Lettuce mosaic virus: from pathogen diversity to host interactors

INTRODUCTION

Lettuce mosaic disease was first described in Florida (Jagger, 1921) and is now distributed worldwide, probably because seeds have been exchanged internationally over many years (Dinant and Lot, 1992). It occurs in all continents, including Europe, North and South America (Mexico, USA, Argentina, Brazil, Uruguay), the West Indies (Bermuda), Africa, the Middle East (Egypt, Israel, Jordan, Iraq, Iran, Turkey), Asia (China, Japan) and Oceania (Australia, Tasmania, New Zealand). Lettuce mosaic virus (LMV) is a major pathogen of commercial crops in lettuce‐growing areas of the world. Severe losses are recorded mainly in field crops, but the disease may be significant in the greenhouse when seedlings are not grown under insect‐proof conditions (Dinant and Lot, 1992). Elementary sanitary measures (keeping lettuce nurseries away from fields in which crops are grown etc.) may improve control of the disease. The detection of LMV in infected plants or in seed lots is routinely carried out using immunological techniques such as enzyme‐linked immunosorbent assay (Clark and Adams, 1977). Because of the prevalence of seedborne virus (Grogan et al., 1952; Newhall, 1923; Tomlinson, 1970), lettuce seeds can be tested by direct observation of lettuce seedlings. Indeed, seedlings with seedborne virus have misshapen cotyledons, the first true leaf is misshapen and has a dark green mottling appearance. Inoculation of a sensitive indexing host with sap extracted from the ground‐up seed, or a serological technique (Falk and Purcifull, 1983), can also ensure that each lettuce seed lot contains no infected seeds in a sample of 30 000 seeds (MT0, ‘Mosaic Tolerance Zero’ or zero infected seeds in 30 000). More recently, efforts have been made to develop other more sensitive techniques of detection of LMV, based on the polymerase chain reaction (Peypelut et al., 2004). It was shown that expression of a capsid protein transgene protects lettuce against LMV infection (Dinant et al., 1997). However, a more successful control measure is the incorporation of natural virus resistance into the principal lettuce types grown (Walkey et al., 1985).

The only resistance genes currently used to protect lettuce crops worldwide are the recessive allelic genes mo1 ¹ and mo1 ². The mo1 ¹ gene, formerly named g, was first identified in Argentina, in a Latin‐type cultivar named ‘Gallega de Invierno’ (Bannerot et al., 1969). In Europe, lettuce breeders used the Gallega source of resistance to incorporate the g gene in numerous varieties of lettuce, including butterhead, Batavia, cos and crisphead types (Pink et al., 1992a). The mo1 ² gene, identified in three Egyptian wild Lactuca sativa lines and named the recessive gene mo (Ryder, 1970), has been mostly used by North American breeders who introduced it into crisphead and cos types of lettuce (Pink et al., 1992a). Initially considered identical, these genes were later shown to have different specificities and to be either allelic or closely linked and therefore were renamed mo1 ¹ and mo1 ² (Dinant and Lot, 1992). These genes have been recently cloned and sequenced in our laboratory (Nicaise et al., 2003). The resistance alleles mo1 ¹ and mo1 ², as well as the susceptibility allele mo1 ⁰, were found to code for forms of the eukaryotic translation initiation factor eIF4E in lettuce.

Depending on the viral isolate to which they are confronted, these genes can be considered as inducing resistance (no detectable virus accumulation) or tolerance (virus accumulation but failure to induce significant symptoms). Most of the field isolates of LMV are seedborne in susceptible lettuce cultivars, but not in resistant cultivars carrying mo1 ¹ or mo1 ², even in cultivars with low levels of resistance (Dinant and Lot, 1992; Pink et al., 1992a). Therefore, in addition to a reduction of viral infection or symptom expression, the mo1 ¹ and mo1 ² genes also provide a reduction in the dissemination of LMV through seed. For simplicity, only the term resistance will be used to refer the complex set of phenotypes associated with the mo1 ¹ and mo1 ² genes.

In addition to the economic importance of LMV, the large biological diversity and the differences between the biological properties of isolates (symptoms, seed transmission, behaviour towards mo1 genes) make it a very good model to study plant–virus interactions, from both host and pathogen perspectives. Moreover, molecular tools available for LMV, including highly infectious cDNA copies of the LMV genome with the full‐length (FL) cDNA placed under the control of the enhanced CaMV 35S promoter and of the NOS terminator, have been constructed (Redondo et al., 2001; Yang et al., 1998). These infectious clones have provided a tool of tremendous importance to study the molecular genetics of LMV. This article aims to highlight advances made in understanding the lettuce–LMV interactions, at the population, individual and molecular levels.

LMV DIVERSITY: ‘LMV‐MOST ISOLATES’, A THREAT TO DURABLE RESISTANCE?

While LMV was considered appropriately controlled by the use of mo1 ¹ and mo1 ², occasional outbreaks of resistance‐breaking forms of LMV have been described for several decades (Dinant and Lot, 1992; Pink et al., 1992a). Usually, these resistance‐breaking isolates are not seedborne, which has suggested a link between the gain of virulence and the loss of seed transmission. However, since the beginning of the 1990s, mo1‐breaking isolates with high rates of seed‐transmissibility have been described (Dinant and Lot, 1992). In addition to the abilities of being seedborne and of infecting mo1 varieties, LMV isolates differ in the symptoms they induce, which can vary from barely detectable to strongly necrotic or even lethal for a same host (Fig. 1) (Krause‐Sakate et al., 2005; Kyriakopoulou, 1985).

Attempts to link the key biological properties of LMV isolates with their sequence clustering have been carried out (Krause‐Sakate et al., 2002; Revers et al., 1997a). LMV isolates could be clustered in three main groups: a single isolate from Yemen (the upper branch in Fig. 2), a group from the Balkans (Greece and Croatia, named Gr) and a third group with very diverse geographical origins (including the Middle East and Greece, called ‘Rest of the World’ or RoW) (Fig. 2). No seedborne isolate was ever observed in the Yemen and Gr groups. Within the RoW group, two large subclusters of isolates contained all known seedborne LMV isolates: one with isolates unable to infect mo1 ¹ or mo1 ² plants, collectively named ‘LMV‐Common’, and the other with the isolates cumulating mo1 breaking and seed transmission, collectively named ‘LMV‐Most’ for mo1‐breaking, Seed Transmitted. Correlation of the sequence clustering of LMV isolates with their abilities to infect mo1 plants remains less clear than with their seed transmission properties. On the one hand, no mo1‐breaking isolate was ever observed in the LMV‐Common group, and all LMV‐Most isolates and LMV‐Gr assayed, as well as the Yemen isolate, were able to infect mo1 ¹ as well as mo1 ² plants. On the other hand, while the vast majority of LMV isolates outside the LMV‐Common group are able to overcome both mo1 alleles, variations in this respect occur sporadically in the dendrogram (Krause‐Sakate et al., 2002), the most remarkable occurrence being LMV‐1 (overcoming mo1 ¹ but not mo1 ²), which clusters very close to LMV‐E (overcoming both mo1 alleles). Altogether, sequence clustering evidence suggests that the ability to infect mo1 plants did not evolve from isolates unable to do so, such as LMV‐Common.

These studies therefore established a link between biological properties and sequence clustering, and provided the bases for strain‐specific detection of LMV, once the complete nucleotide sequence of an LMV‐Most isolate had been established (Krause‐Sakate et al., 2002; Peypelut et al., 2004). They also enabled a re‐writing of the scenario leading to local outbreaks of resistance‐breaking LMV: instead of the loss of seed transmission upon acquisition of the ability to overcome mo1 (Dinant and Lot, 1992), the available evidence more simply suggests that these outbreaks are caused by non‐Common‐non‐Most isolates occurring locally, perhaps in weed reservoirs, and which are primarily both unable to infect lettuce seed‐to‐seed and able to infect mo1 plants; in this scheme, LMV‐Most probably represents one of these local forms of LMV that occurred to be, or evolved to be, seedborne in lettuce and was therefore spread worldwide through seed trade.

The molecular variability of LMV isolates was also revealed, using monoclonal antibodies directed against the coat protein, but these studies could not reveal any difference between LMV‐Common and LMV‐Most, probably owing to their identical amino‐acid sequence in the immunogenic N‐terminus of the coat protein (Candresse et al., 2007).

The occurrence of LMV‐Most is a concern for lettuce production worldwide, as these isolates are able to overcome the two major modes of control of LMV, namely genetic resistance and seed control. Therefore, specific effort must be made to avoid the spread of LMV‐Most in seeds, by promoting the propagation of lettuce seeds in LMV‐free environments and the dissemination of virus‐free seeds, developing specific detection tools (Peypelut et al., 2004), identifying potential reservoirs specific for LMV‐Most, and understanding the spread of LMV between host species and the molecular bases of the typical biological properties of LMV‐Most.

GENOMIC ORGANIZATION OF LMV

The genome of LMV consists of a single positive‐strand RNA of 10 080 nt in length. The genomic RNA of Potyviruses is covalently linked at its 5′ end to a virus‐encoded VPg protein (Murphy et al., 1991), and is polyadenylated at its 3′ end (Adams et al., 2005). There is a single open reading frame (ORF) flanked by two untranslated regions (UTRs, of 103 and 210 nt at the 5′ and 3′ ends, respectively) that is translated into a single, large polyprotein processed by three virus‐specific proteinases (Reichmann et al., 1992). To date, four full‐length genomic RNA sequences, corresponding to isolates of LMV differing in their biological properties, have been published: two mo1 resistance‐breaking isolates, LMV‐E and LMV‐AF199 (GenBank accession nos. X97705 and AJ278854 respectively) (Krause‐Sakate et al., 2002; Revers et al., 1997b), and two LMV‐common isolates, LMV‐0 from Europe (GenBank X97704) (Revers et al., 1997b) and LMV‐Yuhang from China (GenBank AJ306288) (Zheng et al., 2002).

We focus here on the three molecularly well‐characterized LMV isolates belonging to the same group of isolates: LMV‐E (isolated in Spain by H. Lot, INRA‐Avignon, France), a non‐seedborne resistance‐breaking isolate that provokes symptoms in cultivars carrying mo1 ¹ or mo1 ² genes (Pink et al., 1992b); LMV‐0, an LMV‐common isolate that provokes very mild or no symptoms on cultivars carrying the mo1 ² gene (tolerance) but does not invade systemically the cultivars carrying the mo1 ¹ gene (resistance) (Dinant and Lot, 1992); and LMV‐AF199 (LMV‐Most), which in addition to being seedborne, overcomes the mo1 ¹ and mo1 ² genes in lettuce (Krause‐Sakate et al., 2002).

Regardless of their large differences in pathogenicity, resistance‐breaking and seed‐transmission properties (Table 1), the entire genomic sequences differ from each other only by point mutations, with no deletion or insertions. The overall nucleotide sequence identities between LMV‐AF199 and LMV‐0, LMV‐AF199 and LMV‐E, and LMV‐E and LMV‐0 are 95.9, 93.9 and 94.0% respectively. At the amino acid sequence level, the identities are 98.0, 96.7 and 97.0%, respectively. Generally, the largest variability occurs in the P1 and the N‐terminal region of the coat‐protein (although more conserved between LMV‐0 and AF199) while the NIa protease domain, the NIb protein, the C‐terminus of the helper‐component protease (HcPro) and the 3′ non‐coding region are more conserved. A recombinant LMV isolate resulting from a natural exchange between LMV‐Most and LMV‐Common in a field where both strains occurred has been isolated in only one instance thus far. The putative recombination site was located within the P3 coding region (Krause‐Sakate et al., 2004).

PATTERN OF LMV INVASION IN SUSCEPTIBLE AND RESISTANT MO1 LETTUCE

In order to follow the viral invasion of susceptible and resistant lettuce, GFP and GUS markers have been used for LMV‐E and LMV‐0. When fused to the N‐terminus of the viral protein HcPro, both reporter genes affect the biological properties of recombinant LMV isolates in both susceptible and resistant lettuce varieties (German‐Retana et al., 2000). Upon addition of the NIa cleavage site between the reporter gene and HcPro, in such a way that a nearly wild‐type HcPro is produced upon action of the NIa proteinase, LMV‐0 and LMV‐E recombinant viruses recovered the behaviour of their wild‐type parent (symptoms, viral accumulation) in susceptible plants (German‐Retana et al., 2003) (Fig. 3A–G). In mo1 ² plants, the recombinant LMV‐E modified in this way recovered the breaking properties of its wild‐type counterpart. In mo1 ² plants, the LMV‐0‐derived recombinants showed a severe inhibition in systemic accumulation (3, 4), despite the fact that neither cell‐to‐cell movement nor phloem loading or unloading seemed to be severely affected in an mo1 ² genetic context (German‐Retana et al., 2003). Although infection foci are present in the LMV‐0‐GUS‐inoculated leaves of both quasi‐isogenic lettuce cultivars Salinas (susceptible) and Salinas 88 (mo1 ², resistant) (Fig. 4A,B), LMV‐0‐GUS systemic movement in the upper non‐inoculated systemic leaves is detected only in the Salinas cultivar, in contrast to Salinas 88 (Fig. 4C). This suggests a restriction in long‐distance movement and in systemic accumulation of the tagged LMV‐0 recombinants in mo1 ² lettuce (3, 4). An interesting observation is that the systemic movement to the upper non‐inoculated leaves appears to be more affected than the downward movement to the root system (3, 4). In general, viral invasion of the root system has been poorly studied and in most cases, when analysed, both upward and downward systemic movements were affected. However, similar to the situation reported here with LMV, Guerini and Murphy (1999) showed that in the resistant pepper (Capsicum annuum) variety Avelar, carrying the recessive gene pvr3, downward movement of Pepper mottle virus (PepMoV) to the roots still occurred while systemic movement to upper non‐inoculated leaves was completely blocked due to a block of entry into the internal phloem.

LMV‐GFP: TOOLS FOR LETTUCE BREEDERS

Beside their usefulness for studying viral invasion, both LMV‐0‐GFP and LMV‐E‐GFP can also be very useful tools to facilitate the screening of lettuce plants for LMV resistance, and identification of the resistance alleles present in a particular variety, both in vivo and in vitro (Candresse et al., 2002; Mazier et al., 2004). An evaluation of 101 cultivars of known status was carried out with these recombinant viruses and a 100% correlation was observed between LMV‐0‐GFP behaviour (whose systemic movement is abolished in resistant plants) and the mo1 resistance status. Similarly, the LMV‐E‐GFP (GFP fused to HcPro) allowed the identification of mo1 ² lines because its systemic movement was restricted in mo1 ² lines but not in susceptible or mo1 ¹ lines (Candresse et al., 2002). The use of these recombinant viruses can therefore greatly facilitate LMV resistance evaluation and speed up lettuce breeding programmes. Furthermore, the GFP LMV viruses constitute a simple and efficient tool for testing LMV resistance in in vitro cultivated lettuce, a method which reduces space requirements and improves environmental safety (Mazier et al., 2004).

LMV PATHOGENICITY DETERMINANTS IN SUSCEPTIBLE AND RESISTANT LETTUCE CULTIVARS MAP TO DIFFERENT REGIONS OF THE VIRAL GENOME

MOLECULAR DIALOGUE BETWEEN LMV AND LETTUCE: LOOKING FOR PLANT PARTNERS?

In the case of obligatory parasites such as viruses, absence or inadequacy of a single host factor may lead to the inability of the pathogen to multiply in the host or to invade it systemically (Ishikawa et al., 1997; Yamanaka et al., 2000). Such a mechanism implies that the dominant alleles of the host genes involved would be associated with susceptibility and the recessive alleles encoding non‐functional versions of this host factor with resistance.

Recessive resistance genes used to control Potyviruses agronomically have been estimated to represent about 40% of the known resistance genes (Provvidenti and Hampton, 1992). In the pathosystem lettuce/LMV, based on the observation that VPg was in the domain of the virus genome involved in mo1 breaking and that an interaction between VPg and the eukaryotic translation initiation factor 4E (eIF4E) had been described for two other potyvirus models (Schaad et al., 2000; Wittmann et al., 1997), we isolated three alleles of the lettuce eIF4E in their cDNA form, and obtained circumstantial and functional evidence that two of these alleles correspond to the recessive LMV resistance genes mo1 ¹ and mo1 ² (Nicaise et al., 2003). The immediate consequence of this conclusion is that mo1 ¹ and mo1 ² in lettuce are the mutant alleles of a unique mo1 gene encoding eIF4E.

During the last 5 years, it has been shown that natural mutations of components of the eukaryotic translation initiation complex that result in resistance to specific RNA viruses (especially Potyviruses) occur in a range of plant species (tomato, lettuce, pepper, pea, melon, barley, rice) (for reviews see Diaz‐Pendon et al., 2004; Maule et al., 2007; Robaglia and Caranta, 2006). However, how eIF4E is involved in the infection cycle in plants is currently not fully understood.

Several roles have been proposed for eIF4E in the potyvirus infection cycle based on its known biological and biochemical features. In particular, it was proposed that it could play a role during translation initiation through interaction with the genome‐linked protein VPg at the 5′ end of viral RNA (Lellis et al., 2002), which interacts with eIF4E in several plant–potyvirus systems (Beauchemin et al., 2007; Leonard et al., 2000; Roudet‐Tavert et al., 2007; Schaad et al., 2000; Wittmann et al., 1997). The in vitro interaction between the VPg of LMV and eIF4E from lettuce has been shown (Roudet‐Tavert et al., 2007) and characterized through spectroscopic studies (Michon et al., 2006). The central domain of the LMV VPg is involved in the interaction with the lettuce eIF4E. The VPg forms a ternary complex with both eIF4E and eIF4G, reducing eIF4E affinity for an mRNA cap analogue.

During mRNA translation, eIF4E provides the cap‐binding function and is associated with the protein eIF4G to form the eIF4F complex. Recently, susceptibility analyses of Arabidopsis mutants knocked‐out for At‐eIF4G genes showed that eIF4G factors are also indispensable for LMV infection, and that the eIF4G selective involvement parallels eIF4E recruitment, which suggests recruitment of the whole eIF4F for LMV infection in Arabidopsis (Nicaise et al., 2007).

These results could be simply interpreted as the 5′ VPg of LMV RNA functionally playing a role equivalent to the 5′ cap of cellular mRNAs, as recently shown for an animal calicivirus (Goodfellow et al., 2005). Through its interaction with VPg and possibly other host and virus factors, eIF4E might be involved in the control of the successive fates encountered by the viral RNA, such as intracellular and cell‐to‐cell trafficking (Arroyo et al., 1996; Gao et al., 2004). Another possible implication of eIF4E in the virus cycle could be to allow RNA circularization by interaction of the 5′ VPg with the 3′ poly A, mediated by the same protein complex as in mRNA translation, namely eIF4E‐eIF4G‐PABP. Beside a role in translation, genome circularization may be required for virus RNA replication or other processes of the infection cycle. Indeed, genome circularization is an important feature of the replication of Picornaviruses (Herold and Andino, 2001), relatives of Potyviruses infecting animal hosts.

To date, however, the function of eIF4E and eIF4G during the infection process remains to be elucidated. Roles in the early events of infection are the main candidate hypotheses: viral RNA translation and/or replication, circularization of viral RNA, host protein sequestration, or virus movement from infected to uninfected cells.

CONCLUSIONS

In recent years, our knowledge concerning both LMV and host proteins involved in LMV/lettuce interactions has improved significantly.

LMV pathogenicity determinants in lettuce map to different regions. While the HcPro is involved in the determination of symptom severity in Trocadéro, the 3′ half of the genome including the VPg enables systemic infection and symptom induction on cultivars carrying the genes mo1 ¹ and mo1 ². These results indicate that the ability of LMV to induce severe symptoms and to overcome the protection in lettuce afforded by the recessive allelic genes mo1 ¹ and mo1 ² are independent phenomena. Furthermore, recent data have enabled us to narrow down this viral 3′ region and indicate which protein(s) other than VPg are actually involved in the virulence of LMV.

The identity of the resistance mo1 genes has been clarified since they were shown to encode natural variants of the cap‐binding protein eIF4E, which are unable, in the homozygous state, to provide the cellular and molecular background necessary for LMV accumulation or symptom induction (Nicaise et al., 2003). In order to develop targeted resistance to LMV in lettuce, efforts can focus on screening of a large collection of lettuce natural or artificial mutants in the candidate eIF4E gene using TILLING (targeting induced local lesions in genomes) (McCallum et al., 2000).

All these discoveries should lead to a better understanding of the interactions between Potyviruses and their hosts, a challenge that our laboratory is tackling with the objective of providing new and more sustainable sources of resistance to various Potyviruses in various crops.