Aedes aegypti uses RNA interference in defense against Sindbis virus infection

Abstract

Background: RNA interference (RNAi) is an important anti-viral defense mechanism. The Aedes aegypti genome encodes RNAi component orthologs, however, most populations of this mosquito are readily infected by, and subsequently transmit flaviviruses and alphaviruses. The goal of this study was to use Ae. aegypti as a model system to determine how the mosquito's anti-viral RNAi pathway interacts with recombinant Sindbis virus (SINV; family Togaviridae, genus Alphavirus).

Results: SINV (TR339-eGFP) (+) strand RNA, infectious virus titers and infection rates transiently increased in mosquitoes following dsRNA injection to cognate Ago2, Dcr2, or TSN mRNAs. Detection of SINV RNA-derived small RNAs at 2 and 7 days post-infection in non-silenced mosquitoes provided important confirmation of RNAi pathway activity. Two different recombinant SINV viruses (MRE16-eGFP and TR339-eGFP) with significant differences in infection kinetics were used to delineate vector/virus interactions in the midgut. We show virus-dependent effects on RNAi component transcript and protein levels during infection. Monitoring midgut Ago2, Dcr2, and TSN transcript levels during infection revealed that only TSN transcripts were significantly increased in midguts over blood-fed controls. Ago2 protein levels were depleted immediately following a non-infectious bloodmeal and varied during SINV infection in a virus-dependent manner.

Conclusion: We show that silencing RNAi components in Ae. aegypti results in transient increases in SINV replication. Furthermore, Ae. aegypti RNAi is active during SINV infection as indicated by production of virus-specific siRNAs. Lastly, the RNAi response varies in a virus-dependent manner. These data define important features of RNAi anti-viral defense in Ae. aegypti.

Figure 1

Predicted protein domains of RNAi components used in this study compared to Drosophila orthologs. Ae aegypti orthologs: Ago2, [Vectorbase: SUPP_AEDES003395], supercontig 1.89; TSN, [Vectorbase: AAEL000293], supercontig 1.5; Dcr2, [Genbank: AY713296], supercontig 1.221. Drosophila orthologs: Ago2, [Genbank: NP_648775], chromosome 3L; Dcr2, [Genbank: NP_523778], chromosome 2R; TSN [Genbank: NP_612021], chromosome 3L. "dsRNA", dsRNA binding; "DUF", domain of unknown function; "DEAD", helicase domain; "PAZ", small RNA binding domain; "PIWI", double-stranded RNA guided RNA cleavage; "RNase", RNA nuclease; "SNase", staphylococcal nuclease; "Tudor", domain of unknown function [51].

Figure 2

TR339-eGFP virus infection transiently increased in mosquitoes receiving a Ago2, Dcr2, or TSN dsRNA injection. (A) qRT-PCR analysis. Percent reduction in cognate Dcr2, and TSN mRNA levels, relative to actin, in midguts following Ago2, Dcr2 or TSN dsRNA treatment. "ND", none detected. Asterisks indicate statistical significance (Mann Whitney U test P ≤ 0.05). Error bars indicate standard error of three experimental replicates. (B) Immunoblot. Evidence of Ago2 protein silencing following dsRNA injection. "A" indicates Ago2 dsRNA-injected; "C" indicates βGAL injected controls. "0" days post-infection corresponds to female midguts harvested prior to virus meal. In each lane, equivalent amounts of protein extract from a pool of six midguts were separated by SDS-PAGE, blotted and probed with anti-Ago2 antibody. The loading control is a 19 kDa anti-Ago-2 cross-reacting band. Blot shown is from a single experiment and is representative of three independent replicates. (C) Two-step qRT- PCR showed that TR339 positive strand (+) RNA significantly increased in midguts of Ago2 dsRNA-injected mosquitoes over βGAL dsRNA controls at 1 dpi (P ≤ 0.05, Mann-Whitney U test). Pools of five midguts per group were used. Error bars depict standard error of three independent feedings. Viral (+) strand RNA copy numbers were calculated using the standard curve method. Asterisk indicates statistical significance, Mann-Whitney U test p ≤ 0.05. (D) TR339-eGFP viral titers of individuals at 4 dpi in Ago2, Dcr2, and TSN dsRNA-injected mosquitoes, with βGAL dsRNA controls. "**" indicates infection rate significantly higher than βGAL controls. "^^" above the graph indicates viral titers significantly higher than βGAL controls. Closed squares, Ago2 dsRNA; closed circles, βGAL dsRNA; closed diamonds, DCR2 dsRNA; closed triangles, TSN dsRNA; open squares, nsP3. Horizontal line indicates the median titer; no line indicates a median value of zero. "()" indicates (number of mosquitoes in each group per number positive). (E) TR339-eGFP viral titers at 7 dpi.

Figure 3

Small viral RNAs are detected. Small RNAs (18 to 25 nts) were isolated from SINV-infected mosquitoes at 2 or 7 days post-infection, size-selected by gel electrophoresis and hybridized to pools of strand-specific probes representing the complete viral genome indicated. Eluted products were separated on a 5% acrylamide gel and detected by autoradiography. "Dpi", day, post-infection, "MRE", MRE16-eGFP-infected, "TR339", TR339-eGFP-infected, "B", bloodfed control.

Figure 4

Non-infectious bloodmeal depletes Dcr2 transcripts; TSN transcript levels are enriched in mosquito midguts during SINV infection. (A) Effects of bloodfeeding on Ago2, Dcr2, and TSN transcript levels. Relative midgut transcript levels from bloodfed mosquitoes were normalized to unfed controls. Transcript levels at 1.0 indicate no change over unfed control levels. (B) Effects of MRE16-eGFP infection on midgut transcript levels relative to un-infected bloodfed controls. (C) MRE16-eGFP (-) strand RNA detected in midgut total RNA from panel (B). (D) Effects of TR339-eGFP infection on midgut transcript levels relative to un-infected bloodfed controls. (E) TR339-eGFP (-) strand RNA levels in midgut total RNA from panel (D). Mosquitoes were fed either a bloodmeal or the virus indicated. (A, B, D) Total RNA from pools of 5 midguts were used for qRT-PCR of Dcr2, Ago2, TSN and Act1 transcripts. Changes in relative transcript levels of each RNAi component were determined by using actin as an internal reference standard and normalizing each value to that of the control group indicated. Relative expression levels were determined using the comparative Ct method; transcript levels at 1.0 indicate no change over bloodfed control levels. Significant changes in expression levels are shown by asterisks (p ≤ 0.05, Mann-Whitney U test). Error bars depict standard error of three independent feedings. (C, E) Viral transcript copy numbers per nanogram midgut RNA were determined by two-step quantitative RT-PCR; these were calculated using the standard curve method. Error bars indicate standard error of three independent virus infections.

Figure 5

Ae aegypti Ago2 protein depletion and accumulation varied in a virus-dependent manner. Protein was extracted from pools of ten midguts from unfed, bloodfed or virus/bloodfed adult female Ae aegypti at the time points indicated. Equivalent protein amounts were separated by 4–15% gradient PAGE prior to blotting. Anti-Ae aegypti Ago2 antibody recognizes the carboxy-terminal peptide sequence YERMQIRTEIQDGHPMFFV. The loading control is a 19 kDa Ago2 cross-reacting band. "U", unfed female, "B", non-infectious bloodfed female, "M", MRE16-eGFP-infected, "T", TR339-eGFP-infected. Blots are representative of two independent experiments.