Introduction

Bacterial spot is among the most significant diseases limiting tomato production throughout tropical and subtropical regions (Jones et al. 2005). Attempts to control the disease have included seed decontamination as well as the application of bactericides such as streptomycin and copper-based compounds. However, streptomycin has been ineffective since the 1960s, and copper resistance has also become nearly ubiquitous among pathogen strains in the field (Obradovic et al. 2008; Stall et al. 2009). While new strategies employing systemic acquired resistance and biocontrol have shown promise (Obradovic et al. 2008), most recent efforts are focused on genetic resistance as a sustainable path to the control of bacterial spot disease.

The causative agent of bacterial spot of both tomato and pepper was once considered a single species, Xanthomonas campestris pv. vesicatoria. Jones and colleagues (2004) used DNA:DNA hybridization to reveal four distinct species: X. euvesicatoria, X. vesicatoria, X. gardneri, and X. perforans, which differ in their distribution, metabolic properties, and effector repertoires (Potnis et al. 2011). In the southeastern United States, where nearly half of all fresh market tomatoes in the USA are grown (Glades Crop Care 1999) and where bacterial spot disease is a major problem due to warm, humid conditions, X. euvesicatoria was long the predominant species. Since the mid-1990s, however, it has been replaced by X. perforans (Jones et al. 1998); the competitive advantage of Xp is likely due to its production of bacteriocins that antagonize Xe strains (Hert et al. 2005). The virulence of xanthomonads and other bacterial pathogens depends on a type III secretion system, which delivers an arsenal of effector proteins directly into plant cells. These effectors function to suppress components of the plant immune system. Some plants, however, have evolved resistance proteins that can directly or indirectly recognize pathogen effectors and activate a more robust defense response known as effector-triggered immunity; this recognition is specific, in that a given resistance protein is triggered by only one or two particular effectors (Jones and Dangl 2006).

Cultivated tomato (Solanum lycopersicum) lacks effective genetic resistance to bacterial spot disease, and breeding efforts have employed related wild species as sources of resistance. These have included both quantitative resistance, which is present in the accession S. lycopersicum var. cerasiformae PI 114490 (Hutton et al. 2010), and effector-triggered immunity conferred by single dominant loci, such as Xv3, found in the breeding line S. lycopersicum Hawaii 7981 and in at least two S. pimpinellifolium accessions (Wang et al. 2011). Unfortunately, Xv3 resistance cannot be effectively used in the field in Florida, since many pathogen strains isolated over the past decade do not contain the recognized avrXv3 effector gene. This may be due to selective pressure conferred by grape tomatoes, which likely originated from S. pimpinellifolium (Stall et al. 2009). The avrBs2 effector gene, however, is highly conserved among diverse xanthomonads (Kearney and Staskawicz 1990) and has been shown to play a role in virulence on tomato (Zhao et al. 2011). The Bs2 resistance gene from the pepper species Capsicum chacoense is effective against bacterial spot in transgenic tomato plants (Tai et al. 1999).

A durable disease-resistance strategy will ideally combine quantitative sources of resistance with effector-triggered immunity. A key approach to achieving durability will be the use of multiple resistance genes recognizing conserved pathogen effectors that play a role in virulence, in order to minimize the possibility for pathogen strains to mutate the recognized effector with no cost to fitness. Previous work has shown that the Xp effector XopJ4 [formerly designated AvrXv4 (Potnis et al. 2011)], a putative SUMO protease, plays a role in pathogen virulence (Roden et al. 2004). A preliminary survey of Xp field isolates from throughout the state of Florida suggests that the xopJ4 effector gene is conserved in the pathogen, making it a promising target for durable resistance. The XopJ4 effector is recognized by a dominant resistance locus, RXopJ4, in the wild accession Solanum pennellii LA716. The locus was previously referred to as Xv4 and was reported to map to a ~20-cM segment on chromosome 3 (Astua-Monge et al. 2000). Since the taxonomy of the Xanthomonas genus continues to evolve, and there is overlap among type III effector repertoires of different species, we have chosen to base the name of this resistance locus, RXopJ4, on the recognized pathogen effector (XopJ4), rather than on the pathogen species.

In this study, we sought to further refine the mapping of the RXopJ4 locus to identify cosegregating markers that would aid in the cloning of the resistance gene or genes. Initially, using a collection of introgression lines of S. pennellii in S. lycopersicum, we determined that the resistance locus lies within a 4.2-Mb region on the long arm of chromosome 6. We developed or adapted 14 molecular markers across this region and used these for fine mapping on an F2 population. Analysis of recombinant F3 plants showed that a 190-kb genomic region cosegregates with resistance. Our markers can be used for the detection of resistant recombinants and further fine mapping, and for the identification of candidate genes with the eventual goal of durable resistance to bacterial spot disease of tomato.

Materials and methods

Plant materials and DNA extraction

Solanum pennellii LA716 was obtained from Charles M. Rick at UC Davis. Jay W. Scott at the Gulf Coast Research and Education Center of the University of Florida provided the S. lycopersicum line FL8000, and the M82 line was received from Roger T. Chetelat at UC Davis. S. pennellii introgression lines were obtained from the C.M. Rick Tomato Genetics Resource Center at UC Davis, except IL 6-1, which was provided by Zachary Lippman at Cold Spring Harbor Laboratory. The RXopJ4 8000 OC7 resistant parent was produced from an initial cross between Sp LA716 and Sl H7998 (Astua-Monge et al. 2000), followed by five generations of outcrossing resistant F2 plants to Sl FL216 (a seventh backcross line from the cross between Sl FL7060 and S. pimpinellifolium PI 128216 containing Xv3 resistance), and then two outcrosses to FL8000, which also contains Xv3 resistance. The eighth F1 generation and F2 mapping population were generated from another outcross to FL8000; 910 plants from this F2 population were used for fine mapping of the RXopJ4 resistance locus. Seeds were sown directly in soil and allowed to germinate and grow for 2 weeks in a growth chamber at 24 °C with 12 h light/dark cycle. Seedlings were then transferred to a greenhouse with supplemental lighting.

For DNA extraction, ~60 mg young leaf tissue was collected per plant. The extraction protocol was adapted from one provided by Lauren Headland at UC Davis. Briefly, tissue was frozen at −80 °C overnight. Tissue was then disrupted in a BeadBeater (BioSpec Products, Bartlesville, OK, USA) with a 3-mm glass bead and extraction buffer containing 200 mM Tris–HCl (pH 7.5), 250 mM NaCl, 25 mM EDTA, and 14 μg/ml RNase A. Samples were then treated with SDS followed by extraction with phenol:chloroform:isoamyl alcohol (25:24:1). After phase extraction, DNA was precipitated with isopropanol, washed with 70 % ethanol, and resuspended in 1 mM Tris pH 8.

Bacterial strains, inoculum preparation and inoculation

Bacterial strains used in this study are described in Table 1. The Xp 4BΔxopJ4 strain was constructed using the suicide vector pLVC18 containing an upstream fragment spanning 1,054 bp of the xopJ4 promoter plus 33 bp of its open reading frame (ORF), and a 1,108-bp downstream fragment beginning at 7 bp past the stop codon; a BamHI restriction site was introduced between the two fragments. A double homologous recombination event resulted in deletion of the xopJ4 ORF, which was confirmed by PCR and Sanger sequencing. The 4BΔxopJ4 deletion strain was complemented by selecting for a single crossover with pLVC18 containing 846 bp of the xopJ4 promoter plus the ORF.

Table 1 Bacterial strains used in this study
Full size table

For inoculation, bacteria were grown on nutrient yeast glycerol agar (NYGA) supplemented, as appropriate, with 100 μg/ml rifampicin, 25 μg/ml kanamycin, and 10 μg/ml tetracycline, and were incubated at 28 °C for 48–72 h. Cells were then washed from agar plates with 1 mM MgCl2 and concentration was adjusted as necessary. Inoculation was performed by leaf infiltration using a 1-ml needleless syringe. For HR, two sections of each leaflet were infiltrated with suspensions of Xe 69-1 containing pLAFR6 or pLAFR6 xopJ4, both at 3 × 108 cfu/ml (OD600 = 0.3). For disease assays, whole leaflets were infiltrated with Xp 4B-WT, 4BΔxopJ4, or 4BΔxopJ4/comp at ~104 cfu/ml. For disease assays, inoculated plants were placed in a growth chamber at 24 °C with 12 h light/dark cycle. For HR assays, inoculated plants were placed either in a growth chamber or in a greenhouse. HR was assessed 24–48 h after inoculation, while disease was assessed 10–14 days after inoculation. For the growth assay, four to six 0.8-cm2 punches were taken from leaves of each genotype at each time point and homogenized in 1 mM MgCl2 in a BeadBeater; appropriate dilutions were plated on NYGA supplemented with 50 μg/ml rifampicin and 50 μg/ml cycloheximide (with 12.5 μg/ml kanamycin for the complemented mutant strain).

Marker development and PCR conditions

For mapping of the RXopJ4 resistance locus, we used a total of 14 molecular markers spanning the ~27 cM or 7.3 Mb from TG118 to CT204 (Table 2; Figs. 2, 5) on the long arm of chromosome 6. Thirteen of these markers are cleaved amplified polymorphic sequence (CAPS) markers, for which a short (<1,000 bp) sequence is amplified by polymerase chain reaction (PCR) and then digested with a restriction enzyme to yield product sizes that differ between the two parent genotypes—in this case, Sl and Sp (Fig. S2). One marker, SLM 6-15, is a simple sequence repeat (SSR) marker; it spans a region of tandem TA repeats that is 39 nucleotides longer in Sl than in Sp and can thus be visualized directly by agarose gel electrophoresis without restriction digestion (Fig. S2).

Table 2 Molecular marker information for mapping of the RXopJ4 resistance locus in Solanum pennellii LA716
Full size table

The markers TG118, CT83, C2_At1g21640, TG164, TG352, and CT204 were derived from RFLP or CAPS markers used in previous mapping studies (Fig. 5). Sequences, and in some cases CAPS protocols, for these markers were obtained from the Sol Genomics Network (Bombarely et al. 2011). Primers were optimized for PCR and CAPS visualization and are listed in Table 2. 06g060670 is derived from a predicted gene model in the International Tomato Annotation Group (ITAG) 2.3 genome annotation release (Sato et al. 2012). J350, J351, J352, J353, J366, and J385 markers were adapted from CAPS marker predictions based on preliminary Sl and Sp mRNA sequence information (manuscript in preparation). The SLM 6-15 SSR marker protocol was taken directly from Geethanjali et al. (2010). S. pennellii genome assemblies were provided by Anthony Bolger, Alisdair Fernie, and Björn Usadel at the Max Planck Institute for Molecular Plant Physiology, Golm, Germany.

Primers were designed using Amplify (Bill Engels, University of Wisconsin, USA) and CLC Main Workbench (CLC bio, Aarhus, Denmark) and were obtained from Integrated DNA Technologies (Coralville, IA, USA). PCR was conducted using Klentaq LA (DNA Polymerase Technology, St. Louis, MO, USA) according to the manufacturer’s protocol. Initial denaturation was at 94 °C for 5 min, followed by 35 cycles of 5 s at 94 °C, 30 s at 50–60 °C (Table 2), and 1 min at 68 °C, and a final extension at 68 °C for 7 min. Restriction digestion was carried out for 2 h at the appropriate temperature with enzymes and buffers purchased from New England Biolabs (Ipswich, MA, USA). PCR and restriction digestion were performed in a programmable thermal controller (PTC-100; MJ Research, Inc., Watertown, MA, USA). PCR/restriction products were separated on 3 % (SLM 6-15 and J351) or 2.5 % agarose gels (all other markers) and stained with ethidium bromide (SLM 6-15 and J351) or GelRed (Phenix Research Products, Candler, NC, USA) (all other markers; GelRed stain was included in the loading dye).

Results

Phenotypic analysis of XopJ4-dependent hypersensitive response

In order to determine the chromosomal location of the RXopJ4 resistance locus from S. pennellii, we employed a collection of 50 introgression lines (ILs) from the cross between S. pennellii LA716 and S. lycopersicum M82 (Eshed and Zamir 1994). Each line contains a homozygous introgression of S. pennellii over one segment of one chromosome, while the rest of its genome is S. lycopersicum; introgressed segments are overlapping. Each IL was inoculated with the compatible strain Xe 69-1 containing the empty vector pLAFR6 or pLAFR6 xopJ4 (Fig. 1). IL 6-2 and its sub-line IL 6-2-2 exhibited a XopJ4-dependent hypersensitive response (HR), a rapid cell death associated with pathogen resistance (Jones and Dangl 2006). However, IL 6-1, which contains a S. pennellii introgression that partially overlaps with those of ILs 6-2 and 6-2-2 (Fig. 2), did not exhibit HR. Furthermore, ILs 3-3 and 3-4, which Astua-Monge et al. (2000) identified as containing the RXopJ4 locus, did not exhibit XopJ4-dependent HR. The XopJ4-dependent HR of ILs 6-2 and 6-2-2, but not IL 6-1, indicated that the RXopJ4 locus was within a 4.2-Mb region on the long arm of chromosome 6 (Fig. 2).

Fig. 1
figure 1

The effector XopJ4 elicits a hypersensitive response in the S. pennellii introgression lines (ILs) 6-2 and 6-2-2 and RXopJ4 8000 OC7. Plants were inoculated with the compatible strain Xe 69-1 carrying pLAFR6 empty vector (EV) or pLAFR6 xopJ4 at a concentration of ~3 × 108 cfu/ml. Photographs were taken at 24 h post inoculation

Full size image
Fig. 2
figure 2

Initial mapping of the RXopJ4 resistance locus using S. pennellii introgression lines (ILs), showing molecular marker genotypes and phenotypes of parent lines and ILs. White bars indicate the marker genotype of S. lycopersicum M82 and black bars indicate S. pennellii LA716. Distances between markers are not drawn to scale but are indicated in megabases (Mb). A black bar indicates the 4.2-Mb region containing the RXopJ4 locus

Full size image
Fig. 3
figure 3

Fine mapping of the RXopJ4 resistance locus using an FL8000 × RXopJ4 8000 OC7 F2 population, showing molecular marker genotypes and phenotypes of parent lines and recombinant individuals. White bars indicate the marker genotype of S. lycopersicum FL8000 homozygotes and black bars indicate S. pennellii LA716 homozygotes. Recombinant phenotypes were assessed in homozygous individuals from the F3 generation. Distances between markers are not drawn to scale but are indicated in megabases (Mb). A black bar indicates the 190-kb region containing the RXopJ4 locus

Full size image

Cosegregation of HR and resistance with Sp introgression in an F2 mapping population

The ILs 6-2 and 6-2-2 proved undesirable as resistant parents of an F2 mapping population due to a linkage drag that conferred poor fertility and germination, small leaves, and necrosis of leaves upon exposure to cold or low light intensity, as well as an autogenous necrosis as the plants grew older. Instead, a resistant seventh outcross line from crosses between various commercial tomato lines and Sp LA716 was used (see “Materials and methods”). This line displayed a XopJ4-dependent HR (Fig. 1) but was missing part of the S. pennellii introgression present in ILs 6-1, 6-2, and 6-2-2 (Figs. 2, 3) and, importantly, also lacked the necrotic phenotype of ILs 6-2 and 6-2-2. A mapping F2 population was generated from a cross between FL8000 and this outcrossed line (designated RXopJ4 8000 OC7). To assess the segregation of HR in this mapping population, we inoculated a collection of 17 F2 plants with Xe 69-1 containing pLAFR6 or pLAFR6 xopJ4, and found that XopJ4-dependent HR segregated as a single locus (Fig. S1). Plants homozygous for Sp LA716 at the RXopJ4 locus exhibited strong HR, while plants homozygous for FL8000 did not exhibit HR. Heterozygous plants occasionally displayed a strong HR, but most produced weak or no HR; only some heterozygous plants showed disease resistance (data not shown), suggesting that RXopJ4 resistance is semi-dominant.

To verify the disease resistance of the RXopJ4 8000 OC7 parent line, we conducted a bacterial growth assay on this line, the susceptible parent FL8000, and Sp LA716. In order to assess the XopJ4 dependence of the resistance, we inoculated each line with the bacterial strains Xp 4B-WT, 4BΔxopJ4, and 4BΔxopJ4/comp. Sp LA716 and RXopJ4 8000 OC7 restricted the growth of Xp 4B-WT (Fig. 4). This resistance was dependent on XopJ4 recognition, since these lines were susceptible to the 4BΔxopJ4 deletion strain; complementation of this strain restored recognition by the resistant lines. The Sl FL8000 line, however, was susceptible to Xp 4B-WT, 4BΔxopJ4, and 4BΔxopJ4/comp strains. The overall lower levels of bacterial growth seen in Sp LA716 are likely due to factors other than the RXopJ4 resistance.

Fig. 4
figure 4

Solanum pennellii LA716 and RXopJ4 8000 OC7 display XopJ4-dependent resistance to X. perforans infection. Whole leaflets of Sl FL8000, Sp LA716, or RXopJ4 8000 OC7 plants were infiltrated with Xp 4B-WT (light gray bars), 4BΔxopJ4 (white bars), or 4BΔxopJ4/comp (dark gray bars). Bacteria were syringe-infiltrated into leaves at 104 cfu/ml. Bacterial counts were determined on the day of infiltration (Day 0) and 9 days post infiltration (Day 9). Error bars indicate the standard deviation from the mean of 4–6 samples

Full size image

Mapping of the RXopJ4 resistance locus

Initially, F2 plants from a cross between Sl FL7060 and IL 6-2-2 were scored at the genetic markers TG352, CT83, TG164, and CT204. This analysis revealed that the marker positions were not as previously mapped; for instance, among 347 F2 individuals, no recombinants were found between TG352 and TG164, which are 12.9 cM apart on the Tomato-EXPEN 1992 map (Fig. 5a). This could be partially explained by a suppressed homeologous recombination rate between FL7060 and the relatively short Sp introgression in IL 6-2-2 (Canady et al. 2006), but could also be due to the difficulty of accurately genotyping with RFLP markers. To remedy the discrepancy between the 1992 linkage map and the observed recombination rate, and to avoid the possible suppressed recombination resulting from a short Sp introgression, we mapped five CAPS markers—TG118, CT83, C2_At1g21640, TG352, and TG164—on a Sl M82 × Sp LA716 F2 population. This revealed, in particular, a location for the marker TG352 that differed from the EXPEN-1992 and 2000 linkage maps (Fig. 5b). At the same time, a fully assembled version of the tomato genome sequence became available (Bombarely et al. 2011; Sato et al. 2012). This provided a more accurate picture of the distances between markers, and also showed that TG164 and TG352 were in the opposite order from the M82 × Sp LA716 linkage map—this order was based on a single recombinant (Fig. 5b, c).

Fig. 5
figure 5

Maps of the RXopJ4 region on tomato chromosome 6. Marker names are on the top with map positions and distances in centimorgans (cM, a, b) or megabases (Mb, c) between each marker on the bottom. Map distances are not to scale. a Linkage maps based on F2 individuals from a cross between S. lycopersicum cv. VF36 and S. pennellii LA716. The Tomato-EXPEN 1992 map is based on RFLP markers and 67 F2 individuals, while the 2000 map is based on RFLP and CAPS markers and 80 F2 individuals; the 2000 map also includes conserved ortholog set (COS) markers such as C2_At1g21640 (Tanksley et al. 1992; Fulton et al. 2002). b Linkage map from this study based on CAPS markers and 62 F2 individuals from a cross between S. lycopersicum cv. M82 and S. pennellii LA716. c Sequence map based on tomato genome sequence release SL2.40 (Sato et al. 2012)

Full size image

Subsequently, the markers CT83, TG352, and J385 were scored on 910 individuals from an FL8000 × RXopJ4 8000 OC7 F2 population; 27 recombinant plants were identified—a recombination rate of ~3 % across the 4.8 Mb spanned by these three markers. All F2 plants were recombinant only on one copy of chromosome 6, and thus were heterozygous on one side of the recombination breakpoint. Since, as previously discussed, the resistance phenotype was inconsistent in heterozygotes, the F3 progeny of these plants were collected and homozygous recombinants were scored for both HR and disease resistance. The genotypes and phenotypes of key recombinant individuals are shown in Fig. 3. The left boundary of the RXopJ4 resistance locus is defined by the recombinants 5C3 and 18B10, which have a crossover from the Sp genotype to the Sl genotype between J350 and 06g060670, and are susceptible. The right boundary is defined by the recombinant 60B2, which has a crossover from the Sp to the Sl genotype between J351 and J352 and is resistant. The Sp introgressions in these recombinants indicate that the RXopJ4 locus lies within a 190-kb region between the markers J350 and J352.

Discussion

In the present study, we have defined the RXopJ4 resistance locus from S. pennellii LA716 to a 190-kb segment on the long arm of chromosome 6. We first narrowed the resistance to a 4.2-Mb region within the Sp introgressions of ILs 6-2 and 6-2-2. Then, we used a combination of previous linkage maps, our own linkage map of the RXopJ4 region, and new molecular markers to map the resistance on a segregating FL8000 × RXopJ4 8000 OC7 F2 population; this enabled fine mapping to a region of 190 kb. Although this resistance locus was originally reported to map to chromosome 3 (Astua-Monge et al. 2000), upon re-characterization of the S. pennellii ILs, we were unable to reproduce this result (Fig. 1).

The emergence of an annotated genome sequence for S. lycopersicum and preliminary genome and mRNA sequence data for S. pennellii greatly facilitated the development of precise molecular markers for fine mapping. Recently, the mapping of both a bacterial spot resistance quantitative trait locus (QTL) from S. lycopersicum var. cerasiformae PI 114490 (Hutton et al. 2010) and a major resistance gene, Rx4, from S. pimpinellifolium PI 128216 (Pei et al. 2011) was also aided by genome sequence data. The sequencing of more Solanum accessions and the identification of SNPs from cultivated varieties will continue to accelerate the process of marker development and thus the mapping of disease resistance and other desirable traits in the future.

Linkage drag is a common problem resulting from the introgression of disease resistance into crop plants by conventional breeding, and has been characterized by Lewis and colleagues (2007) in backcrossed tobacco lines containing the N gene, which confers resistance to tobacco mosaic virus. Notably, transgenic lines expressing the N gene exhibited significant increases in yield and cash return compared to backcrossed resistant lines. Similarly, the S. pennellii LA716 introgression lines 6-2 and 6-2-2 contain the RXopJ4 resistance locus, but suffer from a linkage drag that confers low fruit yield, small fruit, and an autogenous leaf necrosis. A seventh outcross resistant line showed improved traits and was suitable for mapping of the RXopJ4 locus, but may still have a disadvantage in the field when compared to the parent line FL8000. Transgenic approaches to disease resistance may mitigate the effects of linkage drag seen in backcrossed lines; alternatively, because of barriers to the acceptance of transgenic crops, backcrossed resistant lines can be developed that contain a minimal genomic region from the wild accession. We are currently working to identify such a recombinant individual containing RXopJ4 resistance.

Another common feature of disease-resistance genes is incomplete dominance. This has been observed for the Xv3 and Rx4 resistance genes, which were identified in different accessions but map to the same region of chromosome 11. Based on allelism tests that failed to identify susceptible F2 plants from a cross between Xv3 and Rx4 resistant parents, the genes are either closely linked or are alleles of the same gene (Wang et al. 2011). While Xv3 heterozygous plants developed HR more slowly than homozygous resistant plants (Wang et al. 2011), Rx4 heterozygotes showed similar disease ratings to homozygous plants under field conditions (Robbins et al. 2009). RXopJ4 heterozygotes exhibited slow, weak, or nonexistent HR, and also showed inconsistent phenotypes in disease assays. This could be problematic in the field, since hybrids are usually grown for their increased vigor. However, the heterozygous resistance phenotype may be improved in transgenic plants or in a different genetic background; moreover, the additive effect of multiple resistance genes in a single line may outweigh any reduced resistance due to heterozygosity.

Our current efforts are focused on finding new resistant recombinant plants from the F2 population and employing both the tomato genome sequence and preliminary S. pennellii sequence to identify and test candidate resistance genes within the region we have defined here. Once the RXopJ4 gene has been identified, it can be combined with Bs2, perhaps in an elite line containing the bacterial spot resistance QTL identified by Hutton et al. (2010), to produce a line that we expect will possess durable resistance to X. perforans in the southeastern United States and other regions where similar strains predominate. Next-generation sequencing now makes it possible to survey pathogen strains from the field for their type III effector repertoires and other virulence components. This approach will aid the development of durable resistance, as the search for new sources of resistance can be targeted to resistance genes that recognize the most conserved pathogen effectors (Bart et al. 2012). Surveying pathogen populations from different geographic areas could also guide strategies for both broadly applicable and region-specific disease resistance. Genetic resistance to bacterial spot disease has heretofore met with limited success in the field. Now, though, the use of genome sequence data for both pathogen and host will enable targeted resistance through precise marker-assisted selection and gene identification. We anticipate that these developments will help to accelerate the process of breeding and gene discovery to produce durably resistant varieties.