An In Vivo Transfection Approach Elucidates a Role for Aedes aegypti Thioester-Containing Proteins in Flaviviral Infection

Mosquitoes transmit pathogens that cause infectious diseases of global importance. Techniques to easily introduce genes into mosquitoes, however, limit investigations of the interaction between microbes and their arthropod vectors. We now show that a cationic liposome significantly enhances delivery and expression of plasmid DNA in Aedes aegypti and Anopheles gambiae mosquitoes. We then introduced the genes for Ae. aegypti thioester-containing proteins (AeTEPs), which are involved in the control of flaviviral infection, into mosquitoes using this technique. In vivo transfection of AeTEP-1 into Ae. aegypti significantly reduced dengue virus infection, suggesting that the approach can further our understanding of pathogen-mosquito interactions.


Introduction
Mosquitoes transmit pathogens of medical importance throughout the world. Approximately 700 million people develop mosquito-borne infectious diseases each year, and more than 2 million individuals die of these illnesses [1]. The ability of mosquitoes to transmit viruses and protozoa varies widely. Dengue virus (DENV) and Yellow Fever virus (YFV) are primarily transmitted by Aedes aegypti [2,3]; West Nile virus (WNV) is carried by Culex spp. and Aedes spp. [4][5][6]; and the genus Anopheles is the dominant vector for Plasmodium spp. [7]. Approved human vaccines are not yet available for many of these infectious diseases [8]. An comprehensive understanding of the survival of microbes in mosquitoes may facilitate the development of new strategies to control transmission and reduce infection. The limitation of current approaches to introduce genes into mosquitoes, however, impairs efforts to investigate the interactions between pathogens and these important arthropod vectors.
Studies in mosquitoes currently rely on several techniques, including RNA interference (RNAi) [9] and germline transformation [10]. RNAi is the most popular approach, and has been broadly applied to functional studies in mosquitoes [11]. Another common strategy is to produce stable transgenic mosquitoes through germline transformation. There are, however, some drawbacks in these approaches. A serious concern about RNAi is that dsRNA/siRNA may promote off-target silencing, leading to ambiguous conclusions of gene function as it relates to microbial infection [12,13]. Germline transformation is inefficient and time consuming, and the maintenance of transgenic strains is costly and laborious [14]. Development of a simple approach to efficiently express a gene in mosquitoes is, therefore, urgently needed and will accelerate the study of pathogen-mosquito interactions. We now report on the efficient introduction of plasmid DNA into mosquitoes and use this technique to elucidate the role of Ae. aegypti thioester-containing proteins (AeTEP) in flaviviral infection.

Results
An efficient approach to deliver genes into mosquitoes is a powerful tool for mosquito genetic studies. We used a GFPexpressing plasmid, pAc-GFP, which has a Drosophila actin promoter and GFP gene, to optimize the experimental condition. To transfect the GFP gene into mosquitoes, we directly microinjected pAc-GFP or empty vector DNA into the thorax of mosquitoes, as described in a previous study [15]. GFP expression, however, could not be readily detected either by RT-QPCR (Figure 1a) or by fluorescence microscopy (data not shown). This suggests that direct inoculation of DNA is inefficient and that protein expression is insufficient for functional studies.

In vivo transfection of pAc-GFP plasmid with Cellfectin IIH
A previous study showed that an exogenous gene was transfected into dissected Ae. aegytpi salivary glands by an in vitro liposome-based transfection [16]. We therefore reasoned that a cationic-liposome could also deliver plasmid DNA with high efficiency into live mosquitoes. Liposome-based reagents are widely used for DNA transfection into insect cell lines. Cellfectin IIH [17] and FuGene 6H [18] are common transfection reagents for cellular studies. We therefore examined whether these reagents could deliver plasmid DNA into adult female mosquitoes with high efficiency.
Liposome is potentially harmful to mosquitoes. FuGene 6H was lethal to Ae. aegypti, as after thoracic inoculation, only 1 out of 120 mosquitoes survived. Cellfectin IIH, however, was well tolerated by mosquitoes (Figure 1b), suggesting that it may be suitable for in vivo transfection studies. To evaluate the ability of Cellfectin IIH to deliver plasmid DNA, we mixed Cellfectin IIH with the pAc-GFP plasmid and microinjected the liposome-DNA mixture into Ae. aegypti. Empty DNA vector and pAc-GFP in medium served as controls. GFP mRNA was quantified by RT-QPCR at various time points. The empty vector and pAc-GFP alone groups expressed very little GFP; however, Cellfectin IIH significantly enhanced GFP expression (Figure 1a). Consistent with the QPCR results, GFP protein was dramatically enhanced by Cellfectin IIH post microinjection (Figure 2a). We then titrated the plasmid DNA via 2-fold dilutions, and microinjected this material with Cellfectin aegypti. 200 ng pAc-GFP and the empty vector DNA were microinjected per mosquito. Total RNA was isolated from whole mosquitoes and the samples were decontaminated by Dnase I treatment. The GFP expression was determined by SYBR GreenH QPCR, and normalized using Ae. aegypti actin (AAEL011197). Each dot represents 1 mosquito. The experiment was repeated twice with similar results. (b) Cellfectin IIH does not influence mosquito survival. Each liposome reagent mixed with vector DNA was microinjected into the mosquito (200 ng DNA each mosquito). The mosquito survival was recorded at 6 hours after materials inoculation. Control mosquitoes were treated with PBS. The number at the top of bar represents the number of live mosquitoes/total mosquitoes in a treatment. doi:10.1371/journal.pone.0022786.g001 IIH into mosquitoes. As seen in Figure 2b, a minimum of 25 ng pAc-GFP per mosquito produced substantial GFP expression in Ae. aegypti.
Anopheles spp. transmit Plasmodium, the etiologic agent of malaria [7,19]. We therefore extended this approach to An. gambiae, and assessed whether this method was effective in another mosquito species that transmits important diseases. Cellfectin IIH also significantly increased GFP expression in An. gambiae ( Figure 2c). Moreover, the amount of GFP protein was 100-500 folds greater in An. gambiae than in Ae. aegypti (Figure 2c). In the titration analysis, GFP expression was dramatically decreased when the inoculated DNA was less than 50 ng per mosquito (Figure 2d). To determine the distribution of expression, GFP was assessed in mosquito tissues by microscopy. Fluorescence was observed in the An. gambiae salivary glands and midgut. Figure S1 shows fluorescence at days 3 and 6 post injection in the liposome-DNA treated groups, confirming that Cellfectin IIH facilitates the in vivo expression of DNA in mosquitoes.

In vivo overexpression of AeTEP-1 influences flaviviral infection
The Thioester-containing protein (TEP) gene exists in numerous mosquito species, and forms a distinct clade of a multigene family. TEP includes several vertebrate complement domains, and is considered a complement-like protein in mosquitoes [20]. A TEP gene from An. gambiae mediates recognition and killing of Plasmodium berghei ookinetes [21]. In invertebrates, TEPs are a component of the innate immune system to pathogens. The role of TEPs in flavivirus infection is, however, unknown. We therefore determined the role of Ae. aegypti TEP (AeTEP) genes in DENV and WNV infection by RNA interference, and validated the phenotype by in vivo overexpression.
We identified 6 TEP homologues from the Ae. aegypti genome and designated them as AeTEP-1 to AeTEP-6. AeTEPs share 21-39% amino acid identity with Anopheles TEP-1. Of these 6 genes, 5 AeTEPs were expressed in the adult female Ae. aegypti (Table S1). AeTEPs genes were silenced in Ae. aegypti by gene-specific dsRNA. Compared to the mock group, expression of all the AeTEPs was significantly reduced from days 3 through 9 ( Figure S2a, S2b, S2c, S2d, S2e). WNV and DENV were therefore inoculated into mosquitoes on day 3 following dsRNA-treatment, and the viral load on day 6 post-infection was quantified by RT-QPCR. The viral loads in AeTEP-1 and AeTEP-2 silenced mosquitoes were enhanced by 2-3 fold when compared with those of mock-treated mosquitoes ( Figure 3 and Figure S3), suggesting that AeTEPs may play an important role in resistance to flaviviral infection of Ae. aegypti.
To further confirm the function of AeTEPs in flavivirus infection, we selected AeTEP-1 and AeTEP-3 for in vivo overexpression investigation. Meanwhile, a truncated AeTEP-1 (AeTEP-1-BstBI) was generated as a control for functional study. These genes were cloned into pAc5.1/V5-His A vector and the recombinant plasmids were named as pAc-AeTEP-1, pAc-AeTEP-1-BstBI and pAc-AeTEP-3 respectively. Then, these genes were expressed in Drosophila S2 cells. Immunoblot confirmed the expression of a ,200 kD recombinant protein encoded by the AeTEP-1 (Figure 4a) and ,100 kD proteins by AeTEP-3 and truncated AeTEP-1-BstBI genes ( Figure S4a). Subsequently, these recombinant plasmids were microinjected into mosquitoes with Cellfectin IIH to assess the efficiency of in vivo over-expression. The same amount of empty pAc vector or pAc-AeTEP plasmids alone served as controls. The mosquitoes were collected at various time points after microinjection to isolate the total RNA for the detection of AeTEPs. AeTEP-1 (Figure 4b) and AeTEP-3 expression ( Figure S4b) were significantly higher in the Cellfectin IIH transfected groups than in controls, at all time points. Then, the AeTEPs transfection mosquitoes were used to elucidate the role in flaviviral infection. The Cellfectin IIH alone and empty vector inoculated mosquitoes served as negative controls. Compared to the controls, the DENV burden was significantly decreased in the AeTEP-1 expressing group (p,0.0001), but not in the AeTEP-3 or AeTEP-1-BstBI expressing groups (Figure 4c), suggesting AeTEP-1 has a specific role against DENV infection. However, the overexpression of AeTEP-1 did not influence WNV infection of mosquitoes ( Figure  S4c), implying that the AeTEP-1 function is ambiguous in WNV infection. In conclusion, our results demonstrate the usefulness of this in vivo transfection strategy for functional study of mosquito genes.

Discussion
Mosquitoes transmit numerous infectious diseases. Development of techniques to rapidly manipulate mosquitoes will facilitate our understanding of the pathogenesis of vector-borne diseases. Genetic studies on mosquitoes are generally conducted in cellculture. These studies provide important information; however, understanding the pathogenesis of vector-borne microbes requires in vivo studies. Here, we used a transfection reagent to introduce plasmid DNA into mosquitoes through thoracic microinjection. This procedure enabled us to successfully express an important gene in vivo. Our results demonstrated that the method worked in both An. gambiae and Ae. aegypti, albeit better in An. gambiae. However, Cellfectin IIH did not efficiently transfect plasmid DNA into Ixodes scapularis (data not shown), a tick that transmits diverse pathogens including the agent of Lyme disease. These data indicated that different arthropod species has varying sensitivity to this liposome-based transfection.
In this study, we assessed the antiviral function of the family of Ae. aegypti complement-like factors using RNAi and in vivo transfection approaches. The complement system combats viral infection in mammals [22]. Complement components are associated with viral surface proteins during early infection and are activated to promote the destruction of viral particles and initiate innate immune responses [23,24]. In flaviviral infection, complement components can activate B cells, assist in B cell maturation, and enhance viral neutralization in an antibody-mediated manner [25]. Recent studies have shown that a complement-like factor, Thioester-containing protein (TEP) genes, exists in An. gambiae to mediate parasite killing [21,26]. To determine the role of TEP genes in flavivirus infection, we identified five expressed TEP homologues in female Ae.aegypti. Silencing certain AeTEP by RNAi led to a significant enhancement of WNV and DENV burdens. To further confirm the role of AeTEPs in flaviviral infection, we chose AeTEP-1 and AeTEP-3, and also generated an AeTEP-1 truncation, for in vivo over-expression. Compared to liposome alone and vector controls, AeTEP-1 expression dramatically decreased the DENV infection, however, the same phenomenon was not observed in AeTEP-3 and truncated AeTEP-1-transfected mosquitoes. Together with the gene silencing results, we concluded AeTEP-1 plays a role in resistance to DENV infection of mosquitoes. Ae. aegypti is a primary mosquito species for DENV infection and transmission in nature. Similar to mammals, mosquitoes also evolve specific defense mechanism, including complement-like factors, to efficiently control DENV infection. AeTEP-1 silencing also showed a strong resistance in WNV infection. However, the in vivo transfection of AeTEP-1 did not elucidate a similar phenomenon, suggesting AeTEPs-based resistance system is important, but not essential in WNV infection. In summary, we have developed an efficient method to express an exogenous gene in mosquitoes in vivo. The technique was validated by assessing the importance of an Ae. aegypti TEP gene in flaviviral infection. This method is potentially useful in mosquitopathogen interaction studies in many ways: validation of RNAi silencing results, in vivo biochemical studies and mosquito genetic investigations. This approach should facilitate an understanding of the interaction between mosquitoes and the infectious agents that they transmit.

Mosquitoes, cells and viruses
Ae. aegypti and An. gambiae mosquitoes were maintained in a sugar solution at 27uC and 80% humidity according to standard rearing procedures [27,28]. The Aedes albopictus C6/36 cell line was grown at 30uC in minimal essential medium (MEM) for WNV and DENV production. The Drosophila melanogaster S2 cell line was cultured in Schneider's Medium. All mediums were supplemented with 10% heat-inactivated fetal bovine serum, 1% L-glutamine, 100 IU of penicillin and streptomycin per ml. WNV strain 2741, which is identical at the protein level to the NY99 strain [29], and DENV-2 (DENV New Guinea C strain) were used in this study. The viruses for in vivo experiments were titrated in mosquitoes through thoracic microinjection. The procedure was described in a previous study [11].

Cellfectin IIH in vivo transfection
Cellfectin IIH (Invitrogen, Cat No# 10362-100) was mixed with S2 Schneider's medium (serum/antibiotics-free) as 1:1 ratio (vol/vol). The mixture was kept at room temperature for 10 min. 2 vols of the diluted plasmid DNA were then combined with diluted Cellfectin IIH and gently mixed. The transfection mixture was further incubated for 30 min at room temperature prior to thoracic microinjection.
Microinjection dsRNA synthesis was performed as described previously [30]. The primers are shown in the Table S1. Plasmid DNA was isolated by an EndoFree Plasmid Maxi Kit (Qiagen, Cat. No# 12362). For thoracic microinjection, 1 week old adult female mosquitoes were kept on ice for 15 min, and then transferred to a cold tray to receive a systemic injection of dsRNA or plasmid DNA into the hemocoele. In the gene silencing studies, 2 mg of dsRNA/300 nl in PBS was microinjected into the thorax of each mosquito. Following a 3-day recovery period, the mosquitoes were microinjected with WNV or DENV 10 M.I.D 50 /300 nl (50% Mosquito Infective Dose) [11] for functional studies. In the in vivo transfection, 200 ng plasmid DNA/300 nl mixture was inoculated into each mosquito. The mosquitoes were further challenged by 10 M.I.D 50 of DENV and WNV at day 6 following the inoculation of the transfection mixture.

Quantitative PCR
The specific RNAs of the WNV-E gene, DENV-E gene and AeTEPs gene were quantified by RT-QPCR. The primers and probes for WNV-E and DENV-E gene were described previously [11]. The primers for AeTEPs are shown in Table S1. The amount of viruses were normalized using Ae. aegypti actin (AAEL011197).

ELISA
Three mosquitoes were pooled together and homogenated in 300 ml T-PER tissue protein extraction buffer (Thermo Scientific, Cat. No# 78510) with a Pestle Grinder System (FisherSci, Cat. No# 03-392-106). The samples were centrifuged for 10 min at 4uC and the supernatant were collected for ELISA assay. The experimental details are described in the product manual of GFP ELISA kit (Cell Biolabs, Cat. No# AKR-121).

Isolation and imaging of mosquito tissues
Salivary glands and midgut were dissected as previously described [31]. Tissues were isolated, placed on sialylated slides (PGC Scientific, Gaithersburg, MD), washed in PBS, and fixed in 4% PFA at 37uC for 1 hr. Slides were imaged with a Single-Track mode of a Zeiss LSM 510 meta confocal microscope. Figure S1 GFP fluorescence in mosquito tissues. The tissues were dissected at day 3 and day 6 post microinjection. The fluorescence was detected by confocal microscopy. (a) salivary glands; (b) midgut. Images were examined using a Zeiss LSM 510 meta confocal 106objective lens. (TIF) Figure S2 AeTEPs RNAi efficiency. The mock group was treated with the same amount of GFP dsRNA. AeTEPs-dsRNA or GFP-dsRNA treated mosquitoes were sacrificed to isolate total RNA at 3 days and 9 days post-inoculation. mRNA of AeTEPs was determined by SYBR GreenH QPCR, and normalized using Ae. aegypti actin.