Phylogenetic analysis of HSP70 genes from An. gambiae and D. melanogaster indicate that Anopheles HSC70B gene is evolutionarily more conserved with Drosophila Hsc70-1 and Hsc70-4 genes than other Anopheles homologues (Fig 1.). For example, Anopheles HSC70B is most tightly liked to Drosophila Hsc70-4 (Fig. 1). Multiple sequence alignments indicated that there are high polymorphisms near the 5' and 3'-end coding and non-coding regions of the Anopheles HSC70 family (Fig. 2). Indeed, based on these sequence polymorphisms among the Anopheles HSC70 family, we were able to detect unique transcripts of the Anopheles HSC70B gene among the Anopheles HSP70 genes by using the RT-PCR and qRT-PCR primers based on the 3'and 5'-end regions, respectively (Fig. 3 and Table 1). In addition, the dsRNA of Anopheles HSC70B was designed based on the 5' end that was specific to the HSC70B gene, which enabled HSC70B specific knockdown in Anopheles gambiae.

First, efficiency of dsRNAi was assessed by using semi-quantitative RT-PCR and qRT-PCR analyses. In contrast to high induction of HSC70B in dsβ-gal and ONNV/ONNV-eGFP coinjected mosquitoes, only traces of HSC70B mRNA were detected in dsHSC70B mosquitoes using semi-quantitative RT-PCR and primers corresponding to the 3' end of the HSC70B gene (Fig. 3A). This result shows that the injection of specific dsHSC70B successfully reduced endogenous target transcripts of HSC70B gene after ONNV injection. Similarly, RT-PCR results also showed the successful post-transcriptional inhibition in agglutinin and EF-1α genes (Fig. 3A).

Mosquitoes were coinjected with ~3.1 × 10² pfu of virus and 625 ng of the respective dsRNAs targeting the HSC70B, agglutinin and EF-1α gene or the β-gal gene as an internal control (Table 1). To quantify ONNV in the respective dsRNAi-treated mosquitoes, each An. gambiae was triturated and ONNV titer was quantified by plaque assay from the treatment groups. Mosquitoes coinjected with ONNV and dsHSC70B had significantly more plaques than those mosquitoes coinjected with ONNV and dsβ-gal (P = 0.00045) at 6 d.p.i. (Fig. 4). No significant differences in the number of plaques were observed among the mosquitoes injected with ONNV and each of ds-agglutinin, dsEF-1α, and dsβ-gal (P ≥ 0.74) (Fig. 4).

To investigate whether up-regulation of HSC70B gene impedes ONNV replication, An. gambiae was infected with eGFP-tagged ONNV whilst HSC70B transcripts were silenced by dsRNAi. ONNV titers were then indirectly estimated by visually comparing eGFP expression among three mosquito groups that were coinoculated with dsHSC70B, or dsβ-gal or buffer. In control mosquitoes injected with dsβ-gal or buffer, expression of eGFP was usually weak in head, thorax and abdomen tissues (Fig. 5 and Table 2). Mosquitoes coinjected with ONNV-eGFP and dsHSC70B typically had stronger expression of eGFP in all three tissues (Fig. 5). At 6 dpi, 47% (n = 32) of mosquitoes receiving dsβ-gal expressed eGFP in thoracic tissues. However, 87% (n = 23) expressed eGFP in thoracic tissues when dsHSC70B was silenced (Table 2). In a similar way, the mosquitoes receiving dsβ-gal showed 38% (n = 32) and 22% (n = 32) of eGFP expression in head and abdomen, respectively (Table 2). In contrast, mosquitoes receiving dsHSC70B showed 70% (n = 23) and 65% (n = 23) of eGFP expression in head and abdomen, respectively (Table 2).

The ribosomal proteins S4 (RpS4) and S7 (RpS7) were used as internal controls for infection studies of ONNV and the malaria parasite, Plasmodium berghei, respectively 714. When RpS7 was analyzed by qRT-PCR and cDNA microarray studies, Student's t test determined that the mRNA expression levels of RpS7 in uninfected and ONNV-infected mosquitoes collected at 24 h, 48 h and 14 days p.i. were not significantly different (data not shown). As shown in Figure 3, RT-PCR analysis of RpS7 and qRT-PCR analysis of RpS4 transcript levels at 6 days p.i., detected no significant difference between the relative mRNA levels derived from treated mosquitoes. The results indicate that the low expression levels observed for candidate genes were related to the observed knockdown effect by dsRNAi, and not attributed to variation in sample loading.

Co-injection of ONNV and the dsHSC70B significantly shortened the lifespan of adult mosquitoes compared with the co-injected mosquitoes of ONNV and dsβ-gal, causing a significant reduction of survival rates from 7 days p.i. (Fig. 6, P ≤ 0.019). In addition, An. gambiae mosquitoes with down-regulated HSC70B alone lead a reduced survival rate (~80%) at 6 days post injection, though it is much less harmful than co-injection of ONNV and dsHSC70B. This suggests that both reduced expression levels of HSC70B gene and increased ONNV infection level synergistically shorten the lifespan of An. gambiae (Fig. 4 and Fig. 6).

The 4arr strain of An. gambiae was reared at 27°C and 80% relative humidity under a 16 h light: 8 h dark photoperiod, as previously described 34. Adults were supplied with a cotton wool pad soaked in a 10% sucrose solution ad libitum and fed on anaesthetized guinea pigs once per week for egg production.

The SG650, strain of ONNV was obtained from the World Reference Center for Arboviruses at the University of Texas Medical Branch, Galveston, TX. Strain SG650 was isolated from human serum in Uganda in 1996 35 and has been passed once in Vero cells (GenBank Accession Number AF079456). Stock virus was produced following a single passage in Vero cells maintained at 37°C in Leibovitz L-15 medium with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 μg/mL streptomycin. Cell supernatant was harvested when 75% of the cells showed cytopathic effect (3+ CPE). Supernatant containing the virus was collected and titrated. The virus stock contained 2 × 10⁷ plaque-forming units (pfu)/ml, diluted to 2 × 10⁶ pfu/ml, aliquoted, and stored at -80°C.

The eGFP gene was amplified from pEGFP plasmid (Clontech Laboratories, Inc) using primers Onn-GFP-Asc-F (5'-GACCTATGGTG

AGCAAGGGCGAGGAGCTGTTC-3'

TTACTTGTACAGCTCGTCCAT

Templates for the preparation of dsRNA for each candidate gene were PCR-derived fragments flanked by two T7 promoter sequences (TAA TAC GAC TCA CTA TAG) (Table 1). Each PCR-derived fragment were sequenced and blasted against the genomic database of An. gambiae 38 to validate the redundancy of the sequence and all confirmed as a unique sequence. Single-stranded RNAs were then synthesized by using the MEGAscript T7 transcription kit (Ambion, Austin, TX) according to the manufacturer's instructions. Annealed dsRNAs were ethanol precipitated and dissolved in injection buffer (0.1 mM sodium phosphate, pH 6.8; 5 mM KCl). In knockdown experiments, ~0.5 μl of a 2:1 mix of dsRNAs (2 μg/μl), ONNV or ONNV-eGFP (~2 × 10⁶ pfu/ml) were coinjected into the thorax of CO₂-anesthetized adult females by using a IM 300 Microinjector (Narishige, Japan). Thus, each mosquito was coinjected with ~3.1 × 10² pfu of virus and ~625 ng of dsRNA.

Each mosquito was triturated in 1 ml of DMEM, and large particulates were pelleted by centrifugation at 300 r.p.m. and then titrated by standard plaque assay in Vero cells 39. The plaques were counted and the differences in viral titers were analyzed by Pairwise t-tests.

Total RNA samples were extracted from three batches of 15 adult female mosquitoes at 6 d.p.i. with ONNV or ONNV-eGFP and the respective dsRNA, using the Trizol Reagent (Invitrogen, CA). To remove genomic DNA contamination, RNA samples were treated with 1.0 μl DNase I following the manufacturer's instructions (50–375 units/μl; Invitrogen, CA). For reverse transcription, 5 μg of total RNA were reverse transcribed with Superscript III RNase H-reverse transcriptase (Invitrogen). Single-stranded cDNAs of different dilution were amplified by PCR using recombinant Taq DNA polymerase (Invitrogen). To show the RNAi efficiency, primers were made to amplify endogenous agglutinin, EF-1α and HSC70b genes of An. gambiae. RpS7 gene of An. gambiae was used as an internal control for 23 cycles (Table 1). To characterize the ONNV-eGFP transcript, primers were made to amplify endogenous E1 and NS1 genes of for 23 cycles (Table 1).

qRT-PCR was performed using an ABI 7700 Sequence Detection System (Applied Biosystems, CA). Standard curves were generated for each transcript tested using 10-fold serial dilutions of An. gambiae genomic DNA ranging from 116 to 0.0116 ng per reaction. All reactions were performed in triplicate in a total volume of 25 μl containing 12.5 μl of SYBR Green PCR Master Mix, 300 nmol of each primer at the following conditions: 50°C for 2 min, 95°C for 10 min followed by 50 cycles of denaturation at 95°C for 15 s, annealing and extension at 60°C for 1 min. RNA samples were extracted from mosquitoes at 6 days p.i. Sequences of gene-specific primer sets are given in Table 1. Statistical significance of differences in the expression of individual genes was determined by using a Student's t-test between the relative transcript values derived from the dsHSC70B-injected and dsβ-gal injected mosquitoes across 3 replicates for each gene.

To evaluate the knockdown effect of HSC70B gene on the survival rate of ONNV infected An. gambiae, 15 females per cohort were intrathoracically co-injected with dsHSC70B and ONNV. For control, 15 females per cohort were intrathoracically coinjected with dsβ-gal and ONNV. Each treated cohort was kept in 8 cm (diameter) ×12 cm cages with a cotton wool pad soaked a 10% sucrose solution. The cages were placed at 27°C and 80% relative humidity under a 12 h light: 12 h dark photoperiod, and mosquito survival was assessed at 24 hours. Survival was defined as the ability of the mosquito to right itself. Experiments for each of the two groups were replicated six times.

The An. gambiae genome has 10 genes containing the HSP70 domain 40. Among these, two genes which have a short fragment of HSP70 domain were excluded for further analysis (ENSANGG00000023531, ENSANGG00000023619). The HSP70 gene sequences for Drosophila were obtained from the Berkeley Drosophila Genome Project 4142.

Multiple sequence alignments were performed by using ClustalW v1.81 43. Phylogenetic trees were constructed by the Neighbor-Joining (NJ) and maximum parsimony (MP) methods, both included in MEGA3 44. The accuracy of reconstructed trees was examined by the bootstrap test with 10,000 replications.