Research ArticlePLANT DEFENSES

An engineered pathway for N-hydroxy-pipecolic acid synthesis enhances systemic acquired resistance in tomato

See allHide authors and affiliations

Science Signaling  22 Oct 2019:
Vol. 12, Issue 604, eaay3066
DOI: 10.1126/scisignal.aay3066

Enhancing defenses in crop plants

Localized exposure of a plant tissue to a pathogen stimulates systemic acquired resistance (SAR), a plant-wide immune response that protects it from further infection. SAR is induced by soluble factors, such as hormones and metabolites. Holmes et al. identified the minimal set of Arabidopsis thaliana enzymes required to synthesize the SAR-inducing metabolite NHP in a tobacco relative. Localized expression of these enzymes in tomato plant leaves or NHP treatment of sweet pepper plant leaves protected the plants from infection at distant sites. Thus, engineering crop plants to produce NHP constitutively or under the control of an inducible system may be useful for enhancing endogenous defenses, thus improving agricultural productivity.


Systemic acquired resistance (SAR) is a powerful immune response that triggers broad-spectrum disease resistance throughout a plant. In the model plant Arabidopsis thaliana, long-distance signaling and SAR activation in uninfected tissues occur without circulating immune cells and instead rely on the metabolite N-hydroxy-pipecolic acid (NHP). Engineering SAR in crop plants would enable external control of a plant’s ability to mount a global defense response upon sudden changes in the environment. Such a metabolite-engineering approach would require the molecular machinery for producing and responding to NHP in the crop plant. Here, we used heterologous expression in Nicotiana benthamiana leaves to identify a minimal set of Arabidopsis genes necessary for the biosynthesis of NHP. Local expression of these genes in tomato leaves triggered SAR in distal tissues in the absence of a pathogen, suggesting that the SAR trait can be engineered to enhance a plant’s endogenous ability to respond to pathogens. We also showed tomato produces endogenous NHP in response to a bacterial pathogen and that NHP is present across the plant kingdom, raising the possibility that an engineering strategy to enhance NHP-induced defenses could be possible in many crop plants.


Plants resist infection using a sophisticated innate immune system to detect and respond to pathogens in their environment (1). Integration of conserved signaling pathways with species-specific defense mechanisms is central for mounting an effective plant immune response (2). The modular nature of the plant immune system has enabled the engineering of desired resistance traits in different plant species. For example, transferring the unique pathogen detection capabilities from one plant family to others by genetic engineering has led to the generation of lines with broad-spectrum and enhanced disease resistance relative to wild-type plants (3).

In addition to local immune responses, plants use a circulating, chemical immune system to provide protection throughout the plant. One classic example is systemic acquired resistance (SAR). During SAR, pathogen infection of primary (local) plant tissue leads to the production of mobile signals that move long distances to activate immune responses in secondary (distal) uninfected tissues. The importance of SAR has long been documented (4), and many plant species inoculated with various microorganisms are protected against subsequent infections with the same or even unrelated pathogens (5). A plant’s ability to induce quantitative, broad-spectrum resistance provides a selective advantage for survival in a changing environment. Elucidating the molecular mechanisms that govern long-distance defense priming is an intense area of research given that SAR appears to be conserved in plant families across the plant kingdom (6) and there is potential for enhancing plant protection through engineering strategies.

Among the several metabolites that have been connected to the SAR response (salicylic acid, methyl salicylate, azelaic acid, glycerol-3-phosphate, and dehydroabietinal) (6, 7), N-hydroxy-pipecolic acid (NHP) has emerged as the salient secondary metabolite in Arabidopsis thaliana that is required to initiate SAR (811). Direct treatment of Arabidopsis leaves with synthetic NHP renders distal, untreated leaves more resistant to pathogen growth, with increased protection comparable to that observed in a classical SAR response initiated by a primary infection (10, 11). Furthermore, application of NHP provides protection against the oomycete pathogen Hyalopersonospora arabidopsisdis (11) and the bacterial pathogen Pseudomonas syringae (10, 11). NHP is derived from lysine, and previous work has associated three enzymes with the NHP biosynthetic pathway: the aminotransferase AGD2-like defense response protein 1 [ALD1 (12)], which generates the α-keto acid dehydro-pipecolic acid (dehydro-Pip) from lysine; a reductase, such as SAR-deficient 4 [SARD4 (13)], to convert dehydro-Pip to Pip; and flavin-dependent monooxygenase 1 [FMO1; (10, 11)], an enzyme that adds the hydroxylamine to Pip to generate NHP and is known to be a critical regulator of SAR (Fig. 1A) (1416). Given that NHP alone is sufficient to initiate SAR and provide broad-spectrum disease resistance in Arabidopsis, engineering inducible or constitutive NHP biosynthesis into crop plants would be an attractive approach to enhance a plant’s endogenous ability to respond to pathogens.

Fig. 1 Expression of Arabidopsis NHP biosynthesis enzymes in N. benthamiana.

(A) Canonical pathway for biosynthesis of N-hydroxy-pipecolic acid (NHP) from l-lysine (l-Lys). PLP, pyridoxal phosphate; Pip, pipecolic acid. (B) Extracted ion abundance (EIC) of Pip (m/z = 156) and NHP (m/z = 172), as determined by GC-MS, in N. benthamiana leaves transduced with Agrobacterium strains harboring the indicated combinations of GFP and Arabidopsis ALD1, SARD4, and FMO1 (filled boxes). Bars represent the means ± SD. n = 3 independent biological replicates. Amounts reported as zero indicate no detection of metabolites.


Expression of two genes enables NHP biosynthesis in N. benthamiana

As a first step toward engineering SAR into a crop plant, we built a metabolic pathway for production of NHP on demand. We tested reconstitution of the biosynthetic pathway using transient expression of the three canonical Arabidopsis NHP biosynthetic enzymes (ALD1, SARD4, and FMO1) in Nicotiana benthamiana leaves. We leveraged the Agrobacterium tumefaciens–mediated transient gene expression system (17) to simultaneously express multiple Arabidopsis gene products in intact leaves. During infection, Agrobacterium transfers DNA containing a gene of interest into the nucleus of leaf cells, which is then transcribed and translated. In preliminary experiments, we coinfiltrated a mixture of three Agrobacteria strains, each harboring one of the three Arabidopsis NHP biosynthesis genes, into leaves of N. benthamiana and observed significant accumulation of NHP using gas chromatography–mass spectrometry (GC-MS) (Fig. 1B). Subsequent optimization revealed that SARD4 was not necessary for NHP production in N. benthamiana leaves because NHP accumulation was similar in leaves expressing ALD1, SARD4, and FMO1 (ALD1-SARD4-FMO1) and those expressing only ALD1 and FMO1 (ALD1-FMO1) (Fig. 1B). This result suggests that N. benthamiana leaves provide sufficient endogenous reductase activity for the conversion of dehydro-Pip to Pip.

To determine biosynthetic pathway specificity and the effect of coexpressing ALD1 and FMO1 on leaf metabolism, we performed untargeted metabolite analysis to identify compounds aside from NHP that changed after transgene expression. Tandem mass spectrometry (MS/MS) analysis of metabolites in untreated leaves and those transiently expressing ALD1 and FMO1 revealed two sets of compounds that appeared upon expression of the transgenes. We propose that the first group is structurally related to NHP based on MS/MS analysis (fig. S1, A to C, compounds 1 to 4). Several of these (compounds 1 to 3) also accumulated when an authentic synthetic standard of NHP was heated in vitro (fig. S1, D and E), suggesting that they may be degradation products and potentially formed during the generation of leaf extracts. The second group was likely derived from proline and depended on FMO1, but not ALD1, expression. MS/MS analysis supported structures related to hydroxyproline (fig. S1, B and C, compounds 5 to 7). One of these metabolites (fig. S1B, compound 3) has previously been reported as an FMO1-dependent product in pathogen-elicited responses in Arabidopsis (10). Collectively, these data show that two genes, ALD1 and FMO1, from Arabidopsis can support production of NHP in a heterologous system with minimal disruption of native metabolism.

Application of exogenous NHP induces SAR in solanaceous plants

Given that NHP has only been studied in the context of the Arabidopsis model system, we next asked whether NHP is the key metabolite initiating SAR in tomato, a valuable vegetable crop worldwide and a member of the same family as N. benthamiana (Solanaceae). We first tested whether NHP was present in tomato plants and whether exogenous application of synthetic NHP provoked an SAR-like response in adult tomato leaves. We detected significantly increased amounts of NHP in hydroponically grown tomato seedlings after exposure to P. syringae pathovar tomato (Pst), a virulent strain that infects VF36 tomato plants, whereas amounts in adult leaves remained low after a similar Pst treatment (fig. S2). These data show that tomato plants have an endogenous capacity to produce NHP at the seedling stage and imply that a mechanism to detect and respond to NHP has also evolved in tomato. Furthermore, the low background abundance of NHP in adult tomato leaves suggested that supplementation of plants with exogenous NHP has the potential to enhance the endogenous SAR response.

To test this, we assessed the bioactivity of purified NHP in fully expanded leaves of a 4- to 5-week-old tomato plant (Solanum lycopersicum cultivar VF36). Tomato plants have compound leaves consisting of five leaflets attached to a common stem (Fig. 2A). We used 1 mM NHP because previous studies have reported that 1 mM of the precursor Pip provides the most protection in an SAR assay in Arabidopsis (12, 18). Solutions of MgCl2 alone (mock) or MgCl2 and 1 mM NHP were infiltrated into the two leaflets closest to the main stem of the plant (referred to as the “bottom” leaflets) (Fig. 2A). Twenty-four hours later, the remaining three leaflets (referred to as the “top” leaflets) were challenged with Pst (Fig. 2A). At 4 days post-inoculation (dpi), the titer of Pst in the top leaflets was quantified (Fig. 2B), and leaf symptoms were documented (Fig. 2C). The titer of Pst was significantly lower in the top leaflets of leaves with NHP-treated bottom leaflets compared to those of leaves with mock-treated bottom leaflets (Fig. 2B). In addition, disease symptoms (such as bacterial speck, leaf yellowing, and leaf necrosis) were reduced in the Pst-infected top leaflets of leaves with NHP-treated bottom leaflets compared to those of leaves with mock-treated bottom leaflets (Fig. 2C). These data show that treatment of NHP in bottom leaflets is sufficient to induce disease resistance in neighboring top leaflets of the same leaf, providing evidence that NHP functions as a bioactive SAR signal in tomato.

Fig. 2 Local NHP application induces systemic defense in tomato leaves.

(A) Diagram showing design of tomato (S. lycopersicum cultivar VF36) leaf treatment for SAR experiments. Two bottom leaflets of a tomato leaf were treated with 1 mM NHP or vehicle only (mock) for 24 hours, and then, the three top leaflets of the same leaf were inoculated with Pseudomonas syringae pathovar tomato strain DC3000 (Pst). (B) Titer of Pst in top leaflets at 4 days post-infection (dpi). Bars represent means ± SD. n = 3 independent biological replicates per condition per trial. Trials 1 and 2 indicate batches of plants grown months apart. Asterisks denote the significant differences between indicated samples using a one-tailed t test (**P < 0.01). (C) Disease symptoms of top leaflets infected with Pst at 4 dpi. For each treatment, panels show symptoms of a whole leaflet (left) and an enlarged region of the respective leaflet (right). Scale bars, 1 cm. Images are representative of n = 3 independent biological replicates per condition per trial. Trials 1 and 2 indicate batches of plants grown months apart.

To extend our analysis to an additional solanaceous crop plant, we tested the bioactivity of NHP in sweet pepper (Capsicum annuum cultivar Early Calwonder). We infiltrated two lower, fully expanded leaves of 4- to 5-week-old peppers with MgCl2 solution (mock) or MgCl2 solution containing 2 mM NHP. We used 2 mM NHP for this assay because it was previously reported that 2 mM Pip induces SAR in Arabidopsis (18), and our SAR assay for pepper plants examines signaling between simple leaves on separate branches. Twenty-four hours later, an upper leaf was inoculated with a suspension of Xanthomonas euvesicatoria strain 85-10 (Xe 85-10), a virulent strain that infects pepper plants (fig. S3A). At 10 dpi, there was a significant decrease in Xe 85-10 growth (fig. S3B) and symptom development (fig. S3C) in pepper plants treated with NHP compared to mock-treated plants. In combination with our results in tomato, these data provide strong evidence that component(s) that detect NHP and activate SAR are present in the Solanaceae.

Expression of the Arabidopsis NHP biosynthetic pathway is protective in S. lycopersicum

Transient expression of the NHP biosynthetic pathway in N. benthamiana presents an opportunity to quickly determine whether biosynthetically produced NHP could protect against pathogen infection to crop plants naïve to the infecting pathogen. In other words, can SAR be established with the constitutive expression of just two NHP biosynthetic genes, ALD1 and FMO1? We chose to test this concept in tomato, given that these plants are amenable to Agrobacterium-mediated transient gene expression (19) and our finding that they were responsive to exogenous NHP.

Similar to the SAR assay developed to test the activity of synthetic NHP (Fig. 2A), we used an individual branch of a fully expanded tomato leaf. For each leaf, two bottom leaflets were infiltrated with a suspension of Agrobacteria carrying green fluorescent protein (GFP) alone or GFP and the two Arabidopsis NHP biosynthesis genes ALD1 and FMO1 (Fig. 3A). At 48 hpi, we performed metabolite profiling using GC-MS to monitor Pip and NHP in Agrobacteria-infected bottom leaflets and uninfected top leaflets. We detected both Pip and NHP in infected bottom leaflets (Fig. 3B). Significantly higher amounts of NHP were present in the bottom leaflets treated with Agrobacteria carrying GFP, ALD1, and FMO1 compared to Agrobacteria carrying GFP alone (Fig. 3B and fig. S4). During simultaneous transient expression of GFP, ALD1, and FMO1, NHP accumulated to amounts similar to those in tomato leaves treated with 1 mM NHP 24 hours earlier (fig. S4). Only low amounts of Pip were detected in the uninfected top leaflets (Fig. 3B).

Fig. 3 Engineering the NHP biosynthesis pathway in tomato leaves.

(A) Diagram showing conditions for metabolite analysis. Two bottom leaflets of a tomato leaf were incubated with Agrobacteria carrying the gene encoding GFP or a combination of Agrobacteria carrying GFP, Arabidopsis ALD1, and Arabidopsis FMO1 (GAF) for 48 hours, and then, the two bottom treated leaflets and the three top untreated leaflets were harvested and analyzed by GC-MS. (B) EIC of Pip (m/z = 156) and NHP (m/z = 172), as determined by GC-MS, in leaflets from (A). Bars for bottom leaves represent means ± SD (one to two leaflets each from n = 3 independent plants). Bars for top leaves represent means ± SD (three leaflets each from n = 3 independent plants). Amounts reported as zero indicate no detection of metabolites. (C) Diagram showing conditions for the bacterial growth assay. Two bottom leaflets of a tomato leaf were treated with Agrobacteria strains as in (A). At 48 hpi, the three top untreated leaflets were inoculated with Pst. (D) The titer of Pst in the top leaflets was determined 4 days later. Bars represent means ± SD (three leaflets each from n = 3 independent plants). Asterisks denote the significant differences between indicated samples using a one-tailed t test (*P < 0.05 and ***P < 0.001). Trials 1 and 2 indicate batches of plants grown months apart.

Increasing NHP production by transgene expression in the bottom leaflets of tomato leaves provided protection in distal, top leaflets challenged with Pst (Fig. 3, C and D). Transient expression of ALD1 and FMO1 provided similar protection as expression of ALD1, SARD4, and FMO1 (fig. S5, A and B), indicating that overexpression of SARD4 is not necessary for initiating defense priming in tomato leaves under these conditions. Infection of bottom leaflets by Agrobacteria expressing GFP alone did not significantly affect Pst growth in top leaflets compared to the mock control (Agrobacterium induction medium), further indicating that ALD1 and FMO1 expression was responsible for enhanced protection (fig. S5, C and D). Together, these data demonstrate that engineering a minimal set of enzymes for the Arabidopsis NHP biosynthetic pathway into tomato not only enhanced NHP production but also primed SAR in distal leaf tissue.

NHP biosynthetic enzymes are conserved in the plant kingdom

Our results in tomato prompted us to determine whether NHP is likely to function as a SAR signal throughout the plant kingdom. To analyze conservation of the NHP biosynthetic pathway, we used the National Center for Biotechnology Information (NCBI) protein Basic Local Alignment Search Tool (BLAST) ( to compare the percent amino acid identity between the three proteins associated with NHP biosynthesis in Arabidopsis (ALD1, SARD4, and FMO1) and their closest homologs in 50 other plant species (Fig. 4A and table S1). We then performed a reverse BLAST of each homolog back to the Arabidopsis proteome to identify reciprocal BLAST pairs, a common criterion for determining orthologous proteins (20). We found that 37 of the 50 plant species analyzed contain an entire orthologous NHP biosynthetic pathway (ALD1, SARD4, and FMO1) and that all but one (the moss, Physcomitrella patens) harbors an orthologous FMO1 protein (Fig. 4A).

Fig. 4 Evidence for NHP production across the plant kingdom.

(A) Phylogenetic tree of sequenced plant genomes created using the PhyloT tree generator ( Circles are scaled by the percent amino acid identity of the best BLAST (blastp) hit between the Arabidopsis thaliana NHP biosynthetic proteins (yellow) and the ALD1, SARD4, and FMO1 homologs of the respective plant proteome (black). Solid circles signify that the homolog returned the respective A. thaliana protein in a reverse BLAST into the A. thaliana proteome (best reciprocal BLAST). (B) Detection of endogenous NHP production and FMO1 homolog activity in six plant species. Dashed lines label the species and common name, respectively, for plants examined: N. benthamiana, tobacco relative; S. lycopersicum, tomato; A. thaliana, thale cress; B. rapa, mustard; G. max, soybean; Zea mays, corn. To measure endogenous NHP, seedlings were grown hydroponically and then mock-treated (−Pst; gray) or treated with Pseudomonas syringae pathovar tomato strain DC3000 to induce SAR (+Pst; red). NHP was measured 48 hours later by GC-MS. Bars indicate EIC for NHP (m/z = 172). Asterisks indicate a significant NHP increase in Pst-treated plants (one-tailed t test; *P < 0.05). Pip and NHP-Glc abundances for this experiment were also measured (fig. S7). To measure enzyme activity, the FMO1 homologs from each plant with highest identity with Arabidopsis FMO1 were transiently expressed in N. benthamiana leaves in the presence of 1 mM Pip. Bars (blue) indicate EIC of NHP measured in leaves expressing the respective FMO1 homolog. Pip amounts were also measured, and an Arabidopsis FMO1 mutant (10) was included as a control (fig. S8). Bars represent the means ± SD (n = 3 independent biological replicates). Values reported as zero indicate no detection of metabolites.

To understand how the conservation of NHP biosynthesis compares to other signaling and defense metabolites, we performed a similar analysis using the biosynthetic enzymes for salicylic acid (SA) and jasmonic acid (JA), two widely conserved defense hormones, and indole glucosinolate (GSL), a Brassicaceae-specific defense metabolite. We tracked conservation of isochorismate synthase 1 (21) for SA, 12-oxophytodienoate (OPDA) reductase (22) and jasmonate resistance 1 (23) for JA, and S-alkyl-thiohydroximate lyase (24) and cytochrome P450 monooxygenase CYP83B1 (25) for GSL and compared the profiles of homologous proteins across the plant kingdom with those in the NHP biosynthetic pathway (fig. S6). The pattern and relative degree of sequence similarity of best reciprocal BLAST orthologs in the NHP pathway corresponds with the pathway patterns for SA and JA but differs from the Brassicaceae-specific GSL pathway (fig. S6). This analysis provides evidence that NHP biosynthesis is not restricted to the Brassicaceae but is conserved throughout the plant kingdom.

NHP is present in species widely distributed across the plant kingdom

In addition to bioinformatic analysis, we also compared the abundance of endogenous and pathogen-elicited NHP pathway metabolites in Arabidopsis with that of four dicot plants [Brassica rapa (field mustard), Order Brassicales; N. benthamiana (tobacco relative) and S. lycopersicum (tomato), Order Solanales; and Glycine max (soybean), Order Fabales] and one monocot plant [Zea mays (corn), Order Poaceae] (Fig. 4B). Seedlings were grown axenically in hydroponic media and treated with Pst. We then quantified basal and Pst-elicited biosynthesis of Pip, NHP, and NHP-glucoside (NHP-Glc), a glycosylated conjugate of NHP, present in seedling extracts (10).

We detected NHP in untreated seedlings of N. benthamiana, tomato, soybean, and corn, but not B. rapa (Fig. 4B), and the amounts of NHP increased in seedlings of N. benthamiana, tomato, and B. rapa after Pst elicitation (Fig. 4B). By contrast, NHP abundances in Pst-treated seedlings of soybean and corn were similar to those in the untreated seedlings (Fig. 4B). Amounts of Pip also increased in response to Pst treatment (fig. S7). NHP-Glc only accumulated in Arabidopsis and B. rapa (fig. S7). These data reveal that, in addition to Arabidopsis, five diverse species produce NHP. They also provide evidence that accumulation of NHP increases in tissues in response to pathogens outside of the Brassicaceae, whereas the production of NHP-Glc may be unique to the Brassicaceae (fig. S7).

FMO1 orthologs from across the plant kingdom catalyze the N-hydroxylation of pipecolic acid

In parallel to metabolite profiling, we also tested the in planta biochemical activity of the putative FMO1 orthologs identified from five plant species. We cloned each FMO1-like gene (BraA06g014860 from B. rapa, Glyma13g17340 from G. max, Solyc07g042430 from S. lycopersicum, Niben101Scf05682g00009 from N. benthamiana, and AC191071.3_FGP001 from Z. mays) and transiently expressed each in N. benthamiana leaves using Agrobacteria. Arabidopsis FMO1 was used as a positive control. Leaves were also infiltrated with 1 mM Pip to provide ample substrate for the reaction. Expression of each putative FMO1 ortholog led to the production of NHP (Fig. 4B and fig. S8). NHP was the major metabolite identified by GC-MS for all plant FMO1-like enzymes tested, and no differences in metabolite profiles were apparent in the respective leaf extracts in comparison to those expressing Arabidopsis FMO1. Leaves expressing GFP (negative control) or an inactive Arabidopsis FMO1 mutant (FMO1 G17A-G19A) (10) did not produce any NHP (fig. S8). These data show that FMO1-like genes from diverse plant species encode orthologous proteins that can catalyze the N-hydroxylation of Pip when expressed heterologously, providing biochemical evidence that monooxygenases capable of producing NHP are conserved across the plant kingdom.


Several aspects of plant defense, including pathogen detection, immune signaling, and antimicrobial production, have been the targets of engineering in an effort to enhance endogenous immunity in crops. For example, since the discovery that the Arabidopsis elongation factor Tu receptor could function in solanaceous plants (26), many immune receptor genes have been transferred between plant families to expand native pathogen detection (3). An alternative strategy for engineering disease resistance has focused on introducing biosynthetic genes that produce small-molecule phytoalexins with direct antimicrobial activities (27). This approach has been successful, especially with the well-studied phytoalexin resveratrol, which has been engineered into numerous plants to enhance pathogen resistance (28). Other small-molecule–based approaches have used direct chemical priming, wherein application of chemical inducers enhances endogenous disease resistance mechanisms in a manner akin to SAR. This approach allows bypassing of species-specific pathogen detection instead of relying on external induction of a plant’s native defense response. Such a strategy could enable protection against the emergence of a pathogen before contact with a crop plant and/or enhanced resistance against adapted pathogens that actively suppress the plant host defense responses directly downstream of detection. Several synthetic chemical inducers of this type have met success in commercial agriculture and are being used in combination with antimicrobials in the field (29). It is notable that these commercial inducers, such as benzothiadiazole (30), are synthetic and cannot be made in plants using an engineered biosynthesis approach similar to that demonstrated here for NHP. Deployment of a combination of multiple strategies may be the most effective way to achieve durable disease resistance in agriculture.

By transiently expressing Arabidopsis ALD1 and FMO1 in N. benthamiana, NHP production can be increased 100- to 1000-fold over its accumulation in native plants (Fig. 1 and fig. S7), suggesting that engineering plants to overproduce NHP could induce disease resistance. As a proof of concept, we have shown that the NHP biosynthetic pathway can be engineered under a high-expression constitutive promoter. In the future, stable engineering approaches could use transgenic plants in which expression of the NHP biosynthesis genes can be activated by exogenous chemical inducers or natural plant regulatory systems to minimize growth-yield trade-offs that may come with constitutive NHP production (31).

In conclusion, our results show that NHP is not only synthesized in diverse species across plant phylogeny but also provides SAR protection in agriculturally important plants. Our work demonstrates that metabolic engineering of the NHP pathway is a promising approach to induce innate immune responses in crops. Furthermore, we anticipate that NHP pathway engineering will also enable new studies to probe fundamental questions in SAR, including how small molecules interact to establish functional SAR, how signals are transported, perceived, and attenuated, and how SAR mechanisms have been conserved and diversified throughout plant evolution.


Plant materials and growth conditions

For seedling hydroponic experiments, 15 ± 1 A. thaliana ecotype Col-0 seeds, two rapid-cycling B. rapa seeds, two S. lycopersicum Heinz 1706 seeds, 10 ± 1 N. benthamiana Nb-1 seeds, one G. max Pella 86 seed, and one Z. mays B73 seed were placed into 3 ml of sterile 1× Murashige-Skoog medium with vitamins (PhytoTechnology Laboratories) (pH 5.7) and sucrose (5 g/liter) in the wells of a six-well microtiter plate. Plates were kept in the dark at room temperature for 48 hours and then transferred to a growth chamber at 50% humidity, 22°C, and photon flux (100 μmol/m2 per s) under a 16-hour light/8-hour dark cycle. After 1 week of growth, spent medium was replaced with 3 ml of fresh Murashige-Skoog medium with sucrose (5 g/liter). B. rapa, S. lycopersicum, G. max, and Z. mays seedlings were elicited at 1 week. A. thaliana and N. benthamiana seedlings were elicited at 2 weeks. For adult plant experiments, tomato (S. lycopersicum cultivar VF36), pepper (C. annuum cultivar Early Calwonder), and N. benthamiana plants were grown in a greenhouse (16-hour light/8-hour dark, 25° to 28°C) and used at 4 to 6 weeks of age.

Bacterial strains and growth conditions

Escherichia coli strains DH5-α and 10-β, Pst strain DC3000, X. euvesicatoria strain Xe 85-10, and A. tumefaciens strains GV3101 and C58C1 pCH32 were used in this study. E. coli strains were grown in lysogeny broth (LB) agar containing appropriate antibiotics at 37°C. Pseudomonas and Xanthomonas strains were grown at 28°C on nutrient yeast glycerol agar (NYGA) medium containing rifampicin (100 μg/ml). Agrobacteria strains were grown at 28°C on LB agar containing rifampicin (100 μg/ml), tetracycline (5 μg/ml), and kanamycin (50 μg/ml) for C58C1 pCH32 and gentamycin (100 μg/ml) and kanamycin (50 μg/ml) for GV3101.

Elicitation methods

For seedling hydroponic experiments, Pst was grown overnight on LB agar plates at 30°C. A single colony was picked and grown to mid-log phase in liquid LB at 30°C. Cells were centrifuged, washed three times with Murashige-Skoog with sucrose (5 g/liter), and resuspended in Murashige-Skoog with sucrose (5 g/liter) to an optical density at 600 nm (OD600) of 0.2. Twenty microliters of sterile 10 mM MgCl2 was added to wells of mock-elicited samples, and 20 μl of the Pst suspension was added to each treatment well. Plants were elicited 48 hours before sample harvest. For adult tomato experiments, one leaflet of the third compound leaf of 4- to 5-week-old S. lycopersicum VF36 plants was infiltrated with mock buffer (10 mM MgCl2) or 10 mM MgCl2 containing suspension of Pst (1 × 107 colony-forming units (CFU)/ml). Two days later, the treated leaves were harvested for metabolites profiling by GC-MS analysis. Three biological repeats were performed per treatment in this experiment.

Plant tissue extraction for metabolite profiling

All plant tissue was flash-frozen, lyophilized to dryness, and homogenized at 25 Hz for 2 min using a ball mill (Retsch MM 400). Samples were resuspended in 20 μl of 80:20 MeOH:H2O per milligram of dry tissue and incubated at 4°C for 10 min. The liquid fraction of each sample was split, and one half was passed through a 0.45-μm polytetrafluoroethylene (PTFE) filter and directly analyzed with liquid chromatography–mass spectrometry (LC-MS). The other half was dried under an N2 stream and further processed for GC-MS analysis. Dried samples were resuspended in 250 μl of a 10% (v/v) N-methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA) with 1% trimethylchlorosilane (Sigma-Aldrich) solution in hexanes, briefly vortexed, and incubated at 70°C for 30 min. Samples were then allowed to cool to room temperature, filtered through 0.45-μm PTFE filters, and analyzed with GC-MS.

Cloning of plant FMO1 homologs and A. thaliana NHP biosynthetic genes

FMO1 homologs from S. lycopersicum cultivar VF36 and N. benthamiana were polymerase chain reaction (PCR)–amplified using complementary DNA (cDNA) isolated from plant leaves and gene-specific primers (table S2) and then cloned into the pCR8/GW/TOPO vector (Life Technologies, Carlsbad, CA). cDNA sequences were confirmed by DNA sequence analysis and then subcloned into pEAQ-HT-DEST3 (32) to create C-terminal 6× His-tagged fusion proteins. Plasmids were then transformed into E. coli DH5-α and A. tumefaciens strain C58C1 pCH32 by heat shock transformation.

FMO1 cDNA homologs from B. rapa (NCBI: XP_009149401.1; BraA06g014860), G. max (NCBI: XP_003541317.1; Glyma13g17340), and Z. mays (NCBI: XP_008660479.1; AC191071.3_FGP001) were codon-optimized for expression in N. benthamiana using Integrated DNA Technologies’ codon optimization tool (, which eliminates codons with a usage rate of <10%, assigns codons with a frequency matching that of the target organism, and further optimizes to streamline synthesis of the gene. The optimized genes were synthesized as gBlock fragments (Integrated DNA Technologies), cloned into pEAQ-HT (32) under the control of the cauliflower mosaic virus (CaMV) 35S promoter, and transformed into E. coli 10-β and A. tumefaciens GV3101 by heat shock transformation.

A. thaliana ALD1, SARD4, and FMO1 cDNAs were amplified by PCR using A. thaliana cDNA and gene-specific primers (table S1). These products were cloned into pEAQ-HT (32) under control of the CaMV 35S promoter and transformed into E. coli 10-β, A. tumefaciens GV3101, and A. tumefaciens C58C1 pCH32 by heat shock transformation.

Transient expression in N. benthamiana

A. tumefaciens strains engineered with NHP biosynthetic gene constructs were grown for 48 hours on LB agar plates with appropriate antibiotics. Cells were scraped from the surface of the plate using an inoculation loop, washed three times, resuspended in Agrobacterium induction medium [10 mM MES buffer, 10 mM MgCl2, and 150 μM acetosyringone (pH 5.7)], and incubated at room temperature with shaking for 2 hours. For transient expression of plant FMO1 homologs, cells were diluted to a final OD600 of 0.3 in induction medium and 1 mM Pip. For transient expression of A. thaliana NHP pathway genes, cells were combined in equal proportions of OD600 of 0.1 with a GFP expressing strain for a final OD600 of 0.4 in induction medium. These solutions were infiltrated into the underside of N. benthamiana leaves using a needleless 1-ml syringe. Plants were returned to the growth shelf for 24 hours before tissue harvest and processing.

LC-MS metabolite profiling

NHP-Glc was measured using a previously documented reversed-phase (RP) chromatography method on an Agilent 1260 HPLC coupled to an Agilent 6520 quadrupole time-of-flight electrospray ionization (Q-TOF ESI) mass spectrometer (10). Experimental parameters for compounds reported in this study are 3 [RP LC-MS; retention time (rt) = 0.9 min, collision-induced dissociation (CID) = 40 V]; 4 (RP LC-MS; rt = 1.84 min, CID = 10 V); 7 (RP LC-MS; rt = 0.9 min; CID = 40 V); NHP [CID = 10 V, included for (10)] (fig. S1, A to C). MS data have been deposited into the MetaboLights database ( with accession number MTBLS1099.

GC-MS metabolite profiling

Plant extracts derivatized with MSTFA were analyzed on an Agilent 7820A gas chromatograph coupled to an Agilent 5977B mass spectrometer. Samples were flown through a fused silica column (30 m by 0.25 mm by 0.1 μm; Agilent VF-5ht) using the following oven temperature gradient: 70°C for 2 min, ramp to 150°C at 10°C/min, and final hold at 320°C for 2 min. To detect Pip [2-trimethylsilyl (2-TMS) derivative] and NHP (2-TMS derivative), the instrument was run in both scan and selected ion monitor (SIM) modes where the mass/charge ratio (m/z) values of the selected ions were 156, 230, and 273 for Pip and m/z 172, 246, and 274 for NHP. Reported values for ion abundance are peak integration values of extracted ion chromatograms for Pip (m/z = 156) and NHP (m/z = 172) detected in SIM mode. Fragmentation patterns and retention times for in planta Pip and NHP were verified with authentic standards of Pip (Oakwood Chemical) and NHP (10). Experimental parameters for compounds reported in this study are as follows: 1 (GC-MS, rt = 9.41 min); 2 (GC-MS, rt = 6.55 min); 5 (GC-MS, rt = 8.25 min); 6 (GC-MS, rt = 5.83 min); NHP-2TMS (GC-MS, rt = 9.2 min) (fig. S1, A to C). MS data have been deposited into the MetaboLights database ( with accession number MTBLS1099.

Transient expression of Arabidopsis NHP biosynthesis genes in tomato

A. tumefaciens strains C58C1 pCH32 carrying A. thaliana NHP biosynthetic gene constructs (ALD1, FMO1, and/or SARD4) or an Arabidopsis FMO1 inactive flavin adenine dinucleotide–binding mutant [At FMO1 mutant, G17A-G19A (5)] or S. lycopersicum FMO1 (Sl FMO1) were grown on LB agar plates with appropriate antibiotics. Cells were scraped from the surface of the plate, washed one to two times, resuspended in Agrobacterium induction medium [10 mM MES buffer, 10 mM MgCl2, 150 μM acetosyringone (pH 5.7)], and incubated at room temperature for 2 hours. Cells of a single strain were inoculated at a final OD600 of 0.1 in induction medium. Cells of different strains were combined for a final OD600 of 0.3 in induction medium. Two leaflets of third or fourth compound leaves of 4- to 5-week-old S. lycopersicum VF36 plants were infiltrated with each suspension (one suspension per plant). Plants were placed under light (16-hour light/8-hour dark or continuous light) for 24 or 48 hours before tissue harvest or bacterial growth assays. For bacterial growth assays, three terminal leaflets on the same branch were inoculated with a suspension of Pst (1 × 105 CFU/ml). Plants were then incubated in a greenhouse (16-hour light/8-hour dark, 25° to 28°C). The titer of Pst in the terminal leaflets was quantified 2 or 4 dpi by homogenizing leaf discs in 1 ml of 10 mM MgCl2, plating appropriate dilutions on NYGA medium with rifampicin (100 μg/ml), incubating plates at 28°C for 2 days, and then counting bacterial colonies. Three biological repeats were performed per treatment in two independent experiments.

SAR assay in tomato

Two bottom leaflets (closest to main stem) of third or fourth compound leaves of 4- to 5-week-old S. lycopersicum VF36 plants were infiltrated with 10 mM MgCl2 (mock) or 10 mM MgCl2 containing 1 mM NHP. After chemical incubation for 24 hours, three top leaflets (distal to main stem) on the same branch were inoculated with a suspension of Pst (1 × 105 CFU/ml). Plants were incubated in greenhouse (16-hour light/8-hour dark, 25° to 28°C). The titer of Pst in the bottom leaflets was quantified at 4 dpi by homogenizing leaves discs in 1 ml of 10 mM MgCl2, plating appropriate dilutions on NYGA medium with rifampicin (100 μg/ml), incubating plates at 28°C for 2 days, and then counting bacterial colonies. Symptoms of infected leaves were photographed at 4 dpi. Three biological repeats were performed per treatment in two independent experiments.

SAR assay in pepper

Two bottom leaves of 4- to 5-week-old pepper plants were infiltrated with 10 mM MgCl2 or 10 mM MgCl2 containing 2 mM NHP. After chemical incubation for 24 hours, one upper leaf was inoculated with a suspension of Xe 85-10 (1 × 104 CFU/ml). Plants were incubated in a greenhouse (16-hour light/8-hour dark, 25° to 28°C). The titer of Xe 85-10 in the infected leaves was quantified 5 and 10 dpi by homogenizing leaves discs in 1 ml of 10 mM MgCl2, plating appropriate dilutions on NYGA medium with rifampicin (100 μg/ml), incubating plates at 28°C for 2 days, and then counting bacterial colonies. Symptoms of infected leaves were photographed at 10 dpi. Four biological repeats were performed per treatment in two independent experiments.

Sequences of codon-optimized FMO1 genes

B. rapa FMO1 (optimized from XP_009149401.1):


G. max FMO1 (optimized from XP_003541317.1):


Z. mays FMO1 (optimized from XP_008660479.1):



Fig. S1. Abundances of unknown compounds during transient expression of Arabidopsis NHP biosynthesis genes in N. benthamiana.

Fig. S2. Production of Pip and NHP in tomato plants in response to Pst.

Fig. S3. Local NHP treatment induces systemic defense in pepper plants.

Fig. S4. Production of NHP during transient expression in tomato leaves.

Fig. S5. Engineering NHP biosynthesis in tomato leaves.

Fig. S6. BLAST analysis of plant defense metabolites.

Fig. S7. Production of Pip, NHP, and NHP-Glc in seedlings in response to Pst.

Fig. S8. Production of NHP catalyzed by FMO1 homologs transiently expressed in N. benthamiana leaves.

Table S1. Percent amino acid identity of best BLAST hits to A. thaliana proteins.

Table S2. Primers used in this study.


Acknowledgments: We thank G. Lomonossoff (John Innes Centre) for providing the pEAQ plasmid. Funding: This work was supported by an HHMI and Simons Foundation Grant 55108565 (to E.S.S.), National Science Foundation Graduate Research Fellowship DGE-1656518 (to E.C.H.), National Science Foundation IOS-1555957 and Binational Science Foundation Grant 2011069 (to M.B.M.), and Ministry of Science and Technology of Taiwan-105-2917-I-564-414 093 (to Y.-C.C.). Author contributions: E.C.H., Y.-C.C., E.S.S., and M.B.M. contributed to the study design; E.C.H. and Y.C.C. performed research; E.C.H., Y.-C.C., E.S.S., and M.B.M. analyzed data; and E.C.H., Y.-C.C., E.S.S., and M.B.M. wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data materials and availability: MS data have been deposited into the MetaboLights database ( with accession number MTBLS1099. A material transfer agreement was required to obtain the pEAQ vector from G. Lomonossoff (John Innes Centre). All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.

Stay Connected to Science Signaling

Navigate This Article