1. Introduction

About 1% of angiosperms are parasitic plants, which evolved independently from 12 or 13 origins (Westwood et al., 2010; Nickrent, 2020). Parasitic plants use a special organ, haustorium, to attach to and penetrate host plants, and through haustoria, parasitic plants extract water and nutrients from hosts. Some of the parasitic plants cause severe economic losses, especially witchweed (Striga, Orobanchaceae), broomrape (Phelipanche and Orobanche, Orobanchaceae) and dodder (Cuscuta, Convolvulaceae). For example, approximately 20% of sorghum yield is lost annually in Africa due to infestation by Striga hermonthica (Kawa et al., 2024). It was estimated that annual economic losses caused by all parasitic weeds, exceed a minimum value of US ＄111 million and are most likely around US ＄200 million, increasing by approximately US ＄30 million annually (Rodenburg et al., 2016). Understanding the molecular mechanisms underlying the interactions between host and parasitic plants and breeding or genetic engineering parasitic plant-resistant crops are essential for controlling these noxious weeds.

Although many parasitic plants have evolved to be able to parasitize wide ranges of host species, many plants are resistant to parasitization, and different crop cultivars or ecotypes of wild species may have varying degrees of resistance to parasitic plants. A growing number of genetic and biochemical studies have implicated plant metabolites and resistance (R) genes in such host plant-parasitic plant interactions (Clarke et al., 2019; Jhu and Sinha, 2022). For example, a maize variety, the genotype NP2222, has dramatically decreased levels of strigolactones zealactol and zealactonoic acid, resulting in highly reduced Striga germination and parasitization, as Striga needs to perceive strigolactones released from roots to germinate and locate hosts for successful parasitization (Li et al., 2023). Parasitization of dodder Cuscuta reflexa leads to increased contents of plumbagin and isoshinanolone in the tropical liana plant Ancistrocladus heyneanus, and in vitro experiments indicated that these metabolites inhibit C. reflexa growth; in addition, a hypersensitive response (HR) is developed at the infection sites of A. heyneanus, leading to abortion of parasitization (Bringmann et al., 1999). Some genotypes of both common vetch (Vicia sativa) and bitter vetch (Vicia ervilia) exhibit HR to dodder (Cuscuta campestris) infestation (Córdoba et al., 2021). Some sorghum (Sorghum bicolor) genotypes can block Striga hermonthica haustorial penetration through certain cell wall-enabled mechanism and some show a HR to Striga hermonthica infection, completely fending off this parasite (Kavuluko et al., 2021; Mutinda et al., 2023). The cowpea (Vigna unguiculata) cultivar B301 is resistant to the races SG4 and SG3 of Striga gesnerioides by developing a HR at the infection sites, and the host VuPOB1, a BTB-BACK domain-containing ubiquitin E3 ligase-like protein, positively regulate HR (Su et al., 2019). Many wild sunflower species (Helianthus) are completely resistant to sunflower broomrape (Orobanche cumana) and different cultivars of common sunflower Helianthus annuus also exhibit varying levels of resistance (Cvejić et al., 2020). HaOr7 gene, originally from a wild sunflower, which encodes a leucine-rich repeat receptor-like kinase protein, confers resistance in sunflower during the early stages of interaction with O. cumana (Duriez et al., 2019).

As a result of effector-triggered immunity, HR is common in R gene-mediated plant resistance to pathogens (Belkhadir et al., 2004). R proteins directly or indirectly interact with microbe-derived effectors, which function to subvert pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI) (Chisholm et al., 2006; Boller and Felix, 2009; Yuan et al., 2021a), and such interactions can induce activation of R protein-mediated effector-triggered immunity (ETI), a process generally includes HR, characterized by rapid apoptotic cell death and local necrosis (Chisholm et al., 2006; Boller and Felix, 2009). In addition to HR, plant defense responses against pathogens also include complex signaling networks involving the mitogen-activated protein kinases (MAPKs) and the phytohormones salicylic acid (SA) and jasmonic acid (JA) (Spoel et al., 2007; Zhang et al., 2022). As part of PTI, in response to PAMPs, MAPKs are rapidly activated and thereby regulate downstream defense responses (Zipfel, 2009; Hirt et al., 2013; Wang et al., 2023). SA is required for the initiation of HR and to enhance the synthesis of an array of antimicrobial phytoalexins and pathogenesis-related proteins (Durrant and Dong, 2004; Pieterse et al., 2009). JA signaling has been demonstrated to crosstalk with the SA signaling and thus optimizes the responses to biotic stresses (Berens et al., 2017; Ku et al., 2018).

Dodders are rootless and leafless obligate holoparasitic plants, which solely rely on host plants to survive. They have vine-like stems which twine around the host stem, forming many haustoria at the contact sites between stems of dodder and host. Dodder parasitization causes severely stunted growth or even death of hosts and dodders are difficult to eradicate due to tight physical connections with their hosts. Much progress has been made in understanding the unique physiology of dodders, including their genome evolution (Sun et al., 2018; Vogel et al., 2018), dodder-host interplant signaling (Shen et al., 2023), and haustorium-mediated transfer of mRNAs, small RNAs, and proteins between dodder and host and between different dodder-connected plants (Kim et al., 2014; Shahid et al., 2018; Liu et al., 2020; Zhang et al., 2024). However, how dodder parasites locate host plants, establish parasitism, and how hosts respond to dodder parasitization remain not well understood.

Most solanaceous plants are susceptible to dodders, including the wild tomato Solanum pennellii (hereafter wild tomato). However, the stem of cultivated tomato (hereafter tomato) S. lycopersicum develops a strong HR at the haustorial penetration sites; thus, tomato is completely resistant to dodder (Sahm et al., 1995). Runyon et al. (2010) demonstrated that Cuscuta pentagona infestation causes HR in petioles of tomato, and the SA-deficient NahG tomato plants or JA-insensitive jasmonic acidinsensitive1 (jai1) tomato mutants have highly decreased resistance to C. pentagona. Extract of the Cuscuta reflexa induces ethylene in the leaves of tomato, but not in the leaves of wild tomato; importantly, using genetic mapping and biochemical assays, Hegenauer et al. (2016) identified a cell surface leucine-rich repeat receptor-like protein Cuscuta Receptor 1 (CuRe1), whose ortholog in wild tomato had pseudogenized, and the wild tomato or Nicotiana benthamiana ectopically expressing CuRe1 responded to C. reflexa extract with ethylene emission and exhibited a HR and increased resistance to C. reflexa. Later, a glycine-rich protein (GRP) and its minimal peptide epitope Crip21 were identified in C. reflexa, and GRP and Crip21 specifically bind to and activate CuRe1 to induce defense responses (Hegenauer et al., 2016, Hegenauer et al., 2020). Coimmunoprecipitation analysis demonstrated interaction between CuRe1 and suppressor of BAK1-interacting receptor kinase (SlSOBIR1) or SlSOBIR1-like (Hegenauer et al., 2016). Through RNA-seq analysis, Jhu et al. (2022) identified three possible resistance genes PR1, CuRe1-like, and NLR in tomato, and the tomato plants parasitized by C. campestris exhibited decreased heights, when these genes were respectively knocked out.

How does tomato perceive dodder parasitization and activate a HR that is often observed during ETI in the interaction between plants and pathogens? In this study, we generated multiple mutants of tomato, whose AOC (JA biosynthesis), EIN2 (ethylene signaling), CuRe1, SOBIR1 and SOBIR1-like, MPK1 and MPK2 (orthologs of MPK6 in Arabidopsis) (Stulemeijer et al., 2007), or RBOH1 (respiratory burst oxidase homolog 1, a ortholog of RBOHF in Arabidopsis) (Zhou et al., 2014) were respectively knocked out, and a transgenic tomato (NahG, encoding a salicylate hydroxylase) to investigate the regulation of tomato resistance to the dodder Cuscuta australis. Through analyses of phytohormones, HR phenotype, dodder biomass, and transcription data, we discovered that the JA and SA pathway both play pivotal roles in mediating HR-based resistance of tomato to C. australis, and these two phytohormones also regulate non-HR-based resistance. Importantly, the JA and SA pathway transcriptionally regulate CuRe1. Our data also suggest that certain receptor, but not CuRe1, perceives C. australis parasitization and regulates JA and SA responses. This study reveals a possible linear model of tomato resistance to C. australis.

2. Materials and methods 2.1. Plant materials

AOC, ISC1, CuRe1, SOBIR1 and SOBIR1-like, MPK1 and MPK2, EIN2, and RBOH1 were knocked out using the clustered regularly interspaced short palindromic repeats CRISPR/Cas9 system in tomato (Solanum lycopersicum, cv. Ailsa Craig). Gene accession numbers and the single guide RNA (sgRNA) sequences are listed in Table S1. The NahG sequence under the control of a cauliflower mosaic virus 35S promoter was transform into tomato (cv. Ailsa Craig). All the above transgenic materials were screened to obtain homozygous plants for subsequent experiments through antibiotic screening and sequencing the PCR fragments of the target-flanking regions (Fig. S1 and the primers are listed in Table S1). The tomato mutant spr8 is in a Castlemart background and the Castlemart tomato was used as the wild-type plants for comparisons. The wild tomato Solanum pennellii (accession LA0716) was used in this study. Cuscuta australis was a strain maintained in our laboratory (Sun et al., 2018).

2.2. Plant cultivation and Cuscuta australis infestation

Cuscuta australis was germinated following a method by Zhang et al. (2024). The wild tomato plants were used to infest C. australis seedings, giving rise to C. australis stocks. The seeds of the cultivated tomato, wild tomato, and all transgenic tomato plants were germinated on sodden filter paper for three days and then sown in soil. All the plants were grown in a glasshouse under a condition of ca. 12 h of light and 12 h of dark with the temperature of ca. 25 ℃ during the day and 22 ℃ at night. Thirty days old tomato plants and 50 days old wild tomato plants were used for all the experiments. To infest C. australis on tomato or wild tomato, five segments of C. australis stem tips (~8 cm long) excised from the stocks were placed next to each stem of tomato or wild tomato plants, and the next day, one of the C. australis stems which successfully twined around tomato stem were kept and the rest C. australis stems were removed.

2.3. Plant treatment

Solutions of 2 mM methyl jasmonate (MeJA) and 2 mM SA were prepared by directly dissolving MeJA (Sigma, CAS: 392707) or SA (Sigma, CAS: 69-72-7) in water. For exogenous supplementation of these hormones, one month old tomato plants were infested with Cuscuta australis, and after three days, water, 2 mM MeJA, 2 mM SA, or both 2 mM MeJA and SA were sprayed to the whole plants once a day for 14 days. To wound the plants, a pattern wheel was rolled along the midrib three times on each side, and in 1 h, the leaves were harvested for phytohormone measurement.

2.4. Phytohormone extraction and measurement

For quantification of phytohormones, stem segments of tomato or wild tomato where C. australis parasitized or wounded leaves were harvested and ground in liquid nitrogen. Approximately 100–150 mg of each sample was extracted with 1 mL of ice-cold ethyl acetate spiked with the internal standards (20 ng of D₆-JA, 5 ng of D₆-JA-Ile, 10 ng of D₄-SA, 5 ng of D₆-ABA, and/or 5 ng of D₆-IAA). Detailed extraction and measurement of plant hormones was performed according to a previously published method (Setotaw et al., 2024).

2.5. RNA extraction and qPCR

Total RNA was extracted using the TRIzol reagent (ThermoScientific) following the manufacturer's instructions. Total RNA (0.5 μg) was reverse transcribed to cDNA using Hifair Ⅲ 1st Strand cDNA Synthesis Kit (Yeasen). Quantitative real-time PCR (qPCR) was performed on a CFX Connect Real-Time PCR Detection System (Bio-Rad) using the iTaq Universal SYBR Green Supermix kits (Bio-Rad). For each gene expression analysis, a linear standard curve was first constructed using a series of dilutions of cDNA and the threshold cycle number, based on which the extrapolated concentration of transcripts was calculated. The tomato EF1a was used as the internal standard for normalizing the variations in cDNA concentrations. The qPCR primers pare listed in Table S2.

2.6. RNA-seq data acquisition and analysis

Three stem samples were pooled as one replicate, and three replicates were used for RNA library construction and sequencing. For RNA-seq data acquisition, Illumina TruSeq RNA Sample Prep Kit (Illumina) was used to construct cDNA libraries. Based on the genome sequences of tomato (https://jgi.doe.gov/) and Cuscuta australis (Sun et al., 2018), DESEQ2 was used to identify differentially expressed genes (DEGs) (Love et al., 2014). DEGs with P-values < 0.05 and the absolute value log₂ foldchang ≥1 were selected for further analysis. Venn diagram analysis was performed using an online tool (http://bioinformatics.psb.ugent.be/webtools/Venn/). KEGG pathway analysis was performed using the TBtools (Chen et al., 2023). The RNA-seq data can be retrieved from the Beijing Institute of Genomics under the BioProject: PRJCA034507.

2.7. Statistical analysis

Statistical analysis was performed using R (v.4.2.1) (https://www.r-project.org/). Student's t-test was applied for comparisons of two groups. One-way ANOVA with Duncan's multiple comparison was performed when there were more than two groups.

3. Results 3.1. JA and SA signaling pathway play key roles in tomato resistance to Cuscuta australis

Previously, it was found that the dodder Cuscuta pentagona grew larger on the NahG and jai1 tomato plants, which are respectively deficient in SA and insensitive to JA, than on the wild-type (WT) tomato plants, and no HR was observed in all the NahG plants, while the jai1 mutants exhibited strong HR, partial HR, or no HR (Runyon et al., 2010). To clarify the roles of JA and SA pathway in tomato resistance to C. australis, we used the CRISPR/Cas9 system to respectively knock out ICS1 (isochorismate synthase 1, a SA biosynthesis gene) and AOC (allene oxide cyclase, a JA biosynthesis gene) in tomato. However, the ics1 mutants exhibited a lethal phenotype soon after true leaves emerged. Thus, the NahG tomato plants were generated to repress the SA levels. Quantification of JA, jasmonoyl-isoleucine (JA-Ile, the bioactive form of jasmonate), and SA in the wounding-treated leaves of WT, aoc-1 and aoc-2 (two aoc mutant lines), NahG-1 and NahG-2 (two transgenic NahG lines) plants indicated successful knocking out of AOC and suppression of SA levels by NahG (Fig. S2A-C). C. australis stems were infested on the WT and these mutant/transgenic tomato plants. Among these, aoc-1 and aoc-2 and NahG-1 and NahG-2 plants exhibited susceptibility to C. australis parasitization without a HR even 14 days post infestation (DPI) (Figs. 1A and S3), and C. australis could grow normally on these mutants until setting seeds, indicating that the SA and JA pathway are both critical for tomato resistance to C. australis parasitization. Similarly, the JA biosynthesis mutant spr8, which has a mutation in the JA biosynthesis gene LoxD (Yan et al., 2013), was also susceptible to C. australis without activation of HR (Fig. S4A).

To gain insight into whether the JA and SA pathway interact during tomato resistance reaction to C. australis, we further investigated the changes of JA, JA-Ile, and SA in the stems of WT, aoc-1, and NahG-1 plants at 0 and 4 DPI, a time that HR was not yet developed. On 4 DPI, we detected respectively 1.9-, 2.4-, and 2.6-fold increase of JA, JA-Ile and SA contents in the stems of WT plants (Fig. 1B–D). As expected, the stems of aoc-1 mutants had very low levels of JA and JA-Ile (Fig. 1B and C), which did not change after C. australis parasitization, while the SA contents in the stems of aoc-1 plants increased to levels that were even greater than in the stems of WT plants (Fig. 1D). In the stems of spr8 mutants, there were no significant changes in the levels of JA and JA-Ile after C. australis parasitization as well, and the SA levels also increased, although slightly lower than in the WT plants (Fig. S4B). There were no increases of the SA levels in the stems of NahG-1 plant after C. australis parasitization, and notably the JA and JA-Ile contents remained at very low levels (Fig. 1B–D). Furthermore, the contents of phytohormones abscisic acid (ABA) and indole acetic acid (IAA) in the stems of WT, aoc-1, and NahG-1 were not correlated with C. australis resistance (Fig. S5A and B), ruling out their involvement in the resistance.

To further study whether there is any interaction between the SA and JA pathway in tomato resistance to Cuscuta australis parasitization, we exogenously applied MeJA, SA, or both to the WT, aoc-1, and NahG-1 plants, which had been infested with C. australis for three days. Only the aoc-1 plants, but not the NahG-1 plants, when being exogenous applied with MeJA alone or both MeJA and SA, regained the ability to induce HR, while applying SA to aoc-1 mutants did not activate HR (Fig. S6). Next, we measured the biomasses of C. australis parasites after 14 days of exogenously applying these hormones. On the aoc-1 plants which were supplied with MeJA or both MeJA and SA, compared to the control group, the biomasses of C. australis decreased 93.2% and 96.4%, respectively, because of the HR (Fig. 1E); although supplying SA alone to the aoc-1 mutants did not induce HR and influence the survival rates of C. australis (Figs. S6 and 1F), the biomasses of C. australis parasites decreased 47.2% (Fig. 1E). Supplementation of these hormones failed to impact the survival rates of C. australis parasites or induce HR on the NahG-1 plants (Figs. 1F and S6). However, supplying MeJA or both MeJA and SA respectively reduced the biomass of C. australis by 70.2% and 72.8%, while supplementation of SA had no impact on the C. australis biomasses (Fig. 1E). We speculated that the reason why supplementation of SA could not rescue the HR-based defense phenotype of NahG plants is that the exogenously supplied SA was rapidly degraded by the NahG enzyme, reducing to low levels that were insufficient to activate HR. Based on these data, we propose that both JA and SA signaling are both required for HR-based resistance of tomato plants; moreover, these data indicate that even in tomato plants whose JA or SA pathway is compromised and C. australis can successfully establish parasitism, JA and SA pathway also play positive roles in regulating certain non-HR-based resistance.

Applying the extract of Cuscuta reflexa to tomato activated ethylene emission (Hegenauer et al., 2016). MAPKs and ROS produced by RBOHs are involved in plant PTI and ETI (Ngou et al., 2021; Yuan et al., 2021b). Thus, additionally the ein2, mpk1 mpk2, and rboh1 mutants, whose EIN2 (ethylene signaling), MPK1 and MPK2 (two important MAPK genes), and RBOH1 (a RBOH gene) were respectively knocked out, were used to examine their resistance to C. australis. All these mutants exhibited normal HR as did the WT tomato plants (Fig. S7A), and on these plants, all the C. australis parasites died within two to three weeks (Fig. S7B). Thus, signaling pathways mediated by ethylene, MPK1 and MPK2, and RBOH1 are not required for tomato resistance to C. australis.

3.2. CuRe1, SOBIR1, and SOBIR1-like are indispensable in tomato resistance to Cuscuta australis parasitization

Previously, CuRe1 was identified in tomato by map-based cloning and expressing CuRe1 in wild tomato and Nicotiana benthamiana led to a HR at the parasitization sites and thus these transgenic plants exhibited highly enhanced resistance to Cuscuta reflexa; biochemical assays indicated that CuRe1 physically interacts with SOBIR1 and SOBIR1-like, two adaptor kinases (Hegenauer et al., 2016). However, whether CuRe1 and SOBIR1/SOBIR1-like are involved in tomato resistance to C. australis was unclear. Thus, we respectively knocked out CuRe1 and both SOBIR1 and SOBIR1-like in tomato.

Upon Cuscuta australis parasitization, the cure1-1 and cure1-2 line exhibited no HR and C. australis could well establish parasitism (Figs. 2A and S8), indicating the importance of CuRe1 in tomato resistance to C. australis. Next, we asked whether the JA and SA pathway are involved in CuRe1-mediated activation of HR. Thus, the levels of JA, JA-Ile, and SA in cure1-1 mutants were measured. The JA, JA-Ile, and SA contents in the cure1-1 mutants all increased to similar levels as those in the WT plants (Fig. 2B–D). Next, the cure1-1 mutants were applied with MeJA, SA, or both to examine whether these hormones can rescue the HR phenotype. However, none of these treatments could activate HR in cure1-1 mutants (Fig. 2A), and importantly, the biomasses and survival rates of C. australis were not impacted by exogenous application of MeJA, SA, or both (Fig. 2E and F). These data suggest that certain receptor, but not CuRe1, functions to perceive parasitization of C. australis and thereby activates the JA and SA pathway to deploy HR.

The sobir1 sobir1-like double mutants were also infested with Cuscuta australis, and similar to the cure1-1 mutants, these sobir1 sobir1-like double mutants did not have a HR against C. australis parasitization (Fig. 3A), and the JA, JA-Ile and SA contents in sobir1 sobir1-like mutants were also increased to similar levels as those in the WT plants (Fig. 3B–D). After sobir1 sobir1-like mutants were exogenously applied with MeJA, SA, or both, these mutants similarly failed to activate of HR (Fig. 3A) and the C. australis biomasses and survival rates were not altered either by exogenous supplementation of JA or SA (Fig. 3E and F).

These data not only indicate that CuRe1 and SOBIR1/SOBIR1-like play the key roles and are indispensable in tomato HR- and non-HR-based resistance to Cuscuta australis but also imply a possible scenario that JA and SA pathway may be located upstream of CuRe1 and SOBIR1/SOBIR1-like.

3.3. JA and SA pathway transcriptionally regulate CuRe1

To determine whether the JA and SA pathway regulate CuRe1 and SOBIR1/SOBIR1-like, we subsequently assessed the relative expression levels of these genes in the WT, aoc-1, and NahG-1 plants. On four DPI, the transcript levels of CuRe1 were highly induced in the stems of WT tomato plants (348-fold), but it was only induced 16.2-fold and 4.1-fold, respectively, in the stems of aoc-1 and NahG-1 plants (Fig. 4A). To investigate whether CuRe1 is specifically transcriptionally regulated by C. australis parasitization, JA, SA, or both were applied to unparasitized tomato plants once a day for seven days, and the stem samples were harvested at early (6 h) and late (7 days) times. No obvious changes of the transcript levels of CuRe1 were detected (Fig. 4B). These data indicate that in addition to JA and SA, certain factor from C. australis is also required for the transcriptional upregulation of CuRe1.

Both SOBIR1 and SOBIR1-like were only slightly upregulated upon Cuscuta australis infestation in the stems of WT, aoc-1, and NahG-1 plants (Fig. 4A), and notably, SOBIR1 and SOBIR1-like were negatively regulated by the JA pathway, as the transcript levels of SOBIR1 and SOBIR1-like transcripts were even greater in the aoc-1 plants than in the WT plants. The transcript levels of CuRe1 in the sobir1 and sobir1-like mutants and the transcript levels of SOBIR1 and SOBIR1-like in the cure1-1 mutants were analyzed, but we found no evidence of mutual transcriptional regulation between CuRe1 and SOBIR1/SOBIR1-like (Fig. 4C and D).

3.4. JA and SA signaling pathway do not regulate the resistance of wild tomato to Cuscuta australis

The wild tomato Solanum pennellii is susceptible to Cuscuta australis. To gain insight into the underlying mechanisms of the susceptibility of wild tomato to C. australis, we first quantified the JA and SA levels in the stems of wild tomato after C. australis infestation. The JA and SA contents in the wild tomato plants remained relatively low and did not alter over 6 DPI, while the tomato plants exhibited increased levels of JA and SA, especially on 4 DPI (Fig. 5A and B). Next, we exogenously applied MeJA, SA, or both to the wild tomato plants to examine whether these two hormones can enhance wild tomato resistance to C. australis. However, no HR was detected (Fig. S9). Furthermore, these hormone treatments did not affect the growth of C. australis either (Fig. 5C and D). These findings suggest that the JA and SA signaling pathway do not influence the resistance of wild tomato to C. australis parasitism, and given that JA and SA levels in wild tomato did not change in response to C. australis parasitization, it is likely that the receptor specific for perception of C. australis parasitization is missing or nonfunctional in the wild tomato.

3.5. Transcriptomes of WT, aoc-1, NahG-1, and cure1-1 plants suggest possible involvement of metabolites in tomato resistance to Cuscuta australis

Next, transcriptome analyses were used to gain further insight into the underlying mechanisms of tomato resistance to Cuscuta australis. Since the JA and SA levels were the most upregulated on 4 DPI, when HR was not yet developed, 4 DPI is likely a suitable time to investigate the transcriptional regulation of tomato resistance to C. australis (Fig. 1, Fig. 5A). Thus, we conducted a transcriptome analysis using the stems of WT, aoc-1, NahG-1, and cure1-1 plants, which had been parasitized by C. australis for 4 days or uninfected (controls). The differentially expressed genes (DEGs) in stems of WT, aoc-1, NahG-1, and cure1-1 plants were found to be 2372 (1231 up-, 1141 downregulated), 4299 (2510 up-, 1789 downregulated), 2495 (632 up-, 1863 downregulated), and 1432 (669 up-, 763 downregulated), respectively (Fig. 6A and Table S3).

Venn diagram was used to depict the specific and common DEGs among the stems of WT, aoc-1, NahG-1, and cure1-1 plants (Fig. 6B and Table S4). Given that the WT tomato plants are resistant to C. australis, while the other aoc-1, NahG-1, and cure1-1 plants are not, it is possible that certain DEGs specific for WT are involved in the resistance. The 626 WT-specific DEGs (Fig. 6C) were therefore used for enrichment analysis of pathways based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) tool. It was found that the KEGG pathways related to secondary metabolites were enriched, such as "biosynthesis of other secondary metabolites", "flavonoid biosynthesis", and "phenylpropanoid biosynthesis" (Fig. 6C). Next, transcription factors (TFs) were specifically inspected due to their important roles in transcriptional regulation of downstream genes. Among these 626 WT-specific genes, 20 TFs were identified, including five belonging to the MYB family (4 up-, 1 downregulated), two belonging to the TCP family (both downregulated), two from the bHLH family (both downregulated), two from the ethylene-responsive TF family (both upregulated), and two from the zinc finger family (both upregulated) (Fig. 6D and Table S5).

There are 159 common genes among the stem samples of WT, aoc-1, NahG-1, and cure1-1 plants (Fig. 6B), and the genes having different regulations between the WT and all the other plants (aoc-1, NahG-1, and cure1-1) may be involved in the resistance of tomato plants. Therefore, these common genes were specifically inspected for those which have the same patterns of regulation (up- or downregulation) in all the aoc-1, NahG-1, and cure1-1 plants but have the opposite directions of regulation in the WT plants. Three genes were identified in this manner: a lipoxygenase (LOX2S), which may be involved in JA biosynthesis, and aspartic protease inhibitor 1 and wound-induced proteinase inhibitor 1 precursor, two genes encoding protease/proteinase inhibitors (Fig. S10A). In addition, the genes that have the same patterns of regulation (up- or downregulation) in all aoc-1, NahG-1, and cure1-1 plants but are respectively more strongly up- or downregulated in the WT were identified: nine genes were more strongly upregulated in the WT than in all aoc-1, NahG-1, and cure1-1 plants, and among these, there are two polygalacturonase genes, which encode enzymes with functions of degradation of cell walls (Fig. S10B and C). Another scenario is that certain genes commonly regulated in aoc-1, NahG-1, and cure1-1 plants but not in WT may also be part of the reason for the susceptibility of these plants to C. australis. Thus, the DEGs in all aoc-1, NahG-1, and cure1-1 plants but not regulated in the WT plants were screened among the 76 common genes. Sixty-six genes were obtained, which were commonly up- or downregulated in all the aoc-1, NahG-1, and cure1-1 plants (Table S6), including two TFs, one belongs to the GATA type zinc finger TF family and one belongs to the bHLH DNA -binding superfamily (Fig. S11A), and KEGG pathway analysis indicated that metabolism was enriched from these 66 genes (Fig. S11B).

These transcriptome analyses point to the possibility that certain tomato metabolites may confer resistance to Cuscuta australis.

4. Discussion

How host plants resist parasitization is central to the understanding of host plant-parasitic plant interactions and is of great interest for breeding or genetically engineering parasitic plant-resistant crops. This study reveals that Cuscuta australis parasitization is likely perceived by certain receptor in tomato, which subsequently activates the JA and SA pathway (Fig. 7), and JA and SA serve as the primary regulatory phytohormones in tomato to transcriptionally regulate CuRe1, and CuRe1 and SOBIR1/SOBIR1-like thereby together activate both HR- and non-HR-based resistance to C. australis parasitism (Fig. 7).

A large body of evidence has indicated that SA and JA, especially the former, play important roles in disease resistance (Powers et al., 2024; Tian et al., 2024), but little is known about the roles of phytohormones in plant resistance to parasitic plants (Jhu and Sinha, 2022). The resistance of tomato to dodder by activating a locally restricted HR has been reported decades ago (Bernd et al., 1988). The HR appears about ten days after Cuscuta australis parasitization, when haustoria have penetrated tomato stems. Obviously, the HR-based resistance is initiated post-attachment (Jhu and Sinha, 2022), as the first barrier and the most effective means of resistance to dodders. Runyon et al. (2010) found that the jai1 mutants, which have a defect in the JA receptor gene COI1 (coronatine-insensitive1), partially lost the HR-based resistance to C. pentagona: among ten jai1 mutants, 40% showed a strong HR, 20% showed relative weak HR, while 40% had no HR. In contrast, all our aoc mutants lost HR throughout all the experiments we performed and so did all the spr8 mutants. We speculate that in addition to COI1, another COI-independent pathway regulated by JA is also involved in activation of HR in tomato. Using the NahG plants, both current study and Runyon et al. (2010) revealed the importance of SA pathway in activating HR in tomato. Furthermore, by supplying JA or SA to aoc-1 and NahG-1 tomato plants, our data revealed that JA and SA function independently but convergently to activate HR during C. australis infection (Fig. 1E and F), as neither supplying JA to NahG-1 plants nor supplying SA to aoc-1 plants could restore the HR phenotype (Fig. S6). It is worth noting that in the NahG-1 plants, C. australis parasitization-induced JA and JA-Ile levels were lower than in the WT plants (Fig. 1B and C). Thus, the SA pathway is required for normal induction of JA and JA-Ile after C. australis attack. Similarly, in the transgenic tobacco plants expressing NahG, the JA levels were not elevated during a HR elicited by Pseudomonas syringae pv phaseolicola (Mur et al., 2006). Early accumulation of JA after inoculation of Pseudomonas syringae avrRpt2 was also absent in the Arabidopsis NahG plants (Heck et al., 2003). However, despite the decreased JA/JA-Ile levels, supplementation of JA to NahG-1 plants could not restore its HR, indicating that JA does not function downstream of SA. Supplementation of JA to aoc-1 plants perfectly restored the HR phenotype, and this result ruled out the possibility of too slow diffusion/transport of exogenous JA into tomato stem tissues. After supplying SA to the aoc-1 plants or JA to the NahG plants, the aoc-1 and NahG-1 plants exhibited increased resistance – C. australis parasites were smaller than those grew on mock-treated aoc-1 and NahG-1 plants. Thus, JA and SA not only control HR-based resistance but also modulate certain non-HR-based resistance in response to C. australis parasitization (Fig. 7).

Hegenauer et al. (2016) showed that infecting transgenic wild tomato and Nicotiana benthamiana, which expressed tomato CuRe1, with Cuscuta reflexa could induce HR. In this study, we knocked out CuRe1 in tomato and our data indicated the critical function of CuRe1 in tomato resistance to C. australis. Furthermore, by knocking out SOBIR1 and SOBIR1-like, we show that these two adaptor proteins are also required for activation of HR in tomato response to C. australis parasitization. These genetic data well support the biochemical analysis indicating that CuRe1 forms a complex with SOBIR1 or SOBIR1-like (Hegenauer et al., 2016). The JA and SA pathway both transcriptionally positively regulate CuRe1 (Fig. 4A). However, similar to the WT tomato plants, the cure1-1 mutants still responded to C. australis infestation normally with elevated JA and SA levels (Fig. 2B–D). Thus, CuRe1 is likely not the receptor of C. australis-derived molecular pattern, but certain unknown receptor functions to perceive the C. australis molecular pattern and thereby activate the JA and SA pathway, which in turn transcriptionally upregulate CuRe1 to initiate HR (Fig. 7). In addition to HR, CuRe1 is also required for JA- and SA-mediated non-HR-based resistance (Fig. 7), as applying JA or SA to the cure1-1 mutants did not have any impact on the biomasses of C. australis (Fig. 2A). Sequence analysis revealed that CuRe1 in the wild tomato S. pennellii had been pseudogenized during evolution (Hegenauer et al., 2016) and consistently, applying JA or SA to wild tomato did not affect the growth of C. australis. In contrast to tomato plants, the wild tomato plants exhibited no changes of JA and SA levels during C. australis infestation (Fig. 5A and B). Thus, in addition to a nonfunctional CuRe1, it is likely that the wild tomato S. pennellii also lacks the receptor for C. australis molecular pattern or the receptor had been pseudogenized as well, like the CuRe1.

During plant interaction with pathogens, HR/programmed cell death often involves MAPKs and ROS, which is often produced by RBOHs (Kabbage et al., 2017). However, our genetic analyses indicated that MPK1, MPK2, and RBOH1 are not required for Cuscuta australis parasitization-induced HR in tomato (Fig. S7A). In this study, we only focused on RBOH1, which was reported to be involved in tomato resistance to the root-knot nematode Meloidogyne incognita (Wang et al., 2019). Given that RBOHs are a family of multiple members in plants, it is worth examining the functions of other RBOHs in HR-based resistance of tomato. In plant–pathogen interactions, ethylene pathway may have positive or negative effect on plant resistance, depending on the plant and pathogen species and genotypes (Wang et al., 2019). Although ethylene is highly induced in tomato by the total extract of C. australis (data now shown) or C. reflexa (Hegenauer et al., 2016), knocking out EIN2, which is critical for ethylene signaling, did not compromise the HR in tomato (Fig. S7A). Thus, ethylene signaling pathway is also not important for tomato resistance to C. australis. By contrast, during Phtheirospermum japonicum (Convolvulaceae) infection on Arabidopsis thaliana, the ethylene pathway of the Arabidopsis host plants contributes to parasite invasion (Cui et al., 2020). Thus, the ethylene pathway seems to play different roles in different host-parasitic plant interaction systems.

Jhu et al. (2022) found that three tomato genes PR1, CuRe1-like, and NLR are involved in the transcriptome response of tomato to Cuscuta campestris parasitization. Our transcriptome analyses suggested that certain metabolites and TFs may be involved in the resistance of tomato to C. australis. Genetic analyses using knock-out mutants of these genes should be conducted to verify their functions, and such analyses will further deepen our understanding of the mechanisms underlying host-parasitic plant interactions.

Acknowledgements

This work was supported by National Natural Science Foundation of China (32270314 (GS)), the Key Project of Applied Basic Research Program of Yunnan (202201AS070056 (JW), 202301AS070064 (GS)), Yunnan Revitalization Talent Support Program "Yunling Scholar" Project (JW), Chinese Academy of Sciences (CAS) Light of West China Program (GS), Yunnan Revitalization Talent Support Program "Young Talents" Project (XDYC-QNRC-2022-0001 (GS).

CRediT authorship contribution statement

Jianxiang Yang: Writing – original draft. Guojing Shen: Writing – review & editing. Jianqiang Wu: Writing – review & editing.

Declaration of competing interest

None declared.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2025.03.003.