Romoting pseudoexon inclusion was experimentally verified by deleting this sequence in the pT-FGG-IVS6-320A.T plasmid. Transient transfection of the 25-bp-deleted construct (pT-FGG-M-del25) in HeLa cells resulted in a change in pseudoexon inclusion from 96 to 44 , as quantified by fluorescent RT-PCR (Figure 4A). The marked reduction in pseudoexon inclusion confirmed that the deleted nucleotides are necessary to reach full efficiency in pseudoexon recognition. Similar results were obtained by qRT-PCR analysis (see Figure S2). To confirm that hnRNP F acts by interacting with the 25-bp region, hnRNP F silencing was performed in cells expressing the pT-FGG-M-del25 plasmid. In contrast with what observed in the presence of the whole pseudoexon sequence (see Figure 2A), silencing of hnRNP F in the absence of the 25-bp Ebselen price region significantly promoted pseudoexon inclusion (Figure 4B). This result suggests that: 1) the role of hnRNP F in enhancing pseudoexon recognition is strictly dependent on the 
presence of the 25-bp region; 2) the two G-run motifs located outside this region may act as ESSs. Since the predicted hnRNP F binding site within the 25-bp region is partially overlapped to a SRp40 binding site (Figure 3A) and that indeed a weak binding of SRp40 was evidenced bypulldown experiments (Figure 3B), we attempted to modulate SRp40 level in HeLa cells. While we could not reach a sufficient level of SRp40 silencing, overexpression experiments showed that, at least in our experimental conditions/system, the percentage of pseudoexon inclusion is insensitive to SRp40 upregulation (Figure S3).Different G-run Elements Exhibit Opposite Effects in Regulating Pseudoexon InclusionTo further dissect the functional elements within the splicingpromoting 25-bp region, as well as to map all hnRNP F binding sites within the pseudoexon sequence at a higher resolution, we decided to test the effect of the single (G1 and G2) and combined (G1+G2, G1+G3, G2+G3, and G1+G2+G3) deletion of the three different G-runs in the pT-FGG-IVS6-320A.T plasmid. Moreover, as pseudoexon regulation might depend on the cellular context, transient transfections of the mutant constructs were performed also in human hepatoma HepG2 cells, which endogenously express fibrinogen and therefore represent a more physiological model system than HeLa. Experiments in HepG2 showed no physiological expression of transcripts including the FGG pseudoexon (data not shown), and a higher level of FGG Madrasin web wildtype splicing (23 ) in the presence of the IVS6-320A.T mutation, thus allowing a more accurate analysis of the effects of the different deletion constructs (Figure 5). In HepG2, the expression of the G2-deleted construct (pT-FGGM-delG2), lacking the only G-run element located within the 25bp region, resulted in a significant reduction in pseudoexon inclusion (from 77 to 68 ) (Figure 5), confirming our hypothesis that this hnRNP binding site functions as an ESE. Surprisingly, the ablation of G2 had no effect on splicing in HeLa cells, indicating a certain degree of cell type-specific responsiveness of this element. This discrepancy might be due either to differences in the basal level of expression of hnRNP F between the two analyzed cell lines, or to an additional trans-acting factor only present in HepG2. The first possibility was explored by real-time RT-PCRG-runs Regulating FGG Pseudoexon InclusionFigure 2. Effect of hnRNP H and F modulation on the regulation of FGG pseudoexon splicing. (A) Knoc.Romoting pseudoexon inclusion was experimentally verified by deleting this sequence in the pT-FGG-IVS6-320A.T plasmid. Transient transfection of the 25-bp-deleted construct (pT-FGG-M-del25) in HeLa cells resulted in a change in pseudoexon inclusion from 96 to 44 , as quantified by fluorescent RT-PCR (Figure 4A). The marked reduction in pseudoexon inclusion confirmed that the deleted nucleotides are necessary to reach full efficiency in pseudoexon recognition. Similar results were obtained by qRT-PCR analysis (see Figure S2). To confirm that hnRNP F acts by interacting with the 25-bp region, hnRNP F silencing was performed in cells expressing the pT-FGG-M-del25 plasmid. In contrast with what observed in the presence of the whole pseudoexon sequence (see Figure 2A), silencing of hnRNP F in the absence of the 25-bp region significantly promoted pseudoexon inclusion (Figure 4B). This result suggests that: 1) the role of hnRNP F in enhancing pseudoexon recognition is strictly dependent on the presence of the 25-bp region; 2) the two G-run motifs located outside this region may act as ESSs. Since the predicted hnRNP F binding site within the 25-bp region is partially overlapped to a SRp40 binding site (Figure 3A) and that indeed a weak binding of SRp40 was evidenced bypulldown experiments (Figure 3B), we attempted to modulate SRp40 level in HeLa cells. While we could not reach a sufficient level of SRp40 silencing, overexpression experiments showed that, at least in our experimental conditions/system, the percentage of pseudoexon inclusion is insensitive to SRp40 upregulation (Figure S3).Different G-run Elements Exhibit Opposite Effects in Regulating Pseudoexon InclusionTo further dissect the functional elements within the splicingpromoting 25-bp region, as well as to map all hnRNP F binding sites within the pseudoexon sequence at a higher resolution, we decided to test the effect of the single (G1 and G2) and combined (G1+G2, 
G1+G3, G2+G3, and G1+G2+G3) deletion of the three different G-runs in the pT-FGG-IVS6-320A.T plasmid. Moreover, as pseudoexon regulation might depend on the cellular context, transient transfections of the mutant constructs were performed also in human hepatoma HepG2 cells, which endogenously express fibrinogen and therefore represent a more physiological model system than HeLa. Experiments in HepG2 showed no physiological expression of transcripts including the FGG pseudoexon (data not shown), and a higher level of FGG wildtype splicing (23 ) in the presence of the IVS6-320A.T mutation, thus allowing a more accurate analysis of the effects of the different deletion constructs (Figure 5). In HepG2, the expression of the G2-deleted construct (pT-FGGM-delG2), lacking the only G-run element located within the 25bp region, resulted in a significant reduction in pseudoexon inclusion (from 77 to 68 ) (Figure 5), confirming our hypothesis that this hnRNP binding site functions as an ESE. Surprisingly, the ablation of G2 had no effect on splicing in HeLa cells, indicating a certain degree of cell type-specific responsiveness of this element. This discrepancy might be due either to differences in the basal level of expression of hnRNP F between the two analyzed cell lines, or to an additional trans-acting factor only present in HepG2. The first possibility was explored by real-time RT-PCRG-runs Regulating FGG Pseudoexon InclusionFigure 2. Effect of hnRNP H and F modulation on the regulation of FGG pseudoexon splicing. (A) Knoc.
