Fig 2: Effect of UFM1 and PDK1 on prognosis of patients with gastric cancer. a The expression of PDK1 protein in gastric tumor tissues and respective adjacent tissues were analyzed using IHC, magnification, × 100 and × 400. b PDK1 expression scores are shown as box plots. The expression of PDK1 in gastric tumor tissues and respective adjacent tissues was compared using the t-test. n = 120 (*, P < 0.005). c Kaplan-Meier survival curve of gastric cancer patients with “High PDK1” or “Low PDK1” (P < 0.05, log-rank test). d Kaplan-Meier analysis of the correlation between the combined expression of UFM1 and PDK1 with the overall survival of gastric cancer patients (P < 0.05, log-rank test). e Working model of the role of UFM1 in PDK1 signaling and gastric cancer cell invasion

Fig 3: UFM1 suppresses the metastatic potential and epithelial-to-mesenchymal transition of gastric cancer in PDK1-dependent manner. a UFM1 associates with PDK1 in gastric cancer. Immunoprecipitation using PDK1 antibody was performed in AGS cell lysates (up panel). Stable AGS cells (down panel) were collected, lysed, and cell lysates were applied to immunoprecipitation with UFM1 antibody. b UFM1 promoted PDK1 ubiquitination. 293 T cells were cotransfected with constructs as indicated. PDK1 was immunoprecipitated with an anti-PDK1 antibody, and the ubiquitinated PDK1 was visualized by Western blot analysis using an anti-Ub antibody. c Immunofluorescence images showing the changes in PDK1 in stable AGS cells. d Stable AGS cells were treated with Control siRNA or PDK1 siRNA then cell lysates were applied in western blot analysis. e The stimulatory effect of UFM1 downregulation on AGS cell migration and invasion was rescued by PDK1 siRNA transfection; scale bar, 50 μm. f Quantitative results of (e) is show. The data are presented as the mean ± SD (*P < 0.05; **P < 0.01; ns, no significance)

Fig 4: AKT3-174aa interacts with p-PDK1 and prevents AKT thr308 phosphorylation. a RNA sequencing was conducted in U251-circ-AKT3 IRES mut, U251-circ-AKT3, U373-circ-AKT3 IRES mut, U373-circ-AKT3 cells. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed to find the most differentially changed signaling pathways. b Upper, U251, U373, SW1783 and Hs683 cells were overexpressed or knocked down circ-AKT3, according to their endogenous AKT3-174aa expression level. The AKT3-174aa, AKT1, AKT2, AKT3, AKT-pan, AKT-thr308, AKT-ser473, p-EGFR, PTEN level were determined. Lower, AKT1, AKT2 and AKT3 mRNA level were determined in above cells, respectively. c IP assay using AKT1, AKT2 and AKT3 antibodies were performed in indicated cells followed by immunoblot using AKT1, AKT2, AKT3, AKT-thr308, AKT-ser473 antibodies. d Total proteins from Flag-AKT3-174aa plasmid-transfected HEK293T cells were separated via SDS-PAGE. P-PDK1 was identified by LC/LC-MS in AKT3-174aa protein complex. e Mutual interaction of p-PDK1 and Flag-AKT3-174aa were determined by IP. f Left, prokaryotic purified His-tagged-PDK1, GST-tagged-AKT3-174aa and GST was subjected to GST pull down or His-IP. Right, eukaryotic purified His-tagged-PDK1, prokaryotic purified GST-tagged-AKT3-174aa and GST was subjected to GST pull down or His-IP. g Flag-tagged AKT3-174aa was transfected into U251 cells and immunofluorescence was performed using anti-Flag and anti-p-PDK1 antibody. Scale bar, 20 µm

Fig 5: Overexpression of miR-138 protects cardiac cells against hypoxia through targetting PDK1(A) AC16 cells were transfected with control mimic, control mimic plus PDK1, miR-138 mimic alone, or cotransfection of miR-138 and PDK1 for 48 h. Western blot analysis of PDK1 protein expression. A representative image is presented using ß-actin as the loading reference. (B) AC16 cells were transfected with control mimic, control mimic plus PDK1, miR-138 mimic alone, or cotransfection of miR-138 mimic and PDK1 overexpression plasmid for 48 h. The glucose uptake, (C) lactate product, (D) oxygen consumption, (E) mitochondrial respiration, and (F) intracellular ATP were measured. (G) AC16 cells were transfected with control mimic, control mimic plus PDK1, miR-138 mimic alone, or cotransfection of miR-138 and PDK1 for 48 h. Cells were exposed to hypoxia at 0, 24, 48, 72, and 96 h, followed by the detection of cell viability by MTT assay. All experiments were performed in triplicate. Data are presented with the indication of mean ± S.D. *: P<0.05; **: P<0.01; ***: P<0.001.