RNA Interference-Based Pesticides and Antiviral Agents: Microbial Overproduction Systems for Double-Stranded RNA for Applications in Agriculture and Aquaculture

Abstract

RNA interference (RNAi)-based pesticides are pest control agents that use RNAi mechanisms as the basis of their action. They are regarded as environmentally friendly and are a promising alternative to conventional chemical pesticides. The effective substance in RNAi-based pesticides is double-stranded RNA (dsRNA) designed to match the nucleotide sequence of a target essential gene of the pest of concern. When taken up by the pest, this exerts an RNAi effect and inhibits some vital biochemical/biological process in the pest. dsRNA products are also expected to be applied for the control of viral diseases in aquaculture by RNAi, especially in shrimp farming. A critical issue in the practical application of RNAi agents is that production of the dsRNA must be low-cost. Here, we review recent methods for microbial production of dsRNAs using representative microorganisms (Escherichia coli, Pseudomonas syringae, Corynebacterium glutamicum, Chlamydomonas reinhardtii, and others) as host strains. The characteristics of each dsRNA production system are discussed.

1. Introduction

As the global population and living standards grow, production of abundant, high-quality, safe food is increasingly important. However, pests and pathogens pose serious threats to agriculture and aquaculture. The protection of agricultural products from crop pests, especially pest insects, by chemical pesticides has played a central role in securing food resources. However, various disadvantages of many conventional chemical pesticides have become apparent. For example, because of their broad-spectrum of action, they have a negative impact on beneficial insects such as pollinators and predatory ladybeetles, and this poses a challenge to biodiversity. In general, synthetic chemical pesticides are slow to decompose in the natural environment, and thus there are concerns about the health of ecosystems because of residual chemical pesticides. The use of extremely large amounts of chemical pesticides over a long time has led to the emergence of pests with acquired resistance to the chemicals. Moreover, synthetic chemical pesticides have been extensively studied and commercialized, hence, the development of innovative synthetic pesticides with novel chemical structures is no longer as easy as it used to be. Furthermore, in aquaculture, especially shrimp farming, concerns have been raised about the ecological impact of the use of large amounts of chemical antibiotics and synthetic antiviral agents that can induce the emergence of drug-resistant pathogens, and thus the development of more natural and environmentally friendly disease-control agents is coveted. In the light of this situation, the 2020 report “Farm to Fork Strategy” [1] of the European Union (EU), called for halving the overall use and risk of chemical pesticides, and the use of antibiotics in livestock animals and aquaculture by 2030 [2,3]. One of the most promising alternatives is RNAi-based pesticides and RNAi reagents for viral diseases.

The functional ingredient of such RNAi-based products is an RNA molecule with a double-stranded structure (dsRNA). Its action is based on the mechanism of RNA interference (RNAi), which specifically represses the expression of a target gene in a nucleotide sequence-specific manner [4,5] when the dsRNA is taken up into a cell with RNAi capability (Figure 1). In other words, RNAi-based pesticides function by inhibiting the expression of a gene that is essential for the survival of the target crop pest through the RNAi effect, thereby suppressing vital biology of the pest and even causing death [6,7]. Thus, such pesticides are also referred to as “RNA-induced gene silencing pesticides”. RNAi effects by topical dsRNA application have also been studied to control the infection and proliferation of pathogens including fungi, viruses, and viroids in plants [8,9,10,11,12].

The application of RNAi-based pesticides and antiviral agents can be roughly divided into two categories: the establishment of genetically modified plants expressing dsRNAs that act on target pests, and the exogenous (topical) application of dsRNAs to the plant body (e.g., through foliar spray, irrigation, or trunk injection). The former is a method of pest management using transgenic crops, and the latter is a non-transformative delivery of dsRNA. The focus of this review is on the production of dsRNAs for topical administration, especially for use in foliar spraying of crops. We also discuss the production of dsRNAs in forms that can be applied in feed in shrimp farming [13].

The first advantage of using RNAi-based pesticides as insecticides is that they essentially inhibit the growth of only the target pest(s). As a nucleotide sequence that specifically suppresses the expression of a target gene in the pest to be killed or inhibited is designed, the RNAi-based pesticide should not affect unintended (non-target) insects; this approach is favorable for ecosystem biodiversity. A second advantage is that the RNA, the effective factor of the insecticide, degrades reasonably quickly in the environment because of its inherent chemical instability and the ubiquitous presence of nucleases in the bioecological community [14,15]. Again, this feature of RNAi-based pesticides is fundamentally consistent with protection of ecosystems. Nevertheless, moderate control of degradability is an important aspect of the practical application of dsRNAs in pesticides applied topically to crops [16,17]. A third valuable characteristic of RNAi-based pesticides is that their application allows for flexible and fast responses to emerging pest diseases, and pests that have become resistant to pesticides already in use. Even if the target pest becomes resistant to the RNAi-based pesticide used, a new RNAi-based pesticide can, in principle, be developed quickly by simply targeting another region of the same gene or by changing the target gene in the pest [18]. In this regard, however, the emergence of a formidable resistant pest has recently been reported due to genetic mutations involved in the underlying mechanism of RNAi action [19]; this is discussed specifically below.

It remains essential that the effects of the practical application of these innovative RNAi-based pesticides and antiviral agents on human health and ecosystems are evaluated (just as with conventional pesticides). However, the immediate practical issues for development and application of this technology are: (i) the requirement for substantial reduction of the production costs; (ii) improvement of administration methods to target organisms such as crop pests (this is also related to the product cost issue), and (iii) selection of RNA sequences that do not cause off-target effects, i.e., that do not exert effects on genes other than the specific gene of the target pest. The most urgent challenge is reduction of dsRNA production costs to a level where RNAi-based products can be put into practical commercial use. The amount of RNA required for use in crop fields and aquaculture ponds is enormous, and thus, ways must be found to produce extremely large amounts of dsRNA (with designed nucleotide sequences of a few hundred base pairs) in an economically viable way.

This review presents an overview of recent production methods of dsRNA (including hairpin-type RNA species, hpRNA) using the representative microorganisms Escherichia coli, Pseudomonas syringae, Corynebacterium (C.) glutamicum, and Chlamydomonas (Chl.) reinhardtii as host strains, as well as research results using some other microorganisms. In E. coli, there are many reports of the production of hpRNAs with partial double-stranded structures and dsRNAs using the coliphage T7 transcription system. A unique system using the combination of P. syringae and its bacteriophage phi6 has been developed to increase the stability of dsRNA products. A high-level dsRNA production system has recently been established using the industrial microorganism C. glutamicum, a “Generally Recognized as Safe” (GRAS) strain. Research results for dsRNA production using the photosynthetic microalgae Chl. reinhardtii have been reported, especially for application in the aquaculture industry. Besides those organisms, recent research cases of dsRNA-production using various yeasts, lactic acid bacteria, and symbiotic microbes have been reported for applications in pest control.

2. Production of Recombinant dsRNA in Microorganisms

2.1. Escherichia coli

The combination of E. coli strain HT115(DE3) (which has a transposon inserted into the rnc gene that encodes the dsRNA-degrading enzyme RNase III) and expression plasmid L4440, which was first used for a successful RNAi experiment in the nematode Caenorhabditis elegans in 1998 [20,21], has since become one of the standard methods for dsRNA preparation in the laboratory. Deletion of the gene encoding RNase III was a prerequisite for dsRNA of interest to accumulate in the bacterium. The HT115(DE3) strain carries the λDE3 lysogen to allow the T7 RNA polymerase gene to be expressed under the control of the lacUV5 promoter. Therefore, the addition of IPTG (isopropyl β-D-1-thiogalactopyranoside), an inducer of gene expression from the lacUV5 promoter, results in the production of T7 RNA polymerase, which in turn leads to expression of the target RNA gene from the T7 promoter in the RNA expression unit carried on plasmid L4440, thus resulting in the accumulation of target RNA in the cells. In general, there are two modes of expression in dsRNA production in microorganisms (Figure 2). One is to insert the target dsRNA coding region between two oppositely directed promoters (usually T7 promoters are used) and produce the target dsRNA molecule by convergent transcription, as in the case of RNA expression in L4440. The other expression mode is to design the nucleotide sequence to form a hairpin-loop structure within a single-stranded RNA transcript after transcription from a single strong promoter to produce an hpRNA containing the dsRNA segment of interest.

One of the pioneering works using in vivo-produced dsRNA against plant viral diseases was performed by Tenllado et al. in 2003 [22]. To suppress the infectivity of Pepper mild mottle virus (PMMoV) that infects the plant Nicotiana benthamiana, they constructed a system in which dsRNA (hp-type RNA, 977-bp long) corresponding to the partial sequence of the PMMoV replicase gene was highly expressed under the control of the T7 promoter using E. coli HT115(DE3) as the host. The productivity of the hpRNA was 4 μg per milliliter of culture of the producer strain. The producing E. coli cells were disrupted using a French press, and the centrifugal supernatant (as crude extract) was tested for its ability to interfere with viral infection. It was shown that the infectivity of the virus was significantly decreased when plants were sprayed with the lysate several days before viral inoculation [22]. There are now numerous reports on preparation of various dsRNAs or hpRNAs using the combination of strain HT115(DE3) and the T7 promoter-directed expression system [23,24,25,26,27,28]. For example, the productivity of recombinant hpRNA (including a 400-bp dsRNA region) targeted to silence the expression of a protease gene in yellow head virus (YHV), which infects the edible shrimp Penaeus monodon, was 1.5 mg per 5 mL of culture medium [29]. The productivity of dsRNA for silencing the wing development vestigial (vg) gene in Aedes aegypti mosquitoes was reported to be 500 µg per 25 mL of culture medium [30].

RNAi-mediated silencing targeting the HvRPS18 and HvRPL13 genes encoding ribosomal proteins [31] and V-ATPase subunit A and E genes [32] has been investigated for application to the 28-spotted ladybeetle Henosepilachna vigintioctopunctata; in these studies, dsRNA expressed using the L4440 vector with E. coli HT115(DE3) as the host was employed. The use of RNAi-based products against tree pests is also being explored. Oral administration of a suspension of dsRNA-expressing E. coli to larvae of the invasive forest pest emerald ash borer (EAB) was shown to inhibit growth and reduce survival of the pest [27]. EAB larvae feed on the cambial tissue of ash trees, thus killing them and causing significant damage to forest resources in North America. As the target genes, the shibire gene (shi), which is involved in neuronal chemotransduction, and the hsp gene, which encodes a 70 kDa-heat shock protein, were selected, and dsRNA corresponding to each gene was produced by convergent transcription using the T7 promoter of L4440 in E. coli HT115(DE3), although the productivity was not specified.

Zhu et al. [33] showed in 2011 that heat-sterilized E. coli cells expressing dsRNA could be orally administered to Colorado potato beetle (CPB), a threatening crop pest, to induce RNAi and cause significant mortality in the target pest. In this case, the combination of E. coli HT115(DE3) and vector L4440 was employed to produce dsRNAs targeting five genes, including genes encoding actin and subunits of V-ATPase. Recently, Necira et al. [34] showed that topical application of sterilized E. coli cells containing a target dsRNA produced by a recombinant RNA expression system could induce resistance of N. benthamiana to potato viruses. The effectiveness of the E. coli-encapsulated dsRNA in providing RNAi-mediated protection against dual infection by two potato viruses was shown to be comparable to that of naked dsRNA crudely extracted from the producing bacteria. In addition, the oriental fruit fly Bactrocera dorsalis, which is a damaging pest that infects important fruits and vegetables, was fed with E. coli cells equipped with the same dsRNA production system as above as part of an artificial diet [35]. The treated maggots, emerged pupae, and adults exhibited severe mortality and deformities.

The application of dsRNA as an antiviral agent to protect shrimp from damage caused by white spot syndrome virus (WSSV) has been studied. For RNAi action in the shrimp, Thammasorn et al. [36] produced two kinds of dsRNA in E. coli to target expression of VP28, the major structural protein of the virus, and WSSV015, a hub-protein that interacts with many other proteins. They constructed E. coli strains producing each dsRNA using strain HT115(DE3) as the host with a T7 promoter expression system, and co-cultured them in Terrific Broth by fed-batch culture with the addition of glycerol. The accumulation of total dsRNA reached approximately 95 mg per liter of culture broth. The use of such multiple-target dsRNAs will be of considerable interest as a method to effectively control the complex WSSV that has marked genetic variation.

In an effort to further improve dsRNA productivity in the L4440-based recombinant RNA expression system in E. coli, Papić et al. [37] analyzed the kinetics of target dsRNA (480 bp-segment) production in fed-batch fermentation and found that after 12 h of fermentation, the dsRNA/biomass yield was 0.06 g/g, the maximum productivity of dsRNA was 15.2 mg/L/h, and the final dsRNA concentration achieved was 182 mg/L. Thus, it was indicated that the dsRNA production by the standard expression system is growth-associated with the producing bacteria.

E. coli strains other than HT115(DE3) have also been generated as host strains for dsRNA production. One such strain is E. coli M-JM109lacY, which was constructed by deleting the rnc and lacY (encoding lactose permease) genes from strain JM109(DE3), a derivative of strain JM109 [38]. The authors examined the productivity of a target hpRNA (approximately 480 bp long) for suppression of gene expression of tobacco mosaic virus (TMV) coat protein, with several E. coli strains as hosts including M-JM109lacY, using a transcription system based on IPTG-induced expression of T7 RNA polymerase. It was shown that the productivity of the target hpRNA using strain M-JM109lacY was superior to that using strains such as HT115(DE3) as the host. It was speculated that mutations in endA (encoding endonuclease I, which is a periplasmic enzyme that cleaves within duplex DNA) and recA (encoding a DNA strand exchange protein that is central to the process of homologous recombination) in M-JM109lacY might have contributed to the stability of the expression plasmid and the dsRNA products in the host cells, but details of the underlying factors remain unclear. In addition, although it was suggested that lacY mutation allows uniform entry of IPTG into all cells in the population of the culture and produces a homologous induction level of gene expression by the inducer, the involvement of the lacY mutation in improving dsRNA productivity in the strain is also not certain. However, there have been some reports on successful dsRNA production using M-JM109lacY as the host strain [38,39,40,41,42]. Recently, Ma et al. [43] attempted to construct a production system that is more efficient than the conventional system for dsRNA production, L4440-HT115(DE3). When an rnc-deficient strain of E. coli BL21(DE3) was used as the host and a hairpin-loop-type dsRNA expression system using pET28 (equipped with a single T7 promoter) was adopted as the dsRNA transcription system, it was reported that the productivity of the target dsRNA (4.23 mg/L) was approximately three times that in L4440-HT115(DE3).

A unique system that allows microorganisms to produce and partially purify short interference RNA (siRNA) molecules of approximately 21 nucleotides for RNAi has been developed. Huang et al. [44,45] used an RNA-binding protein named p19 to stabilize target siRNAs produced within E. coli and to purify the target siRNAs. p19 is a 19-kDa protein encoded by the plant RNA virus Tombusvirus, which selectively binds to siRNA molecules and is originally a suppressor of RNA silencing function in plants. Specifically, siRNA-embedded short hairpin RNA (shRNA) was co-expressed fused with glutathione S-transferase (GST) and His-tagged p19 (GST-p19-His) (Figure 3). The target siRNA segment from the produced shRNA was separated by the action of endogenous RNase III within E. coli, and the siRNAs bound to GST-p19-His to form a stable complex. The siRNA-GST-p19-His complex was easily purified from the total protein extracted from the producing E. coli by nickel affinity chromatography. Subsequently, the siRNA was isolated from the purified complex by SDS treatment, and then finally purified by anion exchange chromatography. Using this preparation method, the authors succeeded in producing a few kinds of functional siRNA, but the overall yield and productivity were low (several tens of micrograms per liter of E. coli culture). This indicates that while this approach is very elegant for siRNA production, challenges remain for its use in agriculture and aquaculture where extremely low-cost dsRNA is required.

When expressing recombinant RNA in a microorganism, it is critical to avoid attack on the RNA molecule by endogenous RNA-degrading enzymes in the cell. Besides reducing the activity of endogenous RNA-degrading enzymes, some measures have been developed to make the produced RNA molecules less susceptible to the action of ribonucleases. One approach is to produce RNAs that have a stable and compact three-dimensional structure that is resistant to ribonucleases in vivo, and to carry the target RNA segment on this scaffold. Such scaffolds for the production of recombinant RNA molecules of interest include tRNA [46,47,48], 5S rRNA [49,50], tRNA/pre-miRNA structures [51,52], and viroid-derived RNA structures [53,54]. The concept of using the structures of tRNA and 5S rRNA as scaffolds is based on the rationale that these structures will be recognized as endogenous molecules by the RNA metabolizing machinery in the producing bacteria, resulting in suppression of rapid degradation of the additional RNA segment of interest. Ponchon et al. [46,47] designed a method that uses tRNA as a protective scaffold for the production of desired RNA molecules in E. coli and for convenient isolation of the target chimeric RNA. Specifically, the RNA segment of interest was inserted into the anticodon stem of the tRNA-scaffold, and an aptamer-tag (e.g., a Sephadex aptamer-based affinity tag) was also incorporated into the scaffold for efficient purification of the chimeric RNA molecules. The authors succeeded in obtaining a chimeric RNA molecule containing the target RNA segment, which could be easily purified. They produced several RNA fragments of pharmacological interest by this method, and tens of milligram quantities of the RNA (as purified chimeras) were obtained from 1 L scale culture. To finally obtain the target RNA from the chimera with the scaffold, an in vitro RNase H-cleavage method directed by two DNA oligonucleotides complementary to the invariable part of the tRNA-scaffold was employed. Subsequently, Nelissen et al. [48] improved this tRNA-scaffold approach. First, to further enhance RNA expression, the promoter was changed from the constitutive promoter of the lpp gene (encoding the major outer membrane lipoprotein in E. coli) to the inducible T7 promoter. Then, to enable efficient and simple excision of the target RNA segment from the produced tRNA-scaffold chimera, a cis-acting hammerhead (HH) ribozyme was incorporated into the chimeric structure, and a method of isolating the target RNA molecule by HH ribozyme-mediated cleavage was devised. In this system, several of the chimeric RNAs showed a yield of approximately 10–30 mg per liter of culture medium. These RNA expression systems are suitable for the production of relatively short RNAs, such as therapeutic siRNAs, for applications that require structurally homogeneous and highly pure RNAs. For agricultural applications, high structural homogeneity of the target RNA is not necessarily required, but such technology is also expected to be applied for the higher accumulation of RNAs of interest for agricultural uses in bacteria.

Subsequently, based on tRNA-scaffold technology, Chen et al. [51] further investigated the expression of chimeric RNAs by fusing several kinds of pre-miRNAs to the tRNA-scaffold with E. coli as the host. They developed an RNA expression system that uses the tRNA/pre-miR-34a scaffold as a new, optimal non-coding RNA scaffold (named OnRS) for production of functional small RNAs; the fusion construct with pre-miR-34a was particularly highly expressed in E. coli. This scaffold was also resistant to nucleolytic digestion in E. coli, and multi-milligrams of the chimeric RNAs were readily obtained per liter of bacterial culture. They further developed this system and designed a new scaffold, tRNA/pre-miR-34a/pre-miR-34a, by adding an additional human pre-miR-34a to the OnRS structure, and used it to produce multiple small RNAs of interest [52]. The chimeras containing the target RNAs were successfully expressed in E. coli at high levels (>40% of the total bacterial RNA). Thus, a new platform for recombinant RNA production using the novel RNA scaffold as a carrier for the accommodation of multiple small RNAs of interest, such as miRNAs, siRNAs, and RNA aptamers, has been constructed.

Viroids are infectious agents of higher plants that exclusively consist of relatively small non-protein-coding, naked, circular RNA molecules. Daròs et al. [53] found that co-expression of Eggplant latent viroid (ELVd)-derived RNAs with eggplant tRNA ligase (an enzyme involved in viroid RNA-circularization in viroid-infected plants) in E. coli resulted in monomeric cyclization of the majority of the RNAs produced and significant accumulation of the product in the host cells. Then, a heterologous desired RNA segment, several dozens of nucleotides long, was integrated into a specific site of the viroid-scaffold (named ELVd-scaffold), and when the chimeric RNA construct was expressed in the same way, the cyclized chimeric RNA accumulated to several tens of milligrams per liter of culture of the E. coli cells. The authors aimed to apply this viroid-based expression system to produce the large amounts of dsRNA required to combat the pest Western corn rootworm (WCR, Diabrotica virgifera), using RNAi [54]. The smooth septate junction 1 (DvSSJ1) gene, which is essential for the growth of WCR, was selected as the target for RNAi action. When constructing the transcription system for the ELVd-scaffold with 83-nt inverted repeats homologous to the DvSSJ1 gene, the group-I autocatalytic intron was inserted as a spacer to avoid structural instability of the expression plasmid. When the ELVd chimeras were co-expressed with eggplant tRNA ligase in an RNase III-deficient E. coli strain, it was observed that the inserted introns were self-spliced from the nascent biosynthesized transcript, and the chimeric ELVd containing the target RNA segment was intramolecularly cyclized and accumulated in large amounts in the cells. Then, in larval-feeding bioassay, it was demonstrated that the circular RNAs containing a dsRNA moiety (hpRNA) homologous to the DvSSJ1 gene exhibited excellent insecticidal activity against WCR. Thus, an intron-assisted, viroid-based production system for circular dsRNA in E. coli was developed.

2.2. Pseudomonas syringae

For in vivo dsRNA production, a novel method that uses the RNA-dependent RNA polymerase (RdRp) function of bacteriophage phi6, a dsRNA virus infecting Pseudomonas syringae cells, was reported by Aalto et al. [55] RdRp in phi6 performs de novo initiation of transcription and produces full-length dsRNAs from given template ssRNAs. P. syringae is a Gram-negative bacillus with subpolar pili, to which phi6 phages typically attach. Many P. syringae strains are plant pathogens, but Wilson et al. constructed a dsRNA production system using nonpathogenic P. syringae strain Cit7 [56] (containing plasmid pLM1086 constitutively expressing T7 RNA polymerase) as the host microbe. The phi6 genome consists of three segments (L, M, and S), and the genes on each genomic segment are clustered into functional groups, which are flanked by 5′-packaging and 3′-replication signals [57]. The L-segment encodes the components of the polymerase complex (to form a procapsid (PC) with RNA polymerase activity), and the PC particles formed in the host microbe incorporate ssRNAs with 5′-packaging signals and convert them to dsRNAs. pLM991 is designed to produce phage PC components and kanamycin resistance (as a selection marker) under the control of the T7 promoter [58]. Aalto et al. [55] employed the egfp gene (encoding green fluorescent protein) as a target dsRNA example for genetic knockdown experiments, allowing the egfp gene flanked by 5′-packaging and 3′-replication signals to be expressed by T7 RNA polymerase on plasmid pPS9, and introduced these into the host microbial cells. egfp-ssRNAs were converted to dsRNAs inside empty PCs, and finally, the production of capsids containing egfp-dsRNAs (1.6 mg dsRNA/g wet cells) was achieved in P. syringae.

The dsRNA production system consisting of P. syringae and phi6 was subsequently reported to produce dsRNA (named dsRNA^rep-MP) against TMV genomic DNA [59] (Figure 4). Two dsRNAs corresponding to sequences encoding the replicase (Rep) and movement protein (MP) partial regions of TMV were produced. By inserting the DNA fragments for dsRNA production into the DNA regions corresponding to the M- and S-segments of the phi6 genome, two plasmids were constructed in which the ssRNAs were expressed under the control of the T7 promoter, and the plasmids were introduced into P. syringae strain LM2691 expressing T7 RNA polymerase along with a plasmid expressing the L segment of the phi6-polymerase complex. The resulting strain was cultured to produce dsRNAs, and accumulation of approximately 700 µg dsRNA^rep-MP per 100 mL of culture (4 × 10⁹ cells/mL) was achieved. The TMV-derived dsRNAs produced in this way, when applied to TMV-infected N. benthamiana plants, were shown to inhibit the propagation of TMV and to provide efficient protection against local spread of the pathogen. Compared with other in vivo dsRNA production systems, this system allows the efficient production of very long dsRNAs (>2.6 kbp). Such long dsRNAs have the advantage of giving a large pool of siRNAs for a single target (i.e., a viral genome), although the issue of off-target effects should be considered. Long dsRNAs will also be able to link multiple target sequences to protect plants against several pathogens simultaneously. In addition, the formation of target dsRNAs inside the procapsid structure is expected to efficiently avoid degradation by ribonucleases present in the producing host microbial cells, which is advantageous for the preparation of structurally more homogeneous dsRNAs and will contribute to stable dsRNA storage.

2.3. Corynebacterium glutamicum

C. glutamicum is a non-sporulating Gram-positive soil bacterium with a high G + C mol% content in its genomic DNA. This microorganism is nonpathogenic and does not produce endotoxins, making it a safe strain for humans. In 1965, this strain was reported as a microorganism capable of overproducing L-glutamic acid in the culture medium [60]. Since then, C. glutamicum, metabolically engineered, has been used as a workhorse for the production of various L-amino acids, proteins, and, recently, a variety of other useful biochemicals [61,62,63]. In particular, C. glutamicum has been employed as an industrial L-amino acid-producing microorganism in large-scale cultivation for several decades, and is a GRAS host for the industrial production of biochemicals including L-glutamate and L-lysine [64,65]. The advantages of this robust strain are also being exploited to make it a host for overproduction of recombinant RNA molecules.

Hashiro et al. first examined the possibility of overexpression of RNA molecules with a desired nucleotide sequence in a mutant C. glutamicum strain (named 2256LΔrnc, in which the rnc gene encoding RNase III was deleted from wild-type strain ATCC13869) [66,67]. To enhance transcription of recombinant RNA molecules, the F1 promoter derived from bacteriophage BFK20 [67,68], which infects Corynebacterium, was chosen as a strong promoter that functions in C. glutamicum. A newly acquired high-copy-number mutant plasmid carrying the transcription system was also employed [69]. Overproduction of a recombinant RNA named U1A*-RNA (approximately 160 nucleotides long) was examined as a model RNA for expression studies in strain 2256LΔrnc (Figure 5a). U1A*-RNA consists of a stem/loop II (SL II) structure from U1 small nuclear RNA (snRNA) and is an RNA molecule designed to contain a domain that binds to U1A protein to form a U1 small nuclear ribonucleoprotein (snRNP) involved in pre-mRNA splicing. The producer microbe showed prominent accumulation of U1A*-RNA (approximately 300 mg/L of culture medium), which opened the door to the production of recombinant RNAs using C. glutamicum as the host microorganism [67].

C. glutamicum-driven production of dsRNAs that can be applied as RNAi-based pesticides was subsequently investigated. The diap1 gene (encoding death-associated inhibitor of apoptosis protein 1) [70] was selected as a target for RNAi action to combat the crop pest H. vigintioctopunctata, and production of target dsRNA (named diap1*-dsRNA) corresponding to a region of approximately 360 nucleotides long was undertaken [71]. This production system in C. glutamicum was designed to produce the target diap1*-dsRNA by convergent transcription using two strong F1 promoters (Figure 5b). This expression system led to the accumulation of diap1*-dsRNA to approximately 75 mg per liter of culture medium in jar fermentation of the producing microorganism [71]. To further increase the productivity of dsRNA, the employment of T7 RNA polymerase within C. glutamicum was considered, since T7 RNA polymerase has the advantage of avoiding collisions between the RNA polymerases in the transcriptional process from T7 promoters in such a convergent transcription system (Figure 5c). Convergent transcription in C. glutamicum using the combination of T7 RNA polymerase and the T7 promoter allowed the production of longer dsRNA strands than was possible using the endogenous multisubunit RNA polymerase and the F1 promoter; it was possible to achieve the production of dsRNAs as long as ~1 kbp without a significant decrease in productivity [72]. In addition, by installing T7 terminators in the transcription system and further improving the high-copy-number plasmid vector [73] carrying the expression system, diap1*-dsRNA was produced at a yield of approximately 1 g per liter of culture medium in jar fermentation of C. glutamicum strain 2256LΔrnc [72]. As far as we know, this is the highest reported microbial production level of dsRNA so far. The capacity to produce long dsRNAs means that many dsRNA segments for various regions of the targeted essential gene of a desired pest can be carried on a single dsRNA molecule. This would improve the effectiveness of the pesticide, because even if the long dsRNA is somewhat degraded in the process of being taken into cells of the target pest, the probability that some of the remaining dsRNA segments are able to exert RNAi action will increase. In addition, since dsRNA segmented for RNAi induction against several different target pests can be carried on a single long dsRNA, in principle, one RNAi-based pesticide can be designed that works on several target species at once.

A delivery method of dsRNA accumulated in C. glutamicum cells to pests was also devised to facilitate the use of RNAi-pesticides at low cost. Generally, E. coli (and other bacteria) that have accumulated dsRNA are sterilized by heat treatment and then the killed bacteria containing the dsRNA are administered to the target pest [74,75]. To promptly sterilize dsRNA-producing C. glutamicum and to eliminate the risk of latent degradation of the target dsRNA within the cells during heat treatment, a dsRNA preparation and delivery method using common alcohols (such as methyl and ethyl alcohols) was proposed [71]. Sterilization of the C. glutamicum strain was achieved by simply suspending the bacterial cells in each alcohol solution at a concentration of >50%; no noticeable degradation of dsRNA in the cells was observed. When larvae of the target pest H. vigintioctopunctata ingested the dsRNA encapsulated in killed C. glutamicum cells prepared in this way, the feeding activity of the pest on potato leaves was reduced and weight gain was significantly suppressed, indicating that diap1*-dsRNA-containing sterilized C. glutamicum cells were effective in inhibiting the growth of the pest [71,72]. A similar method of ethanol sterilization of a dsRNA-producing microorganism for use as an RNAi-based pesticide has recently been reported in E. coli producing dsRNA for protection against potato virus infecting N. benthamiana [34]. The antiviral protective effect of using killed bacteria containing the target dsRNA was comparable to that of using dsRNA extracted from the producing strain. These findings suggest that the sterilization of dsRNA-producing microbes by alcohol treatment and the resultant encapsulation of dsRNA may be effective for the formulation of RNAi-based pesticides in a variety of dsRNA-producing microbes. Encapsulating dsRNA in bacterial cell bodies and using it as a pesticide in the field will decrease the risk of environmental UV damage to, and RNA-degrading enzyme attacks on, the dsRNA compared with using naked dsRNA, and it will also lower the total production cost of the RNAi-based pesticides.

2.4. Chlamydomonas reinhardtii

Chl. reinhardtii, a eukaryotic unicellular green microalga, is a potential industrial bioproduction platform for various products including recombinant proteins, enzymes, and biofuels [76,77,78]. Recently, the expression of recombinant dsRNA in Chl. reinhardtii has also been investigated. Since Chl. reinhardtii typically inhabits freshwater environments such as ponds and lakes, research has been conducted to use this microbe producing target dsRNAs to confer resistance to viral diseases in edible shrimp in aquaculture. Viral infection such as by WSSV or YHV results in extensive production losses in shrimp farming [79].

The expression of target recombinant genes in the microalga can be achieved by incorporating the desired dsRNA expression unit into the nuclear DNA, or into the chloroplast DNA [76,80]. For the control of YHV, an RNA virus that infects penaeid shrimp including Panaeus vannamei, Somchai et al. produced dsRNA (hpRNA) that can act specifically on the expression of RNA-dependent RNA polymerase (RdRp) of YHV, through integration of the dsRNA expression unit into the nuclear genome of Chl. reinhardtii [81]. The expression of hpRNA-YHV was approximately 45 ng per 100 mL of culture (~1 × 10⁸ cells of the engineered Chl. reinhardtii strain). A feed containing Chl. reinhardtii cells expressing YHV-dsRNA (1 × 10⁸ cells per gram of feed) was formulated; oral administration of the feed to post-larval P. vannamei improved the survival rate after YHV infection. Specifically, at 9 days post-treatment, all shrimp fed the normal feed had died, while approximately 22% of the shrimp fed the feed containing the dsRNA-containing cells survived [81]. This showed the potential effectiveness of dsRNA produced from the nuclear genome of Chl. reinhardtii in protecting shrimp from YHV infection. However, because the nucleus of the microalga is equipped with RNAi machinery for dsRNA processing, which could have degraded the target dsRNA produced, the researchers then tried to produce the target dsRNA within the chloroplast, which lacks any such functions [82]. Target dsRNA for the RdRp-encoding gene of YHV was produced in the chloroplast using a convergent transcription mode with the promoter of the psaA gene (which encodes the PsaA subunit of photosystem I in the chloroplast). In this transformation system, a selection system eliminating the need for antibiotic resistance markers was used. The engineered Chl. reinhardtii produced a low level of dsRNA (approximately 16 ng-dsRNA per liter of culture medium). However, when the generated strain was used in a shrimp feeding trial to evaluate the efficacy of the microalgal RNAi against YHV, the mortality rate of shrimp pretreated with dsRNA-expressing microalgae was approximately 50% at 8 days post-YHV infection, compared with approximately 84% mortality in the control group, suggesting that the use of dsRNA-producing microalgae may be a promising strategy to reduce mortality caused by YHV in shrimp farming [82].

Chlamydomonas was also used as a host to produce hpRNA as an RNAi-based microalgal larvicide for mosquito control. Mosquitoes transmit many life-threatening parasitic and viral diseases, including filaria, yellow fever, dengue fever, and malaria. Many chemical pesticides and biological control agents have been developed and used to control mosquitoes, but there is a need for control methods that are safer for humans and ecosystems. Furthermore, the long-term use of these conventional chemical pesticides has resulted in the development of pesticide-resistant mosquito species. Microalgae are a major source of nutrition for mosquito larvae in water bodies that serve as breeding grounds for mosquitoes. Hence, Kumar et al. [83] designed an hpRNA expression system in Chlamydomonas chloroplasts to suppress expression of the gene encoding 3-hydroxykynurenine transaminase (3-HKT), which is involved in tryptophan catabolism and is essential for mosquito growth, by RNAi. Although the productivity of this system was not explicitly specified, by delivering the recombinant microalgae to Anopheles stephensi larvae, a significant increase in larval mortality was observed compared with the control treatment.

So far, compared with the commonly used dsRNA expression systems in E. coli, target dsRNA productivity in Chl. reinhardtii remains low. However, the use of such microalgae may be more acceptable than the continuous release of large amounts of E. coli into shrimp farms and wetlands, which would raise concerns about environmental contamination. The GRAS status of Chl. reinhardtii and the lack of any endotoxin production make this microbe attractive as a host for dsRNA production.

2.5. Other Microorganisms for dsRNA Production

There have been studies on the use of the probiotic bacterium Lactobacillus plantarum and the oleaginous yeast Yarrowia lipolytica as production hosts for dsRNAs that can be used to help alleviate viral and bacterial diseases of marine shrimp [84,85]. Saksmerprome’s research group expressed in L. plantarum a hairpin-type dsRNA to function against shrimp YHV. They had already orally administered YHV-dsRNA expressed in E. coli to shrimp and confirmed its effectiveness as a potential antiviral agent [86], but lactic acid bacterium are more suitable for application in the shrimp farming environment. L. plantarum is a plant-derived lactobacillus that is found in many fermented food products and is nonpathogenic to humans. Lactobacilli have already been applied to farmed shrimp to enhance the immunity of shrimp toward Vibrio and other pathogenic bacteria. The target RNA was expressed in L. plantarum using the xylose-induced xylA promoter; the production was approximately 1.66 ng per 7 × 10¹¹ colony-forming units of the lactobacillus [84], although the bacterial strain used for the dsRNA production was not mutated in dsRNA-degrading enzyme genes. Thus, it was shown that oral administration of L. plantarum containing YHV-specific dsRNA in shrimp improved their survival and reduced YHV propagation. Although this dsRNA accumulation was estimated to be lower than that of the production of similar hairpin-type RNAs in E. coli, it is still a promising method for disease prevention in cultured shrimp because the lactobacilli themselves, which are administered with the produced RNA, are expected to have a probiotic effect on the shrimp.

Another host microorganism for the production of hpRNA for disease prevention in cultured shrimp is Y. lipolytica, a GRAS microorganism that does not produce endotoxins and can be grown on low-cost substrates. In addition, the high protein content of Y. lipolytica is expected to contribute to the growth of shrimp that feed on the yeast cells. Furthermore, yeasts including Y. lipolytica have been used as immunostimulants to enhance the immune response of aquaculture organisms to pathogenic bacteria and viruses. Álvarez-Sánchez et al. produced hpRNA in Y. lipolytica to suppress the virulent activity of WSSV in the white leg shrimp Litopenaeus vannamei [85]. The selected target was the orf89 gene of WSSV, a region of approximately 0.4 kbp, of which was used as the dsRNA portion to be expressed. The promoter for the hpRNA expression was the constitutive promoter of Y. lipolytica XPR2, a gene encoding an extracellular alkaline protease (the major protein secreted by this yeast), and the hpRNA expression system was constructed on a plasmid. The RNase III gene (Rnt1) of the host Y. lipolytica strain remained intact. In the end, the maximum amount of hpRNA produced was 182 ng/L, but when the total RNA extracted from the producing strain was injected intramuscularly into shrimp, the WSSV-dependent mortality rate improved by approximately 25%.

Furthermore, there is research where microorganisms symbiotic in the gut of a target pest were made to produce dsRNA for RNAi action. Whitten et al. [87] selected the blood-sucking bug Rhodnius prolixus and the globally invasive polyphagous agricultural pest Western thrips (Frankliniella occidentalis) as target pests, expressed dsRNAs within symbiotic bacteria of each pest, and administered the symbiotic bacteria to the corresponding pests. For the former, Rhodococcus rhodnii strain LMG5362, which has evolved to survive in the midgut of Rhodnius prolixus, was employed as the host microbe, and dsRNA targeting the vitellogenin (Vg) gene of the pest was expressed in the strain lacking the RNase III gene. For the latter pest, an RNase III-deficient strain of the Gram-negative facultative symbiotic bacterium, named BFo2 [88,89], was created to produce dsRNA for RNAi of the α-tubulin gene (Tub) of F. occidentalis. When the respective dsRNA-producing symbiotic strains were orally administered to the respective pests, a significant decrease in fecundity was observed in R. prolixus, and a significant increase in mortality was confirmed in F. occidentalis. This method of using symbiotic microorganisms producing a desired dsRNA may result in a continuous supply of dsRNA in the gut of that pest. This could be a very effective way to ensure RNAi action against the target pest. However, since this method also involves the release of recombinant bacteria into the environment, there are many issues from a biosafety perspective that will need to be addressed.

The baker’s yeast Saccharomyces cerevisiae has also been used to produce target RNA for RNAi pesticides to combat mosquitoes. Duman-Scheel and colleagues studied the production in S. cerevisiae of short hpRNA targeting the Aedes aegypti synaptotagmin (syt) gene (a gene involved in the release of neurotransmitters in the larval neural synapse) [90,91]. The yeast cells containing the produced target dsRNA were then heat-sterilized, dried, and used as a mosquito larvicide. Yeast acts as an oviposition attractant for gravid adult females of mosquitoes and S. cerevisiae is also a source of nutrition and a strong odorant attractant for hatched mosquito larvae; therefore, the use of such formulations is expected to attract mosquito larvae and induce them to ingest the formulations, resulting in the death of the targeted mosquito larvae caused by the dsRNA contained in the yeast. The DNA region encoding the target hpRNA was placed under the control of the strong inducible Gal1 promoter, and a cyc1-terminator was used to ensure termination of transcription [90,91]. To construct the hpRNA-producing strain, two copies of the hpRNA-expression unit were integrated into the genome of the strain S. cerevisiae CEN.PK. Although S. cerevisiae lacks endoribonuclease Dicer, which cleaves dsRNA into short dsRNA fragments, it has a gene encoding Rntp1, which shows RNase III-like activity; however, stem-loop type RNAs may not be recognized as substrates by Rntp1. The authors prepared yeast cells containing the expressed target hpRNA in a dried, inactivated tablet formulation and delivered it to mosquito larvae, which resulted in suppression of syt gene expression, defects in the neural synapse, and high rates of larval A. aegypti mortality. Interestingly, the sequence they employed for RNAi in A. aegypti also induced RNAi effects in related mosquito larvae with the same sequence of the corresponding gene, leading to their death. This suggests that yeast expressing biorationally-designed hpRNAs may be used to combat multiple mosquito vectors of human diseases.

The methanol-using yeast Pichia pastoris has been employed as a host for the production of hpRNAs (ca. 120 bp) that induce RNAi effects against the expression of the juvenile hormone (JH) acid methyl transferase gene (jmtA) involved in JH biosynthesis in A. aegypti [92]. Although the productivity of the target hpRNA in the yeast was not specified, mosquito larvae orally administered heat-sterilized yeast cells containing the hpRNA showed delayed hatching into adults and high mortality. P. pastoris is an excellent host for producing heterologous proteins, and it has already been reported that it is possible to express Bacillus thuringiensis Cry toxin in this host, which made it toxic to A. aegypti larvae [93]. Hence, P. pastoris co-expressing Cry toxin and a dsRNA of interest is expected to be effective in killing target mosquito larvae.

3. Conclusions and Future Perspectives

The RNAi phenomenon [2] was first reported in 1998 in Caenorhabditis elegans. Since then, RNAi technology has made great progress and is applied in basic research, such as for analysis of gene function, and in applied research in fields from medicine to agriculture. In 2007, a potential innovative pesticide, an RNAi-based pesticide that inhibited the growth of target agricultural pests, was developed [94]. However, a major barrier to the practical use of RNAi-based agents is the production cost. The amount of these pesticides required for use in practical situations has not yet been studied in detail, but recent estimates suggest that approximately 2–10 g of RNA per hectare may be needed [95,96] (although this amount depends significantly on the efficiency of the RNAi action of the dsRNA in the target species and the stability of the formulated dsRNA product in the environment in which it is applied). Thus, it is also crucial to develop more effective application methods and formulations of dsRNA [6,16,97]. Recently, Mehlhorn et al. [98] examined effective RNA doses for CPB from several regions of Europe to determine their differential susceptibility to RNAi action in foliar spray application. At the laboratory level, dsRNA targeting the actin-encoding gene of CPB resulted in almost 100% mortality with a dose equivalent to 0.96 g dsRNA/ha for the most susceptible sample species, although the least susceptible required approximately 7.5 times that amount of dsRNA.

The production cost of RNAs has been decreasing year-by-year as manufacturing technology improves [17,99]. A biotechnology company called RNAgri (formerly APSE) has developed a technology that can produce dsRNA at a cost of USD 1–2/g RNA. Their production system claims to be based on accumulation of target dsRNA in virus-like particles (VLPs) in E. coli. The coat protein (CP) of bacteriophage MS2 and the precursor of the target dsRNA containing packaging sites (PSs) are simultaneously expressed. The CP and PS bind and encapsulate the RNA in VLPs [100]. RNA production in GreenLight Biosciences’ cell-free system for the synthesis of biomolecules is also attracting attention. In general, the main barrier to RNA production in cell-free systems is the high cost of the four ribonucleoside triphosphates (NTPs) that are the building blocks of RNA. This company has enabled low-cost production of target RNAs by degrading nucleic acids derived from yeast biomass into the constituent monomers using nucleases; phosphorylating and regenerating the NTPs, and then using them as substrates for RNA synthesis in the cell-free system [101].

The microbial production methods for dsRNAs described in this review can be used to mass-produce target dsRNAs from sustainable raw materials. Some representative reports describing the productivity of target dsRNAs are summarized in

Table 1. Representative examples of microbial dsRNA production.

Species of Host Microbe	Strain	RNA Type	Length of the dsRNA Region	Targeted Protein or Gene for RNAi	Targeted Organism or Virus	Expression System; Mode, Promoter, Vector	Titer (Culture Scale)	Reference
Escherichia coli	HT115(DE3)	hpRNA	977 bp	Replicase	Pepper mild mottle virus	T7 promoter on pGEM	4 mg/L (20 mL)	[22]
	HT115(DE3)	hpRNA	400 bp	Protease	Yellow head virus	T7 promoter on pET3a	30 mg/L (50 mL)	[29]
	HT115(DE3)	dsRNA	422 bp	Vestigial gene (vg)	Aedes aegypti	Convergent transcription, T7 promoters on pLitmus28i	20 mg/L (25 mL)	[30]
	HT115(DE3)	dsRNA	480 bp	Not specified	Not specified	Convergent transcription, T7 promoters on plasmid L4440	182 mg/L (initial volume of fed-batch culture, 2 L)	[37]
	M-JM109M-JM109lacY	hpRNA	480 bp	Coat protein (CP)	Tobacco mosaic virus	T7 promoter on pET-22b or pGEM-T	Not specified	[38]
	T7 Express Iq	siRNA	~21 bp	Enhanced green fluorescent protein (EGFP) Lamin A/C (LMNA)Polo-like kinase 1 (PLK1) Tumor protein p53 (TP53) Viral infectivity factor (vif) Group-specific antigen (gag)	Aequorea Victoria Homo sapiens Homo sapiens Homo sapiens Human immunodeficiency virus 1 (HIV1) Human immunodeficiency virus 1 (HIV1)	T7 promoter and GST-p19-His expression cassette on pGEX-4T-1	~4 µg/L (300 mL)	[44]
	HT115(DE3)	ELVd-hpRNA	87 bp	Diabrotica virgifera smooth septate junction 1 (DvSSJ1)	Diabrotica virgifera	Murein lipoprotein (lpp) promoter on pLELVd-BZB and eggplant tRNA ligase expression cassette on p15LtRnlSm	Not specified	[54]
Pseudomonas syringae	Cit7	dsRNA	~4.0 kbp	Enhanced green fluorescent protein (eGFP)	HeLa-eGFP cells	Conversion of ssRNA to dsRNA by RNA-dependent RNA polymerase of phi6	1.6 mg/g wet cells (10 L fermenter)	[55]
	LM2691	dsRNA	2.6 kbp, 3.5 kbp	Replicase and movement protein (rep-MP)	Tobacco mosaic virus	Conversion of ssRNA to dsRNA by RNA-dependent RNA polymerase of phi6	7 mg/L (100 mL)	[59]
Corynebacteirum glutamicum	2256LΔrnc	dsRNA	360 bp	Death-associated inhibitor of apoptosis protein 1 (Diap1)	Henosepilachna vigintioctopunctata	Convergent transcription, F1 promoter on pVC7H2	75 mg/L (jar fermenter)	[71]
	2256LΔrnc	dsRNA	360 bp	Death-associated inhibitor of apoptosis protein 1 (Diap1)	Henosepilachna vigintioctopunctata	Convergent transcription, T7 promoter on pPK4H1	1.0 g/L (jar fermenter)	[72]
Chlamydomonas reinhardtii	CC-503 (cw92 mt+)	hpRNA	368 bp	RNA-dependent RNA polymerase (RdRp)	Yellow head virus	psaD promoter on nuclear expression plasmid pSL18-YHV	450 ng/L (100 mL)	[81]
	CC-5168 (cw15, ΔpsbH, SpecR)	dsRNA	374 bp	RNA-dependent RNA polymerase (RdRp)	Yellow head virus	Convergent transcription, psaA promoters on pSR-PYP integrated into the chloroplast	Not specified	[82]
	CC-4147 (FUD7 mt+)	hpRNA	328 bp	3-Hydroxykynurenine transaminase (3-HKT)	Anopheles stephensi	atpA promoter on pCVAC108 integrated into the chloroplast	Not specified	[83]
Yarrowia lipolytica	P01a	hpRNA	416 bp	Putative regulatory protein (Orf89)	White spot syndrome virus	XPR2 promoter on pRRQ1	182 ng/L (1 L)	[85]

. The basic characteristics of the production methods of dsRNA by the representative microbes discussed here are summarized in

Table 2. Characteristics of typical microbial dsRNA expression systems.

Host Microbe	Advantages	Disadvantages
Escherichia coli	- High production yield of dsRNA - Easy genetic manipulation - Many examples of use as host strain for dsRNA production	- Gram-negative bacterium with endotoxins
Pseudomonas syringae	- Enables the efficient production of long RNAs of uniform length - Preparation of RNA with high purity	- Low production yield of dsRNA
Corynebacteirum glutamicum	- Very high production yield of dsRNA - Robustness and proven performance in large-scale fermentation - Non-pathogenic, non-endotoxic microorganism (GRAS status)	- Thick cell membrane structure that makes RNA extraction difficult
Chlamydomonas reinhardtii	- Non-pathogenic, non-endotoxic microalga (GRAS status) - Use as live dsRNA-producing alga considered (mainly for use in aquaculture feed)	- Low production yield of dsRNA - Concerns about release of living GMOs into the environment
Saccharomyces cerevisiae	- Non-pathogenic, non-endotoxic eukaryote (GRAS status) - Proven performance in large-scale culture - Beneficial properties of microbial ingredients containing pest attractants	- Low production yield of dsRNA
Symbiotic bacteria (Rhodococcus rhodnii, BFo2)	- Capable of persistent production of dsRNA while coexisting in the gut of the target pest	- Difficulties in constructing dsRNA-producing bacteria with the limited available genetic engineering tools - Concerns about release of living GMOs into the environment

Abbreviations: GRAS, generally recognized as safe; GMO, genetically modified organism.

. The methods applied must be tailored to specific needs; it will be necessary to establish production methods optimized for each RNAi-based product according to its use and practical requirements. Aspects such as using the producing microorganism as a capsule to protect the target dsRNA from unintended degradation, or using the microorganism itself as a supplemental nutrient for aquaculture, may add value to RNAi-based agents. RNAi-based products could contribute to solutions to various problems related to safe and reliable production of farm products now and in the future; in other words, the realization of “sustainable farming”. RNAi-based pesticides will be helpful in conserving and maintaining natural ecosystems and will greatly assist in establishing “sustainable agriculture”.

In aquaculture, the antiviral effects of dsRNA species will reduce the extensive use of chemical antivirals, which have a wide spectrum of activity, helping with the establishment of “sustainable aquaculture”.

Successful implementation of RNAi-based pest control strategies requires high efficacy of the dsRNA in inhibiting the growth of the target pest and also easy operation in the field. Therefore, optimal delivery systems for the dsRNA to the target pest are essential. Particularly in lepidopteran species, oral administration of dsRNA may not fully elicit the RNAi effect. Parson et al. [102] addressed this problem by synthesizing a cationic polymer compound, poly[N-(3-guanidinopropyl)-methacrylamide], to bind to RNA. When dsRNA was bound to this compound and administered to larvae of the fall armyworm, it showed moderate larval mortality. This carrier compound is thought to bind to the dsRNA and make it resistant to the attack of high-activity RNA-degrading enzymes in the insect gut. It has also been suggested that the complex of the polymer and dsRNA facilitates its transport into cells in the digestive tract. Similarly, Christiaens et al. [103] showed an increase in mortality after oral administration to Spodoptera exigua, a lepidopteran insect, of a dsRNA formulation using guanylated polymer nanoparticles to ensure the stability of the complex in the alkaline environment of the digestive tract. Recently, a new nanocarrier polymer/detergent formulation has been developed that allows dsRNA to quickly penetrate the insect body wall. In plant-chewing pests such as hemipteran aphids, dsRNA ingested orally is rapidly degraded by enzymes in the gut and hemolymph, making it difficult to achieve the full effect of RNAi. However, it was reported that the novel delivery system (dsRNA/nanocarrier/detergent formulation), in which the target dsRNA was bound to a cation dendrimer and covered with a detergent, promoted transdermal penetration of the target dsRNA into soybean aphids and induced the expected inhibitory RNAi effect [104]. Refinement of compounds for the optimal delivery of RNA products is expected to continue, but, in the future, the total manufacturing cost of RNAi-based products, including these delivery formulations, will have to be reduced to a level where they can be used commercially. We have mentioned that high productivity of dsRNA by the microbe is crucial for economic efficiency of the production of target dsRNA for RNAi pesticides. An ideal production strain might be a dsRNA-producing microbe that enables the RNAi effect to be effectively demonstrated against pests when dsRNA-containing killed-microbes are fed to the target pest.

RNA molecules are inherently chemically labile and are susceptible to degradation by the ribonucleases of various organisms. In addition, when naked dsRNA is topically applied to plants by foliar spraying, the RNA does not persist on the leaf surface and the period during which the dsRNA functions effectively in RNAi may be short. To address this issue, it has been demonstrated that by loading the target dsRNA onto a special layered double hydroxide (LDH) clay nanosheet, the stability of the dsRNA on the leaf surface is maintained, and sustained dsRNA release from the clay complex is achieved [16,105,106]. Thus, LDH can extend the effective period of topically applied dsRNA for inducing RNAi against plant viruses. Such innovations for more effective application of dsRNAs will contribute greatly to reduction of the total costs of RNAi-based products.

RNAi-based pesticides are characterized by the principle that the dsRNA only works on mRNA with the identical nucleotide sequence. Therefore, when a mutation occurs in the target gene of a target pest (but substantial functional expression of the mutated gene is retained), the originally designed dsRNA may no longer be effective against the pest. To cope with such a situation, it is essentially sufficient to change the dsRNA sequence such that RNAi can be performed on the altered sequence, although the issue of off-target effects should be revisited. However, a case was reported in which genetic mutation in more fundamental aspects of the target pest rendered RNA-based pesticides ineffective; the occurrence of such mutations had been predicted. A study of acquired resistance to RNAi-based pesticides by Khajuria et al. [107] using WCR indicated that the resistance was due to a defect in dsRNA uptake function in gut cells of the pest. It was also suggested that dsRNA uptake into cells other than gut cells and systemic spread of RNAi silencing signals may also have been impaired. In addition, it is conceivable that mutations in genes related to the core RNAi pathway leading to defects in dsRNA processing and/or RNA-induced silencing complex formation could contribute to the acquisition of resistance to RNAi-based pesticides. To address these issues, further elucidation of the mechanism of action of RNAi in pests will be crucial. In the meantime, it will be important to delay the emergence of dsRNA-resistant pests as much as possible by developing appropriate Insecticide Resistance Management strategies, including the use of other compounds with different modes of action (e.g., Bacillus thuringiensis (Bt)-derived agents and some existing biological pesticides) and physical crop segregation. Indeed, to enhance the insecticidal activity of Bt-based biopesticides, Caccia et al. exploited the use of E. coli as a delivery vector for dsRNA that targeted the immune system of a pest of interest, the noctuid moth Spodoptera littoralis [108]; dsRNA-induced immunosuppression was shown to enhance the insecticidal effect of the Bt-based biopesticides.

Another issue in the application of RNAi-based products is the possibility of off-target effects of siRNA. This is when short dsRNAs generated from a long dsRNA silence genes other than the target gene within the target pest [109]. As the dsRNAs for RNAi action are only approximately 21 to 23 bp long, the possibility of off-target binding sites in genomic DNA is real. This is especially important if the off-target effects are likely to occur in organisms other than the target pest. To address this issue, in silico analysis of sequence homology between all possible nucleotide sequences in the designed dsRNAs and the genomic or transcriptomic sequences of relevant organisms is required [110,111]. In this regard, because it is impossible that the genome sequences of all organisms on Earth will be revealed, off-target effects can never be completely ruled out. Nevertheless, it is essential to carefully design dsRNA sequences to avoid unintended RNAi effects on non-targeted organisms using bioinformatic analyses of the latest available gene databases.

This review has principally summarized the first technical barriers to the practical application of RNAi-based products. However, wider issues must also be considered. To realize the benefits of RNAi-based products, communication among agricultural product consumers, RNAi-based product users (farmers), and producers of RNA pesticides, in cooperation with public institutions, will be essential to ensure accurate information sharing and mutual understanding based on scientific evidence of RNAi-based products. Then, appropriate management, supervision, and legislation will be necessary, and RNAi-based products must go through appropriate safety evaluation and approval processes. Sprayed or externally applied dsRNA-based pesticides are under development, but there are currently no legal regulations specific to such RNAi-based pesticides [112]. There are no specific guidance documents defining the data requirements for the approval of dsRNA-based pesticides in the European Union or the United States. The first review and recommendations on this matter have been presented by the Organisation for Economic Co-operation and Development (OECD) [113].

Once low-cost dsRNAs can be procured, various RNAi-based products can be applied practically not only in agriculture and aquaculture, but also in the control of sanitary pests such as cockroaches, fire ants, termites, and vectors of infectious disease [114,115,116,117,118]. In this way, sustainable RNAi-based products are expected to have a beneficial effect on the lives of humankind. Climate change will promote the propagation and spread of intimidating crop pests. Environmentally friendly pest control methods that can deal with the potential food crisis caused by crop pests are extremely important, and RNAi-based products are expected to play a critical role in addressing this challenge.