Src Family Kinase Inhibitors Block Translation of Alphavirus Subgenomic mRNAs

Alphaviruses are arthropod-transmitted RNA viruses that can cause arthralgia, myalgia, and encephalitis in humans. Since the role of cellular kinases in alphavirus replication is unknown, we profiled kinetic changes in host kinase abundance and phosphorylation following chikungunya virus (CHIKV) infection of fibroblasts.

alphavirus replication should facilitate the development of effective antiviral therapies to treat individuals suffering from alphavirus infections.
Alphavirus virions are internalized into host cells by clathrin-mediated endocytosis (13,14), and the glycoproteins initiate viral fusion in endosomal compartments, releasing the viral genome into the cytoplasm. The nonstructural protein precursors P123 and P1234 are translated from the incoming capped and polyadenylated genomic RNA (15,16). The nonstructural proteins form the replication complex in spherules at the plasma membrane (17)(18)(19)(20)(21)(22). Minus-stranded RNA is synthesized by P123 and cleaved nsP4, while full-length positive-sense viral genomic RNA and the subgenomic mRNA (sgmRNA) are synthesized by cleaved nonstructural protein P1234 (23,24). The translated subgenomic polyprotein is cleaved by virus-and host-mediated proteolytic activity into the mature structural proteins capsid (C), E3, E2, 6K/TF, and E1 (25)(26)(27). Translation of the sgmRNA is efficient, even though CHIKV infection induces inactivating phosphorylation of the ␣ subunit of eukaryotic translation initiation factor 2 (eIF2␣) (28). In addition, translation of the SINV sgmRNA does not appear to require eIF4G (29) or the noncanonical translation initiation factors eIF2D and eIF2A (29,30). Translation of SINV sgmRNA is facilitated by a GC-rich structural element called the downstream loop (DLP), which allows for translation in the absence of eIF2 (28,31,32). However, CHIKV and VEEV have no predicted DLP structure (31), but rather, they may have an alternative mechanism for efficient sgmRNA translation.
Identifying the specific cellular kinases and signaling pathways that promote alphavirus replication could aid in the development of antialphavirus compounds. In order to identify kinase pathways important for alphavirus replication, we profiled changes in kinase abundance and/or activity following CHIKV infection. We then tested the antiviral activities of inhibitors of cellular pathways that were altered by infection, including the SFK-phosphatidylinositol 3-kinase (PI3K)-Akt-mTOR pathway. SFK inhibitors (PP2 and dasatinib) and mTORC1/2 inhibitors (PP242 and Torin 1) block alphavirus replication in human fibroblasts. While dasatinib and Torin 1 had minimal effects on viral RNA (vRNA) synthesis, each treatment blocked structural protein accumulation during CHIKV and VEEV replication. Furthermore, dasatinib decreased the amount of CHIKV RNA associated with polysomes, indicating that CHIKV relies on SFKs and mTORC1/2 for structural protein synthesis. Our novel findings provide insight into molecular processes that comprise the CHIKV replication cycle, including the requirement of SFK activity for alphavirus replication.

RESULTS
Kinome pathway analysis of CHIKV-infected fibroblasts. To identify kinase pathways that are kinetically altered following CHIKV infection, we utilized multiplexed kinase inhibitor beads coupled with quantitative mass spectrometry (MIB-MS) kinome profiling to quantify changes to the cellular kinome over a time course of infection (57,58). Kinome profiling uses broad-spectrum kinase inhibitors coupled to a matrix as an affinity reagent to capture kinases from cell lysates. The identity and abundance of the recovered kinases are then determined by quantitative mass spectrometry. Changes in the amounts of kinase recovered under different conditions reflect changes in the expression or activity of the kinase. As depicted in Fig. 1A, serum-starved normal human dermal fibroblasts (NHDFs) were infected with CHIKV 181/25 (multiplicity of infection [MOI] ϭ 3 PFU/cell), and lysates were collected at 4, 8, 12, and 24 h postinfection (hpi) for kinome profiling. Infection induced multiple changes to the kinome (Fig. 1B). Examples of kinases significantly altered after infection are shown in Fig. 1C. These kinases include the transforming growth factor ␤ (TGF-␤) receptor subunits and members of the ephrin family. Other kinases showed consistent trends but failed to achieve significance due to variability. A comparative analysis of kinome changes revealed that kinases clustered into three distinct groups by k-means clustering. Kinases in cluster 1 (118 kinases) remained stable across the course of infection. Levels of kinases in cluster 2 (32 kinases) were decreased by infection, and those of kinases in cluster 3 (17 kinases) were increased ( Fig. 1D; see also Table S2 in the supplemental material). These data suggest that CHIKV infection significantly alters the expression and/or activity of the cellular kinome. Changes in kinase recovery measured by kinome profiling can arise from changes in either kinase expression or kinase activity due to phosphorylation. To more specifically identify changes in phosphorylation, CHIKV-infected cell lysates were collected at 0, 1, 2, 4, 6, 8, and 24 hpi and analyzed via PathScan receptor tyrosine kinase (RTK) signaling antibody arrays spotted with RTK phosphoantibodies (Fig. S1). Increased phosphorylation of Akt, platelet-derived growth factor receptor (PDGFR), and extracellular signalregulated kinase 1/2 (ERK1/2) was detected at early times postinfection (1 to 2 hpi). Phosphorylation of the ephrins (EphA1, EphB4, and EphB1), Axl, Src, Lck, and IRS-1 occurred after 2 hpi. These findings confirm the kinome profiling results and support the finding that the SFK-PI3K-Akt-mTOR signaling pathway is activated following CHIKV infection.
The Src family kinase inhibitor dasatinib blocks CHIKV replication and reduces CHIKV-induced cell death. Both of the kinome profiling experiments showed that phosphorylated Src (p-Src) and p-Akt levels change over time during CHIKV infection. To investigate the functional importance of the SFK-PI3K-Akt-mTOR signaling pathway, chemical inhibitors of this pathway were tested for their ability to block CHIKV replication. To ensure that the inhibitors were used at concentrations that were not cytotoxic, cell viability assays were performed in uninfected NHDFs 24 h after drug treatment (Fig. S2). Inhibitor activity was verified in NHDFs by Western blotting for the presence of phosphorylated ribosomal protein S6 (p-rpS6) (Fig. 2C), a common marker of mTOR activation (59,60). To assess the effects of the SFK-PI3K inhibitors on CHIKV replication, NHDFs were pretreated for 2 h with the inhibitor and infected with CHIKV SL15649 (MOI ϭ 1 PFU/cell) in the presence of drug. At 2 hpi, the inoculum was removed, cells were washed with phosphate-buffered saline (PBS), and new drugcontaining medium was added. Supernatants were collected at 20 hpi, and virus titers were determined by serial-dilution plaque assays on Vero cells. Dasatinib and Torin 1 treatments reduced yields of infectious virus over 10-fold ( Fig. 2A) and reduced CHIKV-mediated cytotoxicity (Fig. 2B). PP242 and LY294002 treatments mildly reduced CHIKV yields. In contrast, treatment with GSK 690693, GSK 650394, and CGP 57380 had no significant effect on CHIKV replication. Treatment with RAD001 and rapamycin slightly increased CHIKV replication in NHDFs ( Fig. 2A), which is consistent with data from previous reports (61)(62)(63).
Of the inhibitors tested, dasatinib treatment resulted in the strongest inhibition of CHIKV replication. Dasatinib is a potent inhibitor of SFKs (64,65). As confirmation of this activity in NHDFs, the activating phosphorylation of Src at Tyr416 was analyzed in cell lysates at 4, 8, and 24 hpi (66). In the absence of dasatinib, there was an increase of Src Tyr416 phosphorylation in CHIKV-infected cells compared with uninfected cells. Dasatinib reduced Src phosphorylation in both infected and uninfected cells (Fig. 2D).
Dasatinib has many cellular targets, including Src, Lck, Lyn, Yes, Fyn, ephrin receptors, and Bcr/Abl (67). To confirm the importance of Src in CHIKV replication, CHIKV growth was measured in SYF Ϫ/Ϫ mouse embryonic fibroblasts (MEFs) (lacking Src, Yes, and Fyn) and Src ϩ SYF Ϫ/Ϫ MEFs, in which Src was stably reintroduced (68). At each time point evaluated, yields of infectious CHIKV from Src ϩ SYF Ϫ/Ϫ MEFs were consistently higher than those from SYF Ϫ/Ϫ MEFs at both high and low MOIs (Fig. 3). These data are consistent with the hypothesis that Src is required for efficient CHIKV replication.
To validate the inhibition of CHIKV replication, immunofluorescence staining was performed on CHIKV-infected cells treated with dasatinib or another SFK inhibitor, PP2. Cells were fixed at 24 hpi and stained with antibodies directed against CHIKV E2 and actin as well as the DNA stain DAPI (4=,6-diamidino-2-phenylindole). Foci of CHIKVinfected cells were observed in nontreated cells, whereas dasatinib and PP2 treatments resulted in the presence of single infected cells (Fig. 4). Similar findings were observed in VEEV TC83 -infected cells, where SFK inhibitors blocked the ability of the virus to spread through the culture (Fig. S3). Combined, these data indicate that SFK inhibitors block virus spread by reducing infectious virus production from infected cells.
Dasatinib and Torin 1 block alphavirus infection in a type I interferonindependent manner. CHIKV infection activates interferon (IFN) regulatory factor 3 (IRF3), resulting in the transcription of type I interferon, which blocks viral replication (69). Src can promote the phosphorylation of TANK-binding kinase 1 (TBK1), which activates IRF3 and promotes type I IFN production (70). Therefore, SFK inhibitors have the potential to augment CHIKV replication via repression of type I IFN responses. To test whether dasatinib-mediated inhibition of CHIKV replication was dependent on IRF3, human fibroblasts deficient in IRF3 (THF-ISRE-ΔIRF3) (71) were infected with CHIKV 181/25 or VEEV TC83 and treated at 0 hpi or 2 hpi with dasatinib. Titers in supernatants collected at 24 hpi were determined by a serial-dilution plaque assay, and the results showed that replication of CHIKV and VEEV was inhibited by dasatinib treatment in IRF3 Ϫ/Ϫ cells (Fig. S4A). Another component of the interferon response is the protein signal transducer and activator of transcription 1 (STAT1), which acts downstream of the type I interferon receptor and activates the transcription of interferon-stimulated genes. STAT1 knockout telomerized human fibroblasts (THFs) (THF-ISRE-ΔSTAT1) (71) were infected with CHIKV 181/25 or VEEV TC83 and treated with dasatinib or Torin 1. Quantification of viral yields at 24 hpi revealed that CHIKV or VEEV maintained sensitivity to dasatinib in STAT1 Ϫ/Ϫ THFs (Fig. S4B). Together, these data indicate that the mechanism of dasatinib and Torin 1 inhibition of CHIKV replication is independent of their potential effects on IRF3-and STAT1-mediated innate immunity.
Torin 1 and dasatinib differentially regulate autophagy. Autophagy is an important downstream pathway of mTOR, which is induced when mTOR is inhibited (72). There are conflicting reports regarding whether CHIKV replication is facilitated or impaired by autophagy (61,63,73), but both Torin 1 and dasatinib block mTOR signaling. Therefore, we next determined whether Torin 1 and dasatinib treatments induced autophagy in NHDFs and whether this impacted alphavirus replication. CHIKV 181/25 -infected cells were treated with dasatinib, Torin 1, or 3-methyladenine (3-MA) (1 mM), an inhibitor of autophagy. At 7 hpi, cell lysates were harvested and evaluated by Western blotting to examine the proteolytic cleavage of LC3-I to LC3-II, a conventional indicator of the induction of autophagy (74). Torin 1 induced LC3-II accumulation, but dasatinib and 3-MA did not alter LC3 status compared with untreated infected cells (Fig. S5A). Blotting cellular lysates for E2 showed that while dasatinib and Torin 1 reduced E2 levels, 3-MA increased CHIKV E2 accumulation compared with the untreated control (Fig. S5B). Our data indicate that Torin 1 and dasatinib inhibit CHIKV replication via different mechanisms: Torin 1 may block virus through the induction of autophagy, while dasatinib may have an alternative mechanism.
Dasatinib does not block CHIKV RNA replication. To further identify the step in the virus replication cycle sensitive to SFK inhibition, viral RNA (vRNA) accumulation was monitored in dasatinib-or Torin 1-treated NHDFs infected with CHIKV 181/25 or VEEV TC83 (MOI ϭ 3 PFU/cell). Viral RNA levels were measured by reverse transcriptionquantitative PCR (qRT-PCR) at 12 hpi. Dasatinib and Torin 1 modestly reduced CHIKV RNA levels ( Fig. 5A) but had no discernible effect on VEEV RNA accumulation (Fig. 5B). To validate and further explore the lack of effects of dasatinib on vRNA production, we performed a Northern blot analysis on RNA isolated from treated NHDFs infected with CHIKV (MOI ϭ 3 PFU/cell). A double-stranded DNA (dsDNA) digoxigenin (DIG)-labeled probe recognizing E2-6K-E1 was generated to distinguish between genomic vRNA and subgenomic vRNA. Dasatinib treatment minimally reduced the levels of genomic and subgenomic CHIKV RNAs to 79% and 98% of the viral RNA levels of the untreated control, respectively (Fig. 5C). qRT-PCR analysis confirmed that the ratios of CHIKV nsP2 and E2 RNA levels were unaffected by dasatinib treatment (Fig. S6A and S6B). Combined, these data indicate that CHIKV RNA amplification in the presence of dasatinib is largely unaffected, and the minimal effects on RNA levels could be attributed to the abrogation of subsequent rounds of replication due to the block in the production of infectious virus.
Next, we determined whether dasatinib disrupts CHIKV replication complex formation. At 8 hpi, replication complexes were identified by immunofluorescence staining with an antibody specific for dsRNA (21,75). The dsRNA-positive complexes appeared indistinguishable between dasatinib-treated cells and untreated cells (Fig. 5D). Additionally, there were no visual differences in the amounts of dsRNA staining at 8 hpi ( Fig.   FIG 4 CHIKV replication and spread are limited by dasatinib and PP2. NHDFs were infected with CHIKV 181/25 (MOI ϭ 1 PFU/cell) and treated with dasatinib or PP2 at 2 hpi. At 24 hpi, cells were fixed, stained for virus-specific envelope glycoproteins (green), and counterstained with DAPI (blue) and phalloidin (red) to highlight host DNA and actin filaments. S6C), which is consistent with our data demonstrating that dasatinib affects CHIKV replication downstream of vRNA synthesis.
Dasatinib blocks CHIKV and VEEV structural protein production. We hypothesized that dasatinib does not influence the synthesis of viral nonstructural proteins since viral RNA levels and replication complex formation appeared normal in treated cells. To directly test this hypothesis, Western blot analysis for nsP3 was performed on lysates from treated CHIKV-infected cells at 8 hpi. We observed no difference in nsP3 levels with dasatinib treatment (Fig. 6A) in the presence of a strong reduction in E2 protein accumulation (Fig. 6B). Similarly, levels of VEEV E2 glycoprotein (Fig. 6C) and capsid (Fig. 6D) were also dramatically reduced in the presence of the dasatinib. Together, these data suggest that dasatinib blocks the synthesis of alphaviral structural proteins but has minimal effects on the production of vRNA or nonstructural proteins.
Next, we sought to determine the mechanism by which dasatinib blocks CHIKV structural protein accumulation. Ribosomal profiling was performed on lysates from CHIKV-infected, inhibitor-treated NHDFs at 12 hpi. Lysates were fractionated by centrifugation over a linear sucrose gradient, and 17 fractions were collected and analyzed for the optical density at 264 nm (OD 264 ) as well as for the level of vRNA by qRT-PCR. Absorbance readings of the fractions revealed that dasatinib caused an increase of the absorbance in the 80S monosomal peak and a decrease in the absorbance in polysome fractions (fractions 9 to 15) (Fig. 7A). In dasatinib-treated cells, the vRNA was shifted into fractions containing monosomes (Fig. 7B). The translation efficiency (ratio of viral RNA levels in the polysome fractions to those in the fractions containing ribosomal subunits and monosomes) was reduced by approximately 80% in dasatinib-treated samples (Fig.  7C). These results indicate that dasatinib treatment disrupted CHIKV structural protein synthesis by preventing the translation machinery from associating with viral sgmRNA.
Dasatinib does not inhibit alphavirus-mediated host translational shutoff. The precise mechanism of translation of CHIKV and VEEV sgmRNAs during host transcription and translation shutoff is not known (76). Translation of the sgmRNA may not require many of the canonical translation initiation factors, as was previously documented for SINV (29,30,77). To determine whether dasatinib affects the virus-induced host translation shutoff process, we performed a puromycin pulse-chase assay in NHDFs infected with CHIKV or VEEV. Although dasatinib significantly decreased viral E2 accumulation, there was no effect of treatment on virally induced host translation shutoff at 24 hpi (Fig. 8A). Additionally, dasatinib did not affect puromycin incorporation into newly synthesized protein in uninfected cells, whereas cycloheximide (CHX) treatment abolished protein production, and Torin 1 reduced global protein production (Fig. 8A).
A marker for the shutoff of cellular translation during viral infection is the phosphorylation status of the alpha subunit of eukaryotic initiation factor 2 (eIF2␣), a host initiation factor essential for the translation of capped mRNAs. Phosphorylation inactivates eIF2␣, thereby preventing cap-dependent cellular translation. Translation of CHIKV sgmRNA likely does not require eIF2 because the sgmRNA is translated efficiently when eIF2␣ is phosphorylated (69). One of the kinases responsible for eIF2␣ phosphorylation during CHIKV replication is the cytoplasmic dsRNA sensor protein kinase R (PKR), which is activated by CHIKV infection (69). Therefore, we determined the effects of dasatinib and Torin 1 treatments on CHIKV replication in cells lacking PKR. Compared with the parental cell line, dasatinib and Torin 1 still reduced E2 levels in PKR Ϫ/Ϫ THFs (Fig. 8B). Consistent with previously reported data, CHIKV infection induced eIF2␣ phosphorylation in a PKR-dependent manner. In addition, dasatinib treatment in- creased eIF2␣ phosphorylation in both the parent THF cell line and PKR Ϫ/Ϫ THFs, suggesting that dasatinib induces eIF2␣ phosphorylation in a PKR-independent manner. Since dasatinib induces eIF2␣ phosphorylation, we tested whether activating transcription factor 4 (ATF4) levels were affected, as they are expected to increase when eIF2 is phosphorylated/inactivated (78)(79)(80). There was an increase in ATF4 levels following the addition of dasatinib, suggesting that dasatinib induces a stress response in the cell. Despite inducing eIF2␣ phosphorylation and increasing ATF4 levels, dasatinib did not have an effect on global protein synthesis, as shown by puromycin incorporation (Fig. 8A).
We next tested whether dasatinib could block translation from in vitro-transcribed and capped mRNAs transfected into NHDFs. mRNA was generated by in vitro transcription from the CHIKV 181/25 cDNA clone or a CHIKV replicon (pMH42), where mCherry replaces the CHIKV structural genes (Fig. 8C). Transfection of these mRNAs into cells results in nsP production, replication complex formation, and sgmRNA translation, leading to mCherry (replicon derived) or the viral structural proteins (cDNA clone derived). NHDFs were treated with dasatinib, and the production of E2 or mCherry was quantified at 20 h posttransfection (hpt) by Western blotting. Levels of pE2 were reduced following dasatinib treatment in mRNA-transfected NHDFs (Fig. 8D), and mCherry expression was decreased in pMH42-transfected cells, confirming that dasatinib blocks the translation of the sgmRNA.
To determine whether dasatinib could block sgmRNA outside the context of the replicon, another mRNA was generated, which contained the CHIKV 5= subgenomic untranslated region (sgUTR) followed by a hemagglutinin (HA)-tagged version of enhanced green fluorescent protein (eGFP) (sg-HA) (Fig. 8C). mRNAs derived from the pMH42 mCherry-expressing replicon, sg-HA, or a combination of both were transfected into NHDFs with and without dasatinib treatment. Consistent with our above-described data, levels of mCherry generated from the replicon were reduced after treatment with dasatinib (Fig. 8E). In contrast, samples transfected with sg-HA mRNA demonstrated increased HA levels in the presence of dasatinib. When both mRNAs were transfected together, mCherry levels were reduced relative to those with replicon transfection alone, but expression of mCherry was blocked in the presence of dasatinib, while HA levels were increased relative to those in untreated cells. These data demonstrate that dasatinib blocks the translation of sgmRNAs generated from replication-derived complexes but does not block translation of mRNAs outside the context of the viral replication machinery. However, we cannot rule out the possibility that the effect of dasatinib on translation is linked to virus-mediated changes in the cellular environment that may not be fully recapitulated in this system. Dasatinib and Torin 1, but not rapamycin, block eIF4E phosphorylation. SFKs modulate a complex array of signaling pathways, including the Raf/MEK/Erk and the Akt/mTOR pathways (81)(82)(83). The Raf/MEK/Erk pathway and p38 mitogen activated protein (p38 MAP) kinases contribute to activation of eIF4E, which is a rate-limiting factor in the translation initiation complex of capped mRNAs (84). Figure S7A in the supplemental material confirms that dasatinib blocks Erk1/2 phosphorylation in NHDFs. To test whether signaling through the Raf/MEK/Erk/MNK cascade (MNK is mitogenactivated protein kinase interacting protein kinase) was responsible for the reduction in viral structural protein synthesis, NHDFs were infected with CHIKV and treated with U0126, a specific inhibitor of MEK1/2 (85). U0126 reduced the phosphorylation of Erk1/2, but the treatment failed to affect viral titers or CHIKV E2 protein levels (Fig. S7B). Similarly, the MNK inhibitor CGP 57380, which blocks eIF4E phosphorylation, did not block CHIKV replication (Fig. 2A). These results indicate that the dasatinib-mediated decrease in viral structural protein synthesis is not related to its effects on the Raf/MEK/Erk signaling pathway.
ONNV, MAYV, and RRV are sensitive to dasatinib and Torin 1. We next examined the phylogenetic breadth of viruses sensitive to dasatinib and Torin 1. NHDFs were infected with ONNV, MAYV, or RRV and treated with dasatinib or Torin 1. At 20 hpi, titers in the supernatants from infected cells were determined by serial-dilution plaque assays (Fig. 9A). Titers of ONNV, MAYV, and RRV were all significantly reduced by dasatinib and Torin 1, suggesting a common mechanism of action against a broad range of alphaviruses. To determine if a more distantly related mosquito-transmitted RNA virus was also sensitive to dasatinib and/or Torin 1, we tested the ability of these inhibitors to block the replication of the flavivirus Zika virus (ZIKV). Dasatinib treatment of ZIKV-infected NHDFs blocked viral replication; however, Torin 1 failed to block ZIKV infection (Fig. 9B). Consistent with the viral titers, ZIKV E protein levels were reduced by dasatinib but only moderately decreased by Torin 1. Together, these data indicate that CHIKV, VEEV, ONNV, RRV, and MAYV are sensitive to both dasatinib and Torin 1. However, ZIKV exhibits different sensitivities to Torin 1 and dasatinib, indicating a unique requirement for the Akt/mTOR pathway.

DISCUSSION
Alphaviruses such as CHIKV, VEEV, MAYV, and ONNV are emerging pathogens that cause severe arthritis or encephalitis. Since there is no FDA-licensed antiviral therapy or vaccine, there is an urgent need for basic antiviral research to facilitate the creation of new therapeutics. In this study, we sought to understand the importance of SFK signal transduction during alphavirus infection. We found that the SFK inhibitor dasatinib and the mTORC1/2 inhibitor Torin 1 were effective in blocking virus replication at the level of structural protein synthesis (Fig. 10). Mechanism-of-action studies demonstrated that both dasatinib and Torin 1 had minimal to no effect on the formation of viral replication complexes or levels of vRNA. However, both drugs dramatically inhibited viral structural protein synthesis. Polysome profiling revealed that there were fewer CHIKV viral mRNA transcripts associated with translation-active ribosomal complexes in dasatinib-treated cells. These results demonstrate that dasatinib-mediated inhibition of alphaviruses occurs at a replication stage between viral RNA transcription and structural protein synthesis.
Dasatinib is a broadly acting SFK inhibitor that affects several cellular processes, and many of these processes may be required for effective alphavirus replication. We show that SFK inhibition blocks infection in the absence of functional interferon signaling in cells lacking IRF3 or STAT1. We also demonstrate that dasatinib, unlike Torin 1, does not induce autophagy, as indicated by LC3-II accumulation. Interestingly, inhibition of autophagy by 3-MA increased E2 accumulation, suggesting that autophagy has a negative impact on viral replication. Thus, promotion of autophagy by Torin 1 inhibition may contribute to the impairment of alphavirus translation and replication. These conclusions contradict other reports using HEK293 cells (61), but these disparities may reflect differences in cell types. Thus, these results demonstrate that dasatinibmediated inhibition of alphavirus replication in human fibroblasts is independent of interferon signaling and autophagy processes.
We hypothesized that SFK signaling events are required for translation of viral sgmRNAs. In addition to inhibiting viral structural protein translation during infection, dasatinib treatment also blocked the translation of replicon-derived sgmRNAs. Importantly, we show that dasatinib did not inhibit translation of host mRNAs or exogenously transfected sgmRNAs but that its effects were specific to the replicon-derived sgmRNAs. These results are reminiscent of those from previous studies with SINV replicons showing that the translational behavior of transfected subgenomic mRNAs is different from that of subgenomic mRNAs derived from replicons (86,87). SINV replicon-derived sgmRNAs were reported to be translated efficiently in the presence of the eIF4 inhibitor arsenite, and the replicon-derived sgmRNA had reduced requirements for eIF2␣, eIF4A, and eIF4G (29,77). In contrast, translation of the exogenously transfected sgmRNA was sensitive to arsenite and required eIF4G. Therefore, we found that dasatinib targets the translation processes unique to the arsenite-insensitive sgmRNA but not other mRNAs.
During alphavirus infection, translation of host mRNAs is strongly inhibited, while translation of viral sgmRNAs is very efficient (69, 88, 89). Although host translation inhibition can occur as a result of virus-induced inactivation of eIF2␣ by the dsRNA- sensing protein kinase R (PKR), we observed host translation shutoff in PKR Ϫ/Ϫ cells, where eIF2 is still active, suggesting an alternative mechanism (69,90). Consistent with those previous findings, we show that CHIKV induced eIF2␣ phosphorylation in a PKR-dependent manner. In addition, dasatinib enhanced eIF2␣ phosphorylation in PKR ϩ/ϩ and PKR Ϫ/Ϫ cells. eIF2␣ is phosphorylated by one of four kinases: hemeregulated inhibitor (HRI), PKR-like endoplasmic reticulum kinase (PERK), general control nondepressible 2 (GCN2), or PKR (91). As such, dasatinib may induce eIF2␣ phosphorylation through one of these kinases. One hypothesis based on our results is that the virus requires eIF2 for its replication and that dasatinib tips the balance of active versus inactive eIF2 available for the viral sgmRNA. Alternatively, the virus does not require active eIF2 for translation of its sgmRNA, and dasatinib mediates another effect on the cell that blocks the synthesis of viral structural proteins.
Previous findings regarding the translation of alphavirus sgmRNA support a model in which alphaviruses do not require activated eIF2 (28,32,92). SINV possesses a structural element in the 5= end of the sgmRNA called a downstream loop (DLP), which promotes stalling of the ribosome and negates the need for an activated eIF2 (28,31,32). Other alphaviruses either possess similar DLP structures with different stabilities and stem lengths or have no predicted DLP-like structure (31). CHIKV, VEEV, and ONNV do not possess a predicted DLP structure and may have a different feature that serves the same function. These alphaviruses may have differential requirements for translation initiation. Conversely, MAYV has a predicted DLP structure, but it was still sensitive to dasatinib treatment. More experiments are needed to determine whether dasatinib-mediated inhibition of eIF2 is the mechanism of action for inhibiting translation of the viral sgmRNA or whether a different feature of translation is blocked.
To identify an SFK-dependent signaling pathway that modulates CHIKV replication, we utilized a panel of inhibitors specific for downstream molecules of the Ras/Raf/MEK/ ERK and PI3K/Akt/mTORC1 signaling pathways. We show that treatment with U0126, an inhibitor of MEK1/2, did not affect viral protein synthesis or levels of infectious virus. This indicated that the effect of dasatinib on viral protein synthesis was mediated by another signaling event, such as the PI3K/Akt/mTOR pathway. Previous studies found that mTORC1 inhibition by rapamycin and other rapalogs could enhance CHIKV repli- In this study, we found that CHIKV is sensitive to the SFK inhibitor dasatinib, the PI3K inhibitor LY294002, and the mTORC1/2 inhibitors Torin 1 and PP242. We show that dasatinib and Torin 1 block CHIKV sgmRNA translation. In contrast, inhibitors of the Ras/Raf/MEK/ERK pathway did not block CHIKV sgmRNA translation. cation, presumably by enhancing the phosphorylation of eIF4E (62). Rapamycin inhibition of mTOR may result in feedback activation of Akt and increase eIF4E phosphorylation, while PI3K inhibitors and dual mTORC1-mTORC2 inhibitors decrease eIF4E phosphorylation (93)(94)(95). This feedback loop could explain the discordant effects of Torin 1 and rapamycin treatments on alphavirus replication. We found, however, that CGP 57380, an MNK1/2 inhibitor that blocks eIF4E phosphorylation, did not block CHIKV replication, suggesting that eIF4E may not be necessary for translation of the CHIKV sgmRNA.
Our data indicate that SFK signaling events play a crucial role during alphavirus replication at the step of structural protein translation but not at the level of viral RNA synthesis. Therefore, further research on these pathways will facilitate the development of novel therapeutics that target the translation step of alphavirus replication and provide insight into the signaling pathways required for efficient replication.
(Thermo Scientific). Peptides were separated on an Easy Spray PepMap C 18 column over a 2-h gradient of 5 to 32% mobile phase B at a 250-nl/min flow rate, where mobile phase A was 0.1% formic acid in water and mobile phase B consisted of 0.1% formic acid in acetonitrile. The QExactive HF mass spectrometer was operated in the data-dependent mode, and the 15 most-intense precursors peptides were selected for subsequent MS/MS analysis. To identify kinases, peptide data were analyzed against the current version of the human UniProt database using MaxQuant software, as previously described (57,58). Peptide data are available in Table S1 in the supplemental material. Temporal kinome shifts and clustering were determined by k-means clustering performed in the R statistical programming language. Kinase clustering data are presented in Table S2.
Quantitative RT-PCR analysis. Viral RNA in infected NHDFs was quantified using real-time quantitative RT-PCR. Primers and probes include forward primer GAGGTGTGGGACTGGTTGTTG, reverse primer CAAGTTAGTGCCTGCTGAACGA, and probe AATCGTGGTGCTATGCGT for CHIKV-LR E1; forward primer CGGCGTCTACCCATTTATGT, reverse primer CCCTGTATGCTGATGCAAATTC, and probe AAACACGCAGTTG AGCGAAGCAC for CHIKV E1; forward primer AGAGACCAGTCGACGTGTTGTAC, reverse primer GTGCGCA TTTTGCCTTCGTA, and probe ATCTGCACCCAAGTGTACCA for CHIKV nsP2; and forward primer TCCCGAC GCCTTGTTCAC, reverse primer CGCCAAAGTCGGATGAATACA, and probe CACCGACACTTTCAGCGG for VEEV TC83 E2. Total RNA was prepared from infected NHDFs by the TRIzol method at 0, 2, 4, 6, 8, 12, and 24 hpi. RNA was treated with RNase-free DNase, and single-stranded cDNA was then generated using random hexamers and Superscript III RT (Invitrogen). Gene amplicons served as quantification standards (the limit of detection is approximately 10 to 100 copies). Quantitative RT-PCR results were analyzed using an ABI StepOne Plus real-time PCR system (Applied Biosystems).
In vitro mRNA synthesis and mRNA transfections. A CHIKV subgenomic mRNA construct expressing an eGFP-HA-tagged fusion protein (sg-HA) was generated as follows: the CHIKV 5= sgUTR was amplified from pMH56 using forward primer 5=-ATATAAGCTTCGTCATAACTTTGTACGGCGG-3= and reverse primer (5=-CTCCTCGCCCTTGCTCACCATTGTAGCTGATTAGTGTTTAG-3=. The eGFP-HA fragment was amplified using forward primer 5=-CTAAACACTAATCAGCTACAATGGTGAGCAAGGGCGAGGAG-3= and reverse primer 5=-ATATGGATCCTTAGGCGTAGTCGGGCACATCGTACGGGTACTTGTACAGCTCGTCCATGC-3=. The sg-HA fusion gene was synthesized by overlapping PCR and cloned into the pSP64 poly(A) vector. For vRNA synthesis, the CHIKV 181/25 cDNA clone and pMH42 were first linearized with NotI (NEB), or pSP64 was digested with EcoRI (NEB). Capped and polyadenylated mRNA transcripts were generated from the linearized plasmids using the mMessage mMachine SP6 transcription kit (Thermo Fisher). Samples were digested with Turbo DNase I, and mRNAs were purified using the RNeasy Plus minikit (Qiagen). mRNAs were transfected into fibroblasts using Lipofectamine MessengerMAX (Invitrogen) according to the manufacturer's instructions.
PathScan RTK antibody array. PathScan RTK signaling phosphoantibody array analysis (Cell Signaling) was performed according to the manufacturer's protocol. Briefly, lysates from NHDFs infected with CHIKV SL15649 at an MOI of 1 PFU/cell were collected at 0, 1, 2, 4, 6, 8, and 24 hpi. Protein concentrations were normalized, and the samples were incubated on an RTK antibody array chip for 2 h. The plates were washed and incubated with a detection antibody cocktail for 1 h. The plates were washed and incubated with a DyLight 680-linked streptavidin secondary antibody, after which they were washed, dried, and scanned using an Odyssey infrared imaging system (Li-Cor). Spot intensities were quantified with Odyssey quantification software and are reported relative to values for the uninfected samples.
Immunofluorescence. Cells were plated onto 4-well Lab-Tek chambered cover glass slides (Fisher) and infected as indicated. Cells were fixed with 4% paraformaldehyde for 15 min and washed with PBS. Cells were incubated for 1 h with permeabilization buffer containing 0.2% saponin and 1% BSA. Cells were incubated overnight, as indicated, with anti-CHIKV E2, anti-VEEV gp, or anti-dsRNA (J2; Scicons) diluted 1:1,000 in the presence of saponin and BSA. Cells were washed and incubated with goat anti-mouse Alexa Fluor 488 (Abcam), DAPI, and/or anti-mouse Texas Red-X phalloidin for 1 h. Cells were washed and visualized using a confocal fluorescence microscope.
Polysome analysis. NHDFs were mock infected or infected with CHIKV 181/25 (MOI ϭ 3 PFU/cell). At 2 hpi, medium was removed, and cells were washed twice with PBS. The medium was replaced with or without 10 M dasatinib. At 12 hpi, 0.1 mg/ml cycloheximide (Sigma-Aldrich) was added to the medium, and cells were incubated for 10 min at 37°C. Following incubation, cells were washed twice with ice-cold PBS containing 0.1 mg/ml cycloheximide. The cells were scraped, pelleted for 10 min at 1,130 ϫ g, and frozen at Ϫ80°C. Samples were resuspended in polysome buffer (20 mM Tris-HCl [pH 7.4], 140 mM KCl, 5 mM MgCl 2 ) containing 0.1% Triton X-100, 10 mM DTT, and 100 M CHX. Cells were disrupted by five passes through a 27-gauge needle, and nuclei were removed by centrifugation for 5 min at 1,150 ϫ g. The resulting cytoplasmic lysates were layered onto 10 to 50% linear sucrose gradients prepared in polysome buffer and centrifuged at 35,000 ϫ g for 2 h at 4°C with no brake. The gradients were fractionated with continuous absorbance monitoring at an OD 254 using a gradient fractionation system (Brandel). RNA was isolated from each fraction and quantified by qRT-PCR using the absolute quantification method. Translation efficiency was calculated as the sum of CHIKV RNA copies in the polysomecontaining fractions (fractions 9 to 15) divided by the sum of CHIKV RNA copies in the fractions containing ribosomal subunits and monosomes (fractions 4 to 7) (n ϭ 2).

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.1128/AAC .02325-18.