Eukaryotic translation initiation factor 4E‐mediated recessive resistance to plant viruses and its utility in crop improvement

SUMMARY

The use of genetic resistance is considered to be the most effective and sustainable approach to the control of plant pathogens. Although most of the known natural resistance genes are monogenic dominant R genes that are predominant against fungi and bacteria, more and more recessive resistance genes against viruses have been cloned in the last decade. Interestingly, of the 14 natural recessive resistance genes against plant viruses that have been cloned from diverse plant species thus far, 12 encode the eukaryotic translation initiation factor 4E (eIF4E) or its isoform eIF(iso)4E. This review is intended to summarize the current state of knowledge about eIF4E and the possible mechanisms underlying its essential role in virus infection, and to discuss recent progress and the potential of eIF4E as a target gene in the development of genetic resistance to viruses for crop improvement.

INTRODUCTION

Plant pathogens are a major constraint to agriculture and threaten global food security (Dodds, 2010). The use of genetic resistance is considered to be the most effective and sustainable strategy to control plant pathogens in agricultural practice, as it is environmentally friendly and target specific, and provides reliable protection against pathogens without additional labour or material costs (Kang et al., 2005a; Maule et al., 2007). On the basis of inheritance, resistance genes that confer qualitative traits can be classified into two types: dominant and recessive. The majority of dominant R genes confer resistance to bacteria and fungi. R gene‐mediated resistance, often associated with extreme resistance (ER) and a hypersensitive response (HR), is triggered by either direct or indirect interactions between the R gene‐encoded protein of the host and the avirulence factor produced by the corresponding avirulence (Avr) gene of the invading pathogen (Bonas and Lahaye, 2002; Soosaar et al., 2005). Interestingly, published data suggest that recessive resistance is more frequently found for plant viruses than for other plant pathogens (Truniger and Aranda, 2009).

Viruses have small genomes that encode a very limited number of proteins. For instance, the genome of potyviruses, the largest group of plant viruses that includes many agriculturally important viruses, is a single‐stranded, positive‐sense RNA molecule of approximately 10 kb encoding a single polyprotein and a truncated frameshift peptide (Chung et al., 2008; Shukla et al., 1994). Therefore, viruses must depend on host factors to complete their infection cycle. Loss or mutation of an essential host factor required by the virus may induce recessive resistance to this virus (Diaz‐Pendon et al., 2004). To date, all recessive genes involved in plant–virus interactions that have been characterized encode eukaryotic translation initiation factors (eIFs), mainly eIF4E, eIF4G and their isoforms. These include at least 14 natural recessive resistance genes, e.g. mo1 in lettuce (Lactuca sativa); nsv in melon (Cucumis melo); pot‐1 in tomato (Solanum lycopersicum); pvr1, pvr2 and pvr6 in pepper (Capsicum annuum and C. chinense); rym4, rym5 and rym6 in barley (Hordeum vulgare); sbm1, wlv and cyv2 in pea (Pisum sativum); and tsv1 and rymv‐1 in rice (Oryza sativa). Except for tsv1 and rymv‐1, which encode a mutated form of eIF4G or a defective form of eIF(iso)4G, and confer resistance to Rice tungro spherical virus (RTSV, Waikavirus) and Rice yellow mottle virus (RYMV, Sobemovirus) (Albar et al., 2006; Lee et al., 2010), all others encode either eIF4E or eIF(iso)4E. Given the unique status of eIF4E, both as a key regulator of cellular translation and a controller of resistance/susceptibility to plant viruses, a better understanding of the molecular mechanisms underlying eIF4E‐mediated resistance may assist in the development of novel strategies against plant pathogens. In the past several years, excellent reviews have been devoted to plant R and/or recessive resistance genes, including those encoding eIFs (Kang et al., 2005a; Le Gall et al., 2011; Maule et al., 2007; Robaglia and Caranta, 2006; Truniger and Aranda, 2009). In this review, we focus on eIF4E and eIF4E‐mediated resistance, and its deployment through advanced biotechnology for application in agriculture.

BIOLOGICAL FUNCTIONS OF eIF4E

eIF4E is a cap‐binding protein that plays an essential role in the initiation of cap‐dependent mRNA translation

eIF4E was originally named the ‘cap‐binding protein’ as it interacts specifically with the 5′‐terminal cap of mRNA (Sonenberg et al., 1978). Now, it is best known for its essential function in the initiation of mRNA translation, a rate‐limiting step of protein synthesis (see the recent review by Jackson et al., 2010). In eukaryotes, mRNA translation is predominantly cap dependent, with an exception for a small portion of mRNAs whose translation is mediated by internal ribosome entry sites (IRESs). Cap‐dependent translation initiation is a multistep process requiring the assembly of an mRNA–protein complex by different eIFs. The first step is the binding of eIF4E to the 5′‐7mGpppN‐cap of mRNAs. Subsequently, eIF4E recruits eIF4G (the scaffold protein) and eIF4A (the DEAD‐box RNA helicase), leading to the formation of the eIF4F protein complex (Goodfellow and Roberts, 2008). Translation initiates on recruitment of the eIF4F complex to the small (40S) ribosomal subunit carrying eIF3 and the ternary initiation tRNA‐eIF2‐GTP complex through the interaction of eIF3 with eIF4G. Then, 40S starts to scan the mRNA in the 5′ to 3′ direction to the initiation codon. The scaffold protein eIF4G interacts with the 3′‐bound poly(A)‐binding protein (PABP), bringing the mRNA ends into proximity for efficient translation. The formation of the complete translation initiation complex requires the eIF5B‐bridged coupling of the 40S with the 60S ribosomal subunit (Kawaguchi and Bailey‐Serres, 2002).

eIF4E is also a nuclear protein and is involved in the nuclear export of mRNA

In addition to its essential function in protein synthesis in the cytoplasm, approximately 68% of all eIF4E is present in the nucleus of mammalian cells (Iborra et al., 2001). In plants, the distribution of eIF4E seems to be dynamic. The Arabidopsis eIF4E has been shown to accumulate in the nucleus in quiescent cells, but is preferentially present in the cytoplasm in proliferating cells (Bush et al., 2009). As a nuclear protein, eIF4E is involved in the nuclear export of a subset of mRNAs containing eIF4E‐sensitivity elements. eIF4E is targeted to the nucleus via interactions with the 4E‐transporter (4E‐T), an eIF4E‐binding protein (Dostie et al., 2000). In the nucleus, a number of homeodomain proteins can bind to eIF4E to regulate its cap binding and mRNA export capacity (see the review by Culjkovic et al., 2007).

Plants have a second form of eIF4E, termed eIF(iso)4E, and both forms have similar activities, but may have distinct physiological functions

In contrast with other eukaryotes, plants have a second form of eIF4F, named eIF(iso)4F (Bush et al., 2009). The eIF(iso)4F complex is composed of the corresponding isoforms of eIFs, including eIF(iso)4E and eIF(iso)4G. eIF(iso)4F and eIF4F complexes have complementing activities, but their respective components are differentially expressed, suggesting divergent roles during development or in response to different environmental (biotic or abiotic) cues (Dinkova et al., 2011; Gallie et al., 1998; Rodriguez et al., 1998;). Functional redundancy has been observed between eIF4E and eIF(iso)4E isoforms under regular growth conditions (Combe et al., 2005; Duprat et al., 2002; Lellis et al., 2002). For example, Arabidopsis thaliana has three eIF4E genes, i.e. eIF4E1 (At4g18040, commonly designated eIF4E), eIF4E2 (At1g29690) and eIF4E3 (At1g29550), and one eIF(iso)4E gene (At5g35620) (Robaglia and Caranta, 2006). In contrast with eIF4E, which is differentially subcellularly distributed dependent on the cell type, eIF(iso)4E localizes to both the cytoplasm and nucleus in either quiescent or proliferating cells (Bush et al., 2009). In Arabidopsis, no phenotypic or fertility difference is evident among eIF4E or eIF(iso)4E mutants and wild‐type plants (Duprat et al., 2002; Lellis et al., 2002; Sato et al., 2005), demonstrating that eIF4E or eIF(iso)4E is dispensable for plant growth and development by being complemented by the other form. In Arabidopsis eIF(iso)4E knockout (KO) mutants, an increase in the concentration of the eIF4E protein was observed, suggesting the existence of a cellular mechanism that regulates eIF4E expression to compensate for the loss of eIF(iso)4E (Duprat et al., 2002). Antisense depletion of either eIF4E or two copies of eIF(iso)4E in tobacco plants does not alter the phenotype of the plant, whereas antisense depletion of both eIF4E and eIF(iso)4E results in a semi‐dwarf phenotype and reduces polysome loading, demonstrating their complementary and additive effect in plant growth (Combe et al., 2005).

DISCOVERY OF eIF4E AS A HOST FACTOR IN VIRUS INFECTION

Potyviruses encode a genome‐linked viral protein, VPg, which is covalently linked to the 5′ end of the genomic RNA (Murphy et al., 1996). This viral protein is absolutely required for virus infection; host proteins physically associating with VPg are likely to play an essential role in virus infection. Wittmann et al. (1997) pioneered the research on the identification of VPg‐interacting host proteins by conducting a yeast two‐hybrid screen. The Turnip mosaic virus (TuMV) VPg was used as a bait to screen a cDNA expression library derived from A. thaliana. eIF(iso)4E was identified as an interacting partner of the viral protein (Wittmann et al., 1997). Point mutagenesis of the VPg domain that binds to eIF(iso)4E suggested that the VPg–eIF(iso)4E interaction is essential to preserve virus infectivity in planta (Léonard et al., 2000). Lellis et al. (2002) screened a mutant population of A. thaliana for decreased susceptibility to TuMV and found three loss‐of‐susceptibility mutants. Map‐based cloning identified eIF(iso)4E as the gene responsible for the susceptibility change in all three mutants. Inspired by these studies, Ruffel et al. (2002) used a candidate gene approach to clone the recessive resistance gene pvr2 in pepper, which confers resistance to Potato virus Y (PVY). Restriction fragment length polymorphism (RFLP) analysis on the PVY‐susceptible and PVY‐resistant pepper cultivars showed a perfect map co‐segregation between a polymorphism in the eIF4E gene and the pvr2 alleles (Ruffel et al., 2002). Susceptibility of the pvr2‐genotype cultivar was restored by transient expression of eIF4E from a susceptible cultivar, providing conclusive evidence that eIF4E is a host factor required for PVY infection and that pvr2 encodes an eIF4E variant that confers recessive resistance to PVY (Ruffel et al., 2002). The pvr2 gene became the first natural recessive resistance gene to be cloned in plants.

NATURAL RECESSIVE RESISTANCE GENES ENCODE eIF4E OR eIF(iso)4E

Cloned natural recessive genes encoding eIF4E

There are two amino acid substitutions in the pvr2‐encoded eIF4E (Ruffel et al., 2002). Later, it was found that the pvr2 alleles are actually alleles at the pvr1 locus which also encode eIF4E variants (Kang et al., 2005b). The pvr2 or pvr1 alleles also confer resistance to other potyviruses, including Pepper veinal mottle virus (PVMV) and Tobacco etch virus (TEV) (Charron et al., 2008; Kang et al., 2005b; Ruffel et al., 2006). Recently, five new pvr2 variants conferring PVY resistance have been identified in Capsicum species (Ibiza et al., 2010). Shortly after the cloning of pvr2, the lettuce recessive gene mo1, which confers resistance to Lettuce mosaic virus (LMV), was identified using the same candidate gene approach (Nicaise et al., 2003). It encodes an eIF4E mutant with amino acid variations specifically located near the predicted mRNA cap recognition pocket.

An orthologue of pvr2 in tomato, pot‐1, was isolated using comparative genomic mapping (Ruffel et al., 2005). Complementation assays confirmed that small amino acid mutations in the pot‐1‐encoded eIF4E protein account for resistance to PVY and TEV. Almost at the same time, two other groups reported independently that recessive resistance genes rym4 and rym5 in barley are alleles of eIF4E (Kanyuka et al., 2005; Stein et al., 2005). A few amino acid substitutions in barley eIF4E can alter the outcome of infections by Barley yellow mosaic virus (BaYMV) and Barley mild mosaic virus (BaMMV) from susceptibility to resistance (Kanyuka et al., 2005; Stein et al., 2005). Three‐dimensional (3D) modelling of the barley eIF4E suggests that all the polymorphic residues are located near the mRNA cap‐binding pocket (Kanyuka et al., 2005). This group also determined that rym6 is an eIF4E mutant (Kanyuka et al., 2005).

In pea, sbm1‐mediated recessive resistance to Pea seed‐borne mosaic virus (PSbMV) is also linked to an eIF4E allele (2004a, 2004b). Based on the predicted 3D structure, mutations in sbm1‐encoded eIF4E potentially disrupt both translational and trafficking functions (Gao et al., 2004b). In two other independent reports, resistance genes wlv and cyv2 in pea, conferring resistance to Bean yellow mosaic virus (BYMV) and Clover yellow vein virus (ClYVV), respectively, were shown to correspond to the sbm1 allele of eIF4E (Andrade et al., 2009; Bruun‐Rasmussen et al., 2007). In common bean (Phaseolus vulgaris), another legume species, recessive resistance to Bean common mosaic virus (BCMV), controlled by the recessive gene bc‐3, co‐segregates with an eIF4E variant (Naderpour et al., 2010). This variant contains a set of mutations closely resembling a pattern of eIF4E mutations determining potyvirus resistance described above. Therefore, bc‐3 is very probably an allele of eIF4E.

In melon, the recessive resistance to Melon necrotic spot virus (MNSV), a member of the genus Carmovirus, family Tombusviridae, is conferred by nsv (Nieto et al., 2006). A combination of positional cloning and microsynteny analysis concluded that nsv is an eIF4E variant in melon. Further complementary experiments suggested that a single amino acid change at position 228 of the protein is responsible for MNSV resistance (Nieto et al., 2006).

Cloned natural recessive genes encoding eIF4(iso)E

Comparative genetic mapping demonstrated that recessive gene pvr6 in pepper corresponds to the eIF(iso)4E locus (Ruffel et al., 2006). In comparison with the wild‐type eIF(iso)4E transcript, pvr6 contains a deletion of 82 nucleotides (Ruffel et al., 2006). The deletion starts from nucleotide 89 of the eIF(iso)4E cDNA and modifies the open reading frame by a frame shift to introduce a stop codon after amino acid 51 (Ruffel et al., 2006). Conditional on the presence of pvr1 or pvr2, pvr6 confers resistance to PVMV and Chilli veinal mottle virus (ChiVMV) (Hwang et al., 2009; Rubio et al., 2009; Ruffel et al., 2006). Consistently, simultaneous mutations in both eIF4E and eIF(iso)4E are required for ChiVMV and PVMV resistance in pepper (Hwang et al., 2009; Rubio et al., 2009; Ruffel et al., 2006). A genetic study has been carried out to examine whether other eIF4E alleles, in combination with pvr6, could confer resistance to PVMV. Among 12 eIF4E alleles cloned and tested, three alleles have a complementary effect with eIF(iso)4E for resistance (Rubio et al., 2009).

MOLECULAR MECHANISMS UNDERLYING eIF4E/eIF(iso)4E‐DEPENDENT INFECTION

Recruitment of eIF4E/eIF(iso)4E as a host factor in virus infection

On the basis of the molecular identification of these natural recessive genes, eIF4E and eIF(iso)4E act as host factors for a number of viruses, including both VPg‐containing and VPg‐free viruses, with the former as a major group. It is worth mentioning that human calicivirus also has a VPg which interacts with eIF4E, and such an interaction is required for the viral genome translation in vitro (Goodfellow et al., 2005). In the case of VPg‐free viruses, in addition to MNSV (Tombusviridae) (Nieto et al., 2006), Cucumber mosaic virus (CMV; Bromoviridae) also requires eIF4E for infection (Yoshii et al., 2004). The accumulation of CMV is strongly reduced in an Arabidopsis eIF4E1‐defective mutant (Yoshii et al., 2004). Unlike potyviruses, the genomic RNAs of these two viruses are capped at their 5′ termini and do not have a poly(A) tail at the 3′ ends. For such viruses, translation initiation is conferred at the 5′ proximal AUG by direct base pairing of a stem loop in the 3′ untranslated region (UTR) with the corresponding structure in the 5′ UTR. The recruitment of eIF4E by MNSV or CMV probably occurs through the 3′ UTR (Nieto et al., 2006; Yoshii et al., 2004).

For the cloned recessive resistance genes to VPg‐containing viruses, which are exclusive in the family Potyviridae, the requirement of eIF4E in virus infection is highly virus specific, and a given virus may prefer one form of eIF4E to the other, or use both for compatible infections (Nicaise et al., 2007). For example, Arabidopsis eIF(iso)4E‐defective mutants are resistant to TuMV, LMV, TEV and Plum pox virus (PPV), whereas mutations of eIF4E1 in Arabidopsis inhibit ClYVV infection (Decroocq et al., 2006; Duprat et al., 2002; Lellis et al., 2002; Sato et al., 2005). Arabidopsis eIF(iso)4E mutants resistant to TuMV and LMV are susceptible to ClYVV, but the eIF4E mutants resistant to ClYVV have no resistance to TuMV (Sato et al., 2005). The preference for a specific form of eIF4E to the other by viruses can also vary with the host. LMV depends on eIF4E to infect lettuce, but switches to eIF(iso)4E in Arabidopsis (Duprat et al., 2002; Nicaise et al., 2003; Ruffel et al., 2005). Both eIF4E and eIF(iso)4E are involved in ChiVMV and PVMV resistance in pepper (Hwang et al., 2009; Rubio et al., 2009; Ruffel et al., 2006). TuMV can use both eIF4E and eIF(iso)4E of Brassica rapa for replication in the Arabidopsis eIF(iso)4E KO mutant (Jenner et al., 2010). Taken together, these findings suggest that, although eIF4E and eIF(iso)4E can compensate for their function in plant growth and development, such a functional redundancy does not necessarily extend to virus infection (Robaglia and Caranta, 2006).

Apparently, the recruitment of eIF4E or eIF(iso)4E by VPg‐containing viruses is through its direct interaction with VPg. The physical interaction of VPg and eIF4E/eIF(iso)4E correlates with compatible infections in many potyvirus–plant pairs (Charron et al., 2008; Gao et al., 2004b; Grzela et al., 2006; Kang et al., 2005b; Yeam et al., 2007). For instance, in the pepper pvr1–TEV, pvr1–PVY and pvr2–PVY pathosystems, a few amino acid mutations within the pvr1‐ or pvr2‐encoded eIF4E abolish or impair VPg–eIF4 interactions and inhibit virus infection (Charron et al., 2008; Yeam et al., 2007). Interestingly, among the eIF4E variants encoded by recessive genes from seven different crop species, most of the amino acid substitutions are located in two regions near the cap recognition pocket (Charron et al., 2008; Robaglia and Caranta, 2006; Truniger and Aranda, 2009) and probably mediate interactions of eIF4E with VPg. Consistently, variations in the VPg domain may counter‐compromise the resistance conferred by recessive genes encoding defective forms of eIF4E/eIF(iso)4E. In most resistance‐breaking potyviral isolates that break down recessive genes, the virulence determinant is mapped to VPg (Borgstrom and Johansen, 2001; Bruun‐Rasmussen et al., 2007; Charron et al., 2008; Duprat et al., 2002; Gallois et al., 2010; Grzela et al., 2006; Moury et al., 2004). Most mutations in these VPgs are located in the central region that is involved in the interaction with eIF4E (Roudet‐Tavert et al., 2007). These data establish the status of the VPg and eIF4E/eIF(iso)4E interaction as a key determinant of virulence or avirulence in many, but not all, potyvirus infections.

Indeed, the pea sbm1 gene encoding a defective form of eIF4E confers resistance to PSbMV (2004a, 2004b). No detectable interaction was found between the PSbMV VPg and eIF4E from a susceptible pea genotype (Gao et al., 2004b). More recently, Gallois et al. (2010) performed a systematic yeast two‐hybrid assay to examine interactions between all the Arabidopsis eIF4Es and virulent VPgs responsible for overcoming resistance in eIF(iso)4E KO Arabidopsis mutants. No significant interactions were found between eIF4Es and virulent/avirulent VPgs, other than between eIF(iso)4E and virulent/avirulent VPgs. In these cases, how the eIF4F complex is recruited for viral genome translation has yet to be investigated (Gallois et al., 2010).

Proposed roles of eIF4E/eIF(iso)4E in virus infection

A favourable model is that VPg functions as an analogue of the cap structure of mRNA to recruit the translation complex for viral genome translation (Beauchemin et al., 2007; Khan et al., 2008; Michon et al., 2006). Thus, the VPg and eIF4E/eIF(iso)4E interaction is pivotal to virus infection. This model is supported by several observations (Fig. 1A). First, viral VPg competes and binds to its preferential form of eIF4E more rapidly than does the m7G cap (Khan et al., 2006; Miyoshi et al., 2006). The VPg‐binding site is located within or near the cap‐binding region (Ashby et al., 2011; Miyoshi et al., 2006). Second, the presence of eIF4G/eIF(iso)4G, another component of the eIF4F complex, enhances VPg–eIF4E/eIF(iso)4E interactions (Khan et al., 2006). eIF4G is also an essential host factor for potyvirus infections (Albar et al., 2006; Decroocq et al., 2006; Nicaise et al., 2007). The recruitment of the form of eIF4G and eIF4E by potyviruses is coordinated (Nicaise et al., 2007). Third, viral replicase proteins, eIF4E/eIF(iso)4E and other eIFs, such as eEF1A and PABP, are present in the virus translation/replication complex (Beauchemin et al., 2007; Thivierge et al., 2008; Wei et al., 2010a). A limitation of this model is that it cannot explain why mutations in the eIF4G‐binding area do not affect the ability of eIF4E to restore LMV susceptibility in mol¹ plants (German‐Retana et al., 2008a). If VPg is a cap substitute, VPg should promote translation of VPg‐linked RNAs and inhibit translation of the capped mRNAs, given that eIF4E is the rate‐limiting factor of translation in the cell (Jackson et al., 2010). In wheat germ extracts, the presence of VPgs indeed inhibits cap‐dependent translation, but stimulates translation of uncapped IRES‐containing RNAs (including viral RNA) (Khan et al., 2008). The 5′ leader of the TEV genomic RNA confers cap‐independent translation in an eIF4G‐dependent manner (Gallie, 2001). In agreement with this finding, Potato virus A (PVA) VPg represses the expression of the capped mRNA and upregulates the expression of both replicating and nonreplicating PVA RNAs (Eskelin et al., 2011). Enhancement of viral RNA expression by VPg functions through the 5′ UTR. Therefore, it cannot be excluded that potyviral genome translation initiation may occur at the 5′ UTR, independent of the VPg–eIF4E interaction.

Figure 1.

Proposed roles of eukaryotic translation initiation factor 4E (eIF4E) during potyvirus infection. (A) eIF4E binds to the genome‐linked viral protein (VPg) and recruits the translation initiation apparatus for viral genome translation. (B) eIF4E, cylindrical inclusion (CI) protein and eIF4G may form a complex that binds to VPg to mediate intracellular trafficking of the viral genome for targeting to plasmodesmata for cell‐to‐cell movement and, further, for systemic infection. (C) The VPg–eIF4E complex may be involved in the suppression of eIF4E‐mediated transport of mRNA from the nucleus to the cytoplasm for translation and in the disturbance of siRNA and microRNA processing in the nucleus. eIF4E, P1, VPg and HC‐Pro (helper component–proteinase) may form a complex that functions as an RNA‐silencing suppressor to safeguard virus translation/replication in the cytoplasm. 4A, 4E, 4G, 2, 3 and 5 represent eIF4A, eIF4E, eIF4G, eIF2, eIF3 and eIF5, respectively. P1, first protein; PABP, poly(A)‐binding protein.

In addition to VPg, the potyviral cyclindrical inclusion (CI) protein acts as a virulence determinant responsible for overcoming eIF4E‐mediated recessive resistance (Abdul‐Razzak et al., 2009). The CI protein, having RNA‐binding, RNA helicase and ATPase activities, has been shown to be essential in virus intra‐ and intercellular movement and virus replication (Carrington et al., 1998; Kekarainen et al., 2002; Wei et al., 2010b). In lettuce, the LMV‐0 isolate is restricted for systemic infection in mol¹‐genotype lettuce, but systemically infects the near‐isogenic variety without the resistance gene (2003, 2008b). The LMV‐E isolate breaks down mol¹‐ or mol²‐mediated resistance. Replacement of the VPg of LMV‐0 with that of LMV‐E confers the recombinant virus ability to overcome mol¹‐mediated resistance, whereas the CI protein of the isolate LMV‐E allows both eIF4E alleles mol¹ and mol² to be overcome (Abdul‐Razzak et al., 2009). The LMV CI protein interacts with the VPg and the lettuce eIF4E (2007, 2012). Such interactions are stronger for LMV‐E than for LMV‐0 (Roudet‐Tavert et al., 2012). Therefore, it is possible that the eIF4E‐mediated resistance may be associated with the impairment of virus intracellular trafficking, cell‐to‐cell movement or long distance movement via its interaction with VPg, CI protein and eIF4G (Fig. 1B). This proposal is supported by previous findings that PVY and PSbMV cell‐to‐cell movement is inhibited in pvr2‐genotype pepper and sbm1‐genotype pea, respectively (Arroyo et al., 1996; Gao et al., 2004b).

In view of the nuclear localization of both eIF4E and VPg, it is tempting to speculate a possible role of VPg in the suppression of eIF4E‐mediated mRNA transport from the nucleus to cytoplasm for translation (Fig. 1C). This is consistent with the observation that VPg inhibits translation of the capped mRNAs (Eskelin et al., 2011; Khan et al., 2008). It is also possible that the eIF4E–VPg complex is involved in RNA silencing, a counter‐defensive mechanism that enables systemic infection. The potyviral protein HC‐Pro (helper component–proteinase) is the major RNA‐silencing suppressor with the first protein (P1) and VPg as an accessory factor (Kasschau and Carrington, 1998; Rajamäki and Valkonen, 2009). VPg is a nuclear‐ and nucleolar‐targeting protein; mutations that reduce nuclear and nucleolar localization compromise its RNA‐silencing ability, possibly through disturbance of siRNA and microRNA processing in the nucleus, and prevent or diminish virus replication (Rajamäki and Valkonen, 2009). eIF4E may be involved in this process via its interaction with VPg (Fig. 1C).

P1 exhibits nonspecific RNA‐binding activity and enhances HC‐Pro‐mediated suppression of RNA silencing (Brantley and Hunt, 1993; Kasschau and Carrington, 1998; Rajamäki et al., 2005). In the pea–ClYVV pathosystem, a point mutation in P1 is responsible for breaking down the resistance conferred by cyv2 (an eIF4E allele) (Nakahara et al., 2010). VPg, P1, HC‐Pro and eIF4E are interacting partners (Ala‐Poikela et al., 2011; Merits et al., 1999; Nakahara et al., 2010; Roudet‐Tavert et al., 2007). P1 localizes in the cytoplasm (Nakahara et al., 2010). In infected cells, HC‐Pro is present in proximity to the eIF4E‐ and VPg‐containing virus replication complex (Ala‐Poikela et al., 2011). It would be interesting to determine whether P1 physically associates with the replication complex in virus‐infected cells and forms, with VPg, HC‐Pro and eIF4E, a functional complex to suppress RNA silencing and safeguard virus replication in the cytoplasm (Fig. 1C).

GENERATION OF RECESSIVE RESISTANCE THROUGH BIOTECHNOLOGY

Although the molecular mechanisms underlying recessive gene‐mediated resistance to viral pathogens and susceptibility to resistance‐breaking isolates are not fully understood, many recessive genes, such as the pepper pvr1, have been used successfully for several decades. On the basis of the aforementioned discussion, genetic resistance may be developed by silencing a host factor, such as a specific eIF4E or eIF4G member essential for viral infection. The resulting ‘recessive resistance’ is actually controlled by a dominant transgene. For instance, eIF(iso)4E is required for PPV infection in plum, and transgenic plum plants in which eIF(iso)4E is silenced by the expression of a short hairpin RNA specifically targeting eIF(iso)4E are highly resistant to PPV (X. Wang et al., unpublished, Agriculture and Agri‐Food Canada, London, Ontario). Genetic resistance can also be generated by the overexpression of recessive genes identified for the target virus (Cavatorta et al., 2011; Kang et al., 2007; Yan et al., 2010; Yeam et al., 2007). One example is that ectopic expression of the recessive Capsicum gene pvr1, an eIF4E allele, in tomato results in dominant broad‐spectrum resistance to TEV and Pepper mottle virus (PepMoV) (Kang et al., 2007).

The advent of new technologies, such as Targeting‐Induced Local Lesions IN Genome (TILLING), high‐resolution melting (HRM), KeyPoint and next‐generation sequencing, has opened up a new era for the development of genetic resistance in crops (Hofinger et al., 2009; Piron et al., 2010; Rigola et al., 2009). These technologies allow the identification of target gene mutants from an artificially induced mutation population (Fig. 2). For instance, TILLING was employed to screen for eIF4E or eIF4G mutants in an ethyl methanesulphonate (EMS)‐induced tomato mutant population (Piron et al., 2010). A splicing mutant Sl‐eIF4E1 (G1485A) was identified to be immune to PVY and PepMoV. The mutated gene Sl‐eIF4E1 (G1485A) encodes a truncated protein that is impaired in cap‐binding activity (Piron et al., 2010). A TILLING‐based method to identify natural nucleotide diversity, termed EcoTILLING, has also been employed successfully to identify allelic variants of eIF4E against MNSV in melon (Nieto et al., 2007) and against PVY in Capsicum species (Ibiza et al., 2010). This approach may be particularly effective for heterozygous species in which recessive alleles may exist, but cannot be screened out by the conventional phenotypic resistance assay.

Figure 2.

A flowchart showing the process of generation of recessive‐based resistance in crops to a target virus via artificially induced mutagenesis and subsequent screening using new technologies, i.e. TILLING (Targeting‐Induced Local Lesions IN Genome) and next‐generation sequencing. EMS, ethyl methanesulphonate.

CONCLUDING REMARKS

In the past 15 years, significant progress has been made in the understanding of genetic resistance mediated by recessive genes, particularly eIF4E and eIF4G. This makes it possible to manipulate them for the control of relevant viruses. However, there still remain difficulties for the use of recessive resistance in some crop species, such as those strictly clonally propagated crops, highly allogamous species and perennial plants with a long life cycle. Moreover, genetic diversity resulting from error‐prone replication and RNA recombination in the existing and continuously evolving viral population, and the selection force acting on this variability, lead to the occurrence of resistance‐breaking isolates, posing a serious challenge to the deployment of eIF4E‐ and eIF4G‐mediated resistance. It is worth noting that the recessive genes that have been cloned so far only account for a very small portion of the known recessive genes (Kang et al., 2005a; Truniger and Aranda, 2009). As suggested by Le Gall et al. (2011), the predominant overrepresentation of eIF4E/eIF(iso)4E in the cloned recessive genes may be a result of the fact that the first cloned recessive gene against a plant virus was eIF4E, which prompted the isolation of other recessive genes using a candidate gene approach. The cloning and characterization of additional recessive genes against diverse viruses will certainly help to gain a better understanding of recessive resistance and to more efficiently deploy recessive resistance into breeding programmes against plant viruses as well as other pathogens.