Selection of miRNA targets
The first miRNA selected for this work, miR8, was highly upregulated in both the Ae. aegypti and An. gambiae fat body following blood feeding and targets the 3′ UTR region of secreted wingless-interacting molecule (Swim), a molecule involved with the Wnt/Wingless signaling pathway (Table 1) [40,41,42,43]. When miR8 was depleted in Ae. aegypti, Swim levels remained high following blood feeding and egg development was inhibited [41]. Another miRNA, miR34, showed differential expression in several different mosquito species during pathogen infection, including in An. gambiae where midgut expression was decreased following an infectious Plasmodium berghei (P. berghei) blood meal [22, 34, 44,45,46]. Specifically, miR34 was predicted to bind to Relish-like transcription factor 1 (Rel1) and Caspar transcripts, important factors in the Toll and IMD immune pathways, respectively [37, 46]. As Caspar is a negative regulator of the Relish-like transcription factor 2 (Rel2) and Cactus is a negative regulator of Rel1, Rel1 and Rel2 transcript levels were also assessed in target gene quantitative PCR (qPCR) reactions (Table 1). These transcripts, as well as mating induced stimulator of oogenesis (MISO), were predicted target genes of miR34 via the now defunct miRNA–mRNA binding prediction webtool Insectar (http://www.insectar.sanbi.ac.za/) [47]. Previously, knockdown of MISO transcripts using RNAi resulted in reduced egg production, indicating a potential role for miR34 in reproduction [48]. The third selected miRNA, miR305, was elevated in the ovaries and midgut of An. gambiae following blood feeding and was higher in midguts following an infectious Plasmodium-containing blood meal [36, 37]. Inhibition of miR305 decreased the midgut microbiota and increased resistance to P. falciparum, whereas enhancement of miR305 increased P. falciparum infection levels and led to higher levels of midgut microbiota [37]. This miRNA was predicted to target the 3′ UTR of APL1C and as APL1 is part of a complex that stabilizes the immune factor thioester-containing protein 1 (TEP1), which binds to the surface of Plasmodium leading to parasite destruction, miR305 may impact Plasmodium infection [37]. Supporting this, miR305 depletion in An. gambiae led to increased resistance to both P. falciparum and Plasmodium berghei infection and altered the levels of many immunity or anti-Plasmodium genes in mosquito midguts [43]. The final miRNA, miR375, was only detected in blood fed Ae. aegypti mosquitoes and was predicted to bind to the 5′ UTR of Toll pathway immune genes Cactus and Rel1 [49]. Expression of a miR375 mimic in Ae. aegypti mosquitoes or cells led to binding of the 5′ UTRs of Cactus and Rel1 and the upregulation of Cactus and downregulation of Rel1 [49]. Similar changes in target genes and an increase in Dengue virus type 2 titers were observed in Ae. albopictus Aag2 cell lines [49]. Although miR375 has not been studied in An. gambiae, this miRNA has an identical sequence to miR375 in Ae. aegypti and has also been predicted by Insectar to target An. gambiae Cactus and Rel1 along with other gene transcripts including Caspar and Rel2 [47].
Plasmid preparation and production
Sure 2 supercompetent E. coli cells (Agilent Technologies, 200152) were transformed as per kit instructions (SOC media was substituted for NZY + media) with pAcEGFP and pWTAgDNV plasmids [18]. Transformed colonies were plated on Luria broth agar plates with 100 µg/mL ampicillin and incubated at 37 °C overnight. Colony PCR was used to verify transformations and selected colonies were grown in 5 mL of Luria broth in a 37 °C shaker overnight and then preserved as glycerol stocks. Purified plasmids were produced by growing glycerol stocks in liquid culture as before, extracted using an Omega Bio-tek E.Z.N.A. Plasmid DNA Kit (D6942-02), and quantified using a NanoDrop ND-1000 spectrophotometer.
Intron design
A potential splice acceptor site in WT AgDNV was identified at position 463 of the gene encoding the viral protein using the neural-network-based NetGene2 predictive splicing server, which identifies transition sequences between introns and exons [50,51,52]. This sequence was converted from AG^ACGCAGACAG (with “^” indicating the predicted splicing site) into a splice donor site by replacing the intronic portion with the starting sequence of the second intron of An. gambiae RPS17 such that the new sequence was AG^GTAGGCGCGC. This sequence was further modified by two base pairs to AG^GTAAGTGCGC to match the An. gambiae U1 small nuclear RNA conserved region (Fig. 1A) [53]. This U1 sequence (GTAAGT) represents the binding site for the U1 small nuclear ribonucleoprotein which helps to form the spliceosome [53]. A splice acceptor with the sequence TACTGACATCCACTTTGCCTTTCTCTCCACAG was created to accompany this splice donor at position 464 of the gene encoding the viral protein by adding in the branch point, polypyrimidine tract, and intronic portion from the 3′ end of a chimeric human intron (last 32 nucleotides) preceding an immunoglobulin gene heavy chain variable region that is commonly found in commercial vectors such as in the pRL-CMV plasmid from Promega (Fig. 1A) [54,55,56]. This splice donor and splice acceptor site were initially chosen within AgDNV to attempt the creation of a nondefective recombinant AgDNV, as previously described for AaeDNV, but we later decided to use a co-transfection system with pWTAgDNV and a transducing plasmid with the artificial intron to create an EGFP reporter phenotype [20]. These developed splice donor and splice acceptor sites were placed within the EGFP gene of pAcEGFP at positions 334 and 337, respectively, to create a reporter phenotype such that improper intronic splicing or a lack of splicing would result in a stop codon within the EGFP gene and correct splicing would result in EGFP expression (Fig. 1B). Predicted splicing was examined at all steps using NetGene2 and the created splice donor and splice acceptor sequences both had a confidence scores of 1.0, indicating a high confidence in splicing [50].
miRNA and sponge selection
For intronic miRNA expression, endogenous pre-miRNA sequences were inserted into the created intron so that upon splicing, the pre-miRNA hairpin would be co-transcriptionally processed alongside EGFP transcripts [22,23,24]. Selected pre-miRNA sequences for An. gambiae miR8, miR34, miR305, and miR375, as well as a miRNA sponge against miR8 (miR8SP), were added to this developed intron to test the co-expression system and intronic splicing mechanism. A random nonsense RNA sequence (NS) was added to the intron as a control. These miRNAs and the miRNA sponge were chosen, as described above, on the basis of known or predicted effects on genes involved with immunity, pathogen defense, or reproduction in An. gambiae, Ae. aegypti, or relevant mosquito cell lines (Table 1). To test intron functionality and demonstrate that splicing is sequence-dependent, altered splice donor and splice acceptor site sequences were developed using site-directed mutagenesis of the pAcEGFPmiR8 plasmid [50]. When the splice donor site was changed by a single nucleotide (in bold) from AGGTAAGTGCGC to AGATAAGTGCGC, NetGene2 no longer identified this as a splice donor site. Similarly, when the splice acceptor site was changed by one nucleotide (in bold) from TACTGACATCCACTTTGCCTTTCTCTCCACAG to TACTGACATCCACTTTGCCTTTCTCTCCACAT, this site was no longer predicted to be a splice acceptor.
Cloning and intronic cargo
MluI and BstBI sites were introduced into the EGFP-encoding gene of pAcEGFP using site-directed mutagenesis to create synonymous mutations. A MluI site was created by altering position 327 of EGFP from C to G, and position 330 from C to T. A BstBI site was created in EGFP by switching position 348 from G to A. Endogenous An. gambiae pre-miRNA sequences from miRbase (https://www.mirbase.org/) were converted to DNA and used to order g-blocks from Integrated DNA Technologies (IDT) [22]. The mir8SP sequence contained ten repeated blocks of the reverse complement of mature An. gambiae miR8. Each block was separated by four spacer nucleotides and the entire sponge sequence was placed within the intron as with pre-miRNA sequences. A nonsense RNA (NS) was created using a random sequence with no matches to the An. gambiae genome or transcriptome when searched using the Basic Local Alignment Search Tool (BLAST). Each pre-miRNA, miRNA sponge, or nonsense RNA was coded on IDT g-blocks synthesized with flanking MluI and BstBI sites, EGFP segments to replace those removed during digestion, and the splice donor and splice acceptor sites (Table S1; Fig. 1A). G-blocks were subcloned into pJet using a CloneJet PCR Cloning Kit (ThermoFisher Scientific, K1231) and later digested using MluI and BstBI. These inserts were ligated into pAcEGFP that had also been digested with MluI and BstBI, and the resulting plasmid sequences were verified.
Cell culture and transfections
Sua5B and Moss55 An. gambiae cells were grown in 25 cc plug cap flasks at 28 °C and passaged once per week at a 1:5 dilution with Schneider’s Drosophila media with 10% fetal bovine serum (FBS) v/v. For transfections, cells were quantified using a hemocytometer and 6 × 106 cells were added to each well of a 6-well plate along with 3 mL of complete media and incubated overnight. Cells were transfected at ~70–80% confluence with a 1:2 ratio of pWTAgDNV to transducing plasmid with 830 ng pWTAgDNV and 1660 ng transducing plasmid per well using a Lipofectamine LTX with Plus Reagent kit (ThermoFisher Scientific, 15338030). Briefly, plasmids were added to a mix of 500 µL OptiMem media with 3 µL Plus reagent and incubated at room temperature for 10 min. Then, 5 µL Lipofectamine was added and tubes were incubated at room temperature for 25 min before transfecting each well with 500 µL of this mixture. Transducing plasmids were pAcEGFPmiR8, pAcEGFPmiR8SP, pAcEGFPmiR34, pAcEGFPmiR305, pAcEGFPmiR375, pAcEGFPNS, pAcEGFPSA, and pAcEGFPSD. Cells were incubated and imaged at 3 d post-transfection. RNA for splicing validation was also gathered 3 d post-transfection. Preliminary in vitro miRNA and target gene expression experiments harvested RNA at 5 d post transfection. For Sua5B in vitro miRNA and target gene expression, cells transfected with pWTAgDNV and pAcEGFPNS served as controls, whereas in Moss55 in vitro miRNA and target gene expression experiments, cells transfected with pWTAgDNV alone served as a control.
Viral production and quantification
To produce virus particles for mosquito infections, Moss55 cells were transfected with selected transducing and helper viruses (as described) and virions were extracted 3 d post-transfection by removing the media, washing cells with 1× phosphate-buffered saline (PBS), and suspending cells in 1 mL 1× PBS. Cells were lysed using three cycles of freeze-thawing and centrifuged at 5000 rpm for 5 min to pellet debris. The virus-containing supernatant was collected and plasmid DNA and free viral genomes were removed using an Ambion TURBO DNA-free kit (AM1907). DNA was extracted using an Omega Bio-tek E.Z.N.A Tissue DNA kit (D3396-02) and viral genome equivalents were determined using standard curves created using AgDNV-coding plasmids with a single copy of each gene-of-interest. Samples and standards were run using PerfeCTa SYBR Green FastMix (Quantabio, 95,072–012) on a Qiagen Rotor-Gene Q at 95 °C for 2 min followed by 40 cycles of 95 °C for 10 s, 60 °C for 40 s, and 72 °C for 30 s. Runs were finished with a melt step using a ramp of 55–99 °C rising by 1 °C each step. WT AgDNV was quantified using primers against AgDNV nonstructural gene 1 (NS-RT-IIIF: CATTCGATCACGGAGACCAC, NS-RT-IIIR: GCGCTTGTCGCACTAAGAAAC) and a standard curve of pWTAgDNV. Selected transducing viruses (vAcEGFPmiR8, vAcEGFPmiR8SP, vAcEGFPmiR34, vAcEGFPmiR305, vAcEGFPmiR375, vAcEGFPNS, vAcEGFPSA, and vAcEGFPSD) were quantified using primers against EGFP (GFP-RT-II-F497: TCAAGATCCGCCACAACATC, GFP-RT-II-R644: TTCTCGTTGGGGTCTTTGCT) and a standard curve of pAcEGFP. Each production of virus consisted of a mixture of vWTAcEGFP and a transducing virus.
Mosquito injections
Female An. gambiae mosquitoes (Keele strain) that were 3 d post-emergence were injected intrathoracically with 200 nL densovirus mixture containing both wild-type vWTAgDNV and transducing virus (either vAcEGFPmiR8, vAcEGFPmiR8SP, vAcEGFPmiR34, vAcEGFPmiR305, vAcEGFPmiR375, vAcEGFPNS, vAcEGFPSA, or vAcEGFPSD) using a Drummond Scientific Nanoject III (3-000-207) and Drummond Scientific 10 µL microcapillary tubes (3-000-210-G) pulled using a Sutter Instrument Co. Model P-2000 (Heat 400, Fil 4, Vel 40, Del 140 Pul 140). Three biological replicates in mosquitoes were completed. For each replicate, mosquitoes were injected with ~106–107 transducing virus particles and 106–108 WT DNV particles (Table S2). Mosquito treatment groups were kept in separate cardboard cup cages with 10% sugar solution w/v ad libitum until RNA extraction or imaging. RNA was harvested and tested from three biological replicates.
RNA extractions and cDNA production
For both in vitro and in vivo experiments, RNA was extracted using an Omega Bio-tek MicroElute Total RNA Kit (R6831-02). For in vitro experiments, RNA was extracted 3 d post-transfection for intronic splicing assessments or 5 d post-transfection for miRNA and target gene quantification. For in vivo experiments, mosquitoes were individually homogenized 10 d post-injection in lysis buffer using zinc-plated steel BB pellets (Daisy 0.177 cal or 4.5 mm) and a Qiagen TissueLyser II with a lysis program lasting 2 min with a frequency of 30 Hz. Following homogenization, RNA was extracted and DNase treated either on the column using an Omega Bio-tek RNase-free DNase Set I kit (E1091) or following RNA extraction using an Ambion DNA-free DNA Removal Kit (AM1906). For target gene quantification or assessment of intronic splicing, cDNA was synthesized using a Quantabio qScript cDNA synthesis kit (95047–500); whereas for miRNA quantification, samples were converted to cDNA using the HighSpec option in the Qiagen miScript II RT kit (218161) and diluted 1:10.
Intronic splicing, miRNA expression, and target gene quantification
In vitro and in vivo intronic splicing was assessed using primers spanning the intronic region (GFP-COLPCRF: CTGACCTACGGCGTGCAGTGC, RGFP-COLPCRR: CGGCCATGATATAGACGTTGTGGC). PCR products were run on 2% agarose gels and imaged using a UVP GelDoc-It transilluminator. Spliced transcripts resulted in a product of 274 bp, whereas PCR reactions using DNA plasmid controls or unspliced transcripts produced variably sized amplicons depending on insert size with most being ~480 bp.
Target gene qPCR reactions were run on a Qiagen Rotor-Gene Q using PerfeCTa SYBR Green FastMix (Quantabio, 95072–012) or an Applied Biosystems 7900HTFast Real-Time PCR System with Applied Biosystems PowerUp SYBR Green Master Mix (A25724), with conditions of 95 °C for 2 min, 40 cycles of 95 °C for 10 s, 60 °C for 40 s, and 72 °C for 30 s, and a melt curve with a ramp from 55 °C to 99 °C with 1 °C change per step. Primers for An. gambiae Swim cDNA were developed during this study, whereas others came from published studies (Table S3) [48, 57,58,59,60].
Reactions to quantify miRNAs used Qiagen miScript SYBR Green PCR kits (218075) and a Qiagen Rotor-Gene Q with a universal reverse primer and forward primers consisting of the sequences of each mature miRNA (Table S4) [40]. An. gambiae U6 levels served as a reference with which to compare miRNA levels. Conditions for miRNA qPCR reactions were 95 °C for 15 min followed by 40 cycles of 95 °C for 15 s, 60 °C (for all miRNAs during cell culture replicates as well as for in vivo miR34) or 55 °C (all U6 reactions and in vivo miR375) for 60 s, and 72 °C for 20 s. All reactions ended with a melt curve consisting of a ramp from 55 °C to 99 °C that increased 1 °C per step.
Data analysis
All qPCR data was analyzed using the delta-delta Ct method to calculate the fold change in expression relative to reference genes (S7 for mRNA transcripts and U6 for mature miRNA quantification unless otherwise noted). The fold change expression data was log2 transformed and a D’Agostino–Pearson omnibus K2 test was used to assess normality in Graphpad Prism 9. If both the control and experimental groups passed the normality test, a parametric unpaired two-tailed t-test assuming equal standard deviations was used to measure statistical significance. If either or both groups failed the D’Agostino–Pearson normality test, a nonparametric two-tailed Mann–Whitney test was used to compare ranks and to assess significance. Significant P values (< 0.05) were reported on graphs. All graphs report fold change expression using a log2 scale. The mean and standard error of the mean was reported for groups analyzed using an unpaired t-test, whereas median and 95% confidence intervals were shown for groups compared using a nonparametric two-tailed Mann–Whitney test.