Six different tomato (Solanum lycopersicum) genotypes were used in this study. Four of them are wild-type cultivars: C32 (susceptible), KG52201 (susceptible), 90E402F (I-1), and E779 (I-3) (Kroon and Elgersma, 1993; Simons et al., 1998; Mes et al., 1999; Scott and Jones, 2004), three are transgenic lines; KG324 and KG325 are I-2-containing derivatives of cv. KG52201, and MM+I-7 is a derivative of cv. Moneymaker (Simons et al., 1998; Gonzalez-Cendales et al., 2016). All plants were cultivated in a controlled greenhouse environment, maintained at a temperature of 25°C, 65% relative humidity, and a 16-hour photoperiod. The fungal isolates Fol004 (race 1), Fol007 (race 2), and Fol029 (race 3) were used (Rep et al., 2005).

Sixteen-day-old tomato seedlings were grafted at the hypocotyl using Grafting Cassettes™ (GRA&GREEN Inc, Japan; https://www.gragreen.com/en). The seedlings remained in the grafting cassettes for three weeks inside a propagator to sustain high humidity permitting the formation of graft junctions. The roots of the grafted plants were covered with soil to stimulate root growth. During the first week, the propagators were covered with paper tissue to minimize evaporation. In the third week, propagator humidity was gradually reduced from 90% to 65% to match greenhouse conditions (Aalders et al., 2024).

Fungal isolates were inoculated from glycerol stocks and cultivated on Czapek’s Dox Agar (CDA) plates at 25°C for a minimum of seven days. From the plate, a single plug was transferred to a 250ml flask containing 100 ml of minimal nitrate medium (100mM KNO3, 3% sucrose, and 0,17% Yeast Nitrogen Base without amino acids or ammonia) and incubated for five days at 25°C, 150 rpm. Fungal spores were filtered through sterile miracloth (22-25μm pore size), pelleted for 10 min at 2000 rpm, and washed with sterile Milli-Q (MQ). Before inoculation tomato roots were washed with water. The concentration of the spore suspension, the extent of root trimming, and inoculation time depended on the age of the plant and the type of experiment. For the fungal recovery assay, the roots of ten-day-old seedlings were trimmed to ~1 cm length and inoculated with 10⁷ spores/ml suspension for two minutes. Untrimmed tomato roots of three-week-old grafted tomato plants were inoculated for five minutes with a 0.5x10⁷ spores/ml suspension. For xylem sap proteomics, the main root and lateral roots of four-week-old tomato plants were trimmed to ~1 cm and inoculated for five minutes with 0.5x10⁷ spores/ml suspension.

Three weeks post-inoculation (wpi) of grafted tomato plants, disease index (DI), and fresh weight (FW) were determined. Disease index was scored on a scale of 0-5 (0, no brown vascular bundles; 1, only browning of vascular bundles at the crown level; 2, less than half of the bundles are brown at cotyledon level; 3, at least half of the bundles are brown; 4, 3/4 of the bundles are brown; 5, dead plants) (de Lamo et al., 2018).

For the fungal recovery assay from grafted plants, stems were cut at the crown level and above the first true leaf node at three wpi. The cut stems were surface-sterilized with 70% ethanol for four minutes and rinsed with sterile MQ (de Lamo et al., 2020). From the sterilized stems, slices (~0,5 cm thick) were collected from the first true leaf node, cotyledon, and right below and above the grafting point. For the fungal recovery assay from MM+I-7 plants, the stems of 19-day-old tomato MM+I-7 seedlings were submerged in ethanol for 5 seconds, rinsed with sterile MQ, and cut into ~8 mm long fragments (van der Does et al., 2019). Both stem slices and stem fragments were placed on potato dextrose agar plates supplemented with penicillin (100μg/ml) and streptomycin (200μg/ml) and incubated for three to five days at 25°C. For quantifying colonization of stem slices, the following numbering was used; 1= fungal outgrowth from a stem slice right below the grafting point, 2 = right above the grafting point, 3= fungal outgrowth at the level of the cotyledon, and 4= first true leaf node. The surface area of fungal outgrowth from stem slices was quantified using Fiji (https://imagej.net/Fiji) plugin tool.

Xylem sap was collected one-week post inoculation of four-week-old tomato plants. Plants were generously watered eighteen and three hours before xylem sap collection. The plants were cut just below the second true leaf node, positioned horizontally, and attached to a polystyrene tube placed on ice. The stump exuded xylem sap for four to six hours. Five repetitions of Fol-inoculation and xylem sap collection were performed during subsequent weeks. Xylem sap from four to six plants per repetition was pooled and centrifuged at 800 x g for 10 minutes to eliminate potential spores and soil particles. Next, 12 ml of the pooled xylem sap was concentrated using Amicon Ultra-15 Centrifugal Filter Units (Millipore) at 2500 x g for 20-25 minutes. The concentrated samples were transferred to low-binding tubes (Protein LoBind microcentrifuge tubes, Eppendorf). Protein quantification was performed using a BCA kit (BCA1-1KT Sigma Aldrich). Samples were treated with TCEP and CAA at final concentrations of 10 mM and 30 mM, respectively, and incubated at 70°C for 30 minutes for a 1 step reduction and alkylation of disulfide bridges. Subsequently, samples were prepared for mass spectrometry analysis using the single-pot, solid-phase-enhanced sample preparation (SP³) protocol (Hughes et al., 2014), with modifications to optimize protein recovery for xylem sap proteins. In short, no detergents were added to the samples to enable optimal precipitation of the soluble xylem sap proteins, and the precipitation time was extended to 30 minutes at room temperature. Following two washes with 70% ethanol and a single wash using acetonitrile, beads were air-dried and resuspended in 100 mM ammonium bicarbonate (Sigma) after which trypsin was added at a protease-to-protein ratio of 1:50 (w/w) at 37°C. Following overnight digestion formic acid was added to achieve a final concentration of 1%, resulting in an approximate pH of 2. The samples were then placed on a magnetic separator device and the peptides were recovered for LC-MS analysis.

Resulting samples were separated by reversed phase chromatography using an Ultimate 3000 RSLCnano UHPLC system (Thermo Scientific, Germeringen, Germany). Peptide separation was performed on a 75 μm × 250 mm analytical column (C18, 1.6 μm particle size, Aurora, Ionopticks, Australia), which was maintained at a temperature of 50°C and operated at a flow rate of 400 nL/min with 3% solvent B for 3 minutes (solvent A: 0.1% formic acid in water, solvent B: 0.1% formic acid in acetonitrile, ULCMS-grade, Biosolve). Following this, a multi-stage gradient was applied, with 17% solvent B at 21 minutes, 25% solvent B at 29 minutes, 34% solvent B at 32 minutes, 99% solvent B at 33 minutes, kept at 99% solvent B till 40 minutes. The system was returned to initial conditions at 40.1 minutes and was held until 58 minutes for equilibration. The eluted peptides were electrosprayed by a captive spray source via the column-associated emitter and were analyzed by a TIMS-TOF Pro mass spectrometer (Bruker, Bremen, Germany). The instrument was operated in PASEF mode for standard proteomics acquisition, with quadrupole isolation widths set at 2 Th at 700 m/z and 3 Th at 800 m/z, and collision energy values varying between 20 and 59 eV across the TIMS scan range. Precursor ions possessing an m/z range from 100 to 1700 and a TIMS range from 0.6 to 1.6 Vs/cm^2 were chosen for fragmentation. PASEF MS/MS scans were initiated ten times with a total cycle time of 1.16 seconds, a target intensity of 2e4, an intensity threshold of 2.5e3, and a charge state range of 0-5. Active exclusion was enabled for a period of 0.4 minutes, with precursors being reevaluated if the ratio of the current intensity to the previous intensity exceeded 4.

LC-MS data were processed using MaxQuant software (version 1.6.10.43) using standard settings, i.e. trypsin/p as the enzyme allowing for 2 missed cleavages with carbamidomethylation at cysteine as a fixed modification and oxidation at methionine as a variable modification searching the proteome databases of Solanum lycopersicum (UP000004994) and Fol race 2 (UP000009097) from Uniprot (03-2022). Several discrepancies exist in the current tomato Ensembl genome, wherein a multi-exon gene has been erroneously assigned for the single-exon PR5x gene (gene ID: Solyc08g080620; protein ID: Q8LPU1) and NP24 (gene ID: Solyc08g080640; protein ID: P12670). Additionally, the N-terminal sequence of P23 (gene ID: Solyc08g080650; protein ID: Q01591) is lacking nine amino acids. In the interim, we manually included Q8LPU1, modified Q01591, and deleted the incorrect A0A3Q7HTH3 and A0A3Q7HVV0 entry. The fungal databases were improved by adding non-annotated sequences of Fol Avr proteins. Bioinformatics analyses of MaxQuant output were performed with Perseus 2.03. Principle component analyses labeled one mock- and one Fol-inoculated I-1 plant repetition as outliers, which were excluded from further investigations. Before statistical analysis contaminants, reversed hits, and proteins identified in only one of the five repetitions were excluded. Subsequently LFQ values were log2 transformed and missing values were imputed using the lowest value in the dataset (-9.32). Differential accumulation of proteins between Fol- and mock-inoculated plants was determined via a Student’s T-test (p-value) without restricting the p-value correction using FDR or Benjamini Hochberg analysis.

Fol induces wilt disease in susceptible tomato plants by extensively colonizing xylem vessels and subsequently spreading to the adjacent tissues in the root and stem (Gordon, 2017). In an incompatible interaction, Fol also colonizes above-ground tissue but remains confined within the vessels, without causing visible disease symptoms (Mes et al., 2000). To investigate the potential of a resistant rootstock in preventing Fol-induced disease and colonization of susceptible foliage, Fol disease assays were conducted using chimeric plants composed of Fol007 (race 2)-susceptible (Sus) (KG52201) and I-2 resistant (Res) tissue (KG325). It was hypothesized that resistance is xylem-based, allowing a resistant rootstock to confer resistance to a susceptible scion, despite the presence of the fungus in a susceptible scion. Four grafting combinations were employed (rootstock | scion): Sus | Sus, Res | Res, Res | Sus, and Sus | Res. Plants were grafted at the hypocotyl and three weeks after mock- or Fol007-inoculation, fresh weight and disease index at the cotyledon level were assessed. All mock-inoculated graft combinations per repetition exhibited similar fresh weights showing that the tissues are compatible (Figure 1A). As expected, Fol007-inoculated Sus | Sus grafts displayed a significant decrease in fresh weight in all three repetitions, with most of the chimeric plants showing severe vascular browning (DI=4) (Figures 1A, B). Similar results were obtained for Sus | Res grafts in the first and second repetition, showing that a susceptible rootstock results in severe disease symptom development in the resistant scion. In the third repetition, Fol-inoculated Sus | Res grafts did not display a significant decrease in fresh weight but showed clear vascular browning (DI=3) at the cotyledon level (resistant tissue). Conversely, the fresh weight of both Fol007-inoculated Res | Res and Res | Sus grafts resembled those of mock-inoculated grafts. The disease indexes of Res | Res and Res | Sus results were very similar to each other, with the majority exerting none or minor vascular browning (DI=0-2). Merged disease index scores show significant differences between Sus | Sus and Res | Res or Res | Sus, while no significance was found between Sus | Sus and Sus | Res. These observations demonstrate that a resistant rootstock prevents disease development in a susceptible scion.

The absence of disease development in Res | Sus plants can be caused by; 1) Fol007 proliferation being contained in the resistant rootstock, preventing colonization of the susceptible scion, or 2) Fol007 colonizes susceptible scions, but a mobile compound or signal from the resistant rootstock suppresses its ability to cause disease. To distinguish between these options the presence of Fol007 in the scion was examined. Stem pieces were excised at various plant levels (below and above the grafting point, at the cotyledon, and first true leaf node) and cultured on agar plates to monitor fungal outgrowth (Figure 2). Fol presence was confirmed above the grafting point in all grafting combinations in two out of three repetitions (Figure 2B). The amount of Fol outgrowth from stem slices collected below and above the grafting point was quantified from two independent repetitions (Figures 2C, D). A significantly lower amount of fungal growth was observed from stem slices collected from below the grafting point of Res | Sus stems compared to Sus | Sus, whereas Sus | Res showed a similar trend as Sus | Sus (Figure 2D). For stem slices collected from Res | Res, there appears to be less fungal outgrowth in comparison to Sus | Sus, although this difference is not statistically significant. There were no differences observed in the fungal outgrowth from stem slices collected above the grafting point. The pathogen reached the first true leaf node of all Sus | Sus and Sus | Res grafts in every repetition (Figure 2B). The stem colonization of Res | Res and Res | Sus grafts were very similar to each other, with 30-40%, 70-80%, and 75% being colonized till the first true leaf node in the first, second, and third repetition, respectively. Combining fungal stem colonization data from three repetitions revealed significant differences between Sus | Sus and Res | Res or Res | Sus, while no significant difference between Sus | Sus and Res | Sus. These data show that I-2 rootstocks do not prevent the colonization of susceptible scions by Fol but do reduce the colonization and prevent disease development. As the fungus is contained in the vessels of the susceptible scion we hypothesized that resistant roots signal immune responses to the shoot that may involve translocation of specific antifungal compounds via the xylem to restrict Fol proliferation and pathogenicity.

Avirulent Fol strains colonize the stems of resistant plants, and in I-2 plants Fol has been shown to colonize the xylem vessels (Mes et al., 2000; van der Does et al., 2019). To test whether the absence of disease symptom development is due to a lack of effector secretion by the fungus, the xylem sap collected from I-1, I-2, I-3, and I-7 lines infected with avirulent Fol was examined for the presence of Fol proteins (Supplementary Table 1). For xylem sap collection, four-week-old tomato plants were mock- and Fol-inoculated with the corresponding avirulent Fol race. Concentrated xylem sap samples were subjected to LC-MS analysis to quantify and identify fungal proteins. Secreted Into Xylem (SIX) proteins were found in all samples and were most abundant in the xylem sap collected from I-1, I-2, and I-7 lines. SIX1, also known as Avr3 (A0A0C4DI36) was identified as a Differentially Accumulating Protein (DAP) in I-1, I-2, and I-7 when compared to mock-inoculated plants (Figure 3). SIX3, or Avr2 (A0A0C4DI32), and SIX6 (A0A0C4DHX7) were identified as DAPs in I-2 and I-7 plants. SIX1 and SIX3 were also detected in I-3 xylem sap samples but only in one out of the five repetitions. These findings corroborate the presence of the fungus in the vasculature of resistant plants and show that it secretes effector proteins, albeit at varying abundances.

The difference in the SIX protein amounts in the sap likely reflects the amount of fungal biomass and/or the extent of stem colonization. Previous studies already reported different degrees of stem colonization of I-1, I-2, and I-3 plants by avirulent Fol strains (van der Does et al., 2019). However, it remains unknown whether in I-7 lines Fol race 3 (Fol029) can reach and colonize the stem, and if so, to what degree. To address this, ten-day-old resistant (MM+I-7) and susceptible (C32) tomato seedlings were mock- and Fol029-inoculated. At nine days post inoculation (dpi), stem fragments were harvested and fungal presence was assessed. I-7 plants were colonized by Fol up to 63% of their height, while susceptible plants were colonized almost fully (96%) (Supplementary Figure 1). These findings confirm the presence of Fol in the xylem system, secreting SIX proteins, albeit with varying abundance across resistance lines.

To identify proteins associated with Fol resistance, quantitative proteome analysis was conducted on the xylem sap of the resistant lines following mock- or Fol-inoculation. The availability of four structurally distinct R-receptors (I-1, I-2, I-3, or I-7) within a single pathosystem provides the opportunity to identify unique and shared proteins that differentially accumulate upon Fol inoculation. LC-MS analysis of concentrated xylem sap revealed on average 860 plant proteins per sample (Supplementary Table 1). A minority of the identified proteins were classified as DAPs, as depicted in the volcano plots: 4.7% in I-1, 3.3% in I-2, 2.8% in I-3, and 7% in I-7 (Figures 4A-D).

To identify shared and unique responses induced by the different R-proteins to Fol, the upregulated proteins in the DAP list were compared using an UpSet plot (Figure 5A). These comparative analyses revealed a unique and shared set of upregulated DAPs among I-1, I-2, I-3, and I-7. All four Fol-inoculated I-gene lines exhibited differential accumulation of four proteins: PR-5x (Q8LPU1), PR-P2 (P32045), and two glucan endo-1,3-β-D-glucosidases (A0A3Q7IKF2, A0A3Q7FVX4) (Figure 5B). The p-value and protein abundance ratio of these four DAPs vary between the resistant lines (Figures 4A-D). Apart from this common set of pathogenesis-related proteins, subsets were observed that were shared between two or three resistant lines. For example, Fol-inoculated I-1 and I-7 lines show differential accumulation of an endochitinase (A0A3Q7IHS3), PR-P23 (Q01591) and an aspartic protease (A0A3Q7HJX2); I-2 and I-3 both accumulate a serine carboxypeptidase-like protein (A0A3Q7FBF5) and a chitinase (A0A3Q7H377); I-2 and I-7 share an uncharacterized protein (A0A3Q7H8L4) and two endochitinases (A0A3Q7GMW3, A0A3Q7FJ40); I-3 and I-7 an endochitinase (Q05539) and a subtilisin-like protease, P69B (A0A3Q7HVI4); I-1, I-2, and I-7 a PR-6 (P04284) and a leucine-rich repeat-like protein (A0A3Q7ET47); and I-1, I-3 and I-7 an uncharacterized protein (A0A3Q7FD15). In addition, DAPS unique to a single resistant line were identified (Table 1). For instance; I-1 plants showed differential accumulation of two chitin-degrading enzymes, a peroxidase, and a xylose synthase; I-2 a glycoprotein, a redox protein, and a rapid alkalization factor; I-3 a pathogenesis-related 5 protein, a sugar-degrading enzyme, and a protease; I-7 two pathogenesis-related protein 1 isoforms, tobacco stress-induced protein 1, four complex carbohydrate degrading enzymes, and three proteases. The comparative analysis suggests that in addition to distinct immune responses against Fol, I-1, I-2, I-3, and I-7 lines accumulate a common set of four PR proteins as DAPs in the xylem sap of inoculated plants.