Evidence for different, host‐dependent functioning of Rx against both wild‐type and recombinant Pepino mosaic virus

Summary

The potato Rx gene provides resistance against Pepino mosaic virus (PepMV) in tomato; however, recent work has suggested that the resistance conferred may not be durable. Resistance breaking can probably be attributed to multiple mutations observed to accumulate in the capsid protein (CP) region of resistance‐breaking isolates, but this has not been confirmed through directed manipulation of an infectious PepMV clone. The present work describes the introduction of two specific mutations, A‐T78 and A‐T114, into the coat protein minimal elicitor region of an Rx‐controlled PepMV isolate of the EU genotype. Enzyme‐linked immunosorbent assay (ELISA) and phenotypic evaluation were conducted in three Rx‐expressing and wild‐type solanaceous hosts: Nicotiana benthamiana, Nicotiana tabacum and Solanum lycopersicum. Mutation A‐T78 alone was sufficient to confer Rx‐breaking activity in N. benthamiana and S. lycopersicum, whereas mutation A‐T114 was found to be associated, in most cases, with a secondary A‐D100 mutation to break Rx‐mediated resistance in S. lycopersicum. These results suggest that the need for a second, fitness‐restoring mutation may be dependent on the PepMV mutant under consideration. Both mutations conferred Rx breaking in S. lycopersicum, whereas neither conferred Rx breaking in N. tabacum and only A‐T78 allowed Rx breaking in N. benthamiana, suggesting that Rx may function in a different manner depending on the genetic background in which it is present.

The plant immune system is multilayered, consisting of both broad‐spectrum and specific lines of defence. Dominant resistance (R) genes constitute an important component of these specific defence mechanisms. The products of R genes recognize pathogen avirulence (Avr) molecules and trigger a highly effective resistance response in a race‐specific manner. Recognition and resistance depend on factors expressed from both the pathogen and the host, and are therefore described as a gene‐for‐gene interaction system (Flor, 1971). The triggered resistance commonly involves the induction of the hypersensitive response (HR) (Jones and Dangl, 2006), a form of programmed cell death resulting in necrosis at the site of infection, thereby preventing systemic viral spread.

The Rx gene from potato provides resistance to Potato virus X (PVX) in commercial potato accessions (Cockerham, 1970). It encodes a nucleotide‐binding site‐leucine‐rich repeat (NBS‐LRR)‐type protein with a coiled‐coil (CC) domain at the N‐terminus (CC‐NBS‐LRR) (Bendahmane et al., 1999). The C‐terminus of the LRR domain is thought to be involved in the specific recognition of the pathogen elicitor (Dangl and Jones, 2001; Farnham and Baulcombe, 2006). Co‐expression studies have demonstrated intramolecular interactions between the CC‐NBS and LRR domains to be integral in the functioning of the Rx protein. The presence of the pathogen elicitor disrupts these interactions, leading to Rx activation and defence signalling initiation (Moffett et al., 2002).

The resistance conferred by Rx is unusual in that it does not involve an HR. Viral replication has been reported to be halted in the initially infected cell and cannot therefore be detected at tissue level. For these reasons, the term ‘extreme resistance’ (ER) has been coined to describe it (Bendahmane et al., 1999; Tozzini et al., 1991). The PVX capsid protein (CP) is the sole elicitor of the Rx‐based resistance response (Bendahmane et al., 1995, 1999; Goulden et al., 1993). The resistance conferred is described as durable as only a single resistance‐breaking isolate is known: PVX_HB (Jones, 1985; Moreira et al., 1980). Mutational analysis has shown that the mutation of a conserved CP residue is sufficient to overcome Rx‐mediated resistance (Goulden et al., 1993). However, PVX isolates carrying this mutation are severely affected in their ability to mount a systemic infection in potato. A second, fitness‐restoring mutation in the CP is necessary for full infectivity in resistant potato varieties. This need for two mutations to gain a full resistance‐breaking phenotype probably accounts for the durability of Rx resistance (Goulden et al., 1993).

A number of R genes have been shown to retain their effectiveness when transgenically introduced into heterologous plant species (Baurès et al., 2008; Rommens et al., 1995; Song et al., 2003; Spassova et al., 2001; Whitham et al., 1996). Rx has been shown to be active against a range of potexviruses in transgenic Nicotiana spp., even when as little as 40% homology exists between the CPs of the viruses concerned (Baurès et al., 2008; Candresse et al., 2010). As a result of this unusually broad activity range, it has been proposed that Rx‐based recognition is dependent on conserved structural elements of the viral CP, rather than on a linear amino acid sequence (Baurès et al., 2008; Chapman et al., 1992; Goulden and Baulcombe, 1993). Transient expression of CP fragments from PVX, White clover mosaic virus (WClMV) and Narcissus mosaic virus (NMV) has allowed the identification of a 90‐amino‐acid ‘minimal elicitor’ region required for Rx‐based recognition (Baurès et al., 2008).

Pepino mosaic virus (PepMV), an emergent potexvirus that presents a major threat to tomato production, also possesses the ability to infect a number of other solanaceous crops. Despite control efforts, tomato‐infecting isolates of PepMV have gained a worldwide distribution in just over 10 years (Hanssen et al., 2010). A high level of conservation is displayed between all sequenced PepMV isolates in the Rx minimal elicitor region of the CP (Candresse et al., 2010). Rx has been shown to be active against PepMV, providing initial hopes that it may provide a valuable source of resistance in susceptible crop species. However, recent evidence has indicated that Rx‐based resistance against PepMV in tomato may not be durable. Candresse et al. (2010) passaged PepMV through Rx‐expressing tomato and reported the frequent selection of resistance‐breaking isolates. Sequence analysis of the CP of these variants showed the accumulation of a number of point mutations in the Rx minimal elicitor region that were proposed to affect Rx‐mediated recognition (Candresse et al., 2010). However, the precise impact of the observed mutations could not be confirmed in the absence of an infectious clone (IC) in which to introduce the suspected point mutations.

Using our recently described ICs of an EU genotype of PepMV (Duff‐Farrier et al., 2015; GenBank accession; KJ018164), we report the analysis of the impact of two of these candidate point mutations (A‐T78, A‐T114) in the Rx minimal elicitor region. The Rx‐breaking activity of the resulting PepMV mutants was investigated in three different transgenic Rx‐expressing hosts: Nicotiana benthamiana, Nicotiana tabacum and Solanum lycopersicum.

The desired mutations (A‐T78 and A‐T114) were introduced into the CP region of a wild‐type PepMV EU IC, constructed as described previously and contained within a pYES2 vector, pYES2_EU (Duff‐Farrier et al., 2015). pYES2_EU was used as the template in Phusion PCR (Thermo Scientific, Wilmington, DE, USA) to amplify the PepMV CP region (primers 1 + 2, Table 1), which was cloned into pJET1.2 (Thermo Scientific), forming pJET1.2_CP. This was entered into two site‐directed mutagenesis reactions using a GeneArt® Site‐Directed Mutagenesis System (Invitrogen, Carlsbad, CA), according to the manufacturer's protocol; mutations A‐T78 and A‐T114 used primer sets 3 + 4 and 5 + 6, respectively (Table 1). The mutagenized CP regions were excised and entered into a yeast recombination reaction with digested pYES2_EU and a linearized pYES2 backbone, following the protocol of Gietz and Woods (2002). Restriction digestion and sequencing confirmed successful construct generation for both pYES2_EU_A‐T78 and pYES2_EU_A‐T114.

Table 1.

Primers and annealing temperatures used in this investigation

Primer number	Sequence (5′–3′)	Annealing temperature (°C)
1	CAATCAACTTCTCCCCTTGGAACGG	58
2	CTCACTATAGGGAATATTAAGCTTGGTACCAATTGGTACCACGCGTTTTTTTTTTTTTTTTTTTTTTTTTT	58
3	AATGAGACTGGTCCGACCATGTGGGATCTAG	57
4	CTAGATCCCACATGGTCGGACCAGTCTCATT	57
5	ATCACGCCGAGCCCTTGCTACTCAGTTTGATCGAATCAAT	57
6	ATTGATTCGATCAAACTGAGTAGCAAGGGCTCGGCGTGAT	57

A Riboprobe® SP6 System (Promega, Madison, WI, USA), in conjunction with a Ribo m7G Cap Analog (Promega), was used to generate infectious transcripts in vitro from wild‐type, A‐T78 and A‐T114 KpnI linearized templates (Foster and Turner, 1998; Turner et al., 1994, 1999). Each reaction was immediately inoculated onto the surface of two N. benthamiana plants at the three‐leaf stage. Plants were kept under glasshouse conditions (18 °C with a 16 h/8 h light/dark cycle). Enzyme‐linked immunosorbent assay (ELISA) at 21 days post‐inoculation (dpi) showed high absorbance values for all constructs, indicating successful establishment of infection in all instances. All ICs displayed systemic phenotypes akin to that of the mild EU IC, characterized by light mosaics (data not shown). For each of the ICs, the CP region was amplified by Phusion PCR (Thermo Scientific) (primers 1 + 2, Table 1) and directly sequenced. Retention of the desired mutations was confirmed in progeny of the two mutant ICs; however, secondary mutations in the CP region were also observed (Fig. 1). The progenies of the parental IC and of the A‐T78 mutant were found to contain an additional V‐A230 mutation, whereas the A‐T114 mutant progeny contained an additional E‐K236 mutation.

Figure 1.

Consensus sequences of the Pepino mosaic virus (PepMV) capsid protein (CP) regions obtained from systemic leaves of Nicotiana benthamiana after inoculation with in vitro‐generated RNA representing each infectious clone (IC): EU, A‐T78 and A‐T114. Grey crosses indicate intended mutations and white stars indicate secondary mutations.

Homogenates from the sequenced primary N. benthamiana infections were used to inoculate wild‐type and Rx‐expressing N. benthamiana, S. lycopersicum (cv. Microtom) and N. tabacum (cv. Samsun) (Bendahmane et al., 1999; Candresse et al., 2010). Triplicate plants of each host type were inoculated with sap representing each IC. Plants were grown as outlined previously. A phenotypic analysis and evaluation of systemic viral accumulation were carried out at 21 dpi. ELISA readily detected the wild‐type EU IC in all wild‐type hosts, indicating full capacity for systemic movement and accumulation (Fig. 2). The systemic infection phenotypes were characterized by light mosaics in N. benthamiana (Fig. 3A), but by asymptomatic infection in S. lycopersicum (Fig. 3B) and N. tabacum (data not shown). In contrast with the wild‐type plants, background ELISA values were observed and a general absence of symptoms on upper non‐inoculated leaves for the Rx hosts, indicating an Rx‐specific inhibition of viral systemic infection (Fig. 2B,C). However, necrotic local lesions were observed on the inoculated leaves of Rx‐expressing N. benthamiana (Fig. 3C, panel A). Necrosis around the site of inoculation was also observed in S. lycopersicum (Fig. 3D), but not in N. tabacum (data not shown).

Figure 2.

Double antibody sandwich‐enzyme‐linked immunosorbent assay (DAS‐ELISA) data displaying systemic viral titres at 21 days post‐inoculation (dpi) of Pepino mosaic virus (PepMV) infectious clones (ICs), A‐T78, A‐T114 and unmutated EU, in Rx‐expressing (Rx) and wild‐type (WT) solanaceous hosts: (A) Nicotiana benthamiana; (B) Solanum lycopersicum cv. Microtom; (C) Nictotiana tabacum cv. Samsun.

Figure 3.

(A) Representative phenotypes displayed in wild‐type and Rx‐expressing Nicotiana benthamiana by the wild‐type EU infectious clone (IC) and capsid protein (CP) point mutation ICs A‐T78 and A‐T114. All ICs presented mild phenotypes in the wild‐type hosts. Both EU and A‐T114 were asymptomatic in the Rx host, whereas A‐T78 presented vascular necrosis in the upper parts of the plant. Symptoms were viewed at 21 days post‐inoculation (dpi). (B) Representative phenotypes displayed in wild‐type and Rx‐expressing (Rx) Solanum lycopersicum by ICs EU, A‐T78 and A‐T114. All ICs displayed asymptomatic phenotypes in the wild‐type hosts. In the Rx‐expressing hosts, the EU IC caused basal necrosis around the site of inoculation, whereas both A‐T78 and A‐T114 displayed trailing necrosis over the plant surface. Symptoms were viewed at 21 dpi. (C) Responses observed in the inoculated leaves of Rx‐expressing N. benthamiana, challenged with: A, EU IC; B, mutant A‐T78; C, mutant A‐T114. Symptoms were viewed at 21 dpi. (D) Close up of basal necrosis around the site of infection in Rx‐expressing S. lycopersicum challenged with the wild‐type EU IC. Symptoms were viewed at 21 dpi. (E) Close up of vascular necrosis in Rx‐expressing N. benthamiana challenged with mutant A‐T78. Symptoms were viewed at 21 dpi.

RNA was extracted from systemically infected leaves of the wild‐type plants using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. The CP region was amplified as described above and cloned into pJET1.2 (Thermo Scientific). At least two clones were sequenced for each sample. As it is widely known that potexviral CP is the sole elicitor of the Rx‐based resistance response (Bendahmane et al., 1995, 1999; Goulden et al., 1993), only the CP region was analysed in this work. In all three hosts, the V‐A230 mutation previously observed in the inoculum source was retained in the sequenced progenies (Table 2).

Table 2.

Mutational composition of clones sequenced from systemic infections generated in both Rx‐expressing and wild‐type (WT) hosts from challenge with both WT and mutant infectious clones (ICs). Intended mutations (italics), secondary mutations and infection phenotypes are indicated

Host		Genotype/phenotype	EU WT	A‐T78	A‐T114
Nicotiana benthamiana	WT (inoculum source)	G	100% V‐A230	100% A‐T78, V‐A230	100% A‐T114, E‐K236
		P	Systemic light mosaics	Systemic light mosaics	Systemic light mosaics
	WT	G	100% V‐A230	100% A‐T78, V‐A230	100% A‐T114, V‐A230
		P	Systemic light mosaics	–	Systemic light mosaics
	Rx	G	–	66.6% A‐T78, V‐A230 33.3% D‐E3, V‐A230	–
		P	Local necrosis	Severe local necrosis Systemic vascular necrosis	Severe local necrosis
Nicotiana tabacum	WT	G	100% V‐A230	100% A‐T78, V‐A230	100% D‐E3, V‐A230
		P	Asymptomatic	Asymptomatic	Systemic light mosaics
	Rx	G	–	–	–
		P	–	–	–
Solanum lycopersicum	WT	G	100% V‐A230	100% A‐T78, V‐A230	100% A‐T114, V‐A230
		P	Asymptomatic	Asymptomatic	Systemic light mosaics
	Rx	G	–	100% A‐T78, V‐A230	75% A‐T114, A‐D100, V‐A230 25% A‐T114, V‐A230
		P	Local necrosis	Severe systemic trailing necrosis	Trailing systemic necrosis

Similar to the wild‐type EU IC parent, mutant A‐T78 was able to systemically infect all wild‐type host species, as indicated by ELISA values comparable with those observed in the wild‐type infections (Fig. 2). Again, light mosaics were observed in N. benthamiana (Fig. 3A), but asymptomatic infection in both S. lycopersicum (Fig. 3B) and N. tabacum (data not shown). High ELISA values were observed in the non‐inoculated tissues of all Rx‐expressing N. benthamiana and S. lycopersicum, indicating a breakdown of Rx resistance in these hosts (Fig. 2A,B). Infection phenotypes were characterized by vascular necrosis in the upper parts of the plant in N. benthamiana (Fig. 3A,E), and by trailing necrosis over the entire plant in S. lycopersicum, with the plant showing a very stunted phenotype (Fig. 3B). A phenotype of spreading necrosis was also observed in the inoculated leaves of N. benthamiana (Fig. 3C, panel B). In contrast with the situation in Rx‐expressing N. benthamiana and S. lycopersicum, no symptoms or systemic viral accumulation could be detected in the Rx‐expressing N. tabacum plants, indicating that mutant A‐T78 could not evade the action of Rx in this host.

RNA was extracted from systemically infected leaves of both wild‐type and Rx‐expressing plants where infection had established (one plant representing each infection event), and the CP regions of the viral progenies were sequenced as described above; the results are given in Table 2. The introduced A‐T78 mutation and the V‐A230 secondary mutation previously detected in the inoculum were retained in all progenies sequenced. However, in a third of progeny clones obtained from the systemic leaves of Rx‐expressing N. benthamiana, the A‐T78 mutation was lost and, instead, a D‐E3 mutation was observed.

Similar to the A‐T78 mutant, the A‐T114 mutant possessed full systemic accumulation capacity in all wild‐type hosts, indicated by positive ELISA values comparable with those of the parental isolate (Fig. 2). The infection phenotypes were also similar to those observed for the wild‐type EU IC (Fig. 3A,B). The Rx resistance‐breaking capability of this mutant was also found to differ between the three tested Rx‐expressing hosts. Systemic accumulation levels similar to those in the wild‐type host were only observed in S. lycopersicum, indicating Rx breaking in this host (Fig. 2B), and the plants showed a trailing necrosis phenotype (Fig. 3B). However, no symptoms in non‐inoculated tissues and no systemic accumulation could be observed for this mutant in N. benthamiana or N. tabacum Rx‐expressors (Fig. 2A,C). A local necrotic response was observed in the inoculated leaves of Rx‐expressing N. benthamiana (Fig. 3C, panel C), characterized by circular necrotic lesions within a background of complete leaf necrosis. Sequencing of CP regions was conducted for all progenies as described above; the results are given in Table 2. The E‐K236 secondary mutation that had been identified in the inoculum was lost from all progenies. Instead, the same V‐A230 mutation present in the inoculum source, progeny of the wild‐type parent and progeny of the A‐T78 mutant was observed. In tomato, the introduced A‐T114 mutation was retained in all progenies irrespective of the Rx status of the plants, but was accompanied by an A‐D100 secondary mutation in 75% of progeny clones obtained from Rx‐expressing plants. In wild‐type N. benthamiana, the A‐T114 mutation was retained in all instances. On the contrary, it was absent in all progeny clones obtained from N. tabacum and, instead, a secondary D‐E3 mutation was observed.

This work describes the analysis of the impact of the introduction of two point mutations on the infection phenotype in wild‐type and Rx‐expressing plants of three host species. These mutations in the Rx minimal elicitor region of an Rx‐sensitive PepMV IC of the EU genotype were selected because they were expected to confer Rx‐breaking properties (Candresse et al., 2010). The wild‐type EU IC possessed full capacity for systemic movement and accumulation in all three wild‐type hosts tested but, as expected from previous reports (Baurès et al., 2008; Candresse et al., 2010), it was efficiently and specifically restricted in all Rx‐expressing hosts, further confirming that PepMV is recognized by the Rx‐sensing mechanism. It is interesting to note that localized necrotic responses were observed both at and around the site of inoculation for N. benthamiana and S. lycopersicum, whereas no such reaction was observed in N. tabacum, possibly as a consequence of Rx functioning more efficiently in this host. Previous work has shown that Rx confers a complete ER phenotype when confronted with a range of different avirulent potexviruses, including PepMV (Baurès et al., 2008; Candresse et al., 2010). One possibility for this discrepancy is that work carried out by Candresse et al. (2010) concerned the CH2 genotype of PepMV, whereas the IC used in the present investigation was of the EU strain.

A secondary mutation, V‐A230, was observed in all wild‐type EU IC progeny, as well as in almost all progenies derived from the two mutants, the sole exception of which was the first progeny obtained in wild‐type N. benthamiana for mutant A‐T114. This mutation was observed irrespective of the Rx status of the host species, suggesting that its highly reproducible accumulation probably reflects the reversion of a detrimental mutation present in the parental IC. In keeping with this interpretation, the alanine at position 230 is highly conserved among PepMV isolates and only absent in three of 82 PepMV CP sequences present in GenBank, all three deriving from the EU IC used in the present experiments. However, the E‐K236 mutation observed in the A‐T114 inoculum, but lost on further propagation, is probably the result of unselected genetic drift. The same could be true for the D‐E3 mutation observed in the progeny of the same mutant on propagation in wild‐type tobacco, but this remains to be conclusively demonstrated.

Mutation A‐T78 was sufficient to confer Rx‐breaking properties in both N. benthamiana and S. lycopersicum, but not in N. tabacum, without a need for any additional secondary mutation in the CP. This result is in line with previous work, where the A‐T78 mutation was identified alone in the CP region of Rx resistance‐breaking variants of PepMV in tomato (Candresse et al., 2010). In contrast with the A‐T78 mutant, A‐T114 was only able to overcome Rx in S. lycopersicum and not in N. benthamiana or N. tabacum. The role of mutation A‐T114 in conferring Rx‐breaking activity in tomato is less clear‐cut than for A‐T78, as a secondary A‐D100 mutation was also observed in the majority of clones sequenced.

It would appear that mutation A‐T114 confers Rx‐breaking activity in S. lycopersicum as it was observed alone in one of the progenies. However, its frequent association with a second compensatory mutation, such as A‐D100 reported here or A‐V71 previously observed together with A‐T114 in spontaneous Rx‐breaking mutants (Candresse et al., 2010), suggests that these secondary mutations may either improve the Rx‐breaking ability or the fitness of the A‐T114 mutant. In all cases, Rx breaking in S. lycopersicum was accompanied by a spreading necrosis phenotype (Candresse et al., 2010; present work), which is Rx mediated, as it is not observed in wild‐type tomato. The secondary A‐D100 mutation was observed in the majority of sequenced clones from Rx tomato, yet was absent from all sequences obtained from wild‐type tomato, indicating that its compensatory role in Rx breaking or in restoring the fitness of the A‐T114 mutant is Rx dependent. Interestingly, this same A‐D100 mutation has been observed in the CP region of two PepMV resistance‐breaking variants in S. lycopersicum (Candresse et al., 2010), alongside Q‐R125 in one variant, and with both A‐T78 and Q‐R125 in another.

The results of this investigation show that Rx possesses a high level of recognition in tobacco, as no resistance breaking is observed and systemic movement of the virus is halted. In N. benthamiana, Rx recognition is intermediate. A‐T114 is recognized (local lesions), but not localized, whereas A‐T78 is not recognized and displays complete systemic movement capability. In S. lycopersicum, recognition is weakest and both mutants, although still recognized, evade localization and overcome resistance. Evidence for intermediate elicitor recognition phenotypes observed in CP–Rx‐based systems shows that the intensity of the response may vary, clearly based on the strength of protein–protein interactions (Baurès et al., 2008; Sturbois et al., 2012). Indeed, the findings of this investigation nicely parallel those of Sturbois et al. (2012), whereby different tomato mutants were found to possess different interaction phenotypes when confronted with a mutant PVX isolate of intermediate Rx elicitor activity. Work by Harris et al. (2013) concerning the artificial evolution of Rx found that extending the range of Rx recognition to include Poplar mosaic virus (PopMV) could come with a cost of systemic trailing necrosis. The Rx resistance response was demonstrated to consist of separate recognition and activation phases, with PopMV being recognized, but a delayed or incomplete activation of Rx resulting in an inability to suppress viral movement and in a trailing HR phenotype. The necrotic symptoms caused by PepMV mutants in Rx‐expressing hosts in this work suggest that the mutants are similarly recognized, but not localized, because of host‐specific differences in the sensing or downstream signalling in the heterologous hosts studied.

Another point to consider may be host‐dependent fitness penalties associated with each mutation. However, fitness penalties associated with the debilitating T‐K121 mutation in PVX mutants are seen in non‐Rx hosts (Goulden et al., 1993). On the contrary, the equal levels of accumulation displayed by the various mutants in the wild‐type hosts in this work makes this mechanism unlikely.

In conclusion, the results of this investigation support the guard hypothesis of R‐gene functionality (Dangl and Jones, 2001), which implies the existence of host adaptors that may contribute to resistance efficiency and durability. Understanding the mechanisms underlying the increased durability of Rx‐based resistance in the N. tabacum host may be integral if the Rx gene is to provide a suitable form of resistance against PepMV in tomato. This is a much more complex system than previously thought; the cellular environment in which Rx is expressed is integral in its functionality.