Identification of I-7 expands the repertoire of genes for resistance to Fusarium wilt in tomato to three resistance gene classes

Abstract

The tomato I-3 and I-7 genes confer resistance to Fusarium oxysporum f. sp. lycopersici (Fol) race 3 and were introgressed into the cultivated tomato, Solanum lycopersicum, from the wild relative Solanum pennellii. I-3 has been identified previously on chromosome 7 and encodes an S-receptor-like kinase, but little is known about I-7. Molecular markers have been developed for the marker-assisted breeding of I-3, but none are available for I-7. We used an RNA-seq and single nucleotide polymorphism (SNP) analysis approach to map I-7 to a small introgression of S. pennellii DNA (c. 210 kb) on chromosome 8, and identified I-7 as a gene encoding a leucine-rich repeat receptor-like protein (LRR-RLP), thereby expanding the repertoire of resistance protein classes conferring resistance to Fol. Using an eds1 mutant of tomato, we showed that I-7, like many other LRR-RLPs conferring pathogen resistance in tomato, is EDS1 (Enhanced Disease Susceptibility 1) dependent. Using transgenic tomato plants carrying only the I-7 gene for Fol resistance, we found that I-7 also confers resistance to Fol races 1 and 2. Given that Fol race 1 carries Avr1, resistance to Fol race 1 indicates that I-7-mediated resistance, unlike I-2- or I-3-mediated resistance, is not suppressed by Avr1. This suggests that Avr1 is not a general suppressor of Fol resistance in tomato, leading us to hypothesize that Avr1 may be acting against an EDS1-independent pathway for resistance activation. The identification of I-7 has allowed us to develop molecular markers for marker-assisted breeding of both genes currently known to confer Fol race 3 resistance (I-3 and I-7). Given that I-7-mediated resistance is not suppressed by Avr1, I-7 may be a useful addition to I-3 in the tomato breeder's toolbox.

Keywords: Fusarium oxysporum f. sp. lycopersici; Fusarium wilt; Solanum lycopersicum; Solanum pennellii; leucine-rich repeat; plant disease resistance gene; receptor-like protein.

Figure 1

Plot of single nucleotide polymorphism (SNP) frequencies for Tristar and M82 transcripts relative to the tomato reference transcriptome for genes on the long arm of chromosome 8. The SNP cluster corresponding to the introgressed region of S olanum pennellii  DNA on Tristar chromosome 8 is highlighted in blue. SNP frequencies were calculated as the number of SNPs per transcript unique to Tristar or M82 divided by the length of the transcript.

Figure 2

(A) Disease assays on T2 plants carrying the S olyc08g077740 gene from Tristar in a Moneymaker (MM) or M82 background. T2 plants, together with Tristar (resistant), Moneymaker (susceptible) and M82 (susceptible) control plants, were inoculated with F ol race 3. Photographs were taken at 18 days post‐inoculation. (B) Disease scores for the T2 plants shown in (A). Probability values were obtained using the non‐parametric Kruskal–Wallis test to determine significant differences in disease scores between plant lines.

Figure 3

(A) Disease assays on T2 plants carrying the S olyc08g077740 gene from M82 in a Moneymaker (MM) background. T2 plants, together with Tristar (resistant) and Moneymaker (susceptible) control plants, were inoculated with F ol race 3. Photographs were taken at 21 days post‐inoculation. (B) Disease scores for the T2 plants shown in (A). Probability values were obtained using the non‐parametric Mann–Whitney test to determine significant differences in disease scores between plant lines.

Figure 4

(A) Disease assays on Tristar × sun1‐1 EDS 1/eds1 and eds1/eds1  F2 plants (segregating for I ‐7) and parent lines inoculated with F ol race 3. F2 plants were genotyped for I ‐7 by cleaved amplified polymorphic sequence (CAPS) marker analysis and plants carrying I ‐7 are indicated. Photographs were taken at 21 days post‐inoculation. (B) Disease scores for the T2 plants shown in (A), excluding F2 plants not carrying I ‐7. Probability values were obtained using the non‐parametric Mann–Whitney test to determine significant differences in disease scores between plant lines.

Figure 5

Sequence and graphic representation of the predicted I‐7 protein divided into eight domains: signal peptide, N‐terminal flanking region, typical extracellular leucine‐rich repeat (LRR) region (with LRRs numbered 1–30), divided into two blocks by a loop out region or island domain, C‐terminal LRR‐flanking region, acidic domain, transmembrane domain and basic cytosolic domain. The locations of predicted 3₁₀ (xxLxx), concave β1 (xxLxLxx) and convex β2 (xLxGx) motifs are shown in blue above the LRR domain. The positions of amino acid differences between I‐7 and i‐7^M82 are shown in red, whereas those between I‐7 and i‐7^LA716 are highlighted in yellow. Putative N‐glycosylation sites are underlined and a putative endocytosis motif is highlighted in green. A possible SOBIR1‐dimerization motif (GxxxGxxxG; Bi et al., 2015) is highlighted in light blue. Amino acids present in I‐7, but missing from i‐7^M82, owing to deletions in i‐7 relative to I ‐7, are double underlined, whereas those missing from i‐7^LA716 are overlined.

Figure 6

Alignment of I‐7 and orthologous and paralogous proteins from Tristar, Heinz 1706 and LA716. Sequences were aligned using the EMBL‐EBI MAFFT server at http://www.ebi.ac.uk/Tools/msa/mafft/, and the alignment was shaded using the ExPASy BoxShade server at http://www.ch.embnet.org/software/BOX_form.html. Amino acid identities are highlighted in black. Amino acid similarities are highlighted in grey. Putative N‐glycosylation sites encoded by orthologous and paralogous genes on chromosome 8 (labelled o8 and p8, respectively) and paralogous genes on chromosome 6 (labelled p6) are indicated by black boxes above the sequence alignment. Putative N‐glycosylation sites encoded by genes at only one or two of these three loci are indicated by red boxes (o8) or green boxes (p8) above the alignment or blue boxes (p6) below the alignment. Residues unique to I‐7 are highlighted in yellow. Numbers above the alignment indicate the leucine‐rich repeat (LRR) number, and asterisks indicate variable sites with three or more different residues, with red asterisks indicating those in predicted solvent‐exposed residues in the β‐strand motif of the LRR. It should be noted that the first 18 amino acids of the Trist_p8 sequence are inferred from the LA716_p8 sequence.

Figure 7

Alignment of N‐ and C‐terminal regions conserved between I‐7 and other tomato and tobacco leucine‐rich repeat receptor‐like proteins (LRR‐RLPs). Sequences were aligned using the EMBL‐EBI MAFFT server at http://www.ebi.ac.uk/Tools/msa/mafft/, and the alignment was shaded using the ExPASy BoxShade server at http://www.ch.embnet.org/software/BOX_form.html. Amino acid identities are highlighted in black. Amino acid similarities are highlighted in grey. In LRR regions, the consensus sequence is shown above the alignment, and amino acid identities for non‐consensus residues are highlighted in bright green and amino acid similarities for non‐consensus residues are highlighted in dark green. Conserved cysteine residues likely to be involved (full connecting lines) or potentially involved (broken connecting lines) in disulfide bonding, based on structural studies of other plant extracellular LRR proteins (Di Matteo et al., 2003; Hothorn et al., 2011), are highlighted in blue. Conserved N‐glycosylation motifs are highlighted in red. LRR‐RLP sequences were derived from GenBank accessions AAA65235 (Cf‐9), AAC15779 (Cf‐2), BAA88636 (EILP), ACR33106 (Ve1) and AAR28378 (Eix2).