Fig 2: PIAS4 upregulated KDM5B to promote hypoxia adaption. (A) Relative cell viability of SGC7901 was determined by the CCK8 assay after co-transfected with PIAS4 siRNA or NC siRNA and FLAG-KDM5B or corresponding empty vector as indicated for 72 h under 1% O2 conditions. The NC group in vehicle was normalized to calculated the relative viability (mean ± SD, n = 3, one-way ANOVA, *p < 0.05). (B) The protein level of KDM5B was detected in SGC7901 cells after co-transfected with PIAS4 siRNA or NC siRNA and FLAG-KDM5B or corresponding empty vector as indicated under 1% O2 conditions by Western blotting. (C) Relative cell viability of BGC823 cells was determined by the CCK8 assay as described in (A). (D) The protein level of KDM5B was detected in BGC823 cells as described in (B). (E) SGC7901 cells was co-transfected with PIAS4 siRNA or NC siRNA and Flag-KDM5B or corresponding empty vector as indicated for 24 h before replating as single-cell suspension for colony formation for 10 days under 1% O2 conditions. (F) The number of colonies in (E) with more than 50 cells was counted and quantified (mean ± SD, n = 3, one-way ANOVA, *p < 0.05). (G) Colony formation assay was performed as described in (E) in BGC823 cells. (H) The number of colonies in (G) with more than 50 cells was counted and quantified (mean ± SD, n = 3, one-way ANOVA, *p < 0.05). (I) SGC7901 cells were co-transfected with PIAS4 siRNA or NC siRNA and FLAG-KDM5B or corresponding empty vector as indicated for 24 h, followed by 1% O2 treatment for 24 h. The cells were collected for PI/Annexin V flow cytometry analysis of cell apoptosis, and the percentage of apoptotic cells was statistically analyzed (mean ± SD, n = 3, one-way ANOVA, *p < 0.05).

Fig 3: The metabolism of nuclear WT and mutant Blimp-1 proteins is triggered by PML-mediated Sumo-2/3 modification. a WT or mutant Blimp-1-expressing plasmids were co-transfected with empty plasmid (pcDNA3.0 (-) B) or PIAS1-, PIAS2a-, PIAS2ß-, PIAS3-, PIAS4-, CBX4-, or PML-expressing plasmids into 293T cells. Total lysates were analyzed by western blotting. b WT or mutant Blimp-1-expressing plasmids were co-transfected with empty plasmid (pCMV6-AC-DDK) or PML-, CBX4-, or PIAS1-expressing plasmid into 293T cells, which were then treated with MG132 or left untreated for 12 h. The lysates were immunoprecipitated with anti-flag antibody and then immunoblotted with antibodies against Sumo-1 or Sumo-2/3. c PML -/- HeLa cells supplemented with mCherry-tagged PML were transduced with GFP-tagged WT or mutant Blimp-1-expressing plasmids and then treated with MG132 for 12 h before being inspected under a confocal microscope. Scale bar: 7.5 µm. d PML -/- HeLa cells with or without PML supplementation were treated with MG132 or left untreated for 24 h before whole-cell lysates were collected for immunoprecipitation and western blotting. e PML -/- HeLa cells with or without PML supplementation were transduced with flag-tagged WT or mutant Blimp-1-expressing plasmids for 48 h before being analyzed via western blotting. f WT or mutant Blimp-1-expressing plasmids were co-transfected with empty plasmid (pCMV6-AC-DDK) or PML-expressing plasmids into 293T cells and treated with MG132 or left untreated for 24 h. The nuclear lysates were immunoprecipitated with an anti-flag antibody and then immunoblotted with an antibody against Ub. g WT or mutant Blimp-1-expressing plasmids were transfected alone or together with si-SUMO-2/3 into PML-overexpressing 293T cells, which were then treated with MG132 or left untreated for 24 h. The nuclear lysates were immunoprecipitated with an anti-flag antibody and then immunoblotted with an antibody against Ub

Fig 4: KDM5B increased HIF1a expression in response to hypoxia. (A and B) The protein levels of HIF1a in SGC7901 cells (A) and BGC cells (B) with treatment of 1% O2 (left panel) or CoCl2 (100 µM) (right panel) for 6 h after knocking down KDM5B with siRNA for 48 h were determined by Western blotting. (C and D) JIB04 was added to SGC7901 cells (C) and BGC cells (D) with two concentrations for 18 h before the treatment of 1%O2 (left panel) or CoCl2 (100 µM) for 6 h. The protein level of HIF1a was determined by Western blotting. (E and F) qPCR was conducted to determine the expression of HIF1a target genes (VEGFA and GLUT1) after knocking down KDM5B with siRNA (E) or JIB04 treatment (F) in SGC7901 cells cultured under 1%O2 for 12 h (mean ± SD, n = 3, one-way ANOVA, *p < 0.05). (G and H) The protein level of HIF1a in SGC7901 cells (G) and BGC823 cells (H) with treatment of 1%O2 (left panel) or CoCl2 (100 µM) for 6 h after knocking down PIAS4 with siRNA for 48 h was determined by Western blotting. (I) The HIF1a target gene (VEGFA and GLUT1) expression in SGC7901 cells after PIAS4 knockdown with siRNA was detected by qPCR (mean ± SD, n = 3, one-way ANOVA, *p < 0.05).

Fig 5: PIAS4 was the SUMO E3 ligase for hypoxia-induced KDM5B SUMOylation. (A) The protein level of the KDM5B protein in SGC7901 cells under normoxia or 6 h 1% O2 treatment with or without PIAS4 knockdown by siRNA for 48 h was analyzed by Western blotting. (B) SGC7901 was transiently transfected with PIAS4 siRNA for 48 h, later with cycloheximide (CHX) and 1% O2 treatment for the indicated times, and then the half-life of KDM5B was analyzed by Western blotting and followed by measurement by ImageJ. The p-value is derived from ANCOVA by comparing NC and si-PIAS4 (mean ± SD, n = 3, ANCOVA, *p < 0.05). (C,D) Co-IP was performed to detect the interaction between exogenous expression FLAG-KDM5B and HA-PIAS4 in HEK293T. The FLAG (C) and HA (D) antibody (IgG was used as negative control) was used for immunoprecipitation and the other for Western blotting analysis. (E) SUMOylation assay was employed to HEK293T and co-transfected with indicated plasmids for 48 h, followed by 1% O2 treatment for 6 h. The SUMOylation of KDM5B with or without the exogenous expression of the wild-type PIAS4 was determined by Western blotting. (F) The SUMOylation analysis of KDM5B in HEK293T co-transfected with the E3 activity inactivation mutant PIAS4 (HA-PIAS4 ?RING (Sachdev et al., 2001) and the wild-type PIAS4 under hypoxia conditions. (G) The interaction between exogenous FLAG-KDM5B and HA-PIAS4 under hypoxia conditions was determined by co-IP in HEK293T. (H) The SUMOylation of KDM5B under normoxia and hypoxia (6 h 1% O2) was analyzed as indicated by the SUMOylation assay in HEK293T.