A MANDA C OTTAGE 1 AND P ETER U RWIN 2 **
15.4. Genetic Engineering to Target the Nematode Directly
Direct targeting of a pathogen in order to kill it or impair its function sufficiently to prevent reproduction is a strategy commonly used to engineer resistance to fungi, bacteria and insect pests. Plant-parasitic nematodes present a variety of targets for such an approach, including the exoskeleton, the feeding tube, digestion, essential metabolic processes and reproduction; some of the methods used to attack these tar- gets are discussed in the following sections.
15.4.1. Bacillus thuringiensis (Bt) crystal (Cry) proteins
Bacillus thuringiensis (Bt) crystal (Cry) proteins are best known for their insecticidal properties. The Bt toxin causes the death of insect larvae by binding to receptors in
the epithelial cells of the larval gastrointestinal tract, which leads to pore formation and cell lysis (reviewed in Koziel et al., 1993). Insect resistance conferred by the Bt gene has been used to target cotton bollworm, maize borers and potato beetles. Crops, predominantly maize and cotton, genetically engineered to express the Bt toxin now cover over around 66 million ha worldwide (James, 2011). The nematicidal potential of the Bt toxin was first demonstrated in Caenorhabditis elegans, where exposure resulted in reduced fecundity and viability (Marroquin et al., 2000). However, uptake of the Bt Cry protein as a possible control mechanism for plant-parasitic nematodes was not initially investigated as H. schachtii was shown to be unable to ingest proteins larger than 23 kDa (Bửckenhoff and Grundler, 1994; Urwin et al., 1997a) and Bt Cry (Cry6A) is a 54 kDa protein. However, this ingestion limit does not apply to all other plant-parasitic nematodes and expression of Bt Cry (Cry6A) in tomato hairy root culture was shown to intoxicate M. incognita as demonstrated by a fourfold reduction in progeny (Li et al., 2007). By contrast, Cry5B expressing roots supported signifi- cantly reduced numbers of galls. This was reflected in a reduced total egg production but there were no significant differences in the number of eggs per egg mass between transgenic and control lines (Li et al., 2008). Therefore, Cry5B appears to exert its strongest effect on juvenile stages, whilst reproduction is most sensitive to Cry6A. It is likely that root-knot nematodes are able to ingest larger molecules than cyst nema- todes as their feeding tubes differ significantly (Sobczak et al., 1999). Thus, the use of Bt as a nematicide has limitations due to the range of plant-parasitic nematodes affected but additionally as the resistance may have limited durability; the widespread and intensive use of Bt has resulted in the emergence of resistant pests in some crops in Australia, India and China and resistance occurs frequently under artificial selec- tion pressures in several species, including C. elegans (Barrows et al., 2007).
15.4.2. Plantibodies
Plants can be engineered to produce functional antibodies, or antibody fragments known as plantibodies. The potential use of plantibodies in inactivating pathogen- derived biologically active molecules is self-evident and they have been used to control both viral and bacterial pathogens (reviewed in Schillberg et al., 2001).
They have also been investigated as a means of controlling plant-parasitic nema- todes. Some species of nematodes alter plant cell cycle and cell differentiation to form specialized feeding cells, and the factors that redirect the fate of the plant cells originate from three nematode pharyngeal glands (the sub-ventral glands and the dorsal gland; see Chapter 9). Therefore, inactivation of the active components in these secretions would result in the nematode failing to initiate a feeding cell.
This was first assessed in tobacco with the expression of the heavy and light chains of a murine monoclonal antibody specific to stylet secretions of M. incognita (Baumet al., 1996). Unfortunately, despite the antibodies binding to the nematode pharyngeal glands and stylet secretions, no reduction in the ability of the nematode to parasitize tobacco roots was seen. Although it is not entirely clear why the plantibodies failed to disrupt giant cell formation by inhibiting stylet secretion activity, one feasible explanation is that the antibody and antigen may not have coincided spatially, as the plantibodies accumulated in the endoplasmic reticulum (ER) and apoplastically, while nematode stylet secretions were delivered to the
cytoplasm. It is also possible that the antigen targeted by the antibody was not essen- tial to feeding cell initiation or that the antibody failed to inactivate its function.
Plantibodies have also been used in an attempt to inhibit early nematode coloniza- tion by targeting penetration and migration. Plantibodies were generated that recognize secretory–excretory proteins on the cuticle surface and amphids of G. pallida and G. rostochiensis. The plantibodies were shown to affect nematode movement and resulted in delayed root penetration. However, these effects were temporary as turno- ver of the secreted proteins resulted in loss of the bound antibody (Sharon et al., 2002). To date, plantibodies are yet to be a proven nematode control strategy.
Additionally they are likely to be of limited use as antibodies raised against one spe- cies are unlikely to cross-react with another.
15.4.3. Lectins
Lectins are sugar binding proteins that bind specific monosaccharides or oligosac- charides; they are found naturally in plants, animals and fungi (reviewed in Lam and Ng, 2011). Lectins can be categorized according to their structure (merolectins, hol- olectins, chimerolectins and superlectins), their carbohydrate specificity or by family (legume lectins, type II ribosome-inactivating proteins, monocot mannose-binding lectins, and other lectins). Lectins have been ascribed various functions, including symbiotic recognition, seed storage, growth regulation, plant development and defence against pathogens. Lectins are thought to play a major role in plant defence as they are produced in response to various pathogens, including nematodes (Jammes et al., 2005; Fuller et al., 2007). Lectins are of particular use against insect pests as many plant lectins bind glycans, such as chitin, which are rarely found in plants, and lectins have been used to engineer control of insect pests in wheat, rice, tobacco and potatoes (Powell et al., 1995). The potential of lectins as nematode control agents was first demonstrated by a reduction in galling of 75% in M. incognita infected tomato roots treated with concanavalin A lectin (Con A) (Marban-Mendoza et al., 1987). Using an engineered approach, constitutive expression of snowdrop (Galanthus nivalis) lectin (GNA) in potato and oilseed rape achieved partial resistance to G. pal- lida and to the migratory nematode Pratylenchus bolivianus; although results were varied, some lines showed 80 and 100% resistance, respectively (Burrows et al., 1997, 1998). Partial resistance to M. incognita was also achieved in A. thaliana expressing GNA with 50% less galling in comparison to controls (Ripoll et al., 2003).
The mechanism of pathogen resistance is not known but it is thought that lectin may bind glycoproteins localized on the surface of the nematodes, on the chemoreceptors in the amphid sensory organs themselves, or in the amphidial secretions, thereby interfering with nematode sensory perception and its ability to establish feeding cells.
Although this has yet to be proven in plant-parasitic nematodes, ConA has been found to be associated with the anterior amphids of the nematode Strongyloides ratti (a mammalian parasite) resulting in disruption of its chemokinetic response (Tobata- Kudo et al., 2005). Lectins are considered a leading substance in engineering phy- topathogen resistance due to their broad-spectrum activity, but unfortunately unfavourable publicity about the use of lectins was generated when adverse effects were reported in rats fed on raw potato containing snowdrop lectin (Ewen and Pusztai, 1999), and even though the report was later discredited by The Royal Society
of London (http://royalsociety.org/uploadedFiles/Royal_Society_Content/policy/pub- lications/1999/10092.pdf), the use of lectins still generates some concern. Additionally, lectin expression levels and degree of pathogen resistance can be highly variable and hyper-susceptibility to nematode infection in some lines (Ripoll et al., 2003) and with some species (Kaplan and Davis, 1991) of nematode has been observed.
15.4.4. Protease inhibitors
Transgenic expression of proteinase inhibitors (PIs) in plant roots is the most widely explored approach to engineered resistance to plant-parasitic nematodes. A range of different inhibitors, most of them naturally-occurring plant proteins, have been shown to be detrimental to feeding nematodes, reducing their growth and fecundity.
Inhibitors of all four main classes of proteinase (serine, cysteine, aspartic and metallo- proteinase) occur in plants and are often induced in response to wounding or her- bivory. Correspondingly, proteinase genes and activity have been identified in plant-parasitic nematodes (Lilley et al., 1996, 1997; Urwin et al., 1997b; Neveu et al., 2003; Fragoso et al., 2009). A digestive role has been proposed for these enzymes, corroborated for some by expression in the intestine. With digestion of protein being a common requirement of nematodes, PI-based control could have efficacy against a wide range of species, irrespective of their parasitic strategy. This would have particu- lar utility in those field situations where a number of different nematode pests occur concurrently.
Cysteine proteinase inhibitors, termed cystatins, have received the most attention.
Initial experiments utilized the rice cystatin Oc-I, modifying its coding region to remove an amino acid and improve its inhibitory activity 13-fold over the native protein. Expression of this engineered variant (Oc-IDD86) in tomato hairy roots using the cauliflower mosaic virus (CaMV35S) promoter resulted in significantly smaller female G. pallida after 6 weeks when compared to control roots (Urwin et al., 1995). Expression of Oc-IDD86 in a second model system A. thaliana, using the same promoter, allowed the cystatin to be tested against additional nematode species. The size of female H. schachtii and M. incognita was considerably reduced relative to controls with growth arrested before egg laying. This effect was correlated with detection of the cystatin in the feeding nematodes and reduced cysteine proteinase activity in the intestine of female H. schachtii recovered from plants (Urwin et al., 1997b). The same A. thaliana plants also suppressed growth and egg production of the reniform nematode, Rotylenchulus reniformis, with cystatin expression level influencing reproductive success (Urwin et al., 2000). Although A. thaliana is not a favoured host for this nematode, the study is an example of a model system providing preliminary data to support later cystatin expression in crops of interest such as pine- apple (Wang et al., 2009b), where transformation is limited by a slow rate of regen- eration. An alternative model host plant, lucerne (alfalfa), was used to demonstrate that the rice cystatins Oc-I and Oc-II expressed at a low level in lucerne under the control of a wound-inducible promoter conferred some resistance to the root-lesion nematodePratylenchus penetrans (Samac and Smigocki, 2003).
A rather different approach was taken to inhibit cysteine proteinases of H. glycines parasitizing transgenic soybean hairy roots (Marra et al., 2009). The propeptides that are cleaved from cysteine proteinase precursors can often act as inhibitors of their
cognate enzymes (e.g. Silva et al., 2004). The propeptide region of the H. glycines HGCP-I cathepsin L enzyme was expressed in roots and caused a reduction in the number and fecundity of female nematodes (Marra et al., 2009). The prodomain inhibitor displays greater specificity for target enzymes than do typical plant PIs (Silvaet al., 2004) and whilst this may limit the utility of the approach to control a wide range of nematode species, it could have biosafety advantages for non-target organisms.
Although less widely studied, serine proteinase inhibitors have also demonstrated potential for nematode control. In another model system study, transgenic expression of the sweet potato serine PI, sporamin, inhibited growth and development of female H. schachtii parasitizing sugar beet hairy roots (Cai et al., 2003). In this case, the severity of the effect was clearly correlated with the level of trypsin-inhibitory activity detected in the transformed root lines.
Engineered resistance based on PIs has been extensively tested in potato, primar- ily against G. pallida. The potential of plant PIs as anti-nematode effectors was first explored using the serine PI cowpea trypsin inhibitor (CpTI). CpTI expressed in transgenic potato influenced the sexual fate of newly established G. pallida (Hepher and Atkinson, 1992) and as a result the population was biased toward a predomi- nance of the much smaller and less damaging males. Subsequent work focused on cystatins and culminated in successful field trials of transgenic potatoes. The best transgenic line of the fully susceptible potato cv Désirée, expressing chicken egg white cystatin from the constitutive CaMV35S promoter, displayed 70% resistance to potato cyst nematodes in the field (Urwin et al., 2001). When the same construct was used to transform two potato cultivars, Sante and Maria Huanca, that each display natural partial resistance to G. pallida, the best transgenic lines of each were enhanced to full resistance (Urwin et al., 2003). Subsequent field trials demonstrated that both the modified rice cystatin (OcIDD86) and a sunflower cystatin expressed in cv Désirée afforded similar levels of protection to chicken egg white cystatin (Urwin et al., 2003). Potato plants in which expression of the OcIDD86 cystatin was limited mainly to the roots and, in particular, to the syncytia and giant cells induced by G. pallida andM. incognita, respectively, were shown to have similar resistance levels to those achieved with constitutive expression for both nematodes (Lilley et al., 2004).
Proteinase inhibitor strategies are being tested in banana and plantain. Cavendish dessert bananas that express the OcIDD86 engineered variant of rice cystatin under the control of the maize ubiquitin promoter displayed 70 ± 10% resistance to Radopholus similis in a glasshouse trial (Atkinson et al., 2004). Plants expressing the same cystatin under the control of a root-specific promoter that is upregulated in giant cells (Green et al., 2002; Lilley et al., 2004) were resistant (83 ± 4%) to M. incognita (H.J. Atkinson, personal communication). The approach is now progressing to cook- ing varieties of Musa. East African Highland banana plants constitutively expressing a maize cystatin support reduced multiplication of R. similis and the plantain cv Gonja has been transformed to express both a cystatin and a repellent peptide (Roderick et al., 2012). Similar additive cystatin plus repellent constructs have been introduced into East African Highland banana varieties (NARO, Uganda). There could be an additional advantage to cystatin-mediated nematode resistance in banana as cystatin impairs feeding and development of banana weevils (Kiggundu et al., 2010).
To date, the only nematode resistance technology introduced into rice is the cys- tatin-based defence. Transgenic plants of four elite African rice varieties constitutively
expressing the modified rice cystatin OcIDD86 displayed 55% resistance to M. incog- nita (Vain et al., 1998). Only a low level of cystatin expression was observed, possibly due to a suboptimal CaMV35S promoter or homology-dependent silencing of the transgene in combination with the endogenous OcI gene. In subsequent work, a maize cystatin has been expressed in the rice variety Nipponbare under the control of a root promoter from A. thaliana (TUB-1) that is known to be upregulated in the feeding cells of M. incognita parasitizing rice (Green et al., 2002).
Inhibitory activity of a potato serine proteinase inhibitor (PIN2) expressed in transgenic wheat showed a positive correlation with plant growth and yield following infestation with the cereal cyst nematode H. avenae (Vishnudasan et al., 2005).
A protective effect on the plant against nematode infection was inferred; however, the effect of the PI on nematode development was not investigated. In a further demon- stration of the potential of PIs, a cystatin from the tropical root crop taro (Colocasia esculenta) was expressed constitutively in a root-knot nematode-susceptible tomato cultivar. There was a 50% reduction in the number of galls formed by M. incognita on the transgenic plants compared with wild-type plants and a larger reduction in the number of egg masses produced per plant (Chan et al., 2010).
15.4.4.1. Biosafety of protease inhibitors
Plant protease inhibitors are attractive as effectors against nematodes because, while they have a negative effect on the parasite, they present no harm to humans and live- stock and are present in a normal diet. Considerable effort has been put into deter- mining the effect on the environment of their use in agriculture and studies have determined no effect on other insect feeders, their predators or soil microorganisms (Cowgillet al., 2002a,b, 2004; Cowgill and Atkinson, 2003).
15.4.5. RNA interference (RNAi)
RNA interference (RNAi) is the process in which double-stranded RNA (dsRNA) trig- gers the silencing of specific target genes through mRNA degradation. It has been adopted as a tool for functional analysis of plant-parasitic nematode genes (Rosso et al., 2009), and delivery of dsRNA from a host plant to bring about RNAi silencing of genes in the feeding nematode is being explored as a nematode resistance strategy (Lilleyet al., 2011). To date, the approach has shown potential against both cyst and root-knot nematodes and screening methods have been developed to allow the evalu- ation of many gene targets. Arabidopsis thaliana plants expressing dsRNA from hair- pin inverted repeat constructs at least partially reduced transcript abundance of targeted parasitism genes in the pharyngeal gland cells of feeding H. schachtii (Patel et al., 2008, 2010; Sindhu et al., 2009). For six of eight genes tested this led to a signifi- cant reduction in female numbers of between 23% and 64%, with considerable varia- tion between lines for some constructs (Sindhu et al., 2009). Variable, non-significant effects were also observed when a fibrilin gene of H. glycines was targeted from chi- meric soybean plants (Li et al., 2010a). In the same study, RNAi of a coatomer subu- nit of this nematode resulted in a significant reduction in egg production. Soybean composite plants derived from hairy root cultures engineered to silence either of two
ribosomal proteins, a spliceosomal protein or synaptobrevin, of H. glycines by RNAi resulted in 81–93% fewer females developing on the transgenic roots (Klink et al., 2009), whilst a similarly high reduction in egg production was achieved by targeting mRNA splicing factor prp-17 or an uncharacterized gene cpn-1 (Li et al., 2010b).
A high level of resistance to root-knot nematode was achieved by targeting a para- sitism gene expressed in the sub-ventral gland cells of M. incognita (Huang et al., 2006).
dsRNA complementary to the 16D10 gene was expressed in transgenic A. thaliana and the resulting lines displayed a significant reduction (63–90%) in the number of galls and their size, with a corresponding reduction in total egg production. The high level of homology between the 16D10 sequences of different Meloidogyne species led to broad- range resistance against M. incognita, M. javanica, M. arenaria and M. hapla. Almost complete resistance to Meloidogyne infection was reported in tobacco plants expressing dsRNA corresponding to splicing factor or integrase (Yadav et al., 2006) and of four genes targeted from transgenic soybean roots, two (encoding a tyrosine phosphatase and mitochondrial stress-70 protein precursor) reduced gall number by >90% (Ibrahim et al., 2011). However, not all host-delivered RNAi targeting of Meloidogyne genes results in a resistance phenotype. Partial silencing of a putative transcription factor of M. javanica (MjTis11) did not significantly affect either nematode development or fecundity (Fairbairn et al., 2007). Partial resistance resulted from targeting either a dual oxidase gene with a probable role in cuticle formation or a subunit of signal peptidase, a protein complex required for the processing of secreted proteins. Crossing transgenic lines expressing these two defences provided higher levels of resistance to M. incognita than either parent plant (Charlton et al., 2010). Such additive resistance may raise the efficacy and durability of RNAi-based defences but a combinatorial RNAi targeted at H. glycines did not deliver that benefit (Bakhetia et al., 2008). Possibly short interfering RNAs may sometimes compete for the RNA-induced silencing complex as occurs in mammals. One strength of the RNAi approach is that ongoing (Lilley et al., 2010) and completed genome sequencing for plant-parasitic nematodes (Abad et al., 2008;
Oppermanet al., 2008) provide a large range of potential targets that can be screened in vitro to select those for plant transformation constructs.
Cyst and root-knot nematodes are the two most important economic nematode groups but some crops such as banana are also severely damaged by migratory endo- parasites including R. similis, which causes rotting of banana roots. It is susceptible to RNAi, although the extent of silencing can vary according to the region of the nema- tode gene targeted and the experimental occasion. Its infection of Medicago truncatula was reduced by 60% after soaking it in dsRNA homologous to a nematode gland cell xylanase gene (Haegeman et al., 2009). It remains to be seen whether or not the efficacy of host-generated RNAi will work efficiently against all species of nematodes in the field.