Virus-derived DNA drives mosquito vector tolerance to arboviral infection

Abstract

Mosquitoes develop long-lasting viral infections without substantial deleterious effects, despite high viral loads. This makes mosquitoes efficient vectors for emerging viral diseases with enormous burden on public health. How mosquitoes resist and/or tolerate these viruses is poorly understood. Here we show that two species of Aedes mosquitoes infected with two arboviruses from distinct families (dengue or chikungunya) generate a viral-derived DNA (vDNA) that is essential for mosquito survival and viral tolerance. Inhibition of vDNA formation leads to extreme susceptibility to viral infections, reduction of viral small RNAs due to an impaired immune response, and loss of viral tolerance. Our results highlight an essential role of vDNA in viral tolerance that allows mosquito survival and thus may be important for arbovirus dissemination and transmission. Elucidating the mechanisms of mosquito tolerance to arbovirus infection paves the way to conceptualize new antivectorial strategies to selectively eliminate arbovirus-infected mosquitoes.

Figure 1. Mosquito cells produce host reverse transcriptase-dependent arbovirus-derived DNA.

(a) Schematic of CHIKV viral genome. Top arrows indicate the position of the genomic and subgenomic promoters. Bottom arrows indicate the position of the primers used for vDNA detection. (b) Kinetics of vDNA synthesis. C6/36, U4.4 and Aag2 cells were infected with CHIKV at a MOI of 0.1 and cells were harvested at the indicated time points. Cells were analysed by PCR (upper panel) for vDNA detection. RT-PCR (lower panel) was used to follow viral infections. Non-infected cells (n.i) were used as a negative control and cellular 18S rRNA was used as a housekeeping gene loading control. (c) AZT inhibits vDNA synthesis in vitro. C6/36, U4.4 and Aag2 cells were treated with increasing concentration of AZT for 2 days. At the indicated time point, cells were harvested and vDNA and RNA were assessed as described in b. (d–e) Endogenous reverse transcriptase activity in insect cells. (d) Dose-dependent AZT inhibition was tested in insect cell extracts and (e) quantification of reverse transcriptase inhibition expressed as an activity percentage of non-treated extracts. Heat inactivated samples (heat) were used as negative controls. Drosophila S2 cells were used as a positive control. Each experiment was completed at least 3 times. Error bars correspond to the s.d. t-test with Welch's correction was used to determine statistical significance compared with the untreated control as a reference (**P<0.01; ***P<0.001; ****P<0.0001).

Figure 2. Chikungunya vDNA is required for small viral RNA production in cultured cells.

vDNA contributes to the production of vpiRNAs in mosquito cell lines. Infected U4.4 cells were (a) left untreated (clear bars) or (b) treated with 5 mM AZT (green bars) for 3 days. Graph represents the size distribution of the total number of CHIKV-specific small RNA reads (ranging from 18 to 33 nts) normalized to the total number of reads. (c) vsiRNAs in U4.4 cells show homogenous coverage over the entire genome. (d) vsiRNA coverage in AZT-treated cells. Uncovered regions are represented as grey lines. The coverage of vpiRNAs on the CHIKV genome in (e) untreated cells or (f) AZT-treated cells. Most of the CHIKV vpiRNA reads belong to the 3′-terminal region on the viral genome. The sense and anti-sense small RNAs are in red and green, respectively. (g-h) Relative nucleotide frequency per position of the 27–29 nt viral small RNAs that map to the sense and anti-sense strand of the viral genome, red and green respectively. The intensity varied in correlation with the frequency. A nucleotide bias (U1 and A10) is observed. (i,j) Accumulation of viral and cellular small RNAs in U4.4 cells 3 days post CHIKV infection in the presence (AZT pos) or absence (AZT neg) of AZT, assessed as (i) mapping of small RNAs corresponding to CHIKV vsiRNA (green), endo-siRNA (purple) belonging to the gene GAPW01000199 or to miRNA (blue). For miRNA each dot represents one miRNA. For siRNAs each dot represents the coverage of a region of 20 bases of the target RNA. The lines for miRNA and endo-siRNA are superposed. (j) Mapping of small RNA corresponding to CHIKV vpiRNA (orange), endo-piRNA (red) belonging to the gene GAPW01000199 or to miRNA (blue). For miRNAs each dot represents one miRNA. For piRNAs each dot represents the coverage of a region of 20 bases of the target RNA. For (i,j) lines represent the linear trendline of each set of values. The equation and R² value of each regression are also mentioned.

Figure 3. Arbovirus-infected mosquitoes produce vDNA that contributes to the production of vsiRNAs in vivo.

(a) Ae. albopictus mosquitoes were infected with CHIKV and analysed for vDNA 9 days post infection. Three non-infected (n.i.) and nine CHIKV-infected mosquitoes were tested. 18S rRNA was used as a control. (b) Mosquitoes were untreated (clear bars) or (c) treated with 5 mg ml⁻¹ AZT (green bars) and harvested 3 dpi. Five whole mosquitoes were pooled from each condition for the generation of small RNA libraries. Graphs represent the size distribution of the total number of CHIKV-specific small RNA reads (corresponding to the positive and negative strand orientation of the viral genome) ranging from 18 to 33 nts normalized by the total number of reads. (d–g) Panels show the coverage of CHIKV genome at 3 dpi using the 21 or 27–29 nts-long small RNAs at the different conditions (untreated: d and f, or AZT-treated: e and g). The sense and anti-sense small RNAs are in red and green, respectively. Grey lines represent uncovered regions. (h,i) Relative nucleotide frequency per position of the 27–29 nt viral small RNAs that map to the sense and anti-sense strand of the viral genome, red and green respectively. The intensity varied in correlation with the frequency. No nucleotide bias (U1 and A10) is observed. (j,k) Accumulation of viral and cellular small RNAs in Ae. albopictus 3 days post CHIKV infection in the presence (AZT pos) or absence (AZT neg) of AZT, assessed as (j) mapping of small RNAs corresponding to CHIKV vsiRNA (green), endo-siRNA (purple) belonging to the gene GAPW01000199 or to miRNA (blue). The lines for miRNA and endo-siRNA are superposed. (k) Mapping of small RNA corresponding to CHIKV vpiRNA (orange), endo-piRNA (red) belonging to the gene GAPW01000199 or to miRNA (blue). For (j,k) each blue dot represents one miRNA. For siRNAs and piRNAs each dot represents the coverage of a region of 20 bases of the target RNA. Lines represent the linear trendline of each set of values. The equation and R² value of each regression are also mentioned.

Figure 4. vDNA is required for mosquito tolerance to CHIKV infection.

(a) Survival curve of CHIKV-infected mosquitoes treated daily with 0, 5 or 10 mg ml⁻¹ of AZT. After the infectious blood meal survival of mosquitoes was monitored daily for 14 days. Continuous lines show the lifespan of AZT-treated and infected mosquitoes while dotted lines show lifespan of AZT-treated but non-infected (n.i.) mosquitoes. Grey lines: 0 mg ml⁻¹ AZT; blue lines: 5 mg ml⁻¹ AZT and green lines: 10 mg ml⁻¹ AZT. At least 40 mosquitoes were used for each condition. P-values were calculated using a Gehan–Breslow–Wilcoxon test using uninfected mosquitoes with the same AZT treatment as a reference. (b,c) Frequency of vDNA in mosquitoes treated with 10 mg ml⁻¹ AZT (green, squares) or untreated mosquitoes (grey, circles) in (b) bodies or (c) legs and wings. (d) Individual mosquito viral titres quantified by plaque assay. Viral titres were determined in bodies (left panel), legs and wings (middle panel) and saliva (right panel) for untreated mosquitoes (grey dots) or mosquitoes treated with 10 mg ml⁻¹ AZT (green dots) at the indicated time points. Standard deviation and geometric mean are shown. At least 10 mosquitoes were used for each time point. Error bars represent the s.d. A Wilcoxon rank-sum test was used to determine statistical significance (*P<0.05, ****P<0.0001). Absence of P-value represents non-statistical significance. (e) Number of viral RNA molecules were calculated by single mosquito qRT-PCR in bodies (left panel) and legs and wings (right panel) for untreated mosquitoes (grey dots) or mosquitoes treated with 10 mg ml⁻¹ AZT (green dots) at the indicated time points. Absence of P-value represents non-statistical significance. Each experiment was completed at least three times.