A natural recessive resistance gene against potato virus Y in pepper corresponds to the eukaryotic initiation factor 4E (eIF4E)


*For correspondence (fax + 334 327 227 02; e-mail caranta@avignon.inra.fr).


We show here that the pvr2 locus in pepper, conferring recessive resistance against strains of potato virus Y (PVY), corresponds to a eukaryotic initiation factor 4E (eIF4E) gene. RFLP analysis on the PVY-susceptible and resistant pepper cultivars, using an eIF4E cDNA from tobacco as probe, revealed perfect map co-segregation between a polymorphism in the eIF4E gene and the pvr2 alleles, pvr21 (resistant to PVY-0) and pvr22 (resistant to PVY-0 and 1). The cloned pepper eIF4E cDNA encoded a 228 amino acid polypeptide with 70–86% nucleotide sequence identity with other plant eIF4Es. The sequences of eIF4E protein from two PVY-susceptible cultivars were identical and differed from the eIF4E sequences of the two PVY-resistant cultivars Yolo Y (YY) (pvr21) and FloridaVR2 (F) (pvr22) at two amino acids, a mutation common to both resistant genotypes and a second mutation specific to each. Complementation experiments were used to show that the eIF4E gene corresponds to pvr2. Thus, potato virus X-mediated transient expression of eIF4E from susceptible cultivar Yolo Wonder (YW) in the resistant genotype YY resulted in loss of resistance to subsequent PVY-0 inoculation and transient expression of eIF4E from YY (resistant to PVY-0; susceptible to PVY-1) rendered genotype F susceptible to PVY-1. Several lines of evidence indicate that interaction between the potyvirus genome-linked protein (VPg) and eIF4E are important for virus infectivity, suggesting that the recessive resistance could be due to incompatibility between the VPg and eIF4E in the resistant genotype.


Virus resistance genes, long a mainstay of the plant breeder, provide a highly efficient barrier to virus infection in plants. The best characterized resistance genes include dominant genes (R-genes) that trigger a hypersensitive response (HR) or extreme resistance (ER). These two resistant phenotypes are conferred through matched specificity between the R-gene and a pathogen avirulence gene, and operate on the basis of ‘gene-for-gene’ mechanism (Keen, 1990). Four virus R-genes have been cloned (N, Rx1, Rx2 and Sw-5) and all belong to the nucleotide binding site, leucine-rich repeat (NBS-LRR) super-family of R-genes (Whitham et al., 1994; Bendahmane et al., 1999; Bendahmane et al., 2000; Brommonschenkel et al., 2000).

Other types of genetic resistance include virus resistance genes that are not associated with HR or ER (Fraser, 1990). The two dominant genes RTM1 and RTM2, involved in restriction of long-distance movement of tobacco etch virus (TEV) in Arabidopsis thaliana fall into this class (Chisholm et al., 2000; Whitham et al., 2000). Resistance in this case is not caused by activation of known defense pathways and it remains unclear how these genes operate. Recessive resistance genes also belong to this class but, up to now, most information about the molecular mechanisms involved in recessive resistance was obtained using mutagenesis approach. The recessive mlo mutation in barley which controls broad spectrum resistance against several isolates of the fungus Erysiphe graminis f.sp. hordei, and the Arabidopsis recessive mutant, edr1, which provokes resistance to some bacterial and fungal pathogens (Büschges et al., 1997; Frye and Innes, 1998), illustrate mutations affecting the control of the defense response and/or cell death. Concerning viruses, it is widely believed that incompatibility occurs when the plant lacks one or several genes required for virus infection. Such recessive mutations suppressing efficient multiplication of tobamoviruses and potyviruses were recently identified in Arabidopsis (Yamanaka et al., 2000; Yamanaka et al., 2002; Lellis et al., 2002).

Plant–potyvirus interactions constitute an interesting model to investigate recessive resistance because an unusually high frequency of occurrence of genes conferring recessive resistance to potyviruses has been observed (40% versus 20% for resistance against other viruses) (Provvidenti and Hampton, 1992). pvr2-mediated resistance of pepper (Capsicum spp.) to potato virus Y (PVY) falls into this type of resistance. The pvr2 resistance locus consists of two alleles, pvr21 and pvr22 (Kyle and Palloix, 1997): pvr21 is effective only against PVY-0, while pvr22 is effective against PVY-0 and -1 but is overcome by PVY-1,2. Further studies revealed another characteristic feature of potyvirus resistance genes: resistance is often not restricted to a single potyvirus. Thus, pvr22 is also associated with resistance to a second potyvirus, the TEV. Whether a single locus or two tightly linked genes controls PVY and TEV resistance remains to be determined.

In this paper, we describe isolation of the pvr2 gene using a candidate gene approach, an alternative to map-based cloning and insertional mutagenesis when assumptions can be made regarding the biological function of the gene of interest (Pflieger et al., 2001). We demonstrate map co-segregation between the pvr2 resistance locus and a eukaryotic translation initiation factor 4E (eIF4E) gene. eIF4E was chosen as a plausible candidate gene because (i) eIF4E proteins bind with the viral protein genome linked (VPg) to turnip mosaic virus (TuMV) and TEV (Wittmann et al., 1997; Léonard et al., 2000; Schaad et al., 2000); (ii) the resistance-breaking determinants for several recessive potyvirus resistance genes localize in the VPg-coding domain (Nicolas et al., 1997; Keller et al., 1998; Masuta et al., 1999; Hjulsager et al., 2002, Ramajäki and Valkonen, 2002) ; and (iii) the VPg–eIF4E interaction is necessary for virus infectivity and upregulates genome amplification (Léonard et al., 2000; Schaad et al., 2000). Moreover, it was recently shown that disruption of the eIF(iso)4E gene in Arabidopsis resulted in resistance to potyviruses (Lellis et al., 2002). Confirmation that pvr2 is indeed an allele of eIF4E was obtained using a potato virus X (PVX)-based transient expression assay. Two nucleotide substitutions inducing amino acid differences between alleles in resistant and susceptible plants were identified.


An eIF4E gene co-segregates with the pvr2 locus

eIF4E belongs to a small multigenic family (Rodriguez et al., 1998; Manjunath et al., 1999). Therefore, conventional RFLP was chosen as the mapping strategy because it facilitates mapping of multiloci components. To look for map co-segregation between eIF4E and the pvr2 locus, tobacco eIF4E cDNA was used as a probe on genomic DNA of pepper cultivars with distinct alleles at the pvr2 locus. With EcoRI-digested DNA, hybridization revealed an RFLP of approximately 7 kb between the PVY-susceptible cultivar, Yolo Wonder (YW) and the PVY-resistant cultivars, Yolo Y (YY) and Florida VR2 (F) (Figure 1). Four other major EcoRI fragments, monomorphic between YW, YY and F, were also detected confirming the occurrence of several eIF4E genes in the genome.

Figure 1.

Map co-segregation between the pvr2 locus and a RFLP marker detected with a tobacco eIF4E cDNA probe.

Yolo Wonder (YW) contains the pvr2+ allele for susceptibility to PVY; Yolo Y (YY) contains the pvr21 recessive allele for resistance to PVY-0 and FloridaVR2 (F) contains the pvr22 recessive allele for resistance to PVY-0 and -1. F1 corresponds to the F1 hybrid between YW and YY. Lanes R correspond to DNA from PVY-0 resistant plants of the BC1 progeny (YW × YY) × YY and lane S to DNA from PVY-0 susceptible plants from the BC1. The EcoRI DNA fragment linked to the resistant alleles is indicated by A and the EcoRI DNA fragment linked to the susceptible allele is indicated by B. MW indicates DNA size markers.

Segregation observed for PVY-0 resistance in the BC1 progeny of 440 plants was consistent with a single recessive gene model (224 resistant : 216 susceptible; χ2 for 1 : 1 = 0.145; = 0.7). DNA from each BC1 was digested with EcoRI and subjected to Southern blot hybridization with the tobacco eIF4E cDNA probe. The 224 BC1 plants resistant to PVY (homozygous for pvr21) presented the same RFLP pattern as YY whereas the 216 BC1 plants susceptible to PVY (heterozygous pvr2+/pvr21) presented the same RFLP pattern as the F1 (YW × YY), indicating perfect map co-segregation between the eIF4E related-sequence and the pvr2 locus.

Molecular cloning of the pepper eIF4E gene

3′ and 5′ RACE experiments were used to obtain the full-length sequence of an eIF4E cDNA from YW. The resulting cDNA was 870-nucleotides long, with a 5′ untranslated region of 28 nucleotides preceding the first in-frame AUG. The open reading frame (ORF) is 687 nucleotides in length and encodes a protein of 228 amino acids, assuming that translation initiates at the first AUG codon, as in tobacco and tomato eIF4E mRNA. Moreover, the sequence AA−3CAAUGG+4C flanking the putative initiation codon exactly matches the consensus sequence reported for initiation codons in plants (Lutcke et al., 1987). The pepper eIF4E ORF displays 73.4% nucleotide sequence identity with the tobacco eIF4E sequence and 86.1% identity with that of tomato (GenBank accession nos. AF259801). The pepper eIF4E sequence is more similar to the ArabidopsiseIF4E ORF sequence (EMBL accession no. Y10548) than it is to the Arabidopsis eIF (iso) 4E ORF sequence (EMBL accession no. Y10547) with 70.4 and 55.3% identity, respectively, suggesting that the cDNA codes for a protein of the eIF4E subfamily.

To determine whether the cloned cDNA corresponds to the eIF4E gene that co-segregates with pvr2, a pepper BAC library was screened using primers P1 and P2. PCR amplified a single fragment of 493 bp from YW cDNA and a single fragment of 1800 bp on YW genomic DNA, indicating the presence of at least one intron in this part of the gene. If the BAC clones identified with these primers contain the eIF4E gene which co-segregates with pvr2, they should contain a 7-kb EcoRI fragment corresponding to the RFLP (Figure 1) and this 7-kb fragment should contain at least part of the eIF4E gene. Four BAC clones were identified, a number consistent with the library's estimated representation based on the average insert size and the total number of clones. Restriction analysis indicated that the four BAC clones correspond to the same locus in the genome. EcoRI digestion of the clones produced a 7-kb fragment. This fragment was cloned and amplified using primers P1 and P2. The sequence of the resulting 1800-bp PCR product demonstrated that the eIF4E gene which co-segregated with the pvr2 locus had been cloned.

Two amino acid substitutions within the eIF4E protein distinguish the susceptible and resistant pepper genotypes

Nucleotide sequences of the eIF4E cDNA of YW, DDL, YY and F were obtained from three independent RT-PCR experiments. Amino acid sequence alignment revealed that the two unrelated susceptible genotypes YW and DDL were 100% identical in their eIF4E sequences (Figure 2). Compared to the susceptible genotypes, the two resistant genotypes, YY and F, each contained two amino acid substitutions. A Glu instead of a Val was present at position 67 of both resistant genotypes (Figure 2). In YY (pvr2-1), a second specific mutation occurred at position 79 (Arg for Leu) whereas F (pvr22) contained an Asn for Asp mutation at position 109 (Figure 2). Alignment of the pepper eIF4E protein sequences with eIF4E sequences of Arabidopsis (EMBL accession no. Y10548), Saccharomyces cerevisiae (GenBank accession no. P07260), Mus musculus (GenBank accession no. NP_031943), and Homo sapiens (GenBank accession no. XP_017925) showed that these mutations do not target highly conserved amino acids, or amino acids implicated in recognition of the mRNA 5′ cap structure (m7GpppN, where N is any nucleotide), or in interactions with the eIF4G and 4E binding proteins (Marcotrigiano et al., 1997).

A PVX-mediated transient expression assay confirms that pvr2 corresponds to the eIF4E gene

The ORF corresponding to eIF4E from susceptible YW and resistant YY were cloned into PVX vector pPVX201 (Figure 3). To obtain a high-titer inoculum for application to pepper leaves, pPVXeYW, pPVXeYY and empty vector pPVX201 were propagated in Nicotianabenthamiana. Because genetic instability of the PVX vector insert was reported by Sablowsky et al. (1995), inoculum from N. benthamiana inoculated leaves and systemically infected leaves was tested on pepper plants. PCR analysis indicated that PVX eIF4E inserts were maintained in all parts of N. benthamiana. However, pepper plants inoculated with sap from systemically infected tissue developed severe symptoms and PCR analysis revealed PVX vectors without inserts or with shortened inserts. In contrast, inoculum from inoculated N. benthamiana leaves resulted in only weak symptoms on pepper and RT-PCR revealed that the eIF4E inserts were largely maintained in full-length form. Therefore, only inoculated leaves of N. benthamiana were used as inoculum source in experiments with pepper.

Based on the hypothesis that the dominant eIF4E allele is required for susceptibility to PVY, we predicted that transient expression of the dominant susceptible eIF4E allele in pepper plants resistant to PVY would lead to PVY genome amplification in the PVX-infected cells expressing the full-length eIF4E insert. In preliminary experiments, the effect of double-inoculation of pPVX201 and PVY on the phenotype of the pvr2 alleles was tested. Both RT-PCR and doubled antibody sandwich enzyme-linked immunosorbent assay (DAS-ELISA) analysis of PVY accumulation indicated that PVX systemic infection of pepper genotypes YW, YY and F followed by infection with PVY, did not alter infectivity of PVY-0 and PVY-1, compared to plants infected with PVY-0 or PVY-1 alone (Figure 4; Table 1).

Table 1.  Number of PVY-infected leaves/number of PVY-inoculated leaves in the PVX-based transient expression assay
Inoculated withPepper genotypes
  1. n.d.: not determined.

pPVX201 + PVY-020/200/200/20
pPVXeYW + PVY-020/204/204/17
pPVXeYY + PVY-020/200/20n.d.
pPVX201 + PVY-120/2030/300/30
pPVXeYW + PVY-120/20n.d.n.d.
pPVXeYY + PVY-120/2030/306/30

To test if PVY-0 infection of resistant YY (pvr21) could be supported by expression of eIF4E from susceptible YW, YY plants were inoculated with pPVXeYW and PVY-0. Of 20 double-inoculated leaves, four displayed significant accumulation of PVY coat protein and RNA (Figure 4a,b, lane 10). Similarly, PVY-0 accumulation was observed in four of 17 resistant genotype F (pvr22) plants double-inoculated with pPVXeYW and PVY-0 (Table 1). In parallel experiments, it was shown that expression of eIF4E from PVY-0 resistant YY did not support PVY-0 infection in YY plants (Figure 4a,b, lane 9). Susceptible genotype YW inoculated with PVY-0 or with pPVX201, pPVXeYW or pPVXeYY followed by inoculation with PVY-0, accumulated PVY capsid protein and RNA (Figure 4, lanes 3–6). These results indicate that PVX vector expression of eIF4E from the dominant allele pvr2+ from YW in resistant genotypes YY (pvr21) and F (pvr22) can lead to a high level of virus accumulation.

YY is resistant to PVY-0 but susceptible to PVY-1. Thus, we hypothesized that transient expression of the YY eIF4E allele in F, which is resistant to PVY-1, should lead to PVY-1 accumulation in double-inoculated F leaves. DAS-ELISA and RT-PCR experiments showed an accumulation of PVY-1 in six inoculated leaves among the 30 assessed; F inoculated with only PVY-1 never presented PVY accumulation and YY inoculated with PVY-1 or inoculated with pPVXeYY and PVY-1 presented a high level of accumulation of the PVY capsid antigen (Figure 4; Table 1). Thus, complementation experiments demonstrate that expression of the eIF4E susceptible alleles in resistant pepper genotypes is both necessary and sufficient to restore susceptibility to PVY.

PVY accumulation was never observed in upper uninoculated leaves of double-inoculated resistant pepper plants, suggesting that PVX expression of the eIF4E susceptible alleles did not support systemic infection with PVY. In inoculated leaves, the absorbance values observed in leaves from resistant genotypes expressing eIF4E susceptibility allele were significantly lower than in the susceptible controls (Figure 4a, lane 10 and 16). This can be explained by the requirement that the preliminary infection by PVX and subsequent eIF4E expression must occur in the same cells which are infected by PVY. Furthermore, even if a given cell becomes infected with both the PVX vector and PVY, it is evident that loss of the eIF4E insert from the vector in such a cell will abolish PVY accumulation. Presumably, the low percentage of resistant plants which become PVY-infected in the above experiments is due to this sort of phenomenon.

It is well known that PVX carrying host-derived sequence inserts can silence expression of the corresponding gene in infected plants in a process known as virus-induced gene silencing (VIGS, Burton et al., 2000). If infectivity of PVY-0 on YY plants depended on silencing of eIF4E by pPVXeYW, it would be expected that pPVXeYY would also induce silencing of eIF4E in YY plants and thus allow PVY-0 infection. However, virus accumulation was never detected by RT-PCR or by DAS-ELISA assays in YY leaves inoculated with pPVXeYY and PVY-0 (Figure 4a,b). RT-PCR and sequence analysis of YY plants inoculated with pPVXeYW and PVY-0 confirmed the presence of eIF4E mRNA corresponding to the YW allele expressed from pPVXeYW (data not shown). Thus, it was concluded that PVX functioned as an expression vector rather than a silencing vector.


We reported the molecular characterization of the first natural recessive resistance gene against a plant virus and showed that the pvr2 locus of pepper corresponds to a gene encoding a protein of the eIF4E subfamily. This work, thus, underlines the key role of eukaryotic initiation factor 4E not only in the viral life cycle as suggested by the known interaction between eIF4E and VPg (Wittmann et al., 1997; Léonard et al., 2000; Schaad et al., 2000), but also in the resistance strategy developed by the host to combat potyvirus infection.

eIF4E is a component of the eIF4F complex and provides the 5′ cap-binding function during formation of translation initiation complexes on most eukaryotic mRNAs. In plant cells, this complex is composed of only two proteins, eIF4E and eIF4G (Browning, 1996). Another specific feature of plants is the existence of an additional cap-binding complex, eIF (iso) 4F, in which a second cap-binding protein (eIF (iso) 4E) binds with eIF (iso) 4G (Bailey-Serres, 1999). In Arabidopsis, different genes code for these isoforms; three genes code for proteins of the eIF4E subfamily and one for eIF (iso) 4E (Rodriguez et al., 1998). Southern hybridization experiments in pepper also show that eIF4E encoding genes belong to multigenic family.

Potyviral RNA differs from host mRNAs in that the 5′ cap structure is replaced by the 5′ covalently linked VPg. As noted above, the VPg has been shown to bind to eIF4E proteins in several plant–potyvirus systems (Wittmann et al., 1997; Léonard et al., 2000; Schaad et al., 2000). The VPg may intervene at several steps in the virus infection cycle, including viral RNA replication, and viral cell-to-cell and long-distance movement (Revers et al., 1999), but the biological significance of the interaction between eIF4E proteins and the VPg remains to be determined. The most likely hypothesis is that interaction is important for translation and/or replication of the viral genome. Thus, pvr2-resistant plants never developed visible symptoms when challenged with the virus, and virus accumulation in inoculated leaves could not be detected by ELISA, back-inoculation, or RT-PCR (data not shown). Furthermore, pvr2 resistance is active at the protoplast level (Deom et al., 1997) and following graft inoculation. We cannot rule out the possibilities that eIF4E–VPg interaction could also interfere with translation of host mRNAs by sequestration of translation initiation factor, as observed during infection of animal cells by adenovirus and poliovirus (Feigenblum and Schneider, 1993).

Only two amino acid substitutions in the eIF4E protein determine the resistance phenotype against PVY: a substitution common to both resistant genotypes and two genotype-specific substitutions. The three substitutions do not involve highly conserved amino acids among the eukaryotes, nor amino acids implicated in the recognition of the mRNA 5′ cap structure or in recognition of eIF4G and 4E binding proteins (Figure 2), suggesting that the key functions of eIF4E are not disrupted. However, a direct effect on the structure of protein is possible because both YY mutations involve replacement of an amino acid containing a non-polar side chain with an amino acid having an acidic or basic side chain. Thanks to the high conservation of the eIF4E sequence among eukaryotes and to the availability of 3D structures of eIF4E from mouse and yeast (Marcotrigiano et al., 1997; Matsuo et al., 1997), the positions of the three substitutions can be predicted to be physically close and at the surface of the protein. The two alleles display different levels of resistance towards potyvirus strains and potyviruses (both pvr21 and pvr22 control the PVY pathotype 0, but pvr22 also controls resistance to PVY-1 and to TEV). Moreover, the YW allele confers susceptibility to PVY-0 in both YY and F, and the YY allele confers susceptibility to PVY-1 in F, matching the specificity of the pvr2 alleles. Thus, it can be hypothesized that the genotype-specific substitutions are correlated with the resistance phenotype conferred by the locus. Further structure–function analysis should provide insight into the importance of the amino acid differences detected between the different pvr2 alleles.

An interesting possibility, in accordance with the generally accepted hypothesis about the nature of recessive resistance genes against viruses, is that the eIF4E amino acid substitutions in YY and F abolish or decrease the affinity between the pepper eIF4E protein and the VPg of PVY, leading to resistance. The recent finding that PVY-containing amino acid substitutions in the VPg can overcome the pvr21 and pvr22 alleles from pepper and the pot-1 resistance gene from tomato corroborates this hypothesis (Moury et al., submitted). pot-1 confers a recessive resistance against PVY and TEV in tomato and was shown to map in a colinear genomic region of the pepper genome, suggesting that resistance to the same pathogen could be attributed to evolutionarily related loci (Parrella et al., 2002). These results, together with the recent demonstration that eIF(iso)4E is required for susceptibility to TuMV, TEV and lettuce mosaic virus (LMV) in Arabidopsis (Lellis et al., 2002; Duprat et al.,2002), indicate that eIF4E-mediated resistance occurs in several other plant–potyvirus interactions. This conclusion is strengthened by the unusually high frequency of recessive resistance genes against potyviruses in plants and by the key role of the central domain of the VPg in overcoming several unrelated host resistance genes from distinct plant families. Finally, it is of interest to note that the eIF4E and eIF (iso) 4E isoforms are functionally distinct in their interaction with the mRNA 5′ cap structure (Gallie and Browning, 2001) and in their ability to promote viral genome amplification. Thus, a protein from the eIF4E subfamily is involved in PVY resistance of pepper (this work) whereas a protein from the eIF (iso) 4E subfamily promotes TuMV, TEV and LMV amplification in Arabidopsis (Duprat et al.,2002). In conclusion, the demonstration that pvr2 corresponds to an isoform of eIF4E is a major step in understanding plant–potyvirus interactions; it should contribute to the elucidation of the molecular mechanisms underlying plant susceptibility to potyviruses, the frequent occurrence of recessive resistance and the main features of this class of natural resistance genes.

Experimental procedures

Plants, virus isolates and disease resistance scoring

C. annuum cultivars used were Yolo Wonder (YW), Doux Long Des Landes (DDL), Yolo Y (YY) and FloridaVR2 (F). Both YW (from California) and DDL (from southern France) contain the pvr2+ dominant allele for susceptibility to PVY and are unrelated cultivars; YY contains the pvr21 recessive allele for resistance to PVY pathotype 0 and F contains the pvr22 allele for resistance to PVY pathotypes 0 and 1 (Kyle and Palloix, 1997). A back-cross progeny (BC1) of 440 plants originating from the F1 hybrid (YW × YY) back-crossed with YY was developed to search for map co-segregation between the candidate gene eIF4E and the pvr2 locus. The parental lines, the F1 hybrid and the 440 BC1 plants were evaluated under growth chamber conditions for resistance to PVY isolate LYE72 (pathotype 0). Inoculum and mechanical inoculation procedures were as described previously (Caranta and Palloix, 1996). Thirty days post-inoculation (dpi), plants without obvious symptoms were evaluated for presence/absence of PVY coat protein by DAS-ELISA.

Genomic DNA extraction and mapping

Plant genomic DNA isolation and restriction fragment length polymorphism (RFLP) were performed as described previously (Lefebvre et al., 1995). Hybridization was done at 55°C. DraI, EcoRI, EcoRV, HindIII and XbaI(GIBCO/BRL, Life Technologies) were used to digest genomic DNA. The tobacco eIF4E cDNA probe used for RFLP mapping was kindly provided by D. Twell (University of Leicester, UK).

Isolation and analysis of the eIF4E cDNA from pepper

Total RNA was isolated from YW, DDL, YY and F leaf tissue using Tri-Reagent (Sigma). The full-length eIF4E cDNA of YW was obtained by rapid amplification of cDNA ends (RACE) using the GIBCO/BRL Life Technologies 3′ and 5′ RACE System (Version 2.0). The 3′ end of YW eIF4E cDNA was obtained using two degenerate oligonucleotides designed from an alignment of the tobacco (sequence from D. Twell) and tomato eIF4E cDNA (GenBank accession no. AF259801), in combination with the adapter primer (AUAP) of the kit. The full-length YW eIF4E cDNA was amplified with oligonucleotides designed from the sequence of the 3′ RACE product. All amplifications were performed with High Fidelity Platinum Taq polymerase (GIBCO/BRL, Life Technologies). The RACE products were cloned into the pGEM-T easy vector (Promega) and 10 independent 5′ and 3′ terminal clones were sequenced by Genome Express (Grenoble, France). Two primers based on sequence in the 5′ and 3′ NTR regions of the YW eIF4E cDNA were used in RT-PCR experiments to obtain the full length eIF4E cDNA of DDL, YY and F. All primer sequences are available from the authors on request. Genetics Computer Group (Madison, WI, USA) software package version 10.3 was used for protein and nucleic acid sequence analysis. The pepper eIF4E cDNA sequences have been deposited in the GenBank database (accession nos. AY122052, AF521963, AF521964, AF521965).

Identification and analysis of the BAC clones

A doubled haploid line, HD208, containing the pvr2+dominant allele of YW was the source of genomic DNA for the library construction. High molecular weight DNA was separately digested with EcoRI, BamHI and HindIII(GIBCO/BRL, Life Technologies) and cloned into the BAC vector pCUGIBAC1. The library consisted of 239,232 clones with an average insert size of 125 kb, representing approximately ten haploid genome equivalents (Ruffel, S., Caranta, C. and Bendahmane, A., unpublished). Two oligonucleotides, designed from the full-length eIF4E cDNA sequence of YW (P1: 5′-AGACTTTCATTGTTTCAAGCATAA-3′ and P2: 5′-GATTAGAAAGTGCAAACACCAATAC-3′) were used to screen the pepper BAC library as described (Kanyuka et al., 1999). Maxipreps of four candidate BAC clones were performed as described (http://www.plpa.agri.umn.edu/∼neviny/labsite/protocolsf/bacmini.html). For isolation of the 7-kb fragment, 30 μl of BAC DNA was digested using EcoRI (GIBCO/BRL, LifeTechnologies). Following electrophoresis, fragments of approximately 6–8 kb were eluted and cloned into pGEM3Zf+ (Promega) digested with EcoRI and dephosphorylated. Positive clones were selected by PCR using primers P1 and P2.

PVX-mediated transient expression assay

Primers corresponding to the 5′ and 3′ ends of the YW eIF4E ORF and containing ClaI and XhoI sites were used to amplify the YW and YY eIF4E ORFs. The digested PCR products were transferred to ClaI- and SalI-digested pPVX201 (kindly provided by DC Baulcombe, Sainsbury Laboratory, UK, Baulcombe et al., 1995) to produce plasmids pPVXeYW and pPVXeYY. The plasmids were manually inoculated to carborundum-dusted leaves of N. benthamiana (2.5 μg of DNA per leaf). Ten dpi, the inoculated N. benthamiana leaves were used as inoculum source for transient expression assays in pepper. First, YW, YY and F leaves were manually inoculated with the N. benthamiana leaf extract. Ten days later, the same leaves were manually inoculated with PVY-LYE72 (pathotype 0) or PVY-CAA16 (pathotype 1). In each experiment, pepper genotypes with specific behavior against PVY pathotypes were inoculated to check that no contamination occurred. Stability of the PVX constructs was checked by RT-PCR with primers sited on both sides of the cloning site (5′-CCGATCTCAAGCCACTCTCCG-3′ and 5′-CCTGAAGCTGTGGCAGGAGTTG-3′). PVY accumulation in inoculated pepper leaves was determined 10 dpi by DAS-ELISA using PVY capsid antigen and RT-PCR with two primers corresponding to the PVY capsid sequence (5′-CATCGATTATGGCAAATGATACAATTGATGC-3′ and 5′-TGTCGACATTCACATGTTCTTGACTCC-3′).


We thank K. Richards and H. Lecoq for very helpful comments on the manuscript, T. Desnos for help in RACE methodology and M.L. Lesage, G. Nemouchi, T. Phally and A.M. Daubèze for their excellent assistance. We are very grateful to David Twell for providing the unpublished tobacco eIF4E sequence. This work was supported by grants from GENOPLANTE and the French Ministry of Agriculture and Fishing (CI-1999-019 and CI-2001-005). SR was supported by a doctoral fellowship from the "Provence-Alpes Côte d'Azur Region" and Clause S.A. Experiments were carried out in compliance with current French guidelines concerning genetically modified organisms.