Fig 2: AKT1 interaction with UPF1 requires UPF3X. A proximity ligation assay was performed to assess the interactions between AKT1 and UPF1 or UPF3X after downregulation of UPF1 or UPF3X with siRNA. The white arrows indicate typical interaction points. The white squares correspond to a magnification of the background image. The bar plot at the bottom of the figure shows the average number of interaction points per cell as determined for >200 cells for each condition. Error bar = S.D., P-values were calculated with Student's t-test: *<0.05, **<0.01. All the results of this figure are representative of two experiments.

Fig 3: AKT1 interacts with NMD factors. (A) AKT1 immunoprecipitation analysis. AKT1 was immunoprecipitated from HEK293FT WT cells. The immunoprecipitations were performed in the presence of RNase A or BSA to assess the requirement for RNA in the protein interactions. The immunoprecipitate was then analyzed by western blotting for the presence of co-immunoprecipitated factors. MDM2 as a partner of AKT1 was used as positive control. (B) UPF1 immunoprecipitation analysis. UPF1 was immunoprecipitated from HEK293FT WT cells and the immunoprecipitate analyzed by western blotting for the presence of co-immunoprecipitated factors. (C) UPF1 immunoprecipitation analysis under NMD activation. UPF1 was immunoprecipitated from HEK293FT WT cells transfected with a construct expressing PTC-carrying globin mRNA in the presence of RNase A or BSA to assess the requirement for RNA in the protein interactions. The immunoprecipitate was then analyzed by western blotting for the presence of co-immunoprecipitated factors. The right panel is the result of an RT-PCR performed on RNA extracted from the immunoprecipitates to assess the efficiency of the RNase A treatment. (D) UPF3X immunoprecipitation analysis. UPF3X was immunoprecipitated from HEK293FT WT cells and the immunoprecipitate analyzed by western blotting for the presence of co-immunoprecipitated factors. GFP immunoprecipitation was performed as a negative control to assess the specificity of the immunoprecipitations. The immunoprecipitations were performed in the presence of RNase A or BSA to assess the requirement for RNA in the protein interactions. The immunoprecipitate was then analyzed by western blotting for the presence of co-immunoprecipitated factors. The right panel is the result of RT-PCR performed on RNA extracted from the immunoprecipitates to assess the efficiency of RNase A treatment. The three leftmost lanes in each panel of the figure correspond to serial dilutions of HEK293FT whole cell extract. The results presented in this figure are representative of two experiments.

Fig 4: UPF1 knockdown and WT and NSUN2 KO cells.(A) Genome tracks showing the RNA levels of NSUN2 (left) and UPF1 (right) genes in four experimental groups. N = 3. (B) Clusters of replicates from 4 treatment groups. Z-score is the scaled cpm. (C) Scatter plot showing the relative RNA levels upon knockdown of UPF1 in WT (left) and NSUN2 KO (right) HeLa cells. P-values were calculated as t test. (D) Scatterplot showing the relative RNA levels upon UPF1 KD, in WT (x axis) and NSUN2 KO (y axis) HeLa cells. mRNAs modified with a “in region” (±50 nt from UPF1 footprint) NSUN2-mediated m5C sites are highlighted in red. Candidates chosen for RT-qPCR validation (Fig 4D) are shown in the figure. (E) Scatterplot showing the relative RNA levels upon UPF1 KD, in WT (x axis) and NSUN2 KO (y axis) HeLa cells. mRNAs modified with a “in region” (±50 nt from UPF1 footprint) m5C sites characterized by m5CUCCA sequence motif typical of NSUN6-dependent sites are highlighted in blue.

Fig 5: UPF1 function is affected by the lack of NSUN2.(A) Western blot of extracts from wild-type and NSUN2 KO HeLa cells after siRNA-mediated UPF1 knockdown (siUPF1). A scrambled siRNA (siSCR) was used as control. Shown are probings for UPF1, NSUN2; ACTB was used as loading control. A representative image from replicated experiments (N = 3) is shown. (B, C) Box plot showing the relative RNA levels (measured by RNA-seq) of UPF1 targets with m5C sites located within (B) or outside (C) a ±50 nt interval around the mid-point of UPF1 footprints, upon UPF1 knockdown in WT (solid line), and NSUN2 KO (dashed line) HeLa cells. mRNAs groups are colour-coded depending on their enrichment level with UPF1 protein. P-values were calculated using t test. (B, D) RT-qPCR measurement of UPF1 mRNA target levels (selected from panel (B)), upon UPF1 knockdown in WT and NSUN2 KO HeLa cells. PEA15 and ITM2B were used as non-methylated, positive and negative controls for UPF1 binding, respectively. Values were normalised over ACTB and “siUPF1” values expressed relative to “siSCR” treatment set as 1. P-values were calculated using two-tailed t test. t test-derived P-values: *P < 0.05; **P < 0.01; ***P < 0.001. N = 3. (E) The HeLa mRNA transcriptome was subdivided into five categories (Roman numerals) and overlapped, based on UPF1 binding, m5C sites targeted by NSUN2, and distance between the two. (F) Box plot showing RNA levels (measured by RNA-seq) in NSUN2 KO and WT HeLa cells in control (siSCR, solid line) and UPF1 knockdown (siUPF1, dashed line) conditions. (E) RNA groupings and colour-coding are defined in the Venn diagram in panel (E). P-values were calculated using t test.