Johnson et al. Parasites & Vectors (2025) 18:355 Parasites & Vectors https://doi.org/10.1186/s13071-025-06994-7 RESEARCH Open Access Expression of endogenous Anopheles gambiae microRNAs using an Anopheles gambiae densovirus (AgDNV) intronic expression system Rebecca M. Johnson1, Hillery C. Metz2, Yasutsugu Suzuki3, Kyle J. McLean4 and Jason L. Rasgon2,5,6,7* Abstract Background Anopheles gambiae densovirus (AgDNV) is a highly species-specific parvovirus that reaches high titers in adult Anopheles gambiae mosquitoes with few transcriptomic effects and minimal significant fitness effects. Given these characteristics, AgDNV has been proposed as a viral vector for basic research and mosquito control. Previous work created an AgDNV co-expression system with a wild-type AgDNV helper plasmid and a transducing plasmid expressing enhanced green fluorescent protein (EGFP) that can be used to co-transfect cells to generate infec- tious recombinant transducing AgDNV virions. Generated virions infect the An. gambiae midgut, fat body, and ova- ries, yet this viral vector system is limited in the size of transgenes that can be expressed due to capsid packaging limitations. Methods Considering these size constraints, we created an artificial intron within the EGFP gene of the transducing construct that can express small pieces of genetic material such as microRNAs (miRNAs), microRNA sponges, or other small sequences. Placement of this intron in EGFP created a fluorescent reporter such that incorrect splicing produces a frameshift mutation in EGFP and an early stop codon, whereas correct splicing results in normal EGFP expression and co-transcription of the intronic genetic cargo. A selection of miRNAs with predicted or demonstrated importance in mosquito immunity and reproduction with expression localized to the fat body or ovaries were chosen as intronic cargo. Construct expression and splicing was evaluated, and the impact of miRNA expression on putative miRNA targets was measured in vitro and in vivo. Results The created intron was correctly spliced in cells and mosquitoes; however, miRNA delivery resulted in incon- sistent changes to miRNA and predicted target gene transcript levels—possibly due to organ-specific miRNA expres- sion or inaccurate putative target predictions leading to miRNA–target gene sequence mismatch. Conclusions Although our results on target gene expression were inconsistent, with optimization this viral vector and developed intron have potential as an expression tool within An. gambiae mosquitoes or cell lines. Keywords miRNA, miRNA sponge, RNAi, Aedes aegypti, Densovirus, AgDNV, Viral vector *Correspondence: Jason L. Rasgon jlr54@psu.edu Full list of author information is available at the end of the article © The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat ivecom mons. org/l icen ses/ by/4.0 /. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons.o rg/ publi cdomai n/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Johnson et al. Parasites & Vectors (2025) 18:355 Page 2 of 16 Background make AgDNV ideal for use as a late-life acting bioinsecti- Anopheles gambiae is the major vector of Plasmodium cide or as a viral vector to express genes against the para- falciparum in Sub-Saharan Africa, where most malaria sites themselves [16]. cases occur [1]. While current malaria control efforts Previously a co-plasmid expression system was devel- rely heavily on bite prevention via bed nets and mosquito oped consisting of the unaltered AgDNV genome in a reduction through insecticides, insecticide resistance is pBluescript cloning vector (pWTAgDNV) and a trans- increasing [2]. Similarly, the efficacy of antimalarials used ducing pBluescript construct containing an Actin5C to treat disease in humans is threatened by drug-resistant promoter, an EGFP reporter sequence, and a SV40 ter- parasites. As such, there is an increasing need for novel mination sequence (pAcEGFP) [18]. From here on, “p” tools to better investigate An. gambiae biology, as well as designates the plasmid form of the construct, whereas new methods for mosquito and pathogen control (such as “v” designates the viral form; constructs lacking a label the development of malaria-resistant mosquitoes capable denote the general sequence. As both plasmids in this co- of replacing susceptible populations) [3]. expression system possess the AgDNV terminal hairpins Although CRISPR-Cas9 editing has great promise, such that are crucial for genome packaging, both sequences experiments typically require specialized microinjection get packaged into capsids produced by the wild-type equipment, alterations to the mosquito genome, and the construct. Virions produced by this co-expression sys- time-consuming establishment of mosquito lines. The tem localize to mosquito tissues important for pathogen introduction of genetic material without modifying the transmission and immunity such as the midgut, ova- genome, such as through genetically modified microbes, ries, and fat body when injected into adult mosquitoes offers an alternative approach that is useful for altering [13, 18]. Although AgDNV has potential as a viral vec- the expression of existing genes, introducing new genes, tor, the small genome and capsid size limits the length of or, in the case of paratransgenesis, using a symbiont to transgenes that densoviruses can express; for example, in express a transgene within the vector that acts against AaeDNV, a size increase of 8% over that of the wild-type the pathogen. While the bacterium Wolbachia has been genome resulted in a 10% reduction in packaging effi- proposed for paratransgenesis, and select strains have cacy [17, 19]. While less deleterious, a shorter sequence been successful at blocking pathogens in Aedes aegypti, can also negatively alter packaging efficacy [18]. These Wolbachia has yet to be genetically modified, and gener- size constraints have led to the development of various ating stable infections in An. gambiae has proven difficult expression strategies using AaeDNV, including the use [4–6]. Although there have been reports of Wolbachia of an artificial intron to express microRNAs or sponges infections in wild populations of An. gambiae in Africa, [20]. These constraints and the success of various modi- these findings need further verification, and the impact fication strategies in AaeDNV led us to modify the trans- of these infections on P. falciparum in An. gambiae are ducing construct of AgDNV to express microRNAs or currently unclear [5, 7–10]. Viruses also have potential miRNA sponges that would increase the size of AcEGFP for use in paratransgenesis; however, a limited number from 3994 nt to closer to the 4139 nt size of WT AgDNV, infect An. gambiae and few are ideal for genetic manipu- but would not exceed the size of the WT genome [18]. lation or for the introduction of new genes owing to off- MicroRNAs (miRNAs) are small, non-coding RNAs target effects, or the danger of transmission to humans that act as post-transcriptional regulators of gene and the resultant disease [11, 12]. One of the only non- expression through the RNA interference (RNAi) path- pathogenic insect-specific viruses known to infect An. way [21]. The RNAi pathway and miRNAs are highly gambiae was discovered in 2008 in An. gambiae Sua5B conserved and, in An. gambiae, over 163 miRNAs cells: the An. gambiae densovirus (AgDNV) [13]. AgDNV involved in a variety of processes including mosquito is closely related to other mosquito densoviruses includ- reproduction, immunity, and development have been ing Culex pipiens pallens densoviruses (CppDNV) and identified to date [22–28]. Endogenous miRNAs are Aedes aegypti densovirus (AaeDNV), and consists of a coded in introns or intergenic sections of DNA and 4139 nucleotide (nt) ssDNA genome with terminal hair- form short, hairpin-shaped, secondary RNA structures pins at each end that allow for viral genome packaging following transcription [29]. After processing by Dro- [13]. AgDNV is highly specific to An. gambiae and has sha, pre-miRNA hairpins are exported from the nucleus poor infectivity even within closely related mosquitoes to the cytoplasm where they are cut further by Dicer [14]. While some densoviruses in other mosquito spe- into ~22 nt duplexes [30]. One strand of each duplex is cies do cause mortality, AgDNV is not pathogenic to degraded, and the remaining, more stable strand forms An. gambiae and increases in titer over the course of the the mature miRNA that is brought to target mRNA adult lifespan while having little impact on An. gambiae sequences by the RNA-induced silencing complex fitness or gene expression [15–17]. These characteristics (RISC). The mature miRNA sequence binds to regions J ohnson et al. Parasites & Vectors (2025) 18:355 Page 3 of 16 of mRNAs where there is sequence complementarity enhanced, whereas expression of miRNA sponges will (often the 3′ UTR) and the degree of complementarity lead to depletion of endogenous mature miRNAs. controls whether the mRNA transcript is cut and tar- As miRNAs are often encoded in introns or intergenic geted for degradation or whether binding simply blocks regions, and we wanted to express small RNA cargo in a translation [21, 31]. Endogenous An. gambiae miRNAs way that allowed for tracking of expression, we developed or in silico designed miRNA sponges that are com- an artificial intron with a reporter phenotype within the plementary to mature miRNA sequences and “soak” EGFP gene of the transducing AgDNV construct (Fig. 1A up endogenous miRNAs are ideal for expression via and B). To test this intron, we identified miRNAs for AgDNV owing to their small size [32, 33]. Through the expression and manipulation that are putatively involved expression and endogenous processing of pre-miRNAs in mosquito functions such as immunity and egg devel- into mature miRNAs, levels of mature miRNAs can be opment that are tied to organs that AgDNV is known to infect [34–37]. Selected miRNAs all had predicted or Fig. 1 Intron layout and splicing scheme. A Splicing donor and acceptor sites within the EGFP gene flank the pre-miRNA region. MluI and BstBI cut sites allow for swapping of g-block sequences containing EGFP bases removed during digestion, splice sites, and intronic cargo (pre-miRNA sequences, miRNA sponge, or nonsense RNA). Arrows mark cut sites within the splice donor and acceptor sequences. B During transcription, WT AgDNV transcripts are expressed from the WT construct (pWTAgDNV), while EGFP-encoding transcripts containing the intron are expressed by the transducing construct. Intronic splicing of the pre-mRNA (or miRNA sponge) transcript from the transducing construct results in the rejoining of EGFP-encoding transcript halves and the intronic cargo being processed via the RNAi pathway. Translation of WT DNV transcripts results in capsid formation, whereas translation of EGFP mRNA results in EGFP expression if intronic splicing occurred correctly Johnson et al. Parasites & Vectors (2025) 18:355 Page 4 of 16 observed functions within An. gambiae, yet many have was predicted to bind to Relish-like transcription factor not been purposefully manipulated in An. gambiae, and 1 (Rel1) and Caspar transcripts, important factors in the functional studies identifying specific mRNA transcript Toll and IMD immune pathways, respectively [37, 46]. As targets and downstream effects are lacking (Table 1) [36, Caspar is a negative regulator of the Relish-like transcrip- 38, 39]. tion factor 2 (Rel2) and Cactus is a negative regulator of The lack of validation of these selected miRNAs within Rel1, Rel1 and Rel2 transcript levels were also assessed in An. gambiae proved challenging when assessing changes target gene quantitative PCR (qPCR) reactions (Table 1). in miRNA or mRNA target levels, yet transcript splicing These transcripts, as well as mating induced stimulator of patterns indicate that the developed intron delivery sys- oogenesis (MISO), were predicted target genes of miR34 tem functioned as expected. This intron represents a new via the now defunct miRNA–mRNA binding prediction AgDNV viral vector expression strategy and may be use- webtool Insectar (http:// www.i nsect ar. sanbi. ac. za/) [47]. ful for the expression of sequences including endogenous Previously, knockdown of MISO transcripts using RNAi miRNAs, miRNA sponges, synthetic short interfering resulted in reduced egg production, indicating a potential RNAs, or guide RNAs within genetically modified Cas9 role for miR34 in reproduction [48]. The third selected mosquito lines. miRNA, miR305, was elevated in the ovaries and midgut of An. gambiae following blood feeding and was higher in Methods midguts following an infectious Plasmodium-containing Selection of miRNA targets blood meal [36, 37]. Inhibition of miR305 decreased the The first miRNA selected for this work, miR8, was highly midgut microbiota and increased resistance to P. falcipa- upregulated in both the Ae. aegypti and An. gambiae fat rum, whereas enhancement of miR305 increased P. falci- body following blood feeding and targets the 3′ UTR parum infection levels and led to higher levels of midgut region of secreted wingless-interacting molecule (Swim), microbiota [37]. This miRNA was predicted to target the a molecule involved with the Wnt/Wingless signaling 3′ UTR of APL1C and as APL1 is part of a complex that pathway (Table 1) [40–43]. When miR8 was depleted in stabilizes the immune factor thioester-containing protein Ae. aegypti, Swim levels remained high following blood 1 (TEP1), which binds to the surface of Plasmodium lead- feeding and egg development was inhibited [41]. Another ing to parasite destruction, miR305 may impact Plasmo- miRNA, miR34, showed differential expression in several dium infection [37]. Supporting this, miR305 depletion different mosquito species during pathogen infection, in An. gambiae led to increased resistance to both P. fal- including in An. gambiae where midgut expression was ciparum and Plasmodium berghei infection and altered decreased following an infectious Plasmodium berghei (P. the levels of many immunity or anti-Plasmodium genes berghei) blood meal [22, 34, 44–46]. Specifically, miR34 in mosquito midguts [43]. The final miRNA, miR375, Table 1 Selected miRNAs and miRNA sponge along with selected target gene transcripts Construct Target gene transcripts and Target gene functions Target gene binding Study species directionality miR8 ↓ Swim Egg development Demonstrated Ae. aegypti An. gambiae miR8SP ↑ Swim Egg development Demonstrated Ae. aegypti ↓ miR8 An. gambiae miR34 MISO Egg development Predicted Ae. albopictus Cactus Toll immune pathway regulation Ae. aegypti Rel1 An. gambiae An. stephensi Caspar IMD immune pathway regulation Rel2 miR305 APL1C TEP1 regulation and P. falciparum immunity Predicted An. gambiae Ae. albopictus miR375 ↑ Cactus Toll immune pathway regulation Demonstrated Ae. aegypti ↓ Rel1 Caspar IMD immune pathway regulation Predicted Rel2 Many miRNA targets are predicted as in vitro and in vivo functional studies are lacking. Directionality, as noted through arrows, indicates an increase or decrease in target gene transcript levels following miRNA or sponge (SP) expression. Predicted target transcripts have no directionality as they have not yet been studied. The binding column notes whether this binding has been predicted or whether it has been observed in functional studies. Study species are noted to indicate which species these miRNAs and targets have been examined or predicted in J ohnson et al. Parasites & Vectors (2025) 18:355 Page 5 of 16 was only detected in blood fed Ae. aegypti mosquitoes variable region that is commonly found in commercial and was predicted to bind to the 5′ UTR of Toll path- vectors such as in the pRL-CMV plasmid from Promega way immune genes Cactus and Rel1 [49]. Expression (Fig.  1A) [54–56]. This splice donor and splice accep- of a miR375 mimic in Ae. aegypti mosquitoes or cells tor site were initially chosen within AgDNV to attempt led to binding of the 5′ UTRs of Cactus and Rel1 and the creation of a nondefective recombinant AgDNV, as the upregulation of Cactus and downregulation of Rel1 previously described for AaeDNV, but we later decided [49]. Similar changes in target genes and an increase in to use a co-transfection system with pWTAgDNV and a Dengue virus type 2 titers were observed in Ae. albopic- transducing plasmid with the artificial intron to create an tus Aag2 cell lines [49]. Although miR375 has not been EGFP reporter phenotype [20]. These developed splice studied in An. gambiae, this miRNA has an identical donor and splice acceptor sites were placed within the sequence to miR375 in Ae. aegypti and has also been pre- EGFP gene of pAcEGFP at positions 334 and 337, respec- dicted by Insectar to target An. gambiae Cactus and Rel1 tively, to create a reporter phenotype such that improper along with other gene transcripts including Caspar and intronic splicing or a lack of splicing would result in a Rel2 [47]. stop codon within the EGFP gene and correct splicing would result in EGFP expression (Fig.  1B). Predicted Plasmid preparation and production splicing was examined at all steps using NetGene2 and Sure 2 supercompetent E. coli cells (Agilent Technologies, the created splice donor and splice acceptor sequences 200152) were transformed as per kit instructions (SOC both had a confidence scores of 1.0, indicating a high media was substituted for NZY + media) with pAcEGFP confidence in splicing [50]. and pWTAgDNV plasmids [18]. Transformed colonies were plated on Luria broth agar plates with 100  µg/mL miRNA and sponge selection ampicillin and incubated at 37 °C overnight. Colony PCR For intronic miRNA expression, endogenous pre-miRNA was used to verify transformations and selected colonies sequences were inserted into the created intron so that were grown in 5 mL of Luria broth in a 37 °C shaker over- upon splicing, the pre-miRNA hairpin would be co- night and then preserved as glycerol stocks. Purified plas- transcriptionally processed alongside EGFP transcripts mids were produced by growing glycerol stocks in liquid [22–24]. Selected pre-miRNA sequences for An. gambiae culture as before, extracted using an Omega Bio-tek miR8, miR34, miR305, and miR375, as well as a miRNA E.Z.N.A. Plasmid DNA Kit (D6942-02), and quantified sponge against miR8 (miR8SP), were added to this devel- using a NanoDrop ND-1000 spectrophotometer. oped intron to test the co-expression system and intronic splicing mechanism. A random nonsense RNA sequence Intron design (NS) was added to the intron as a control. These miRNAs A potential splice acceptor site in WT AgDNV was iden- and the miRNA sponge were chosen, as described above, tified at position 463 of the gene encoding the viral pro- on the basis of known or predicted effects on genes tein using the neural-network-based NetGene2 predictive involved with immunity, pathogen defense, or reproduc- splicing server, which identifies transition sequences tion in An. gambiae, Ae. aegypti, or relevant mosquito cell between introns and exons [50–52]. This sequence was lines (Table  1). To test intron functionality and demon- converted from AG^ACG CAGA CAG (with “^” indicat- strate that splicing is sequence-dependent, altered splice ing the predicted splicing site) into a splice donor site by donor and splice acceptor site sequences were developed replacing the intronic portion with the starting sequence using site-directed mutagenesis of the pAcEGFPmiR8 of the second intron of An. gambiae RPS17 such that the plasmid [50]. When the splice donor site was changed by new sequence was AG^GTAG GCG CGC. This sequence a single nucleotide (in bold) from AGGTAA GTGC GC to was further modified by two base pairs to AG^GTA AGATAAG TG CGC, NetGene2 no longer identified this AGTGCGC to match the An. gambiae U1 small nuclear as a splice donor site. Similarly, when the splice acceptor RNA conserved region (Fig. 1A) [53]. This U1 sequence site was changed by one nucleotide (in bold) from TAC (GTAA GT ) represents the binding site for the U1 small TGA CATC CAC TTT GC CTTT CTC TC CACA G to TAC nuclear ribonucleoprotein which helps to form the spli- TGA CAT CCA CTT TGC CTT TCTC TC CACA T, this site ceosome [53]. A splice acceptor with the sequence TAC was no longer predicted to be a splice acceptor. TGA CATC CA CTT TGC CTT TCT CTC CACA G was cre- ated to accompany this splice donor at position 464 of the Cloning and intronic cargo gene encoding the viral protein by adding in the branch MluI and BstBI sites were introduced into the EGFP- point, polypyrimidine tract, and intronic portion from encoding gene of pAcEGFP using site-directed mutagen- the 3′ end of a chimeric human intron (last 32 nucleo- esis to create synonymous mutations. A MluI site was tides) preceding an immunoglobulin gene heavy chain created by altering position 327 of EGFP from C to G, Johnson et al. Parasites & Vectors (2025) 18:355 Page 6 of 16 and position 330 from C to T. A BstBI site was created and target gene expression experiments, cells transfected in EGFP by switching position 348 from G to A. Endog- with pWTAgDNV alone served as a control. enous An. gambiae pre-miRNA sequences from miR- base (https:// www. mirba se.o rg/) were converted to DNA Viral production and quantification and used to order g-blocks from Integrated DNA Tech- To produce virus particles for mosquito infections, nologies (IDT) [22]. The mir8SP sequence contained ten Moss55 cells were transfected with selected transduc- repeated blocks of the reverse complement of mature An. ing and helper viruses (as described) and virions were gambiae miR8. Each block was separated by four spacer extracted 3  d post-transfection by removing the media, nucleotides and the entire sponge sequence was placed washing cells with 1× phosphate-buffered saline (PBS), within the intron as with pre-miRNA sequences. A non- and suspending cells in 1  mL 1× PBS. Cells were lysed sense RNA (NS) was created using a random sequence using three cycles of freeze-thawing and centrifuged at with no matches to the An. gambiae genome or transcrip- 5000 rpm for 5 min to pellet debris. The virus-containing tome when searched using the Basic Local Alignment supernatant was collected and plasmid DNA and free Search Tool (BLAST). Each pre-miRNA, miRNA sponge, viral genomes were removed using an Ambion TURBO or nonsense RNA was coded on IDT g-blocks synthe- DNA-free kit (AM1907). DNA was extracted using an sized with flanking MluI and BstBI sites, EGFP seg- Omega Bio-tek E.Z.N.A Tissue DNA kit (D3396-02) ments to replace those removed during digestion, and the and viral genome equivalents were determined using splice donor and splice acceptor sites (Table S1; Fig. 1A). standard curves created using AgDNV-coding plasmids G-blocks were subcloned into pJet using a CloneJet PCR with a single copy of each gene-of-interest. Samples and Cloning Kit (ThermoFisher Scientific, K1231) and later standards were run using PerfeCTa SYBR Green FastMix digested using MluI and BstBI. These inserts were ligated (Quantabio, 95,072–012) on a Qiagen Rotor-Gene Q at into pAcEGFP that had also been digested with MluI and 95  °C for 2 min followed by 40 cycles of 95  °C for 10 s, BstBI, and the resulting plasmid sequences were verified. 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 Cell culture and transfections AgDNV nonstructural gene 1 (NS-RT-IIIF: CAT TCGA TC Sua5B and Moss55 An. gambiae cells were grown in 25 cc ACGG AG ACCA C, NS-RT-IIIR: GCGC TTG TC GCA plug cap flasks at 28  °C and passaged once per week at CTA AGAA AC) and a standard curve of pWTAgDNV. a 1:5 dilution with Schneider’s Drosophila media with Selected transducing viruses (vAcEGFPmiR8, vAcEGFP- 10% fetal bovine serum (FBS) v/v. For transfections, miR8SP, vAcEGFPmiR34, vAcEGFPmiR305, vAcEGFP- cells were quantified using a hemocytometer and 6 × 106 miR375, vAcEGFPNS, vAcEGFPSA, and vAcEGFPSD) cells were added to each well of a 6-well plate along with were quantified using primers against EGFP (GFP-RT- 3 mL of complete media and incubated overnight. Cells II-F497: TCA AGA TCCG CC ACAA CA TC, GFP-RT-II- were transfected at ~70–80% confluence with a 1:2 ratio R644: TTCT CG TTG GGG TCT TTGC T) and a standard of pWTAgDNV to transducing plasmid with 830  ng curve of pAcEGFP. Each production of virus consisted of pWTAgDNV and 1660 ng transducing plasmid per well a mixture of vWTAcEGFP and a transducing virus. using a Lipofectamine LTX with Plus Reagent kit (Ther- moFisher Scientific, 15338030). Briefly, plasmids were Mosquito injections added to a mix of 500 µL OptiMem media with 3 µL Plus Female An. gambiae mosquitoes (Keele strain) that were reagent and incubated at room temperature for 10 min. 3  d post-emergence were injected intrathoracically with Then, 5  µL Lipofectamine was added and tubes were 200  nL densovirus mixture containing both wild-type incubated at room temperature for 25 min before trans- vWTAgDNV and transducing virus (either vAcEGFP- fecting each well with 500 µL of this mixture. Transduc- miR8, vAcEGFPmiR8SP, vAcEGFPmiR34, vAcEGFP- ing plasmids were pAcEGFPmiR8, pAcEGFPmiR8SP, miR305, vAcEGFPmiR375, vAcEGFPNS, vAcEGFPSA, or pAcEGFPmiR34, pAcEGFPmiR305, pAcEGFPmiR375, vAcEGFPSD) using a Drummond Scientific Nanoject III pAcEGFPNS, pAcEGFPSA, and pAcEGFPSD. Cells were (3-000-207) and Drummond Scientific 10  µL microcap- incubated and imaged at 3 d post-transfection. RNA for illary tubes (3-000-210-G) pulled using a Sutter Instru- splicing validation was also gathered 3  d post-transfec- ment Co. Model P-2000 (Heat 400, Fil 4, Vel 40, Del 140 tion. Preliminary in vitro miRNA and target gene expres- Pul 140). Three biological replicates in mosquitoes were sion experiments harvested RNA at 5 d post transfection. completed. For each replicate, mosquitoes were injected For Sua5B in vitro miRNA and target gene expression, with ~106–107 transducing virus particles and 106–108 cells transfected with pWTAgDNV and pAcEGFPNS WT DNV particles (Table  S2). Mosquito treatment served as controls, whereas in Moss55 in vitro miRNA groups were kept in separate cardboard cup cages with J ohnson et al. Parasites & Vectors (2025) 18:355 Page 7 of 16 10% sugar solution w/v ad libitum until RNA extraction (Table S4) [40]. An. gambiae U6 levels served as a refer- or imaging. RNA was harvested and tested from three ence with which to compare miRNA levels. Conditions biological replicates. for miRNA qPCR reactions were 95  °C for 15  min fol- lowed by 40 cycles of 95 °C for 15 s, 60 °C (for all miRNAs RNA extractions and cDNA production during cell culture replicates as well as for in vivo miR34) For both in vitro and in vivo experiments, RNA was or 55  °C (all U6 reactions and in vivo miR375) for 60 s, extracted using an Omega Bio-tek MicroElute Total and 72 °C for 20 s. All reactions ended with a melt curve RNA Kit (R6831-02). For in vitro experiments, RNA consisting of a ramp from 55  °C to 99  °C that increased was extracted 3 d post-transfection for intronic splicing 1 °C per step. assessments or 5 d post-transfection for miRNA and tar- get gene quantification. For in vivo experiments, mosqui- Data analysis toes were individually homogenized 10  d post-injection All qPCR data was analyzed using the delta-delta Ct in lysis buffer using zinc-plated steel BB pellets (Daisy method to calculate the fold change in expression relative 0.177 cal or 4.5 mm) and a Qiagen TissueLyser II with a to reference genes (S7 for mRNA transcripts and U6 for lysis program lasting 2  min with a frequency of 30  Hz. mature miRNA quantification unless otherwise noted). Following homogenization, RNA was extracted and The fold change expression data was log2 transformed DNase treated either on the column using an Omega Bio- and a D’Agostino–Pearson omnibus K2 test was used to tek RNase-free DNase Set I kit (E1091) or following RNA assess normality in Graphpad Prism 9. If both the con- extraction using an Ambion DNA-free DNA Removal trol and experimental groups passed the normality test, Kit (AM1906). For target gene quantification or assess- a parametric unpaired two-tailed t-test assuming equal ment of intronic splicing, cDNA was synthesized using standard deviations was used to measure statistical sig- a Quantabio qScript cDNA synthesis kit (95047–500); nificance. If either or both groups failed the D’Agostino– whereas for miRNA quantification, samples were con- Pearson normality test, a nonparametric two-tailed verted to cDNA using the HighSpec option in the Qiagen Mann–Whitney test was used to compare ranks and to miScript II RT kit (218161) and diluted 1:10. assess significance. Significant P values (< 0.05) were reported on graphs. All graphs report fold change expres- Intronic splicing, miRNA expression, and target gene sion using a log2 scale. The mean and standard error quantification of the mean was reported for groups analyzed using an In vitro and in vivo intronic splicing was assessed using unpaired t-test, whereas median and 95% confidence primers spanning the intronic region (GFP-COLPCRF: intervals were shown for groups compared using a non- CTG ACCT AC GGCG TG CAG TGC, RGFP-COLPCRR: parametric two-tailed Mann–Whitney test. CGGC CA TGAT AT AGA CGT TGT GGC ). PCR products were run on 2% agarose gels and imaged using a UVP Results GelDoc-It transilluminator. Spliced transcripts resulted In vitro EGFP expression in a product of 274  bp, whereas PCR reactions using When An. gambiae Sua5B or Moss55 cells were co-trans- DNA plasmid controls or unspliced transcripts produced fected with pWTAgDNV and the original transducing variably sized amplicons depending on insert size with plasmid pAcEGFP that lacked the created intron, strong most being ~480 bp. EGFP expression was observed (AcEGFP; Fig.  2A, B). Target gene qPCR reactions were run on a Qiagen Sua5B and Moss55 cells that were co-transfected with Rotor-Gene Q using PerfeCTa SYBR Green FastMix pWTAgDNV and transducing plasmid pAcEGFPmiR8, (Quantabio, 95072–012) or an Applied Biosystems pAcEGFPmiR8SP, pAcEGFPmiR34, pAcEGFPmiR305, or 7900HTFast Real-Time PCR System with Applied Biosys- pAcEGFPmiR375 had visible EGFP expression indicative tems PowerUp SYBR Green Master Mix (A25724), with of intronic splicing 3 d post-transfection (miR8, miR8SP, conditions of 95 °C for 2 min, 40 cycles of 95 °C for 10 s, miR34, miR305, and miR375; Fig. 2A, B). EGFP expres- 60 °C for 40 s, and 72 °C for 30 s, and a melt curve with sion was also present when cells were transfected with a ramp from 55  °C to 99  °C with 1  °C change per step. pWTAgDNV and the nonsense-RNA-encoding transduc- Primers for An. gambiae Swim cDNA were developed ing plasmid pAcEGFPNS (NS RNA; Fig.  2A, B). When during this study, whereas others came from published splice donor or splice acceptor sites were mutated, EGFP studies (Table S3) [48, 57–60]. expression was not detectable in Sua5B cells 3  d post- Reactions to quantify miRNAs used Qiagen miScript co-transfection with pWTAgDNV and pAcEGFPSA, SYBR Green PCR kits (218075) and a Qiagen Rotor-Gene while greatly reduced splicing was observed in cells co- Q with a universal reverse primer and forward prim- transfected with pWTAgDNV and pAcEGFPSD, indi- ers consisting of the sequences of each mature miRNA cating that splicing was largely dependent on splice site Johnson et al. Parasites & Vectors (2025) 18:355 Page 8 of 16 Fig. 2 EGFP expression in vitro 3 d after co-transfection with pWTAgDNV and selected transducing plasmids. A EGFP expression in co-transfected Sua5B cells. B EGFP expression in co-transfected Moss55 cells. Panels are labeled with the transducing construct that was co-transfected with pWTAgDNV. Scale = 100 µm J ohnson et al. Parasites & Vectors (2025) 18:355 Page 9 of 16 sequences (SA mutant and SD mutant; Fig.  2A). In An. pAcEGFPmiR34, pAcEGFPmiR305, pAcEGFPmiR375, gambiae Moss55 cells, similar expression patterns were and pAcEGFPNS (cDNA samples 1–6; Fig.  3A). observed, with less EGFP expression occurring in Moss55 Unspliced plasmid DNA samples had PCR product sizes cells co-transfected with pWTAgDNV and pAcEGFPSD dependent on intronic length. This was ~480 bp for most than in Sua5B cells co-transfected with the same con- constructs, although the miR8SP-expressing plasmid had structs (Fig.  2B). EGFP signals were generally weaker a larger insert and an amplicon of 572 bp (plasmid sam- in Moss55 compared with Sua5B cells; however, results ples; Fig.  3A). Sua5B cells co-transfected with pWTAg- were consistent from both cell lines and indicate that DNV and constructs containing mutated splice acceptor intronic splicing is induced in a sequence specific manner or splice donor sequences (pAcEGFPSA or pAcEGFPSD) for a wide variety of pre-miRNAs, miRNA sponges, and exhibited some level of intron splicing despite absent or small RNAs in An. gambiae cell lines of varied lineage. greatly reduced visible EGFP expression (cDNA samples 7 and 8, Fig.  3A; SA mutant and SD mutant, Fig.  2A). Confirmation of in vitro intronic splicing This indicates that some transcripts are spliced despite In vitro intronic splicing was further validated using the lack of predicted splicing via NetGene2, but that this primers spanning the intronic insert. A 274 bp PCR prod- splicing may be incomplete or in a location that causes uct consistent with splicing was observed in Sua5B cells a disruption in EGFP expression due to a stop codon. 3  d post-co-transfection with pWTAgDNV and trans- Faint ~480  bp bands, indicating the presence of some ducing constructs pAcEGFPmiR8, pAcEGFPmiR8SP, unspliced transcript, were also observed in PCR reactions Fig. 3 Intronic splicing in vitro 3 d post co-transfection alongside unspliced DNA plasmid control samples. A Intronic splicing and plasmid controls in co-transfected Sua5B cells. B Intronic splicing and plasmid controls in co-transfected Moss55 cells. In both A and B, matching numbers indicate cDNA and plasmid versions of the same construct. Samples are as follows: (1) miR8, (2) miR8SP, (3) miR34, (4) miR305, (5) miR375, (6) nonsense RNA, (7) SA mutant, (8) SD mutant, (9) EGFP lacking the intron, (10) no template control. Splicing of the intron in cDNA samples resulted in a PCR product of 274 bp, whereas a lack of splicing, as observed in plasmid controls on the right side of the gel, resulted in bands of ~480 bp for most constructs and 572 bp for miR8SP in well 2 Johnson et al. Parasites & Vectors (2025) 18:355 Page 10 of 16 using cDNA from cells co-transfected with pWTAgDNV with vWTAgDNV and vAcEGFPSD exhibited weak and pAcEGFPNS (cDNA sample 6; Fig. 3A). This points EGFP expression that remained localized to the mos- to some level of splicing disruption in these constructs, quito thorax (SD mutant; Fig. 4A). In vivo intronic splic- yet, given the visible EGFP expression in cells co-trans- ing was measured as before via PCR of cDNA made from fected with pWTAgDNV and pAcEGFPNS, this may be 10  d post-injection mosquitoes. Spliced 274  bp bands explained by splicing intermediates (NS RNA; Fig.  2A). were observed in cDNA samples taken from mosqui- Both cDNA and plasmid versions of EGFP lack the intron toes co-injected with vWTAgDNV and vAcEGFPmiR8, sequence and have the same PCR product size of 274 bp vAcEGFPmiR8SP, vAcEGFPmiR34, vAcEGFPmiR305, (cDNA sample 9, plasmid sample 9; Fig. 3A). vAcEGFPmiR375, or vAcEGFPNS (cDNA samples 1–6; Similar splicing patterns were observed in Moss55 cells Fig.  4B). Faint unspliced bands were also observed in (Fig.  3B). Cells co-transfected with pWTAgDNV and mosquitoes co-injected with vWTAgDNV and vAcEGFP- transducing plasmid pAcEGFPmiR8, pAcEGFPmiR8SP, miR34, vAcEGFPmiR375, or vAcEGFPNS (cDNA sam- pAcEGFPmiR34, pAcEGFPmiR305, pAcEGFPmiR375, ples 3, 5 and 6; Fig.  4B). Plasmid controls resulted in or pAcEGFPNS all had 274 bp bands, indicative of splic- larger bands of ~480 bp for all constructs except for the ing (cDNA samples 1–6; Fig.  3B). Cells co-transfected larger miR8SP insert, which produced a band of 572 bp with pWTDNV and pAcEGFPmiR8, pAcEGFPmiR34, (plasmid samples 1–6; Fig. 4B). or pAcEGFPNS also had larger ~480  bp bands, consist- ent with some level of unspliced transcript or splicing intermediates (cDNA samples 1,3, and 6; Fig.  3B). In Preliminary in vitro work to select miRNA targets cells co-transfected with pWTAgDNV and pAcEGFPNS, Preliminary work in Sua5B cells showed that transfec- another intermediate-sized band was also present tion with pWTAgDNV and pAcEGFPmiR8 or pAcEGFP- (cDNA sample 6; Fig.  3B). When cells were co-trans- miR375 led to higher levels of miR8 and miR375, fected with pWTAgDNV and the splice acceptor mutant respectively, 5  d post-transfection (Fig. S1A and C). pAcEGFPSA, a ~480 bp band, representative of a lack of miR34 levels were reduced rather than elevated when splicing, as well as a 274 bp band, consistent with splic- Sua5B cells were transfected with pWTAgDNV and ing, was observed despite of a lack of EGFP expression pAcEGFPmiR34, perhaps indicating the processing of in transfected cells (cDNA sample 7, Fig. 3B; SA mutant, this transcript into an anti-miRNA rather than expres- Fig.  2B). Cells co-transfected with pWTAgDNV and sion of the predicted mature miRNA (Fig. S1B). Transfec- the splice donor mutant pAcEGFPSD produced a single tion with pWTAgDNV and pAcEGFPmiR305 or miR8SP strong  ~480 bp band representative of a lack of splicing did not result in any significant changes in miRNA levels despite faint EGFP expression observed in transfected (data not shown). We also observed significant upregu- cells (cDNA sample 8, Fig.  3B; SD mutant, Fig.  2B). As lation of the predicted target gene Cactus in cells trans- before, PCRs of plasmid DNA resulted in larger ~480 bp fected with pWTAgDNV and pAcEGFPmiR375 (Fig. bands for most constructs (plasmid samples 1 and 3–8, S2A). Given the lack of significant change in miRNA lev- Fig. 3B). A band of 572 bp, reflective of a larger intronic els for miR8SP and miR305, and difficulty in interpreting segment, was detected for pAcEGFPmiR8SP (plasmid results from miR8 in light of the miR8SP results, these sample 2; Fig. 3B). Plasmid DNA from pAcEGFP, as well miRNAs were not evaluated in later experiments. A pre- as cDNA from cells co-transfected with pWTAgDNV liminary in vivo experiment was also carried out where and pAcEGFP, lacked the intron and produced bands of mosquitoes 10  d post-injection with vWTAgDNV and 274 bp (plasmid sample 9, cDNA sample 9; Fig. 3B). vAcEGFPmiR34 showed significant elevation in miR34 target gene transcripts MISO, Caspar, and Rel2 despite a lack of change in the levels of miR34, Cactus, and Rel1A In vivo EGFP expression and intronic splicing (Fig. S3A–F). Isolations of vWTAgDNV and each transducing virus purified from Moss55 cells were injected into adult female mosquitoes that were 3 d old, with images of these In vivo miRNA and target gene expression mosquitoes taken 10  d later. Punctate EGFP expression After down-select, larger-scale in vivo experiments indicative of splicing was observed in the thorax and focused on mosquitoes injected with vWTAgDNV and abdomen of mosquitoes co-injected with vWTAgDNV either vAcEGPFmiR34 or vAcEGFPmiR375. Mosquitoes and vAcEGFPmiR8, vAcEGFPmiR8SP, vAcEGFPmiR34, injected with vWTAgDNV and vAcEGPFmiR34 were vAcEGFPmiR305, vAcEGFPmiR375, or vAcEGFPNS evaluated for changes in MISO, Caspar, and Rel2 tran- (Fig.  4A). Little to no EGFP expression was observed script levels, whereas mosquitoes injected with vWTAg- in mosquitoes co-injected with vWTAgDNV and DNV and vAcEGPFmiR375 were tested for differences in vAcEGFPSA (SA mutant, Fig.  4A). Mosquitoes injected Cactus and Rel1A transcript levels. J ohnson et al. Parasites & Vectors (2025) 18:355 Page 11 of 16 Fig. 4 EGFP expression and intronic splicing in 10 d post-injection mosquitoes. A Punctate expression is visible in mosquitoes co-injected with vWTAgDNV and vAcEGFPmiR8, vAcEGFPmiR8SP, vAcEGFPmiR34, vAcEGFPmiR305, vAcEGFPmiR375, and vAcEGFPNS (panels miR8, miR8SP, miR34, miR305, miR375, and NS miRNA). Little to no EGFP expression was present in mosquitoes co-injected with vWTAgDNV and vAcEGFPSA (SA mutant panel). Weak expression is visible in mosquitoes co-injected with vWTAgDNV and vAcEGFPSD (SD mutant panel). B PCR product gel measuring intronic splicing in female An. gambiae 10 d post-injection with cDNA samples (left side) alongside unspliced DNA plasmid control samples (right side). Matching numbers indicate cDNA and plasmid versions of the same construct. Samples are as follows: (1) miR8, (2) miR8SP, (3) miR34, (4) miR305, (5) miR375, (6) nonsense RNA, and (7) no template control. In vivo splicing resulted in a PCR product of 274 bp for wells 1–6 on the cDNA side. A lack of splicing in plasmid controls for wells 1–6 on the plasmid side resulted in bands of ~480 bp, or 572 bp in the case of pAcEGFPmiR8SP Similar to preliminary in vivo results, levels of miR34 The direction of this change in target gene transcript were not significantly different between mosquitoes abundance may indicate miRNA signaling through the injected with vWTAgDNV and vAcEGFPmiR34 and lessor known RNA activation pathway rather than the control mosquitoes injected with vWTAgDNV and RNAi pathway, or that this miRNA acts on an upstream vAcEGFPNS (Fig.  5A). In addition, no differences were inhibitor of Rel2 [61–63]. seen in Caspar or MISO transcript levels 10 d post-injec- When mosquitoes were injected with vWTAgDNV tion (Fig. 5B and C). However, Rel2 transcripts were more and vAcEGFPmiR375, miR375 levels were reduced abundant 10 d post-infection in mosquitoes injected with 10  d post-infection (Fig.  6A). This was unexpected vWTAgDNV and vAcEGFPmiR34 compared with those given the strong increases seen in preliminary in vitro injected with vWTAgDNV and vAcEGFPNS (Fig.  5D). Johnson et al. Parasites & Vectors (2025) 18:355 Page 12 of 16 Fig. 5 miR34 and target gene transcript expression in mosquitoes 10 d post-injection with vWTAgDNV and vAcEGFPmiR34. A Expression of miR34 was unchanged in mosquitoes that were injected with vWTAgDNV and vAcEGFPmiR34. B Expression of Caspar transcripts was not altered following injection. C Levels of MISO were also not changed 10 d post-injection. D Rel2 expression was enhanced in mosquitoes following injection of vWTAgDNV and vAcEGFPmiR34. Dashed lines indicate a fold change of 0. Green dots represent individual mosquitoes co-injected with vAcEGFPmiR34 and pWTAgDNV, whereas blue squares indicate individual mosquitoes co-injected with control vAcEGFPNS and vWTAgDNV. Data in A–D were normal as assessed by a D’Agostino–Pearson normality test and were analyzed using a two-tailed unpaired t-test experiments and may indicate that the introduced increase in Cactus and decrease in Rel1A from Ae. miRNA is not processed in vivo as expected, and aegypti (Table 1). instead produces an anti-miRNA that binds to endog- enous miR375. Both Cactus and Rel1A transcript lev- Discussion els were elevated in mosquitoes 10  d post-infection These experiments show that splicing of the developed (Fig.  6B and C). This contrasts with the predicted AgDNV-delivered intron is robust in vitro within two different An. gambiae cell lines of varied lineages, as J ohnson et al. Parasites & Vectors (2025) 18:355 Page 13 of 16 Fig. 6 miR375 and target gene transcript expression in mosquitoes 10 d post-injection with vWTAgDNV and vAcEGFPmiR375. A Expression of miR375 was slightly decreased in mosquitoes injected with vWTAgDNV and vAcEGFPmiR375. B Expression of Cactus transcripts was elevated in mosquitoes 10 d post-injection. C Levels of Rel1A expression was slightly increased 10 d post-injection with vWTAgDNV and vAcEGFPmiR375. Dashed lines indicate a fold change of 0. Green dots represent individual mosquitoes co-injected with vAcEGFPmiR375 and pWTAgDNV, whereas blue squares indicate individual mosquitoes co-injected with control vAcEGFPNS and vWTAgDNV. Data in A and B was not normally distributed as assessed by a D’Agostino–Pearson normality test and was analyzed using a nonparametric two-tailed Mann–Whitney test. Data in C was normally distributed and analyzed using a two-tailed unpaired t-test well as in vivo for a variety of endogenous pre-miRNA B). While not investigated in this study, this difference in sequences, one developed miRNA sponge sequence, and observed EGFP expression between cell lines may be due one random RNA sequence. Occasionally, unexpected to chronic AgDNV infection already present in Sua5B unspliced transcripts were observed in PCR assays. These cells that could enhance the viral replication and pack- likely represent splicing intermediates or pre-splicing aging efficiency of transducing constructs [13]. Moss55 transcripts, as the presence of these larger bands did cells lack natural AgDNV infection, and thus, rely solely not correlate with altered EGFP expression in vitro or in on AgDNV constructs expressed from transfecting plas- vivo (Figs.  2A and B, 4A). This demonstrates that splic- mids. Alternatively, Sua5B cells are considered to have ing is specific to the developed intronic sequence and not hemocyte-like properties, whereas Moss55 cells have an altered by the cargo sequence. Further supporting this, epithelial origin, and differences in cell lifecycle or rate of intronic splicing can be eliminated or reduced through transcription may explain the variation in EGFP intensity alteration of splice donor or splice acceptor sequences. [64–66]. Despite these differences in AgDNV infection Although mutated splice donor and splice acceptor con- status and EGFP intensity, no differences were observed structs within the two cell lines sometimes produced in the viral titers produced by the different cell lines for bands seemingly consistent with some level of splicing, mosquito injections. To better control the genetic diver- reduced or absent in vitro EGFP expression observed for sity of purified viruses, Moss55 cells were used to grow these constructs indicates that any splicing that occurs is viral stocks used for in vivo experiments. greatly suppressed, modified, or results in a stop codon as In vivo experiments largely focused on mosquitoes 10 d predicted (cDNA samples 7 and 8, Fig. 3A; cDNA sam- post-injection with vWTAgDNV and either vAcEGFP- ples 7 and 8, Fig.  3B; SA and SD mutants, Fig.  2A and miR34 or vAcEGFPmiR375. While experiments with B). Although splicing was not assessed by PCR in vivo mosquitoes injected with vWTAgDNV and vAcEGFP- for splice donor and splice acceptor mutants, mosquitoes miR34 did not result in any differences in miR34 levels, injected with vWTAgDNV and vAcEGFPSA had little to similar to preliminary results, Rel2 transcript levels were no EGFP expression, and mosquitoes injected with vWT- slightly elevated 10  d post-infection (Fig.  5A and D). AgDNV and vAcEGFPSD exhibited weak EGFP expres- Although a directionality to Rel2 changes was not pre- sion (SA and SD mutants, Fig. 4A). Thus, this intron and dicted prior to experiments, this increase in Rel2 tran- expression strategy represents a promising new method script abundance indicates that miR34 may act on Rel2 for introducing small RNAs both in vitro and in vivo. through the RNA activation pathway or that miR34 Despite the success of this expression strategy in vitro, acts on another transcript that in turn influences Rel2 some differences between cell types were observed. abundance, possibly via an upstream inhibitor of Rel2 Sua5B cells are larger in size and produced noticeably (Table  1) [61–63]. Similarly, although miR375 was pre- stronger EGFP expression than Moss55 cells (Fig. 2A and dicted to bind to Cactus and Rel1 and cause an increase Johnson et al. Parasites & Vectors (2025) 18:355 Page 14 of 16 in Cactus and a decrease in Rel1, mosquitoes that were for meaningful intervention in biological or field set- injected with vWTAgDNV and vAcEGFPmiR375 had tings. Additional studies of AgDNV viral replication slightly elevated Cactus and Rel1A transcript levels 10 d dynamics, dosing requirements for injections, and post-infection (Fig.  6B and C). This was unexpected, as organ specificity would greatly aid future work and the Cactus is a negative regulator of Rel1, and if binding of further development of this symbiont and expression miR375 results in transcript depletion though RNAi, in system as a tool for paratransgenesis. Further, it may be theory, binding would decrease the transcript levels of that a nondefective recombinant approach, such as that both. As with miR34, it is possible that miR375 is acting described in AaeDNV by Liu et. al., 2016, may provide through RNA activation or on transcripts upstream of an advantage in expression consistency that could be Cactus or Rel1. In addition, as miR375 levels were slightly explored in future AgDNV studies [20]. decreased in mosquitoes infected with vWTAgDNV and vAcEGFPmiR375, it is also possible that this miRNA is Supplementary Information being processed into an anti-miRNA and that this binds The online version contains supplementary material available at https://d oi. to mature miR375 or operates in a different way than org/1 0.1 186/ s13071- 025- 06994-7. mature miR375 (Fig. 6A). There are many possible reasons for differences Supplementary Material 1. between what we observed from in vivo expression experiments versus what was expected on the basis of Acknowledgements We thank Amelia Roma and Francine McCullough for their supportive services prior predictions (Table  1). Most importantly, although pertaining to mosquito rearing and administrative tasks. We also thank Dr. there have been some prior studies for certain miR- Gang Ning and Missy Hazen from the Pennsylvania State University Micros- NAs, the true direct targets of these An. gambiae miR- copy Core Facility for their imaging assistance. NAs have largely not been identified and most targets Author contributions have only been bioinformatically predicted (Table  1). R.M.J., K.J.M., and J.L.R. conceived the study. Y.S. designed plasmids further Further, muted changes in miRNA or target gene tran- modified in this study. K.J.M. designed initial intronic sequences that were further modified in this study. R.M.J. performed experiments. J.L.R. provided scripts could be due to organ-specific effects that are supervision and experimental oversight. R.M.J. wrote the initial draft of the diluted when analyzing the whole mosquito body. Incon- manuscript. R.M.J., H.C.M., and J.L.R. edited the initial manuscript. All authors sistencies could also be due to a lack of effective miRNA reviewed and approved the final manuscript. expression in vivo or indicate that infection levels of Funding transducing viruses were lower than those needed to This research was supported by NIH/NIAID grant R01AI128201, NSF/BIO grant induce a significant alteration in endogenous miRNAs. 1645331, USDA Hatch Project 4769, a grant with the Pennsylvania Department of Health using Tobacco Settlement Funds, and funds from the Dorothy Foehr Despite these inconsistencies between predicted and Huck and J. Lloyd Huck endowment to J.L.R. observed alterations, we did observe changes in miR375 levels and in some miR34 and miR375 target transcripts, Data availability Data supporting the main conclusions of this study are included in the indicating that the developed artificial intron has poten- manuscript. tial for use in miRNA expression and may be more suc- cessful if specific target genes are better characterized Declarations (Figs. 5D, 6A–C). Ethics approval and consent to participate Not applicable. Conclusions Future experiments should examine miRNA and tar- Consent for publicationNot applicable. get gene transcript levels within organs known to sup- port AgDNV infection, such as the midgut, ovaries, and Competing interests fat body. By examining specific organs rather than the The authors declare no competing interests. whole mosquito, noise in the system may be reduced Author details and changes in miRNAs or target genes may be more 1 Department of Entomology, Center for Vector Biology and Zoonotic Diseases, evident. The developed expression system can concep- Connecticut Agricultural Experiment Station, New Haven, CT, USA. 2 Depart- ment of Entomology, The Pennsylvania State University, University Park, PA, tually also be useful for expressing other effectors such USA. 3 Center for Marine Environmental Studies, Ehime University, Matsuyama, as small interfering RNAs (siRNAs) developed against Ehime Prefecture, Japan. 4 Department of Biological Engineering, Massachu- specific genes. By expressing siRNAs with specific setts Institute of Technology, Cambridge, MA, USA. 5 Department of Biochem- istry and Molecular Biology, The Pennsylvania State University, University Park, targets, the true capabilities of this virus and intron PA, USA. 6 Center for Infectious Disease Dynamics, The Pennsylvania State expression system could be better tested and assessed University, University Park, PA, USA. 7 Huck Institutes of the Life Sciences, The Pennsylvania State University, University Park, PA, USA. J ohnson et al. Parasites & Vectors (2025) 18:355 Page 15 of 16 Received: 12 June 2025 Accepted: 30 July 2025 24. Griffiths-Jones S, Grocock RJ, van Dongen S, Bateman A, Enright AJ. miR- Base: microRNA sequences, targets and gene nomenclature. Nucl Acid Res. 2006;34:D140–4. 25. Feng X, Wu J, Zhou S, Wang J, Hu W. Characterization and potential role of microRNA in the Chinese dominant malaria mosquito Anopheles sinensis (Diptera: Culicidae) throughout four different life stages. Cell Biosci. 2018;8:1–17. References 26. Hu W, Criscione F, Liang S, Tu Z. Micrornas of two medically important 1. World Health Organization. World malaria report 2019 WHO Reg. Off. mosquito species: Aedes aegypti and Anopheles stephensi. Insect Mol Biol. Africa. Geneva: World Health Organization; 2019. 2015;24:240–52. 2. World Health Organization. Global report on insecticide resistance in 27. Allam M, Spillings BL, Abdulla H, Mapiye D, Koekemoer LL, Christoffels A. malaria vectors: 2010–2016. Geneva: World Health Organization; 2018. Identification and characterization of microRNAs expressed in the African 3. Peter B Bloland. Drug resistance in Malaria. WHO Dep. Commun. Dis. malaria vector Anopheles funestus life stages using high throughput Surveill. Response. 2001. sequencing. Malaria. 2016;15:1–15. 4. Hughes GL, Koga R, Xue P, Fukatsu T, Rasgon JL. Wolbachia infections are 28. Asgari S. Role of micrornas in arbovirus/vector interactions. Viruses. virulent and inhibit the human malaria parasite Plasmodium falciparum in 2014;6:3514–34. Anopheles gambiae. PLoS Pathog. 2011;7:e1002043. 29. Ul Hussain M. Micro-RNAs (miRNAs): genomic organisation, biogenesis 5. Baldini F, Segata N, Pompon J, Marcenac P, Robert Shaw W, Dabiré RK, and mode of action. Cell Tissue Res. 2012;349:405–13. et al. Evidence of natural Wolbachia infections in field populations of 30. Hammond SM. Dicing and slicing: the core machinery of the RNA inter- Anopheles gambiae. Nat Commun. 2014;5:1–7. ference pathway. FEBS Lett. 2005;579:5822–9. 6. Yen PS, Failloux AB. A review: Wolbachia-based population replacement 31. Valencia-Sanchez MA, Liu J, Hannon GJ, Parker R. Control of transla- for mosquito control shares common points with genetically modified tion and mRNA degradation by miRNAs and siRNAs. Genes Dev. control approaches. Pathogens. 2020;9:1–14. 2006;20:515–24. 7. Gomes FM, Hixson BL, Tyner MDW, Ramirez JL, Canepa GE, Alves e Silva 32. Ebert MS, Sharp PA. Microrna sponges: progress and possibilities. RNA. TL, et al. Effect of naturally occurring Wolbachia in Anopheles gambiae s.l. 2010;16:2043–50. mosquitoes from Mali on Plasmodium falciparum malaria transmission. 33. Ebert MS, Neilson JR, Sharp PA. Microrna sponges: competitive inhibitors Proc Natl Acad Sci U S A. 2017;114:12566–71. of small RNAs in mammalian cells. Nat Methods. 2007;4:721–6. 8. Jeffries CL, Lawrence GG, Golovko G, Kristan M, Orsborne J, Spence K, 34. Winter F, Edaye S, Hüttenhofer A, Brunel C. Anopheles gambiae miRNAs et al. Novel Wolbachia strains in Anopheles malaria vectors from Sub- as actors of defence reaction against Plasmodium invasion. Nucleic Acids Saharan Africa. Wellcome Open Res. 2018;3:113. https:// doi. org/ 10. 12688/ Res. 2007;35:6953–62. wellco meope nres.1 4765.2. 35. Lampe L, Levashina EA. The role of micrornas in Anopheles biology—an 9. Shaw WR, Marcenac P, Childs LM, Buckee CO, Baldini F, Sawadogo SP, et al. emerging research field. Parasite Immunol. 2017;39:e12405. Wolbachia infections in natural Anopheles populations affect egg laying 36. Lampe L, Levashina EA. Microrna tissue atlas of the malaria mosquito and negatively correlate with Plasmodium development. Nat Commun. Anopheles gambiae. G3 Genes|Genomes|Genetics. 2018;8:185–93. 2016;7:11772. 37. Dennison NJ, BenMarzouk-Hidalgo OJ, Dimopoulos G. Microrna-regula- 10. Walker T, Quek S, Jeffries CL, Bandibabone J, Dhokiya V, Bamou R, et al. tion of Anopheles gambiae immunity to Plasmodium falciparum infection Stable high-density and maternally inherited Wolbachia infections in and midgut microbiota. Dev Comp Immunol. 2015;49:170–8. Anopheles moucheti and Anopheles demeilloni mosquitoes. Curr Biol. 38. Chatterjee R, Chaudhuri K. An approach for the identification of 2021;31:2310. microRNA with an application to Anopheles gambiae. Acta Biochim Pol. 11. Nanfack Minkeu F, Vernick KD. A systematic review of the natural virome 2006;53:303–9. of Anopheles mosquitoes. Viruses. 2018;10:1–21. 39. Biryukova I, Ye T, Levashina E. Transcriptome-wide analysis of microRNA 12. Fauver JR, Grubaugh ND, Krajacich BJ, Weger-Lucarelli J, Lakin SM, Fakoli expression in the malaria mosquito Anopheles gambiae. BMC Genomics. LS, et al. West African Anopheles gambiae mosquitoes harbor a taxonomi- 2014;15:557. cally diverse virome including new insect-specific flaviviruses, monon- 40. Bryant B, Macdonald W, Raikhel AS. MicroRNA miR-275 is indispensable egaviruses, and totiviruses. Virology. 2016;498:288–99. for blood digestion and egg development in the mosquito Aedes aegypti. 13. Ren X, Hoiczyk E, Rasgon JL. Viral paratransgenesis in the malaria vector Proc Natl Acad Sci U S A. 2010;107:22391–8. Anopheles gambiae. PLoS Pathog. 2008;4:e1000135. 41. Lucas KJ, Roy S, Ha J, Gervaise AL, Kokoza VA, Raikhel AS. MicroRNA-8 14. Suzuki Y, Barik TK, Johnson RM, Rasgon JL. In vitro and in vivo host range targets the Wingless signaling pathway in the female mosquito fat body of Anopheles gambiae densovirus (AgDNV). Sci Rep. 2015;5:12701. https:// to regulate reproductive processes. Proc Natl Acad Sci. 2015;112:1440–5. doi. org/ 10. 1038/s rep1 2701. 42. Swarup S, Verheyen EM. Wnt/Wingless signaling in Drosophila. Cold 15. Ren X, Hughes GL, Niu G, Suzuki Y, Rasgon JL. Anopheles gambiae denso- Spring Harb Perspect Biol. 2012;4:1–15. virus (AgDNV) has negligible effects on adult survival and transcriptome 43. Dong S, Fu X, Dong Y, Simões ML, Zhu J, Dimopoulos G. Broad spectrum of its mosquito host. PeerJ. 2014;2:e584. immunomodulatory effects of Anopheles gambiae microRNAs and 16. Ren X, Rasgon JL. Potential for the Anopheles gambiae densonucleosis their use for transgenic suppression of Plasmodium. PLoS Pathog. virus to act as an “evolution-proof” biopesticide. J Virol. 2010;84:7726–9. 2020;16:1–20. 17. Johnson RM, Rasgon JL. Densonucleosis viruses (‘densoviruses’) for mos- 44. Jain S, Rana V, Tridibes A, Sunil S, Bhatnagar RK. Dynamic expression quito and pathogen control. Curr Opin Insect Sci. 2018;28:90–7. of miRNAs across immature and adult stages of the malaria mosquito 18. Suzuki Y, Niu G, Hughes GL, Rasgon JL. A viral over-expression system for Anopheles stephensi. Parasit Vectors. 2015;8:1–20. the major malaria mosquito Anopheles gambiae. Sci Rep. 2014;4:5127. 45. Su J, Wang G, Li C, Xing D, Yan T, Zhu X, et al. Screening for differentially 19. Afanasiev BN, Kozlov YV, Carlson JO, Beaty BJ. Densovirus of Aedes aegypti expressed miRNAs in Aedes albopictus (Diptera: Culicidae) exposed to as an expression vector in mosquito cells. Exp Parasitol. 1994;79:322–39. DENV-2 and their effect on replication of DENV-2 in C6/36 cells. Parasit 20. Liu P, Li X, Gu J, Dong Y, Liu Y, Santhosh P, et al. Development of non- Vectors. 2019;12:1–12. defective recombinant densovirus vectors for microRNA delivery in the 46. Liu Y, Zhou Y, Wu J, Zheng P, Li Y, Zheng X, et al. The expression profile of invasive vector mosquito, Aedes albopictus. Sci Rep. 2016;6:1–13. Aedes albopictus miRNAs is altered by dengue virus serotype-2 infection. 21. Bartel DP. MicroRNA target recognition and regulatory functions. NIH Cell Biosci. 2015;5:1–11. Public Access. 2009;136:215–33. 47. Alhaj Ali MAM. Identification and characterization of microRNAs and their 22. University of Manchester. miRBase: The microRNA database. https://w ww. putative target genes in Anopheles funestus s.s. Thesis at University of the mirbas e. org/. 2013. Western Cape. 2013. 1–195. 23. Kozomara A, Griffiths-Jones S. miRBase: integrating microRNA annotation 48. Baldini F, Gabrieli P, South A, Valim C, Mancini F, Catteruccia F. The and deep-sequencing data. Nucleic Acids Res. 2011;39:D152–7. interaction between a sexually transferred steroid hormone and a female protein regulates oogenesis in the malaria mosquito Anopheles gambiae. Johnson et al. Parasites & Vectors (2025) 18:355 Page 16 of 16 PLoS Biol. 2013;11: e1001695. https:// doi. org/ 10. 1371/ journ al. pbio. 10016 95. 49. Hussain M, Walker T, O’Neill SL, Asgari S. Blood meal induced microRNA regulates development and immune associated genes in the dengue mosquito vector, Aedes aegypti. Insect Biochem Mol Biol. 2013;43:146–52. 50. DTU Bioinformatics: Department of Bio and Health Informatics. Net- Gene2. 2017. 51. Hebsgaard SM, Korning PG, Tolstrup N, Engelbrecht J, Rouzé P, Brunak S. Splice site prediction in Arabidopsis thaliana pre-mRNA by com- bining local and global sequence information. Nucleic Acids Res. 1996;24:3439–52. 52. Brunak S, Engelbrecht J, Knudsen S. Prediction of human mRNA donor and acceptor sites from the DNA sequence. J Mol Biol. 1991;220:49–65. 53. Chang LW, Tseng IC, Wang LH, Sun YH. Isoform-specific functions of an evolutionarily conserved 3 bp micro-exon alternatively spliced from another exon in Drosophila homothorax gene. Sci Rep. 2020;10:1–13. 54. Promega Corporation. pRL Renilla luciferase reporter vectors TB550. Madison: Promega Corporation; 2008. p. 1–8. 55. Bothwell ALM, Paskind M, Reth M, Imanishi-Kari T, Rajewsky K, Balti- more D. Heavy chain variable region contribution to the NPb family of antibodies: somatic mutation evident in a y2a variable region. Cell. 1981;24:625–37. 56. Senapathy P, Shapiro MB, Harris NL. Splice junctions, branch point sites, and exons: sequence statistics, identification, and applications to genome project. Methods Enzymol. 1990;183:252–78. 57. Frolet C, Thoma M, Blandin S, Hoffmann JA, Levashina EA. Boosting NF-κB-dependent basal immunity of Anopheles gambiae aborts develop- ment of Plasmodium berghei. Immunity. 2006;25:677–85. 58. League GP, Estévez-Lao TY, Yan Y, Garcia-Lopez VA, Hillyer JF. Anopheles gambiae larvae mount stronger immune responses against bacterial infection than adults: evidence of adaptive decoupling in mosquitoes. Parasit Vectors. 2017;10:1–19. 59. Meister S, Kanzok SM, Zheng XL, Luna C, Li TR, Hoa NT, et al. Immune signaling pathways regulating bacterial and malaria parasite infec- tion of the mosquito Anopheles gambiae. Proc Natl Acad Sci U S A. 2005;102:11420–5. 60. Dong Y, Taylor HE, Dimopoulos G. AgDscam, a hypervariable immuno- globulin domain-containing receptor of the Anopheles gambiae innate immune system. PLoS Biol. 2006;4:1137–46. 61. Huang V, Qin Y, Wang J, Wang X, Place RF, Lin G, et al. RNAa is conserved in mammalian cells. PLoS ONE. 2010;5:e8848. 62. Vasudevan S, Tong Y, Steitz JA. Switching from repression to activation: microRNAs can up-regulate translation. Science. 2007;318:1931–4. 63. Vasudevan S. Posttranscriptional upregulation by microRNAs. WIREs RNA. 2012;3:311–30. 64. Dimopoulos G, Richman A, Müller HM, Kafatos FC. Molecular immune responses of the mosquito Anopheles gambiae to bacteria and malaria parasites. Proc Natl Acad Sci U S A. 1997;94:11508–13. 65. Rasgon JL, Ren X, Petridis M. Can Anopheles gambiae be infected with Wolbachia pipientis? Insights from an in vitro system. Appl Environ Micro- biol. 2006;72:7718–22. 66. Marhoul Z, Pudney M. A mosquito cell line (Mos. 55) from Anopheles gam- biae larvae. Trans R Soc Trop Med Hyg. 1972;66:183. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations.