RNA interference-mediated knockdown of voltage-gated sodium channel (MpNav) gene causes mortality in peach-potato aphid, Myzus persicae

Voltage-gated sodium channels (VGSC) are transmembrane proteins that generate an action potential in excitable cells and play an essential role in neuronal signaling. Since VGSCs play a crucial role in nerve transmission they have become primary targets for a broad range of commercial insecticides. RNA interference (RNAi) is a valuable reverse genetics tool used in functional genomics, but recently, it has also shown promise as a novel agent that could be used to control agricultural insect pests. In this study, we targeted the VGSC (MpNav) gene in the peach-potato aphid Myzus persicae, by oral feeding of artificial diets mixed with dsRNAs. Knock-down of MpNav gene expression caused up to 65% mortality in 3rd instar nymphs. Moreover, significantly lower fecundity and longevity was observed in adult aphids that had been fed with dsMpNav solution at the nymphal stage. Analysis of gene expression by qRT-PCR indicated that the aphid mortality rates and the lowered fecundity and longevity were attributable to the down-regulation of MpNav by RNAi. Taken together, our results show that MpNav is a viable candidate target gene for the development of an RNAi-based bio-aphicide.

degree of specificity, RNAi is also being widely explored as a potential novel pest control strategy for a variety of pest species 10 , with the perceived benefits of increased species discrimination and decreased risk to the environment and non-target species 11 .
Unusually in M. persicae (and other aphids), the functioning VGSC is encoded by two genes (NCBI Accessions FN601405 and FN601406) 12 with some unique properties that are not present in the VGSCs of other insects. Unlike the channels of other insects, the aphid has a unique heterodimeric channel (composed of two subunits, H1 and H2, encoding DI-II and DIII-IV of the VGSC respectively) with an atypical ion selectivity filter (similar that found in the mammalian sodium sensor channel Na x ) 13 , which, atypically for insect VGSCs, is extremely insensitive to tetrodotoxin.
Abd El Halim et al. 14 recently reported that RNAi-mediated knock-down of VGSC gene expression, through application of complementary dsRNA, caused significantly high larval mortalities and severe developmental arrest in the red flour beetle Tribolium castaneum, a coleopteran insect pest that has proved to be particularly amenable to RNAi 15,16 . In comparison levels of gene-knock down and systemic RNAi responses (following injection or ingestion of dsRNA) in many other insect classes are extremely variable 17 . In particular, RNAi outcomes in phloem sap feeding hemipteran species such as aphids, whitefly, psyllids and plant hoppers can be exceedingly disparate, ranging from no phenotype to significant mortality and from very low to complete gene knock-down [18][19][20] . Hemipteran species are therefore more challenging to work with as they appear to be somewhat intractable to RNAi manipulations 21 . In this study, we demonstrate that oral delivery of MpNa v dsRNA to the peach-potato aphid M. persicae successfully down-regulates the expression of the aphids VGSC and causes significant nymphal mortalities. Moreover, lower fecundity and longevity was also observed in dsRNA treated insects. These results suggest that RNAi targeting the VGSC could be a promising novel bio-pesticide against this hemipteran pest.

Materials and Methods
Insect culture. Myzus persicae were collected from a cabbage field in Mardan, Khyber Pakhtunkhwa, Pakistan, and established as a laboratory colony that was maintained on cabbage plants (variety Golden Acre) under standard laboratory conditions (25 ± 2 °C, 70 ± 10% relative humidity, 12:12 (light: dark) photoperiod). total RNA extraction and cDNA synthesis. Total RNAs were isolated from 1 st , 2 nd , 3 rd , 4 th instar nymphs and adults of M. persicae using Wizol ™ Reagent (Wizbiosolutions Inc., Korea) following the manufacturer's recommended protocol. RNA integrity was analyzed on 1.5% (w/v) agarose gels as described in Sambrook and Russell 22 . cDNA's were synthesized using 1 μg total RNA with WizScript ™ First Strand cDNA synthesis kit (Wizbiosolutions Inc. Korea), according to the manufacturer's recommended protocol. synthesis of dsRNA molecules. For RNAi experiments, a 289 bp fragment of the M. persicae MpNa v (heteromer H1) gene (NCBI Accession FN601405) and a 329 bp fragment of the Aequorea victoria green-fluorescent protein (GFP) gene (NCBI Accession M62653) were amplified from cDNA obtained from adult aphids using PCR. The T7 promoter sequence 5′GGATCCTAATACGACTCACTATAGGA3′ was added in front of the forward and reverse PCR primers as required for subsequent dsRNA synthesis (Table 1). Primers were chosen based on the results of GeneScripts primer design software (https://www.genscript.com/tools/pcr-primers-designer). The reaction mixture for PCR contained 500 ng of the cDNA template, 0.5 µm of forward and reverse primers, 1.75 mM Mg 2+ , 0.25 mM of each dNTP (Takara Bio, Japan), 1X PCR buffer, 5U/µl Taq DNA polymerase (Takara Bio) and double-distilled H 2 O to make a 25 µl total reaction volume. The PCR conditions were 95 °C for 5 min, followed by 35 cycles of 95 °C for 30 s, 55-60 °C for 30 s and 72 °C for 30 s, and an additional final polymerization step of 72 °C for 5 min. PCR products were purified using TIANgel Midi Purification Kit (Tiangen, Beijing, China). These purified products were then used to synthesize dsRNA using the T7 RiboMAX ™ Express RNAi System (Promega, US), following the manufacturer's protocol. The dsRNA quality was monitored on agarose gel electrophoresis, and the concentration was determined by spectrophotometry (Nano Drop 1000, Thermo Scientific, US). dsRNA products were stored at −80 °C prior to further use.
Dietary delivery of the double-stranded RNA to nymphs. The artificial feeding diet for M. persicae was prepared as described by Pan et al. 23 , albeit with a 20% lower sucrose content 24 . In the RNAi diet, dsMpNa v or dsGFP RNA, or DEPC water was mixed in and fed to 3 rd instar nymphs. To rear aphids on this artificial diet, MpNa v expression analysis by Quantitative Real-time pCR. Quantitative reverse transcription PCRs (RT-qPCR) was performed to analyze the expression level of MpNa v (heteromer H1). Primers for the MpNa v H1, GFP and Actin genes were designed online using Primer-BLAST 26 (https://www.ncbi.nlm.nih.gov/tools/ primer-blast/). qRT-PCR reactions were performed using iTaq ™ Universal SYBR Green Supermix (Bio-Rad) in a Bio-Rad iCycler (Bio-Rad, Hercules, CA, USA), following the manufacturer's instructions. PCR conditions were 95 °C for 10 min, 40 cycles of 95 °C for 15 s, 60 °C for 30 s and 72 °C for 30 s. All the analyses were repeated in triplicate. Relative expression levels were calculated using 2−△△Ct method 27 . Actin was used as the internal control 28 . All the primers used in this study (Table 1) were designed to avoid the homologous ORF region used for the synthesis of dsRNA.
Longevity and fecundity analysis. After continuous feeding of dsRNA for 7 days, each survivor aphid was transferred to a cabbage leaf placed on a freshly prepared agar plate. Plates were kept under laboratory condition as described above. The number of new born aphid nymphs from each plate was recorded daily until the aphid died. The nymphs produced by the aphids were removed from the dishes after counting. The cabbage leaf was replaced every alternate day with a fresh leaf. statistical analysis. The data were statistically analyzed with GraphPad prism 5.0. The results were analyzed using One-way analysis of variance (ANOVA) with subsequent Tukey Kramer multiple comparison. For all tests, P < 0.05 was considered significant.
Results temporal expression of MpNa v gene. The relative abundance of MpNa v H1 mRNA at specific developmental stages of M. persicae was estimated by qRT-PCR. Transcripts were detected in all life stages investigated. However, MpNa v H1 was most abundantly expressed (P < 0.005) in 3 rd and 4 th instar nymphs and adults (Fig. 1). There was no significant difference in the level of MpNa v gene expression in the 3 rd instar, 4 th instar, and adult aphids. Based on this, 3 rd instar nymphs were selected as a suitable developmental stage for RNAi studies.
Bioinformatic analysis of targeted sequence. Homology of the selected MpNa v dsRNA fragment to other insect Na v sequences, including representatives from important pollinator and beneficial species, was investigated using the NCBI BLASTn algorithm. The BLASTn search indicated a high degree of similarity between the MpNa v dsRNA fragment and the Na v coding sequence in other aphid species (overall homology scores www.nature.com/scientificreports www.nature.com/scientificreports/ ranging between 95% identity, e-value 5e-125 to the pea aphid Acyrthosiphon pisum to 84% identity, e-value 3e-83 to the sugar-cane aphid Sipha flava). For important pollinating and beneficial insects, Blastn scores ranged from 74% identity, e-value 2e-16 identity for the common eastern bumble bee, Bombus impatiens, 73% identity, e-value 1e-13 for the common honey bee, Apis mellifera, and 71% identity, e-value 1e-12 for the parasitic wasp Trichogramma pretiosum.
MpNa v expression after dsRNA feeding. The silencing efficiency of dsMpNa v was examined using real-time quantitative PCR. We analyzed expression of MpNa v at daily intervals following continuous oral delivery of dsRNA. A direct correlation was observed between the amounts of dsRNA ingested and a consequent decrease in the abundance of MpNa v mRNA transcript. At the 2nd day of ingestion, MpNa v dsRNA caused significant (P < 0.01) reduction in MpNa v mRNA abundance. During the 3 rd to 7 th day of dsMpNa v feeding the MpNa v transcript level showed very significant differences (P < 0.001, 2.5-fold decrease) compared with the DEPC water and dsGFP control treatments, indicating a substantial gene knockdown (Fig. 2).
Effects of dsRNA on aphid development and mortality. No significant aphid mortality was observed for the DEPC water, dsGFP and dsMpNa v treatments after the first day of feeding. The mortality differences between treatment become more and more obvious from the 2 nd day of feeding onwards. The average mortality for dsMpNa v treatments reached 34.7%, 43.6%, 58.2%, 60.0%, 63.6% and 65.7% respectively after the second, third, fourth, fifth, sixth and seventh day of continuous feeding, whereas the mortality of the DEPC water treated group was 5.5%, 4.3%, 6.2%, 5.0%, 8.1% and 8.4% respectively, whilst the mortality for dsGFP oral feeding treatments was 3.8%, 6.2%, 4.8%, 5.7%, 7.6% and 8.1% respectively (Fig. 3). Relative to the -ve (DEPC water) and +ve (dsGFP) control diets, which exhibited only a small decrease in survival over an assay period of 7 days, the dsMpNa v containing diet resulted in a significant (p < 0.005) impact on aphid mortality between days 3 and 7. At day 7, any surviving insects were transferred to cabbage leaves and their longevity and fecundity evaluated. Prior dietary delivery of dsMpNa v significantly decreased (p < 0.01, p < 0.05) the longevity and fecundity of M. persicae survivors compared to control treatments (Figs 4 and 5).

Discussion
For electrical signaling in eukaryotes, VGSCs are essential and ubiquitous 29 . Insect VGSCs are normally encoded by a single gene (para in Drosophila melanogaster and its equivalent orthologs in other insect species), that can generate a large number of transcriptional editing and splice variants that are differentially expressed at various developmental stages within the life cycle and in specific cell types 30 . Unusually in M. persicae (and other aphids), the functioning VGSC is encoded by two genes (designated heteromer H1 and H2) with some unique channel properties 13 . These two heteromer's are hypothesized to have arisen due to a gene inversion event having occurred at some time during the evolution of aphids, resulting in a novel two-subunit channel.
Abd El Halim et al. 14 recently demonstrated that RNAi-mediated knockdown of a VGSC gene (TcNa v ) by oral feeding causes up to 51.34% larval mortality and developmental arrest in the red flour beetle T. castaneum. This dose dependent larval mortality was observed after the beetles were continuously fed dsRNA for 6 days. Moreover, adult emergence was also significantly reduced in insects fed with dsRNA. Our results for the peach-potato aphid channel MpNa v are on a par with this study, with up to 63.6% mortality observed for aphid nymphs by day 6 of exposure to dsRNA targeted against the H1 subunit. In our experiments, knock-down of the MpNa v gene also significantly reduced adult longevity (by up to 7 days) and the fecundity (a decline of approximately 45%) of survivors. A previous study in Drosophila melanogaster has reported that decreased levels of the www.nature.com/scientificreports www.nature.com/scientificreports/ para VGSC in mle napts (no-action potential temperature sensitive mutation of the maleless (mle) gene) mutant flies resulted in decreased longevity and fecundity 31 . mle napts is a recessive gain-of-function mutation of mle that results in a splicing defect of the Na + channel transcript and a severe reduction of VGSC RNA levels and channel activity. Another recent RNAi study targeting the VGSC in the bird cherry-oat aphid Rhopalosiphum padi 32 , a global pest of wheat, indicated significant suppression of the transcript levels of heteromers H1 and H2, in this case by direct injection (rather than oral feeding) of their respective dsRNA, and a significant cross-suppressions in the transcript levels between H1 and H2 subunit genes. Although in our investigations we did not analyse expression of the M. persicae H2 subunit in response to application of H1 dsRNA, the likelihood is that significant cross-suppression will also have occurred.
If we compare the regions targeted for RNAi suppression in these three individual studies, whereas in T. castaneum a 239 nt dsRNA covering the membrane spanning regions S1 and S2 in Domain I of the channel (GenBank accession NM_001165908.1) was used, the two separate aphid studies (M. persicae and R, padi) employed dsRNAs encompassing the DI-DII linker region of heteromer H1. For R. padi H1 (GenBank accession KJ872633) a larger 485 bp Rpvgsc1 fragment was amplified. However, in both aphid studies a common region of 71 amino acids (214 nucleotides) was covered in the amplified fragments. In both cases similar levels of suppression were reported, suggesting that this region is a good target for RNAi suppression technology. In the R. padi study a 358 bp dsRNA fragment encompassing the DIV S4-S5 and S5-S6 linker region (Rpvgsc2 GenBank  www.nature.com/scientificreports www.nature.com/scientificreports/ accession no. KP966088) of heteromer H2 (which is the more evolutionarily divergent gene when compared to classical monomeric VGSCs) was also selected for suppression studies. It is not, however, clear from these experiments which region of the gene (coding region 3′ or 5′ end) is ideal for dsRNA design, and in fact it may not really be that important e.g. in the pea aphid Acyrthosiphon pisum, no difference in mortality was observed in groups of insects fed with dsRNA matching either the 5′ or 3′ end of the hunchback (hb) gene 25 .
Unintentional gene silencing in non-target species 33 is the primary risk posed by pesticidal RNAi. Bioinformatic analyses that compares the pesticidal RNAs to non-target genomes has been recognised as an useful initial screen that can help to predict potential non-target risks posed by RNAi 33,34 . In the present study, we used in-silico searches to determine whether our 289 bp dsRNA shared prohibitive sequence similarities with the genes from key insect pollinators and beneficial insects. We obtained somewhat similar homology scores in the range 73-74% for the bee species represented in the NCBI nucleotide database and the highest score for a beneficial insect (predatory wasp) was 71%. In comparison, other aphid species returned high homology scores of 84-95%, suggesting that in practice the dsRNA used in this study may suppress the VGSC expression of more than one aphid type. No significant hits were obtained when the off-target search algorithm dsCheck 35 was used to identify exact and near nucleotide matches of the 289 bp M. persicae dsRNA to the genome of the model insect D. melanogaster. However, even considerable sequence divergence between an mRNA and the dsRNA does not rule-out unintentional gene silencing since, once ingested, the dsRNA is cleaved into numerous very short (19-23 nucleotides) small interfering RNAs (siRNA). These siRNAs most likely have abundant direct sequence matches throughout most eukaryotic and prokaryotic genomes, thereby increasing the chances for off-target (i.e. silencing of a gene with sufficient sequence similarity to the dsRNA) and non-targeted (silencing of the intended gene in an unintended organism) effects 36 .
E-RNAi 37 analysis used for design of the dsRNA used in the T. castaneum study 14 predicted over two hundred 19 nt siRNAs could be generated from the 239 bp dsRNA fragment (albeit with an average efficiency score of 54.16, and no off-target predictions); there is thus considerable scope for unforeseen off-target effects. Most studies report that dsRNA ranging from 140 to 500 nucleotides in length are required for successful RNAi in insects, although successful suppression with dsRNA of 50 bp has been reported 38 . This finding could be helpful to further optimize the dsRNA specificity by using shorter amplified dsRNA fragments, thereby decreasing the likelihood of non-target and off-target effects. The success of RNAi is also dependent on the molecules stability and an effective uptake of the dsRNA by the target species 39 . A rapid degradation of dsRNA by extracellular ribonucleases in the insect haemolymph and gut is increasingly recognised as a fundamental factor influencing RNAi efficiency in several insect orders 40 . This is exacerbated in hemipteran species since extra-oral salivary degradation of dsRNAs provides an additional block to cellular uptake 17,21,41 . To translate RNAi to field applicability, RNAi silencing elicitors will most likely need to be combined with biotic or abiotic systems that mediate both protection and uptake of the eliciting RNAi trigger 39 .
To conclude, our study demonstrates that silencing of a voltage-gated sodium channel gene (MpNa v ) via RNAi in M. persicae caused larval mortality. It provides evidence to propose that MpNa v is a feasible candidate gene to target for the control of this insect pest using RNA interference technology and provides the foundation to design dsRNA molecules and associated delivery systems with a high degree of species specificity that could ultimately be used as commercial bio-aphicides.

Data Availability
All data generated or analysed during this study are included in this published article.