ICLs. The other approach identified and established relationships between an interaction partner and NCAPH. Our results indicate that NCAPH responds rapidly to the damage caused by DNA-ICLs, limits UFB formation, and plays an essential role in proper chromosome segregation and cytokinesis during sister-chromatid separation. Furthermore, we confirm that the interaction between GEN1 and NCAPH in the chromatin was promoted by the induction of DNA-ICLs. Our results show that NCAPH provides GEN1 with access to the chromatin and helps to maintain its position in the chromatin. Thus, NCAPH acts as an essential regulator that recruits and stabilizes GEN1 in chromatin, wherein the NCAPH–GEN1 interaction repairs damage caused by DNA-ICLs. These results provide further insight into how condensin I links the processes that repair DNA damage.

The two metazoan condensin complexes, condensin I and II, play individual roles in chromosome compaction. Condensin I regulates the lateral compaction of mitotic chromosomes, whereas condensin II modulates chromosome shortening (Green et al., 2012; Shintomi and Hirano, 2011). Ultra-high resolved imaging and experimental Hi-C (Lee and Seo, 2021) data on mitotic chromosomes show that condensin II builds long chromatin loops, whereas condensin I forms smaller DNA compaction within the larger condensin II-mediated DNA compaction, suggesting its involvement in chromosome separation (Gibcus et al., 2018; Walther et al., 2018). Condensin I causes shortening of mitotic chromosomes even in the absence of condensin II, whereas condensin II is essential for the decatenation and formation of a straight chromosome axis (Elbatsh et al., 2019). Thus, condensin mediates chromosome condensation and proper chromosome segregation (Lamothe et al., 2020; Thadani et al., 2012), and resolves cellular DNA damage reactions (Lehmann, 2005). Data on the role of condensin in repairing damaged DNA show that condensin does not play an important role in DSB repair but rather interacts strongly with the PARP1-XRCC1 complex in single-strand break repair (Heale et al., 2006). However, condensin II does not affect G2/M checkpoint activation, and condensin II depletion only affects HR recovery in IR-induced DSBs. Condensin II participates in the HR repair of DSBs to maintain genomic integrity (Wood et al., 2008).

We hypothesized that condensin I maintains the chromatids in a stable state until cell division ends by recovering the DNA damage caused by DNA-ICLs. ICL is an adduct that forms in DNA, which inhibits the processes involved in DNA unwinding. The ICL removal process can be simplified into three major steps: (i) DNA repair-associated proteins are mobilized in response to ICL-induced cell cycle arrest; (ii) the ICL is excised from the DNA by nucleotide excision repair and Fanconi anemia-related proteins; and (iii) DSBs are restored by the induction of DNA-ICLs to HJs (Lopez-Martinez et al., 2016). To repair ICLs formed by cisplatin or MMC, HR must proceed, and the covalently linked intermediates between the sister chromosomes must be removed before anaphase proceeds for stable chromosomal separation (Sarbajna et al., 2014). We observed that MMC-induced DNA-ICLs significantly enhanced NCAPH expression at the site of DNA damage, as shown by increased expression of DNA damage markers. G2/M-phase arrest was increased and UFBs were formed in NCAPH-depleted cells with cisplatin-induced DNA-ICLs, leading to defective chromosome segregation and cytokinesis failure. NCAPH reacted sensitively to DNA lesions caused by DNA-ICLs, and ensured normal cell division by completely removing the remaining DNA strands to repair DNA damage until the cells separated into two daughter cells. This indicates the presence of an NCAPH-interacting protein whose expression strongly correlated with that of condensin I, as well as the major HR mechanisms underlying DNA-ICL repair.

In ICL repair, the activities of the XPF-ERCC1 complex and other structure-selective endonucleases (SSEs), termed HJ resolvases, act collaboratively in the processes of translesion DNA synthesis and HR-mediated DNA repair (Lopez-Martinez et al., 2016; Pichierri et al., 2002). These SSEs, including MUS81–EME1 (Chen et al., 2001), SLX1–SLX4 (Fekairi et al., 2009), and GEN1 (Ip et al., 2008), describe two distinct HJ resolution pathways in human cells. SLX1, MUS81–EME1, and Fanconi-associated nuclease 1 endonucleases are known to be involved in repairing ICLs; however, the underlying mechanisms of regulation remain unclear (Zhang and Walter, 2014). To our knowledge, no studies have explored the mechanisms by which condensin I repairs the damage caused by DNA-ICLs and its relationship to HJ resolvases. We found that NCAPH regulates DNA-ICL damage repair via cooperation with HJ resolvases. HJ resolvases are required to prevent the intertwinement of sister chromosomes, thus ensuring correct chromosome condensation when sister chromatids separate (Wechsler et al., 2011). GEN1 (Yen1) can resolve HJs before chromosome separation, the last point of mitosis (Dehé and Gaillard, 2017). In the absence of HJ resolvases (GEN1–/– knockout cells deficient of MUS81), anaphase bridges are formed that persist until the end of mitosis, and these HRs, known as UFBs, delay the cell cycle and cause large-scale cell death (Chan et al., 2018). Our results of the effects of NCAPH depletion are consistent with this process, and therefore, an association between NCAPH and HJ resolvases may be expected. We demonstrated that NCAPH physically interacts with GEN1 but not with MUS81. The NCAPH–GEN1 interaction in the chromatin was significantly increased by induced DNA-ICLs, suggesting that this interaction is essential for repairing DNA-ICLs. The absence of NCAPH restricted the localization of GEN1 in chromatin during the G2/M-phase, indicating that NCAPH is a regulator of the recruitment and stabilization of GEN1 at DNA damage sites. NCAPH may play an essential role in repairing DNA damage caused by ICLs by interacting with GEN1. In the absence of NCAPH, the expression of other components of condensin I in the cytoplasm, nucleus, and chromatin fractions decreased, but this did not significantly affect the expression of other components of condensin II. Collectively, these data demonstrate a close correlation between NCAPH and GEN1, and that the regulation of the recruitment and stabilization of GEN1 to the DNA damage site by NCAPH may be the mechanism through which condensin I rapidly detects and repairs the damage caused by DNA-ICLs. Although this study is limited, as the results were validated only in cell culture models, it identifies a binding protein of condensin I and suggests a novel DNA-ICL repair mechanism.

Our findings show that the recruitment of NCAPH and GEN1 to DNA damage lesions is essential for DNA damage repair and normal cell division. In the absence of NCAPH, GEN1 is not stably maintained in the chromatin to repair ICLs, leading to the formation of chromosome bridges, lagging chromosomes, and late cytokinesis failures. Significantly, the interaction between NCAPH and GEN1 repairs DNA-ICLs. When DNA damage is detected, NCAPH and GEN1 react rapidly to increase their interaction at the site of DNA damage. In particular, NCAPH is involved in the recruitment and stabilization of GEN1 to chromatin in the G2/M-phase, thereby completely removing the intertwined DNA strands to facilitate efficient cell division; this reveals a mechanism of repairing DNA damage caused by DNA-ICLs. This study therefore provides new insight into the maintenance of chromosome stability via complete chromosome segregation facilitated by a DNA-ICL repair mechanism involving condensin I and HJ resolvase. These findings offer a new foundation for exploring the potential link between condensin and chromosome stability in future research.

This work was supported by grant No. NRF-2021R1A2C 200708811 from the National Research Foundation of Korea (NRF). We would like to thank Editage (www.editage.co.kr) for English language editing.

J.H.K. performed the experiments, carried out the data analysis, and wrote the manuscript. J.H.H. discussed the results and contributed to the manuscript revision. Y.Y. performed the reviewed experimental works. All of the authors discussed and approved the final manuscript.

The authors have no potential conflicts of interest to disclose.