Facile method for delivering chikungunya viral replicons into mosquitoes and mammalian cells

Reverse genetics is an important tool in the elucidation of viral replication and the development of countermeasures; however, these methods are impeded by laborious and inefficient replicon delivery methods. This paper demonstrates the use of a baculovirus to facilitate the efficient delivery of autonomous CHIKV replicons into mosquito and mammalian cells in vitro as well as adult mosquitoes in vivo. The efficacy of this approach was verified via co-localization among an eGFP reporter, nsP1, and dsRNA as well as through the inhibition of an RNA-dependent RNA polymerase (RdRp) null mutation (DDAA) in nsP4, or the treatment of a known antiviral compound (6-azauridine). We also investigated the correlation between CHIKV replicon-launched eGFP expression and the effectiveness of CHIKV replicon variants in inducing IFN-β expression in human cell lines. This delivery method based on a single vector is applicable to mosquito and mammalian cells in seeking to decipher the mechanisms underlying CHIKV replication, elucidate virus–host interactions, and develop antivirals. This study presents an effective alternative to overcome many of the technological issues related to the study and utilization of autonomous arbovirus replicons.


Results
In vitro functional analysis. We established a method for the efficient delivery of autonomous CHIKV replicons into mosquito and mammalian cells in order to assess the self-replicating CHIKV replicon-mediated expression of eGFP (Fig. 1a). Recombinant baculovirus bears a CMV promoter-directed wild type (WT/eGFP) or CHIKV replicon variant (G1332V/eGFP and DDAA/eGFP) containing the eGFP gene under the transcription control of CHIKV 26S subgenomic promoter (Fig. 1b). To determine whether the CHIKV replicon was functional, two mammalian cells (U2OS and HEK293T) and two mosquito cells (AP61 and C6/36) were serially transduced with WT/eGFP. The results (Figs. 2, 3) indicate the strong dose-dependency of eGFP expression in mosquito and mammalian cells, which was particularly pronounced in mosquito cells. Note that eGFP levels obtained from mosquito cells should far exceed those obtained from mammalian cells. The most pronounced eGFP expression was detected in WT/eGFP-transduced AP-61 cells. All transduced cells except C6/36 cells presented eGFP expression with dose-dependency proportional to the MOI. Note that at MOIs of 5 and 10, WT/ eGFP-transduced C6/36 cells presented a rounded morphology; however, at an MOI of 10, WT/eGFP-transduced AP-61 cells appeared normal. This is an indication that C6/36 cell are perhaps more sensitive to WT/eGFP transduction than are AP-61 cell. Next, to examine the correlation of eGFP expression and CHIKV replication, immunofluorescence assay (IFA) was conduct to detect eGFP, dsRNA (intermediate of CHIKV replication) and nsP1(viral protein of CHIKV) in transduced AP-61 cells. We also observed the specific co-localization of eGFP, dsRNA and nsP1 in WT/eGFP-transduced AP-61 cells (Fig. 4), indicating RNA replication. To confirm the identity of the autonomous CHIK replicon, we performed functional analysis of IVT-RNA from WT/eGFP. The results (Fig. S1) revealed a high eGFP-positive ratio in IVT-RNA-electroporated BHK-21 cells, indicating the successful expression of eGFP by self-amplified RNA transcripted from the transfer vector of WT/eGFP. Taken   Evaluation of antiviral targeting CHIKV replicon. In order to confirm the correlation of eGFP expression and CHIKV replication, we assessed the efficacy of this proposed system in antiviral assays by examining the inhibition effect of 6-AU on CHIKV replicon-launched eGFP expression. The eGFP expression was assessed after culturing WT/eGFP-transduced AP-61 cells as well as WT/eGFP-or CMV-GFP-transduced U2OS cells (as a control) in the presence of 6-AU at specified dilutions. As shown in Fig. 5a, 6-AU treatment reduced eGFP expression in WT/eGFP-transduced U2OS cells in a dose-dependent manner; however, it had no effect on eGFP expression in CMV-GFP-transduced U2OS cells. We also calculated the 50% inhibitory concentration (IC50) and 50% cytotoxicity concentration (CC50) of 6-AU in U2OS cells at concentrations of 1 μg/ml and > 10 μg/ml. Interestingly 6-AU was shown not to affect CHIKV replicon-launched eGFP expression in AP-61 cells (Fig. 5b).
Overall, 6-AU exhibited a mammalian cell-biased inhibitory effect on baculovirus-mediated CHIKV replicon replication. Note that these results are in line with those obtained in previous studies on mammalian cells. Taken together, our findings further support the CHIKV replicon-mediated eGFP expression and suggest that this proposed replicon system could be utilized for the evaluation of antiviral treatments.

Evaluating the functionality of CHIKV replicon variants in dual-host cells.
To evaluate the functionality of CHIKV replicon for eGFP expression, recombinant baculoviruses bearing either a null functional  . Interestingly, eGFP intensity in G1332V/eGFP-transduced HEK293T and C6/36 cells was significantly higher than WT/eGFP transduced those both cells. By contrast, the expression level of eGFP of WT/eGFP was extremely significant higher than eGFP level of G1332V/eGFP in U2OS cells.
The eGFP intensity of G1332V/eGFP was lower than that of WT/eGFP at 5 and 7 dpt in AP-61 cells (Fig. 8a). As expected, no eGFP signal was observed in either DDAA/eGFP-transduced mosquito and mammalian cell lines, which indicates that eGFP expression can be attributed to the functional CHIKV replicon. In time-course assays ( Fig. 8), we observed the accumulation of intense eGFP signals in all transduced mosquito cells throughout incubation from 1 to 7 dpt. Note that in transduced mammalian cells, eGFP signals of peak intensity were observed at 2 dpt. Levels of eGFP expression were higher in transduced mosquito cells (10-20 × 10 3 RFU) than in transduced mammalian cells (< 10 × 10 3 RFU) at 7 dpt. To determine whether the pattern of CHIKV replicon-launched eGFP expression was associated with innate immunity status, we measured the level of IFN-β mRNA in transduced human cells (U2OS and HEK293T) at 4 and 24 post-transduction (hpt) using qRT-PCR. As shown in Fig. 9, at 4 hpt, IFN-β mRNA levels in WT/eGFP-and G1332V/eGFP-transduced human cells were significantly higher than in DDAA/eGFP-transduced human cells; however, we did not observe a statistically significant difference between WT/eGFP-and G1332V/eGFP-transduced human cells. At 24 hpt, IFN-β mRNA levels in G1332V/eGFP-transduced U2OS and HEK293T cells were significantly higher than in WT/ eGFP transduced cells; however, we did not observe a statistically significant difference between WT/eGFPand DDAA/eGFP-transduced human cells. This indicates that G1332V/eGFP is associated with the potency of IFN-β mRNA induction. Taken together, these results demonstrate the efficacy of the proposed replicon system in characterizing CHIKV replicon variants in terms of replicon-launched eGFP expression (replication) and the potency of IFN-β induction.
Functional analysis of in vivo delivery system into mosquitos. The efficacy of the proposed in vivo delivery system was evaluated by intrathorically injecting WT/eGFP, G1332V/eGFP, or DDAA/eGFP into adult mosquitoes of two species (A. aegypti and A. albopictus). Those eGFP signals were observed in both species of mosquito following injection with WT/eGFP or G1332V/eGFP. Time course results (data not shown) revealed that eGFP signals can be attributed to gene turn-on rather than carry over. As shown in Fig. 10a, the eGFP signals at 5 dpt were more pronounced in WT/eGFP-injected mosquitoes than in G1332V/eGFP-injected mosquitoes. Of note, the eGFP signals were more pronounced in A. aegypti mosquitoes than in A. albopictus mosquitoes. As expectedly, no eGFP signals were detected in DDAA/eGFP-injected mosquitoes. Quantitative results (Fig. 10b) of eGFP intensity confirm that levels of eGFP signal in WT/eGFP-injected both mosquitoes were significantly

Discussion
Reverse genetics is an important tool in the elucidation of viral replication and the development of novel vaccine strategies 22,[30][31][32] . Note however that nearly all reverse genetics systems are based on the cells of vertebrates, despite the fact that the lifecycle of arboviruses involves mosquitoes as well as mammals (i.e., invertebrates and vertebrates). Thus, a simple universal method by which to introduce genomes or replicons into both mammalian and mosquito cells could facilitate the study of functional genomics pertaining to arboviruses as well as the prediction of outbreaks and the development of countermeasures. Viral vectors that are accessible to mammalian cells in vitro may provide an efficient system for the autonomous delivery of replicons into target cells 24,33,34 ; however, no such system has been reported specifically for mosquito cells. Furthermore, most reverse genetics systems designed for positive-sense single-stranded RNA viruses associated with the Aedes mosquito cell line are based on the transfection or electroporation of DNA-and RNA-based replicons. Note that the efficiency of the RNA delivery systems is very low (0.2-10%) 27 . One study reported an improved transfection protocol to  www.nature.com/scientificreports/ overcome these technological issues for insect cells 28 . Efforts to use reverse genetics for arboviruses will require a new approach to the delivery of replicons into mosquitoes and mammalian cells.
In the current study, we developed a baculovirus-mediated CHIKV replicon delivery system (Fig. 1), in which a CHIKV cDNA replicon is delivered into mosquitoes and mammalian cells via transduction under the control of a CMV promoter. The use of a single baculovirus vector eliminates the trouble and expense of capped IVT-RNA preparation and electroporation. Baculovirus/CHIKV replicon-launched eGFP expression was validated via co-localization among an eGFP reporter, nsP1, and dsRNA (Fig. 4) as well as inhibition via 6-AU treatment (Fig. 5) or the introduction of a DDAA mutation in RdRp (Figs. 6, 7). CHIKV replicon-launched eGFP expression was more pronounced in mosquito cells than in mammalian cells (Figs. 2, 3, 8), due perhaps to a lower rate of transduction in mammalian cells 29 . eGFP expression was more pronounced in transduced-U2OS cells than in transduced-HEK293T cells, due largely to the efficiency of baculovirus transduction 35 . Alphavirus nsP2 virulence can be characterized in terms of the effect on the host antiviral response via two down-regulatory pathways: (i) IFN-β transcription via global host shutoff and (ii) JAK-STAT signaling via inhibition of STAT1 phosphorylation and/or nuclear translocation 13,15,[36][37][38][39][40][41] . Interestingly, this study demonstrated that the G1332V mutant (nsP2) is associated with an increase in eGFP expression in innate immunity-deficient C6/36 (from 1 to 7 dpt) and HEK293T cells (at 2 and 3 dpt) 42,43 and a decrease in eGFP expression in AP-61 (at 5 and 7 dpt) and U2OS cells (at 2 and 3 dpt) (Fig. 6, 7, 8). This is consistent with previous reports that the replication/transcription of CHIKV G1332V mutant did not alter activity in mammalian cells, but induced a 60-fold increase in C6/36 cells 12 . Researchers have reported a variety of defects in the innate immune responses to adenovirus infection in U2OS cells 44 . However, RNA viral infections tend to upregulate IFN-β expression in U2OS cells 45 . The nsP2 G1332V mutant is a potent inducer of IFN-β rapidly activated ISGs in U2OS cells, resulting in the inhibition of CHIKV replicon-launched eGFP expression. G1332V/eGFP-transduced cells (C6/36 and HEK293T) presented a significant increase in replication (eGFP expression), due perhaps to the delayed cleavage of nsP2-nsP3 9 . In C6/36 cells (incompetent innate immunity), the restriction of innate immunity could be relieved resulting in eGFP expression levels of G1332V/eGFP higher than WT/eGFP. HEK293T cells retained their transcriptional pro-inflammatory response but lost their ability to secrete type 1 IFN 43   www.nature.com/scientificreports/ (P726G) in SINV also resulted in elevated IFN-β production despite limited virus propagation in vivo 13 . Functional nsP2 in WT/eGFP was shown to induce global transcriptional shutoff via RPB1 degradation 38 , which prevented activation of the transcription-dependent IFN-β response and JAK-STAT signaling and also induced cell death. Conversely, the nsP2 G1332V mutant without the above functionalities was shown to strongly induce IFN-β expression. In cases of SINV infection, the G1332V mutant associated nonfunctional nsP2 strongly promoted α/β interferon production by preventing infected cells from downregulating cellular transcription 46,47 . Note however that replication of the SINV G1332V mutant was defective in mosquito cells but not in vertebrate cells 46 . It appears that nsP2 VLoop mutants are incapable of interfering with the induction of type I interferon attenuated chikungunya virus replication in vitro and in vivo 14 . The nsP1/nsP2/ nsP3 cleavage domains play a critical role in regulating cellular antiviral response induced by the Ross River Virus or Sindbis virus [46][47][48][49] . However, the role of the nsP1/nsP2/ nsP3 cleavage domains of CHIKV has yet to be elucidated. In the current study, we demonstrated that the nsP2/ nsP3 cleavage domain drives the CHIKV-induced IFN-β response.
In CHIKV-eGFP based assays, the IC 50 value for 6-AU was approximately 1 μg/ml (4 μM) (Fig. 5a), which is in good agreement with previous reports on reporter-based stable CHIKV replicon cell lines and CHIKV replication in Vero cells 23,50 . Note that 6-AU did not affect eGFP expression in mosquito cells. Previous studies reported that 6-AU exhibits a moderate anti-arboviral activity in mammalian cells 51,52 but not in mosquito cells 53 , indicating fundamental differences in the host-CHIKV or drug-host interactions. The transient replicon system developed in the current study could be a valuable tool by which to perform high-throughput antiviral screening, optimize antiviral activities, and elucidate antiviral mechanisms. Figure 10. In vivo functional analysis of CHIKV replicon variants in mosquitoes. Adults of A. aegypti (upper panels) and A. albopictus (lower panels) were intrathoracically injected with either WT/eGFP (left panels), G1332V/eGFP (middle panels), or DDAA/eGFP (right panels) at a dose of 5 × 10 6 pfu/mosquito. The eGFP expression (indicated by white arrows) was photographed (a) and quantified (b). Statistical significance was analyzed using Student's t-test (*p < 0.05 and **p < 0.01). www.nature.com/scientificreports/ To the best of our knowledge, this is the first study to report the in vivo delivery of functional CHIKV replicons into adult mosquitoes (Fig. 10). Following injection with WT/eGFP or G1332V/eGFP, eGFP expression was more pronounced in A. aegypti than in A. albopictus (Fig. 10a), due to differences in vector competence and/or CMV activity. Those eGFP signals in WT/eGFP-injected both mosquitoes were significantly higher than eGFP level in corresponding G1332V/eGFP-injected mosquitoes (Fig. 10b), due perhaps to the functional nsP2 restriction of innate mosquito immunity to WT CHIKV replication. Note that we have also conducted in vivo assays using another reporter of a near-infrared fluorescent protein (iRFP682) in mice; however, the results to date have been unsuccessful, due perhaps to the sensitivity of baculovirus and/or the restriction of innate immunity 54 . Efforts to elevate MOI and thereby permit the detection of RFP682 expression in mice are currently on-going.
Alphavirus replicons have previously been used to induce heterologous protein expression in mammalian cells 55,56 . This is the first study to demonstrate the use of a baculovirus for the efficient delivery of autonomous CHIKV replicons into mammalian and mosquito cells in vitro as well as adult mosquitoes in vivo. The peak CHIKV replicon-launched eGFP expression levels were observed at 2 dpt in mammalian cells and 7 dpt in mosquito cells. Persistent replicon-launched eGFP expression exceeding 30 days (data not shown) highlights the potential of the proposed replicon system for heterologous protein expression using mosquito cells. Previous study has demonstrated that baculovirus does not replicate in mosquito cell line 29 . The continuous increase in eGFP expression in transduced mosquito cells likely results in persistency of CHIKV replicon. Additionally, combining baculovirus/CHIKV systems may make it possible to create a single-round infectious virus from dual-host cells as an alternate approach to vaccine development. The proposed scheme allows the efficient delivery of replicons into mosquito and mammalian cells using a single vector to facilitate analysis of CHIKV replication and the development of treatments. The proposed replicon system was shown to improve the utility of alphavirus expression systems, particularly in mosquito cells. Due to inconsistencies in the persistent infectivity of alphaviruses in mosquitoes and mammalian cells, the proposed system could conceivably be used to develop a new generation of systems for alphavirus replicon-based protein expression in mosquito cells. This system also opens the door to the development of alphavirus VRP packages based on mosquito cells, while facilitating the in vivo delivery of alphavirus replicons into mosquitoes. www.nature.com/scientificreports/ IFA. In generating rabbit anti-nsP1 serum, a 911-bp NcoI-SmaI fragment (AF369024, 637-1548) of partial nsp1 gene from the pFastBac1-VSVG-CHIKV replicon-GFP was subcloned into the restriction enzyme sites of pET32c. The expression, purification, and rabbit immunization of recombinant nsp1 were performed in accordance with methods described in a previous report 57 . AP-61 cells were transduced using WT/eGFP at an MOI of 5. At 4 dpt, cells were fixed and incubated using the above-mentioned rabbit anti-nsp1 serum (1:100) or Mab anti-GFP antibodies (Santa Cruz Biotechnology, Inc.) (1:100) and mouse anti-dsRNA antibody (SCICONS) for 1 h. Cells were subsequently washed three times using PBS and incubated with either Alexa Fluor 594-conjugated secondary antibodies or Alexa Fluor 488-conjugated secondary antibodies as well as Hoechst 33342 (10 µM) for 1 h. After a final wash using PBS, images of cells were captured using an inverted fluorescence microscope (DMi8 Microscope, Leica Microsystems). eGFP quantification. Cells in triplicate collected from 24-well plates were lysed using 200 µL/well RIPA buffer (Thermo Scientific™). The eGFP expression was measured using DS-11 FX + spectrophotometer/fluorometer (DeNovix, USA) in accordance with the manufacturer's instructions. Three independent experiments in time course assay were performed with very similar results, and the data presented represent the results from one independent experiment.

Cells and viruses.
Inhibition assay. U2OS and AP-61 cells were seeded in 24-well plates at a density of 1 × 10 5 cells/well. On the following day, U2OS cells were transduced using either WT/eGFP or CMV-GFP (CellLight™ Early Endosomes-GFP, Invitrogen) as a negative control, respectively at on MOI of 5 or 1. AP-61 cells were transduced using WT/ eGFP at an MOI of 1. Transduced cells were cultured in triplicate using growth medium in the presence 6-AU (Sigma-Aldrich) at various indicated concentrations. Following incubation for 24 h, cells were subjected to eGFP quantification assay. In cell viability assays, U2OS or AP-61 cells were seeded in 96-well plate at a density of 2 × 10 4 cells/well. On the following day, cells were cultured in triplicate using growth medium in the presence 6-AU at various concentrations for 24 h. Cell viability was determined using the Cell Counting Kit-8 kit (Enzo) in accordance with the manufacturer's instructions. Absorbance at 450 nm was measured using a microplate reader Infinite ® 200 PRO (Tecan Trading, Ltd., Switzerland). Cell viability was evaluated in terms of absorbance at various doses divided by the absorbance at 0 ug/ml. Three independent experiments were performed with very similar results, and the data presented represent the results from one independent experiment. mosquitoes were reared at 27 °C under 80% humidity with a 12-h light/dark cycle. Cold-anesthetized mosquitoes were intra-thoracically inoculated with indicated viruses at a dose of 5 × 10 6 plaque forming units (pfu)/ mosquito. At 5 dpt, cold-anesthetized mosquitoes were photographed using an IX71 inverted fluorescence microscopy (Olympus). The levels of eGFP intensity of images was quantitated with Image J 1.53e version. We narrowed the hue tone of the color threshold (54 to 112) for the specific intensity of green color. The methods of quantification are outlined in the Image J documentation: ''Image → Adjust → Color Threshold → Measure'' . The eGFP intensity was defined as (Area X Mean fluorescence of readings). For calculation of relative intensities, the pixel intensity was normalized with the fluorescent signal of PBS-injected mosquitoes. Two mosquitoes were analyzed for each genotype.

Quantitative analysis of IFN
Statistical analysis. Data were analyzed using GraphPad Prism 6.01 software. Data was assessed for statistically significant differences using a two-tailed, unpaired t-test. A p value < 0.05 was considered statistically significant. p values were indicated as follows: ns, not significant; *p < 0.05, significant; **p < 0.01, highly significant; ***p < 0.001 and ****p < 0.0001, extremely significant.