Development of large-scale mosquito densovirus production by in vivo methods

Background Mosquito-borne diseases (MBDs) cause a significant proportion of the global infectious disease burden. Vector control remains the primary strategy available to reduce the transmission of MBDs. However, long-term, wide-scale and large-scale traditional chemical pesticide application has caused significant and increased negative effects on ecosystems and broader emerging insecticide resistance in vectors; therefore, the development of a novel alternative approach is urgently needed. Mosquito densoviruses (MDVs) are entomopathogenic viruses that exhibit a narrow host range and multiple transmission patterns, making MDVs a great potential bioinsecticide. However, the application process has been relatively stagnant over the past three decades. The major obstacle has been that viruses must be produced in mosquito cell lines; therefore, the production process is both expensive and time-consuming. Methods In our study, two wild-type (wt) MDVs, AaeDV and AalDV-3, and a recombinant rAaeDV-210 were used to infect the Aag2 and C6/36 mosquito cell lines and the 1st–2nd-instar and 3rd–4th-instar larvae of Ae. albopictus, Ae. aegypti and Cx. quinquefasciatus. Viral titers and yields in cells, media, larvae and rearing water and total viral yield were evaluated. Three kinds of virus displayed higher maximum virus titers in vivo than in vitro, and they displayed higher maximum viral yields in rearing water. Results The three viruses displayed higher total maximum viral yields in C6/36 cells than in Aag2 cells. The three viruses displayed higher total maximum viral yields in Aedes mosquitoes than in Culex mosquitoes. Higher viral yields were produced by 1st–2nd-instar larvae compared to 3rd–4th-instar larvae. The recombinant viruses did not display significantly lower yields than wt viruses in nearly all samples. In summary, by using 100 1st–2nd-instar Aedes mosquito larvae in 200 ml of rearing water, more than 1013 genome equivalents (geq) MDV yield can be obtained. Conclusions Considering the lower production cost, this in vivo method has great potential for the large-scale production of MDVs. MDVs exhibit promising prospects and great potential for mosquito control in the future. Electronic supplementary material The online version of this article (10.1186/s13071-019-3509-5) contains supplementary material, which is available to authorized users.


Background
Mosquito-borne diseases (MBDs), such as malaria, dengue fever, yellow fever, chikungunya, Zika, hemorrhagic fevers and lymphatic filariasis, cause a significant fraction of the global infectious disease burden; indeed, nearly 700 million people worldwide are infected with MBDs each year, resulting in over one million deaths [1]. Vector control remains the primary strategy available to reduce the transmission of MBDs. Synthetic chemical insecticides have been traditionally used as a standard method for vector control and play a key role in public health as part of integrated mosquito management (IMM) programs for disease control and prevention and dramatically reduce the risk of MBDs, particularly in the case of malaria [2]. However, long-term, wide-scale and largescale traditional chemical pesticide application has certainly shown significant and increased negative effects on ecosystems [3]. Additionally, the broadly emerging insecticide resistance of vectors poses a major threat to effective prevention and control of MBDs. In view of these pending issues, the development of novel alternative approaches that can be applied to mosquito control is urgently needed. Biological control of vectors using insect specific pathogens or symbiotic microorganisms, such as fungi, bacteria or viruses, can reduce vector populations or vectoral capacity, and hence disease transmission, and be produced as eco-friendly, highly effective, specific, economic, biodegradable bioinsecticides [4].
Entomopathogenic viruses have been identified in more than one thousand species from at least 13 different insect orders [5]. These insect pathogenic viruses can be divided into two categories: those that produce occlusion bodies in the insect host cells are occluded viruses (OVs), whereas those that are not occluded in protein bodies are no-occluded viruses (NOVs) [6]. The most common OVs of mosquitoes are the nuclear polyhedrosis viruses (NPVs) (family Baculoviridae), such as Culex nigripalpus nuclear polyhedrosis virus (CuniNPV) and cytoplasmic polyhedrosis viruses (CPVs) (family Reoviridae, genus Cypovirus) [7,8]. The main NIVs in mosquitoes are mosquito iridoviruses (MIVs) (family Iridoviridae, genus Chloriridovirus) and densoviruses (DVs) [9].
Mosquito densoviruses (MDVs) are currently included in the genus Brevidensovirus of the subfamily Densovirinae in the family Parvoviridae. The virion is a small icosahedral, non-enveloped DNA virus with a diameter of 20 nm. The viral genome is a linear, single-stranded DNA (ssDNA) molecule approximately four kilobases long. MDVs exhibit strong host specificity that is restricted to members of the family Culicidae in the order Diptera. MDVs can invade and proliferate in multiple organs and tissues of mosquitoes, and infection with MDVs can cause larval death or deformation [10]. Moreover, MDVs extend the larval stage and decrease the lifespan and body size of adults [11]. Furthermore, their pathogenicity can be significantly improved by genetic engineering techniques, such as altering the vector used to express the insect-specific toxin, gene or specific shRNA or artificial miRNA that targets genes essential for the development, growth or physiology of mosquitoes [12][13][14]. MDVs can be transmitted in mosquito populations through horizontal transmission in larval habitats and/or by vertical transmission. Thus, these biological and pathogenic characteristics confer the potential for MDVs as biological control agents [11]. Considering that the Biokiller Cockroach Killer Bait Gel, the first densovirus biopesticide product, which mainly contains Periplaneta fuliginosa densonucleosis virus (PfDNV, genus Ambidensovirus, family Parvoviridae), was commercialized in China [15], the use of MDVs and/or recombinant MDV biopesticides is within reach.
Despite promising application prospects, the standardization of the large-scale production of MDVs remains a major barrier for the progression from laboratory to field. In this study, we analyzed the accumulation of wild-type and recombinant MDVs in mosquito cell lines, including the Ae. albopictus C6/36 cell lines and Ae. aegypti Aag2 cell lines, as well as in three different species of mosquito larvae, Ae. aegypti, Ae. albopictus and Cx. quinquefasciatus. The maximum virus titer and viral yield were evaluated and compared. Finally, we developed a novel and more economical method for the mass production of MDVs in mosquito larvae, which will provide sufficient amounts of virus for bioassays, future moderate-scale field testing and even large-scale field application.

Mosquito cell transfection and virus production
pUCA and pUCP were the infectious clones of AaeDV (GenBank: M37899) and AalDV-3 (or APeDNV, Gen-Bank: AY310877) that contained the full genomic DNA of AaeDV (3981 nt) and AalDV-3 (4006 nt), respectively, and were kindly provided by Professor Erica Suchman and Jonathan Carlson (Colorado State University, Fort Collins, CO, USA). pUCA-210 is an infectious clone of non-defective recombinant densovirus vectors that can express endogenous miRNA-210 in Ae. albopictus. The construction of plasmids has been previously described in detail [16,17].
Wild-type viruses AaeDV and AalDV-3 and recombinant virus rAaeDV-210 were generated by transfecting the corresponding infectious clones pUCA, pUCP and pUCA-210 into C6/36 cell lines according to the manufacturer's protocol. Before transfection, cells were plated at a density of 2.5 × 10 6 cells per 25 cm 2 flask. The cells were incubated at 28 °C until the cells were 80 to 90% confluent. The transfection of plasmids was performed using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer's protocol. Supercoiled plasmids used for transfection were prepared using a GeneJET Endo-Free Plasmid Max-iPrep Kit (Thermo Fisher Scientific). Cells were typically harvested 5 days post-transfection using cell scrapers, lysed by repeat freezing-thawing treatment, and then centrifuged for 10 min at 1000×g. The supernatants were kept as viral stocks.

Viral accumulation in mosquito cells
To explore the proliferation capability of AaeDV, AalDV-3 and recombinant virus rAaeDV-210 in mosquito cells, C6/36 and Aag2 cells were seeded in 24-well plates (C6/36 cells: 1 × 10 5 cells/well, Aag2 cells: 1 × 10 6 cells/well) and incubated with 1 ml of media containing virus at a final concentration of 3.00 × 10 9 genome equivalents per ml (geq/ml) at 28 °C. After incubation for 24 h, the medium was removed, and the cells were washed 3 times with fresh RPMI-1640 medium (Schneider's Drosophila medium for Aag2 cells) to remove residual virus. Then, 1 ml of fresh medium was added, and the cells were returned to the incubator for regular culture.
Viral titers in the cell culture supernatant and in the cells themselves were determined by quantitative realtime PCR (qPCR). First, to determine the titer of the extracellular viruses, samples from the cell culture were collected at serial time points post-infection [0, 2, 4, 6, 8 and 10 days post-infection (dpi)] and then centrifuged for 10 min at 1000× g. Supernatants were used to extract viral genomic DNA using a Viral DNA Kit (Omega Biotek, Norcross, GA, USA). To determine the titer of the intracellular virus, the supernatants were removed, and the infected cells were washed 3 times with 1× PBS. The cells were collected at serial time points post-infection (0, 2, 4, 6, 8 and 10 dpi); the genomic DNA was then isolated for qPCR.

Larval infection and densovirus detection
To explore the percentage of infected mosquito larvae when exposed to MDV, the 1st-2nd-instar and 3rd-4thinstar larvae (n = 100 per group) of Ae. albopictus, Ae. aegypti and Cx. quinquefasciatus larvae were infected with AaeDV, AalDV-3 and rAaeDV-210 by exposure to 1 ml of sterile water in a beaker containing virus with a final concentration of 2.00 × 10 10 geq/ml. After incubation for 24 h at 28 °C, the larvae were washed three times with deionized water and then transferred to pans containing 200 ml sterile water. After 1 dpi, larvae were washed three times with deionized water again, and the DNA was isolated using a TIANcombi DNA Lyse&Det PCR Kit (Tiangen Biotech, Beijing, China). PCR amplification was then performed using Maxima Hot Start Green PCR Master Mix (Thermo Fisher Scientific) with gene-specific primers. The final washed water was used as a control. Ae. albopictus ribosomal protein 7 gene (AalRpS7, GenBank: JN132168), Ae. aegypti RpS7 (AaeRpS7, VectorBase accession number: AAEL009496) and Cx. quinquefasciatus RpL8 (CquRpL8, GenBank: XM_001841875) were used as controls, respectively.

Viral accumulation in mosquito larvae
The 1st-2nd-instar and 3rd-4th-instar Ae. albopictus, Ae. aegypti and Cx. quinquefasciatus larvae (n = 100 per group) were exposed to AaeDV, AalDV-3 and recombinant virus rAaeDV-210 at a concentration of 2.00 × 10 10 geq/ml in 1 ml of sterile water. The blank control group was exposed to virus-free C6/36 cell culture medium under conditions identical to the treatment groups. After incubation for 24 h at 28 °C, the larvae, along with water, were transferred back to the pans, and then sterile water was added until the viral concentration was 1.00 × 10 8 geq/ml; larvae were fed regularly. The 1st-2nd-instar larval rearing water was collected at serial time points post-infection (1, 3, 5, 7, 9 and 11 dpi), whereas the 3rd-4th-instar larval rearing water was collected at 1, 3, 5 and 7 dpi. The virus in the rearing water was titered by qPCR. The larvae were also collected at the same time point, and the virus in vivo was also titrated by qPCR.

Quantitative real-time PCR
Non-encapsidated genomic DNA was removed by treatment with TURBO DNase (Ambion, Austin, TX, USA) at 37 °C for 1 h. The total encapsidated genomic DNA was extracted using a Viral DNA Kit (Omega Biotek). The virus genome copy numbers in cell culture media, cells, rearing water and larvae in vivo were determined by a SYBR green-based real-time quantitative PCR assay with FastFire qPCR PreMix (Tiangen Biotech). A series of known concentrations of linear plasmids with pUCA were used to construct a standard curve. The set of primers used annealed to the conserved region of the viral NS1 gene. These procedures were performed essentially as described previously [18]. The sequences of primers used in the PCR and qPCR are shown in Additional file 1: Table S1.

Statistical analysis
Comparison of the percentage of infected larvae between groups was performed using a Chi-square test. Virus titer and viral yield in cells, cell culture media, larvae and rearing water and total viral yield at different times and in different groups were compared using one-way analysis of variance (ANOVA) followed by Fisher's least significant difference test (LSD). Student's t-test was performed to determine differences between two groups. The P-value for statistical significance is defined as P < 0.05. SPSS software v. 19.0 (SPSS Inc. Chicago, IL, USA) was used for data analysis.
Moreover, the yield sums of three kinds of viruses in cells and medium were quantified. For virus production, viral yield reached a peak at 6 or 8 dpi using Aag2 cells, whereas viral yield gradually increased over a prolonged time in C6/36 cells. The total yield of AaeDV, AalDV-3 and rAaeDV-210 in C6/36 cells was nearly 5, 115 and 5 times higher than that of Aag2, respectively. In addition, the maximum yield of recombinant rAaeDV-210 virus did not exhibit a significant difference compared with its original wild-type AaeDV (Fig. 3) both in C6/36 cells (t (4) = 1.58, P = 0.189) and Aag2 cells (t (4) = 1.95, P = 0.124).

The trend of viral titers in infected larval rearing water
The viral titers in the rearing water of three species of larvae were also quantified, and the samples were collected at the same time points as described for larvae. The result is exciting; the virus infection was prolonged, and all three viruses showed a similar consecutive growth pattern in both 1st-2nd-instar and 3rd-4th-instar larvae to that of all three species in vivo. Although the highest concentration of virus in the water was slightly lower than that in vivo, viral titers still presented the highest values at 11 dpi in rearing water of Ae. albopictus 1st-2nd-instar larvae (Fig. 5a), at 5 dpi (AaeDV) and 7 dpi (AalDV-3 and rAaeDV-210) in rearing water of 3rd-4th-instar larvae (Fig. 5b); at 11 dpi in rearing water of Ae. aegypti 1st-2nd-instar larvae (Fig. 5c), at 7 dpi in rearing water of 3rd-4th-instar larvae (Fig. 5d); at 11 dpi (AaeDV and AalDV-3) and 9 dpi (rAaeDV-210) in Cx. quinquefasciatus 1st-2nd-instar larvae (Fig. 5e), at 5 dpi (AaeDV) and 7 dpi (AalDV-3 and rAaeDV-210) in rearing water of 3rd-4th-instar larvae (Fig. 5f ).
Regarding the total product, the maximal viral yields in the 1st-2nd-instar and 3rd-4th-instar larvae of all three species of mosquitoes were compared. The results showed that for all viruses, significantly higher viral yields were obtained using 1st-2nd-instar larvae rather than 3rd-4th-instar larvae, regardless of whether in vivo (Fig. 7a), in the water (Fig. 7b) or the combined total (Fig. 7c). Moreover, when comparing the 1st-2nd-instar larvae of the three mosquito species, higher yields were detected in Aedes mosquitoes than in Culex mosquitoes in vivo, in the water body or combination. However, we did not detect significant differences between Ae. aegypti and Ae. albopictus, except for the rAaeDV-210, which showed higher yields using Ae. aegypti 1st-2nd-instar larvae.

Discussion
Chemical pesticides have been widely used over the past 60 years in agricultural pest control and in response to epidemics [3,19]. Today, over 2.3 billion kilograms of pesticides are used annually worldwide [20]. However, the excessive use of chemical pesticides is accompanied by a growing general concern for the negative impacts on the ecosystem and human health. Globally, chemical insecticides directly contribute to more than 100,000 deaths and two million hospital admissions every year [21]. Additionally, intensive prophylactic application causes the pests to rapidly develop resistance to most traditional insecticides and causes widespread concern over Fig. 5 The accumulation of MDV in larval rearing water. The viral yields from larval rearing water were quantified at serial time points. Each time point represents the average titer obtained from three independent experiments with the respective standard deviations. The titer of AaeDV, AalDV-3 and recombinant rAaeDV-210 in Ae. albopictus 1st-2nd-instar larval rearing water (a) and 3rd-4th-instar larval rearing water (b); Ae. aegypti 1st-2nd-instar larval rearing water (c) and 3rd-4th-instar larval rearing water (d); and Cx. quinquefasciatus 1st-2nd-instar larval rearing water (e) and 3rd-4th-instar larval rearing water (f) chemical pesticide abuse [22][23][24], resulting in the imperative requirement of alternative and ecologically friendly biocontrol approaches.
Entomopathogenic viruses have a narrow host spectrum, are non-toxic to humans and non-target organisms, have a long shelf life and are eco-friendly; these favorable properties make entomopathogenic viruses a potentially great bioinsecticide. Currently, more than 240 insect virus isolates have been isolated from 196 species from 35 families and seven orders in China. Since the first viral insecticide, Helicoverpa armigera nucleopolyhedrovirus (HaNPV), was registered in 1993 [25], a total of 69 products have been authorized as commercial pesticides by the Ministry of Agriculture of China (MAC) through June 2018 [26]. In the last five years, it has been estimated that 1600 metric tons of viral insecticidal formulations were produced annually, representing nearly 0.2% of the total pesticide yield in China [27].
With the widespread use of viral insecticides in agriculture, the process of propagating viruses for vector control is tortuous, especially in mosquitoes. At present, the major entomopathogenic viruses that display pathogenicity to mosquitoes are NPVs, CPVs, MDVs and MIVs. Although baculovirus application has been successful for agricultural insects, the host specificity of CuniNPV is confined to species of the mosquito genus Culex [7]. Moreover, the virus' slow speed of action limits its application in mosquitoes [28]. MIVs showed only moderate infection in mosquito larvae, and mortality rates were typically lower than 25% in first-and secondinstar larvae [29]. Although relatively higher rates were observed in CPV-infected larvae, the weak pathogenicity was the major obstacle for their development and application. For example, even heavily infected larvae could often survive to pupation and emergence [29]. As a result, the only prospect for potential commercial production is MDV. MDVs have displayed several promising characteristics, such as a higher infection rates for all developmental larval stages, especially molting larvae. Furthermore, newly proliferated virus can be sustained and released into water from the infected and dead mosquito larvae, thereby persisting or increasing the viral The maximum viral yield obtained from larval rearing water. c The maximum viral yield sum of the combination titers and allowing for horizontal transmission to uninfected or newly hatched larvae. Additionally, MDVs can be transmitted vertically to their offspring by infected females [30][31][32]. We previously reported the development of a non-defective recombinant AaeDV, which can express small interference RNAs using an intronic sRNA expression strategy that lays the foundation for further enhancing the virulence of MDVs for bioinsecticide use [12,18]. AaeDV as a product (Viroden) was evaluated in small-scale field studies in the former Soviet Union in 1979 [30]. However, to date, substantial progress has not been made. The major obstacle was that viruses must be produced in mosquito cell lines, so the production process was therefore both expensive and time-consuming. A suspension culture of C6/36 cells in serum-free protein-free Sf-900 II media was previously developed by Suchman et al. [33], and a higher density of cells was collected than that from the conventional process. However, the maximum AaeDV titer only reached 3.8 × 10 8 geq/ml for AeDNV contained in 45 ml of culture medium. Moreover, the price of commercial Sf-900 II was too high to mass produce for large-scale field applications. Considering the effect of MDVs on the mosquito larvae in a concentration-dependent manner, applied viral titers should also be taken into account [34][35][36].
In our study, two wt MDVs, AaeDV and AalDV-3, and a recombinant rAalDV-210 were used as "seeds"; two mosquito cell lines, Aag2 and C6/36, and the larvae of three species, Ae. albopictus, Ae. aegypti and Cx. quinquefasciatus, were used as the viral production "factories", and the viral yield was compared. Two criteria, virus titer and viral yield, were used to evaluate the production capacity of these "factories". Considering that a feature of MDVs is that they can sustain release into the culture medium and rearing water from the infected cells, we also counted the two criteria in vivo, in vitro and combined. Our results showed that (i) the three viruses displayed higher maximum virus titers in cells and larvae than in culture medium and rearing water; (ii) the three viruses displayed higher maximum viral yields in medium and water; (iii) the three viruses displayed higher total maximum viral yields in C6/36 cells than in Aag2 cells; (iv) the three viruses displayed higher total maximum viral yields in Aedes mosquitoes than in Culex mosquitoes; (v) higher viral yields were produced by infection of 1st-2nd-instar larvae compared to 3rd-4th-instar larvae; and (vi) the recombinant virus did not display significantly lower yields than that of the wt viruses in nearly all the samples. In summary, by using 100 1st-2nd-instar Aedes mosquito larvae in 200 ml of rearing water, more than 10 13 Fig. 8 The percentage of MDV infected larvae. Viral DNA was detected by the traditional PCR method with gene-specific primers for the genomic conserved region. The percentage of MDV infected larvae was determined using AaeDV, AalDV-3 and rAaeDV-210 to infect the 1st-2nd-instar and 3rd-4th-instar larvae of Ae. albopictus, Ae. aegypti and Cx. quinquefasciatus. The data are representative of three independent experiments, and the values are expressed as the mean ± SD. Different letters with the same color above bars represent significant differences in relative expression levels at the P ≤ 0.05 level