Skip to main content
NCBI home page
Molecular Plant Pathology logoLink to Molecular Plant Pathology
. 2026 May 29;27(6):e70275. doi: 10.1111/mpp.70275

The Role of Two Glycoalkaloid Metabolism Genes in α‐Tomatine Biosynthesis and Basal Defence in Tomato

Yaohua You 1, Aishwarya Balaji 1, Andrea Lorena Herrera Valderrama 1, Marie‐Emma Denarié 1, H M Suraj 1, Miguel Ramirez Gaona 2, Katharina Hanika 2, Francel Verstappen 3, Iris F Kappers 3, Jan A L van Kan 1,✉
PMCID: PMC13239889  PMID: 42216500

ABSTRACT

Steroidal glycoalkaloids and saponins are plant cholesterol‐based steroid metabolites with antimicrobial activities and potential pharmacological value. The saponin uttroside B from black nightshade ( Solanum nigrum ) plays an important role in defence against herbivorous insects and exhibits anti‐hepatocellular carcinoma activity. The tomato ( Solanum lycopersicum ) glycoalkaloid α‐tomatine has been studied because of its antinutritional effects; however, its role in protecting plants from fungal pathogens remains understudied. The biosynthetic pathway of α‐tomatine involves multiple clustered genes designated as glycoalkaloid metabolism (GAME) genes. In this study, we generated single knockout mutants of SlGAME4 and SlGAME2 by CRISPR/Cas9‐based genome editing. The SlGAME4 mutants did not accumulate glycoalkaloids but instead redirected resources towards steroidal saponin (uttroside B) synthesis. SlGAME2 mutants contained unaltered α‐tomatine contents, indicating that the SlGAME2 gene, previously reported to catalyse the transfer of xylose to β1‐tomatine, is not involved in α‐tomatine biosynthesis. Infection assays with four fungal tomato pathogens demonstrated that the SlGAME4 mutant plants were slightly more susceptible to Botrytis cinerea, but equally susceptible to the other three fungi. Up‐regulation of α‐tomatine‐responsive genes in B. cinerea was observed during infection on SlGAME4 mutant tomato, as well as on S. nigrum , suggesting that uttroside B induces a fungal transcriptional response similar to α‐tomatine. Furthermore, we observed that tolerance mechanisms to plant saponins mediated by glycosyl hydrolases and a glycosyltransferase contribute to the virulence of B. cinerea on SlGAME4 mutant plants and S. nigrum . This indicates that uttroside B also contributes to defence against fungal pathogens and can be detoxified by B. cinerea .

We generated knockout mutants of two tomato genes proposed to be involved in α‐tomatine biosynthesis by genome editing. SlGAME2 mutants contained unaltered α‐tomatine contents; this gene is not involved in its synthesis. SlGAME4 mutants did not accumulate any glycoalkaloids but instead redirected resources towards steroidal saponin synthesis.

graphic file with name MPP-27-e70275-g003.jpg

1. Introduction

Tomato ( Solanum lycopersicum ) from the Solanaceae (nightshade) family is an economically important vegetable crop and serves as a rich source of nutrition worldwide. The yield of tomato can be affected by pathogenic fungi such as grey mould (Botrytis cinerea), leaf mould (Cladosporium fulvum), Verticillium wilt (Verticillium dahliae), early blight (Alternaria solani) or by oomycetes such as late blight ( Phytophthora infestans ), as well as by herbivorous insects (Arie et al. 2007; Blancard 2012; Panthee and Chen 2010; Nowicki et al. 2012). It highlights the necessity to study basal resistance traits of tomato, in particular its reservoir of endogenous antimicrobial metabolites with defensive roles (Bednarek 2012).

Steroidal glycoalkaloids (SGA) are a subgroup of saponins constitutively produced by plant species in the Solanaceae and Liliaceae families (Cárdenas et al. 2015). They not only possess antinutritional properties (bitterness) but are also considered phytoanticipins protecting the plants from attack by pathogens and herbivores due to their high concentration and broad‐spectrum antimicrobial as well as insecticidal activities (Sandrock and VanEtten 1998; Sun et al. 2021; Zhao et al. 2021). α‐tomatine is the major SGA in tomato and accumulates in vegetative tissues and green fruit to concentrations exceeding 1.5 mg/g fresh weight in leaves and 15 mg/g fresh weight in fruit (Kozukue et al. 2004). The biosynthetic pathway of SGAs starts from the precursor cholesterol and is mediated by enzymes encoded by GLYCOALKALOID METABOLISM (GAME) genes (Itkin et al. 2013). The GAME4 gene product, a cytochrome P450 protein, catalyses the first dedicated step from furostanol towards SGAs (Itkin et al. 2013). The intermediate alkaloid tomatidine, the aglycone of α‐tomatine, does not have antimicrobial activity but is toxic to plants (Ökmen et al. 2013). This phytotoxicity can be mitigated by four consecutive glycosylation steps (catalysed by GAME1, GAME17, GAME18 and GAME2), ultimately resulting in production of α‐tomatine (Itkin et al. 2013). During fruit ripening, α‐tomatine is exported from the vacuole to the cytosol by the tonoplast transporter GORKY and converted into a less bitter, non‐toxic SGA, named esculeoside A (Cárdenas et al. 2015; Kazachkova et al. 2021).

The toxicity of α‐tomatine to fungi is attributed to the disruption of fungal plasma membranes through complexing with 3β‐hydroxysterol (Steel and Drysdale 1988; You and van Kan 2021). Enzymatic detoxification of α‐tomatine has been reported in many tomato pathogens, and the most studied mechanism involves secreted glycosyl hydrolases (GH) referred to as ‘tomatinase’. Tomatinase activity in bacteria and fungi that are pathogenic on tomato has been reported for glycosyl hydrolases from three distinct families (GH10, GH3 and GH43) (You and van Kan 2021; You et al. 2024). Besides hydrolytic detoxification, B. cinerea also possesses multiple non‐degradative mechanisms for tolerance to α‐tomatine which are mediated by proteins involved in fungal membrane repair and modification (You et al. 2024).

The constitutive presence of antimicrobial metabolites makes an important contribution to basal plant defence (Osbourn 1996; Zaynab et al. 2021). The most compelling evidence comes from saponin‐deficient mutants of Avena strigosa (wild diploid oat) that exhibited compromised resistance to fungi that normally cannot infect oat, such as Gaeumannomyces graminis var. tritici, Fusarium culmorum and Fusarium avenaceum (Papadopoulou et al. 1999). In a study with 23 fungi, Sandrock and Vanetten (2001) observed that seven taxa that are non‐pathogenic on tomato were all sensitive to α‐tomatine, while 14 out of 16 tomato pathogens were tolerant to it. We recently reported that B. cinerea isolate M3a from grape is sensitive to α‐tomatine and could barely colonise tomato leaves, and that its virulence was enhanced by overexpression of genes that confer tolerance to α‐tomatine (You et al. 2024). Here we describe the generation of knockout mutants in the tomato SlGAME4 and SlGAME2 genes via CRISPR/Cas9 and report the effects of these deletions on saponin profiles and on interactions with fungal pathogens.

2. Results

2.1. CRISPR/Cas9‐Mediated Mutagenesis of SlGAME4 and SlGAME2 Genes in Tomato

We selected the SlGAME4 and SlGAME2 genes for mutagenesis for the following reasons. The SlGAME4 gene product catalyses the first dedicated step in the alkaloid biosynthetic pathway, using furastonol as substrate (Itkin et al. 2013; Grzech et al. 2025). Inactivating this gene would abolish synthesis of all SGAs. By contrast, the SlGAME2 gene encodes a glycosyl transferase that was reported to catalyse the final step in α‐tomatine synthesis, the transfer of xylose to β1‐tomatine (Itkin et al. 2013). Inactivating this gene would result in the accumulation of β1‐tomatine, which is non‐toxic to fungi (Quidde et al. 1998).

CRISPR/Cas9 genome editing was employed to generate single knockout (KO) mutants of SlGAME2 and SlGAME4 with four sgRNAs targeting the open reading frame of each gene, in the background of S. lycopersicum ‘MoneyMaker’ (MM), a genotype that is universally susceptible to most fungal, bacterial and viral pathogens. A high incidence of biallelic mutations was observed in the T0 generation of primary transformants: six out of seven SlGAME2‐KO lines and all six SlGAME4‐KO lines tested carried mutations in both alleles of the target genes. Deletions ranging from 1 to 1358 bp were detected in both SlGAME2 and SlGAME4 in the T0 generation and were inherited by T1 plants. Figure 1 illustrates the positions and sizes of deletions in each gene from four independent homozygous T1 mutants that were analysed in detail. In most cases, deletions were near to the target sequences of either of the sgRNAs. However, the SlGAME4‐KO line #6–7 contained a 617 bp deletion that starts 226 bp downstream of the predicted cleavage site of sgRNA1 and ends 229 bp downstream of the predicted cleavage site of sgRNA2.

FIGURE 1.

FIGURE 1

Open in a new tab

Scheme of CRISPR/Cas9‐mediated knockout of SlGAME2 (A) and SlGAME4 (B). Positions of small guide (sg) RNAs and their predicted cleavage sites are indicated by coordinates and red arrows, respectively. Deletions are illustrated by red dotted lines, the coordinates of the deleted nucleotides are provided, starting counting from the start codon.

2.2. Analysis of Tomatidine and α‐Tomatine Contents in Tissues of KO Plants

Relative concentrations of α‐tomatine and tomatidine were analysed in young and mature leaves, stems and roots sampled from SlGAME2‐KO, SlGAME4‐KO and wild‐type MM plants. α‐Tomatine and tomatidine were detected in trace amounts in SlGAME4‐KO plants (Figure 2), possibly by cross‐contamination from previous runs in LC‐QqQ‐MS. Strikingly, all four independent SlGAME2‐KO lines produced α‐tomatine at a concentration similar to the MM recipient (Figure 2), and β1‐tomatine was not detected in these plants. Among the tissues sampled, young leaves contained the highest concentration of α‐tomatine but not tomatidine (Figure 2A). Roots contained 10 times less α‐tomatine than young leaves but accumulated the highest content of tomatidine, with almost equal levels of α‐tomatine and tomatidine (Figure 2D). Stem tissues contained the lowest concentrations of both α‐tomatine and tomatidine (Figure 2C).

FIGURE 2.

FIGURE 2

Open in a new tab

Content of α‐tomatine and tomatidine in tissues of SlGAME2‐KO, SlGAME4‐KO and tomato cultivar MoneyMaker (MM) plants. (A) Young leaves; (B) mature leaves; (C) stems; (D) roots. Numbers underneath the columns indicate plant genotypes tested. Error bars are standard error of mean (SEM) of five biological replicates. Scale bars in the images indicate 2 cm.

We analysed the saponin profiles in SlGAME4‐KO plants in more detail using LC–MS2. Leaves from four separate SlGAME4‐KO mutant genotypes contained no detectable α‐tomatine but instead accumulated uttroside B, while wild‐type MM leaves contained α‐tomatine but no uttroside B (Figure S1). Fruit from MM and two independent SlGAME4‐KO mutant genotypes was analysed using green and red ripe fruit from three individual plants per genotype (Figure 3). MM green fruit contained a high level of α‐tomatine (compound 3), which in red ripe fruit was converted into esculeoside A (compound 1), as well established (Kazachkova et al. 2021). SlGAME4‐KO mutant fruit contained neither α‐tomatine nor esculeoside A but instead accumulated uttroside B (compound 4) and a compound annotated as uttroside B + pentose (compound 2). The ratio between compounds 4 and 2 decreased upon ripening in the GAME4‐KO #12–2 line, while it remained similar in the GAME4‐KO #25–2 line.

FIGURE 3.

FIGURE 3

Open in a new tab

Metabolite analysis in green and red ripe fruit from wild‐type tomato cultivar MoneyMaker (MM) and SlGAME4 mutant tomato. Samples are labelled on the right‐hand side of the graph. Dotted lines mark peaks of the following metabolites: 1. esculeoside A (m/z 1270.6, RT = 7.7 min); 2. uttroside B + pentose (m/z 1329.6, RT = 9.4 min); 3. α‐tomatine (m/z 1034.55, RT = 9.7 min); 4. uttroside B (m/z 1197.59, RT = 9.8 min).

2.3. Analysis of Expression Levels of α‐Tomatine Biosynthetic Genes in Tomato

We analysed the expression levels of several GAME genes, reported by Itkin et al. (2013), in young leaves, undeveloped leaves, mature leaves and stems in MM as well as the wild tomato Solanum habrochaites LYC4 (Figure S2). Remarkably, expression of GAME2 was barely detected in MM by reverse transcription‐quantitative PCR (RT‐qPCR). In LYC4, GAME2 transcription was mainly detected in young and undeveloped leaves.

2.4. Expression Profile of Fungal α‐Tomatine‐Responsive Genes During Plant Infection

Transcriptional up‐regulation of multiple α‐tomatine‐responsive genes was first demonstrated during in vitro growth of B. cinerea in the presence of α‐tomatine (You et al. 2024). The genes BcTom1 and BcGT28a play important roles in tomato infection and their expression was induced upon inoculation on tomato leaves, but not on Nicotiana benthamiana as this species does not accumulate α‐tomatine (You et al. 2024). We analysed the transcript levels of nine α‐tomatine‐responsive genes during infection on leaves of SlGAME4‐KO plants by RT‐qPCR. Despite the absence of α‐tomatine in this plant, up‐regulation of most of the tested genes was observed from 24 h post‐inoculation (hpi) onwards (Figure 4), which coincided with the onset of plant cell death induction and development of necrotic lesions.

FIGURE 4.

FIGURE 4

Open in a new tab

Relative expression of α‐tomatine‐responsive genes in Botrytis cinerea during infection on SlGAME4‐KO#6–7 leaves. Levels were quantified by reverse transcription‐quantitative PCR and normalised to BcTUBA (Bcin01g08040) and BcSMT3 (Bcin11g03430) transcripts. Error bars indicate standard error (SE) of three biological replicates. hpi, hours post‐inoculation.

2.5. Susceptibility of SlGAME4 ‐KO Plants to Different Tomato Pathogens

SlGAME4‐KO plants were inoculated with different pathogens including the necrotrophic fungi B. cinerea and A. solani, the soil‐borne vascular pathogen V. dahliae, and the biotrophic leaf mould Fulvia fulva. Three independent SlGAME4‐KO mutant plants showed a slight increase in susceptibility to B. cinerea (Figure 5A) but were not significantly altered in susceptibility to the other three fungi (Figure 5B–D).

FIGURE 5.

FIGURE 5

Open in a new tab

Susceptibility to fungal pathogens of SlGAME4‐KO mutant plants. Symptoms and lesion diameters of Botrytis cinerea (A) and Alternaria solani (B) after inoculation on leaves of SlGAME4‐KO were measured at 3 and 5 days post‐inoculation (dpi), respectively. (C) Symptoms and canopy areas (cm2) of tomato seedlings inoculated at the roots with water or with Verticillium dahliae strain JR2 at 15 dpi. Canopy areas were calculated with ImageJ, from photographs that were taken from above, using the diameter of the pots (14 cm) as reference. (D) Relative fungal biomass at 3, 6, 10 and 14 dpi as determined by the abundance ratio of Fulvia fulva DNA (based on fungal actin gene) versus tomato DNA (based on tomato actin gene). Error bars indicate the standard error (SE) of three biological replicates. The asterisks indicate significant differences between SlGAME4‐KO plants and wild‐type (WT) plants, determined by Student's t test (****p < 0.0001). ns indicates no significant difference.

2.6. Fungal Tolerance Mechanisms to α‐Tomatine Contribute to Virulence on SlGAME4 ‐KO

Inoculations were performed with B. cinerea isolates B05.10 and M3a on leaves of SlGAME4‐KO mutant plants. M3a was previously reported to infect MM poorly, due to the absence in its genome of the BcTom1 and Bcgt28a genes that confer tolerance to α‐tomatine (You et al. 2024). As on a wild‐type MM host, isolate M3a caused a low incidence of expanding lesions on SlGAME4‐KO plants that accumulate uttroside B instead of α‐tomatine (Figure 6A). We then studied the role of fungal tolerance mechanisms, mediated by α‐tomatine‐responsive genes, in the virulence on SlGAME4‐KO. A B05.10 mutant in which BcTom1 was deleted showed significantly reduced fungal virulence (Figure 6B). Conversely, the overexpression of three distinct types of tomatinase genes including BcTom1, SlTom1 and CfTom1 as well as the glycosyltransferase gene Bcgt28a promoted M3a infection on SlGAME4‐KO (Figure 6C–F).

FIGURE 6.

FIGURE 6

Open in a new tab

The role of fungal genes conferring tolerance mechanisms to α‐tomatine in virulence on leaves of SlGAME4‐KO. Disease incidence refers to the proportion of inoculation droplets that resulted in expanding lesions; diameters of expanding lesions were measured at 3 days post‐inoculation. (A) Disease incidence of B05.10 compared with M3a. (B) Lesion diameters of wild‐type B05.10 compared with two BcTom1‐KO mutants. (C–F) Disease incidence of wild type M3a compared with tomatinase gene overexpression transformants BcTom1‐OE (C) CfTom1‐OE (D) SlTom1 (E) and overexpression of glycosyltransferase gene BcGT28a (F). The asterisks indicate significant differences, as determined by Student's t test (**p < 0.01*, ***p < 0.001, ****p < 0.0001). Scale bar indicates 1 cm.

2.7. BcTom1 Contributes to Virulence of B. cinerea Isolate B05.10 on S. nigrum

Because uttroside B is the main saponin in S. nigrum (black nightshade), we characterised the importance of tolerance to α‐tomatine in the interaction between S. nigrum and B. cinerea . As a control, we analysed the interaction with potato ( Solanum tuberosum ) which also produces alkaloid saponins, however, with a different oligosaccharide chain. We first analysed the expression profile of α‐tomatine‐responsive genes after inoculation on leaves of S. nigrum and potato. Expression of the tested genes was strongly up‐regulated at 12 h after inoculation on S. nigrum (Figure 7A), whereas most genes exhibited unaltered transcript levels at any time point after inoculation on S. tuberosum (Figure S3). M3a formed significantly smaller lesions than B05.10 (Figure 7B), while the B05.10 mutant in which BcTom1 was deleted showed significantly reduced virulence on S. nigrum (Figure 7C).

FIGURE 7.

FIGURE 7

Open in a new tab

Characterisation of the interaction between Botrytis cinerea and Solanum nigrum . Relative expression of α‐tomatine‐responsive genes in B. cinerea during infection on leaves of S. nigrum (A). Lesion diameters of B05.10 compared with M3a (B) and with a BcTom1‐KO mutant (C) as measured at 3 days post‐inoculation. The asterisks indicate significant differences, as determined by Student's t test (**p < 0.01, ****p < 0.0001). Scale bar indicates 1 cm. hpi, hours post‐inoculation.

3. Discussion

In order to study the contribution of α‐tomatine to basal resistance of tomato to microbial pathogens, we generated knockout lines in two different genes, SlGAME4 and SlGAME2, which were reported by Itkin et al. (2013) to catalyse the first and the last dedicated steps of the α‐tomatine biosynthetic pathway. Four independent homozygous SlGAME4‐KO lines indeed lost the ability to produce α‐tomatine, as well as tomatidine. Unexpectedly, all four independent homozygous SlGAME2‐KO lines produced normal levels of α‐tomatine. Deletions in the coding sequence of SlGAME2 were close to the start codon and all caused a frameshift that should abolish the production of GAME2 protein in homozygous mutant lines. The observation that SlGAME2‐KO plants accumulated normal levels of α‐tomatine demonstrates that SlGAME2 is not essential for α‐tomatine biosynthesis in tomato. Itkin et al. (2013) described that SlGAME2 clusters in the tomato genome on chromosome 7 with three other glycosyltransferase genes (SlGAME1, SlGAME17, SlGAME18) that are reported to be responsible for the first three steps in the glycosylation cascade of tomatidine. The tomato SlGAME1/17/18/2 cluster is highly syntenic with the potato SGA biosynthetic cluster (Itkin et al. 2013). The SlGAME2 gene is orthologous to potato SGT3, which encodes an enzyme that catalyses the transfer of UDP‐rhamnose (a hexose) to generate the potato SGAs α‐solanine and α‐chaconine (Itkin et al. 2013). On the contrary, the tomato enzyme that completes α‐tomatine biosynthesis should use UDP‐xylose (a pentose) as the donor‐substrate and it seems unlikely that this could be achieved by the protein annotated by Itkin et al. (2013) as SlGAME2. Evidence of the existence of β1‐tomatine rhamnoside in tomato is lacking, despite numerous studies that have analysed SGA profiles in tomato. The fact that SlGAME2‐KO plants produce normal levels of α‐tomatine, do not produce any detectable β1‐tomatine, and that the SlGAME2 protein is orthologous to a potato rhamnosyltransferase thus questions the functional annotation of the SlGAME2 gene by Itkin et al. (2013). The tomato genome contains multiple glycosyltransferase genes, but none of these shows significant levels (> 50%) of protein sequence identity to SlGAME2. Furthermore, transcripts of SlGAME2 were only detected in young, developing vegetative tissues of tomato but were undetectable in adult expanded leaves, both by RT‐qPCR and RNA‐seq (Figure S2 and unpublished data). By contrast, transcript levels of the three genes that mediate the first steps of glycosylation of tomatidine (SlGAME1, SlGAME17, SlGAME18) are relatively high in all tissues. We thus conclude that SlGAME2 cannot be involved in α‐tomatine synthesis. Interestingly, Wolters et al. (2023) identified in Solanum commersonii a glycosyltransferase ScGTR2 that catalyses attachment of xylose to triose SGAs, generating dehydrodemissine, which can confer resistance to Colorado potato beetle and A. solani. A recent study by Gharat et al. (2026) showed that the tomato gene Solyc12g009930, ortholog of S. commersonii ScGTR2, indeed mediates the transfer of xylose to β1‐tomatine and thereby completes the synthesis of α‐tomatine. We hypothesise that knocking out the Solyc12g009930 gene in tomato would not only abolish the production of α‐tomatine but would also affect uttroside B synthesis, as these saponins share an identical tetrasaccharide moiety with a terminal xylose (Grzech et al. 2025). A failure to add the terminal xylose would result in a strong reduction of fungitoxicity of uttroside, analogous to the loss of toxicity that is exhibited by β1‐tomatine.

The observed up‐regulation of α‐tomatine‐responsive B. cinerea genes during infection on SlGAME4‐KO leaves, which do not contain α‐tomatine, suggested the presence of other inducers. The LC–MS analysis confirmed that SlGAME4‐KO plants accumulate the steroidal saponin uttroside B instead of SGAs (Figures 3 and S1). We previously reported that expression of α‐tomatine‐responsive genes can also be induced by the steroidal saponin digitonin from Digitalis purpurea , which contains a pentasaccharide moiety that structurally resembles the tetrasaccharide moiety of α‐tomatine and also possesses fungitoxic activity (You et al. 2024). The chemical identity between the tetrasaccharide units of α‐tomatine and uttroside B makes it plausible that uttroside B also serves as a target of fungal tomatinases and can also induce expression of α‐tomatine‐responsive genes in B. cinerea during infection. This assumption is supported by the observation that α‐tomatine‐responsive genes were transcriptionally induced during B. cinerea infection on SlGAME4‐KO mutant and on S. nigrum , which both accumulate uttroside B (Figures 4 and 7A). Moreover, tomatinase significantly contributed to the virulence of B. cinerea on both SlGAME4‐KO and S. nigrum leaves (Figures 6 and 7B).

A recent study by Boccia et al. (2024) showed that uttroside B plays an important role in plant defence against insect pests. In our study, the change of tomato saponin composition through inactivation of SlGAME4 slightly increased susceptibility to B. cinerea but did not compromise the basal resistance against the fungal pathogens A. solani , V. dahliae and F. fulva (Figure 5). Redirecting saponin biosynthesis towards uttroside B might compensate for the reduced basal defence in absence of α‐tomatine and thereby (partly) restore the protection against fungal pathogens. Tomato fruits are generally quite susceptible to fungal diseases because α‐tomatine levels strongly decrease during fruit ripening because of its conversion into the non‐bitter, non‐toxic esculeoside A (Bai et al. 2025; Kazachkova et al. 2021; Nakayasu et al. 2021; Sonawane et al. 2023). Our results also showed the disappearance of α‐tomatine in ripe MM tomato fruit, concomitant with the appearance of esculeoside A (Figure 3), which may have affected the basal defence in mature MM fruit. The presence of uttroside B in mature fruit of transgenic SlGAME4‐KO lines might decrease their susceptibility to pathogens during postharvest. Furthermore, the secretion of α‐tomatine from roots was reported to influence the tomato rhizosphere microbiome (Trivedi et al. 2020; Nakayasu et al. 2021). Therefore, it will be interesting to use SlGAME4‐KO plants in future studies for characterising effects of altered tomato saponin composition on its rhizosphere microbiome and on plant fitness.

4. Experimental Procedures

4.1. Generation of Knockout Lines in Tomato

The GAME mutant lines were generated using a CRISPR/Cas9 genome editing strategy as described in Hanika et al. (2021). In brief, four sgRNAs were designed to target each gene (Table S1). Transformation was carried out according to Huibers et al. (2013). T0 generations were screened and the lines carrying mutations in the coding sequences of target genes were used for seed production. Homozygous mutants were identified in T1 generation, and their seeds were used for large‐scale experiments.

4.2. Inoculation Assays

Tomato plants used for fungal inoculations were grown at 20°C during daytime (16 h) and 19°C at night (8 h) with 60% humidity. Leaves from 5‐ to 6‐week‐old tomato plants were used for inoculation. Inoculation with B. cinerea was performed as described by You et al. (2023). Spores were suspended at 1 × 106 spores/mL in 3 g/L Gamborg's B5 basal salt mixture supplemented with 15 mM sucrose and 10 mM potassium phosphate, adjusted to pH 6.0. Six 2 μL‐droplets were inoculated on one leaflet and lesion diameters were measured at 3 days post‐inoculation (dpi).

Alternaria solani strain (altNL03003/CBS 143772) was grown on potato dextrose agar (PDA) plates and spores were collected as described by Wolters et al. (2019). Spores were suspended at 1 × 105 spores/mL in potato dextrose broth (PDB, 12 g/L) supplemented with 0.3% agar. Six 10 μL‐droplets were inoculated on one leaflet and lesion diameters were measured at 5 dpi.

Fulvia fulva inoculation assays and fungal biomass quantification were performed as in Ökmen et al. (2013). Spores of F. fulva (race 0WU; CBS131901) were resuspended at 2 × 106 spores/mL in tap water. Leaves of 4‐ to 5‐week‐old plants were sprayed on the lower side with spore suspensions. Eight plants per line were inoculated with F. fulva , while two plants per line served as uninoculated control group. Leaf samples were taken at 3, 6, 10 and 14 dpi, freeze‐dried and used for DNA isolation. qPCR was performed to assess the ratio of plant DNA to fungal DNA, using primers for amplifying the fungal Ffactin gene or the plant Slactin gene (primers in Table S1).

Verticillium dahliae inoculation assay was performed as described in Santhanam et al. (2013). Spores from V. dahliae strain JR2 were resuspended at 1 × 106 spores/mL in demi water. Roots of 10‐day‐old tomato seedlings were dip‐inoculated in V. dahliae spore suspension. Eight to 10 plants per line were inoculated with JR2, and two to five plants per line served as uninoculated control group. Canopy areas were measured at 15 dpi.

4.3. Metabolite Analysis and Quantification of Glycoalkaloids by LC‐QqQ‐MS

α‐Tomatine and tomatidine concentrations were analysed in distinct tissues including roots, stems, young leaves, mature leaves, green or red ripe fruit in four or five biological replicates. Frozen tissue samples were freeze‐dried, ground into powder and 5 mg from each sample was used for extraction in methanol containing 0.1% formic acid using 15 min sonication. After centrifugation, supernatant was filtered and used for LC‐QqQ‐MS analysis as described in You et al. (2024).

4.4. Metabolite Analysis by UHPLC–HRMS

Metabolites were analysed by ultra‐high‐performance liquid chromatography‐high‐resolution mass spectrometry (UHPLC–HRMS). Analyses were performed using a Vanquish Horizon UHPLC system interfaced with an Exploris 120 Orbitrap mass spectrometer (Thermo Fisher Scientific). Frozen leaf and fruit tissue samples were freeze‐dried, ground into powder and 5 mg from each sample was used for extraction in methanol containing 0.1% formic acid using 15 min sonication. For each sample, 5 μL of extract was injected onto an Acquity UPLC BEH C18 column (1.7 μm particle size, 2.1 × 150 mm; Waters) maintained at 40°C. Chromatographic separation was conducted at a constant flow rate of 400 μL min−1 using a binary mobile phase system consisting of 0.1% (v/v) formic acid in water (mobile phase A) and 0.1% (v/v) formic acid in acetonitrile (mobile phase B). The gradient elution programme was as follows: isocratic 5% B from 0.0 to 1.0 min; linear increase to 75% B from 1.0 to 22.0 min; linear ramp to 90% B from 22.0 to 23.0 min; isocratic hold at 90% B from 23.0 to 26.0 min; linear ramp to 5% B from 26.0 to 27.0 min followed by re‐equilibration at 5% B from 27.0 to 30.0 min.

For compound putative annotation, samples were analysed using combined full‐scan MS and data‐dependent MS/MS acquisition (Full MS/dd‐MS2, top‐n), separately in negative‐ and positive‐ionisation modes. Full‐scan MS data were acquired over an m/z range of 90–1350 at a resolving power of 60,000 (m/Δm at m/z 200), an automated gain control (AGC) target of 1 × 106, and a maximum injection time of 100 ms. The ESI source parameters were set to a spray voltage of 3.0 kV (negative‐ionisation mode) and 2.5 kV (positive‐ionisation mode), and a capillary temperature of 290°C. Data‐dependent MS2 spectra were acquired for the four most intense precursor ions per scan at a resolving power of 15,000 (m/Δm), using an AGC target of 1 × 105, a maximum injection time of 118 ms, and an isolation window of 1.0 m/z. Fragmentation was achieved using stepped normalised collision energies (NCE) of 20, 40 and 100 eV (%) to generate diagnostic fragment ions supporting the annotation.

4.5. Gene Expression Analysis

Total RNA was extracted from different tissues of tomato, potato or S. nigrum using Maxwell 16 RNA Purification Kits (Promega). Relative expression of fungal genes was quantified by RT‐qPCR according to Qin et al. (2023), using primers described in Table S1. The transcript levels of two B. cinerea housekeeping genes BcTUBA (Bcin01g08040) and BcSMT3 (Bcin11g03430) were used for gene expression normalisation.

Author Contributions

Katharina Hanika: investigation. Yaohua You: conceptualisation, investigation, formal analysis, data curation, visualization, writing – original draft, writing – review and editing. Jan A. L. van Kan: conceptualisation, writing – review and editing, project administration, supervision. Marie‐Emma Denarié: investigation. Aishwarya Balaji: investigation. Miguel Ramirez Gaona: investigation. Iris F. Kappers: methodology, formal analysis. HM Suraj: investigation, formal analysis, visualization. Francel Verstappen: methodology. Andrea Lorena Herrera Valderrama: investigation, formal analysis.

Funding

This work was supported by the China Scholarship Council.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: LC–MS/MS analysis of leaves from wild‐type tomato cultivar MoneyMaker (A) and SlGAME4‐KO mutant genotypes (B). Standards for uttroside B and α‐tomatine are shown in the panels (C) and (D). The profile of MM is shown for one leaf and was representative for that of leaves from three other plants. The profile of the mutant shown here is for SlGAME4‐KO #6–7, and was representative for that of leaves from the independently generated mutant genotypes SlGAME4‐KO #8–1, SlGAME4‐KO #12–2 and SlGAME4‐KO #25–2.

MPP-27-e70275-s003.pdf (165.1KB, pdf)

Figure S2: Relative expression of GAME genes in different tissues from Solanum lycopersicum ‘MoneyMaker’ (MM) and Solanum habrochaites LYC4. Levels were quantified by reverse transcription‐quantitative PCR and normalised to the tomato house‐keeping genes CAC and GAPDH. Error bars are standard error (SE) of three biological replates.

MPP-27-e70275-s002.pdf (303.2KB, pdf)

Figure S3: Relative expression of α‐tomatine‐responsive genes in Botrytis cinerea during infection on leaves of Solanum tuberosum . Levels were quantified by reverse transcription‐quantitative PCR and normalised to BcTUBA (Bcin01g08040) and BcSMT3 (Bcin11g03430) transcripts. Error bars indicate standard error (SE) of three biological replicates.

MPP-27-e70275-s001.pdf (264.9KB, pdf)

Table S1: Primers used in this study.

MPP-27-e70275-s004.docx (33.2KB, docx)

Acknowledgements

The research of Yaohua You was financed by the China Scholarship Council. The authors acknowledge Bert Essenstam (Unifarm, Wageningen UR) for excellent transgenic plant care and Ioannis Thanos for performing a part of the RT‐qPCR analysis. We acknowledge the support by Dr. Carlos Sanchez Arcos and Bert Schipper (Wageningen UR) in performing LC–MS2 runs of plant samples, and by Dr. Marianna Boccia and Dr. Sarah O'Connor (Max Planck Institute for Chemical Ecology, Jena, Germany) who provided uttroside B as LC‐MS standard.

Data Availability Statement

Raw data are available from the corresponding author upon request.

References

  1. Arie, T. , Takahashi H., Kodama M., and Teraoka T.. 2007. “Tomato as a Model Plant for Plant–Pathogen Interactions.” Plant Biotechnology 24: 135–147. [Google Scholar]
  2. Bai, F. , Wu M., Huang W., et al. 2025. “Removal of Toxic Steroidal Glycoalkaloids and Bitterness in Tomato Is Controlled by a Complex Epigenetic and Genetic Network.” Science Advances 11: eads9601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bednarek, P. 2012. “Chemical Warfare or Modulators of Defence Responses—The Function of Secondary Metabolites in Plant Immunity.” Current Opinion in Plant Biology 15: 407–414. [DOI] [PubMed] [Google Scholar]
  4. Blancard, D. 2012. Tomato Diseases. Academic Press. [Google Scholar]
  5. Boccia, M. , Kessler D., Seibt W., et al. 2024. “A Scaffold Protein Manages the Biosynthesis of Steroidal Defense Metabolites in Plants.” Science 386: 1366–1372. [DOI] [PubMed] [Google Scholar]
  6. Cárdenas, P. D. , Sonawane P. D., Heinig U., Bocobza S. E., Burdman S., and Aharoni A.. 2015. “The Bitter Side of the Nightshades: Genomics Drives Discovery in Solanaceae Steroidal Alkaloid Metabolism.” Phytochemistry 113: 24–32. [DOI] [PubMed] [Google Scholar]
  7. Gharat, S. A. , Jozwiak A., Rogachev I., et al. 2026. “Complex Engineering of Solanum Alkaloids Structural Diversity in Nicotiana benthamiana .” Plant Biotechnology Journal 24: 328–345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Grzech, D. , Smit S. J., Alam R. M., et al. 2025. “Incorporation of Nitrogen in Antinutritional Solanum Alkaloid Biosynthesis.” Nature Chemical Biology 21: 131–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Hanika, K. , Schipper D., Chinnappa S., et al. 2021. “Impairment of Tomato WAT1 Enhances Resistance to Vascular Wilt Fungi Despite Severe Growth Defects.” Frontiers in Plant Science 12: 621674. 10.3389/fpls.2021.721674. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Huibers, R. P. , Loonen A. E., Gao D., Van den Ackerveken G., Visser R. G., and Bai Y.. 2013. “Powdery Mildew Resistance in Tomato by Impairment of SlPMR4 and SlDMR1 .” PLoS One 8: e67467. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Itkin, M. , Heining U., Tzfadia A. J., et al. 2013. “Biosynthesis of Antinutritional Alkaloids in Solanaceous Crops Is Mediated by Clustered Genes.” Science 341: 175–179. [DOI] [PubMed] [Google Scholar]
  12. Kazachkova, Y. , Zemach I., Panda S., et al. 2021. “The GORKY Glycoalkaloid Transporter Is Indispensable for Preventing Tomato Bitterness.” Nature Plants 7: 468–480. [DOI] [PubMed] [Google Scholar]
  13. Kozukue, N. , Han J. S., Lee K. R., and Friedman M.. 2004. “Dehydrotomatine and α‐Tomatine Content in Tomato Fruits and Vegetative Plant Tissues.” Journal of Agricultural and Food Chemistry 52: 2079–2083. [DOI] [PubMed] [Google Scholar]
  14. Nakayasu, M. , Ohno K., Takamatsu K., et al. 2021. “Tomato Roots Secrete Tomatine to Modulate the Bacterial Assemblage of the Rhizosphere.” Plant Physiology 186: 270–284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Nowicki, M. , Foolad M. R., Nowakowska M., and Kozik E. U.. 2012. “Potato and Tomato Late Blight Caused by Phytophthora infestans : An Overview of Pathology and Resistance Breeding.” Plant Disease 96: 4–17. [DOI] [PubMed] [Google Scholar]
  16. Ökmen, B. , Etalo D. W., Joosten M. H., et al. 2013. “Detoxification of α‐Tomatine by Cladosporium fulvum Is Required for Full Virulence on Tomato.” New Phytologist 198: 1203–1214. [DOI] [PubMed] [Google Scholar]
  17. Osbourn, A. E. 1996. “Saponins and Plant Defense—A Soap Story.” Trends in Plant Science 1: 4–9. [Google Scholar]
  18. Panthee, D. , and Chen F.. 2010. “Genomics of Fungal Disease Resistance in Tomato.” Current Genomics 11: 30–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Papadopoulou, K. , Melton R. E., Legget M., Leggett M., Daniels M. J., and Osbourn A. E.. 1999. “Compromised Disease Resistance in Saponin Deficient Plants.” Proceedings of the National Academy of Sciences of the United States of America 96: 12923–12928. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Qin, S. , Veloso J., Puccetti G., and van Kan J. A. L.. 2023. “Molecular Characterization of Cross‐Kingdom RNA Interference in Botrytis cinerea by Tomato Small RNAs.” Frontiers in Plant Science 14: 1107888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Quidde, T. , Osbourn A. E., and Tudzynski P.. 1998. “Detoxification of α‐Tomatine by Botrytis cinerea .” Physiological and Molecular Plant Pathology 52: 151–165. [Google Scholar]
  22. Sandrock, R. W. , and VanEtten H. D.. 1998. “Fungal Sensitivity to and Enzymatic Degradation of the Phytoanticipin α‐Tomatine Fungal Sensitivity to and Enzymatic Degradation of the Phytoanticipin α‐Tomatine.” Phytopathology 88: 137–143. [DOI] [PubMed] [Google Scholar]
  23. Sandrock, R. W. , and Vanetten H. D.. 2001. “The Relevance of Tomatinase Activity in Pathogens of Tomato: Disruption of the β2‐Tomatinase Gene in Colletotrichum coccodes and Septoria lycopersici and Heterologous Expression of the Septoria lycopersici β2−Tomatinase in Nectria haematococca, a Pathogen of Tomato Fruit.” Physiological and Molecular Plant Pathology 58: 159–171. [Google Scholar]
  24. Santhanam, P. , van Esse H. P., Albert I., Faino L., Nürnberger T., and Thomma B. P. H. J.. 2013. “Evidence for Functional Diversification Within a Fungal NEP1‐Like Protein Family.” Molecular Plant‐Microbe Interactions 26: 278–286. 10.1094/mpmi-09-12-0222-R. [DOI] [PubMed] [Google Scholar]
  25. Sonawane, P. D. , Gharat S. A., Jozwiak A., et al. 2023. “A BAHD‐Type Acyltransferase Concludes the Biosynthetic Pathway of Non‐Bitter Glycoalkaloids in Ripe Tomato Fruit.” Nature Communications 14: 4540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Steel, C. C. , and Drysdale R. B.. 1988. “Electrolyte Leakage From Plant and Fungal Tissues and Disruption of Liposomemembranes by α‐Tomatine.” Phytochemistry 27: 1025–1030. [Google Scholar]
  27. Sun, L. L. , Hou W. H., Zhang J. J., et al. 2021. “Plant Metabolites Drive Different Responses in Caterpillars of Two Closely Related Helicoverpa Species.” Frontiers in Physiology 12: 460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Trivedi, P. , Leach J. E., Tringe S. G., Sa T., and Singh B. K.. 2020. “Plant–Microbiome Interactions: From Community Assembly to Plant Health.” Nature Reviews Microbiology 18: 607–621. [DOI] [PubMed] [Google Scholar]
  29. Wolters, P. J. , de Vos L., Bijsterbosch G., et al. 2019. “A Rapid Method to Screen Wild Solanum for Resistance to Early Blight.” European Journal of Plant Pathology 154: 109–114. [Google Scholar]
  30. Wolters, P. J. , Wouters D., Tikunov Y. M., et al. 2023. “Tetraose Steroidal Glycoalkaloids From Potato Provide Resistance Against Alternaria solani and Colorado Potato Beetle.” eLife 12: RP87135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. You, Y. , Astudillo‐Estévez I., Essenstam B., Qin S., and van Kan J. A. L.. 2023. “Leaf Resistance to Botrytis cinerea in Wild Tomato Solanum habrochaites Depends on Inoculum Composition.” Frontiers in Plant Science 14: 1156804. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. You, Y. , Suraj H. M., Matz L., et al. 2024. “ Botrytis cinerea Combines Four Molecular Strategies to Tolerate Membrane‐Permeating Plant Compounds and to Increase Virulence.” Nature Communications 15: 6448. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. You, Y. , and van Kan J. A. L.. 2021. “Bitter and Sweet Make Tomato Hard to (B)Eat.” New Phytologist 230: 90–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Zaynab, M. , Sharif Y., Abbas S., et al. 2021. “Saponin Toxicity as Key Player in Plant Defense Against Pathogens.” Toxicon 193: 21–27. [DOI] [PubMed] [Google Scholar]
  35. Zhao, D. K. , Zhao Y., Chen S. Y., and Kennelly E. J.. 2021. “ Solanum Steroidal Glycoalkaloids: Structural Diversity, Biological Activities, and Biosynthesis.” Natural Product Reports 38: 1423–1444. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1: LC–MS/MS analysis of leaves from wild‐type tomato cultivar MoneyMaker (A) and SlGAME4‐KO mutant genotypes (B). Standards for uttroside B and α‐tomatine are shown in the panels (C) and (D). The profile of MM is shown for one leaf and was representative for that of leaves from three other plants. The profile of the mutant shown here is for SlGAME4‐KO #6–7, and was representative for that of leaves from the independently generated mutant genotypes SlGAME4‐KO #8–1, SlGAME4‐KO #12–2 and SlGAME4‐KO #25–2.

MPP-27-e70275-s003.pdf (165.1KB, pdf)

Figure S2: Relative expression of GAME genes in different tissues from Solanum lycopersicum ‘MoneyMaker’ (MM) and Solanum habrochaites LYC4. Levels were quantified by reverse transcription‐quantitative PCR and normalised to the tomato house‐keeping genes CAC and GAPDH. Error bars are standard error (SE) of three biological replates.

MPP-27-e70275-s002.pdf (303.2KB, pdf)

Figure S3: Relative expression of α‐tomatine‐responsive genes in Botrytis cinerea during infection on leaves of Solanum tuberosum . Levels were quantified by reverse transcription‐quantitative PCR and normalised to BcTUBA (Bcin01g08040) and BcSMT3 (Bcin11g03430) transcripts. Error bars indicate standard error (SE) of three biological replicates.

MPP-27-e70275-s001.pdf (264.9KB, pdf)

Table S1: Primers used in this study.

MPP-27-e70275-s004.docx (33.2KB, docx)

Data Availability Statement

Raw data are available from the corresponding author upon request.

Articles from Molecular Plant Pathology are provided here courtesy of Wiley

RESOURCES