Fig 2: Working model.Sororin suppresses WAPL and maintains sister chromatid cohesion in Sororin cohesin cells. In contrast, in PD-L1 overexpressed cells, PD-L1 compensates for the loss of Sororin and competes with WAPL binding to PDS5B, thereby regulating sister chromatid cohesion.

Fig 3: Expression correlation relationship of menin/MAPs genes and OXPHOS/glycolytic genes. (A) Workflow of the in silico correlation analysis of gene expression in The Cancer Genome Atlas (TCGA) breast cancer cohort. (B) Heatmaps of the expression correlation between MEN1/selected 4 MAP genes and OXPHOS genes (upper) or glycolytic genes (lower) in normal (N) and tumor (T) samples. The genes are arranged from the highest to the lowest according to gene expression correlation coefficients of MEN1–OXPHOS genes or MEN1–glycolytic genes in breast tumors. (C) and (D) Scatter plots and linear regression analyses of MEN1/selected MAPs expression and mean expression of OXPHOS genes (C) or glycolytic genes (D) in normal and tumor samples. (E) Violin plots (lower panel) shows the average expressions of the genes of OXPHOS complexes I-V and glycolysis in the samples of each of the corresponding 4 groups are shown as violin plots. Based on the median values (where ≥median is “high” and KMT2A, MEN12, WAPL, and GATA3) and MEN1, the TCGA breast tumor samples were divided into 4 groups—1: high-low, 2: high-high, 3: low-low, and 4: low-high (upper panel). Letters on top of the violin plot denote statistical significance, where two groups with different letters are significantly different (p < 0.05) and those with the same letter are not.

Fig 4: Bioenergetic dynamics are regulated by menin and MAPs in T47D and MCF-7 cells. (A) and (B) Glycolytic and OXPHOS ATP productions in T47D (A) or MCF-7 (B) cells infected with vehicle, shMEN1, shKMT2A, shMED12, shWAPL, or shGATA3 lentivirus. Statistics represented the difference of glycolytic or OXPHOS ATP production between shRNA knockdown and vehicle controls. (C) and (D) Bar charts representing mitochondrial functions in the single knockdown of MEN1, KMT2A, MED12, WAPL, or GATA3 and their vehicle control in T47D (C) or MCF-7 (D) cells. (E) Schematic summary of mitochondrial dynamics affected by the knockdown of MEN1 or MAPs. (F,G) Bar charts represented the glycolytic functions in T47D (F) or MCF-7 (G) cells subject to gene knockdown by shMEN1, shKMT2A, shMED12, shWAPL, or shGATA3 lentivirus. (H) Schematic summary of glycolytic functions affected by the knockdown of MEN1 or MAPs. Data are presented as mean ± S.D. (n = 15–20 technical-replicate wells). Statistical significance was performed by an unpaired two-tailed Student’s t-test between treated groups and corresponding controls. * p < 0.05, ** p < 0.01, and *** p < 0.001.

Fig 5: Nuclear PD-L1 interacts with cohesion complex.a, b PD-L1 interacts with cohesin complex. a Doxycycline-inducible PD-L1 widetype (WT) (+/+) and knockout (KO) (−/−) MDA-MB-231 cells were harvested and fractionated. The nuclear fraction was incubated with monoclonal anti-PD-L1 antibody or corresponding IgG. The immunoprecipitates were blotted with indicated antibodies. b Cells were blocked with thymidine and transfected with indicated siRNAs. After releasing from the second round of thymidine block, cells were released into S phase in the presence of MG132 for 3 h. Cells were harvested and nuclear fractions were processed for immunoprecipitation. c PD-L1 is involved in the regulation of cohesion in the presence of WAPL. MDA-MB-231 cells expressing control shRNA, PD-L1 shRNA, WAPL shRNA, or a combination of PD-L1 and WAPL shRNAs were fixed and dropped on coverslips. Percentages of cells with different sister chromatid cohesion status were quantified. Data are presented as means ± SEM, and were independently replicated three times. d–f PD-L1 YSR-like motif (YKR) is essential for PD-L1 binding with PDS5B in vivo, and competes with WAPL for binding to PDS5B in vitro. PD-L1 wild-type (WT), PD-L1 AIA mutant (mutating both FGF-like motifs to AIA), and PD-L1 AKE mutant (mutating YKR to AKE) plasmids were transfected into cells and nuclear fractions were isolated and incubated with (d) HA-conjugated sepharose beads of (e) protein A beads conjugated with anti-PDS5B antibody. The immunoprecipitates were blotted with indicated antibodies. f Left panel: in vitro competition of binding affinity of PDS5B (purified from insect cells) and purified WAPL (1–30)-GST and PD-L1 (purified from nuclear fraction of MDA-MB-231 cells). Right panel: in vitro interactions between PD1 and wildtype (WT) or AKE mutant PD-L1 purified from cytoplasmic/membrane fraction (C/M) of MDA-MB-231 cells. g–j PD-L1 YSR-like motif is essential for sister chromatid cohesion and cell proliferation. myb gene distance between paired FISH signals (g), and colony formation (i) were measured in MDA-MB-231 cells infected with control shRNA, PD-L1 shRNA, PD-L1 shRNA plus WT shRNA resistant PD-L1 overexpression construct, or PD-L1 shRNA plus shRNA resistant AKE mutant PD-L1 overexpression construct. Scale bar: 10 µm. Data are presented as means ± SEM, and were independently replicated three times with similar results. Statistical significance was calculated using Student’s t-test. **P < 0.01, ***P < 0.001. h Protein levels in cells used for assays in (g and i) were detected by Western blot with indicated antibodies. j B16F10 cells expressing control shRNA (shCtrl), mouse Pd-l1 shRNA (shPd-l1), Pd-l1 shRNA plus shRNA resistant mouse wildtype Pd-l1 (shPd-l1 + Pd-l1 wt), or mouse Pd-l1 shRNA plus shRNA resistant ake mutant Pd-l1 (shPd-l1 + Pd-l1 ake) were injected into C57BL/6 mice. Mice were treated with 5 mg/kg mouse Pd-l1 blocking antibody or corresponding IgG. The tumor growth was monitored by measuring with a caliper every 3–4 days. Data are presented as means ± SEM (n = 5). Two-way ANOVA was used for the comparisons. **P < 0.01, ***P < 0.001.