Poliovirus 2A Protease Triggers a Selective Nucleo-Cytoplasmic Redistribution of Splicing Factors to Regulate Alternative Pre-mRNA Splicing

Poliovirus protease 2A (2Apro) obstructs host gene expression by reprogramming transcriptional and post-transcriptional regulatory events during infection. Here we demonstrate that expression of 2Apro induces a selective nucleo-cytoplasm translocation of several important RNA binding proteins and splicing factors. Subcellular fractionation studies, together with immunofluorescence microscopy revealed an asymmetric distribution of HuR and TIA1/TIAR in 2Apro expressing cells, which modulates splicing of the human Fas exon 6. Consistent with this result, knockdown of HuR or overexpression of TIA1/TIAR, leads to Fas exon 6 inclusion in 2Apro-expressing cells. Therefore, poliovirus 2Apro can target alternative pre-mRNA splicing by regulating protein shuttling between the nucleus and the cytoplasm.


Introduction
Poliovirus (PV), a member of the Enterovirus genus in the Picornaviridae family, contains a single-stranded RNA genome encoding a large polyprotein, which is processed by the viral proteases, 2A pro and 3C pro [1]. These proteases can also cleave host factors and are engaged in the inhibition of host gene expression and also the dismantling of cellular antiviral responses triggered during infection [2]. This can be illustrated by the proteolytic hydrolysis of eukaryotic translation initiation factors 4G (eIF4GI) and eIF4GII, and the poly(A)-binding protein (PABP), which leads to a down-regulation of host cell translation during PV infection [2,3]. Furthermore, both 2A pro and 3C pro also target various host nuclear proteins, including several transcription factors, which results in an inhibition of cellular RNA synthesis [4]. In addition, during PV infection the specific cleavage of Gemin-3 by 2A pro results in decreased assembly of small nuclear ribonucleoprotein (snRNP) U1 [5]. Using transfected cells and in vitro splicing assays, we recently reported that 2A pro activity controls alternative splicing of pre-mRNAs [6]. However, the mechanism by which 2A pro could modulate splicing events is unknown. PV 2A pro has been shown to target the nuclear pore structure by cleaving several nucleoporins (Nups), including Nup62, Nup98, and Nup153 [7,8,9,10]. This event disrupts protein and RNA trafficking between the nucleus and cytoplasm and, as a consequence, proteins normally retained in the nucleus redistribute to the cytoplasm in infected cells, where they are available to participate in viral protein synthesis and RNA replication [11,12]. This nuclear to cytoplasm translocation may also impact host gene expression by modulating post-transcrip-tional regulatory events involving pre-mRNA splicing, mRNA transport and mRNA stability and/or translation. We report here that 2A pro -mediated nucleo-cytoplasmic distribution of splicing factors is a potential mechanism to regulate alternative splicing.

Cell Cultures
Huh7-T7 cells were supplied by R. Bartenschlager (University of Heidelberg, Germany) and grown as described [13].

Results and Discussion
Alternative Splicing of Fas Exon 6 is Regulated by PV 2A pro Our previous data showed that over expression of 2A pro in HeLa cells abolished both constitutive and alternative splicing of Fas and FGFR2 pre-mRNAs, and promoted the exclusion of Fas exon 6 in a time and dose-dependent manner [6]. To begin to explore this mechanism we first assayed the integrity of nuclear factors involved in mRNA splicing after expression of 2A pro . Consistent with results using different cell lines [5,8], we found that plasmid-expressed 2A pro cleaved eIF4GI and eIF4GII proteins completely, and Gemin-3 partially, in Huh-7 cells ( Figure 1A, lane 2). These results were recapitulated with 2A pro expressed using a poliovirus replicon (PV-Rep) ( Figure 1A, lane 5). As expected, expression of a 2A pro point mutant (2AM) lacking proteolytic activity, or wild type 3C pro (3C), failed to cleave these host factors ( Figure 1A, compare lane 2 with lanes 3 and 4), [8]. Importantly, additional RNA-binding proteins, U2AF65, U2AF35, TIA1, TIAR and HuR, which can function as general and auxiliary splicing factors, were not cleaved by 2A pro (Figure 1A). These findings collectively illustrate the selectivity of 2A pro proteolytic activity in Huh7-T7 cells.
To study the relationship between 2A pro activity and alternative splicing we chose the well-characterized model of Fas exon 6 splicing, in which exon 6 is differentially included or skipped depending on the repertoire of RNA-binding proteins (RBPs) present [15,16]. This splicing event is biologically significant because it can regulate the sensitivity of Fas-expressing cells to Fasinduced apoptosis. RT-PCR analysis of co-transfected Fas minigene expression demonstrated an increased frequency of exon 6 skipping, upon 2A pro expression in Huh7-T7 cells (plasmid and PV replicon, Figure 1B and 1C), which was consistent with previous results [6]. As before, expression of 2AM and 3C failed to modify Fas exon 6 minigene splicing ( Figure 1B and 1C). The expression pattern of ß-actin, used as a control, was not affected. These findings suggest that 2A pro promotes changes in the alternative splicing of Fas receptor. U-rich Sequence on Exon 6 (URE6) is Essential for Fas Exon 6 Skipping in PV 2A pro -expressing Cells In order to define the regulatory elements necessary for 2A prodependent splicing alterations, we next utilized mutated Fas minigene reporters in transient transfection assays (Figure 2A) [16]. In agreement with these studies, mutation U-20C (Fas with a uridine to cytidine substitution 20 nucleotides upstream from the 39 splice site of intron 5) resulted in a natural skipping of exon 6, and no additive effect was observed after expression of 2A pro ( Figure 2B, lanes 1-2 vs lanes 4-5 and Figure 2D). Interestingly, variant m0, with replacement of a uridine-rich sequence on exon 6 (URE6), was refractory to 2A pro , and exon 6 was prominently included in all cases ( Figure 2B, compare lanes 1-6 with lanes 7-8 and Figure 2D). This result indicated that the URE6 sequence was important for exclusion of exon 6 in 2A pro -expressing cells.
Furthermore, when sites for the RNA-binding proteins TIA1/ TIAR in the U-rich sequence on Fas intron 6 (URI6) were replaced by an unrelated sequence, increased skipping of exon 6 was observed in mock transfected cells which was enhanced by 2A pro expression ( Figure 2B, compare lanes 10 and 12 with 11 and Figure 2D). These findings indicate that TIA1/TIAR binding regions were important for exon 6 definition [16,17], and substitution of these sequences did not suffice to prevent 2A promediated splicing defects. However, the mutation U1C in the minigene (positions 22 and 23 from 59 splice site of exon were mutated to A and C, respectively, and positions 7 and 8 of intron 6 were mutated to A and T, respectively) promoted inclusion of exon 6, and this was refractory to 2A pro activity ( Figure 2C, lanes 1-3 vs 4-6 and Figure 2D). Similarly, splicing of the minigene combining both mURI6 and U1C mutations was insensitive to 2A pro , suggesting that the 2A-dependent splicing effects can be counter-  Figure 2D). Taken together, these findings indicate a strong dependence of 2A splicing regulation on the U-rich sequence of Fas exon 6 (URE6), and the strength of the 59 splice site of Fas intron 6.

PV 2A pro Selectively Alters the Subcellular Distribution of Splicing Factors
The URE6 sequence of human Fas pre-mRNA is a significant cis-acting regulating element of Fas alternative splicing, which recruits trans-acting factors such as PTB/hnRNPI [16] and/or HuR/ELAVL1/HuA [18] to modulate exon 6 skipping/inclusion. As these splicing factors were not cleaved by 2A pro in our experimental system ( Figure 3A), we thought it possible that 2A pro activity may selectively alter the subcellular distribution/localization of these factors. To test this possibility, nuclear and cytoplasmic fractions were obtained from mock-transfected cells, or cells expressing 2A pro , or 2AM. Interestingly, immunoblot analysis showed that the regulators of Fas splicing: TIA1, TIAR and PTB together with the essential splicing factors U2AF35 and U2AF65 were all partially redistributed to the cytoplasm in cells expressing 2A pro , whereas HuR and hnRNPA1 proteins were retained largely in the nucleus under these conditions ( Figure 3A). Analysis of the cytoplasm/nuclear ratio confirmed this significant redistribution of the affected proteins ( Figure 3B). Similar results were obtained when the PV-rep was used ( Figure 3A, B), whereas expression of the 2AM point mutant failed to promote redistribution of the splicing factors ( Figure 3A). As a complimentary approach, we next assessed the subcellular localization of RBPs by confocal microscopy. As antibodies to the 2A pro work poorly for immunofluorescence we first tagged the proteases with an HAepitope. Immunoblot analysis showed that HA-tagged 2A pro (2A-HA) and HA-tagged 2AM (2AM-HA) had similar expression levels to the untagged proteins in transfected cells, and proteolytic activity against eIF4GI and eIF4GII substrates was also comparable ( Figure S1A). As expected, confocal analysis revealed that both 2A-HA and 2AM-HA exhibited a similar localization in the nucleus and cytoplasm of transfected cells ( Figure S1B). In agreement with the immunoblot analysis ( Figure 3A), inspection of confocal images revealed a redistribution of TIA1 ( Figure 3C) and TIAR ( Figure 3D), but not HuR ( Figure 3E), to the cytoplasm in cells transfected with 2A-HA, which was significantly less apparent in 2AM-HA-expressing cells. Similar expected results were noted for U2AF65, U2AF35 and hnRNP-A1 ( Figure S2A-C) when compared to the localization marker protein Ref1 ( Figure  S2D). Collectively, these results suggest that the protease activity of 2A pro can specifically promote the nucleo-cytoplasmic shuttling of several splicing factors and RNA-binding proteins to tailor a

Overexpression of TIA1 and TIAR Promotes Fas Exon 6 Inclusion in PV 2A pro -expressing Cells
It is known that the HuR protein serves as a repressor, which promotes Fas exon 6 exclusion [17,18]. Given this information, the above results indicate that a nuclear imbalance of HuR protein with respect to TIA1 and TIAR protein levels might mediate Fas  To explore this idea, we applied the experimental strategy outlined ( Figure 4A) to overexpress tagged MS2-TIA1 and MS2-TIAR proteins in Huh7-T7 cells. Interestingly, both proteins showed a predominantly nuclear localization, even in 2A-expressing cells ( Figure 4B), indicating that the nuclear localization signal present in these constructs is enough to retain these proteins in the nucleus. Under these conditions, an increase of Fas exon 6 inclusion was observed in cells expressing MS2-TIA1 or MS2-TIAR, compared with cells expressing control MS2BP ( Figure 4C, compare lanes 4-6 and 7-9, respectively, with 1-3), in a reproducible manner ( Figure 4D) reinforcing the idea that TIA/HuR imbalance modulates Fas alternative splicing in 2A pro -expressing cells.

Depletion of HuR Promotes Fas Exon 6 Inclusion in PV 2A pro -expressing Cells
To further test the hypothesis that an imbalance between HuR and TIA1/R, induced by 2A pro , could modulate Fas alternative splicing we next performed the reciprocal experiment (outlined in Figure 5A) to generate a loss-or gain-of-function of HuR. HuR is an RBP predominantly localized in the nucleus of Huh7-T7 cells (.90%), with nucleolar exclusion; whereas TIA1 and TIAR show a characteristic nuclear-cytoplasmic localization pattern ( Figure 3C-E). Nevertheless, HuR and endogenous TIA proteins are antagonistic in their actions related to the alternative splicing of Fas exon 6 [15,16]. Compared to control siRNA cells, knockdown of HuR in Fas minigene co-transfected Huh7-T7 cells using an siRNA targeted to the 39UTR, resulted in an 80-90% reduction in steady-state levels of HuR expression ( Figure 5B) and, at the same time, led to increased levels of Fas exon 6 inclusion ( Figure 5C, lane 1-3 vs lane 4-6) in a reproducible manner ( Figure 5D). To exclude off-target effects of the siRNA, a GFP-tagged HuR expression plasmid, which was resistant to the 39UTR-specific siRNA (Izquierdo, 2008;Izquierdo, 2010) was used under the same experimental conditions. [17,18]. GFP-HuR was located mainly in the nucleus, even in 2A-expressing cells, whereas GFP showed a combined nucleo-cytoplasmic distribution ( Figure S3A, B). As anticipated, expression of GFP-HuR rescued the exon exclusion phenotype despite knockdown of endogenous HuR ( Figure 5C, compare lanes 1-6 with 7-9 and Figure 5D). Given that PV 2A also relocalizes PTB, another repressor of Fas exon 6 inclusion [16], to the cytoplasm ( Figure 3A) these results strongly suggest that the HuR retained in the nucleus can promote Fas exon 6 skipping in 2A pro -expressing cells. Overall, our results support the notion that the selective re-localization of splicing factors from the nucleus to the cytoplasm in 2A-expressing cells, can account for the effect of 2A protease on human Fas exon 6 splicing.
PV 2A pro Promotes the Nucleo-cytoplasmic Redistribution of Splicing Factors and RNA-binding Proteins: Implications for PV life-cycle Our present work addresses the molecular mechanisms underlying the activity of 2A pro and its regulation of Fas exon 6 splicing. This is the first report to demonstrate that this splicing event could be regulated by an asymmetric distribution of TIA and HuR proteins between the nuclear and cytoplasmic compartments ( Figure 6). Consequently, control of alternative splicing in host cells could emerge as an important target to alter gene expression of infected cells. Indeed, an appreciation of splicing regulation may be important to understand how viruses evade the host antiviral response. Previous work [7,8,9,10] has suggested that the selective cleavage of Nups could facilitate the redistribution of RNAbinding proteins between the nucleus and cytoplasm upon 2A pro expression. In this sense, PV 2A pro can direct the relocalization of the cellular splicing factor SRp20 to the cytoplasm where it interacts with poly(C)-binding protein-2 and stimulates PV translation [11,12]. Our present observations support the idea that PV 2A pro specifically remodel the nucleo-cytoplasmic distribution of a number of RNA-binding proteins which are involved in post-transcriptional regulatory events. The question of why TIA1 and TIAR relocalize to the cytoplasm, whereas HuR remains in the nucleus in cells expressing PV 2A pro could be explained by the differences in the nuclear localization signals harbored by these proteins as well as the distinct nuclear import/ export pathways followed by TIA1/TIAR and HuR [20]. In this sense, it has been described that import/export of TIA1 and TIAR to/from the nucleus significantly differ from that documented for HuR [20,21]. However, further investigations are required to know the mechanism by which PV 2A pro modulates nucleo-cytoplasmic trafficking, which could potentially give insights into differential pathways of protein shuttling between nucleus and cytoplasm. Supporting Information Figure S1 Expression of HA-tagged PV 2A pro proteins in Huh7-T7 cells. (A) Huh7-T7 cells were transfected with pTM1-2A and pTM1-2A-HA. As controls, pTM1-2AM, pTM1-2AM-HA or empty plasmid were transfected also. At 4 hpt, samples were analyzed by Western blotting with antibodies raised to different host proteins as indicated to the right. Molecular mass markers (kDa) for protein are indicated to the left. (B) Distribution of HA-tagged PV 2A pro in Huh-7-T7 cells. Cells were transfected with pTM1-2A-HA or pTM1-2AM-HA or with an empty plasmid as a control. At 4 hpt, cells were fixed and indirect immunofluorescence was carried out using an antibody to HA. Samples were visualized with a confocal microscope. Merge shows the simultaneous a-HA, and Topro-3 to label the nucleus. Scale bar: 10 mm. (TIFF) Figure S2 Subcellular localization of splicing factors. Huh-7-T7 cells were transfected with pTM1-2A-HA or pTM1-2AM-HA or with an empty plasmid as a control. At 4 hpt, cells were fixed and indirect immunofluorescence was carried out using antibodies to U2AF65 (A), U2AF35 (B), hnRNPA1 (C) or Ref1 (D). Localization of 2A-HA and 2AM-HA were analysed as described above. All samples were visualized with a confocal microscope. Merge shows the simultaneous visualization of the splicing factor, a-HA, and Topro-3 to label de nucleus. Scale bars: 10 mm.
(TIFF) Figure S3 Subcellular localization of GFP-HuR. Huh7-T7 cells were transfected with a control or HuR siRNA. At 24 hpt the cells were transfected with plasmids expressing GFP-HuR (A) or GFP (B). The next day cells were co-transfected with the Fas minigene and pTM1-2A. Mutant 2AM cells transfected with the empty pTM1 plasmid were used as controls. At 3 hpt from the last transfection cells were fixed and the localization of GPF-HuR (A) or GFP(B) was analysed by confocal microscopy. Localization of TIA-1 was also analysed using a specific antibody. Merge shows the simultaneous visualization of the GPF/GFP-HuR, a-TIA, and Topro-3 to label de nucleus. Scale bars: 10 mm. (TIF)