Ubiquitination of the DNA-damage checkpoint kinase CHK1 by TRAF4 is required for CHK1 activation

HT29, SW620, HCT-8, FHC, HeLa, U2OS, A549, H1975, and 293T cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured at 37 °C in a humidified incubator with 5% CO₂ according to ATCC protocols. All cell lines were routinely checked for mycoplasma and tested cytogenetically to ensure their authentication. Chemical reagents, including Tris, NaCl, SDS, and dimethyl sulfoxide (DMSO), were purchased from Sigma-Aldrich (St. Louis, MO, USA). Oxaliplatin (cat. #S1224), 5-Fu (cat. #S1209), irinotecan (cat. #S2217), cisplatin (cat. #S1166), pemetrexed (cat. #S1135), gemcitabine (cat. #S1714), VE-821 (cat. #S8007), AZ20 (cat. #S7050), and prexasertib were obtained from Selleck Chemicals (Houston, TX, USA). Lipofectamine™ 2000 (cat. #11668019, Thermo Fisher Scientific) transfection reagent was used for transient transfection experiments following the manufacturer’s instruction. Ctrl siRNA (cat. #sc-93314), UbcH6 siRNA (cat. #sc-61744), BTG3 siRNA (cat. #sc-43644), and ATR siRNA (cat. #sc-29763) were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Three independent experiments were completed for each cell line, treatment, and time point, as indicated.

Whole-cell lysates were extracted with RIPA buffer (cat. #89900, Thermo Fisher Scientific) supplemented with phosphatase inhibitors (cat. #78428, Thermo Fisher Scientific). The lysates were sonicated and centrifuged at 12000×g for 15 min at 4 °C. The BCA Assay Reagent (cat. #23228, Thermo Fisher Scientific) was used to determine protein concentration. For co-immunoprecipitation (co-IP) assays, cells were lysed with IP Lysis Buffer (cat. #87787, Thermo Fisher Scientific). IB and co-IP were performed as previously described [16]. All antibodies for IB analysis were diluted in phosphate-buffered saline (PBS) buffer with 5% non-fat milk. Antibodies against Bax (cat. #5023; IB, 1:1000), Bik (cat. #4592; IB, 1:1000), Bim (cat. #2933; IB, 1:1000), Bid (cat. #2002; IB, 1:1000), Bak (cat. #12105; IB, 1:1000), survivin (cat. #2808; IB, 1:1000), Bcl-2 (cat. #4223; IB, 1:1000), Bcl-xL (cat. #2764; IB, 1:1000), Mcl-1 (cat. #5453; IB, 1:1000), γ-H2AX (cat. #9718; IB, 1:4000), α-tubulin (cat. #2144; IB, 1:10000), ubiquitin (cat. #3936; IB, 1:1000), cleaved-caspase 3 (cat. #9664; IB, 1:2000), cleaved-PARP (cat. #5625; IB, 1:2000), p-(Ser/Thr) ATM/ATR substrate (cat. #2851; IB, 1:1000), p-ATR (S428) (cat. # 2853; IB, 1:1000), p-ATR (Thr1989) (cat. #30632; IB, 1:1000), ATR (cat. # 13934; IB, 1:1000), p-CHK1 (S317) (cat. #12302; IB, 1:1000), p-CHK1 (S345) (cat. #2348; IB, 1:1000), CHK1 (cat. #2360; IB, 1:1000; IP, 1:200), p-CDC25C (Ser216) (cat. #4901; IB, 1:1000), CDC25C (cat. #4688; IB, 1:1000), GST tag (cat. #2624; IB, 1:5000; IP, 1:200), K63-linkage-specific polyubiquitin (cat. #12930; IB, 1:1000), rabbit IgG HRP (cat. #7074; IB, 1:10000), and mouse IgG HRP (cat. #7076; IB, 1:10,000) were obtained from Cell Signaling Technology, Inc. (Danvers, MA, USA). Antibodies against β-actin (cat. #A5316; IB, 1:10000), TRAF4 (cat. #MABC985; IB, 1:4000; IP, 1:200), Flag tag (cat. #F3165; IB, 1:10000; IP, 1:400), and Flag–HRP (cat. #A8592; IB, 1:20000) were from Sigma-Aldrich (St. Louis, MO, USA). Antibodies against HA tag (cat. #ab18181; IB, 1:5000; IP, 1:200) and His tag (cat. #ab18184; IB, 1:5000) were purchased from Abcam (Cambridge, UK). GFP-tag (cat. #TA150032; IB, 1:4000; IP, 1:400) antibody was obtained from OriGene (Rockville, MD, USA). Rabbit anti-TRAF4 (cat. #A302-840A; IB, 1:1000; IP, 1:200) and anti-CHK1 (cat. #A300-298A; IB, 1:1000; IP, 1:200) antibodies were purchased from Bethyl Laboratories (Montgomery, TX, USA). Antibody conjugates were visualized by chemiluminescence (cat. #34076, Thermo Fisher Scientific).

Flag–TRAF4 (cat. #RC200345), GFP–TRAF4 (cat. #RC200345L4), Flag–Chk1 (cat. #RC205094), and GFP–Chk1 (cat. #RC225807L4) were obtained from OriGene. pCDNA3–HA–Akt1 (cat. #73408) was obtained from Addgene (Watertown, MA, USA). His–Ub was a gift from Jianneng Li at Lerner Research Institute, Cleveland Clinic. Flag–Chk1 (DM-N), Flag–Chk1 (DM-C), Flag–TRAF4 (DM-RING), Flag–TRAF4 (DM-Inter), and Flag–TRAF4 (DM-TRAF), Flag–GFP–TRAF4 (C18A), Flag–GFP–TRAF4 (T192A), Flag–GFP-–TRAF4 (T192D), His–Ub (K6, K11, K27,K29, K33, K48, and K63), His–Ub (K48R), His–Ub (K63R), and Flag–Chk1 (S317/345A, K38R, K54R, K145R, K132R, K233R, K244R, K404R, K444R, K451R, and K456/458R) mutants were developed using the Q5 Site-Directed Mutagenesis Kit (cat. #E0554S, NEB) following the manufacturer’s protocol. All mutant constructs were generated using mutagenesis PCR were verified by Sanger DNA sequencing.

To generate CRISPR-Cas9-based TRAF4 and Chk1 knockout constructs, we cloned the annealed single-guide RNAs (sgRNAs) into the Bsm BI-digested lentiCRISPR V2 vector (cat. #52961, Addgene). The sgRNAs were from the Human CRISPR Knockout Pooled Library (GeCKO v2) [21] and are listed as follows: sgTRAF4#1 forward, 5′-AGCCACAAAACTCGCACTTG-3′; reverse, 5′-CAAGTGCGAGTTTTGTGGCT-3′; sgTRAF4#2 forward, 5′-CTCTGCCCATTCAAAGACTC-3′; reverse, 5′-GAGTCTTTGAATGGGCAGAG-3′; sgChk1#1 forward, 5′-AGTCATGGCAGTGCCCTTTG-3′; reverse, 5′-CAAAGGGCACTGCCATGACT-3′; sgChk1#2 forward, 5′-GAGATTCTTCCATCAACTCA-3′; reverse, 5′-TGAGTTGATGGAAGAATCTC-3′. The control or stable knockout cells were generated by transient transfection of sgRNA-inserted CRISPR-Cas9 plasmids and selected with 2 μg/ml puromycin for 2–3 weeks.

Cells were seeded at a density of 3 × 10³ cells/well in 96-well plates. After overnight incubation, fresh medium containing different doses of chemotherapeutic agents was added, and the plates were maintained for 0, 24, 48, or 72 h. The 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) reagent (cat. #G3581, Promega) was used for cell viability assays. Cell proliferation was measured using the 5-ethynyl-2′-deoxyuridine (EdU) assay. Briefly, cells were seeded at a density of 1 × 10⁴ cells in chamber slides (cat. #PEZGS0416, Millipore) and cultured overnight. Cells were then treated with 10 μM 5-Fu for 24 h. Next, 10 μM EdU (cat. #C10339, Thermo Fisher Scientific) was added to the culture medium, which was incubated for an additional 12 or 24 h. Cells were fixed and stained following the manufacturer’s instructions.

The soft-agar colony formation assay was performed as previously described [16]. Briefly, CRC cells were treated with chemotherapeutic agents or treated for 24 h, suspended (1 × 10⁴ cells per well) in 1 mL of 0.3% agar with Eagle’s basal medium containing 10% FBS, and overlaid into six-well plates containing a 0.6% agar base. The cultures were maintained at 37 °C in a 5% CO₂ incubator, and the colonies were counted at 2 weeks.

Cells were treated with chemotherapeutic agents or their vehicle control for 24 h and seeded into a 6-cm plate (300 cells/well). Cells were maintained for 2 weeks at 37 °C in a 5% CO₂ incubator. Colonies were fixed with 4% paraformaldehyde, stained with 0.5% crystal violet, and counted under a microscope. Three independent experiments were performed as indicated.

For the CHK1 IP/kinase assay, stable HT29 cells were treated with 5-Fu or UV light and lysed in IP buffer containing protease and phosphatase inhibitors. CHK1 was immunoprecipitated from cell lysates with an anti-CHK1 antibody (cat. #A300-298A, Bethyl Laboratories) at 4 °C for 4 h and washed twice in IP buffer and twice in 1× kinase buffer (cat. #9802, Cell Signaling Technology, Inc.). The recombinant GST–CDC25C protein (1 μg) was mixed with the immunoprecipitated CHK1 in a 20 μL reaction and incubated with 100 μM ATP in 1× kinase buffer at 30 °C for 30 min. Protein phosphorylation was examined by IB.

293T cells stably expressing Flag–CHK1 were lysed in an IP buffer containing protease and phosphatase inhibitors. IP was performed overnight at 4 °C with the whole-cell lysates using the Flag-tag antibody (cat. #F3165, Sigma-Aldrich) or normal mouse IgG (cat. #5415, Cell Signaling Technology). The IP proteins were resolved by SDS-PAGE, followed by Coomassie blue staining. The desired proteins were then digested and subjected to LC-MS at the Cleveland Clinic Proteomics Core. The peptides were analyzed using all collected collision-induced dissociation (CID) spectra and the search programs Sequest and Mascot. Protein and peptide identifications were validated using the program Scaffold.

Cells were lysed with a lysis buffer (6 M guanidine-HCl, 0.1 M Na₂HPO₄/NaH₂PO₄, 0.01 M Tris/HCl, pH 8.0, 5 mM imidazole, and 10 mM β-mercaptoethanol) supplemented with protease inhibitors and 10 mM N-ethylmaleimide (NEM, cat. #S3692, Selleck Chemicals). After sonication and centrifugation, the supernatant was incubated at room temperature for 4 h with 40 μL Ni-NTA-agarose beads (cat. #30210, Qiagen Inc.). The beads were centrifuged and washed sequentially with five buffers: (A) 6 M guanidine-HCl, 0.1 M Na₂HPO₄/NaH₂PO₄, 0.01 M Tris/HCl, pH 8.0, and 5 mM imidazole plus 10 mM β-mercaptoethanol; (B) 8 M urea, 0.1 M Na₂HPO₄/NaH₂PO₄, 0.01 M Tris/HCl, pH 8.0, 10 mM imidazole, and 10 mM β-mercaptoethanol plus 0.1% Triton X-100; (C) 8 M urea, 0.1 M Na₂HPO₄/NaH₂PO₄, 0.01 M Tris/HCl, pH 6.3, 10 mM β-mercaptoethanol, and 20 mM imidazole plus 0.2% Triton X-100; (D) 8 M urea, 0.1 M Na₂HPO₄/NaH₂PO₄, 0.01 M Tris/HCl, pH 6.3, 10 mM β-mercaptoethanol, and 10 mM imidazole plus 0.1% Triton X-100; and (E) 8 M urea, 0.1 M Na₂HPO₄/NaH₂PO₄, 0.01 M Tris/HCl, pH 6.3, 10 mM β-mercaptoethanol, 10 mM imidazole plus 0.05% Triton X-100. After the last wash, the beads were boiled with 2× SDS sample-loading buffer containing 200 mM imidazole, and the supernatant was separated by SDS-PAGE, followed by IB.

Cells were lysed with a modified RIPA buffer (20 mM NAP, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, and 1% SDS) supplemented with protease inhibitors (cat. #78430, Thermo Fisher Scientific) and 10 mM NEM. After sonication, the lysates were boiled at 95 °C for 15 min, diluted with RIPA buffer containing 0.1% SDS, then centrifuged at 1.6 × 10⁴×g for 10 min at 4 °C. The supernatant was incubated overnight at 4 °C with the primary antibodies and 40 μL protein A-Sepharose beads. After washing with RIPA buffer, the beads were boiled with 2× SDS sample-loading buffer to elute the bound protein. The eluted protein was then separated by SDS-PAGE, followed by IB. The Chromatin Extraction Kit (ab117152, Abcam) was used for chromatin and non-chromatin fractions extraction following the standard instruction.

The in vitro ubiquitination assay was performed as previously described [17]. Briefly, Flag–TRAF4, Flag–TRAF4 (DM-RING), and Flag–TRAF4 (C18A) were expressed in 293T cells, immunoprecipitated with anti-Flag antibody, and eluted with Flag peptide. Flag–TRAF4, Flag–TRAF4 (DM-RING), or Flag–TRAF4 (C18A) protein along with GST–CHK1 protein (cat. #14-346, Millipore) were incubated with 40 nM Ube1 (E1), 0.7 μM UbcH6 (E2), and 10 μM ubiquitin for 3 h at 37 °C in reaction buffer (50 mM Tris-HCl, pH 7.5, 5 mM MgCl₂, 1 mM DTT, and 2 mM ATP). After incubation, the protein mixtures were diluted with RIPA buffer and immunoprecipitated overnight with the anti-CHK1 antibody at 4 °C. CHK1 ubiquitination was then analyzed by IB.

All procedures for the xenograft mouse models were conducted under protocols approved by our Institutional Animal Care and Use Committee (IACUC). Xenograft tumors were established by subcutaneous injection of stable HT29, SW620, and H1975 cells (1 × 10⁶) into the right flank of 6-week-old NSG mice (The Jackson Laboratory, Bar Harbor, ME, USA). Tumors were measured with calipers every 2days, and mice received one of the following treatments by intraperitoneal injection when tumors reached a volume of 100 mm³: (1) vehicle, every 4 days; (2) 5-Fu, 50 mg/kg every 4 days; (3) oxaliplatin, 5 mg/kg every 4 days; (4) irinotecan, 10 mg/kg every 4 days; (5) cisplatin, 5 mg/kg every 4 days; or subcutaneous injection of (6) prexasertib, 16 mg/kg twice every 7 days; (7) the combination of prexasertib, 16 mg/kg twice every 7 days, and 5-Fu, 50 mg/kg every 4 days. Tumor volume was calculated with the following formula: tumor volume (mm³) = (length × width × width/2), where length is the longest diameter and width is the shortest diameter. Mice were euthanized at the endpoint, and xenografted tumors were dissected and analyzed.

CRC patients were diagnosed and classified by the Department of Pathology at The Third Xiangya Hospital and Xiangya Hospital according to the World Health Organization guidelines. All surgical specimens were collected in accordance with an Institutional Review Board-approved protocol. All subjects provided written informed consent for entry into this study. Individuals included 92 cases of primary adenocarcinomas with matched adjacent tissue, 22 cases of CRC liver metastasis, 12 cases of CRC lung metastasis, and 23 cases of relapse after 5-Fu + oxaliplatin + leucovorin combination chemotherapy.

Tumor tissues obtained from CRC patients or xenografts were subjected to hematoxylin and eosin (H&E) staining or IHC staining with specific antibodies. Briefly, the tissue sections were baked at 60 °C for 2 h, deparaffinized, and rehydrated. Slides were submerged into boiling sodium citrate buffer (10 mM, pH 6.0) for 10 min and incubated in 3% H₂O₂ in methanol for 10 min. Slides were blocked with 50% goat serum albumin for 1 h at room temperature and incubated with primary antibodies overnight in a cold room in a humidified chamber. Slides were washed with PBS, followed by incubation with the secondary antibody for 1 h at room temperature. Hematoxylin was used for counterstaining, which was evaluated independently by two pathologists. The percentage of positive cells was scored as follows: 0, no positive cells; 1, ≤ 10% positive cells; 2, 10–50% positive cells; and 3, > 50% positive cells. Staining intensity was scored as follows: 0, no staining; 1, faint staining; 2, moderate staining; and 3, dark staining. Comprehensive score = staining percentage × intensity. For TRAF4/p-CHK1 expression, a comprehensive score ≤ 1.5 is considered low expression and > 1.5 high expression. The association between TRAF4/p-CHK1 levels and patient clinical features, including age, gender, initial clinical stage, tumor stage, and lymph node status, is summarized in Supplementary Table 4 and Supplementary Table 5. The following antibodies were used for IHC staining: Ki67 (cat. #ab16667, Abcam, 1:250), TRAF4 (cat. #MABC985, Sigma-Aldrich, 1:300), p-CHK1 (cat. #ab47318, 1:200), and CHK1 (cat. #ab47574, 1:100). The Click-iT™ TUNEL Colorimetric IHC Detection Kit (cat. #c10623, Thermo Fisher Scientific) was used for apoptosis detection in tissue sections according to the manufacturers’ standard procedures.

Cells in a chamber slide were treated with 10 μM 5Fu for 24 h. The cells were fixed with 4% paraformaldehyde and permeabilized in 0.5% Triton X-100 for 20 min, followed by blocking in 5% BSA for 1 h and overnight incubation with γ-H2AX or p-CHK1 primary antibodies at 4 °C in a humidified chamber. Alexa Fluor 488 dye-labeled anti-rabbit IgG was used as the secondary antibody. Nuclei were counterstained with DAPI (cat. #P36935, Thermo Fisher Scientific). For measuring chemotherapeutic agent-induced γ-H2AX or p-CHK1, cells were seeded onto a chamber slide and treated with the drugs overnight. The chemotherapeutic agent-free medium was then added to the chamber slide, and the cells were fixed at various time points (0, 24, or 48 h) as indicated. The following antibodies were used for IF: γ-H2AX (cat. #ab11174, 1:100), p-CHK1 (cat. #ab47318, 1:200), α-tubulin (cat. #ab7291, 1:100), γ-tubulin (cat. #ab11317, 1:100), goat anti-rabbit IgG Alexa Fluor 488 (cat. #ab150077, 1:500), goat anti-mouse IgG Alexa Fluor 488 (cat. #ab150117, 1:500), and donkey anti-rabbit IgG Alexa Fluor Dye 647 (cat. #ab150075, 1:500).

Statistical analyses were performed using SPSS (version 16.0 for Windows, SPSS Inc., Chicago, IL, USA) and GraphPad Prism (GraphPad 7.0, San Diego, CA, USA). All quantitative data are expressed as the mean ± standard error of the mean (s.e.m) from 3 independent experiments. Differences between means were evaluated with Student’s t test or analysis of variance (ANOVA). The clinicopathologic significance of clinical samples was evaluated with a χ² test or a Fisher exact test for categorical data. The Mann–Whitney U test was used when the data did not fit a normal distribution. Kaplan–Meier analysis and the log-rank test (Mantel–Cox) were used for survival analysis. Pearson rank correlation was used for correlation tests. The Wilcoxon matched-pair signed-rank test was used for evaluating expression differences between tumors and adjacent non-cancerous tissues. A probability value of p ≤ 0.05 was used as the criterion for statistical significance.

To better understand the regulation of CHK1 activation, we performed a mass spectrometry analysis to identify novel CHK1-binding proteins in 293T cells stably expressing Flag-CHK1. In addition to known CHK1-associated proteins, including nucleic acid binding-related proteins and hydrolases, protein ubiquitin E3 ligases, including TRAF4, were identified as major interacting partners of CHK1 compared to the normal IgG negative control (Fig. 1a). The interaction between CHK1 and TRAF4 was confirmed by co-immunoprecipitation (Co-IP) of ectopically expressed proteins in 293T cells (Fig. 1b), and endogenous binding was confirmed via reciprocal IP in both HT29 (Fig. 1c) and SW620 (Fig. S1a) CRC cells. We further confirmed the direct binding of CHK1 and TRAF4 in vitro with purified protein samples (Fig. 1d). Deletion of the N-terminal domain of CHK1 significantly impaired the CHK1-TRAF4 interaction (Fig. 1e, g), while the C-terminal TRAF domain was essential for TRAF4 to bind to CHK1 (Fig. 1f, h). The endogenous interaction between TRAF4 and CHK1 was enhanced in HT29 cells that were treated with various chemotherapeutic agents (Fig. S1b).

Next, we determined whether TRAF4 was required for DNA damage-induced CHK1 activation in CRC cells. Strikingly, 5-Fu-induced phosphorylation of CHK1 at S345 and S317 was drastically decreased in TRAF4-knockout HT29 and SW620 cells with 2 independent sgRNAs when compared to control cells (Fig. 2a, Fig. S1c). However, TRAF4 knockout had no effect on 5-Fu-induced ATR activation (Fig. 2a). A similar reduction in CHK1 phosphorylation was observed in TRAF4-knockout cells treated with oxaliplatin (Fig. S1d) or irinotecan (Fig. S1e), which are two other first-line chemotherapeutic agents for CRC. Consistently, overexpression of TRAF4 enhanced CHK1 phosphorylation induced by 5-Fu in non-cancerous colon epithelial FHC cells, which expressed low levels of endogenous TRAF4 (Fig. 2c). Additionally, an in vitro kinase assay showed that the immunoprecipitated CHK1 from 5-Fu-treated TRAF4-knockout CRC cells exhibited decreased kinase activity (Fig. 2d), which was measured by the phosphorylation of CDC25C, a CHK1 substrate [22]. Moreover, 5-Fu treatment induced phosphorylation of CHK1 at the S345 (p-CHK1) foci in the nucleus of wild-type (WT) HT29 cells but not in TRAF4-knockout HT29 cells (Fig. 2e). Immunofluorescence (IF) staining revealed that exposure to 5-Fu for 24 h significantly enhanced γ-H2AX foci formation in both TRAF4-WT HT29 and TRAF4-knockout HT29 cells. However, TRAF4 deficiency decreased the efficacy of DNA damage repair as γ-H2AX foci formation after 5-Fu withdrawal was significantly reduced in TRAF4-WT HT29 cells but not in TRAF4-knockout HT29 and SW620 cells (Fig. 2f, Fig. S1f). Immunoblotting also revealed that γ-H2AX protein levels were higher in TRAF4-knockout cells when compared to TRAF4-proficient HT29 and SW620 cells (Fig. 2g, Fig. S1g). Similar results of the γ-H2AX foci formation were obtained in HT29 cells treated with oxaliplatin and irinotecan (Fig. S1h and S1i).

Because activation of CHK1 is required to prevent premature mitotic entry [23], we hypothesized that TRAF4 regulated mitosis by promoting CHK1 phosphorylation in CRC. In support of this hypothesis, over 95% of mitotic HT29 cells in the control group developed a bipolar mitotic spindle assembly during mitosis after 5-Fu treatment. In contrast, a loss of TRAF4 led to ~ 45% of 5-Fu-exposed mitotic cells with monopolar or multipolar spindles, chromosome misalignment during metaphase, or chromosome lagging during anaphase (Fig. S2a and 2b). Moreover, TRAF4 deficiency resulted in upregulation of the polyploidy cell population in 5-Fu-treated HT29 cells, and these polyploidy cells had decreased 5-ethynyl-2′-deoxyuridine (EdU) incorporation (Fig. S2c). Our results are consistent with the phenotypes for premature mitotic progression and mitotic catastrophe in CHK1-deficient or inactivated cells [23, 24]. These results support the framework that TRAF4 is required for CHK1 phosphorylation and activation in response to DNA damage.

Next, we investigated the molecular mechanisms by which TRAF4 promoted CHK1 activation. As depletion of TRAF4 had a negligible effect on ATR phosphorylation, but significantly impaired CHK1 activation (Fig. 2a), TRAF4 appeared to be a downstream target of ATR. We hypothesized that TRAF4 functioned as an E3 ligase to mediate CHK1 activation through ubiquitination. TRAF4 promoted CHK1 ubiquitination in 293T cells (Fig. 3b). Notably, the proteasome inhibitor MG132 did not enhance this process substantially (Fig. S3a), suggesting that the ubiquitination did not cause protein degradation. Consistent with this finding, substitution of WT ubiquitin with a mutant K63R markedly suppressed TRAF4-induced CHK1 ubiquitination (Fig. 3b). Moreover, treatment with 5-Fu facilitated the ubiquitination of endogenous CHK1 in SW620 cells (Fig. S3b). Although WT TRAF4 overexpression promoted CHK1 ubiquitination, the activities of the E3 ligase-defective TRAF4 mutants, DM-RING and C18A, were compromised in 293T cells (Fig. 3c). Interestingly, the binding of the DM-RING and C18A mutants to CHK1 was unaffected (Fig. S3c). An in vitro ubiquitination assay suggested that CHK1 was a direct target of TRAF4, but not the DM-RING and C18A TRAF4 mutants (Fig. 3d). Moreover, silencing of UbcH6, the ubiquitin-conjugating enzyme (E2), which is critical for TRAF4 E3 ligase activity [17], impaired TRAF4-mediated CHK1 ubiquitination (Fig. S3d). In addition, 5-Fu-stimulated endogenous CHK1 ubiquitination was diminished in TRAF4-deficient CRC HT29 and SW620 cells (Fig. 3e, Fig. S3e). To corroborate the role of TRAF4 in CHK1 ubiquitination, we introduced WT TRAF4 and its E3 ligase activity-deficient mutants into TRAF4-knockout HT29 cells. As shown in Fig. 3f 5-Fu-induced CHK1 ubiquitination was restored with WT TRAF4, but not with the DM-RING mutant or the catalytically inactive C18A mutant. Moreover, WT TRAF4 rescued 5-Fu- or UV-induced CHK1 phosphorylation (Fig. 3g, Fig. S3f) and 5-Fu-induced p-CHK1 foci formation (Fig. S3g). These data underscore the importance of TRAF4-mediated CHK1 ubiquitination to CHK1 phosphorylation and activation in CRC cells. Collectively, these results support that TRAF4 is a E3 ligase that controls CHK1 ubiquitination in CRC cells.

To identify CHK1 ubiquitination sites, we analyzed the evolutionarily conserved lysine residues among CHK1 proteins from different species. The 11 most highly conserved sites were primarily localized in its N-terminal kinase domain and the C-terminal conserved motif 2 (CM2) (Fig. 4a). In vivo ubiquitination assays demonstrated that TRAF4 promoted the ubiquitination of both WT CHK1 and mutant CHK1 with a C-terminal deletion, but not mutant CHK1 with an N-terminal deletion in 293T cells (Fig. 4b). These data suggested that the ubiquitination sites are localized in the N-terminal kinase domain of CHK1. Using single-site mutants of CHK1, we discovered that only the mutation of K132 to arginine (R) markedly attenuated CHK1 ubiquitination in 293T cells (Fig. 4c). We therefore focused on CHK1 K132 ubiquitination and investigated its role in CHK1 activation. The K132R mutation caused a robust decrease in 5-Fu-induced CHK1 ubiquitination in HT29 cells (Fig. S4a). Moreover, 5-Fu-induced phosphorylation of S317 phosphorylation was reduced in HT29 cells exogenously expressing the K132R CHK1 mutant compared to that with WT CHK1 (Fig. S4b). Reconstitution of WT CHK1, but not the K132R mutant, restored 5-Fu-induced CHK1 S345 phosphorylation in CHK1-knockout HT29 and SW620 cells (Fig. 4d, Fig. S4c). In addition, in vitro kinase assays showed that K132R substantially impaired CHK1-mediated CDC25C phosphorylation in HT29 treated with 5-Fu (Fig. 4e) or in 293T cells treated with UV (Fig. S4d). A previous study suggested that the adaptor protein B-cell translocation gene 3 (BTG3) is required for UV-induced CHK1 ubiquitination in HCT116 cells [13]. We found that BTG3 or TRAF4 silencing decreased CHK1 ubiquitination and that TRAF4 knockdown more effectively inhibited CHK1 ubiquitination when compared to BTG3 knockdown in HT29 cells (Fig. S4e). To determine whether phosphorylation of CHK1 at S345 and S317 was required for ubiquitination, a CHK1 (S317/345A) double mutant was constructed for an in vivo ubiquitination assay in 293T cells. 5-Fu failed to induce phosphorylation of the S317/345A mutant (Fig. S4f), yet TRAF4-mediated ubiquitination of WT CHK1 was similar to that of the S317/345A mutant (Fig. 4f). These results suggest that TRAF4 promotes CHK1 ubiquitination independent of phosphorylation at S317/345, whereas phosphorylation at S317/345 is dependent on CHK1 ubiquitination.

To determine how ubiquitination regulates CHK1 phosphorylation, we isolated the chromatin fraction of HT29 cells treated with 5-Fu. Given that chromatin binding is required for ATR kinase-mediated CHK1 phosphorylation, we examined whether ubiquitination enhanced CHK1 chromatin association. Depletion of TRAF4 robustly attenuated 5-Fu-induced CHK1 ubiquitination on chromatin (Fig. 4g) and in non-chromatin fractions (cytoplasmic proteins and soluble nuclear proteins). Moreover, the total protein levels of CHK1 in the chromatin fraction were decreased in TRAF4-knockout HT29 cells (Fig. 4g), indicating that TRAF4-deficient impaired 5-Fu induced CHK1 ubiquitination and chromatin association. In addition, the Co-IP result indicated that the interaction between ATR and CHK1 in chromatin fraction was decreased in TRAF4-knockout HT29 cells (Fig. 4h). Consistently, mutation of the major ubiquitination site, K132R, reduced chromatin-associated CHK1 ubiquitination and CHK1 expression in response to 5-Fu treatment (Fig. 4i). Loss of ubiquitination on K132 impaired 5-Fu-induced ATR and CHK1 interaction (Fig. 4j). These results indicate that the ubiquitination of CHK1 on K132 is required for DNA damage-induced CHK1 chromatin association and subsequent ATR interaction and activation.

We next investigated whether TRAF4 affects the sensitivity of human CRC cells to chemotherapy. TRAF4 knockout attenuated cell viability in HT29 and SW620 cells in the presence of 5-Fu (Fig. 5a, Fig. S5a). The anchorage-dependent and -independent colony formation potential of both TRAF4-knockout HT29 and SW620 cells was impaired in the presence of 5-Fu (Fig. 5b, c; Fig. S5b and S5c). Similar inhibitory effects were observed in HT29 cells treated with other chemotherapeutic agents, including oxaliplatin and irinotecan (Fig. S5d and S5e). TRAF4 knockout in HT29 cells enhanced chemotherapy-induced apoptosis, which was indicated by the upregulation of cleaved-caspase 3 and -PARP (Fig. 5d). Flow cytometry data showed that the apoptotic cell population was increased in 5-Fu-treated TRAF4-knockout HT29 cells (Fig. S5f); consistently, the comet assay revealed that knockout of TRAF4 enhanced 5-Fu-induced DNA damage (Fig. S5g). These results suggest that TRAF4 deletion promotes the anti-tumor effects of chemotherapeutic agents. Restoration of WT TRAF4, but not the E3 ligase-inactive mutants DM-RING and C18A, promoted the colony formation and anchorage-independent cell growth of TRAF4-knockout HT29 cells receiving 5-Fu treatment (Fig. 5e, f). This in vitro evidence implies that the TRAF4 E3 ligase activity is required for TRAF4-mediated desensitization of CRC cells to chemotherapy. To determine whether CHK1 K132 ubiquitination was required for chemoresistance in CRC cells with high TRAF4 expression, we introduced WT CHK1 or the CHK1 K132R mutant into HT29 cells with stable CHK1 knockdown. Reconstitution of WT CHK1, but not the ubiquitination-deficient K132R mutant, attenuated 5-Fu-induced suppression of cell viability (Fig. 5g) and anchorage-dependent and -independent colony formation (Fig. 5h, i).

Xenograft tumors derived from TRAF4-knockout HT29 cells were treated with 5-Fu and exhibited a significant decrease in tumor growth, tumor mass, and tumor cell proliferation when compared to tumors derived from TRAF4-knockout cells that did not receive the 5-Fu treatment or to tumors retaining WT TRAF4 and treated with 5-Fu (Fig. 5j–m). Moreover, 5-Fu promoted CHK1 phosphorylation in WT-TRAF4 xenograft tumors but attenuated in TRAF4-knockout xenograft tumors (Fig. 5m). In addition, 5-Fu induced increased apoptosis in TRAF4-knockout xenograft tumors when compared to WT-TRAF4 xenograft tumors (Fig. 5m). Similar results were observed in tumors derived from SW620 cells (Fig. S5h and S5i). Other chemotherapeutic agents, such as oxaliplatin and irinotecan, also inhibited TRAF4-knockout tumor development from HT29 cells (Fig. S5j–m). In the TRAF4-knockout HT29 tumors, reintroduction of WT TRAF4 impaired the anti-tumor effectiveness of the chemotherapeutic agents, but reintroduction of the DM-RING and C18A TRAF4 mutants, which both lack the E3 ligase activity, did not (Fig. 5n–p, Fig. S5n). We next determined the chemotherapeutic function of CHK1 K132 ubiquitination in vivo. Xenograft tumors derived from CHK1-knockdown HT29 cells that were treated with 5-Fu exhibited reduced tumor growth (Fig. 5q). Reconstitution of WT CHK1, but not the ubiquitination-deficient mutant CHK1 K132R, attenuated the anti-tumor efficacy of 5-Fu (Fig. 5q). These results suggest that TRAF4 deletion or deficient TRAF4 E3 ligase activity restricts CHK1 activation and inhibits xenograft tumorigenesis during CRC chemotherapy, and that CHK1 K132 ubiquitination is required for maintaining chemoresistance in CRC cells.

To demonstrate the clinical relevance of our findings, we evaluated p-CHK1 (S345) and TRAF4 protein levels in 92 primary CRC tissues. The expression of TRAF4 was significantly increased in tumor tissues when compared to paired adjacent non-cancerous tissues (Fig. 6a, b). Notably, p-CHK1 (S345) was upregulated in colorectal tumor tissues (Fig. 6a, c). A greater percentage of tissues exhibited positive TRAF4 or p-CHK1 staining when tumor tissues were compared to adjacent non-cancerous tissues (Fig. S6a and S6b). Importantly, increased TRAF4 expression was associated with worse overall survival among CRC patients in our cohort or in one from the Human Protein Atlas (Fig. 6d, Fig. S6c). Representative images of low and high TRAF4 and p-CHK1 expression levels in tumor tissues from our cohort were shown in Fig. 6e. We found that TRAF4 protein levels were significantly and positively correlated with p-CHK1 levels (Fig. 6f). Among 92 patients, 59 had tumors with high TRAF4 staining, and 48 of these 59 cases also had high p-CHK1 staining (Fig. 6g). Both TRAF4 (Table S1) and p-CHK1 (Table S2) levels positively correlated with the clinical stage at the initial diagnosis (p < 0.01) and the lymph node status (p < 0.05).

The correlation between TRAF4-p-CHK1 signaling and sensitivity to chemotherapy was further confirmed in 23 cases of paired CRC specimens before chemotherapy and at relapse after 5-Fu + oxaliplatin + leucovorin combination therapy. The relapsed tumors exhibited significantly higher TRAF4 and p-CHK1 levels when compared to the paired untreated tumors. The expression of Ki67 was not different between the paired samples (Fig. 6h–j). Analysis of the pre- and post-treatment paired specimens from individual patients indicated that TRAF4 expression positively correlated with p-CHK1 expression in relapsed specimens (Fig. 6k, l). Collectively, these results suggest that the TRAF4-CHK1 signaling is associated with 5-Fu resistance in CRC.

To determine the relationship between TRAF4–p-CHK1 signaling and chemotherapy resistance in CRC, we constructed 5-Fu-resistant cells using three human CRC cell lines: SW620, HT29, and HCT8. Daughter cell clones with IC₅₀ values at least 20-fold greater than the parental cells were deemed resistant. TRAF4 and p-CHK1 were significantly upregulated in cells resistant to 100 μM 5-Fu, when compared to the parental cell lines (Fig. 7a). Moreover, TRAF4 expression positively correlated with p-CHK1 expression in the 5-Fu-resistant CRC cells (Fig. 7b). To determine whether a high level of TRAF4 was required for maintaining chemotherapy resistance, we generated stable TRAF4-knockout cells in two 5-Fu-resistant cell lines: HT29R and SW620R. High concentrations of 5-Fu substantially attenuated cell viability (Fig. 7c, Fig. S7a) and colony formation (Fig. 7d, Fig. S7b) in the parental HT29 and SW620 cells, both with or without TRAF4. However, the inhibitory effects were only observed in the TRAF4-knockout HT29R and SW620R cells. No inhibition was present in the TRAF4-WT HT29R or the SW620R cells. Treatment with 5-Fu robustly increased the cleavage of PARP and caspase 3 in TRAF4 knockout-resistant cells, but the effect was less pronounced in TRAF4 WT-resistant cells (Fig. 7e, Fig. S7c). These results indicate the apoptosis pathway was activated after 5-Fu treatment.

To determine whether overexpression of TRAF4-mediated chemoresistance was associated with CHK1 activation in CRC cells, we examined the efficacy of a CHK1 inhibitor, prexasertib, in abrogating 5-Fu resistance. Although 5-Fu or prexasertib alone partially decreased cell viability in HT29R (Fig. 7f) and SW620R (Fig. S7d) cells, the combination therapy enhanced the anti-tumor effects. Similar results were also observed for anchorage-dependent and -independent colony formation (Fig. 7g, h; Fig. S7e and S7f). Consistently, prexasertib treatment promoted 5-Fu-induced γ-H2AX expression (Fig. 7i) and increased the cleavage of PARP and caspase 3 in HT29R cells (Fig. 7j). Xenograft tumors that were derived from TRAF4-knockout HT29R cells and treated with 5-Fu exhibited a significant decrease in growth and weight when compared to tumors derived from HT29R cells with WT-TRAF4 (Fig. S7g and S7h). The combination of 5-Fu and prexasertib synergistically reduced tumor growth and weight when compared to each of the single agent treatments (Fig. 7k, l). Together, these results suggest that the TRAF4-p-CHK1 signaling plays a critical role in 5-Fu resistance and that a combination of 5-Fu with CHK1 inhibitors could be a promising strategy to overcome chemoresistance in CRC cells.