The double-stranded RNA-binding protein, Staufen1, is an IRES-transacting factor regulating HIV-1 cap-independent translation initiation

Abstract Translation initiation of the viral genomic mRNA (vRNA) of human immunodeficiency virus-type 1 (HIV-1) can be mediated by a cap- or an internal ribosome entry site (IRES)-dependent mechanism. A previous report shows that Staufen1, a cellular double-stranded (ds) RNA-binding protein (RBP), binds to the 5’untranslated region (5′UTR) of the HIV-1 vRNA and promotes its cap-dependent translation. In this study, we now evaluate the role of Staufen1 as an HIV-1 IRES-transacting factor (ITAF). We first confirm that Staufen1 associates with both the HIV-1 vRNA and the Gag protein during HIV-1 replication. We found that in HIV-1-expressing cells, siRNA-mediated depletion of Staufen1 reduces HIV-1 vRNA translation. Using dual-luciferase bicistronic mRNAs, we show that the siRNA-mediated depletion and cDNA-mediated overexpression of Staufen1 acutely regulates HIV-1 IRES activity. Furthermore, we show that Staufen1-vRNA interaction is required for the enhancement of HIV-1 IRES activity. Interestingly, we find that only Staufen1 harboring an intact dsRNA-binding domain 3 (dsRBD3) rescues HIV-1 IRES activity in Staufen1 CRISPR-Cas9 gene edited cells. Finally, we show that the expression of Staufen1-dsRBD3 alone enhances HIV-1 IRES activity. This study provides evidence of a novel role for Staufen1 as an ITAF promoting HIV-1 vRNA IRES activity.


Luciferase assays
The activities of Firefly luciferase (FLuc) and Renilla luciferase (RLuc) were measured using the DLR® Assay System (#E1960, Promega Corporation) according to the manufacturer's instructions, on 10 l of cell lysates using a Sirius Single Tube Luminometer (Berthold Detection Systems GmbH). Data are expressed as a percentage of the Relative Luciferase Activity (RLA) or as Relative Translation Activity (RTA); the latter corresponding to the FLuc/RLuc ratio, an index of the IRES activity (16,18,24,(61)(62)(63).

Cell viability and proliferation assays
The cell viability assay was performed using the CellTiter 96® Aqueous One Solution Cell Proliferation Assay (MTS) (#G358A, Promega Corporation) according to the manufacturer's instructions. Briefly, HEK 293T cells were seeded at 1.5 × 10 3 cells per well in a 96-well plate and transfected with the indicated concentrations of sc/siRNAs or transfected with the indicated amounts of Stau1 55 -HA 3 expressing plasmid. 24 or 48 h later (as indicated in the text), the CellTiter 96 ® Aqueous One Solution Cell Proliferation Assay was added, incubated at 37ºC for 2 h, and the absorbance was measured at 495nm in a Biochrom EZ Read 400 microplate reader (Biochrom, Holliston, MA, USA). Alternatively, cell viability and cell proliferation were analyzed by flow cytometry (FC) (67). HEK 293T, HCT116 or SKO cells were seeded in 24 well-plate and transfected with the Stau1 55 -HA 3 (200 or 650 ng) expressing plasmid. After 48h of incubation, cells were incubated in human Fc block (#564220 BD Biosciences, San Jose, CA, USA) and Fixed Viability Stain 510 (#564406, BD Biosciences) for 15 min at room temperature, followed by an incubation of 20 min in fixation and permeabilization solution (#554722, BD Biosciences), and incubated with the BUV395 mouse anti-Ki67 antibody (#564071, BD Biosciences) for 30 min at room temperature. The single stain control was prepared using Compensation Beads (#552843, BD Biosciences). An average of 20,000 cells stained for FC was acquired using a BD LSRFortessa (BD Biosciences) at the Flow Cytometry Core Facility, Lady Davis Research Institute. Data were analyzed using FlowJo software (Tree Star).

Fluorescence in situ hybridization, immunofluorescence, and microscopy
Fluorescence in situ hybridization (FISH) and immunofluorescence (IF) analyses on HeLa cells were performed as described previously (36,70). To fix cells onto cover glasses, cells were washed once in D-PBS (Wisent) and fixed with 4% paraformaldehyde for 20 min. Highly Cross-Adsorbed, Alexa Fluor ® 647 #A-31571, by Invitrogen-Thermo Fisher Scientific) were applied at 1:500 for 1 h at 37 • C. Coverslips were washed for 20 min in PBS before being mounted on glass slides using ProLong Gold Antifade Reagent (Life Technologies). Microscopy was performed on a Zeiss LSM5 Pascal laser-scanning confocal microscope (Carl-Zeiss) equipped with a 63× (1.4 numerical aperture, oil immersion) plan-apochromat objective. Scanning was performed at 1024 × 1024 pixel resolution using a multitrack laser scanning protocol. Image files into Imaris software v. 9.7 for generation of colocalization channels. For IF analysis performed in Figure 6, cells were fixed, permeabilized, and blocked as described above, and then incubated with primary antibodies (sheep anti-GFP, Novus Biologicals, #NB100-62622; mouse anti-

Statistical analysis
Graphics and statistical analysis were carried out using the Prism v6.1 or v9.2.0 software (GraphPad), performing the statistical test indicated in the text and figure legends.
Next, we sought to determine if Staufen1 participates in Gag protein synthesis. For this, HEK 293T cell lines were transfected with the pNL-4.3-RLuc DNA ( Figure 1C, upper panel), which has a hemagglutinin (HA)-tagged Renilla luciferase (RLuc-HA) reporter gene inserted in frame with the Gag-protein start codon, generating a Gag-RLuc-HA fusion protein (57). This plasmid allows a direct evaluation of Gag synthesis from the HIV-1 vRNA, using the Gag-RLuc-HA reporter and its luciferase activity as a readout (25,57,69,73). HEK 293T cells were cotransfected with the pNL-4.3-RLuc plasmid and a short interfering RNA (siRNA) si55/63 (50 nM) targeting Staufen1 mRNAs (31,32) or a non-related scrambled control siRNA (scRNA; 50 nM). In agreement with earlier reports (31,33,55,74,75), when using the anti-Staufen1 antibody three bands were observed ( Figure 1C), one band for the 63 kDa isoform and a doublet, associated with an upstream AUG in the N-terminal end of Stau1 55 that enables alternative translation initiation (38), for the 55 kDa isoform. As expected (31,32), treatment of cells with si55/63 reduced the expression of endogenous Staufen1 ( Figure 1C, lower panel). A reduction in the Gag-RLuc-HA expression levels was detected in cells treated with the si55/63 by either using an antibody against the matrix (p17) region of Gag or the HA-tag ( Figure 1C, lower panel). Cytoplasmic RNA was extracted from pNL-4.3-RLuc HIV-1, scRNA, or si55/63 cotransfected cells and used as a template for quantitative analysis of the pNL-4.3RLuc RNA by an RT-qPCR. The relative pNL-4.3-RLuc mRNA content was reduced (∼30% reduction), however, not significantly in si55/63 treated cells, suggesting that the resulting decrease in Gag-RLuc-HA levels ( Figure 1C) was not exclusively associated with a reduction in the abundance of HIV-1 vRNA ( Figure 1D). We also determined the luciferase activity of the fusion protein. Consistent with a decrease in protein levels ( Figure 1C), a significant (P < 0.05) decrease in RLuc activity (∼53% reduction) in cells treated with the si55/63 was also observed (Figure 1E). These results confirm previous reports (31)(32)(33)35) and show that Staufen1 contributes to HIV-1 vRNA translation.

Staufen1 participates in HIV-1 IRES-mediated translation in cells
Stau1 55 colocalizes with the HIV-1 vRNA in the cytoplasm and plays a role in its translation ((31-33,35) and Figure  1). However, Gag-RLuc-HA fusion protein expression from the pNL-4.3-RLuc HIV-1 clone (Figure 1) does not allow the discrimination between cap-dependent or IRESmediated translation initiation (8,9). Knowing that Stau1 55 regulates cap-dependent translation (33) (16) and Figure 2A). The highly structured EMCV sequence, deficient in IRES activity, impedes ribosome reinitiation and readthrough (16,18,76). This well-characterized dl-RNA has proven to be a useful molecular tool to study HIV-1 IRES function (16,18,19,22,(24)(25)(26)(27)(28)69). First, we examined whether the reduction of endogenous Staufen1 influenced HIV-1 IRESmediated translation in cells. HEK 293T cells were chosen for this experiment because they are readily transfectable, and have been used as a model system to study HIV-1 replication (77), HIV-1 IRES activity (69), and the impact of Staufen1 on HIV-1 replication (32,35,72). Thus, HEK 293T cells were cotransfected with the dl HIV-1 IRES plasmid and a scRNA control, or different concentrations of the si55/63 (31,32). Treatment of cells with the si55/63 (50 nM) led to a marked reduction in Staufen1 ( Figure 2B) with no associated effect on cell viability ( Figure 2C). Luciferase activities were measured, and data were expressed as relative luciferase activity (RLA). The expression of RLuc and FLuc obtained from cells transfected with the scRNA(−) was set to 100% ( Figure 2D). A significant decrease (∼60%) of HIV-1 IRES activity (FLuc) with no impact on capdependent translation (RLuc) was evidenced in cells transfected with the dl HIV-1 IRES plasmid treated with the si55/63 (50 nM) ( Figure 2D). As Staufen1 knockdown did not significantly alter RLuc activity, the observed reduction in FLuc activity cannot be attributed to reduced stability of the dl HIV-1 IRES mRNA ( Figure 2D). Analysis of the FLuc/RLuc ratio (relative translational activity, RTA), as an index of IRES activity (16,18,24), confirmed the significant reduction (∼60%) in HIV-1 IRES activity when the endogenous Staufen1 protein was depleted ( Figure 2E), recapitulating what was observed with the pNL-4.3-RLuc mRNA ( Figure 1B-D). Thus, this observation confirms that Staufen1 contributes to HIV-1 IRES-mediated translation in HEK 293T cells Next, we sought to evaluate the impact of Stau1 55 on HIV-1 IRES activity. For this, HEK 293T cells were cotransfected with the dlHIV-1 IRES plasmid, an irrelevant DNA (negative control), or different concentrations (50-650 ng) of plasmid Stau1 55 -HA 3 , encoding for a hemagglutinin (HA 3 )-tagged Stau1 55 (47). The overexpression of Stau1 55 -HA 3 , confirmed by western blot using GAPDH as a loading control ( Figure 2F). As previously described, Stau1 55 -HA 3 migrated at the level of Stau1 63 when resolved by electrophoresis (33,55). The overexpression of Stau1 55 -HA 3 has been reported to impair cell viability and proliferation (46,55,74). However, in agreement with a previous report using asynchronous HEK 293T cells (55), in the time frame of the experiment (24-48 h), the overexpression Stau1 55 -HA 3 did not impair cell viability ( Figure 2G, H) or cell proliferation ( Figure 2I). Luciferase activities were measured, and data expressed as RLA, with the values of RLuc and FLuc obtained from cells transfected with the control DNA(−) set to 100% ( Figure 2J). At lower levels of Stau1 55 -HA 3 expression (50-100 ng of DNA), RLuc and FLuc increased equivalently ( Figure 2J), most probably due to RNA stabilization (33,48). Proteins involved in mRNA stability are known to associate with Stau1 55 (34). However, as the levels of recombinant Stau1 55 -HA 3 increased (200-650 ng of DNA), FLuc activity was enhanced (to a maximum of ∼5-fold) without having a further significant impact on the expression RLuc activity ( Figure 2J). Analysis of the FLuc/RLuc ratio (RTA), better illustrates the concentration-dependent increase and significantly higher activity of the HIV-1 IRES when the Stau1 55 -HA 3 protein is overexpressed (500 and 650 ng of DNA; Figure 2K). These results suggest that Stau1 55 promotes HIV-1 IRES activity.

Overexpression of Staufen1 does not induce alternative splicing nor cryptic promoter activity from the dl HIV-1 IRES reporter in cells
Stau1 55 interacts with splicing factors and is involved in pre-mRNA splicing (34,78). Therefore, Stau1 55 -HA 3 overexpression could generate a monocistronic mRNA encoding a functional FLuc protein from the dl HIV-1 IRES RNA by inducing an alternative splicing event (79). If so, the increase in FLuc activity ( Figure 2J) would not reflect HIV-1 IRES activity but would correspond to the expression of a cap-dependent monocistronic FLuc encoding mRNA. To explore this possibility, the Renilla ORF of the dl HIV-1 IRES mRNA was targeted with an RLuc-siRNA (siRLuc) ( Figure 3A, upper panel), as previously described (24,69). The rationale for this experiment considers that if a monocistronic transcript encoding an active FLuc enzyme is indeed generated from a cryptic promoter or through RNA splicing, using a short interfering RNA (siRNA) that targets the RLuc coding region should knockdown the bicistronic RNA without affecting the expression levels of a monocistronic FLuc transcript. The siRLuc RNA or a scRNA were cotransfected with the dl HIV-1 IRES in HEK 293T cells. The expression of Stau1 55 -HA 3 was confirmed by western blot using an anti-HA or anti-Staufen1 antibody and GAPDH as a loading control ( Figure 3A, lower panel). In the presence of the siRLuc RNA, both RLuc and FLuc activities were significantly reduced, whether Stau1 55 -HA 3 was overexpressed or not ( Figure 3B). When directly compared, the reduction of RLuc and FLuc activities induced by the siRLuc RNA in the presence, or in the absence, of overexpressed Stau1 55 -HA 3 protein was not statistically different ( Figure 3B). This observation indicated that RLuc and FLuc expression levels were associated with a single transcript targeted by the siRLuc RNA and confirmed that the overexpression of Stau1 55 does not induce the generation of FLuc expressing monocistronic transcripts in HEK 293T cells.
The dl HIV-1 IRES reporter plasmid displays cryptic promoter activity in HEK 293T cells (69). In addition, the FLuc reporter gene also exhibits cryptic promoter activity that is detectable in both yeast and mammalian cells (80). However, the identified promoter lies within the FLuc coding sequence generating short mRNAs that do not code for a functional luciferase enzyme (80). Based on these previous reports, we next sought to directly assess if the overexpression of Stau1 55 enhanced cryptic promoter activity present within the used dl-plasmid in HEK293T cells. For  this, cells were transfected with the dl HIV-1 IRES or the SV40 dl HIV-1 IRES plasmid in the presence or absence of the Stau1 55 -HA 3 expressing plasmid. The SV40 dl HIV-1 IRES plasmid lacks the SV40 promoter ( Figure 3C, upper panel) (19,24,69). The expression of the Stau1 55 -HA 3 was confirmed by western blot analysis using an anti-HA or anti-Staufen1 antibody and GAPDH as a loading control ( Figure 3C, lower panel). In the absence of the SV40 promoter ( SV40), both the RLuc and FLuc activities from the reporters were significantly (P < 0.05) diminished in cells ( Figure 3D). However, in agreement with a previous report (69), RLuc activity was detected in HEK 293T, suggesting a leakiness of about 19% existed ( Figure 3D). Our results also showed that FLuc activity was 9% higher than RLuc, confirming the presence of weak cryptic promoter activity in HEK 293T (69). The overexpression of Stau1 55 -HA 3 did not impact FLuc expression from the SV40 dl HIV-1 IRES plasmid, indicating that Stau1 55 -HA 3 does not enhance cryptic promoter activity from the used vector in HEK 293T cells ( Figure 3D).
Based on the above results, we conclude that overexpression of Stau1 55 -HA 3 does not induce the expression of a monocistronic mRNA encoding for an active FLuc enzyme by enhancing splicing (Figure 3A and B) or by increasing cryptic promoter activity ( Figure 3C and D) from the dlHIV-1 IRES reporter in HEK 293T cells. Thus, results in Figure 2 reflect the function of a genuine IRES and indicate that Stau1 55 promotes HIV-1 IRES activity in HEK 293T cells.

The TAR and poly(A) Stem loops of the HIV-1 5 UTR participate in Staufen1 stimulation of HIV-1 IRES activity in cells
Stau1 55 overexpression extends the G2 phase of the cell cycle (46,55,74). Also, Stau1 55 binds to the TAR stem-loop (SL) of the HIV-1 vRNA, enhancing cap-dependent translation initiation (33). As the HIV-1 IRES function is G2/M dependent (16,27), Stau1 55 could enhance HIV-1 IRES activity by binding the TAR or by extending the G2 phase of the cell cycle. Therefore, we determined the impact of  (16,18), translational activity driven by the full-length 5 UTR (1−336) was significantly higher than that obtained with the 5 deletion mutant IRES (104-336) in HEK 293T ( Figure 4B, right panel).

Staufen1−RNA interaction is required to promote HIV-1 IRES activity in HEK 293T cells
Stau1 55 contains four dsRNA-binding domains (dsRBD) consensus sequences, of which the dsRBD3 possesses a strong dsRNA-binding activity, while the dsRBD4 exhibits somewhat weaker binding capacity (38,47). Stau1 55 has two additional RBDs, RBD2 and RBD5, but neither bind RNA (38,47). The dsRBD3 was shown to primarily mediate the Stau1 55 -HIV-1 TAR RNA and Gag interactions (32,33). To further confirm whether vRNA binding by Stau1 55 is required to stimulate HIV-1 IRES activity, the dl HIV-1 IRES plasmid was cotransfected in HEK 293T cells together with different concentrations (50-650 ng) of a plasmid encoding for an HA 3 -tagged Stau1 55 -F135A mutant protein or a control DNA. Stau1 55 -F135A-HA 3 possesses a Phe-to-Ala mutation at position 135 in the dsRBD3 domain that leads to a loss in its dsRNA-binding ability (32,33,35,47). The expression of Stau1 55 -F135A-HA 3 protein was confirmed by western blot ( Figure 5A). The RLuc and FLuc activities were measured, and data were expressed as RLA ( Figure  5B) or RTA ( Figure 5C). The overexpression of Stau1 55 -F135A-HA 3 did not affect RLuc or FLuc activities of the dl HIV-1 IRES mRNA. Hence, the dsRBD3 mutant protein Stau1 55 -F135A-HA 3 was unable to promote HIV-1 IRES activity in HEK 293T cells ( Figure 5C). These observations suggest that the dsRBD3 of Stau1 55 is responsible for mediating the enhancement of HIV-1 IRES activity in HEK 293T cells. Additionally, these observations further indicate that Stau1 55 binding to the RNA is required for its function of stimulating the HIV-1 IRES. Stau1 55 binds the TAR-RNA at the SBS with high affinity (K d 3.5 nM) (33). The SBS is located in the upper stem between the bulge and the loop as determined by filter binding and Northwestern using purified Stau1 55  substitution mutants characterized in previous studies of Stau1 55 -TAR interaction ( Figure 5D) (33). HEK 293T cells were transfected with the dl HIV-1 IRES and the TARmut's harboring dl HIV-1 IRES plasmids. The RLuc and FLuc activities from all dl HIV-1 IRES plasmids were comparable in magnitude ( Figure 5E, upper and middle panel). Nonetheless, RLuc and FLuc activities from the dl HIV-1 IRES TAR-mut1 plasmid were significantly higher than the rest ( Figure 5E). However, the FLuc/RLuc ratio (RTA) analysis revealed no significant differences among the vectors, suggesting that IRES activity in HEK 293T cells was equivalent for all ( Figure 5E, lower panel). Next, the dl HIV-1 IRES and the TAR-mut plasmids were cotransfected in HEK 293T cells together with the Stau1 55 -HA 3 expressing plasmid (650 ng) or a control DNA. The overexpression of Stau1 55 -HA 3 protein was confirmed by western blot (Figure 5F). Luciferase activities were measured, and the RTA of the dl HIV-1 IRES plasmid cotransfected with the control DNA (−), not expressing Stau1 55 , was set to 100% (Figure 5G). In the absence of Stau1 55 -HA 3 , the IRES activity of the dl HIV-1 IRES and dl HIV-1 IRES TAR-mut plasmids was equivalent ( Figure 5G, black bars). As expected, we found that expression of Stau1 55 -HA 3 significantly increased (∼160%) HIV-1 IRES activity from the dl HIV-1 IRES reporter. IRES activity from dl HIV-1 IRES TAR-mut1, harboring a deletion in the apical RNA loop, was also significantly increased (∼77%), yet to lower levels than the no-mutated HIV-1 IRES, by the expression of Stau1 55 -HA 3 ( Figure 5G). This observation is consistent with northwestern assays showing that mutations in the apical loop reduce, but do not abolish, Stau1 55 binding to the TAR RNA (33). Confirming our previous results (Figures 4 and 5A−C), overexpression of Stau1 55 did not impact IRES activity of the dl HIV-1 IRES TAR-mut2, lacking the SBS, or the dl HIV-1 IRES TAR-mut3 lacking the bulge, the upper stem, and the loop ( Figure 5D and G). Together these findings confirm that Stau1 55 -vRNA interaction is required for Stau1 55 to stimulate the activity of the HIV-1 IRES.
Overexpression of Stau1 55 -HA 3 at the highest concentration tested (650 ng of plasmid) reduced HCT116 (by ∼22%) and SKO (by ∼16%) cell viability, but not at lower DNA quantities ( Figure 6A). Stau1 55 -HA 3 overexpression did not affect HCT116 or SKO cell proliferation at any concentration tested ( Figure 6B). Next, HCT116 or SKO cells were transfected with the dl HIV-1 IRES plasmid alone (control DNA) or in combination with plasmids (500 ng) encoding a yellow fluorescent protein (YFP)-tagged Stau1 55 (Stau1 55 -YFP) or Stau1 55 -F135A-YFP protein. The expression of Stau1 55 -YFP and Stau1 55 -F135A-YFP proteins was con-firmed by western blot using an anti-GFP antibody and ␤-actin as a loading control ( Figure 6C). Luciferase activities were measured, and results presented as RTA showed that the expression of Stau1 55 -YFP in HCT116 cells significantly enhanced (∼54% increase) HIV-1 IRES activity (Figure 6D). In contrast, the expression of the dsRNA-binding mutant Stau1 55 -F135A-YFP had no impact on the activity of the HIV-1 IRES ( Figure 6D). These results recapitulate those obtained in HEK 293T cells using HA-epitope tagged Stau1 55 (Figures 2 and 5). HIV-1 IRES was significantly attenuated (∼61%) but not abolished in Staufen1 knockout SKO cells compared to parental HCT116 cells ( Figure 6D), confirming that Stau1 55 plays a role in promoting IRES activity (Figures 4 and 5). In rescue experiments in SKO cells, the overexpression of Stau1 55 -YFP restored and even enhanced (∼90% increase) HIV-1 IRES activity compared to non-transfected SKO cells ( Figure 6D). The expression of Stau1 55 -F135A-YFP in SKO cells also increased HIV-1 IRES activity (∼40% increase) compared to non-transfected SKO cells ( Figure 6D). This observation suggested that in contrast to HEK 293T cells ( Figure 5A-C), in SKO cells, the dsRBD4 might also contribute to HIV-1 IRES activity ( Figure 6D, see below). Nonetheless, HIV-1 IRES activity remained reduced (∼21% lower) in SKO cells expressing Stau1 55 -F135A-YFP compared to HCT116 cells ( Figure 6D). These results show that Stau1 55 -YFP, but not the mutant Stau1 55 -F135A-YFP, fully rescues HIV-1 IRES activity in SKO cells, confirming the role of Stau1 55 as an ITAF of the HIV-1 IRES. The molecular behavior of dsRBD3, including its ability to interact with its targets RNAs and ribosomes, is maintained even when isolated from the rest of the protein (47,58,81,82). We queried whether the expression of Stau1 55 -dsRBD3 alone could impact HIV-1 IRES activity in SKO cells. For this, HCT116 or SKO cells were transfected with the dl HIV-1 IRES plasmid alone or together with plasmids expressing Stau1 55 -GFP-topaz, Stau1 55 -dsRBD3-GFP-topaz, or mutant Stau1 55 -4K-GFP-topaz harboring point mutations in the dsRBD3 and dsRBD4, abrogating both RNA binding domains (59). The overexpression of the Stau1 55 and the Stau1 55 -dsRBD3 domain was confirmed by western blot using an anti-GFP antibody and ␤-actin as a loading control ( Figure 6E). Luciferase activities were measured, and results are presented as RTA ( Figure 6F). As previously observed ( Figure 6D), in this new series of experiments, HIV-1 IRES activity was considerably reduced (∼61%) in SKO cells when compared to HCT116 cells ( Figure 6F). The expression of Stau1 55 -GFP-topaz significantly enhanced HIV-1 IRES activity in HC116 cells (∼40% increase), while the mutant Stau1 55 -4K-GFP-topaz did not affect HIV-1 IRES activity ( Figure  6F). In SKO cells, Stau1 55 -GFP-topaz, but not Stau1 55 -4K-GFP-topaz, restored HIV-1 IRES activity to levels comparable to that found in HCT116 cells ( Figure 6F). This observation confirmed that Stau1 55 binding to the vRNA is a requirement to promote HIV-1 IRES-mediated translation. Results obtained with Stau1 55 -F135A ( Figure 6D) and Stau1 55 -4K-GFP-topaz ( Figure 6F) confirm that the dsRBD4 also contributes to the ability of Stau1 55 to enhance HIV-1 IRES activity in SKO cells. Strikingly, when expressed alone, the Stau1 55 -dsRBD3 domain alone signif- icantly enhanced HIV-1 IRES activity (∼63% increase) in HCT116 cells while again restoring HIV-1 IRES activity in SKO cells ( Figure 6F). Together, these results indicate that Stau1 55 , but not Stau1 55 -F135A or Stau1 55 -4K, fully rescues HIV-1 IRES activity in SKO cells, confirming the role of Stau1 55 as an ITAF of the HIV-1 IRES. These results also suggest that the Stau1 55 -dsRBD3 domain alone is sufficient to restore HIV-1 IRES activity in SKO cells. We next sought to validate the impact of the dRBD3 domain of Stau1 55 on HIV-1 IRES activity in a different cell line and by using an immunofluorescence (IF)-based approach to evaluate protein expression. As HEK 293T cells have limited cytoplasmic space (37,38,59), they were not considered for the assay. HeLa cells were selected as their cytoplasm is suitable for IF analysis, and they support pNL-4.3 replication (Figure 1) and HIV-1 IRES activity (16,24,27,28). HeLa cells were transfected with the dl HIV-1 IRES DNA together with the pStau1 55 -GFP-topaz, mutant pStau1 55 -4K-GFP-topaz or pStau1 55 -dsRBD3-GFP-topaz expressing plasmids. In these assays, the pStau1 55 -4K-GFPtopaz, which does not bind RNA or promote HIV-1 IRES activity ((59) and Figure 6G), and the empty plasmid expressing only eGFP (pe-GFP), were used as a negative control. The expression of the wild-type and mutant Stau1 55 proteins and the dsRBD3 domain alone was confirmed by detecting GFP expression by IF in cells ( Figure 6G). RLuc and FLuc reporter proteins were also detected in HeLa cells by IF as indicated in Materials and Methods. As anticipated, both the cap-dependent RLuc and the HIV-1 IRESdependent FLuc reporter proteins could be readily detected by IF in HeLa cells transfected with the dl HIV-1 IRES plasmid ( Figure 6G). The co-expression of RLuc, FLuc, and pe-GFP, pStau1 55 -GFP-topaz, mutant pStau1 55 -4K-GFP-topaz or pStau1 55 -dsRBD3-GFP-topaz in cells was also confirmed by IF ( Figure 6G). The mean fluorescence intensity (MFI) values for RLuc and FLuc obtained from the imaging data were used to calculate the RTA, which were normalized to GFP expression to account for Stau1 55 and dsRBD3 recombinant protein levels. The Stau1 55 -4K mutant normalized RTA value was set to 100%. Consistent with our observations in HEK 293T, HCT116, and SKO, the overexpression of pStau1 55 -GFP-topaz enhanced HIV-1 IRES activity in HeLa cells ( Figure 6H). Also, confirming our findings in HCT116 and SKO cells ( Figure 6E and F), the expression of Stau1 55 -dsRBD3-GFP-topaz significantly stimulated HIV-1 IRES activity in HeLa cells ( Figure  6H). These observations confirm that the Stau1 55 -dsRBD3 alone is sufficient to promote HIV-1 IRES activity in cells, but the maximum stimulation in HeLa cells is obtained with the whole protein.
In cells, Staufen1 resides in several dynamic RNP complexes (34,40,41). In HIV-1-expressing cells, Staufen1 binds the vRNA and Gag forming different Staufen1-containing complexes (31)(32)(33)(34)(35)(36)52,72). We initiated this study by confirming that Staufen1 associates with the vRNA and Gag in HIV-1-expressing cells and validating that the protein's depletion reduced Gag protein synthesis (Figure 1). The reduction in Gag expression when Staufen1 was depleted using a siRNA strategy was not unexpected (31,33,53). Even so, by using a dl-RNA strategy, we show, for the first time, that the Stau1 55 isoform binds the vRNA and modulates HIV-1 IRES activity (Figures 2, 4-6). Strikingly, Stau1 55 emerges as a unique multifunctional translational regulator of HIV-1 vRNA gene expression, controlling both capand IRES-dependent translation initiation ( (31,33,53) and Figures 2 and 6). How Stau1 55 accomplishes the task of regulating two completely different mechanisms of translation initiation remains unknown. However, binding of Stau1 55 to the HIV-1 vRNA was required to establish control over both cap-and IRES-mediated translation initiation ( (33) and . For HIV-1 IRES activity, we show that Stau1 55 does not stimulate the activity of a 5 deletion mutant HIV-1 IRES (nts 104-336) (Figure 4), lacking the TAR and poly(A) stem-loop. Equally, Stau1 55 does not stimulate the activity of an HIV-1 IRES lacking the SBS ((33) and Figure 5). Thus, to stimulate HIV-1 IRES activity, the Stau1 55 -TAR (SBS) interaction must be attained. Consistent with this conclusion, the Stau1-F135A or Stau1-4K mutants that are restricted in their ability to bind RNA did not promote HIV-1 IRES activity (Figures 5 and 6). As previously indicated (45), the number of Stau1 55 molecules within a cell is fewer than that of ribosomes, suggesting that Stau1 55 would favor translation of only a subpopulation of specialized mRNAs it interacts with, including the HIV-1 vRNAs, characterized for having a SBS within their UTRs. Consistent with this proposed target specificity, Stau1 55 is known to interact with the HIV-1 vRNA uniquely but not with the spliced HIV-1 RNA species (32). The role of Stau1 55 in translation is complex and highly influenced by the position at which the protein is recruited onto the mRNA. When bound to the 3 UTR of its target cellular mRNA Staufen1 induces Staufen1-mediated mRNA decay (44). However, our data shows that when attached to the 5 UTR of the HIV-1 vRNA (Figures 2, 4-6) or mRNAs containing the HIV-1 TAR-SL (33), Staufen1 stimulates translation.
It is tempting to speculate that the function of Stau1 55 as a translational regulator is associated with the natural structural plasticity exhibited by the protein and its ability to induce local RNA structural modification due to its dsRNA binding activity (54,81,82). Also, the ability of Stau1 55 to regulate cap-and IRES-mediated translation initiation of the HIV-1 vRNA could be linked either to the Stau1 55 -binding partners present within the translating Stau1 55 -HIV-1 RNP or to the post-translational modifications (PTMs) of individual protein partners of the RNP, including Stau1 55 itself (25,31,34,69). This possibility is highly attractive as it would suggest that the Stau1 55 -HIV-1 RNP complex would be capable of sensing the environment adapting to favor either cap-or IRES-dependent translation initiation as previously proposed (12,13,20). This would allow viral gene expression to rapidly adapt to physiological changes induced either by the cellular antiviral response or by the viral replication cycle. Even though further studies are required to prove this possibility entirely, several reports already support the notion (12,13,20,22,91). For example, during HIV-1 infection, the viral protein R (Vpr) induces a cell cycle arrest in G2/M (13,92), a cell cycle stage known to compromise canonical cap-dependent translation initiation severely (18,91). Yet, in G2/M, transcription (93) and translation (13,16,27) of the HIV-1 vRNA are promoted. Staufen1 expression is also high in G2/M (55) and is enhanced when cap-dependent translation initiation is compromised (50). Furthermore, the expression of Staufen1 extends the G2 phase of the cell cycle (46,55,74). Therefore, it is expected that Stau1 55 is a key player and a constitutive part of the RNP-complex modulating HIV-1 IRES activity during G2/M. Another point to consider, but not directly evaluated in this study, is that Stau1 55 binds to ribosomes (37,45,47). Specifically, Stau1 55 interacts with the 60S ribosomal subunit (45). Thus, we cannot exclude the possibility that the function of Stau1 55 or the Stau1 55 -HIV-1 RNP complex is to stabilize ribosome assembly, polysome association, and ribosome occupancy of the HIV-1 vRNA (33,47,64), all processes independent from the mechanism by which the HIV-1 vRNA recruits the 40S ribosomal subunit. This last possibility could explain why Stau1 55 enhanced both capand IRES-mediated translation initiation as its function would be shared regardless of the mechanism of translation initiation.
As observed for Stau1 55 , overexpression of the Stau1 55 -dsRBD3 moiety alone was sufficient to promote HIV-1 IRES activity in cells expressing endogenous levels of Staufen1, HCT116, and HeLa, and rescue HIV-1 IRES activity in SKO cells ( Figure 6). These findings support the possibility that, at least in part, RNA structural rearrangement induced by the dsRBD3 binding could assist in stimulating IRES-mediated translation. Studies of the Drosophila dStau1-dsRBD3, which is homologous to the human Stau1 55 -dsRBD3, showed that dsRBD3 binds hair-pin RNA structures with double-helical stems with micromolar affinity (81). Binding of the dStau1-dsRBD3 to its target RNA induced a kink at the stem-loop junction, bending the RNA (81). It is plausible that this RNA structural change associated with dsRBD3 binding is responsible for enhancing HIV-1 IRES activity. Further experiments will be required to validate this last interpretation.
Despite the relevance of our findings, an apparent caveat of the study is that experiments were not conducted in Tcells, natural target cells for HIV-1. Nonetheless, previous reports conclude that the fundamental RNA-dependent mechanism-driving IRES function from the 5 UTR of the HIV-1 vRNA is similar between T-cells and other cell types, albeit with different activities (18). Furthermore, Stau1 55 is ubiquitously expressed in different cell types where it plays a direct role in various stages of the HIV-1 replication cycle in all cells it has been evaluated, including T-cells, HeLa cells and HEK 293T, amongst others (32,(34)(35)(36)(37)52,72). Thus, we expect our results that show that Stau1 55 enhances IRES-mediated translation initiation of the HIV-1 vRNA, to be valid in all cell lines and cell types capable of supporting HIV-1 IRES activity.
In conclusion, this study provides evidence that Stau1 55 acts as a genuine ITAF for the HIV-1 IRES, thereby providing a novel and additional function of this multifunctional RBP, understood to regulate several steps of HIV-1 replication. The results described herein show that Stau1 55 positively regulates HIV-1 vRNA cap-independent translation initiation.