Fig 2: CBFB and hnRNPK regulate translation. a Flow chart showing the in vitro translation assay to test the effect of a translation element (RNA) and a protein. In vitro transcribed RNA was used in the in vitro translation assays using HeLa cells lysate. b In vitro translation assay using the 1-Step Human coupled IVT kit showing the effect of CBFB and hnRNPK binding sites (T14) on RUNX1 translation. c In vitro translation assays using the 1-Step Human coupled IVT showing effect of recombinant CBFB (rCBFB) and hnRNPK (rhnRNPK) on RUNX1 translation. d In vitro translation assay using WT and CBFB_KO HEK 293 T cells lysate. See “Methods” section for details of preparation of 293 T cell lysate. e In vitro translation assay. Un-transfected (control) and hnRNPK knockdown (KD) HEK 293T cells lysate replaced the HeLa cell lysate in the 1-Step Human coupled IVT kit with or without the supplement of rhnRNPK. f IB showing the effect of CBFB deletion on the proteins encoded by bound transcripts. g IB showing the effect of hnRNPK KD on the proteins encoded by bound transcripts

Fig 3: Ablating Hnrnpk in hypertrophic chondrocytes results in enhanced transdifferentiation potential.A Skeletal preparation of P0 (left panel) and P7 (right panel) WT and cCKO mice. Scale bar: 1 mm. B Quantification of limb length of P0 (top) and P7 (bottom) WT and cCKO mice. n = 3 biological replicates. C von Kassa staining of humerus of P0 and P7 WT and cCKO mice. Red arrow: excessive mineral bone in cCKO mice. Scale bar: 100 μm. D Immunostaining of Sp7 (left panel) and Cathepsin K (Ctsk) (right panel) of femur of P0 WT and cCKO mice (left panel) and the quantification of Sp7 and Ctsk positive cells (right panel). Dotted lines: boundary between hypertrophic zone and POC. n = 3 biological replicates. Scale bar: 100 μm. E Immunostaining of Sp7 of P0 Col10a1-Cre;Rosa26-tdTomato and Col10a1-Cre;Hnrnpkfl/fl;Rosa26-tdTomato. White boxes: magnification of metaphysis. White arrow: tdTomato and Sp7 double-positive cells. Scale bar: 100 μm. F Total quantity and ratio of single positive cells and double-positive cells in (E) of P0 WT and cCKO mice. Statistical analysis was exerted upon a ratio of double-positive cells between WT and cCKO mice. n = 3 biological replicates. p-value was calculated by two-tailed unpaired Student’s t-test. Data were shown as mean ± SD. *p < 0.05; **p < 0.01; ns: not significant.

Fig 4: Elevated activity of Hif-1 signaling pathway in Hnrnpk null chondrocytes increases glycolytic intensity.A mRNA fold changes of target genes of Hif-1 signaling pathway of E18.5 CKO growth plate cartilage compared to WT according to RNA-seq results. B mRNA expression level of key enzymes of oxidative phosphorylation and glycolysis of E18.5 WT and CKO growth plate cartilage (left panel). Protein level of Pfkfb3 and Ldha of E18.5 WT and CKO growth plate cartilage (right panel). Red boxes: most active glycolytic enzymes in chondrocytes. n = 3 biological replicates. Densitometry results were expressed as fold change in protein levels compared with WT1 after normalized to β-actin. C mRNA expression level of key enzymes of oxidative phosphorylation and glycolysis of Hnrnpkfl/fl chondrocytes infected with Ad-GFP or Ad-Cre under normoxic or hypoxic conditions. Statistical analysis was exerted between Hypoxia Ad-GFP and Hypoxia Ad-Cre. n = 3 biological replicates. D Protein level of Vegfa, Ldha, and Pfkfb3 of Hnrnpkfl/fl chondrocytes infected with Ad-GFP or Ad-Cre under normoxic or hypoxic conditions. Densitometry results were expressed as fold change in protein levels compared with chondrocytes infected with Ad-GFP under normoxic condition after normalized to β-actin. E Protein level of Hif1α, Ldha, and Pfkfb3 of Hnrnpkfl/fl chondrocytes treated with DMSO or Chetomin after being infected with Ad-GFP or Ad-Cre under hypoxic condition. Densitometry results were expressed as fold change in protein levels compared with chondrocytes infected with Ad-GFP under hypoxic condition after normalized to β-actin. F Glucose consumption, lactate production, and ratio of lactate/glucose of Hnrnpkfl/fl chondrocytes infected with Ad-GFP or Ad-Cre under normoxia or hypoxic conditions. n = 3 biological replicates. G Diagram indicated the strategy of treating Hnrnpkfl/fl chondrocytes with PBS or 3-BrPA after being infected with Ad-GFP or Ad-Cre under hypoxic condition. H mRNA expression level and protein level of Sox9 and Mmp13 in Hnrnpkfl/fl chondrocytes treated with PBS or 3-BrPA after being infected with Ad-GFP or Ad-Cre under hypoxic condition. n = 3 biological replicates. Densitometry results were expressed as fold change in protein levels compared with chondrocytes infected with Ad-GFP under hypoxic condition after normalized to β-actin. p-value was calculated by one-way ANOVA followed by Tukey’s multiple comparisons tests (C, F, H) or two-tailed unpaired Student’s t-test (B). Data were shown as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001; ns: not significant.

Fig 5: CBFB binds to RUNX1 mRNA via hnRNPK. a Silver staining of FLAG pulldown using CBFB KO MCF10A cells expressing empty vector and N-terminally FLAG tagged CBFB. b IB validation of CBFB and hnRNPK interaction. c IB showing the effect of hnRNPK knockdown on RUNX1 protein. d RNA immunoprecipitation (RIP) with CBFB antibody in WT and CBFB_KO MCF10A cells. Error bars are SEM, n = 3 (biological); two asterisks, p value < 0.01. (RIP of CBFB in WT vs. in CBFB KO cells, two-tailed t test). e RIP with hnRNPK antibody. Error bars are SEM, n = 3 (biological); two asterisks, p value < 0.01 (RIP of hnRNPK in WT vs. CBFB KO cells, two-tailed t test). f RNA pulldown assays (RPA) determining the hnRNPK-bound region within RUNX1 mRNA. F1-4, fragment 1 to 4 of 3′ UTR of RUNX1 mRNA. See Methods for details. Numbers indicate the nucleotide positions. g Re-analyses of two public eCLIP datasets (GSM2423241 and GSM2423242) of hnRNPK on the RUNX1 locus. Nucleotide sequences of two poly-C tracts within the binding site of hnRNPK were shown. F3, fragment 3 of 3′-UTR; T14, truncation 14 of 3′-UTR F3; Mu1, mutation 1; Mu2, mutation 2. h RPA using recombinant CBFB in the absence or presence of recombinant hnRPNK. i RPA using recombinant hnRNPK in the absence or presence of recombinant CBFB