The screening platform developed by De Kesel et al. (2020) was used to investigate potential IR-establishing effects of various CCOPEs. Rice cell suspension cultures (RCSCs) were treated with candidate solutions to evaluate the induction of a validated pattern-triggered immunity (PTI)-marker gene set. This in vitro transcriptional stimulation has been shown to be a useful proxy to unveil an IR-inducing capacity of assayed products (De Kesel et al., 2020). Extracts derived from pumpkin (Cucurbita moschata cv. Musquee de Provence; pCOPE), zucchini (C. pepo var. cylindrica; zCOPE) or melon (Cucumis melo var. cantalupensis; mCOPE) peelings were all found to be promising IR stimuli. Indeed, convening with the threshold set by De Kesel et al. (2020), at least four of the six screening genes were significantly upregulated in the RCSCs upon treatment with the evaluated CCOPEs (Figure 1).

Via rice-M. graminicola (Mg) and tomato–M. incognita (Mi) infection studies, it was examined whether CCOPEs could result in effective systemic protection in planta. In line with the in vitro results (Figure 1), foliar CCOPE treatment 1 day before nematode inoculation significantly enhanced the root resistance of rice and tomato against the root-knot nematodes under study (Figure 2). In both hosts, all three evaluated CCOPEs resulted in a significantly lower number of galls per root system when compared to mock-treated control plants. Additionally, the number of females were counted for the infected rice plants: both in pCOPE- and mCOPE-treated plants, this infection parameter was found to be significantly reduced (p = 1.89 × 10^–3 and p = 3.53 × 10^–4, respectively). Interestingly, tomato root lengths increased upon pCOPE or zCOPE treatment (Figure 2B). Rice root or shoot lengths, on the other hand, were not consistently affected upon CCOPE treatment (Figure 2A). However, upon investigation of individual repetitions and/or data obtained in other experiments, treatment with pCOPE or zCOPE was found to be sporadically associated with impaired growth of rice plants (Supplementary Figure 1).

Additional rice-Mg infection assays were done to evaluate the IR-inducing capacity of other CCOPEs. Although a significantly lowered susceptibility could also be observed upon treatment with CCOPEs extracted from peels of butternut pumpkin (C. moschata cv. Butternut) or cucumber (Cucumis sativus), moderate to severe phytotoxicity was induced by these extracts (Supplementary Figure 2). As mCOPE consistently led to significant protection in both crops, and never had a negative effect on plant growth or development, this extract was selected for a more profound investigation.

During the above-discussed infection experiments (Figure 2), plant roots were inoculated 1 day post foliar treatment. As such, potential contact between the applied extracts and the inoculated nematodes could be limited. Nevertheless, various other cucurbit extracts are known to possess nematicidal properties to Mi (Elbadri et al., 2008; Gad et al., 2018; Alam and El-Nuby, 2019). To evaluate the nematicidal activity of mCOPE, second-stage juveniles (J2s) of Mg and Mi were incubated for 6 and 24 h in this extract. Figure 3 illustrates that mCOPE exhibits obvious nematicidal properties. Indeed, near-complete mortality rates were observed upon mCOPE incubation, similarly as observed for the positive control Vertimec, a commercial formulation of the nematicidal compound abamectin (Pitterna et al., 2009; Li et al., 2018).

Considering the major importance of rice as staple crop over tomato and its well-annotated genome, the mode-of-action of mCOPE-IR was further investigated in this monocot species via mRNA sequencing. Root samples were investigated 1 and 4 days post foliar treatment (1 and 4 dpt, respectively), allowing an assessment of systemic effects. 1,306 and 788 genes were found to be induced or repressed at 1 dpt, respectively, whereas these numbers equaled 2,341 and 1,224 for the 4 dpt analysis. 741 and 321 genes were found to be consistently induced or repressed at both time points. Multiple pathways contributing to plant biotic defense responses were found to be induced (Table 1), confirming mCOPE’s IR-triggering capacity. Amongst others, induction of phytohormone signaling through ethylene (ET) and jasmonic acid (JA), as well as cell wall biosynthesis and reactive oxygen species (ROS) metabolism, were found to be induced in roots of mCOPE-treated rice plants (Table 1). Moreover, MapMan and gene ontology (GO)-enrichment analyses revealed that genes involved in “plant–pathogen interactions,” “response to stress,” and “biosynthesis of secondary metabolites” were significantly enriched among the mCOPE-induced gene set (Table 1). Per-gene transcription levels and complete outcomes of MapMan and GO-enrichment analyses can be found in Supplementary Tables 1, 2, respectively.

To independently validate the transcriptomic findings (Table 1), phytohormone, ROS and lignin levels were quantified in rice roots upon foliar mCOPE treatment. In correspondence with the mRNA sequencing experiment, this was done at 1 and 4 dpt. However, as systemic accumulation of ET (Singh et al., 2020b) and ROS (Huang et al., 2015) can occur relatively quickly upon IR stimulus treatment, the two cited metabolites were quantified additionally at 6 h post treatment (6 hpt). In line with the mRNA sequencing results (Table 1), ET levels were increased in roots of mCOPE treated plants at 6 hpt (p = 1.41 × 10^–3) (Figure 4). No other changes in phytohormone levels were detected, except for a significant reduction in abscisic acid (ABA) levels at 4 dpt (p = 2.66 × 10^–2) (Figure 4). Moreover, significantly lower levels of ROS were detected at 1 dpt (p = 9.66 × 10^–3; Figure 4F), while lignin quantification revealed a transient increase of this biomolecule at 1 dpt (p = 6.97 × 10^–4; Figure 4G). The latter two biochemical assessments were also in line with the transcriptomic data presented in Table 1.

As stated by Martínez-Medina et al. (2016), relevant ecological assessments are needed to evaluate fitness-related costs and benefits associated with IR establishment. This may be of particular interest when the IR phenotype under study is characterized by intense defense induction upon contact with the IR stimulus, while not (yet) being challenged by a pest or pathogen (De Kesel et al., 2021). Such direct effects were revealed by our transcriptome analysis (Table 1) and the subsequent biochemical validations (Figure 4). Therefore, growth and yield of rice plants was investigated upon lifelong biweekly mCOPE treatment. Despite the intense defense elicitation as unveiled by mRNA sequencing, no negative long-term effects were observed for repetitively treated plants (Figure 5). Indeed, plant height, number of tillers, number of panicles and seed yield were all unaffected upon comparison with mock-treated control plants. These findings concur with the growth data presented in Figure 2 and Supplementary Figure 1.

To find out whether mCOPE-triggered protection against root-knot nematodes could be agronomically relevant for rice and tomato cultivation, a comparison with commercially available nematicides or IR stimuli was performed. The garlic extract Nemguard^®, marketed for the control of root-knot nematodes specifically (Ladurner et al., 2014; Eder et al., 2021), was used as reference nematicide. The IR stimuli included in this study were Actigard^®, a commercial formulation of benzothiadiazole (BTH; Schurter et al., 1990) and FytoSave^®, a preparation of chito-oligosaccharides and oligogalacturonides (COS-OGA) complexes (van Aubel et al., 2014; Clinckemaillie et al., 2017). Both IR stimuli and/or their active ingredients had previously been shown to result in effective protection of rice against Mg (Nahar et al., 2011; Singh et al., 2019; Soares and Dias-Arieira, 2021). Furthermore, BTH-IR had also been shown to protect tomato against Mi (Melillo et al., 2014). The effect of COS-OGA-IR against Mi, on the other hand, had not yet been studied for tomato. However, van Aubel et al. (2016) reported that this crop is prone to IR-elicitation via COS-OGA, as reduced susceptibility to Leveillula taurica was observed under greenhouse conditions. Our results indicated that, under lab conditions, mCOPE worked as good, or even better, than the commercially available products (Figure 6). Indeed, with respect to rice, Actigard^®-IR was effective in protecting rice from Mg, while Nemguard^® did not affect the infectivity of this nematode. FytoSave^® only led to a near-significant reduction in the number of galls for this pathosystem (p = 5.65 × 10^–2). Moreover, both Actigard^® and FytoSave^® resulted in reduced rice shoot length, while mCOPE did not (Figure 6A). Concerning tomato, all tested compounds led to a significant reduction in nematode infection levels, while not affecting plant growth (Figure 6B).

To investigate the activity spectrum of mCOPE-IR, additional infection experiments were conducted using various pathosystems that included pests or pathogens with different life styles. The migratory nematode Pratylenchus zeae and the cyst nematode Heterodera schachtii were used to challenge rice and sugar beet (Beta vulgaris cv. Amarok), respectively. The former parasite induces cell death and necrotic lesions due to its feeding strategy (Moens and Perry, 2009), while the latter is characterized by a biotrophic life style (Habash et al., 2017). Resistance to Botrytis cinerea and Rhizoctonia solani – both being necrotrophic fungi (Williamson et al., 2007; Foley et al., 2016) – was investigated upon mCOPE treatment of strawberry (Fragaria × ananassa) and fodder beet (B. vulgaris cv. Brunium), respectively. Finally, tomato was challenged by Pseudomonas syringae, a bacterial pathogen with a predominantly biotrophic life style (Kraepiel and Barny, 2016). Effective mCOPE-IR could not be established against any of the studied necrotrophs (Figures 7A,C,D). On the other hand, next to lowering the infectivity of the biotrophs Mg and Mi (Abad et al., 2008; Htay et al., 2016) upon infection of rice and tomato, respectively (Figure 2), mCOPE pretreatment also led to protection in tomato when challenged by the biotrophic pathogen P. syringae (Figure 7E). As a result, H. schachtii was the only investigated parasite with a biotrophic life style whose infection levels were not affected upon mCOPE-IR establishment in the challenged host (Figure 7B). Although more research is needed, these data seem to indicate that mCOPE may have a predominant activity against biotrophs.

Cucurbitaceae COld Peeling Extracts (CCOPEs) were made by blending 100 g of fresh fruit peeling in 200 mL sodium phosphate buffer (0.1 M; pH = 6.5). Only for the in vitro studies on rice cell suspension cultures (RCSCs), a mCOPE concentration of 200 g per 200 mL was used (see section “In vitro Evaluation of IR Activation”). Peelings were blended until a homogeneous mixture was obtained, which was then filtered using Miracloth filtration material. CCOPE preparation and eventual short-term storage was carried out at 4°C. For the in vitro analysis (see section “In vitro Evaluation of IR Activation”) and the nematicidal assays (see section “Nematicidal Assays”), extracts were filter-sterilized to avoid contamination.

To establish IR, plants were foliarly treated with a watery solution of the IR stimulus under study. This was done 1 day before infection/inoculation. Only for the fodder beet-Rhizoctonia solani pathosystem, this procedure was not followed. Here, mCOPE treatment took place on the same day as fodder beet transplanting. All solutions were supplied with 0.02% (v/v) Tween20 as surfactant and sprayed via a fine mist until run-off. Via multiple application rounds with minimal time intervals, a total of 6.25 mL IR stimulus was applied per plant. Only for the infection assays conducted on strawberry plants, 15 mL CCOPE was used per plant. Control plants were treated with an equal volume of the buffer used for CCOPE preparation, also containing 0.02% (v/v) Tween20. When comparing the effectivity of mCOPE-IR in rice and tomato with commercial IR stimuli and/or nematicides, FytoSave^®, and Actigard^® and Nemguard^® were dissolved in the buffer used for CCOPE preparation and contained 0.02% (v/v) Tween20. Fytosave^® was used at the recommended dose of 0.5% (v/v), which corresponds to 12.5 g/L COS-OGA. To study the effects of Actigard^®, solutions with a concentration of 250 μM benzothiadiazole (BTH) were used. Unlike all other treatments, the nematicide Nemguard^® was administered as root drench. This was done 1 day before inoculation using a solution of 0.133 g/L and in accordance with the manufacturer’s instructions.

RCSCs were established and used for IR stimulus detection as described by De Kesel et al. (2020). First, rice seeds were sterilized subsequently for 5 and 30 min with 75% ethanol and a 5% HAZ TABS solution, respectively. After rinsing with sterile water, seeds were cultivated for 30 days on callus inducing medium (CIM) with pH = 5.8. Then, the obtained callus pieces were proliferated in liquid amino acid (AA) medium (pH = 5.8) using disposable CELLSTAR cell culture flasks. RCSCs were grown in the dark at 25°C whilst being shaken at 110 rpm. Every 5–7 days, cells were subcultivated in fresh AA medium. The composition of CIM and AA medium is listed in Supplementary Tables 3, 4, respectively.

At least 10 days after RCSC establishment and approximately 4 days after the most recent subcultivation, cells from multiple RCSCs were separated from the callus clumps using a 70 μm falcon cell strainer to be pooled. Per biological replicate, 37.5 mL cells were transferred to new cell culture flasks and 12.5 mL CCOPE was added hereto. The buffer used for CCOPE preparation was utilized to mock-treat the control RCSCs. After 4 h, cells were transferred to 50 mL tubes, centrifuged for 5 min at 4,000 × g and after removal of the supernatant snap-frozen in liquid nitrogen.

RNA extraction was done using TRI Reagent in combination with Direct-zol RNA purification columns. DNase-treatment was executed with DNase I and cDNA synthesis was done using the Maxima First Strand cDNA Synthesis Kit. RT-qPCR was performed with three technical replicates on two biologically independent replicates. A CFX Connect Real-Time PCR Detection System was used to execute the RT-qPCR cycles. Following PCR protocol was utilized: 10 min at 95°C, followed by 40 cycles of 25 s at 95°C, 25 s at 58°C, and 20 s at 72°C. Finally, a melting curve analysis was performed. Two reference genes were used to normalize for differences in cDNA input across all samples. All primer pairs are listed in Supplementary Table 5. Finally, statistical analysis was done using Rest2009 (Pfaffl, 2001).

Nematicidal effects were evaluated by incubating approximately 100 Mg or Mi J2s in 1 mL of the test solution. As negative control, the buffer used for CCOPE preparation was utilized. Vertimec, a commercial formulation of the strongly nematicidal compound abamectin (Li et al., 2018), was used as positive control in a concentration of 0.2% (v/v). Per time point, at least three biological replicates were included. Nematode mortality was evaluated after six and 24 h of incubation. Nematodes were considered dead when they were not moving upon contact with a small picking needle. After identification and removal of outliers using the IQR-method (Barbato et al., 2011), statistical differences were determined via a heteroscedastic two-sided t-test (p < 0.05).

Systemic transcriptomic alterations were studied via mRNA sequencing in roots of rice plants whose shoots had been treated 1 or 4 days earlier (1 and 4 dpt, respectively). Upon treatment, plants were 14 days old. Same-aged plants treated with the buffer used for CCOPE preparation were used as controls. For each condition, three biologically independent replicates were used, each consisting of at least four pooled root systems. Ground root material was used as input for RNA extraction. This was done using the RNeasy Plant Mini Kit according to the manufacturer’s protocol, with three additional sonication steps of 10 s each after addition of the RLT buffer.

The QuantSeq 3′ mRNA-Seq Library Prep Kit was used for RNA seq library preparation. Quality of the libraries was confirmed using an Agilent Bioanalyzer 2100 and used for sequencing on a NextSeq 500 Illumina sequencing platform. The samples were multiplexed to minimize lane effects. Via single end sequencing, reads of 76 nucleotides in length were generated. Unprocessed sequencing data can be retrieved at the repositories of NCBI (Geer et al., 2010) as BioProject PRJNA767540. Reads were trimmed with Trimmomatic (version 0.36) using following settings: ILLUMINACLIP:TruSeq3-SE.fa:3:30:10, SLIDINGWINDOW:5:20, MINLEN:20 (Bolger et al., 2014) and mapped against the O. sativa ssp. japonica reference genome (build MSU7.0) using STAR (version 2.5.2a) (Dobin et al., 2013). Only uniquely mapped reads were kept for further analysis. BAM files of multiplexed samples were merged using samtools (version 1.3). Count tables were generated by the “Summarize Overlaps” function in the Genomic Alignments R package (version 1.16.0) (Lawrence et al., 2013). Differential expression analysis was performed using the DESeq2 package (version 1.20) (Love et al., 2014). Relative gene expressions with an adjusted p-value < 0.05 were considered as differentially expressed.

Gene ontology (GO)-enrichment analyses were done using the g:Profiler tool on biit.cs.ut.ee/gprofiler/gost (Reimand et al., 2016). As such, GO annotations significantly overrepresented among the significantly differentially expressed genes could be identified (p < 0.05). MapMan analyses were executed to assess general pathway inductions (Thimm et al., 2004) via Wilcoxon signed rank tests using Benjamini and Hochberg corrected p-values (p < 0.05).

To quantify ethylene, foliar treatment with mCOPE was executed in order to collect roots 6 h post treatment (6 hpt), as well as at 1 and 4 dpt. Same-aged plants treated with the buffer used for CCOPE preparation were used as controls. Per biological replicate, four root systems were pooled. Per condition, eight replicates were analyzed. Roots were cut into small pieces and placed in glass vials, which were subsequently sealed to allow ET accumulation in the headspace. After 4 h of incubation at room temperature, the headspace was analyzed by gas chromatography with a flame ionization detector. Sampling was executed using a gastight syringe.

Abscisic acid (ABA), indole-3-acetic acid (IAA), jasmonic acid (JA) and salicylic acid (SA) levels were determined according to the protocol described by Haeck et al. (2018). After sampling, roots were ground in liquid nitrogen. This was followed by a cold solvent extraction, filtration and clean up. Ultimately, phytohormone levels were measured using an ultra-high-performance liquid chromatograph in combination with a tandem high resolution mass spectrometer. Five biological replicates were assayed, each consisting of a pool of at least four root systems.

After identification and removal of outliers using the IQR-method (Barbato et al., 2011), statistical differences were determined via a heteroscedastic two-sided t-test (p < 0.05).

Root ROS content were analyzed at 6 hpt, 1 dpt, and 4 dpt using the xylenol orange assay. Same-aged plants treated with the buffer used for CCOPE preparation were used as controls. Root material was ground in liquid nitrogen and 70–100 mg sample was transferred to an Eppendorf tube. Per mg biological matter, 10 μL of a 5% (m/v) trichloroacetic acid (TCA) solution was added. Next, samples were centrifuged for 10 min at 20,000 × g and at 4°C. Wells of a UV-transparent 96 well plate were filled with 100 μL buffer. This was a mixture of 100 volumes 100 mM sorbitol, 125 μM xylenol orange and 1% ethanol, and 1 volume 25 mM ferrous ammonium sulfate and 2.5 M sulfuric acid. Wells that were used as blanks contained a buffer without xylenol orange. Correspondingly, each sample was analyzed via both of the buffers. Hereto, 10 μL TCA extract was added to the wells. Finally, the plate was incubated in the dark for 30 min at room temperature and absorbance was measured at 560 nm using a Tecan Infinite F200 Pro machine. For the actual ROS measurement, two technical replicates were used, while for the blanks only one. Two replicate studies were conducted for this experiment. Per condition, eight biologically independent replicates were used. Each of them consisted out of three pooled root systems. After identification and removal of outliers using the IQR-method (Barbato et al., 2011), statistical differences were determined via a heteroscedastic two-sided t-test (p < 0.05).

Lignin levels in rice roots were quantified at 1 and 4 dpt using an acetyl bromide assay. Same-aged plants treated with the buffer used for CCOPE preparation were used as controls. Samples were first ground in liquid nitrogen. Then, they were incubated in 1 mL of a series of solvents at a specific temperature, followed by centrifugation for 3 min at 20,000 × g and removal of the supernatant. The solvent - temperature combinations were, in following order, water −98°C, 96% ethanol −76°C, chloroform −59°C, and acetone −54°C. Next, the samples were left to dry, weighed to assure a mass between 2 and 7 mg and dissolved in 40 μL glacial acetic acid per mg dried tissue. The acetic acid contained 25% acetyl bromide and samples were incubated for 2 h at 50°C. Next, 1 mL glacial acetic acid was added to each sample, followed by centrifugation at 20,000 × g for 10 min. 300 μL of the resulting samples was mixed with 300 μL 2 M sodium hydroxide and 300 μL 0.5 M hydroxylamine-hydrochloride. Finally, the absorption of this mixture was measured at 280 nm using a Tecan Infinite F200 Pro machine, using three technical replicates per sample. Per treatment, eight biological replicates were used, each consisting out of four pooled root systems. After identification and removal of outliers using the IQR-method (Barbato et al., 2011), statistical differences were determined via a heteroscedastic two-sided t-test (p < 0.05).