Human embryonic kidney cell (HEK293T), human colon adenocarcinoma cell (HT29) and human cervical cancer cell (HeLa) were purchased from American Type Culture Collection (ATCC) (Manassas, VA, USA). Human rhabdomyosarcoma cell (RD) was obtained from China Center for Type Culture Collection (CTCC) (Wuhan, China). All the cells were cultured in DMEM medium with 10% fetal bovine serum, 1% penicillin–streptomycin and maintained at 37 °C in a 5% CO2 incubator.

For transfection, the cells were seeded into plate and transfected with indicated plasmids and Lipofectamine 2000 (Lipo2000) (Invitrogen, Carlsbad, CA, USA) mixture according to the manufacturer’s instructions.

The EV71 (Xiangyang-Hubei-09) (GenBank accession no. JN230523.1), CVB3 (Nancy strain) and PV1 (vaccine strain) were obtained as described previously (Su et al., 2020). The viruses were propagated in RD cells. For viral infection assays, the cells were infected with virus at indicated multiplicity of infection (MOI). After 2 h, the unabsorbed virus was washed off and the infected cells were cultured in no FBS DMEM medium for specific time.

Rabbit anti-EV71 VP1 (PAB7631-D01P) antibody was purchased from Abnova (Taipei, Taiwan). Rabbit anti-CBX4 (A5532), anti-EV71 3D (A22651) and anti EV71-3C (A20687) antibodies were obtained from ABclonal Technology (Wuhan, China). Mouse anti-β-actin (66009-1-Ig), anti-GAPDH (60004-1-Ig) and anti-Myc (60003-2-Ig) antibodies and rabbit anti-HA (51064-2-AP) antibody were purchased from ProteinTech Group (Wuhan, China). Mouse anti-EV71 3D (GTX630193) antibody was obtained from GeneTex (Irvine, CA, USA). Mouse anti-Flag (F1804) and anti-HA (901514) antibodies were obtained from Sigma-Aldrich (St. Louis, MO, USA) and BioLegend (San Diego, CA, USA), separately. Inhibitors including UNC3866 (HY-100832), 2-D08 (HY-114166), cycloheximide (CHX) (HY-12320) and MG132 (HY-13259) were purchased from MedchemExpress (Monmouth Junction, NJ, USA).

The DNA fragments of EV71 2A, 2B, 2C, 3AB, 3C and 3D were amplified from the cDNA of EV71 and inserted into pCAGGS-HA or pEGFP-C1 vector. The CBX4, SUMO1, SUMO2, SUMO3, UB and Ubc9 genes were obtained from cDNA of HeLa cells and ligated with corresponding vectors.

The titers of EV71 in culture medium supernatants were determined by plaque assay. In brief, RD cells were plated in 24-well plate, and added with different diluent virus samples for 2 h. Then, the cells were washed with PBS and cultured in DMEM medium supplemented with 2% FBS and 1% agarose. Three days later, the cells were fixed with 4% paraformaldehyde for 30 min. After washing three times with PBS, the cells were stained with crystal violet dye. The viral titers (PFU/ml) in supernatants from each group were determined by averaging the plaque counts from three independent wells in a certain volume of diluent.

Total RNAs were extracted from cultured cells with TRIzol reagent (Vazyme, Nanjing, China) and reverse transcription was performed with 1 μg RNA into cDNA using HiScript II Q Select RT SuperMix (Vazyme). Real-time quantitative PCR (qPCR) analysis was executed using ABI 7500 Fast real-time PCR system (Applied Biosystem, Carlsbad, CA, USA). The relative changes in gene expression were determined by 2^–ΔΔCt method. The primers used for qPCR were as follows: GAPDH, forward, 5’-AAGGCTGTGGGCAAGG-3′, reverse, 5’-TGGAGGAGTGGGTGTCG-3′; EV71 VP1, forward, 5’-GAGTTCCATAGGTGACAGC-3′, reverse, 5’-CTGTGCGAATTAAGGACAG-3′; CVB3, forward, 5’-CGGTACCTTTGTGCGCCTGTT-3′, reverse, 5’-GCGGTGCTCATCGACCTGA-3′; PV1, forward, 5’-CCGTATTGAGCCAGTATGTTTGT-3′, reverse, 5’-TAGCGAGTAGGTGGAGGTGTTCT-3′; HoxA10, forward, 5’-ACAAGAAATGTCAGCCAGAAAGG-3′, reverse, 5’-GATGAGCGAGTCGACCAAAAA-3′; HoxB9, forward, 5’-AGGCCGTGCTGTCTAATCAAA-3′, reverse, 5’-CGAGCGTGCAGCCAGTT-3′; HoxC13, forward, 5’-AAGGTGGTCAGCAAATCGAAAG-3′, reverse, 5′- TGGTACAAAGCGGAGACATAAATAGA-3′.

The cells were lysed in RIPA buffer (50 mM Tris–HCl, 150 mM NaCl, 0.25% sodium deoxycholate, 1% NP-40, pH 7.4) and the protein extractions were centrifugated at 12,000 rpm for 10 min in a 4 °C centrifuge. After removing insoluble cell debris, 10% of lysate was collected as input sample, and the rest of supernatants were rotated with indicated antibodies overnight at 4 °C. Then, the samples were added with Protein G agarose (Santa Cruz, CA, USA) for 2 h to collect IP complex. For the SUMOylation assay, HEK293T cells were transfected with indicated plasmids for 24 h. After being lysed in RIPA buffer supplemented with 20 μΜ NEM (Sigma-Aldrich), the samples were immunoprecipitated with anti-GFP immunomagnetic beads (Yeasen, shanghai, China). The concentration of proteins was quantified by BCA Protein Assay Kit (Solarbio, Beijing, China). After loading and separating in 10% SDS-PAGE gel, the proteins were transferred to a NC membrane (Millipore, Billerica, MA, USA). After being blocked with 5% skim milk for 2 h at room temperature, the membrane was incubated with specific primary antibodies and secondary antibodies. The band was analyzed with a luminescent image analyzer (Amersham Imager 600, GE, MA, USA).

The cytoplasmic and nuclear proteins were prepared with a Nucleoprotein Extraction Kit (Solarbio) according to the manufacturer’s instructions. Briefly, the collected cells were lysed with cytoplasmic protein extraction buffer and vortex for 15 s. After centrifugation at 12,000 g for 10 min in a 4 °C centrifuge, the prepared supernatants were collected for cytoplasm protein detection. Subsequently, the precipitations were treated with nuclear protein extraction buffer and vortex for 15 s. The samples were centrifugated at 12,000 g for 10 min in 4 °C and the lysates were collected for nuclear protein detection.

The cell viability was measured by a cell counting kit 8 (CCK8) assay. RD cells with 2,000 cells per well were plated in 96-well plate and treated with inhibitor or virus. After 24 h, each well was added with 10 μL CCK8 solution and incubated for 1 h at 37 °C. The cell viability was detected at 450 nm absorbance.

The cells were plated on 20-mm coverslips and treated with plasmids transfection or virus infection. After indicated time interval, the cells were fixed with 4% paraformaldehyde, permeabilized with 0.4% Triton X-100, and then blocked with 5% bovine serum albumin (BSA). After incubating with indicated primary antibodies overnight at 4 °C, the samples were stained with fluorescein-conjugated secondary antibodies for 1 h at room temperature. Nuclei were stained with DAPI (Roche, Basel, Switzerland). The samples were imaged with confocal laser scanning microscopy (Leica Microsystem, Wetzlar, Germany).

The short-hairpin RNAs (shRNAs) target negative-control (NC) and CBX4 were inserted in pLKO.1 vector. The shRNA sequences were designed as follows, shNC, 5’-CAACAAGATGAAGAGCACCAA-3′; shCBX4, 5’-GAGTGGAGTATCTGGTGAAAT-3′. For lentivirus package, HEK293T cell were plated in 6 cm dishes and transfected together with 1.5 μg psPAX2, 0.5 μg pMD2.G and 2 μg pLKO.1 ligating with shNC or shCBX4 plasmids with Lipo2000. At 24 h and 48 h, the lentivirus supernatants from HEK293T cells were collected. After being infected with lentivirus for 48 h, RD, HEK293T and HT29 cells were selected by puromycin (2.5 μg/mL; Sigma-Aldrich). The knockdown efficiency of shRNA was evaluated by Western blotting.

HEK293T or RD cells were seeded in 6-well plate and subsequently transfected with indicated plasmids or infected with virus. After indicated time interval, the cells were added with CHX at the concentration of 100 μg/mL for different time points. The cell lysates were prepared for Western blot analysis. The relative protein level of 3D polymerase to internal control GAPDH was quantified by Image J software.

All experiments were reproducible and repeated at least three times. The data were presented as mean ± SD. Statistical analysis for comparison of two means was assessed by unpaired Student’s t-test. Statistical significance in protein stability assay was calculated by two-way ANOVA. Analyses were performed with Prism 8 software (GraphPad Software Inc., La Jolla, CA, USA). p-value < 0.05 was considered as statistically significant.

To interpret the function of SUMO E3 ligase CBX4 in EV71 replication and pathogenicity, we firstly assessed whether EV71 infection alters CBX4 expression. Following dose-dependent EV71 infection in RD cells, the levels of VP1 and 3C proteins were increased as expected, in contrast, CBX4 protein expression was maintained at a steady level (Supplementary Figure 1A). We then transfected increasing amounts of CBX4 plasmid in RD cells and found that with the elevation of CBX4 protein level, the replication of EV71 was markedly enhanced (Figure 1A). Furthermore, RD cells were transfected with CBX4 or control vector plasmid, and then infected with EV71 at various time points. Our results showed that the VP1 and 3C proteins and VP1 mRNA were induced upon EV71 infection, and this induction was upregulated by transfection of CBX4 compared with transfection of vector plasmid (Figures 1B,C). Immunofluorescence analyses indicated that overexpression of CBX4 promoted the expression of EV71 replication intermediate dsRNA distinctly (Figure 1D). At the same time, the effect of CBX4 on productive viral release was determined by plaque assay. We observed that CBX4 enhanced the viral titers in cell supernatants (Figures 1E,F). Overall, these results suggested that overexpression of CBX4 obviously facilitates the infection of EV71 in cells and release of infectious virus particles.

Moreover, we constructed lentiviruses expressing shNC or shCBX4 to transduce RD cells. As expected, shNC transduction did not significantly affect EV71 replication compared to non-transduced cells (Supplementary Figure 1B), and shCBX4 effectively attenuated CBX4 expression (Supplementary Figure 1C). Additionally, HT29 stably expressing shNC or shCBX4 cell lines were also generated. After reducing CBX4 expression, the cytopathic effect caused by EV71 was relieved in comparison to shNC expressing cells (Supplementary Figure 1D). The infection of RD and HT29 stable shNC cells with different doses of EV71 led to a robust increase in the VP1 and 3C proteins and VP1 mRNA, but these productions were depressed in shCBX4-treated cells (Figures 2A,B). Furthermore, RD and HT29 cells were infected with EV71 for different time points. Compared with shNC in RD and HT29 cells, the expression of VP1 and 3C proteins and VP1 mRNA were reduced in the presence of shCBX4 across the time course (Figures 2C,D). These results indicated that knockdown of CBX4 inhibits EV71 replication in a does- and time-dependent manner. Immunofluorescence analyses illustrated that knockdown of CBX4 diminished the expression of EV71 dsRNA (Figure 2E). Plaque assay revealed that knockdown of CBX4 restrained viral release in the culture medium supernatants (Figures 2F,G).

CBX4 is critical for EV71 infection, which raises the question of whether other enteroviruses also utilize CBX4 to regulate their replication. We showed that similar with EV71, the replication levels of CVB3 and PV1 decreased upon knockdown of CBX4 (Supplementary Figures 1E,F), suggesting that CBX4 also supports the replication of other enteroviruses. Taken together, the data revealed that CBX4 not only plays a vital role in EV71 replication, but also extends to the other enteroviruses.

To delineate the specific domain of CBX4 involved in EV71 replication, Flag-tagged CBX4 full length, 1-288 aa, 288-405 aa, and 401-560 aa truncation plasmids were constructed (Figure 3A). RD cells were subsequently transfected with these constructs for functional analysis. In contrast to CBX4 full length, which promoted EV71 VP1 and 3C proteins expression, 1-288, 288-405, and 401-560 mutants had no significant effect (Figure 3B), revealing that only full length of CBX4 is associated with the EV71 replication. We then investigated whether the impact of CBX4 on EV71 replication depended on its Polycomb-related or SUMO E3 ligase function. RD cells were treated with an CBX4 chromodomain-histone interaction antagonist (Stuckey et al., 2016), UNC3866 at different concentrations. Cell viability assays showed that the activity of RD cells remained stable throughout these concentration ranges (Figure 3C). Importantly, UNC3866 treatment upregulated known PRC1 target genes such as HoxA10, HoxB9 and HoxC13 (Supplementary Figure 2), providing functional evidence that it effectively inhibits the chromodomain of CBX4. We observed that UNC3866 did not impair the replication of EV71 (Figure 3D). Compared with full-length CBX4, its mutants defective in SUMO ligase activity (lacking SIM1 or SIM2) did not induce the upregulation of VP1 and 3C when overexpressed (Figure 3E). Therefore, the SUMOylation E3 ligase rather than Polycomb-related function of CBX4 is essential for the replication of EV71.

The possible mechanism by which CBX4 regulates EV71 replication was evaluated by a co-IP screening experiment among several EV71 non-structural proteins and CBX4. HEK293T cells were co-transfected with the plasmids expressing Flag-tagged CBX4 and HA-tagged EV71 2A, 2B, 2C, 3AB, 3C, and 3D, respectively. We only detected the expression of Flag-CBX4 in HA-3D-IP complex (Figure 4A), demonstrating that 3D specifically interacts with CBX4. Additionally, the combination of CBX4 and 3D was also proved by co-transfected with Flag-3D and HA-CBX4 plasmids (Figure 4B). Furthermore, RD cells were transfected with Flag-CBX4 expressing plasmid, and then infected with EV71. It was shown that CBX4 was association with 3D rather than 3C generated by EV71 replication (Figure 4C). Previous studies demonstrated that CBX4 mainly takes effect in nuclear compartment (Zhao et al., 2024; Gu et al., 2024), however, immunofluorescence analysis of another study indicated that CBX4 is also expressed in cytoplasm (Liang et al., 2022). As a single positive-stranded RNA virus, the life cycle of EV71 is completed in cytoplasm (van der Linden et al., 2015). Therefore, we performed a nucleus and cytoplasm isolation experiment of RD cells after non-infection and infection with EV71. Our results showed that CBX4 was predominantly distributed in the nucleus, while a small apart was localized in the cytoplasm regardless of EV71 infection (Figure 4D), which provides evidence for CBX4 to regulate EV71 infection in cytoplasm. Next, we found the co-localization of 3D and CBX4 in the cytoplasm by immunofluorescence analysis (Figure 4E). Moreover, the interaction region of 3D and CBX4 was explored, and co-IP results revealed that 3D combined with the 401-560 aa of CBX4 (Figure 4F). Altogether, these findings consistently demonstrated the interaction between CBX4 and EV71 3D protein.

We transfected a gradient increased Flag-CBX4 plasmid together with a constant amount of HA-3D plasmid in HEK293T cells and found that the expression of 3D was markedly upregulated accompanied by the increase of CBX4 expression (Figure 5A), which hinted the possibility of CBX4 in stabilizing 3D protein. Subsequently, CHX chase experiment was carried out to measure the function of CBX4 in 3D stability. HEK293T cells were transfected with HA-3D plasmid alone or co-transfected HA-3D and Flag-CBX4 plasmids jointly, and then treated with the protein synthesis inhibitor CHX at different time points, as well as proteasome inhibitor MG132 and CHX together. The results indicated that co-transfection of CBX4 and 3D prolonged the half-life of 3D, and MG132 treatment attenuated 3D degradation (Figures 5B,C). We next investigated the effect of CBX4 knockdown on the influence of 3D half-life. CHX chase assay demonstrated that knockdown of CBX4 reduced the stability of 3D (Figures 5D,E). Moreover, RD stable shNC or shCBX4 cells were infected with EV71 and treated with CHX, the results showed that knockdown of CBX4 correspondingly diminished the stabilization of the 3D protein induced by EV71 infection (Figure 5F). Collectively, these data disclosed that CBX4 augments the half-life of EV71 3D polymerase.

It is previously reported that SUMO1-mediated SUMOylation and K63-linked ubiquitination upregulates the stabilization of 3D (Liu et al., 2016). Since CBX4 is a SUMO E3 ligase, and SUMOylation interacts with ubiquitination, the role of CBX4 on 3D SUMOylation and ubiquitination was examined. Firstly, we detected the SUMO1-based SUMOylation of 3D by overexpression of Ubc9, and such 3D-linked SUMOylation was notably enhanced upon co-transfection of Ubc9 and CBX4 (Figure 6A). Moreover, 3D-linked SUMO2 and SUMO3 levels were also upregulated by CBX4 (Supplementary Figures 3A,B). The ubiquitination levels of 3D were investigated. We found that overexpression of CBX4 enhanced whole ubiquitination and K63-linked ubiquitination of 3D (Figures 6B,C). Next, we determined the role of CBX4 SIM motifs in regulation 3D SUMOylation and ubiquitination. After the loss of either SIM1 or SIM2, the SUMOylation and k63-linked ubiquitination levels of 3D decreased, and when both SIM1 and SIM2 were absent, the SUMOylation and k63-linked ubiquitination levels of 3D were further reduced (Figures 6D,E). In addition, HEK293T stable shNC or shCBX4 cells were transfection with 3D, SUMO1 and Ubc9 expressing plasmids. The results showed that knockdown of CBX4 inhibited the SUMOylation level of 3D (Figure 6F). Meanwhile, we also detected that knockdown CBX4 significantly reduced the whole ubiquitination and K63-linked ubiquitination levels of 3D (Figures 6G,H). These data together suggested that CBX4 is involved in the regulation 3D SUMOylation and ubiquitination.

Based on the above observations, we want to determine the regulatory relationship between SUMOylation and EV71 replication. We selected a SUMOylation inhibitor 2-D08, which is able to prevent SUMO transfer from Ubc9-SUMO thiol to substrates and inhibit the SUMOylation of substrate proteins (Kim et al., 2014). Firstly, the influence of different concentrations of 2-D08 on cell viability was evaluated by CCK8 assay. The results indicated the cells were stable in these concentrations of 2-D08 (Figure 7A). The function of 2-D08 on the replication of EV71 was dissected in RD and HT29 cells. With the increase of 2-D08 treatment concentration, the cell viability diminished by EV71 were significantly enhanced (Supplementary Figure 4A), the 3D, VP1, and 3C protein levels were decreased progressively (Figure 7B), as well as the VP1 mRNA was gradually declined (Figure 7C). The virus copy number (Supplementary Figure 4B) and viral titers (Figures 7D,E) in cell supernatants were also repressed by 2-D08. Besides, 2-D08 effectively depressed the viral RNA expression of CVB3 and PV1 (Supplementary Figures 4C,D) and restored the cell viability reduced by CVB3 and PV1 (Supplementary Figures 4E,F). We next explored the effect of 2-D08 on ectopically transfected 3D expression. HEK293T were added with different concentrations of 2-D08 and transfected with HA-3D expressing plasmid, the expression level of 3D was reduced by 2-D08 (Figure 7F). In addition, 2-D08 reduced the CBX4-mediated SUMOylation of 3D (Figure 7G). Thus, inhibition of SUMOylation of cell declines 3D protein level and EV71 replication capacity. In general, these findings indicated a mechanism by which CBX4 interacts with EV71 3D polymerase to mediate SUMOylation of 3D and enhance the stability of 3D, leading to facilitating the replication of EV71 in the end (Figure 8).