Dengue subgenomic flaviviral RNA disrupts immunity in mosquito salivary glands to increase virus transmission

Abstract

Globally re-emerging dengue viruses are transmitted from human-to-human by Aedes mosquitoes. While viral determinants of human pathogenicity have been defined, there is a lack of knowledge of how dengue viruses influence mosquito transmission. Identification of viral determinants of transmission can help identify isolates with high epidemiological potential. Additionally, mechanistic understanding of transmission will lead to better understanding of how dengue viruses harness evolution to cycle between the two hosts. Here, we identified viral determinants of transmission and characterized mechanisms that enhance production of infectious saliva by inhibiting immunity specifically in salivary glands. Combining oral infection of Aedes aegypti mosquitoes and reverse genetics, we identified two 3' UTR substitutions in epidemic isolates that increased subgenomic flaviviral RNA (sfRNA) quantity, infectious particles in salivary glands and infection rate of saliva, which represents a measure of transmission. We also demonstrated that various 3'UTR modifications similarly affect sfRNA quantity in both whole mosquitoes and human cells, suggesting a shared determinism of sfRNA quantity. Furthermore, higher relative quantity of sfRNA in salivary glands compared to midgut and carcass pointed to sfRNA function in salivary glands. We showed that the Toll innate immune response was preferentially inhibited in salivary glands by viruses with the 3'UTR substitutions associated to high epidemiological fitness and high sfRNA quantity, pointing to a mechanism for higher saliva infection rate. By determining that sfRNA is an immune suppressor in a tissue relevant to mosquito transmission, we propose that 3'UTR/sfRNA sequence evolution shapes dengue epidemiology not only by influencing human pathogenicity but also by increasing mosquito transmission, thereby revealing a viral determinant of epidemiological fitness that is shared between the two hosts.

Fig 1. Epidemiological fitness positively correlates with higher levels of sfRNA.

Mosquitoes were orally infected with 10⁶ pfu / ml of DENV-2 PR-1 or PR-2B isolates and analyzed 10 days post-infection. (A) Infection rate per isolates and (B) for the same isolates grouped by epidemiological fitness level. (C) DENV-2 gRNA copies per infected mosquito per isolates and (D) for the same isolates grouped by epidemiological fitness level. (E) Plaque forming unit (pfu) per infected mosquito. (F) Ratio of sfRNA:gRNA in whole mosquito per isolates and (G) for the same isolates grouped by epidemiological fitness level. Bars with a different letter were significantly different following Z-test (A) or Tukey’s test (C, E and F). T-test was applied to test significance for isolates grouped according to epidemiological fitness level (B, D, G). N, number of mosquitoes analyzed. Bars show percentages ± s.e. or means ± s.e.m. (H) Mapping of the 3’UTR sequence differences between PR6452 (high EF) and PR315022 (low EF). xrRNA, XRN1-nuclease resistant RNA; pk, pseudoknot; DB, dumb bell; Green line show predicted long-range nucleotide interaction based on the Zika xrRNA1 model [48]. (I) Northern blot with a 3’UTR probe on RNA extracts from whole mosquitoes 10 days p.i. with PR6452 and PR315022.

Fig 2. Ratios of sfRNA:gRNA in human cells and mosquitoes correlate well.

HuH-7 cells were infected with the isolates. (A) gRNA copies relative to GAPDH expression and (B) sfRNA:gRNA ratio were quantified using RT-qPCR. Bars with a different letter were significantly different following Tukey’s test. (C) Correlation between sfRNA:gRNA ratios in HuH-7 cells and whole mosquitoes was plotted. The dotted line represents an exponential regression, for which the R² and equation is provided within the graph. The data in HuH-7 cells have been previously used for comparing averages of PR-1 and PR-2B strains [30] but never shown for each individual strains.

Fig 3. SfRNA:gRNA ratio is higher in salivary glands, peaks at 10 days post-oral infection and is higher for PR6452 than PR315022.

Mosquitoes were orally challenged with 10⁶ pfu /ml of PR6452 (high EF) or PR315022 (low EF) isolates. After 3, 7, 10 and 14 days, mosquitoes were dissected into salivary glands, midgut and carcass (what is left after dissection) and the DENV gRNA and sfRNA copies were quantified. (A) Infection rate, (B) gRNA copies per infected mosquitoes and (C) sfRNA:gRNA ratio in midgut, carcass and salivary glands at different time points post-infection. Two independent experiments each with twenty mosquitoes per condition were conducted. Percentages (infection rate) ± s.e. and mean ± s.e.m (DENV gRNA copies per infected mosquitoes and sfRNA:gRNA ratio) are shown. Asterisks represent significant differences between viral isolates for the same tissue on the same day (p < 0.05).

Fig 4. 3’UTR sequences determine sfRNA:gRNA ratios in chimeric viruses and correlate with epidemiological fitness.

Mosquitoes were orally challenged with 10⁶ pfu / ml of IC6452 (high EF 3’UTR) or IC315022 (low EF 3’UTR) chimeric viruses. After 3, 7, 10 and 14 days, mosquitoes were dissected into salivary glands, midgut and carcass (what is left after dissection) and the DENV gRNA and sfRNA copies were quantified. (A) Infection rate, (B) gRNA copies per infected mosquitoes and (C) sfRNA:gRNA ratio in midgut, carcass and salivary glands at different time points post-infection. Two independent experiments each with twenty mosquitoes per condition were conducted. Percentages (infection rate) ± s.e. and mean ± s.e.m (DENV gRNA copies per infected mosquitoes and sfRNA:gRNA ratio) are shown. Asterisks represent significant differences between viruses for the same tissue within the same day (p < 0.05).

Fig 5. Virus titer in salivary glands is increased for mosquitoes infected with the chimeric virus containing the high epidemiological fitness 3’UTR sequence.

Mosquitoes were orally infected with 10⁶ pfu / ml of the chimeric viruses, IC6452 or IC315022, containing the high or low EF 3’UTR sequence, respectively. At 14 days p.i., salivary glands were dissected and virus titer was quantified using plaque assay. (A) Infection rate in salivary glands. Bars represent percentages ± s.e. (B) plaque forming unit (pfu) per infected salivary glands. Only infected samples were represented, used to calculate the average and to perform t-test. Each point represents a pair of salivary glands. Bars show mean ± s.e.m. N, number of samples analyzed.

Fig 6. Infection rate of saliva is increased for mosquitoes infected with the viruses containing the high epidemiological fitness 3’UTR sequence.

Mosquitoes were orally infected with 10⁶ pfu / ml of viruses, and let to expectorate in blood. DENV gRNA was quantified in mosquitoes and saliva. (A) DENV gRNA copies per infected mosquitoes, (B) DENV gRNA copies per infected saliva and (C) saliva infection rate 10 days p.i. with the high epidemiological fitness (EF) PR6452 or low EF PR315022. (D) DENV gRNA copies per infected mosquitoes, (E) DENV gRNA copies per infected saliva and (F) saliva infection rate 14 days p.i. with IC6452, containing the high EF 3’UTR, or IC315022, containing the low EF 3’UTR. N, number of infected mosquitoes that contained blood in their abdomen. (A, B, D, E) Bars show geometric mean ± 95% C.I. Only infected samples are represented. (C, F) Bars represent percentages ± s.e. Statistical differences between percentages were calculated using Z-test.

Fig 7. The Toll pathway is inhibited in salivary glands by PR6452 and the inhibition depends on 3’UTR sequences.

Expression of (A) activators of Toll (Rel1a), IMD (Rel2), Jak/STAT (Domeless) and TRAF/Rel2 (Vago) immune pathways, and (B) of genes under the control of the Toll (CecG and DefC) or Jak/STAT (TEP22 and vir-1) pathways at 10 days post-oral feeding with 10⁶ pfu / ml of PR6452, PR315022 or non-infectious blood (uninfected). Expression of (C) activators of Toll (Rel1a), IMD (Rel2), Jak/STAT (Domeless) and TRAF/Rel2 (Vago) immune pathways, and (D) of genes under the control of the Toll (CecG and DefC) or Jak/STAT (TEP22 and vir-1) pathways at 14 days post-oral feeding with IC6452, IC315022 or non-infectious blood (blood). Gene expression was quantified by RT-qPCR and normalized to actin. Each point represents one sample grouping 10 salivary glands. Bars show mean ± s.e.m.