Soaking of adult worms

In C. elegans, RNAi can be effectively induced by injection of dsRNA into any site of the worm (Fire et al. Reference Fire, Xu, Montgomery, Kostas, Driver and Mello1998) by feeding on bacteria expressing dsRNA (Timmons et al. Reference Timmons, Court and Fire2001) and by soaking in dsRNA solution (Tabara et al. Reference Tabara, Grishok and Mello1998). Knockdown of target mRNA is depends on several proteins that enable uptake, spreading and processing of exogenous dsRNA. Spreading can be largely ascribed to the transmembrane proteins SID-1 (Systemic RNA Interference Defective) and SID-2. SID-1 forms channels in the membrane to enable dsRNA diffusion (Feinberg and Hunter, Reference Feinberg and Hunter2003) whilst SID-2 is mainly located in the intestine and is involved in the uptake of exogenous dsRNA into the cells (Winston et al. Reference Winston, Sutherlin, Wright, Feinberg and Hunter2007). Together, these proteins facilitate systemic spread of the dsRNA in C. elegans, as demonstrated by effective silencing achieved simply by soaking worms in media containing dsRNA. Indeed, this mode of delivery has been used in a large-scale survey of C. elegans gene function (e.g. Maeda et al. Reference Maeda, Kohara, Yamamoto and Sugimoto2001).

Given this background, initial experiments were performed in Nippostrongylus brasiliensis by soaking adult parasites with dsRNA for AChE B, a secreted acetylcholinesterase, and this treatment resulted in knockdown of both this and two other closely related variants of the enzyme (Hussein et al. Reference Hussein, Kichenin and Selkirk2002). Parasite motility was unaffected, suggesting that neuromuscular AChEs were not significantly affected, and this was also suggested by visualisation of enzyme activity in a polyacrylamide gel-based assay, although this is not quantitative (Hussein et al. Reference Hussein, Kichenin and Selkirk2002). Soaking adult worms with dsRNA for another variant, AChE A, also resulted in suppression of secreted enzyme expression, but we failed to observe knockdown by RT-PCR of six other targets by this method, homologues of ubiquitin, paramyosin (unc-15 in C. elegans), calponin (unc-87), protein disulphide isomerase 3 (PDI-3), phospholipase A2 (PLA2) and β-tubulin (unpublished data). The latter was somewhat surprising in that this gene appears to be particularly susceptible to knockdown in other parasitic nematodes following dsRNA soaking (Aboobaker and Blaxter, Reference Aboobaker and Blaxter2003; Geldhof et al. Reference Geldhof, Murray, Couthier, Gilleard, McLauchlan, Knox and Britton2006; Kotze and Bagnall, Reference Kotze and Bagnall2006; Samarasinghe et al. Reference Samarasinghe, Knox and Britton2011).

As expression of secreted cholinesterases could be suppressed in adult N. brasiliensis worms maintained in vitro, the effect of this treatment on establishment of parasites in the host intestinal tract was examined. Parasites which had been cultured for 24 h in dsRNA for AChE B were used to reinfect rats by oral gavage, and the animals sacrificed 6 days later. A significant reduction in recovery of parasites with suppressed AChE was observed in two independent experiments, but this was matched by two which showed no difference and by others in which no parasites were recovered from control groups (unpublished data). Interpretation of these data was hampered by the very low recovery of adult worms which presumably resulted from extended maintenance in vitro. In addition, in the absence of sequence information for neuromuscular AChEs, we could not rule out effects on these enzymes which, by analogy with C. elegans, would be required for normal movement and presumably the ability to maintain position in the jejunum (Culotti et al. Reference Culotti, Von Ehrenstein, Culotti and Russell1981; Selkirk et al. Reference Selkirk, Lazari, Hussein and Matthews2005).

As a major objective of RNAi is to examine the role of gene products on parasitism (i.e. development in the definitive host), application to infective larval stages is generally more useful. This approach was recently taken to investigate in vivo effects of gene silencing in H. contortus (Samarasinghe et al. Reference Samarasinghe, Knox and Britton2011) and Ascaris suum (Chen et al. 2011). Pre-soaking of infective L3 larvae of H. contortus in dsRNA for an H11 aminopeptidase gene resulted in significant reductions in worm burden, egg output and aminopeptidase activity in recovered worms (Samarasinghe et al. Reference Samarasinghe, Knox and Britton2011). Similarly, pre-soaking of infective larvae of A. suum for 72 h in dsRNA for an enolase gene resulted in shorter worms compared to those derived from control larvae following infection in mice (Chen et al. Reference Chen, Xu, Nisbet, Huang, Lin, Yuan, Song and Zhu2011). These studies indicate that essential gene function can be identified by in vivo RNAi, but by using current methods only those genes susceptible to reliable knockdown can be monitored in this way.

Ingestion of bacterially expressed dsRNA

Feeding C. elegans on bacteria engineered to express dsRNA has been used very effectively for RNAi (Timmons et al. Reference Timmons, Court and Fire2001) and is the method of choice for genome-wide functional screens (e.g. Kamath et al. Reference Kamath, Fraser, Dong, Poulin, Durbin, Gotta, Kanapin, Le Bot, Moreno, Sohrmann, Welchman, Zipperlen and Ahringer2003). Lawns of Escherichia coli HT115, deficient in RNaseIII, support development of N. brasiliensis from newly hatched first stage larvae (L1) through to infective third stage larvae (L3), a process which takes 7 days at 25°C (i.e. similar to incubation in faecal cultures), and these L3 s are fully infectious, completing their life cycle in rats. Both L1 and L2 feed on E. coli, demonstrable by the use of bacteria expressing Green Fluorescent Protein (Fig. 1). Suppression of a number of target mRNAs has been tested in this system by feeding worms with E. coli HT115 expressing dsRNA for N. brasiliensis paramyosin, calponin, protein disulphide isomerase-3 and phospholipase A2. The first two targets were selected as suppression of expression might be predicted to result in alterations in motility (Fraser et al. Reference Fraser, Kamath, Zipperlen, Martinez-Campos, Sohrmann and Ahringer2000; www.wormbase.org) but this was not observed with any consistency, and further investigation indicated there was no reduction in transcript levels for any of these targets. This concurs with the failure of feeding bacteria expressing dsRNA to H. contortus (Geldhof et al. Reference Geldhof, Murray, Couthier, Gilleard, McLauchlan, Knox and Britton2006) and Heligmosomoides polygyrus (Lendner et al. Reference Lendner, Doligalska, Lucius and Hartmann2008) to elicit RNAi. The only species of animal-parasitic nematode in which this method has been documented to have a phenotypic effect is Trichostrongylus colubriformis (Issa et al. Reference Issa, Grant, Stasiuk and Shoemaker2005). In the case of H. polygyrus, it was noted that L1 ingest many fewer bacteria than the corresponding stage of C. elegans (Lendner et al. Reference Lendner, Doligalska, Lucius and Hartmann2008). In an attempt to increase the liberation of dsRNA from bacteria in the parasite intestine, we transfected bacteria with pLysE, a plasmid expressing bacteriophage T7 lysozyme, but this proved to be toxic and inhibited larval development (unpublished data).

Fig. 1.

Nippostrongylus brasiliensis larvae ingest bacteria. Both L1 and L2 (shown) feed on E. coli. Panel A, bright field; Panel B, fluorescence. Parasites do not develop beyond the L1 stage on NGM plates in the absence of bacteria.

Activation of feeding in infective larvae

Infective larvae are developmentally arrested in strongylid nematodes and do not feed, presenting a problem in terms of delivery of dsRNA or siRNA. Chemical exsheathment of N. brasiliensis L3 did not stimulate feeding, but maintenance at elevated temperature (37°C) proved to be sufficient (Huang et al. Reference Huang, Chan, Smyth, Ball, Gounaris and Selkirk2010). Infective larvae begin feeding after a lag period of 3 to 4 h at 37°C, reaching a maximum of 90% of the population feeding by 48 h (Huang et al. Reference Huang, Chan, Smyth, Ball, Gounaris and Selkirk2010). A variety of regimens of soaking temperature-activated larvae with dsRNA or siRNA for up to 72 h have been tested, followed by semi-quantitative RT-PCR for up to 48 h post-treatment. Whilst we could inconsistently detect knockdown of AChE B transcripts, ubiquitin and β-tubulin proved refractory to suppression. Lipofectin, lipofectamine and polyethyleneimine (PEI), the latter a cationic transfection reagent which condenses nucleic acids into positively charged particles, have all been tested without success (Huang, Reference Huang2010).

Resorcinol (Rosso et al. Reference Rosso, Dubrana, Cimbolini, Jaubert and Abad2005) and octopamine (Urwin et al. Reference Urwin, Lilley and Atkinson2002) stimulate uptake of dsRNA by plant-parasitic nematodes. Resorcinol is toxic to N. brasiliensis larvae at similar concentrations used by Rosso et al. (Reference Rosso, Dubrana, Cimbolini, Jaubert and Abad2005) and blocks uptake of fluorescein isothiocyanate-conjugated bovine serum albumin (FITC-BSA). However, serotonin, known to stimulate pharyngeal pumping in C. elegans (Avery and Horvitz, Reference Avery and Horvitz1990) and Ascaris suum (Brownlee et al. Reference Brownlee, Holden-Dye, Fairweather and Walker1995), elicited a dose-dependent increase in uptake of BSA by activated infective larvae up to a maximal concentration of 5 mM (Huang et al. Reference Huang, Chan, Smyth, Ball, Gounaris and Selkirk2010). Nevertheless, inclusion of serotonin in further experiments in which worms were soaked in dsRNA still failed to elicit suppression of ubiquitin transcripts (Fig. 2).

Fig. 2.

Serotonin enhances the rate of feeding in activated L3, but does not enhance RNAi. A. Larvae were activated for 47 h at 37°C, then incubated in the absence (Ctrl) or presence of the indicated concentrations of 5-HT for 1 h before quantitative assessment of feeding by uptake of ¹²⁵I-BSA (Huang et al. Reference Huang, Chan, Smyth, Ball, Gounaris and Selkirk2010). Data are presented as the mean±s.e. of assays in triplicate. Values for uptake which differ significantly from controls (no 5-HT) are indicated (**P<0·005; ***P<0·0001). B. Activated larvae were exposed to 1 mg ml⁻¹N. brasiliensis ubiquitin dsRNA with or without 0·5 mM serotonin (5-HT) for 1 hr before electroporation. Parasite RNA was isolated 48 h after treatment and levels of mRNA for ubiquitin and actin determined on 10x serially diluted cDNA.

Electroporation

Electroporation has been used in a variety of species of parasitic nematodes in RNAi procedures with variable success (Geldhof et al. Reference Geldhof, Murray, Couthier, Gilleard, McLauchlan, Knox and Britton2006; Issa et al. Reference Issa, Grant, Stasiuk and Shoemaker2005; Visser et al. Reference Visser, Geldhof, de Maere, Knox, Vercruysse and Claerebout2006 ; Lendner et al. Reference Lendner, Doligalska, Lucius and Hartmann2008; Chen et al. Reference Chen, Xu, Nisbet, Huang, Lin, Yuan, Song and Zhu2011). We have attempted to use electroporation to introduce dsRNAs and siRNAs into larval and adult N. brasiliensis but, as with soaking experiments, this only effected partial and inconsistent knockdown of AChE expression, with no effect on three other target mRNAs. Luciferase activity has been used effectively as a marker to optimise conditions for electroporation of S. mansoni schistosomula with mRNA for firefly luciferase (Correnti and Pearce, Reference Correnti and Pearce2004). These conditions were used successfully to suppress expression of cathepsin B (Correnti et al. Reference Correnti, Brindley and Pearce2005), and this methodology was further confirmed by others to be more effective than soaking with transfection reagents in this species (Krautz-Peterson et al. Reference Krautz-Peterson, Radwanska, Ndegwa, Shoemaker and Skelly2007). Attempts to introduce mRNA for firefly luciferase into activated L3 or adult N. brasiliensis via a square wave electroporator were unsuccessful however, despite using a range of conditions and efficient recovery of luciferase activity from electroporated mammalian cells (Fig. 3 and Huang, Reference Huang2010). This is perhaps not surprising, as the cuticle would present an impermeable barrier to electroporation, although cuticular pores appear to exist in some nematode species (Martin et al. Reference Martin, Foster, Kester and Donahue1987; Tilney et al. Reference Tilney, Connelly, Guild, Vranich and Artis2005). In the absence of cuticular pores, electroporation would only be possible via an inside-out process into cells which are connected to the exterior, such as those present in intestinal epithelia, secretory or excretory glands or sensory organelles such as amphids, and moreover this process would require electrical conductivity to the outside medium. Electroporation of H. polygyrus using a wide range of conditions also failed to effect RNAi, although the authors did observe focal points of fluorescence between the cuticle and hypodermis when using Cy3-labelled siRNA as a probe (Lendner et al. Reference Lendner, Doligalska, Lucius and Hartmann2008). Electroporation of N. brasiliensis with labelled dsRNA, siRNA or bovine serum albumin (BSA) did not result in internalisation of fluorescence unless preceded by an extended period of incubation, in which case macromolecules accumulated in the pharynx and intestine in a manner indistinguishable from soaking without electroporation (Fig. 3). We thus concluded that, in N. brasiliensis at least, electroporation does not provide a viable means of delivery of RNA.

Fig. 3.

Electroporation does not enhance uptake of dsRNA. A: Adult N. brasiliensis were incubated for 2 h with FITC-labelled bovine serum albumin (A and C) or TRITC-labelled β-tubulin dsRNA (B and D). Parasites in panels C and D were subsequently electroporated. Both dsRNA and BSA accumulated in the pharynx and intestine in a manner indistinguishable from soaking without electroporation. E: Electroporation of N. brasiliensis adult and larval parasites with luciferase mRNA does not result in detectable expression. A range of conditions were used (Huang, Reference Huang2010) and parasites were assayed at 3 h (shown here) and 24 h post-electroporation. T cells (293 cell line) were used as positive controls and assayed 3 h post-electroporation. Results are presented as the mean +s.e. of assays performed in triplicate. F: Electroporation of adult parasites with siRNA or dsRNA for ubiquitin did not result in RNAi. Levels of mRNA for parasite ubiquitin and actin were determined as for Fig. 2 on 10x serially diluted cDNA.

Alternative methods of delivery

Particle bombardment is an alternative approach for delivering nucleic acids into worms (Davis et al. Reference Davis, Parra, LoVerde, Ribeiro, Glorioso and Hodgson1999; Wilm et al. Reference Wilm, Demel, Koop, Schnabel and Schnabel1999; Higazi et al. Reference Higazi, Merriweather, Shu, Davis and Unnasch2002; Jackstadt et al. Reference Jackstadt, Wilm, Zahner and Hobom1999), but this approach is generally harmful to N. brasiliensis (Ball, unpublished data). Microinjection with constructs designed to produce an endogenous hairpin dsRNA has been routinely used for RNAi in C. elegans (Tavernarakis et al. Reference Tavernarakis, Wang, Dorovkov, Ryazanov and Driscoll2000; Johnson et al. Reference Johnson, Behm and Trowell2005). However, microinjection-mediated introduction of RNA was shown to damage H. contortus larvae (Geldhof et al. Reference Geldhof, Murray, Couthier, Gilleard, McLauchlan, Knox and Britton2006), probably due to high internal pressure. This approach is generally not suitable for most parasitic nematodes, as the inability to complete the life cycle in vitro means that it would be very difficult to select transformants, and survival of microinjection procedures tends to be very low. Microinjection of dsRNA has, however, been used in free-living adult female Strongyloides stercoralis, resulting in specific knockdown of mRNA in F1 progeny (Lok, Reference Lok2007). The same effect was not achieved following dsRNA soaking of adult worms, suggesting that delivery of dsRNA directly into nematodes may overcome apparent gene insensitivity to RNAi.

Nanoparticles are an emerging technology with potential application for interfering RNA delivery (Britton et al. 2011). They are very small in size but provide a large surface area for attachment of RNA and current data, in other organisms, suggest their use can enhance dsRNA uptake. Of particular note, C. elegans showed effective uptake of fluorescent silica or polystyrene based nanoparticles, this being evident in the intestine and pharynx inititially. This first phase was followed by translocation to the gonad and even early embryos (Plustoka et al. Reference Pluskota, Horzowski, Bossinger and von Mikecz2009). Nanoparticles have been used with success for the delivery of dsRNA to mosquito larvae intestinal cells following feeding (Zhang et al. Reference Zhang, Zhang and Zhu2010).

RNAi pathway genes

The efficacy of environmental (soaking or bacterial feeding) RNAi in C. elegans seems to be an anomaly even among free-living nematodes. Differences in the SID-2 transporter are responsible for the lack of environmental RNAi in C. briggsae, a deficiency that can be overcome by transgenic expression of the C. elegans sid-2 gene (Winston et al. Reference Winston, Sutherlin, Wright, Feinberg and Hunter2007). Poor conservation of sid-2 and other RNAi pathway genes is therefore likely to influence RNAi success in parasitic species.

The molecular processes and components necessary for a functional RNAi pathway have been extensively investigated in C. elegans and summarized by Grishok (Reference Grishok2005). These are outlined briefly as follows (Geldhof et al. Reference Geldhof, Murray, Couthier, Gilleard, McLauchlan, Knox and Britton2006): the dsRNA is recognized by a protein complex containing a double-stranded RNA-binding protein, Rde-4, as well as the dsRNA-specific ribonuclease, Dicer, which cleaves dsRNA into small inhibitory RNAs (siRNAs). The DexH box helicase Drh-1 and Rde-1 argonaute protein are thought to transport the siRNAs to the RNA induced silencing complex (RISC). This interaction results in the activation of the silencing complex, which targets homologous mRNA transcripts by a base-pairing mechanism. Once bound, the target transcript is degraded by the RISC, resulting in gene knock-down. In C. elegans the silencing effect is systemic, spreading to most cells through the activity of the SID and RSD proteins, and siRNAs are subsequently recognized by the RISC protein complex. Thus, RNAi is induced by dsRNA, the RNAi effect can be systemic and, at least in C. elegans, is heritable (Grishok, Reference Grishok2005).

In a preliminary analysis to examine conservation of RNAi pathway genes in parasitic nematodes, we screened the available H. contortus genome database (at the time with approximately 6-fold coverage). Sequences of 17 proteins with defined functions in the RNAi pathway of C. elegans were used in BLAST searches against the H. contortus genome database (Geldhof et al. Reference Geldhof, Visser, Clark, Saunders, Britton, Gilleard, Berriman and Knox2007). Genes encoding potential homologues of dcr-1, drh-1, rde-1, rde-3, sid-1, eri-1, rrf-1, tsn-1, vig-1, rrf-1, ego-1, mut-7 and rsd-6 were readily identified, but other genes thought to be essential for a functional and systemic RNAi pathway were not found; these included rde-2, rde-4, sid-2 and rsd-2.

The absence of rde-4 was particularly noteworthy, given that the protein product binds double-stranded RNA, and is considered critical for complex formation with Dicer and generation of short interfering RNAs (e.g. Parrish and Fire, Reference Parrish and Fire2001). Standard RNAi is inoperative in C. elegans mutants lacking rde-4 (Tabara et al. Reference Tabara, Grishok and Mello1998). At the time, we tested the efficacy of RNAi against several C. elegans homologues of genes of interest to us in H. contortus (Clark, Geldhof and Knox, unpublished observations). All were susceptible to RNAi in wild-type C. elegans, but not in C. elegans mutants lacking rde-4. Rde-4 was also not obviously apparent in the Brugia malayi genome, nor in available S. mansoni gene datasets. The authors (Geldhof et al. Reference Geldhof, Visser, Clark, Saunders, Britton, Gilleard, Berriman and Knox2007) suggested that perhaps other mechanisms for RNAi activation are operative in helminths. Interestingly, a study in C. elegans using transgenic arrays to express dsRNA showed that when high concentrations of trigger dsRNA were produced within the nematode, the RNAi pathway was activated even in rde-4 mutants (Habig et al. Reference Habig, Aruscavage and Bass2008). This suggested that it is possible to overcome the absence of rde-4 by increasing the levels of dsRNA. It was proposed that this may reflect titration of dsRNA destabilising factors (e.g. adenosine deaminases; Habig et al. Reference Habig, Aruscavage and Bass2008) and/or more effective binding of Dicer to abundant dsRNA, without the need for RDE-4.

Recently, Dalzell et al. (Reference Dalzell, McVeigh, Warnock, Mitreva, Bird, Abad, Fleming, Day, Mousley, Marks and Maule2011) published the outcome of a study which focused on RNAi effector diversity in nematodes. Again, rde-2, rde-4, rsd-2 and sid-2 were not detected in the genome of Haemonchus nor of several other parasitic nematodes using genome datasets which had been considerably expanded since the initial analysis by Geldhof et al. (Reference Geldhof, Visser, Clark, Saunders, Britton, Gilleard, Berriman and Knox2007). The authors (Dalzell et al. Reference Dalzell, McVeigh, Warnock, Mitreva, Bird, Abad, Fleming, Day, Mousley, Marks and Maule2011) also noted that the Meloidogyne species of plant-parasitic nematodes lacked orthologues of SID-1, SID-2 and RSD-2, yet showed systemic RNAi effects after soaking in dsRNA, indicating that there may be alternative uptake proteins in use or even alternative mechanisms of transport. How RNAi works in parasites in which the currently defined RNAi machinery is clearly deficient merits attention, and may be key to achieving consistent RNAi effects for target genes. The absence of rde-2, sid-2 and rsd-2 homologues in the H. contortus genome would suggest that the RNAi effect would not spread through the worm and to the germline. The absence of rde-4, based on findings from C. elegans, suggests that the RNAi pathway would not be active unless high concentrations of dsRNA were present. In these scenarios, it may be anticipated that gene silencing would only be observed in regions of the worm directly accessible to delivered dsRNA and where the concentration of dsRNA was sufficiently high to activate the pathway. Recent findings from H. contortus RNAi (Samarasinghe et al. Reference Samarasinghe, Knox and Britton2011) are consistent with this hypothesis.