G-quadruplex located in the 5′UTR of the BAG-1 mRNA affects both its cap-dependent and cap-independent translation through global secondary structure maintenance

Figure 1.. Scheme of the BAG-1 mRNA organization. (A)The BAG-1 mRNA presents many features in its 5′UTR: an rG4 secondary structure located at its 5′ end; four in-frame start codons with the first being a non-canonical CUG; and, an IRES secondary structure. (B) Translational initiation at the three principal alternative start codons results in the production of three protein isoforms (1L, Long, p50; 1M, Medium, p46; and, 1S, Short, p36) differing from each other by the size of their N-terminal extension. The rarely expressed, shortest BAG-1 isoform p29 is not represented. Its translation starts from the AUG located at position 502. (C) Nucleotide sequence of the complete BAG-1 5′UTR. The rG4 region, the start codons, the IRES region and the ribosome binding sites are highlighted.

Figure 2.. RNA and protein expression levels of BAG-1 in the paired tissues of colorectal tumors at different stages and in their adjacent healthy tissue (margin). (A) Relative expression levels of the BAG-1 mRNA in the paired adjacent healthy tissues (Margin, white circle) and the tumors (black square) at different stages as measured by RT-qPCR. Each tissue pair is connected by a line (n = 8 for adenoma and stage 1, n = 10 for stages 2, 3 and 4). The statistical analysis performed was a paired t-test, with the tumor being compared to the margin. ns = no significant difference, * P≤ 0.05. (B) Protein expression levels, as measured by Western blot, of the three BAG-1 isoforms in the same pairs of margin-tumor tissues as in (A). The positions of the protein ladder are indicated on the left. ERK2 was used as the loading control. The relative densities of the 2 most abundant isoforms, BAG-1M and BAG-1S, are annotated under each lane. The isoforms’ band densities were corrected on the corresponding ERK2 loading control band density and reported relative to the density, of the isoform of the first lane of each blot which was set to 1.00 (n = 3 for adenoma and stage 1, n = 4 for stages 2, 3 and…

Figure 3.. Stabilization of the rG4 with chemical ligands. (A) A representative gel of the BAG-1 endogenous protein isoforms from HCT116 cells after a 24 h treatment with 2 μM of the specific rG4 ligand cPDS compared to that of an untreated control. (B) Relative densitometry of the BAG-1 isoforms’ protein bands after a 24 h treatment of the cells with the ligand (black bar), or of an untreated control (white bar). The ligand concentrations used were: cPDS 2 μM, Phen-DC3 10μM and TmPyP4 2μM. (C) Relative expression levels of the BAG-1 mRNA in the ligand treated cells as compared to the untreated cells. For both (B) and (C), the results are presented as the means and standard deviations of n = 2 (each ligand treatment was performed in triplicate). The statistical analyses performed were an unpaired t-test between the ligand treated and the control untreated cells for each ligand. *P≤ 0.05, **P≤ 0.001.

Figure 4.. Luciferase, RNA and protein isoform expression levels of reporter assays of the complete 5′UTR of BAG-1 with both the mutated rG4 and the mutated 1S start codon. (A) The luciferase assays’ means and standard deviations of the Rluc luminescence levels, normalised over the Fluc luminescence levels, are shown for the WT (black) and rG4 mut (white) psiCHECK-2 constructions that included or not the 1S start codon mutation. The experiments were repeated three times with each of the constructions being transfected in triplicate (n = 3). The statistical analysis performed is a two-way ANOVA with Tukey's multiple comparison. ****P≤ 0.0001. (B) Relative expression (RE) levels of the Rluc RNA normalised over that of the Fluc RNA after the transfections of the different mutated constructions as measured by RT-qPCR. The bar of the RE level of the reporter plasmid without the insertion of the BAG-1 5′UTR is labeled psiCHECK-2. The statistical analysis performed was a two-way ANOVA with Tukey's multiple comparison test (n = 2). ****P≤ 0.0001. (C) Representative immunoblot of the Rluc N-extension protein isoforms’ expression levels after both the rG4 and 1S start codon mutations. The psiCHECK-2…

Figure 5.. Luciferase, RNA and protein expression levels in reporter assays of the 5′UTR of BAG-1 possessing the mutated AUG-254. (A) The luciferase assays’ means and standard deviations of the Rluc luminescence levels, normalised over the Fluc luminescence levels, are shown for all constructions. The statistical analysis performed was a two-way ANOVA with Tukey's multiple comparison (n = 2, each construction was transfected in triplicate). ***P≤ 0.001, ****P≤0.0001. (B) Relative expression (RE) levels of the Rluc RNA normalised over that of the Fluc RNA after the transfections of the different mutated constructions as measured by RT-qPCR. The bar of the RE level of the reporter plasmid without the insertion of the BAG-1 5′UTR is labeled psiCHECK-2. The statistical analysis performed was a two-way ANOVA with Tukey's multiple comparison test (n = 2). ***P≤ 0.001, ****P≤ 0.0001. (C) Representative immunoblot of the Rluc N-extension protein isoforms expression levels after both the rG4 and the AUG-254 start codon mutations. The psiCHECK-2 transfection lane represents the canonical Rluc without any N-terminal extension. Mock indicates the untransfected control. β-actin was used as a loading…

Figure 6.. rG4 mutation impairs the cap-independent translation of the BAG-1 IRES. (A) The ratios of Fluc/Rluc luciferase levels following transfection of the bicistronic plasmid construct are shown. The HCV IRES (gray) bicistronic construct was used as a positive control, with the well characterized IRES placed upstream of Fluc. The WT (black) represent the bicistronic construct with the full length 5′UTR of BAG-1 located upstream of the Fluc. The G4mut (white) represents the bicistronic construct with the full-length 5′UTR of BAG-1 that included the G-to-A mutations abolishing the folding of the rG4. For each experiment, all constructions were transfected in triplicate. The results presented are the means and standard deviations of n = 5 independent experiments. The statistical analysis performed was a paired t-test. *P≤ 0.05. (B) The ratios of the relative RNA expression levels of Rluc and Fluc following transfection. The ratios are close to 1 and are similar between the three constructs, demonstrating the integrity of the bicistronic construct. The bars indicate the means and standard deviations of n = 3. (C) Representation of the WT Stem-loop III secondary structure as defined by…

Figure 7.. Effects of both the rG4 and IRES mutations on the global secondary structure of the BAG-1 5′UTR, as analyzed by SHAPE. (A) Super-n-motif representation of the 114 predicted secondary structures, separated into three clusters (white, cluster 1; gray, cluster 2; black, cluster 3). (B) Distribution in the three clusters of the predicted secondary structure of every sequences analyzed. (C) Arc-plot representation of the most stable predicted secondary structure of the complete WT (G4ss) sequence (green) compared to that of the rG4mut sequence (red). The rG4 region is boxed. (D) Close up of the arc-plot secondary structure of the IRES minimal region from nucleotide positions 313 to 390 (E, F). Most stable secondary structure of the minimal IRES region of the (E) WT (G4ss) sequence, and (F) rG4mut sequence. The color of the nucleotide represents its normalised SHAPE reactivity: black non-reactive; yellow, reactive; and, red, highly reactive. (n = 2 for each of the two primers). Both the RBS and the Stem III region are boxed.