Immunity priming uncouples the growth–defense trade-off in tomato

Plants often face abiotic and biotic stresses, such as pathogen infection or climate/ecological changes. These stresses induce physiological, biochemical and molecular changes, which can be reflected in altered development and reduced productivity. Plants protect themselves against pathogens using several complex mechanisms. In the interaction between the plant and the attacker, elicitor molecules of various types are released (Mishra et al., 2012). Perception of elicitors by the host plant results in signaling that leads to activation of various defense mechanisms (Walters and Heil, 2007).

Following pathogen infection, plants can develop enhanced resistance to further attack. This is known as induced resistance, and is broadly divided into systemic acquired resistance (SAR) and induced systemic resistance (ISR) (Walters and Heil, 2007). SAR, commonly triggered by local infection, can provide long-term resistance to subsequent infection (Klessig et al., 2018). During SAR, pathogenesis-related (PR) genes are activated. Among the best-characterized PR genes is PR-1, which is often used as a marker for SAR (Pieterse et al., 2014). SAR generally relies on salicylic acid (SA), which has been shown to increase in plant tissues during pathogenesis (Yalpani et al., 1991). The SA receptor NPR1 (Non-Expressor of Pathogenesis Related Genes 1) regulates SAR signaling following SA perception. NPR1 is activated by SA, and subsequently activates transcription of genes required for SAR, including PR genes (Cao et al., 1997; Chen et al., 2021; Pieterse et al., 2014).

Unlike SAR, ISR is usually described as being initiated by beneficial soil-borne microorganisms, including fungi. ISR is associated with a physiological state in which plants can react more efficiently to pathogen attack (Shoresh et al., 2005), known as immunity ‘priming’. ISR is regulated by jasmonic acid (JA) and ethylene (ET), and is typically not associated with a direct activation of PR genes, although NPR1 can enhance ISR and, in some cases, PR-1 is accumulated during ISR priming (Gupta and Bar, 2020; Meller Harel et al., 2014; Pieterse et al., 1998). As a result of priming, upon secondary infection, defense responses occur more rapidly and strongly than during the primary infection, enabling a more effective response to the new infection (Heil and Bostock, 2002).

Although the deployment of defense mechanisms is imperative for plant survival, defense activation can come at the expense of plant growth. This is known as the ‘growth–defense trade-off’ phenomenon, and is the result of a balance between growth and defense/adaptation responses (Figueroa-Macías et al., 2021). Growth–defense trade-offs have an important role in plant survival (Huot et al., 2014). In general, mature organs and tissues are no longer growing and, as a result, can be more prepared for defense as they possess more energy and resources that can be mobilized. One recent work suggested that competition for resources could shape the costs of defense induction (Karasov et al., 2017). Growth–defense trade-offs in plants have been increasingly investigated of late (Ke et al., 2020; Margalha et al., 2019; Monson et al., 2022; Saini and Nandi, 2022). Although some works have suggested that growth and defense are antagonistic, the extent of the ‘development–defense trade-off’ and the mechanisms by which it affects plant growth and agricultural yield are largely uninvestigated.

Phytohormones involved in both developmental processes and defense could potentially be involved in regulating development–defense trade-offs. The developmental phytohormone cytokinin (CK) is involved in cell division, leaf senescence, apical dominance, vascular differentiation, root development and stress responses (Zürcher and Müller, 2016). CKs have reported roles in response to biotic stresses (Großkinsky et al., 2011; Gupta et al., 2020; Jiang et al., 2013), and promote resistance through the SA pathway (Choi et al., 2010; Gupta et al., 2020). We have previously shown that CK deficiency results in higher disease susceptibility, whereas high endogenous CK content, as well as external application of CK, confer disease resistance (Gupta et al., 2020, 2021c). Gibberellins (GAs) are plant growth hormones involved in stem elongation, germination, leaf expansion, flowering and fruit development (Davière and Achard, 2013). GAs can negatively regulate JA signaling, potentially affecting disease resistance (Hou et al., 2010). The central roles of CK and GA in plant growth and development, together with their reported connections with biotic stresses, suggest that these two hormones could be involved in growth–defense or development–defense trade-offs.

SA and JA, which were classically defined as ‘defense’ hormones, also regulate developmental processes (van Butselaar and Van den Ackerveken, 2020; Ghorbel et al., 2021; Rivas-San Vicente and Plasencia, 2011; Wasternack et al., 2013). This suggests that growth and defense may not be simply antagonistic, but instead have a more complex balance (Koo et al., 2020). Rather than looking at growth and defense as a binary switch, a more holistic view suggests that plant fitness is optimized when growth and defense are appropriately prioritized in response to both environmental and developmental cues. Researching the molecular mechanisms used by plants to balance growth and defense could lead to the possibility of manipulating these mechanisms, with the goal of increasing plant productivity.

In this work, we aimed to investigate the growth–defense trade-off in tomato, by examining how defense priming affects tomato development and agricultural qualities. We used three different immunity elicitors: ASM (acibenzolar-S-methyl), marketed under the trade name ‘Bion’ (Oostendorp et al., 2001), a SAR elicitor; Trichoderma harzianum, a classical fungal ISR elicitor (Elad, 2000a); and Bacillus megaterium, a bacterial elicitor that likely induces both SAR and ISR (Gupta et al., 2021a, 2022a). These elicitors were employed to investigate differences and commonalities between the SAR and ISR pathways, and examine their effect on tomato development. We found that both SAR and ISR affect plant development, with ISR also generating a positive effect on agricultural qualities. We show that growth and defense can be positively correlated, as mutants that lost the ability to activate defenses also lost the growth effects. In addition, we found that the CK response is activated in response to ISR, and is required for the observed growth and disease resistance resulting from ISR activation. Taken together, our results suggest that growth promotion and induced resistance can be co-dependent, and that defense priming can, within a certain developmental window, not necessitate a trade-off on growth, but rather, promote developmental processes.

The elicitors Trichoderma harzianum T39, a Trichoderma sp. fungus that induces immunity via the JA/ET pathway (Meller Harel et al., 2014), ASM, an SA analog that primes the SA pathway (Huot et al., 2014), and B. megaterium ‘4C’, a phyllosphere Bacillus isolate that promotes plant development and immunity and likely activates both SAR and ISR pathways (Gupta et al., 2021a, 2022a), were used for SAR and ISR induction. Further details on immunity priming methodology are provided in the Materials and Methods section. We conducted Botrytis cinerea infection assays on primed plants. We found that pre-treatment with all three elicitors reduced disease levels (Fig. 1). Foliar spray or soil-drench application of all three elicitors had a similar effect on disease resistance (Fig. S1).

To assess immunity activation, we measured ET and reactive oxygen species (ROS) production in primed plants (Leibman-Markus et al., 2017). We found that pre-treatment with all elicitors amplified wounding ET production and ET production mediated by the fungal microbe-associated molecular pattern (MAMP) ethylene-inducing xylanase (EIX; see Anand et al., 2021) (Fig. 1C,D). Pre-treatment with T39, ASM or 4C also amplified ROS production in response to the bacterial pathogen-associated molecular pattern (PAMP) flg22 (Fig. 1E).

To examine possible growth–defense trade-offs, we treated plants separately with elicitors, inducing immunity via ISR with T39, via SAR with ASM, or via both pathways with 4C. ASM concentrations were calibrated in preliminary assays (Fig. S2). All three elicitors promoted plant growth (Fig. 1F); however, only T39 and 4C were able to significantly promote yield and harvest index (Fig. 1G,H). Total tomato soluble sugars, as measured by Brix refractometry, were unaffected (Fig. 1I), suggesting that immunity elicitation did not affect metabolic processes governing fruit quality.

We next examined whether immunity elicitation could affect seedling development (Fig. 2). All three elicitors affected seedling height (Fig. 2A,B). T39 and 4C also increased seedling fresh weight (Fig. 2C,D) and number of leaves (Fig. 2E,F). ASM did not affect fresh weight or leaf numbers. Representative images of treated seedlings (Fig. 2G-J), meristem and the four youngest leaf primordia (Fig. 2K-N), and fifth leaves (Fig. 2O-R) are provided. A more in-depth analysis of the effect of T39 on development demonstrated that T39 treatment resulted in increases in the height of the shoot apical meristem (SAM), but not the cotyledons (Fig. S3A,E,F), indicating that the treatment affected hypocotyl growth mostly above the SAM. T39 treatment also promoted the number of unfurled leaves (Fig. S3B,G), the developmental plastochron (Fig. S3C,H) and the meristematic differentiation to flowering (Fig. S3D,I). T39 promoted leaf development, increasing the number of leaves produced (Fig. S4A), and leaf complexity (Fig. S4A-D). T39 did not alter the developmental program of leaves 1-2, which are simple and have a short developmental window in tomato (Shleizer-Burko et al., 2011) (Fig. S4C). We previously reported that 4C treatment promotes the leaf developmental program, increasing leaf complexity and patterning (Gupta et al., 2022a). ASM had no effect on leaf development (Fig. 2, Fig. S4).

The balance between CK and GA determines the leaf developmental window in tomato (Israeli et al., 2021). Because the leaf developmental window was extended upon T39 and 4C treatment, with additional organs produced (Fig. 2, Figs S3, S4; see also Gupta et al., 2022a), we hypothesized that changes in CK and/or GA response might be involved. Therefore, we assayed CK and GA pathway gene expression and content, in both source (Fig. 3) and sink (Fig. S5) leaves treated with the different elicitors. The following gene groups were tested (see also Table S3): CK biosynthesis [isopentenyl transferase (IPT) genes; Žižková et al., 2015] (Fig. 3A, Fig. S5A); CK activation [lonely guy (LOG) genes; Naseem et al., 2015] (Fig. 3B, Fig. S5B); CK degradation [cytokinin dehydrogenase (CKX) genes; Cueno et al., 2012] (Fig. 3C, Fig. S5C); CK signaling [tomato response regulator (TRR) genes; Bar et al., 2016] (numbered in accordance with Arabidopsis ARR genes) (Fig. 3D, Fig. S5D); GA biosynthesis: (GA20 oxidase and GA3 oxidase genes; Shohat et al., 2021) (Fig. 3E, Fig. S5E); GA degradation (GA2 oxidase genes; Shohat et al., 2021) (Fig. 3F, Fig. S5F); and GA signaling (procera and GAST1 genes; Shohat et al., 2021) (Fig. 3G, Fig. S5G). Developmental transcription factors were also assayed in sink tissues (Fig. S5H): the KNOX gene TKN2 (Hake and Ori, 2002), the organ boundary determination gene GOBLET (GOB) (Berger et al., 2009), and the differentiation-promoting transcription factor Clausa (CLAU) (Bar et al., 2016). KNOXI proteins modulate compound leaf development by indirectly reducing GA levels (Bolduc and Hake, 2009; Hay et al., 2002; Jasinski et al., 2005; Sakamoto et al., 2001). CLAU is known to reduce CK sensitivity and increase GA sensitivity (Bar et al., 2016; Israeli et al., 2021). To confirm defense pathway activation, we assayed the expression of the SAR gene PR1a (Silverman et al., 2005) and the ISR gene LoxD (Mariutto et al., 2011), and the ratio between the expression of these genes was calculated (Fig. 3I, Fig. S5J). Although we cannot rule out activation of both pathways, a ratio below one is indicative of ISR being the primary defense pathway, whereas a ratio above one suggests that SAR is activated. Our results support SAR as the primary pathway for ASM, as previously reported (Huot et al., 2014), and ISR as the primary pathway for T39, as previously reported (Meller Harel et al., 2014). The close-to-one value for 4C supports combined ISR and SAR activation with this bacterium, as we previously suggested (Gupta et al., 2021a, 2022a).

All three elicitors affected gene expression of CK and GA pathway genes (Fig. 3, Fig. S5). To deduce the possible effect of each elicitor on the CK/GA balance, we calculated the ratio of CK/GA promotion as: (the number of genes of the groups CK biosynthesis, CK activation, CK signaling or GA degradation that were upregulated plus the number of the genes of the groups CK degradation, GA biosynthesis and GA signaling that were downregulated) divided by (the number of genes of the groups CK biosynthesis, CK activation, CK signaling or GA degradation that were downregulated plus the number of the genes of the groups CK degradation, GA biosynthesis and GA signaling that were upregulated). The higher the calculated number, the higher the promotion of the CK/GA ratio. In source leaves, T39 had a 2.66 CK/GA promotion ratio, ASM had a 1.4 CK/GA promotion ratio and 4C had a 1.33 CK/GA promotion ratio. In sink tissue (Fig. S5), T39 and 4C both had a 1.25 CK/GA promotion ratio, whereas ASM has a 0.1 CK/GA promotion ratio. These gene expression results indicate that the CK/GA ratio is likely promoted by T39 and 4C, but it may be inhibited by ASM. In line with these gene expression results, T39 and 4C treatment resulted in an increase of isopentenyl (iP) in source leaf tissue (Fig. 3H), and 4C treatment also increased isopentenyl riboside (iPR) content. ASM treatment did not increase CK content, and resulted in reduction in the content of transZeatin (tZ; Fig. 3H). In sink tissues, CKs and GAs were mostly undetectable, likely owing to the smaller amount of available tissue. tZ was unchanged by elicitor treatments, and 4C treatment effected an increase in GA7 (Fig. S5I), in accordance with the parallel increase in GA20ox1 (Fig. S5E).

The balance between CK and GA, rather than the content of each hormone, has been reported to be important in leaf development (Fleishon et al., 2011; Israeli et al., 2021). Because we observed probable increases in the CK/GA ratio upon elicitor treatments (Fig. 3, Fig. S5), we further examined CK signaling in response to elicitors using tomato plants expressing the synthetic CK-response promoter pTCSv2 (two component sensor) driving VENUS fluorescent protein expression (Steiner et al., 2020). T39 and 4C increased pTCSv2-driven fluorescence (CTF) in meristems and the three youngest leaf primordia (m+3), as well as in older primordia (P4-P5) (Fig. 4). This correlates with the effect of these elicitors on gene expression in sink tissues (Fig. S5), as expected (Steiner et al., 2020; Zürcher et al., 2013).

Because these results indicated that the ratio between CK and GA affects the developmental signal generated by immunity induction with ISR elicitors, we examined the response to T39, ASM and 4C in several altered hormone content and/or signaling lines. Genotypes with high CK content or signaling are generally disease resistant, and genotypes with low CK content or signaling can be disease susceptible (Gupta et al., 2020, 2021c). Because the ratio between CK and GA can determine developmental outcomes (Israeli et al., 2021), we included both high and low CK, and high and low GA genotypes in the study.

We found that the high CK or low GA lines behaved similarly to the M82 background line in growth, development, and B. cinerea resistance upon treatment with all three elicitors (Fig. 5A-D,G,H,K,L, Fig. S6A-D,G,H,K,L). However, T39 and 4C were no longer able to promote growth (Fig. 5E,I), leaf development (Fig. S6E,F,I,J) or B. cinerea resistance (Fig. 5F,J) upon reduced CK (Fig. 5E,F) or increased GA signaling (Fig. 5I,J). ASM promoted disease resistance irrespective of CK or GA signaling (Fig. 5B,D,F,H,J,L), and continued to promote vertical growth upon reduced CK signaling (Fig. 5E), but did not affect leaf development in any of the lines (Fig. S6A-L). Some differences were observed in the level of response of the different lines to the different elicitors (Fig. S7), with responses to T39 and 4C significantly lower in IPT and clau than in M82. Thus, T39 and 4C, which induce ISR, increase CK biosynthesis and response (Figs 3 and 4), and cannot promote growth and development in the absence of a functioning CK pathway (Fig. 5, Fig. S6). ASM, which induces SAR, does not promote CK biosynthesis (Fig. 3), does not increase CK response (Fig. 4), does not require a functioning CK pathway to promote growth (Fig. 5) and does not promote leaf development (Fig. S6). Although 4C induces both SAR and ISR, the ISR effects are dominant in the context of growth and development, and are promoted irrespective of SAR induction.

To examine further the relationship between development and defense, we used defense mutants. Tomato mutants compromised in JA and ET pathways have been previously shown not to respond to Trichoderma, with disease reduction in certain cases (Jogaiah et al., 2017; Korolev et al., 2008), whereas lines compromised in the salicylic acid pathway are expected not to respond to ASM (Oostendorp et al., 2001). Thus, we used the SA-lacking transgene NahG and its background line Moneymaker (Mm) (Brading et al., 2000), the reduced JA mutant spr-2 and its background line Castelmart (Cm) (Li et al., 2003), and the reduced ET sensitivity mutant neveripe (nr) and its background line Pearson (Pn) (Lanahan et al., 1994), and examined their disease response to elicitation with T39, ASM and 4C. The mutant lines responded as expected to B. cinerea infection: spr-2 with increased susceptibility (AbuQamar et al., 2008), and nr and NahG with similar susceptibility (Mehari et al., 2015; Audenaert et al., 2002). All the background lines responded to the three elicitors with reduction in B. cinerea-incited disease (Fig. 6), as was observed for M82 (Fig. 1). However, NahG did not respond to ASM or 4C (Fig. 6A), whereas spr-2 and nr did not respond to T39 or 4C (Fig. 6B,C).

Given that mutating defense pathways prevented disease reduction by elicitor pre-treatment, we examined whether the developmental effects of these elicitors would also be abolished in the same mutants. All background lines responded to the elicitors with increases in height (Fig. 6D-F). NahG did not respond to ASM or 4C (Fig. 6D), whereas spr-2 and nr did not respond to T39 or 4C (Fig. 6E,F).

Next, we examined the ability of the elicitor treatments to promote side shoot production and leaf complexity, which are CK-dependent parameters (Mller and Leyser, 2011; Shani et al., 2010; Werner et al., 2001). ASM did not affect these parameters in wild-type lines, but was included as a control in case it might have effects on the mutant lines. T39 and 4C increased side shoot production and leaf complexity in all background lines (Fig. S8). Although NahG responded similarly to its background line Moneymaker in response to T39, it did not respond to 4C (Fig. S8A,D). spr-2 and nr did not respond to T39 or 4C (Fig. S8B,C,E,F). Although spr-2 had significantly higher side-shoot production than the Cm background line without elicitor treatment, not all buds were activated, with many plants having only two to four activated side-shoots (Fig. S8B); therefore, the lack of activity of T39 and 4C on side shoot activation in spr-2 is not likely to be due to system saturation. ASM had no effect on side shoot or leaflet generation in any of the genotypes (Fig. S8). T39 and 4C promoted yield in all the background lines, but not in spr-2 or nr (Fig. S9A-C). T39 was able to promote yield in NahG, but 4C was not (Fig. S9A). Untreated NahG and its background Mm, and nr and its background Pn, had similar disease resistance and developmental parameters. spr-2 baseline disease resistance was lower than its background Cm, and some of the spr-2 developmental parameters were also altered (Fig. 6, Fig. S8).

We also examined seedling development in defense mutant genotypes (Fig. 7, Fig. S10). All elicitors promoted seedling height in the background lines (Fig. 7A-C). As observed with M82 (Fig. 2, Fig. S6), T39 and 4C increased the number of leaves produced (Fig. S10A-C), as well as leaf complexity (Fig. S10D-F). As observed with the mature plants (Fig. 6, Figs S8 and S9), NahG did not respond to ASM or 4C (Fig. 7A), whereas spr-2 and nr responded to ASM (Fig. 7B-C), but did not respond to T39 or 4C (Fig. 7B-C, Fig. S10B,C,E,F). In seedlings, untreated NahG and its background Mm had similar leaf development, whereas nr and spr-2 had altered leaf development compared with their respective background lines Pn and Cm (Fig. S10). Untreated seedlings of all mutant lines had similar height as their background lines (Fig. 7).

Thus, the competence to respond to each elicitor by increased immunity is the same as the competence to respond by altered development and increased growth and yield.

We observed that tomato lines with altered CK/GA ratios displayed altered responses to SAR and ISR, with a low CK/GA ratio unable to support growth promotion or disease resistance in response to ISR priming (Fig. 5). Given that we previously observed that CK and GA can themselves affect disease resistance and plant development when applied exogenously (Israeli et al., 2021; Marash et al., 2023 preprint), we examined whether they were able to do so in lines with altered defense pathways. As seen in Fig. 8, vertical growth was reduced by CK and promoted by GA irrespective of plant genotype (Fig. 8A), whereas leaf development and resistance to B. cinerea were inhibited by GA and promoted by CK only in the background lines spr-2, NahG, and nr were not able to respond to exogenous CK or GA treatment with respect to leaf development or disease resistance (Fig. 8B,C). Thus, the competence to respond to exogenous CK or GA treatment with alterations in leaf development mirrors the competence to respond to these hormones with alterations in disease resistance.

Cellular defense responses are activated by T39 (Fig. 1; Gupta et al., 2022b; Gupta and Bar, 2020), ASM (Fig. 1; Marash et al., 2022) and 4C (Fig. 1; Gupta et al., 2021a). ASM-treated plants responded to EIX (a JA pathway molecular inducer) more strongly than did 4C-treated plants, whereas T39-treated plants responded to flg22 (an SA pathway molecular inducer) more strongly than did ASM- or 4C-treated plants (Fig. 1). This could indicate crosstalk or cross-potentiation between SAR and ISR, which has been previously suggested (Vlot et al., 2021). To investigate further the relationship between development and defense, we examined whether there could be a genetic overlap between these processes. In previous work (Israeli et al., 2021), we generated a dataset of genes that promote morphogenesis. Our approach, which was validated, relied on published data (Ichihashi et al., 2014) including transcriptomics of Solanum lycopersicum cv M82 at four developmental stages. Here, to examine whether we could identify an overlap between morphogenesis and defense pathways, we re-analyzed the previously published data set using KEGG (Kyoto Encyclopedia of Genes and Genomes), to identify additional significantly differential pathways over-represented in morphogenesis. Interestingly, the KEGG pathways ‘Plant hormone signal transduction’, ‘Plant MAPK signaling pathway’ and ‘Plant-pathogen interaction’ were all found to be significantly enriched in the morphogenetic dataset (Fisher's exact test, P<0.05). Thus, this analysis provides further context to the connections between development and defense that we observed in this work in the context of immunity priming (Figs 1-7) and exogenous hormone treatments (Fig. 8). Genes of the plant-pathogen interaction pathway found to be significantly enriched in the morphogenetic group are provided in Table S1. Most of the cellular immune response pathway is represented in the morphogenesis group, suggesting possible genetic overlap between development and defense (see Table S1 and Fig. S11).

The concept of growth–defense trade-off is based on the notion that the limited resources available to an organism must be prioritized according to internal and ambient conditions the organism experiences. The role of ambient conditions, including biotic stresses, is more prominent in plants, owing to their limited ability to regulate their environment. Early research of growth defense trade-offs was partially prompted by observations that plants with highly active defense pathways or defense hormone imbalances can be highly resistant to pathogens, but have severely impaired growth (Huot et al., 2014; Vos et al., 2013). However, many of these reports investigated severely pleiotropic mutants, and dramatic changes to hormonal pathways and hormone levels can have severe impacts on development, particularly considering that defense hormones also have developmental roles (De Vleesschauwer et al., 2014; Vos et al., 2015; Wasternack et al., 2013). Growth–defense trade-offs have been suggested to rely on crosstalk between different hormones. Cross-talk between CK (Albrecht and Argueso, 2017; Cortleven et al., 2019) or GA (Huot et al., 2014; Monson et al., 2022; Yu et al., 2022), and JA, SA or ET, has been previously suggested to affect the balance between defense pathways and growth development.

Constitutively active plant defense can result in growth inhibition (Gómez-Gómez et al., 1999; Zipfel et al., 2006), although genetically increasing plant immunity was reported to lack negative effects on plant growth or development in several cases (Cao et al., 1998; Heidel et al., 2004; Leibman-Markus et al., 2021; Pizarro et al., 2020). This suggests that the mode of defense activation, whether by external or genetic means, can affect the level of the trade-off, opening the possibility of activating defense without significantly impairing plant development, a highly sought after agricultural advantage.

Previous reports (reviewed by Gupta and Bar, 2020) have shown that immunity priming agents can affect plant growth. The underlying hypothesis of the present work was that priming can in fact improve plant growth in certain cases, breaking the growth defense trade-off. To examine this, we selected three defense priming agents: the SA analog ASM, a molecular priming agent known to activate SAR (Ishii et al., 2019); the fungus T. harzianum T39, a biological priming agent known to activate ISR (Elad, 2000a); and the phyllosphere bacterium B. megaterium 4C, which we previously demonstrated likely activates both SAR and ISR (Gupta et al., 2022a, 2021a).

Our work demonstrates that defense priming can result in both induced resistance and plant growth and development promotion (Figs 1, 2, 5-7). T39, ASM and 4C all induced resistance against B. cinerea. We selected B. cinerea as a model pathogen, although all three inducers have been reported to induce resistance against additional pathogens (Gupta and Bar, 2020; Gupta et al., 2021a; Marolleau et al., 2017). We observed that ASM promoted vertical growth, whereas T39 and 4C promoted several developmental parameters in both seedlings and mature plants, including agricultural traits (Figs 2 and 5, Figs S3, S4, S6, S8-S10). Of note is that treatments with T. harzianum T39 were previously found to activate resistance to downy mildew in grapevine, without negative effects on plant growth (Palmieri et al., 2012). We previously reported that B. megaterium 4C promotes both disease resistance (Gupta et al., 2021a) and plant development and growth, including yield (Gupta et al., 2022a). Although ASM did not positively impact most growth parameters when applied to tomato at 0.001%, it also did not have a negative effect on them at the applied concentration, which could be significant agriculturally under a high disease burden. Higher concentrations of ASM, by contrast, were found to inhibit growth (Fig. S2).

As central growth hormones, it is perhaps expected that CK and GA should be involved in regulating growth–defense trade-offs. In recent years, CK has been established as a priming agent, and its role as a defense hormone is increasingly recognized (Choi et al., 2010; Gupta et al., 2020). In the context of disease resistance, most reports indicate that CK promotes immunity and disease resistance through the SA pathway (Choi et al., 2010; Gupta et al., 2020), although roles in JA-mediated defense cannot be ruled out. CK and GA have both previously been suggested to be involved in growth–defense trade-offs (Cortleven et al., 2019; Yang et al., 2012). CK signaling was shown to be regulated following pathogen infection in several systems (Argueso et al., 2012; Gupta et al., 2020; Igari et al., 2008). CK was suggested to either promote (Choi et al., 2010) or inhibit (Argueso et al., 2012; Choi et al., 2010) growth and defense simultaneously, rather than effecting a trade-off, although experimental evidence of this has thus far been limited.

Notably, JAZ–DELLA interactions integrate JA and GA signaling, such that elevated GA levels or response enhance JAZ repression of defense, and elevated JA levels enhance DELLA-mediated repression of growth (Howe et al., 2018; Yang et al., 2012). Thus, under high GA signaling, plants are more resistant to pathogens that require the SA pathway (Navarro et al., 2008). In many cases, the ratio between CK and GA is the determining factor of the developmental (Israeli et al., 2021) or disease (Fig. 5) outcome.

Here, we observed that priming immunity and growth with T39 or 4C also induced CK signaling (Figs 3 and 4, Fig. S5). CK production was induced by T39 and 4C, but not ASM (Fig. 3G). The promotion of CK signaling by elicitors is unlikely to be explained merely by the changes we observed in hormone levels, and is no doubt a result of increases in CK sensitivity following elicitor application. Notably, the clausa mutant possesses significantly reduced levels of many CKs (Bar et al., 2016), and increased levels of pre-active GAs (Israeli et al., 2021), but displays increased sensitivity to CK, reduced sensitivity to GA, high-CK developmental phenotypes, and increased disease resistance (Bar et al., 2016; Israeli et al., 2021; Gupta et al., 2020), demonstrating that hormone sensitivity, and not hormone quantity, determine growth, development and disease resistance. Thus, the alterations we observed in gene expression following elicitor treatment, could either serve to promote increases in hormone sensitivity, or be the result of feedback relationships aimed at controlling the level of such increases, as previously reported (Müller, 2011; Sun, 2008). Interestingly, ASM promoted GA biosynthesis and signaling, which may underlie its promotion of vertical growth. ASM promoted plant height and induced disease resistance under CK attenuation, and continued to support disease resistance upon increased GA signaling, whereas T39 and 4C were not able to promote increases in plant height or disease resistance under CK attenuation or increased GA signaling (Fig. 5). Leaf development was promoted by T39 and 4C (Fig. 2, Figs S3, S4, S6), but could no longer be promoted in low CK or increased GA genotypes (Fig. S6). Thus, T39 and 4C treatment can uncouple growth–defense trade-offs, in a manner dependent on the CK-GA balance. Interestingly, although CK requires SA to induce immunity and promote disease resistance (Fig. 8; Choi et al., 2010; Gupta et al., 2020), ASM did not appear to require CK in order to induce immunity and promote vertical growth (Fig. 5E,F). Our results demonstrate that both ISR and SAR priming result in transcriptional reprogramming that effect changes in plant developmental programs, and can result in increased growth.

SAR depends on SA signaling, whereas ISR relies on the JA/ET circuit. Although SA has been shown to be antagonistic to JA in defense signaling (Pieterse et al., 2012, 2009), there are cases in which SA and JA have been reported to interact positively (De Vleesschauwer et al., 2013; Ferrari et al., 2003; Liu et al., 2016; Ullah et al., 2022). Dependence of ISR on SA signaling has been reported as well (Martínez-Medina et al., 2013). Distinctions between ISR and SAR are not clear cut in tomato, and overlap between the two has been reported (Betsuyaku et al., 2018; Liu et al., 2016). Interestingly, although Trichoderma spp. are known to potentiate increases in defense via the JA/ET pathway, T39 activated the SA-dependent PR1a (Meller Harel et al., 2014). Furthermore, B. cinerea, a necrotrophic fungal pathogen against which the plant host requires intact JA signaling to defend itself (AbuQamar et al., 2008), can also activate the SA pathway (Meller Harel et al., 2014). To examine whether the same pathways required for defense in response to priming agents are required for plant growth promotion, we used the SA-deficient transgenic line NahG, the reduced JA mutant spr-2 and the reduced ET sensitivity mutant neveripe (nr). As expected, ASM was not able to induce resistance in NahG, but remained able to do so in nr and spr-2, whereas T39 induced resistance in NahG, but not in spr-2 or nr. 4C lost the ability to induce disease resistance in all three mutants (Fig. 6A-C). Investigating the growth-promoting effects in parallel, we observed similar results: T39 promoted growth and increased agricultural and developmental parameters in NahG, but was not able to do so in spr-2 or nr, whereas ASM increased plant height in spr-2 and nr, but not in NahG, and 4C could not promote growth, development or agricultural parameters in any of the mutant lines (Figs 6D-F and 7, Figs S8-S10). The background lines behaved as expected. These results support the notion that development–defense trade-offs can be uncoupled, with priming for defense also serving to promote plant growth, suggesting that components of these two pathways may overlap. In line with these results, the competence to respond to exogenous CK or GA treatment in alterations to leaf development correlated with the plants’ ability to respond to these hormones in alterations in disease resistance (Fig. 8). Interestingly, effects on vertical growth elicited by both hormones were not dependent on the SA/JA/ET circuits (Fig. 8A), supporting the notion that additional factors govern the developmental response to CK and GA or the ratio between them, as expected.

In a meta-analysis of genes previously defined as morphogenesis promoting (Israeli et al., 2021), the plant-pathogen interaction pathway was found to be upregulated (Fig. S11; Table S3). Although is it tempting to speculate that defense pathways are activated in plant tissues undergoing morphogenesis or growth, another alternative explanation could be that genes with similar functions are required in both defense and developmental processes, resulting in their grouping into similar functional pathways. However, our results indicate that some cross-potentiation between defense and developmental pathways is likely.

Based on our results, we propose that immunity elicitors promote plant growth in part through alterations to the CK/GA balance, and that responses to CK and GA, or SA and JA/ET, likely share mutual genetic components. Thus, we suggest that the ‘development–defense’ trade-off concept should be updated to include the idea that, within a defined window, defense priming may actually serve to support developmental processes, rather than inhibit them. Fig. 9 provides a model conceptualizing the involvement of immunity priming in the mitigation of growth–defense trade-offs. When biotic stress is high, growth is naturally inhibited to free up resources for the activation of stress responses and pathogen-resistance mechanisms. Pre-activation of ISR, which primarily requires the JA pathway and promotes the CK response, results in rescue of growth inhibition while promoting defense. Pre-activation of SAR, which primarily requires the SA pathway and may not substantially affect the CK/GA balance, promotes defense but results in a more minor growth inhibition rescue compared with ISR.

Although plants in nature tend to be uniquely adapted to their particular habitat, and can develop and complete their life cycle in the face of environmental and pathogenic stressors, changing climate and intense cropping practices have in some cases rendered such adaptations ineffective. Our work provides a framework for further research towards the prevention of negative effects that increased immunity can have on plant development and growth. Pending future research, this could potentially be exploited agriculturally in certain cases.

Tomato cultivars and mutant and transgenic lines were used throughout the study as detailed in Table S4. Plants were grown from seeds in soil (Green Mix; Even–Ari, Ashdod, Israel) in a growth chamber (seedling assays) under long-day conditions (16 h:8 h, light: dark) at 24°C. Mature plants for disease and immunity assays and yield experiments were grown in a 2 mm nylon mesh nethouse (spring, summer and fall ambient conditions) or in a glasshouse (winter; ambient light supplemented with heating to 24°C).

pBLS>>IPT7 (‘IPT’) overexpresses the rate-limiting enzyme in biosynthesis of the CK isopentenyl driven from the leaf-specific BLS promoter (Lifschitz et al., 2006), and has high CK levels (Márquez-López et al., 2019; Redig et al., 1996; Smigocki and Owens, 1988). pFIL>>CKX3 (‘CKX’) overexpresses a CK-degrading dehydrogenase enzyme driven from the leaf-specific FIL promoter (Bonaccorso et al., 2012), and has low CK levels (Nishiyama et al., 2011; Reid et al., 2016). These lines have been previously characterized (Shani et al., 2010). pFIL>>GFP-proΔ17 (‘proΔ17’) overexpresses a mutated DELLA protein that is constitutively active, driven from the FIL promoter, resulting in inhibition of GA signaling (Nir et al., 2017). procera (‘pro’) is a recessive mutation in the single tomato DELLA gene procera, resulting in constitutively high GA signaling (Livne et al., 2015). clausa (‘clau’) is a recessive MYB transcription factor mutant that has increased CK and reduced GA sensitivity (Israeli et al., 2021). Promoter driver lines were selected to minimize pleiotropic effects; therefore, pBLS>>IPT was used rather than pFIL>>IPT. The leaf developmental phonotypes of CKX and pro, or IPT, clau and proΔ17 were relatively similar in terms of leaf complexity (Fig. S6), suggesting that high CK and low GA, or low CK and high GA, can have similar developmental effects, supporting the notion that the ratio between these hormones, rather than their individual content, determine developmental programs.

For activation of SAR, we used ASM (Marolleau et al., 2017), and for activation of ISR we used T. harzianum isolate T39 (Elad, 2000b). B. megaterium 4C was considered to likely activate both pathways (Gupta et al., 2021a). For disease and immunity assays, elicitors were applied to 4- to 5-week-old plants by spraying the fourth and fifth leaves to drip. Two treatments were given: 3 days before challenge, and 4 h before challenge. For agricultural parameters, elicitors were applied to 3- to 4-week-old plants by soil drench, using 5 ml per pot of the indicated concentrations. Treatments were applied once a week for four consecutive weeks. For developmental analyses, gene expression and hormone content quantification, elicitors were sprayed to drip on seedlings of the indicated ages. Two treatments were given, 1 week apart, in all cases except for TCS visualization, for which one treatment was applied.

ASM, also known as BTH (benzothiadiazole), a synthetic analog of SA, is used commercially to enhance disease resistance (and marketed under the trade name ‘Bion’; Marolleau et al., 2017; Oostendorp et al., 2001). Previously, application of BTH was shown to reduce plant biomass and was correlated with induction of SA-mediated defense responses (Huot et al., 2014). SA analogs have been found to effectively reduce gray mold (B. cinerea) disease in tomato (Meller Harel et al., 2014). ASM has been shown to be processed in vivo by an SA-processing enzyme (Tripathi et al., 2010). As ASM is known to be phytotoxic (Ishii et al., 2019), we calibrated the system using two different ASM concentrations (Fig. S2), in order to ensure that we were working with non-phytotoxic concentrations. A concentration of 0.01% ASM was found to reduce seedling height and number of leaves produced (Fig. S2A,B). In 4-week-old plants, 0.01% ASM reduced the number of side-shoots produced, but did not negatively affect other plant growth parameters. Treatment with 0.001% of ASM had no negative effects on plant growth (Fig. S2A,B,D-G), promoted height in both seedlings and 4-week-old plants (Fig. S2A,D), and also induced immunity to B. cinerea, whereas 0.01% did not (Fig. S2C). Following calibration and verification of lack of toxicity (Fig. S2), ASM was applied to plants at a concentration of 0.001% v/v.

Trichoderma spp. are established plant growth-promoting fungi that enhance disease resistance by inducing ISR in a JA/ET-dependent manner (Meller Harel et al., 2014). Trichoderma spp. have been shown to promote plant growth and activate ISR against a broad spectrum of pathogens. They are ubiquitous filamentous fungi that colonize the rhizosphere and phyllosphere, promote plant growth, and antagonize numerous foliar and root pathogens (Gupta and Bar, 2020). T. harzianum isolate T39 was cultured on PDA plates (Difco lab) at 25°C in natural light. Conidia were harvested in water and applied to plants at a concentration of 10⁷ conidia/ml.

Bacilli spp. are known to promote growth and induce immunity (Miljaković et al., 2020). B. megaterium 4C is a phyllosphere isolate that we previously demonstrated induces immunity and promotes growth and development of mature plants and seedlings (Gupta et al., 2022a, 2021a). Overnight (Luria-Bertani medium) cultures of B. megaterium 4C were washed twice in sterile distilled water, and re-suspended in distilled water. The cell suspension was adjusted to an optical density of OD₆₀₀=0.1 (approximately equal to 10⁸ CFU/ml), and applied to plants via either foliar spray or soil drench as indicated.

CK and GA treatments (100 µM of each) were applied to 3-week-old S. lycopersicum plants three times a week for 2 weeks (six treatments in total), using a hand-held spray bottle. Plants were sprayed to drip and allowed to dry for 30 min before being returned to the growth chamber. Plants were processed for developmental analyses or B. cinerea infection 48 h after the final treatment.

Plants were treated with T39 or ASM once a week by soil drench for 4 weeks, starting from when the plants reached three unfurled leaves, typically 10-14 days after germination. The first three treatments were given at a volume of 5 ml, and the last treatment was given at a volume of 10 ml. Growth parameters were measured once a week: shoot length was measured using a hand ruler from root crown to shoot apical meristem. The number of side shoots was counted, taking into account only shoots of at least 5 cm in length. For leaf complexity, we counted the primary, secondary and intercalary leaflets on the indicated leaf. Primary leaflets are separated by a rachis, and some of them develop secondary and tertiary leaflets. Intercalary leaflets are lateral leaflets that develop from the rachis later than the primary leaflets and between them, and typically have a very short petiolule.

The following parameters were measured for yield: weight of fruits, weight of shoot, calculated harvest index [yield weight/total plant weight (shoot+fruit weight)], and fruit soluble sugar content, measured as Brix percentage, using a ref-85 Refractometer (MRC).

Seedlings were sprayed to drip with elicitors 10 days after germination, and 17 days after germination. Growth and development parameters were analyzed 1 week after each treatment. Parameters analyzed included height, fresh weight, number of unfurled leaves, leaf developmental stage (number of the leaf that is Plastochron 1), formation of floral meristem, and leaf complexity. Shoots were dissected with a surgical blade and No.5 forceps. Analyses following dissection were conducted using a Nikon SMZ-25 binocular stereomicroscope or a Dino-Lite hand-held microscope.

Pathogenesis assays were conducted on detached leaves, with the exception of the assays conducted with the defense mutants and their background lines (Fig. 6), for which whole plants were inoculated. We have previously demonstrated that, in most cases in tomato, relative disease levels are similar on detached leaves and whole plants, although the disease develops faster in detached leaves (Gupta et al., 2020). In the case of the defense mutants, assays were conducted on whole plants because the similarity of detached leaf and whole plant assays has not been verified in all the employed mutant genotypes. B. cinerea isolate BcI16 cultures were maintained on potato dextrose agar (PDA) (Difco Lab) plates and incubated at 22°C for 5-7 days. Leaf discs of 0.4 cm diameter were pierced from 5-day-old mycelia plates and used for inoculation of both leaves and whole plants. Leaves and plants were kept in a humid chamber following inoculation. The area of the necrotic lesions was measured 3-5 days post-inoculation using ImageJ software.

ET production was measured as previously described (Leibman-Markus et al., 2017). Leaf discs, 0.9 cm in diameter, were harvested from plants treated as indicated, and average weight was measured for each plant. Discs were washed in water for 1-2 h. Every six discs were sealed with rubber septa in 10-ml Erlenmeyer flasks, containing 1 ml assay buffer) 10 mM MES pH 6.0, 250 mM sorbitol (with or without the fungal elicitor EIX, 1 µg/ml), overnight at room temperature. Disposable syringes (5 ml) with 21 gage needles were used to extract the gaseous sample, and rubber stoppers were used to hold the sample until it was injected into the gas chromatograph. ET content was measured by gas chromatography (Varian 3350).

ROS levels were measured as previously described (Leibman-Markus et al., 2017). Leaf discs, 0.5 cm in diameter, were harvested from plants that had undergone the indicated treatments. Discs were floated in a white 96-well plate (SPL Life Sciences, Korea) containing 250 μl distilled water for 24 h at room temperature. After incubation, water was removed, and 50 μl ddH₂O were immediately added to each well to prevent the tissue from drying up. The ROS measurement reaction contained: Luminol 150 μM, HRP 15 μg/ml and either 1 μM flg22 or 1 μg/ml EIX. Light emission was measured immediately and over the indicated time, using a microplate reader luminometer (Spark, Tecan, Switzerland).

Stable transgenic M82 tomato pTCSv2::3×VENUS plants expressing VENUS driven by the synthetic two-component signaling sensor pTCSv2 (Bar et al., 2016; Steiner et al., 2020) were sprayed with elicitors 10 days after germination. VENUS expression was analyzed 48 h after treatment using a Nikon SMZ-25 stereomicroscope equipped with a Nikon-D2 camera and NIS Elements v. 5.11 software. ImageJ software was used for analysis and quantification of captured images.

Plants were sprayed with elicitors twice: 10 days after germination, and 1 week later. Plant tissue was collected 48 h after each treatment. For source leaves, L2 and L3 were harvested from four different plants for each biological replicate. Sixteen plants were sampled in total. For sink tissue, ∼70 mg of shoot apices (typically consisting of the meristem and five youngest leaf primordia, m+5) were collected. Five to seven shoots were sampled per biological replicate, for a total of at least 40 plants. Total RNA was extracted from tissues using Tri-reagent (Sigma-Aldrich) according to the manufacturer's instructions. RNA (3 μg) was converted to first strand cDNA using reverse transcriptase (Promega) and oligo-dT. qRT-PCR was performed according to the Power SYBR Green Master Mix protocol (Life Technologies, Thermo Fisher Scientific), using a Rotor-Gene Q machine (QIAGEN). Relative expression was quantified by dividing the expression of the relevant gene by the geometric mean of the expression of three established normalizers: ribosomal protein RPL8 (Solyc10g006580), Cyclophillin CYP (Solyc01g111170) (Mascia et al., 2010; Mehari et al., 2015) and EXPRESSED EXP (Solyc07g025390) (Bar et al., 2016). These genes were previously demonstrated to possess stable expression levels in immunity- or disease-challenged tomato plants (Gupta et al., 2021b; Lacerda et al., 2015; Mascia et al., 2010; Mehari et al., 2015; Pizarro et al., 2020). Although these normalizers are well established, we employed three normalizers to prevent possible effects of lack of stability of any one gene. The geometric mean, rather than arithmetic mean, of the expression levels of these genes was used, in order to account for the possibility that their expression levels are not independent of each other. All primer pairs had efficiencies in the range of 0.97-1.03. Primers used for qRT-PCR are detailed in Table S2. The genes presented in Fig. 3 and Fig. S5 are part of a larger subset of genes tested. The full set of tested genes, and the outcome, are provided in Table S3.

S. lycopersicum cv M82 tomato seedlings were sprayed with elicitors twice, at 10 and 17 days of age, and 48 h after the second treatment, source leaves (L2-L3) and sink leaves (shoot apices) were harvested. Hormone extraction was performed according to Shaya et al. (2019). Briefly, frozen tissue was ground to a fine powder using a mortar and pestle. One gram of the powder was transferred to a 2 ml tube containing 1 ml extraction solvent (ES) mixture (79% IPA:20% methanol:1% acetic acid) supplemented with 20 ng of each deuterium-labeled internal standard (IS, Olomouc, Czech Republic). The tubes were incubated for at 4°C 60 min, with rapid shaking, and then centrifuged at 14,000 g for 15 min at 4°C. The supernatant was collected and transferred to 2 ml tubes, then 0.5 ml of ES was added to the pellet and the extraction steps were repeated twice. The combined extracts were evaporated using a SpeedVac Vacuum Concentrator. Dried samples were dissolved in 200 µl of 50% methanol and, filtered through a 0.22 µm syringe filter with a cellulose membrane; 5-10 µl were injected for each analysis. LC–MS-MS analyses were conducted using a UPLC-Triple Quadrupole MS (WatersXevo TQMS). Separation was performed on a WatersAcuity UPLC BEH C18 1.7 µm 2.1×100 mm column, with a VanGuard pre-column (BEH C18 1.7 µm 2.1×5 mm). The mobile phase consisted of water (phase A) and acetonitrile (phase B), both containing 0.1% formic acid in the gradient elution mode. The flow rate was 0.3 ml/min, and the column temperature was kept at 35°C. Acquisition of LC-MS data was performed using MassLynx V4.1 software (Waters). Quantification was carried out one using isotope-labeled internal standards.

Data are presented as minimum to maximum values in boxplots, or as average ±s.e.m. in bar graphs. For Gaussian-distributed samples, we analyzed the statistical significance of differences between two groups using a two-tailed t-test, with additional post-hoc correction where appropriate: Welch's correction for t-tests between samples with unequal variances, or Holm–Sidak's correction when multiple t-tests were applied to a dataset. We analyzed the statistical significance of differences among three or more groups using analysis of variance (ANOVA). Regular ANOVA was used for groups with equal variances, and Welch's ANOVA for groups with unequal variances. Significant differences between the means of different samples in a group of three or more samples were evaluated using a post-hoc test. Tukey's post-hoc test was used for samples with equal variances, whereby the mean of each sample was compared with the mean of every other sample. Bonferroni's post-hoc test was used for samples with equal variances, whereby the mean of each sample was compared with the mean of a control sample. Dunnett's post-hoc test was used for samples with unequal variances. For samples with non-Gaussian distribution, we analyzed the statistical significance of differences between two groups using a Mann–Whitney U test, and the statistical significance of differences among three or more groups using Kruskal–Wallis ANOVA, with Dunn's multiple comparison post-hoc test as indicated. Gaussian distribution or lack thereof was determined using the Shapiro–Wilk test for normality. Statistical analyses were conducted using Prism9^TM.

Hormone quantifications were conducted with the help of the Volcani Institute Metabolomics Unit. We thank Naomi Ori for helpful discussions, and members of the Elad and Bar groups for ongoing support.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.201158.reviewer-comments.pdf.