Initiation of the RNAi pathway takes place once dsRNA is present within the cell [32]. In the case of artificially introduced exogenous dsRNA, two mechanisms of dsRNA uptake have been described in invertebrates: transmembrane channel-mediated uptake through orthologs of the multi transmembrane protein SID-1 (systemic RNA interference deficient-1) of C. elegans and an endocytosis-mediated uptake mechanism [33].

SID-1 is a non-specific dsRNA transporter, engaged in the passive transport of dsRNA into cells and the spread of an RNAi response throughout an entire organism (systemic RNAi) [34,35]. This transporter protein is surface-expressed and displays diffusion-limited dsRNA transport that is energy independent. Recent evidence shows that SID-1 is instrumental in the import of silencing signals in C. elegans but not for export to neighboring tissues [36], and cells that are known to be refractory to gene silencing with exogenous dsRNA have been shown to be SID-1 deficient. Expression of SID-1 in these cells facilitated dsRNA uptake and subsequently an RNAi response [37]. While homologs of SID-1 have been reported in numerous organisms ranging from humans to various insects, they are noticeably absent from the genomes of others such as D. melanogaster, and this absence seems to correlate with a lack of a strong systemic RNAi response [34,38].

Endocytosis as an uptake mechanism originated from the observation that D. melanogaster cells do not express SID-1 but are still able to internalize exogenous dsRNA. Studies of Drosophila S2 culture cells resulted in the identification of several factors necessary for dsRNA uptake, of which many are also implicated in endocytosis [39]. Owing to the fact that the RNAi response was weakened by blocking endocytosis via pharmaceutical agents, the role of endocytosis in RNAi is evident [40]. Uptake experiments with FITC-labeled dsRNA conducted by Saleh et al. (2006) also showed that dsRNA molecules associate with vesicles for receptor-mediated endocytosis, with components such as vacuolar H⁺ ATPases and scavenger proteins participating. It appears as if H⁺ ATPase plays a crucial role in catalyzing an endocytic pathway for successful RNAi induction. Two scavenger receptors, Eater (which mediates phagocytosis of bacterial pathogens) and Sr-CI (a scavenger receptor class C, type I protein) facilitate uptake of over 90% of the dsRNA in D. melanogaster [39], demonstrating that receptor-mediated endocytosis a major avenue. The involvement of endocytosis in the uptake of dsRNA in C. elegans is also apparent with various “uptake-genes” being involved. RSD3 (RNAi spreading defect 3) seems to be involved in uptake via its ENTH-domains (Epsin N-terminal homology domain), which have been linked to several intracellular trafficking events [33].

The goal of the initiator step is to generate siRNAs from dsRNA. When exploiting RNAi as a reverse genetic tool, the process is artificially induced by the delivery of a dsRNA trigger. In non-mammalian systems long dsRNA (>200 bp) homologous to the target gene effectively triggers RNAi. In mammals, however, the introduction of dsRNA longer than 30 bp results in the activation of an anti-viral interferon response, which causes systemic non-specific inhibition of translation. Therefore in mammalian systems short synthetic siRNAs or DNA constructs, which express short hairpin RNAs (shRNAs), are typically used [41,42]. Since the specificity of RNAi depends on the sequence and structure of the siRNAs [30], it is of great importance to carefully design the dsRNA or shRNA-constructs in order to maximize silencing of the target gene and ensure minimal off-target effects.

The introduction of long dsRNAs into non-mammalian cells cause it to be processed into 21–25 nt double stranded RNAs (siRNAs), which have characteristic structures indicative of ribonuclease III (RNase III) cleavage [14,43,44]. Hannon and colleagues identified and demonstrated that the enzyme responsible for the initiation of RNAi in Drosophila is in fact an RNase III-like protein and owing to its biochemical function, named the enzyme Dicer [45,46]. Dicer acts as a monomer: the two ribonuclease III (RNase III) domains form a single processing center by assembling into a pseudo-dimer producing 2nt 3'-overhangs [47]. Dicer preferentially processes dsRNA from the ends of the substrates, with the Piwi/Argonaute/Zwille (PAZ) domain characterizing the 3'-overhangs [48,49,50,51]

More profound insight into the structural mechanism used by Dicer came from the crystal structure of a Dicer protein derived from the protozoan Giardia intestinalis [52]. This structure revealed that Dicer attains an elongated shape and by this means acts as a molecular ruler, measuring the distance between the terminal-binding PAZ domain and the active site, thereby giving rise to the 21–25 nt products [52,53,54]. Although the structure and mechanism of G. intestinalis Dicer has been studied in detail, much less is known about Dicer proteins from higher eukaryotes [55]. Furthermore, G. intestinalis Dicer is an atypical Dicer in that it is much smaller and simpler than that found in any other organism to date, containing only the two RNase III domains and a PAZ domain. For example, human Dicer (219 kDa) is nearly three times larger than that of G. intestinalis (82 kDa). The difference in molecular mass is accounted for by at least five additional protein domains found in most Dicer proteins [55]. These include a TAR RNA-binding protein (TRBP) [56], an amino terminal DExD helicase-like domain [57], a putative dsRNA-binding domain (dsRBD) named DUF283 [58], an Argonaute-binding domain [59] and a C-terminal dsRBD [60]. These additional domains participate in dsRNA processing, regulate dicing activity and serve as molecular scaffolds for consolidating protein factors involved in the initiation of RNAi [55].

Once a dsRNA trigger has been processed into siRNAs, the guide strand is identified and loaded into an effector ribonucleoprotein complex, the RNA-induced silencing complex (RISC). Since siRNAs cannot catalyze any reaction by themselves, RISC assembly is a key process for these small RNAs to exert their function [61]. To date, a complete understanding of the exact mechanism underlying this process remains elusive. Based on several in vitro studies of RISC assembly using D. melanogaster extracts and siRNAs as substrates, together with new structural studies of Argonaute (Ago), a revised 2-step model of the main features of this process has been proposed [62,63]. The first step, known as RISC-loading, involves the insertion of the siRNA duplexes into the Ago-protein. The second step entails the dissociation or unwinding of the passenger strand from RISC [61].

It is postulated that Dicer and a dsRNA binding partner (e.g. in D. melanogaster these are Dicer-2 and R2D2) are bound to the double stranded siRNAs, to direct them to the RISC complex [64]. Current findings suggest that Dicer-2 and R2D2 form a heterodimer that binds the siRNA, with Dicer binding the thermodynamically less stable end of the siRNA and the dsRBD domain binding the more stable end [61,65]. These thermodynamic and binding asymmetries appear to determine which strand is finally incorporated into RISC. The guide strand is always the strand whose 5’ end is less tightly paired to its complement, implying the strand with the least thermodynamically sTable 5’ end [66,67]. Once the Dicer/dsRBD/siRNA ternary complex is formed it assembles with Ago-protein to form a RISC loading complex (RLC), which is an intermediate from which Dicer/dsRBD is gradually displaced by the Ago-protein to form pre-RISC [61,64,65,68,69]. Once bound to the siRNA duplex, Ago cleaves the passenger strand, triggering its dissociation [70]. After discharge of the passenger strand, possibly in an ATP-dependent step, the guide strand is in close association with Ago, forming a fully mature RISC, also referred to as siRISC [61,64,65,71].

In D. melanogaster, several RISC components have been identified, but only a few have been characterized at the functional level [72]. From current findings it is evident that RISC consists of several proteins and RNA molecules that act as key actors of RNAi promoting mRNA degradation, repression of translation and remodeling of chromatin structure [73]. The central catalytic component of RISC, Argonaute 2, was the first to be identified [74]. Argonaute proteins are characterized by two conserved domains: the PAZ (also present in Dicer) and PIWI domains [75]. Based on a higher degree of homology to either Arabidopsis AGO-1 or Drosophila PIWI, this family of proteins is subdivided into two highly conserved subclasses [76]. Recently an additional subclass, C. elegans-specific group 3 Argonautes, has been identified [77]. Other components which have been identified as part of the RISC-complex include the Vasa intronic gene (VIG)-RNA binding protein, helicase proteins, Tudor- staphylococcal nuclease (Tudor-SN) and Fragile X -protein (dFXR) homologs [78,79]. To date most of these proteins remain functionally uncharacterized, however. Due to the lack of structural data, the exact catalytic mechanism of target mRNA cleavage remains unclear.

In the RISC complex the 3’ end of the siRNA guide strand is bound by the PAZ domain, while the PIWI domain contacts the 5’ phosphorylated end [49,50]. Structural studies of prokaryotic Ago-like protein in complex with siRNA-mimics demonstrated that the sugar-phosphate backbone of the 5’ end of the guide strand (nucleotide 2 to 8) contacts the PIWI domain in such a manner that bases 2 to 8 are presented on the surface [80,81]. This region is known as the ‘seed’ region [14]. Corresponding mRNA targets are initially bound by the seed region of the siRNA and then pairing is extended to the 3’ end. A single cleavage of the target mRNA occurs across from nucleotide 10 and 11 (with respect to the 5’ end) of the guide stand. The catalytic engine is an RNaseH fold present in the PIWI domain [75,81,82]. Although target cleavage by Argonaute does not require ATP per se, release of the cleaved target (RISC recycling) is stimulated by ATP [83,84]. From in vivo studies it is evident that, once released from RISC, the 3’ mRNA fragment is degraded in the cytoplasm by exonuclease Xrn1, while the 5’ fragment is degraded by a complex of exonucleases, the so-called exosome [85]. In the RNAi pathway, siRISC operates as a multiple turn-over enzyme. Once the mRNA target is cleaved, siRISC dissociates from the cleaved mRNA to repeat another cleavage cycle [30].

In plant and nematode species it has been described that the RNAi signal is amplified by an RNA-dependent RNA polymerase (RdRP), which results in a systemic RNAi effect that can spread from the targeted/treated tissue to other tissues [86]. To date, this RdRP-dependent RNAi amplification seems absent in higher eukaryotes, and even from D. melanogaster and mosquitoes. However, based on several RNAi studies in ticks it is evident that this phenomenon does occur in ticks, since application of dsRNA (via body cavity injection, feeding or soaking) results in global and persistent gene silencing in both treated ticks and their progeny [87,88,89,90]. Recent preliminary studies have indicated that this systemic RNAi effect in ticks is most likely attributed to a tick RdRP homolog [26].

In order for RNAi to spread from one cell to another a transport system is required. In C. elegans several proteins, necessary for systemic RNAi have been identified. These include a multi transmembrane protein SID-1 (systemic RNA interference deficient-1(sid-1)) thought to act as a channel for dsRNA uptake [34], as well as RNAi spreading defective proteins (RSD2, RSD3, RSD6) [91]. To date only RSD3 and Endocytic protein (Epn-1) have been identified in ticks, leaving the mechanism of cell to cell dsRNA transport in ticks unresolved. Although more components of the tick RNAi machinery have been identified, the exact mechanism remains to be determined [26]. The current model for most invertebrate parasites RNAi is based on integrated knowledge of the RNAi process of model/other organisms including, D. melanogaster, C. elegans and mosquitoes (Figure 2).

In various plant and nematode species it has been found that the silencing signal is amplified by a process reliant on a RNA-dependent RNA polymerase (RdRP), which in effect permits a systemic RNAi response from the site of origin to eventually most or all other tissues [26,86]. A two-step mechanism has been described for RdRP amplification in C. elegans [92]. Once the primary siRNA has been generated by Dicer, RdRP uses the guide siRNA strand as primer and native mRNA as template to synthesize abundant secondary siRNAs and increasing the effectiveness of RNAi (Figure 2). As RNA synthesis occurs in a 5'–3' direction, this amplification leads to the 5' spreading of the initial RNA interference signal. This phenomenon is known as transitive RNAi [92].

To date RdRP-dependent RNAi amplification seems absent in D. melanogaster, many other insects and higher eukaryotes. Interestingly, a putative homolog of the RdRP EGO-1 protein from C. elegans has also been identified in the tick species I. scapularis. This protein is associated with the amplification of trigger RNA by transcribing additional dsRNA molecules using target mRNA as a template, thereby yielding secondary siRNAs and increasing the effectiveness of RNAi [40,93]. RdRPs appear to be absent in other metazoans except for several nematodes and a lancelet (Branchiostoma floridae) [94]. The tick RNAi pathway may therefore differ from that of other arthropods, but more research is required to clarify the exact RNAi mechanism.

In order for RNAi to spread from one cell to another a transport system is required. The phenomenon of the progression of an RNAi response from the site of origin to neighboring cells and eventually to all, or most, tissues of the organism was first described in plants and C. elegans . The systemic response may even persist through multiple developmental stages, including being transferred through the germline to the progeny. It has been proposed that this response has evolved as a strategy to increase the silencing efficiency [95] and it is known to form part of the immune system in plants. For the functioning of a robust systemic response there are three essential steps: firstly the uptake of an extracellular dsRNA signal into a cell, secondly the amplification and distribution of the said signal to other cells, followed by a successful degradation or blocking of a homologous mRNA transcript [38].

Although the exact mechanisms pertaining to a systemic RNAi response remain unclear, many benefits associated with its presence makes it a highly desirable feature in organisms investigated using RNAi. Firstly, the amplification of the introduced RNAi signal means that lesser/smaller amounts of dsRNA can evoke the needed response throughout the entire organism. Secondly, the fact that silencing is spread and effected throughout all tissues makes target gene selection of the whole genome of the studied organism possible. This is especially relevant in those experiments where the mode of administering dsRNA cannot deliver it to all tissues, for example through oral delivery or direct injection. Another application made possible is the study of genes in life stages such as eggs that may not be amenable to other methods of dsRNA delivery. By delivering the relevant dsRNA to a preceding life stage, in this case the adult female, systemic RNAi will propagate the necessary signal throughout the adult as well as within any eggs produced by that female [88]. This application is known as parental, trans-stadial or transovarial RNA interference. Exploiting systemic RNA interference during RNAi studies is also a relatively simple procedure. For these reasons there is a definite advantage and incentive to elucidate the complete mechanism of systemic RNAi so that it could be induced experimentally in organisms that do not show a systemic RNAi response themselves.

Although more pieces of the RNAi machinery of different disease vectors and pathogens have been identified, the exact mechanism of each remains unknown. Current RNAi models for most invertebrate parasites is based on integrated knowledge of the RNAi process of model/other organisms including D. melanogaster, C. elegans (Figure 2).

Apart from model organisms such as Drosophila, ticks have provided us with a significant understanding of RNAi in arthropod vectors of disease and hence will be discussed in a bit more detail in this section. Ticks are obligate hematophagous parasites and are important vectors of a wide variety of pathogens including viruses, protozoa and bacteria which affect human and animal health worldwide [96]. Elucidating tick-host-pathogen interactions is of interest for the development of novel intervention strategies such as anti-tick and pathogen-transmission blocking vaccines to limit the damage caused by ticks and prevent the spread of pathogens. The increasing availability of genomic data from tick-borne pathogens and ticks such as Ixodes scapularis and Rhipicephalus (Boophilus) microplus has introduced a new era in tick research and greatly advanced studies addressing tick biology and host-vector-pathogen interactions [97,98,99]. The possibility to silence the expression of genes in ticks using RNA interference (RNAi) has proven to be pivotal in numerous studies which focus on deciphering tick gene function. Since the first report of an RNAi experiment in Amblyomma americanum ticks in 2002 [100], this method has been applied successfully in other hard ticks belonging to the Amblyomma, Dermacentor, Haemaphysalis, Ixodes, and Rhipicephalus genera and in Ornithodoros soft ticks [101,102,103].

Mites belong to the same subclass as ticks (Acari) and may also act as vectors of disease, albeit to a lesser extent. The most important mite-borne disease of humans is scrub typhus (Orientia tsutsugamushi), which is primarily transmitted by several members of the genus Leptotrombidium. Parasitic mites living on or burrowing in the skin may also cause dermatitis in humans and animals. The full genome of the two-spotted spider mite Tetranychus urticae, a plant-feeding mite, was recently published and is the first chelicerate genome to be completely sequenced [104]. Successful gene silencing by RNAi has also been described for this mite and the honey bee mite Varroa destructor [105,106].

Two mechanisms of dsRNA uptake have been described in arthropods: a transmembrane channel-mediated uptake through orthologs of the SID-1 protein of C. elegans and an endocytosis-mediated uptake [33]. As SID-1 is not present in tick and mite genome sequence databases, an endocytosis-mediated uptake is hypothesized in Acari. In the endocytosis-mediated uptake of dsRNA in Drosophila, scavenger receptors are thought to mediate dsRNA uptake [39,40]. Evidence for an endocytosis mediated mechanism for dsRNA uptake with an essential role for a scavenger receptor in ticks was recently provided by an ‘RNAi of RNAi’ approach. When the expression of HlSRB, the class B scavenger receptor CD36 from Haemaphysalis longicornis, was silenced 96h before dsRNA coding for a second target gene (HlVg-1 or HlVgR, encoding vitellogenin and vitellogenin receptor) was introduced, subsequent knockdown of the second gene was inhibited. A double knockdown of target genes could however be observed when HlSRB was silenced 96h after the knockdown of HlVg-1 or HlVgR [93]. This suggests that receptor-mediated endocytosis is also relevant for systemic RNAi in ticks, with an essential role for scavenger receptor class B. Orthologs of the class B scavenger receptor CD36 within the superorder Parasitiformes have been annotated for I. scapularis (GenBank Accession number XP002409323) and the western predatory mite Metaseiulus occidentalis (XP003743396). Other key components of intracellular RNAi were recently identified from tick and mite genomic sequence databases by homology searches [26,104]. These include putative Dicers responsible for the cleavage of dsRNA to siRNAs, Argonaute proteins which function as the catalytic component of RISC and others (Table 2).

RNAi is usually accomplished by treatment of ticks, tick tissues, tick cells or mites with dsRNA, but the successful use of siRNA to silence gene expression in these specimens has also been reported [106,107,108,109]. Most laboratories transcribe dsRNA in vitro from a PCR template which has been amplified using primers containing a T7 promoter sequence. The relative simplicity of synthesizing dsRNA and its high efficacy have made it the most commonly used mediator in tick RNAi studies. A drawback of the use of long dsRNA is that it increases the chance of causing off-target effects (OTE), leading to false discovery. OTE are caused by species-specific recognition of transcripts other than the intended target. This can be addressed in the experimental design by preventing the use of dsRNA sequences of which nucleotide stretches may recognize non-target transcripts. Various in silico prediction algorithms to analyze possible OTE have been described [110,111] but their application in tick research is frustrated by the absence of full tick genome data. The use of siRNA, small dsRNA (100–200 bp) or of two non-overlapping dsRNAs per gene to assess whether both induce the same phenotype, may limit the occurrence or facilitate the detection of OTE [110,112]. However, these strategies have thus far found only limited adaptation in tick research using RNAi.

The injection of dsRNA into the hemacoel via the ventral exoskeleton surface using fine-gauge or pulled capillary needles is the most commonly used delivery method for dsRNA in ticks. After a recovery period to monitor for tick mortality resulting from the injection, ticks are fed on experimental animals or an artificial feeding system [100,113]. Parameters such as tick mortality, tick weight after feeding and the amount of eggs oviposited can subsequently be recorded to evaluate the effect of gene silencing on general tick function. Biological material such as hemolymph or saliva and tissues can furthermore be collected for further analysis in functional assays and histological examination [114,115,116,117]. By injecting dsRNA into the hemocoel of freshly engorged female ticks through the spiracular plate, gene silencing can be established in the eggs and larvae hatching from these eggs [88,118,119]. A similar method can be used to silence genes in the eggs of mites [106]. The spread of dsRNA to different tissues following injection in the hemacoel demonstrates that RNAi is systemic in the Acari lineage.

Infection of the I. scapularis cell line ISE6 with the a recombinant Semliki Forest Virus replicon expressing heterologous Nairovirus sequences induced siRNAs which inhibited the replication of Hazara virus, a Nairovirus, by RNAi in this cell line [120]. Besides this viral delivery, the incubation with dsRNA or siRNA is also effective in establishing gene silencing by RNAi in tick cells. Optimum conditions were however shown to vary for different cell lines. DsRNA was more efficient for gene silencing by RNAi in tick cells than siRNA and use of the latter requires a transfection reagent [121]. RNAi can also be used to study the effects of gene silencing on tissue function ex vivo following the immersion of dissected tissues in a dsRNA solution [113,122,123]. Overnight soaking at 4 °C in a 0.9% saline solution containing dsRNA at a concentration of 2,5 µg/µL has been applied successfully to establish gene silencing in whole Varroa destructor mites. The addition of NaCl was apparently necessary as soaking in plain water containing dsRNA was not effective [105]. The soaking of whole ticks, which would reduce the laboriousness and mortality associated with injections and simplify dsRNA delivery especially in juvenile tick stages, has also been reported for Haemaphysalis longicornis nymphs but requires further optimization [124]. Soaking of eggs and nymphs in combination with electroporation was successfully used for dsRNA delivery in I. scapularis ticks [125].

Further evidence for the occurrence of systemic RNAi in ticks was provided by Soares et al., who delivered dsRNA to I. scapularis nymphs using capillary feeding of a dsRNA solution. Ingested dsRNA targeting an anticomplement gene (isac) in the salivary gland spread via the gut to the target organ, where it resulted in a knockdown of isac expression [89].

RNAi has also helped to unravel the molecular complexity of tick physiology and biology. Hard ticks for example concentrate their blood meal during feeding by secreting excess water imbibed with the blood back into the host with saliva. A water channel or aquaporin (AQP) was found to be expressed in the gut and salivary glands, organs which are associated with a high water flux. Gene silencing of AQP in I. ricinus females by RNAi disrupted the water passage from gut to salivary glands, resulting in an accumulation of blood in the gut which could not be concentrated. The gut subsequently became too full to accommodate more blood and the total blood volume ingested was significantly reduced. The hemolymph osmolarity consequently increased and fluid secretion by isolated and ligated salivary glands was abolished [113]. Other examples of biological and physiological processes which have been studied in ticks include blood feeding [89,115,124,126,127,128,129,130,131], blood digestion [132,133], iron metabolism [134], reproduction [135,136] and innate immunity [137].

Another interesting application of RNAi in ticks is its use as a tool to screen for tick protective antigens which may be suitable anti-tick vaccine candidates [138,139]. In these screens, dsRNA is transcribed from individual clones of a cDNA library and screened in increasingly smaller pools for their effect on tick feeding and survival. This may result in the identification of proteins that are critical for the tick which can subsequently be evaluated as potential vaccine antigens. By using this approach, researchers could identify subolesin, a protein which was shown to have efficacy as a recombinant antigen in vaccination trials against various tick species, the poultry red mite Dermanyssus gallinae, mosquitoes and sand flies [140].

Silencing of subolesin produces ticks with diminished reproductive performance which prevents successful mating and the production of viable offspring. The release of subolesin-silenced ticks, alone or in combination with subolesin vaccination, was proposed as an alternative tick control strategy [119,141]. Although interesting, many questions concerning the feasibility and consequences of this approach, such as the reduced fitness of modified ticks and the ecological impact of their release, remain to be answered before this or similar methods can be put into practice.

Injection of adult female mites or engorged female ticks with dsRNA results in gene silencing in their offspring, a phenomenon which is also known as ‘parental’ or ‘transovarial’ gene silencing [88,106,118]. Parental RNAi may be used to study the function of genes in embryogenesis and will facilitate studies in evolutionary developmental biology, an area of tick and mite research which lags behind that of many other invertebrates.

To study tick-pathogen interactions, a typical experimental approach starts with the identification of tick genes which are differentially expressed when ticks are infected with a pathogen. Techniques such as suppression subtractive hybridization (SSH) [142], two-dimensional (2-D) gel electrophoresis [143], 2-D fluorescence difference gel electrophoresis (DIGE) [127], RT-PCR [144] and microarrays [145] have been used for this purpose. The differential expression is usually confirmed by (semi-)quantitative RT-PCR. The expression of genes which are found to be up-regulated following infection can subsequently be silenced by RNAi to study the effect of gene knockdown on the course of the infection. This approach was for instance followed to identify tick salivary gland proteins up-regulated in I. scapularis ticks in response to infection with Borrelia burgdorferi, the causal agent of Lyme disease of which I. scapularis is the predominant vector in North America. DIGE coupled with mass spectrometric analysis using extracts from B. burgdorferi-infected and uninfected salivary glands led to the identification of a histamine release factor (tHRF) which was strongly up-regulated in the salivary glands of B. burgdorferi-infected ticks during engorgement. Gene silencing of tHRF in B. burgdorferi-infected nymphs impaired successful feeding and resulted in reduced spirochete levels in ticks and the mice they had fed upon. It is suggested that the decreased feeding ability of tHRF-deficient ticks may have disrupted the B. burgdorferi replication and dissemination inside the ticks, reducing the efficiency of spirochete transmission [127]. The laboratory of Erol Fikirg, also identified 14 different salivary gland proteins (salp) which are secreted into the host during feeding, by probing an I. scapularis salivary gland cDNA expression library with serum from tick-immune rabbits [146]. The expression of two of these genes, salp15 and salp16, was found to be selectively increased in I. scapularis ticks infected with B. burgdorferi or A. phagocytophilum, respectively [144,147]. Gene silencing of salp15 and salp16 did not impair tick feeding as tHRF did, but did interfere with the pathogen life cycle. While the B. burgdorferi levels in the salivary glands of control and salp15-deficient ticks were similar, the spirochete burden was significantly lower in the mice on which salp15-deficient ticks had fed. This demonstrated that Salp15 facilitates tick-borne B. burgdorferi infection in the host [147]. Silencing the gene expression of salp16 inhibited the acquisition of A. phagocytophilum by the vector tick from infected mice. Pathogen levels in the gut of control and salp16-deficient ticks which had fed on infected mice were similar, but the A. phagocytophilum levels in salivary glands were significantly lower in salp16-silenced ticks, suggesting that A. phagocytophilum requires salp16 to infect the salivary glands [144]. Other functional analysis studies based on RNAi which focused on the tick immune responses have also identified genes which may be involved in the vectorial capacity of ticks for a variety of pathogens [114,142,148,149].These examples demonstrate the intimate relationships between vector-borne pathogens and their arthropod hosts on which RNAi shed light.

Lice are obligate ectoparasites of mammals and birds which may feed on blood (bloodsucking lice of the suborder Anoplura) or on skin debris, hairs and feathers (chewing lice). Three Anoplura species are ectoparasites of humans: the head louse (Pediculus humanus capitis), the human body louse (Pediculus h. humanus) and the crab louse (Phthirus pubis). Of these three, only the body louse is a vector of pathogens, including the causal agents of epidemic typhus (Rickettsia prowazekii), trench fever (Bartonella quintata) and louseborne relapsing fever (Borrelia recurrentis) [150]. The genome of the body louse was recently sequenced and is with 108 Mb the smallest insect genome known to date [151]. A gene-specific reduction in transcript levels following the injection of dsRNA in female lice was recently reported [152] and key components of the RNAi pathway have also been identified in the human body louse genome [150]. Their small genome size, the presence of a functional RNAi pathway and differences in their ability to transmit pathogens make the head and body louse interesting models for comparative functional genomic studies to unravel the molecular mechanisms which determine vector competence [150].

RNAi-based silencing of gene expression has been achieved in mosquitoes using a number of different dsRNA delivery methods. In the majority of studies to date, in vitro synthesized dsRNA have been injected directly into the mosquito hemolymph. Those that survive the micro-injection procedure are then fed or exposed to the relevant physiological condition and/or stress factors and evaluated for the phenotypic effect caused by knockdown of the target gene. In mosquitoes, these often relate to effects on certain aspects of fecundity such as oviposition and egg development, or physiological evidence of cell damage, insect mortality or behavioral changes (reviewed in [155,156]. The micro-injection procedure, however, is not optimal or efficient particularly in light of the mosquito’s small size. One alternative is to use oral introduction, as demonstrated in a study where chitin synthase-targeted dsRNA was fed to An. gambiae larvae, resulting in significant mortality [157]. Another is by topical application of dsRNA, demonstrated in Ae. aegypti adults [158]. In a first, Coy et al. recently demonstrated that knockdown can be achieved in adult mosquitoes by the oral delivery of dsRNA via a liquid sucrose meal. The dsRNA was ingested by adult females and caused a reduction in the level of transcription of the vacuolar ATPase gene target [159]. A recombinant densovirus-mediated RNA interference system has also recently been developed for both in vitro and in vivo use [160]. The efficiency of RNAi has been shown to differ between tissue types, with midgut tissues and hemocytes being amenable to effective knockdown whereas inhibition of genes in the salivary glands required much higher dosages of dsRNA. Nonetheless, the larger quantities appeared to not compromise the silencing specificity [161]. Details of the uptake mechanism of exogenously added dsRNA and its subsequent processing has not been resolved yet, but processing is thought relate to the intrinsic virus-mediated RNAi response of mosquitoes (summarized below and reviewed in [33,162].

Functional studies using RNAi have provided a significant amount of insight into vector biology, such as integrated pathways that respond to hormonal and nutritional signals affecting feeding, vitellogenesis and egg production [153,163,164,165,166,167]; factors involved in lipid biosynthesis [168], vesicle transport between the ER and Golgi [169], steroid metabolism [166,170,171]. A large portion of literature also addresses olfaction in mosquitoes that influences aspects of their behavior, such as host seeking, feeding, mate selection and location of appropriate breeding sites [172,173]. Erdelyan and colleagues (2012) identified two CO₂ receptor genes, AaGr1 or AaGr3 that upon knockdown resulted in a loss of CO₂ sensitivity in Ae. aegypti and Culex pipiens quinquefasciatus mosquitoes, impacting host detection. Interruption of the olfactory system could be used to design compounds that disrupt the mosquito's ability to find a host, thereby limiting the potential for pathogen transmission [174].

Initial studies supporting RNAi-based gene silencing as a suitable tool for understanding the vector-pathogen interface in mosquitoes, include Dengue virus (DENV) replication inhibition by transfection of an Aedes albopictus cell line using DEN1- and DEN2-specific dsRNAs [175,176] and knockdown of the antimicrobial gene defensin by in vivo injection of gene-targeted dsRNA into the thorax of adult Anopheles gambiae [177]. A different approach [178], where dsRNA was expressed in situ from a transgene stably integrated into the mosquito genome and designed to transcribe an intron-spliced RNAi hairpin loop from an upstream promoter, have also become of great use. A myriad of studies have been reported, with most using direct delivery of dsRNA rather than the more demanding approach of transgenesis.

With regard to viral transmission by mosquitoes, the long-standing view that mosquito innate immunity likely plays a role in restricting virus replication has been addressed in a number of studies. Based on the knowledge of immunity in D. melanogaster, orthologs of genes encoding Dcr2, R2D2 and Ago2, components of the intrinsic arthropod RNAi pathway, were silenced in a number of independent studies. Suppression of the mosquito RNAi defense mechanism resulted in an increase of alphavirus- and DENV replication in Aedes aegypti cultures and mosquitoes [27,179] and O’nyong-nyong replication in Anopheles gambiae [180]. These and other studies have provided clear evidence that RNAi is the major defense mechanism against arbovirus infections [181]. Some evidence has also been presented that the RNAi response may be systemic. In cultured Aedes albopictus cells, direct cell-to-cell spread of an RNAi signal inhibited the replication of incoming Semliki Forest viruses [182]. A similar spreading signal has been observed in vivo in D. melanogaster but the mechanism remains to be characterized. Studies have also provided evidence for the production of virus-derived piRNAs following arbovirus infection in mosquito cells, suggesting that this pathway may be another facet of antiviral defense [183]. It has also been possible to generate DENV2-resistant Ae. aegypti mosquitoes by constitutively triggering the RNA interference pathway via long DENV-specific dsRNAs transcribed from an integrated transgene. This transgenic approach provided heritable resistance to virus replication and transmission across 17 generation [184] before expression of the effector gene was then lost for unknown reasons. This encouraging result illustrates the potential of anti-pathogen suppression using RNAi, as explained in more detail in section 7.

Microarray-based transcriptional profiling in combination with RNAi is providing new information on immune signals that respond to virus infection. In one such a large-scale study, DENV infection of Ae. aegypti mosquitoes resulted in the identification of more than 70 genes putatively associated with immune signaling [185]. Knockdown of candidates has shown that both the Toll pathway [185] and JAK-STAT pathway [186], conventionally associated with defense against bacteria and fungi, also contribute to Ae. aegypti virus responses [187,188]. A recent study by Sim and Dimopolous provided evidence that Arboviruses are capable of actively suppressing the immune pathways in cells they infect, rather than simply failing to trigger host immunity [189]. When the transcriptome of Ae. aegypti cells infected with DENV was analyzed using microarrays, the Toll pathway, as in mosquitoes, was activated. In addition, numerous immune signaling molecules and anti-microbial peptides (AMPs) were down-regulated. In a similar study, Colpitts et al. also demonstrated down-regulation of four cecropins and a defensin in Ae. aegypti mosquitoes infected with respectively DENV, WNV and Yellow fever virus [190]. The molecular details of the findings in these studies have yet to be validated by functional RNAi studies.

Finally, mammalian cells infected with DENV undergo apoptosis due to a shutting down of protein synthesis in host cells. In mosquito cells, however, apoptosis is mostly inhibited, likely to allow for the spreading of viral progeny to neighboring cells while evading host inflammatory responses [191]. Chen et al. provided evidence that mosquito cells also have strategies for protecting themselves against stresses induced by virus infection, such as GST that is involved in antioxidant defenses and an inhibitor of apoptosis, IAP. When GST was silenced in DENV-infected C6/36 mosquito cells, there was a higher rate of apoptosis [192]. Furthermore, silencing of IAP rendered mosquito cells highly sensitive to DENV-induced oxidative stress [193].

There is considerable interest in understanding the molecular and cellular interactions between mosquito vectors and protozoan parasites, with many studies focusing on the interaction between An. gambiae and Plasmodium species. In the past decade, RNAi approaches have been applied in numerous studies to unravel the genetic basis of the mosquito’s susceptibility and the immune responses elicited by the parasite. A multitude of mosquito genes that respond to Plasmodium infection have been identified in recent years as reviewed in a number of papers [194,195,196]. These include genes that may either negatively influence Plasmodium development, i.e. antagonists, or have a beneficial effect on parasite development, i.e., agonists. Some of the first An. gambiae genes identified and verified by RNAi to encode anti-parasitic factors were TEP1 [197] and LRIM1 [198]. These are both pattern recognition receptors (PPRs) of the innate immune system, with TEP1 being a complement-like protein and LRIM1 a C-type lectin. RNAi-mediated silencing of TEP1 prevented melanization of parasites, a mosquito response to lower parasite load, while silencing of LRIM1 resulted in a substantial increase in oocyte numbers. APL1, a leucine-rich repeat protein was also verified as an antagonist [199]. Knockdown studies have since indicated that LRIM1 and APL1C occur as a heterodimer in the mosquito hemolymph that interacts with TEP1 [200,201]. It therefore seems as if the LRIM1/ApL1C/TEP1 complex’s function is to target Plasmodium parasites for destruction.

Proteins expressed in the mosquito salivary gland that are involved in parasite invasion have also been characterized. A study of Gosh and colleagues demonstrated that interaction between the salivary-gland-specific surface protein saglin and the parasite surface protein TRAP (Thrombospondin-related anonymous protein) is essential for spoorozoite invasion. In vivo RNAi of saglin results in a strong inhibition of salivary gland invasion by parasite [202]. Many antagonists have since been discovered and their function analyzed using RNAi-based techniques. Most notably, these include the hypervariable PPR AgDscam [203], members of the fibrinogen-related protein family (FREP) [204] and the C-type lectin CLSP2 [205].

Gene silencing approaches have further shown that An. gambiae expresses proteins that protect Plasmodium from its immune defenses, or aid in their transport through membrane barriers. The C-type lectins CLT4 and CTLM2A serve as examples of immune system agonists. Gene knockdown / RNAi-mediated silencing of these genes results in massive melanization of ookinetes [198]. Some examples of mosquito proteins required for Plasmodium midgut and salivary gland invasion include AgESP, epithelial serine protease and upstream regulator of cytoskeleton remodeling [206], PRS1 (Plasmodium responsive salivary 1), a protein up-regulated in infected salivary glands [207], the antibacterial protein lysozyme c-1 [208], and hemolymph lipid transporter, ApoII/I [209]. RNAi-mediated silencing of these genes in some way caused a phenotype that validated their involvement in aiding Plasmodium invasion, reproduction and/or transmission. Proteins with properties like these are obvious candidates that merit further investigation for possible use in future anti-malaria strategies.

Introduction of dsRNA into tsetse flies is done via feeding of flies on dsRNA containing bloodmeals or by in vivo injection [216]. To date a vast amount of genes have been silenced in the tsetse fly to gain insight into their biological functions and contribution to vector fitness. Some examples include silencing of Lipophorin, a lipid transport protein. Silencing of Lipophorin resulted in decreased lipid levels, delayed oocyte development and extended larval gestation periods but the findings still suggest another carrier lipoprotein to also be required for these processes in the tsetse fly [16]. Disruption of a nucleotidase-related protein in the salivary glands reduced anti-thrombotic activities by some 50%, confirming that nucleotidases in tsetse flies, similar to other haematophagous ectoparasites, constitute a major part of the vector’s anti-haemostatic machinery [17]. RNAi studies of two milk proteins GmmMGP2 and GmmMGP3, which are essential for viviparous reproduction, identified only GmmMGP2 as the gene impacting fecundity in tsetse flies. The discrimination power of RNAi revealed that only in GmmMGP2 knock-down flies, were ovulation, oocyte accumulation and degradation disrupted and fecundity reduced [217,218].

Noteworthy is the number of papers in which RNAi was used successfully to disentangle host-pathogen interactions between tsetse flies and the causative agent of African and animal trypanosomiasis, T. brucei. In the past five years, the following insights have been gained: (1) Tsetse ALBA RNA-binding proteins are involved in trypanosome development in vivo [219]; (2) The EP protein is a potent immune-responsive molecule in the tsetse fly capable of reducing T. brucei infection in the vector midgut [15,220]; (3) Although 16 small Rab GTPases are present in T. brucei, three Rab proteins RabX1-3 are novel and play a role in fly infectivity [221]; (4)The bacterial endosymbiont of the tsetse fly, Wigglesworthia glossinidiae, actually protects the vector from being invaded by T. brucei. The interaction between the tsetse peptidoglycan recognition proteins, the bacteria, the level of anti-microbial peptides in the vector, and the mechanism of action are well described [222]. Finally, silencing of transferrin in the fat body tissues of tsetse flies increased the number of trypanosome infections, indicating a role for transferrin in protecting tsetse flies against infection [15].

One paper on understanding the host-vector interactions was published by Caljon and colleagues. In this study, tsetse flies were fed on rabbits to obtain anti-sera for subsequent screening of a cDNA expression library, resulting in the identification of tsetse antigen 5 (TAg5). In patients with allergic reactions to tsetse flies, IgE antibodies against TAg5 were detected, showing that this antigen is involved in the host-vector interface. Silencing of TAg5 in tsetse flies confirmed this observation as saliva from silenced flies displayed a reduced IgE binding potential. This study has contributed in confirming antigen-5 related proteins as functional allergens [223].

A standard PubMed search conducted with the terms Trypanosoma brucei and gene silencing or RNA interference indicate that 495 papers have been published using RNAi in T. brucei. We suggest reading some of the following reviews addressing new understanding of parasite biology and the identification of novel drug targets for parasite control [1,224,225,226].

Transgenesis entails the transformation of vectors with transgenes that impair vector and/or pathogen development. This growing field of controlling vector-borne diseases is reviewed in [227,228,229]. Noteworthy is the contributions that RNAi-studies have made to the identification and evaluation of genes which are now used as transgenes to impact the mosquito vector-pathogen interactions (Table 3). Studies on transgenes affecting vector immune response genes such as Rel regulatory molecules (which affect the expression of numerous immune genes) and anti-microbial peptide genes such as defensins and cecropins, have pioneered this field [230]. With the growth in high-throughput screening studies, it is anticipated that the number of RNAi-discovered transgenes for vector-borne diseases will escalate within the next few years.

An additional strategy to reduce vector competence for disease transmission, involves genetic manipulation of natural endosymbionts of a particular vector species to produce molecules that impedes pathogen transmission or survival [239]. As described in this review, RNAi has already provided much insight into vector-pathogen interactions and paratrangenesis is already being used with success in vectors listed on VectorBase. This strategy was first described in Rhodnius prolixus, the Chagas disease vector, using its endosymbiont, Rhodococcus rhodnii. A gene encoding an antimicrobial peptide, cecropin A (which was evaluated using RNAi), was cloned into R. rhodnii that was stably expressed and released into the gut lumen, killing the disease causing pathogen T. cruzi with limited toxicity to the host tissues. Moreover, the transformed bacteria were stably maintained without the need for antibiotic selection and could be successfully transmitted to hosts with non-transformed endosymbionts due to the coprophagic habits of reduviids. In later studies, functionally active single chain antibody fragments (ScFv) were successfully expressed using a modified endosymbiotic Corynebacterium spiecies in the kissing bug, Triatoma infestans, as well as R. rhodnii in R. prolixus for control of T. cruzi transmission [239].

Other bacterial symbionts that have been successfully transformed and stably transmitted include Sodalis glossinidius (from Glossina sp.) and Asia sp. (from Anopheles spp.) [240,241]. Transformed Asia was also shown to be paternally transmitted in An. stephensi, providing the possibility of using non-hematophagous mosquitoes to spread modified endosymbionts for pathogen transmission control [242]. A reduction in vector competence may also be achieved by using genetically modified densovirusses in mosquitoes. In this regard, transmission of modified virus has been reported for Ae. aegypti and A gambiae, respectively [243,244]. To date, a number of endosymbionts have been identified may be used in paratransgenic control of insect-borne diseases from triatomine, tsetse fly and mosquito vectors [227]. However, few species have been successfully modified and shown to be stably transmitted to non-transformed individuals and their resultant progeny. In theory, genetically modified endosymbionts may not only provide researches with an additional tool for fighting disease but it can also deliver proteins and enzymes that test functional genetic hypotheses formulated from RNAi studies.

The discovery of RNAi, for which Andrew Fire and Craig C. Mello were awarded the Nobel Prize in Physiology or Medicine in 2006, has once more demonstrated the importance of basic science in understanding molecular and physiological mechanisms. With this revelation the door was opened to the development of powerful approaches for investigation of functional genomics in a wide array of organisms, including those of importance in human and animal health. In the post genomic era, with a vast array of sequence information, RNAi provides a method to functionally annotate and subsequently validate the expanding spectrum of putative vaccine and drug targets, this is especially evident in the fields of tick and mosquito research.

RNAi in combination with recently developed technologies, such as high-throughput screening provides a platform to investigate proteins of vital importance in pathogen survival, transmission and the pathogen-vector-host interface. RNAi together with technologies such as deep sequencing revolutionized pathogen and vector research by enabling the comprehensive and refined annotation of the genomes of these organisms, leading to better knowledge of their RNA, proteins and metabolite derivatives. This finally provides an integrated systematic molecular biological approach for new optimized control strategies against these medical and veterinary important organisms [245].

Together with genomic data and RNAi integrated technologies basic biological questions regarding vectors and their pathogens are in the process of being elucidated. High-throughput sequencing provides an invaluable tool to furthermore investigate the full repertoire of RNAi machinery, to identify essential components of the pathway and even perhaps to reveal new classes of small silencing RNAs. This knowledge will allow for fundamental improvements in RNAi efficiency in RNAi-positive organisms and furthermore permit applied research, as the RNAi machinery can be reconstructed in organisms lacking the RNAi pathway, since the proof of concept has been confirmed in Saccharomyces cerevisiae [246].