
==== Front
J Biol Chem
J Biol Chem
The Journal of Biological Chemistry
0021-9258
1083-351X
American Society for Biochemistry and Molecular Biology

S0021-9258(24)02168-9
10.1016/j.jbc.2024.107667
107667
Research Article Collection: Plant Biology
In-depth analysis of isochorismate synthase-derived metabolism in plant immunity: Identification of meta-substituted benzoates and salicyloyl-malate
Scholten Nicola 1
Hartmann Michael 1
Abts Sarah 1
Abts Laura 1
Reinartz Elke 1
Altavilla Angelo 2
Müller Thomas J.J. 2
Zeier Jürgen juergen.zeier@hhu.de
13∗
1 Department of Biology, Institute for Molecular Ecophysiology of Plants, Heinrich Heine University, Düsseldorf, Germany
2 Department of Chemistry, Institute of Organic Chemistry and Macromolecular Chemistry, Heinrich Heine University, Düsseldorf, Germany
3 Cluster of Excellence on Plant Sciences (CEPLAS), Heinrich Heine University, Düsseldorf, Germany
∗ For correspondence: Jürgen Zeier juergen.zeier@hhu.de
12 8 2024
9 2024
12 8 2024
300 9 10766722 6 2024
24 7 2024
© 2024 The Authors
2024
https://creativecommons.org/licenses/by/4.0/ This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Isochorismate-derived metabolism enables biosynthesis of the plant defense hormone salicylic acid (SA) and its derivatives. In Arabidopsis thaliana, the stress-induced accumulation of SA depends on ISOCHORISMATE SYNTHASE1 (ICS1) and also requires the presumed isochorismate transporter ENHANCED DISEASE SUSCEPTIBILITY5 (EDS5) and the GH3 enzyme avrPphB SUSCEPTIBLE3 (PBS3). By comparative metabolite and structural analyses, we identified several hitherto unreported ICS1- and EDS5-dependent, biotic stress-inducible Arabidopsis metabolites. These involve meta-substituted SA derivatives (5-formyl-SA, 5-carboxy-SA, 5-carboxymethyl-SA), their benzoic acid (BA) analogs (3-formyl-BA, 3-carboxy-BA, 3-carboxymethyl-BA), and besides the previously detected salicyloyl-aspartate (SA-Asp), the ester conjugate salicyloyl-malate (SA-Mal). SA functions as a biosynthetic precursor for SA-Mal and SA-Asp, but not for the meta-substituted SA- and BA-derivatives, which accumulate to moderate levels at later stages of bacterial infection. Interestingly, Arabidopsis leaves possess oxidizing activity to effectively convert meta-formyl- into meta-carboxy-SA/BAs. In contrast to SA, exogenously applied meta-substituted SA/BA-derivatives and SA-Mal exert a moderate impact on plant immunity and defence-related gene expression. While the isochorismate-derived metabolites are negatively regulated by the SA receptor NON-EXPRESSOR OF PR GENES1, SA conjugates (SA-Mal, SA-Asp, SA-glucose conjugates) and meta-substituted SA/BA-derivatives are oppositely affected by PBS3. Notably, our data indicate a PBS3-independent path to isochorismate-derived SA at later stages of bacterial infection, which does not considerably impact immune-related characteristics. Moreover, our results argue against a previously proposed role of EDS5 in the biosynthesis of the immune signal N-hydroxypipecolic acid and associated transport processes. We propose a significantly extended biochemical scheme of plant isochorismate metabolism that involves an alternative generation mode for benzoate- and salicylate-derivatives.

Keywords

salicylic acid derivatives
isochorismate synthase
ICS1
EDS5
PBS3
NPR1
salicyloyl-malate
meta-substituted aromatics
Arabidopsis thaliana
plant immunity
Abbreviations

3-CBA 3-carboxy-benzoic acid

3-CMBA 3-carboxymethyl-benzoic acid

3-FBA 3-formyl-benzoic acid

5-CMSA 5-carboxymethyl-salicylic acid

5-CSA 5-carboxy-salicylic acid

5-FSA 5-formyl-salicylic acid

ALD1 AGD2-LIKE DEFENSE RESPONSE PROTEIN1

BA benzoic acid

BSMT1 BENZOIC ACID SALICYLIC ACID METHYLTRANSFERASE

DHBA dihydroxybenzoic acid

EDS5 ENHANCED DISEASE SUSCEPTIBILITY5

FMO1 FLAVIN-DEPENDENTMONOOXYGENASE1

FTIR Fourier transform infrared spectroscopy

GC-MS gas chromatography-mass spectrometry

GH3 Gretchen Hagen 3

hpi hours post-inoculation

hpt hours post-treatment

IC-Glu isochorismoyl-L-glutamate

ICS1 ISOCHORISMATE SYNTHASE1

IPL isochorismate pyruvate lyase

IR infrared

LC-MS liquid chromatography-mass spectrometry

LC-qTOF-MS liquid chromatography-quadrupole time-of-flight mass spectrometry

LOOH lipid hydroperoxides

M+ molecular ion

m/z mass-to-charge

MATE multidrug and toxin extrusion

MeSA methyl salicylate

NHP N-hydroxypipecolic acid

NHPG N-hydroxypipecolic acid glucoside

NHPGE N-hydroxypipecolic acid glucose ester

NPR1 NON-EXPRESSOR OF PR GENES1

PAL phenylalanine ammonia lyase

PBS3 avrPphB SUSCEPTIBLE3

Pip pipecolic acid

PR1 PATHOGENESIS-RELATED1

Psm Pseudomonas syringae pv. maculicola ES4326

Pyr-Glu N-pyruvoyl-L-glutamate

qTOF-MS quadrupole time-of-flight mass spectrometry

rlu relative light unit

S3H salicylic acid-3-hydroxylase

S5H salicylic acid-5-hydroxylase

SA salicylic acid

SA-Asp salicyloyl-aspartate

SA-Mal salicyloyl-malate

SAG salicylic acid-β-glucoside

SAR systemic acquired resistance

SARD4 SYSTEMIC ACQUIRED RESISTANCE-DEFICIENT4

SGE salicylic acid glucose ester

SID2 salicylic acid-induction-deficient2

UGT uridine diphosphate-dependent glycosyltransferase

VAS1 REVERSAL OF SAV3 PHENOTYPE 1

Reviewed by members of the JBC Editorial Board. Edited by Joseph Jez
==== Body
pmcPlants and many microorganisms can synthesize the aromatic amino acids phenylalanine, tyrosine, and tryptophan from erythrose-4-phosphate and phosphoenolpyruvate via the shikimate pathway (1). Chorismate is a key branch-point intermediate of this metabolic route. It can be enzymatically converted to prephenate by chorismate mutase toward the production of Phe and Tyr, to anthranilate for Trp biosynthesis by anthranilate synthase, and to aminodeoxychorismate by aminodeoxychorismate synthase, a precursor of the cofactor tetrahydrofolate (2, 3). In addition, isochorismate synthase (ICS) catalyzes the rearrangement of chorismate to isochorismate, a central precursor for the biosynthesis of salicylic acid (SA) and the photosystem I-resident redox cofactor phylloquinone (4, 5, 6, 7).

SA is a central plant stress hormone that mediates inducible defense responses associated with pattern- and effector-triggered immunity (8). Moreover, in an interplay with the L-Lys-derived immune signal N-hydroxypipecolic acid (NHP), SA guarantees the establishment of systemic acquired resistance (SAR), a state of primed immunity that develops throughout the plant foliage in response to a localized leaf inoculation (9, 10, 11, 12). SA-mediated defense responses require the transcriptional co-activator NON-EXPRESSOR OF PR GENES1 (NPR1) (8). The recognition of pathogens triggers SA biosynthesis (13), and the consequentially increased levels of SA result in the translocation of NPR1 from the cytosol to the nucleus (14). The binding of SA to NPR1 then induces the expression of a battery of defense-related genes for immune activation (15, 16).

Basically, plants are able to synthesize SA by two metabolic routes—the ICS and the phenylalanine ammonia lyase (PAL) pathways (17). The accumulation of SA in pathogen-inoculated Arabidopsis leaves essentially proceeds via the ICS pathway, which is initiated by the conversion of chorismate to isochorismate in the chloroplast via the stress-inducible ICS isoform ISOCHORISMATE SYNTHASE1 (ICS1) (5, 6, 18, 19). Although bacteria directly convert isochorismate to SA via isochorismate pyruvate lyase (IPL) activities (20), plants pursue a circuit strategy for this conversion (21, 22). The Arabidopsis Gretchen Hagen 3 (GH3) acyl acid amido synthetase avrPphB SUSCEPTIBLE3 (PBS3) first catalyzes the ATP-consuming conjugation of isochorismate and L-Glu to generate isochorismoyl-L-glutamate (IC-Glu) as an intermediate. Elimination of N-pyruvoyl-L-glutamate (Pyr-Glu) from IC-Glu, either spontaneously or by the action of the acyltransferase ENHANCED PSEUDOMONAS SUSCEPTIBILTY1, then results in SA formation. Since PBS3 is located in the cytosol, transport of isochorismate out of the chloroplast is necessary for the PBS3-mediated biosynthesis of SA. The chloroplast envelope-resident multidrug and toxin extrusion (MATE) transporter ENHANCED DISEASE SUSCEPTIBILITY5 (EDS5) has been implicated in this transport process (21). Just as ICS1, EDS5 is required for the biotic stress-induced biosynthesis of SA in Arabidopsis (18, 23). Initially, assays with isolated chloroplasts suggested that EDS5 might enable the transport of SA out of the chloroplast (24). However, more recent results on the metabolic phenotypes of Arabidopsis lines expressing differentially targeted ICS1- and PBS3-fusion proteins in the eds5 mutant background strongly suggest a function for EDS5 in isochorismate transport out of the chloroplast (21). Therefore, the current model of SA biosynthesis via the ICS pathway involves the ICS1-catalyzed isomerization of chorismate to isochorismate in the chloroplast, EDS5-mediated transport of isochorismate across the chloroplast envelope, and PBS3-catalysed formation of IC-Glu, which subsequently decomposes to Pyr-Glu and SA (9).

The biochemical modification of a plant hormone generally alters its physiological activity. The glucosylation of SA to the biologically inactive SA-β-glucoside (SAG) and to SA glucose ester (SGE) are predominant SA metabolic pathways. Remarkably, the uridine diphosphate-dependent glycosyltransferase (UGT) UGT76B1 is responsible for the concerted glucosylation of both SA to SAG and NHP to NHP glucoside (NHPG) under biotic stress conditions, respectively, and in this way terminates SAR signaling (9, 25, 26). Two other UGTs, UGT74F1 and UGT74F2, contribute to basal SAG levels, and UGT74F2 also exhibits SGE-forming activity (27). Another prominent option of SA modification is its hydroxylation to 2,3-dihydroxybenzoic acid (2,3-DHBA) and 2,5-DHBA (gentisic acid), which is mediated by the 2-oxoglutarate-dependent dioxygenases SA-3-hydroxylase (S3H) and SA-5-hydroxylase (S5H), respectively (28, 29). The final metabolic products of the SA hydroxylation pathway are DHBA glucosides and xylosides (30, 31, 32). Moreover, SA is converted to its volatile methyl ester methyl salicylate (MeSA) by the jasmonic acid-inducible methyltransferase BENZOIC ACID SALICYLIC ACID METHYLTRANSFERASE1 (BSMT1) (33, 34). MeSA is a significant constituent of the floral scents of many plant species (35). Although MeSA has been suggested as a mobile signal for SAR in tobacco (36), the SAR-inducing abilities of MeSA-deficient Arabidopsis bsmt1 mutant plants argue against a general function for MeSA in SAR establishment (34). Finally, the aspartyl amide conjugate of SA, salicyloyl-aspartate (SA-Asp), was detected as a natural SA derivative in different plant species (37, 38, 39). In vitro biochemical assays suggest a role for the acyl acid amido synthetase GH3.5 in the formation of SA-Asp (40, 41, 42).

In the present study, we identify the occurrence of several hitherto undescribed SA and benzoic acid (BA) derivatives in leaf extracts of Arabidopsis which accumulate upon bacterial inoculation in an ICS1- and EDS5-dependent manner. These include two related sets of meta-substituted SA and BA derivatives carrying formyl-, carboxy- and carboxymethyl-moieties in their aromatic ring. Moreover, we identify the malate ester of SA, salicyloyl-malate (SA-Mal), as a natural SA derivative that accumulates together with SA-Asp and the SA biosynthesis-related metabolites IC-Glu and Pyr-Glu in pathogen-inoculated leaves in dependency of PBS3. We provide information about the biosynthetic origins of these substances, characterize their time-dependent accumulation following bacterial inoculation, and study their immune-inducing abilities. Our data further indicate the existence of an ICS1-dependent but PBS3-independent path to isochorismate-derived SA in Arabidopsis that operates at later stages of bacterial infection to some extent. We further show that ics1, eds5, and pbs3 knockout mutants show similar immune-related characteristics, suggesting that early production of ICS pathway-derived SA functions as a major determinant of plant immune responses. Our results also exclude a previously suggested function for EDS5 in NHP biosynthesis and putatively associated transport processes. By integrating our findings, we propose an extended biochemical scheme of isochorismate synthase-derived metabolism in plants.

Results

Identification of novel meta-substituted salicylic acid and benzoic acid derivatives in Arabidopsis thaliana leaves that accumulate upon pathogen inoculation in dependency of isochorismate synthase

The plastid-resident isochorismate synthase catalyzes the conversion of chorismate to isochorismate in plants (5, 6). The ICS1 gene, which is defective in the SA-biosynthetic mutant salicylic acid-induction-deficient2 (sid2) (18), is the biotic stress-inducible isochorismate synthase isoform of Arabidopsis (4, 19). To get further insights into ICS1-dependent metabolism, we inoculated leaves of wild-type Arabidopsis Col-0 and sid2 mutant plants with the compatible bacterial pathogen Pseudomonas syringae pv. maculicola ES 4326 (Psm), harvested leaves at 48 h post-treatment (hpt), and performed comparative gas chromatography-mass spectrometry (GC-MS)-based metabolite analyses of leaf extracts (Fig. 1). We thereby applied an analytical method that involves sample derivatization with trimethylsilyl-diazomethane to convert analytes with free carboxylic acid groups into methyl esters and thereby facilitate GC-based analyses (43, 44). Analyses of GC-MS ion chromatograms of distinct mass-to-charge ratios (m/z) identified, in addition to the previously analyzed SA and SA glucose conjugates (11), ten peaks of analytes (1a to 10a) that accumulated in the Col-0 wild-type but not in sid2 upon Psm-inoculation (Fig. 1A).Figure 1 Identification of meta-substituted salicylic acid (SA) and benzoic acid (BA) derivatives that accumulate in Arabidopsis leaves in dependence of ICS1.A, segments of overlaid ion chromatograms of extract samples from Arabidopsis thaliana Col-0 wildtype (blue) and sid2 mutant (red) leaves, as analysed by GC-MS and GC-FTIR. The sid2 mutant is defective in the ISOCHORISMATE SYNTHASE1 (ICS1) gene. Leaf samples were harvested 48 h after inoculation with Pseudomonas syringae pv. maculicola (Psm; dark colors) or 48 h after a mock-control infiltration (mock; light colors). Samples were derivatised with trimethylsilyl-diazomethane to convert carboxylic acid groups occurring in analytes into methyl ester groups prior to GC-based analysis. Retention times and m/z values of ion chromatograms for analytes (1a to 10a) accumulating in Col-0 but not in sid2 upon pathogen inoculation are given. B–H, identification of six meta-substituted SA and BA derivatives. Columns from left to right: First column: names, structural formulas, and assigned numerical identifiers (1–6) of analytes with compromised accumulation in sid2. Second column: structural formulas of their derivatised forms (1a to 6a), as detected by GC-MS and GC-FTIR. Methyl groups depicted in grey are introduced by sample derivatisation. Dashed lines in formulas indicate the occurrence of hydrogen bonds in SA derivatives. Third column: Mass spectra of the substances recorded in the GC-MS analysis. Fragmentation patterns and proposed structures of ions are indicated in brown (M+: molecular ion). Forth column: Infrared (IR) spectra of the substances analysed by GC-FTIR. The wavenumbers ṽ (in cm−1) of key IR vibrations and their assignments to functional groups are highlighted. The color assignments correspond to the color shadings of specific functional groups depicted in the molecular structures (second column) to highlight their IR vibrational characteristics [O-H: O-H stretching vibration (blue); C=O: C=O stretching vibration of aromatic methyl ester (O-C=O) (dark red), aliphatic methyl ester (light red), or aromatic aldehyde (H-C=O) (orange); C-H: aromatic C-H-stretching (black), aliphatic C-H-stretching (grey), or C-H stretching of aldehyde group (H-C=O, green), The subscript “hb” indicates the involvement of OH- or C=O-groups in hydrogen bonds]. B, salicylic acid (SA). C, 5-formyl-salicylic acid (5-FSA, 1). D, 5-carboxy-salicylic acid (5-CSA, 2). E, 5-carboxymethyl-salicylic acid (5-CMSA, 3). F, 3-formyl-benzoic acid (3-FBA, 4). G, 3-carboxy-benzoic acid (3-CBA, 5). H, 3-carboxymethyl-benzoic acid (3-CMBA, 6). The identities of analytes (1, 2, 4, 5, and 6) were further confirmed by use of commercially available authentic substances (Fig. S1). Moreover, 1 to 6 were identified in plant extracts using LC-qTOF/MS-based analysis (Fig. S2).

To elucidate the chemical structures of these analytes, we recorded their mass spectra and interpreted molecular ions (M+) and mass spectral fragmentation patterns. In addition, we analyzed the derivatized plant extract samples by GC-Fourier transform infrared spectroscopy (FTIR) and managed to obtain infrared (IR) spectra for most of the analytes (Figs. 1 and 2). The mass spectrum of the well-known ICS1-dependent metabolite SA (which is derivatized to SA methyl ester by the applied method) shows an M+ of 152 and characteristic fragment ions of m/z 120 and m/z 92, carrying predicted lactone- and epoxy-groups at the aromatic ring whose formation is due to subsequent losses of CH3OH and CO, respectively. These fragmentations are favored by the ortho-position of the phenolic OH relative to the methylated carboxy group (Fig. 1B). Moreover, the IR spectrum of methylated SA only shows a weak carbonyl (C=O) band at ṽ ∼1750 cm−1 which is characteristic of benzoic acid methyl esters (44). Instead, the main carbonyl absorption is shifted to ṽ = 1697 cm−1 due to the hydrogen bonding of the phenolic OH group to the C=O functionality of the ortho-positioned methyl ester group (Fig. 1B). This hydrogen bond also lowers the frequency of the phenolic O-H-bond from ∼3490 cm−1 to ∼3260 cm−1. Thus, both mass spectrometric and IR spectroscopic features reveal the ortho-positioning between the carboxy- and the phenolic OH-group in the benzol ring of the SA molecule.Figure 2 Identification of ICS1- and PBS3-dependent SA-malate, SA-aspartate, and SA biosynthesis-related compounds.A–G, chemical characterization of plant-derived compounds 7 to 10 by GC-MS- and/or GC-FTIR-based analyses. A–D, identification of salicyloyl-malate (SA-Mal, 7) as a natural product of Arabidopsis accumulating in Psm-inoculated leaves by GC-MS. A, mass spectra of the plant-derived compound 7a (blue) and of derivatized, synthetic SA-Mal (red) are identical. Fragmentation patterns of derivatized SA-Mal and proposed fragment structures are indicated in brown (M+: molecular ion). Methyl groups appearing in the structural formula in grey are introduced by derivatization. B, retention times of plant-derived, derivatized 7a, and synthetic SA-Mal are identical. Top: Overlaid ion chromatograms (m/z = 120) of GC-MS-analysed samples of extracts from Psm-inoculated Col-0 leaves (blue) and of authentic, derivatized SA-Mal (red). Bottom: Co-injection of plant extract sample and authentic SA-Mal results in a single peak in the ion chromatogram (purple). C and D, Psm-inoculated Arabidopsis leaves synthesize D4-labeled SA-Mal from exogenously supplemented deuterated SA (Fig. S4). C, the mass spectrum of derivatized D4-SA-Mal (D4-7a) generated in leaves after co-infiltration of Psm and 0.5 mM D4-SA shows a characteristic shift of 4 mass units compared with the mass spectrum of unlabeled SA-Mal (A). Proposed fragmentations are indicated in brown. D, ion chromatograms indicating the presence of both unlabeled SA-Mal (7a) (green) and D4-SA-Mal (blue) in extracts of leaves co-infiltrated with 0.5 mM D4-SA and Psm. E, GC-MS-based analysis indicates that compound 8a represents salicyloyl-aspartate (SA-Asp, 8). The mass spectrum of compound 8a (blue) displays an M+ ion at m/z 281, as expected for derivatised (methylated) SA-Asp. Further, the occurrence of α-cleavage-derived main fragments of m/z 121, 160, 222, and 250 confirm the structure of 8a as derivatized SA-Asp. Co-infiltration of D4-SA and Psm into leaves generates D4-labeled SA-Asp (green ion chromatogram and mass spectrum) in addition to unlabeled SA-Asp (blue), indicating that, similar to SA-Mal (C and D), SA-Asp is directly synthesized from SA in plants. The fragmentation pattern of the deuterium-labeled compound (D4-8a) is again consistent with the assigned D4-SA-Asp structure. SA-Mal (7) and SA-Asp (8) were also identified and analyzed in plant extracts via LC-qTOF/MS analysis (Fig. S2). F and G, spectroscopic characterization strongly suggests that the analytes 9a and 10a (Fig. 1A) correspond to derivatized N-pyruvoyl-glutamate and isochorismoyl-glutamate, respectively. F, based on mass spectrometric and IR spectroscopic characterization, analyte 9a was identified as derivatized (dimethylated) N-pyruvoyl-glutamate (Pyr-Glu, 9). Left: The mass spectrum of 9a contains a prominent m/z 174 ion that can result from cleavage of the amide-bond of Pyr-Glu, which decomposes to the main fragment at m/z 142 by loss of CH3OH and possible cyclization into the indicated pyrrolinone. Other plausible fragmentations, including the loss of acetyl- and methoxyl-moieties to yield m/z 202 (M+-43) and 214 (M+-31), respectively, are indicated. Right: The IR spectrum of the proposed dimethylated Pyr-Glu (9a) (blue) confirms the presence of the amide bond [ṽ ∼1710 (C=O), ṽ ∼1500 (C-N) and ṽ ∼3400 N-H; orange], carbonyl groups within methylester and keto moieties [ṽ ∼1760 (C=O); red], and aliphatic methyl (ene) units [ṽ ∼2960 and 2850 (C-H)] in the molecule, and the high similarity of the recorded IR spectrum with the best database hit, the IR spectrum of N-acetyl-glutamate dimethylester (black), further substantiated the notion that 9a represents dimethylated Pyr-Glu. G, the mass spectrum of analyte 10a contained the fragment series m/z 202, 174, and 142 observed for Pyr-Glu, and a plausible M+-H2O ion is consistent with a putative derivatized (trimethylated) isochorismoyl-9-glutamate (IC-Glu).

We noted that the IR spectrum of analyte 1a is remarkably similar to the spectrum of methylated SA, with the frequencies of C=O and O-H-bonds at ṽ ∼1700 cm−1 and ∼3210 cm−1 suggesting a methylated 2-hydroxybenzoate (salicylate) structure (Fig. 1C). In addition, the IR spectrum of 1a reveals the presence of a formyl group in the molecule, as indicated by the occurrence of a characteristic pair of aldehydic C-H vibrations at ∼2810 cm−1 and ∼2725 cm−1 and a C=O vibration at ∼1730 cm−1. The latter overlaps with the methyl ester C=O absorbance at ∼1700 cm−1 but is discernible as a shoulder in the spectrum. Moreover, the mass spectrum of 1a contains an M+ ion at m/z 180, a main fragment ion at m/z 148, and an m/z 120 fragment, which indicates subsequent losses of CH3OH and CO from the M+ ion. An additional strong m/z 119 ion further indicates a loss of a formyl group from the main m/z 148 ion. Together, these spectroscopic data strongly suggested that 1a represents the methylated version of a formyl-salicylic acid isomer (Fig. 1C). A comparison of the GC-MS characteristic of the extract peak and authentic 5-formyl-salicylic acid finally established that the pathogen-inducible and ICS1-dependent compound 1 represents 5-formyl-SA (5-FSA) (Figs. 1C, and S1, A and F).

The IR spectra of analytes 2a and 3a also show the absorptions at ∼1700 cm−1 and ∼3210 cm−1 typical for hydrogen-bonded C=O methyl ester and OH functionalities, again suggesting methylated 2-hydroxybenzoate structures for both analytes (Fig. 1, D and E). The IR spectrum of 2a additionally contains a prominent carbonyl band at ∼ 1750 cm−1, indicating the presence of a second methyl ester group at the benzoate ring that is not involved in H-bonding. In the mass spectrum of 2a, an M+ at m/z 210 and fragment ions at m/z 178, 147, and 119 indicate subsequent losses of CH3OH, a methoxy (CH3O) moiety, and CO (Fig. 1D). Together, these MS and IR features were consistent with a carboxylated salicylic acid structure for 2a in its derivatized, bis-methylated form. Comparison with an authentic 5-carboxy-salicylic acid standard confirmed this assumption and showed that compound 2 represents 5-carboxy-SA (5-CSA) (Figs. 1D, and S1, B and G). In addition to the C=O vibration at ∼1700 cm−1, the IR spectrum of analyte 3a shows a second carbonyl absorption at 1760 cm−1 that is characteristic for C=O vibrations within methyl ester moieties attached to an aliphatic carbon (Fig. 1E; (44)). Moreover, the mass spectrum of 3a contains an M+ at m/z 224 which produces a fragment ion at m/z 192 by loss of MeOH. Both the M+ and m/z 192 can lose a fragment of m/z 59 which corresponds to a carboxymethyl (CO2CH3) group that can be released by α-cleavage from a methylated phenylacetic acid unit (Fig. 1E). Together, these spectroscopic data are consistent with a bis-methylated carboxymethyl-SA structure for 3a, and in analogy to the two other identified 5-substituted SA derivatives, we propose that compound 3 represents 5-carboxymethyl-SA (Fig. 1E).

The IR spectra of analytes 4a to 6a are all characterized by the absence of a phenolic OH-vibration but the presence of a methylated benzoate group, as indicated by the C=O vibration at 1747 cm−1 (Fig. 1, F–H). Similar to 1a, analyte 4a shows aldehydic C-H absorptions at ∼2810 cm−1 and ∼2725 cm−1 and a C=O vibration at 1725 cm−1, consistent with the presence of an additional formyl group in the molecule. The mass spectrum of 4a is characterised by an M+ of m/z 164 which is consistent with a formylated methylbenzoate structure. A main fragment ion at m/z 133 indicates cleavage of a methoxy group from the M+ ion. This dominant “CH3O” cleavage is characteristic of methylbenzoates and is also present in the mass spectra of 5a and 6a (Fig. 1, F–H), while the SA derivatives 1a to 3a preferentially show a loss of CH3OH due to their 2-hydroxy-methylbenzoate structure (Fig. 1, B–D). The loss of two CO units from the m/z 133 ion in the mass spectrum of analyte 4a corroborates its presumed formyl-benzoate structure. A comparison with the authentic substance confirmed that compound 4 is 3-formyl-benzoic acid (3-FBA) (Figs. 1E, and S1, C and H). Further, analyte 5a was identified as derivatized (i.e., bis-methylated) 3-carboxybenzoic acid (3-CBA). Its IR spectrum shows a single, dominant C=O absorption at 1747 cm−1 caused by two chemically equivalent methylbenzoate carbonyls present in bis-methylated phthalate isomers. The M+ of m/z 194 and fragments corresponding to losses of a methoxy group (m/z 163) and subsequently CO (m/z 135) corroborated this phthalate structure. A comparison with the authentic substance then unequivocally identified compound 5 as 3-CBA (isophthalate) (Figs. 1G, and S1, D and I). Finally, the IR spectrum of analyte 6a shows a shoulder of the dominant methyl benzoate-associated C=O absorption (1747 cm−1) at ∼1760 cm−1, indicating, as for 3a, the presence of a methylated carboxyl group attached to an aliphatic moiety. The mass spectral characteristics (M+ of m/z 208, ion fragments of m/z 177 and m/z 149 indicating subsequent CH3O and CO fragment losses) thus strongly suggested that 6a is the derivatized (bis-methylated) form of 3-carboxymethyl benzoic acid (3-CMBA), in analogy with the accumulation of the proposed SA derivative 5-CMSA (3, Fig. 1E). Comparison with the available authentic compound confirmed that compound 6 represents 3-CMBA (Figs. 1H, and S1, E and J).

In summary, our analyses show that Arabidopsis leaves accumulate three 5-substituted SA derivatives carrying formyl, carboxy, and methylcarboxy units (5-FSA, 5-CSA, 5-CMSA), and the three analogues benzoic acid (BA) derivatives (3-FBA, 3-CBA, and 3-CMBA) upon P. syringae inoculation. Since the additional functionalities are located at the meta-position related to the carboxy group of SA or BA, we designate them as meta-substituted SA- and BA-derivatives. We were also able to detect the pathogen-induced accumulation of the six meta-substituted SA/BA-derivatives 1 to 6 by an inherently different analytical method that used liquid chromatography (LC) coupled with quadrupole time-of-flight mass spectrometry (qTOF-MS)-based high-resolution mass spectral analysis of underivatised Arabidopsis extracts (Fig. S2).

Identification of the four remaining ICS1-dependent metabolites as SA-malate, SA-aspartate, N-pyruvoyl-glutamate, and isochorismoyl-glutamate

In the above-described GC-MS analysis, four other Psm-inducible and ICS1-dependent metabolites remained to be characterized (Fig. 1A). The mass spectrum of analyte 7a shows, just like SA methyl ester (Fig. 1B), the m/z 120 and m/z 92 fragment patterning typical for ester-linked SA derivatives (Fig. 2A). The readily detectable M+ ion showed a mass of m/z 282, and an ion at m/z 251 possibly representing an “M+-OCH3”-fragment was also discernible. On this basis, we hypothesized that analyte 7a represents the malate ester of SA (salicyloyl-malate; SA-Mal) in its derivatized, bis-methylated form. We next synthesized SA-Mal chemically (Fig. S3), and recorded the mass spectrum of the derivatized authentic substance. Its GC retention time and mass spectra were identical to those of analyte 7a, confirming that the plant-derived substance 7 is SA-Mal (Fig. 2, A and B). Further, exogenous application of a deuterated SA-form, D4-SA (in which the four hydrogens at the benzol ring are replaced by deuterium; Fig. S4), to Psm-inoculated leaves resulted in the formation of a product (D4-7a) that showed a mass spectrum consistent with a D4-SA-Mal structure, indicating in planta conversion of the applied D4-SA into D4-SA-Mal (Fig. 2, C and D). Therefore, Arabidopsis leaves synthesize the hitherto undescribed plant metabolite SA-Mal from SA in response to pathogen attack.

The mass spectrum of the derivatized compound 8a was identical to a published spectrum of bis-methylated salicyloyl-aspartate (SA-Asp) (Fig. 2E; (37)), indicating that analyte 8 represents SA-Asp. The M+ ion (m/z 281) of bis-methylated SA-Asp was readily detectable in the mass spectrum, and cleavage of the amide-bond resulted in a main salicyloyl-fragment at m/z 121 and a discernible m/z 160 ion representing the bis-methylated aspartyl-moiety. Supplementing Psm-inoculated leaves with exogenous D4-SA resulted in the generation of the corresponding D4-SA-Asp (D4-8a; Fig. 2E). The M+ of this compound fragmented into a main D4-salicyloyl-ion at m/z 125 and the remaining unlabeled aspartyl-associated fragment at m/z 160. In addition, plausible fragments derived from the aromatic molecule part by losses of methoxy or “CH2CO2” differed by 4 mass units in the unlabeled compound 8a and D4-labeled D4-8a, respectively (Fig. 2E). Together, this corroborates a P. syringae-inducible generation of SA-Asp from SA in Arabidopsis leaves. Notably, both SA-Mal and SA-Asp were also detected and characterised by LC-qTOF-MS-based, high-resolution mass spectral analysis of underivatised extracts (Fig. S2).

The IR spectrum of analyte 9a closely resembled the spectrum of a substance present in our IR spectral database, N-acetyl-glutamate dimethyl ester (Fig. 2F). This relates to the presence of an amide bond (N-H vibrations at ∼3400 cm−1, C=O absorption at ∼1710 cm−1, and N-C-vibrations at ∼1500 cm−1), and characteristic absorptions of a bis-methylated glutamate part (e.g., strong absorptions at ∼1760 cm−1 of the 2 C=O vibrations of the methylester moieties). The IR spectrum of 9a therefore strongly suggested that a small organic acid similar to acetic acid is bound to glutamate via an amide linkage in compound 9. The mass spectrum of 9a contained a prominent ion at m/z 174 that was consistent with a (bis-methylated) glutamyl-fragment resulting from amide-cleavage of the presumed conjugate. This fragment could lose MeOH to generate the main ion at m/z 142 with a proposed pyrrolinone ring structure, and a further loss of MeOH would then produce the m/z 110 ion (Fig. 2F). When assuming pyruvate as the amide-linked conjugation partner for glutamate, both the m/z 202 (M+ - acetyl) and the m/z 214 (M+ - methoxy) could be plausibly explained as mass fragments of the derivatized form 9a. A bis-methylated N-pyruvate-glutamate (Pyr-Glu) structure would also be consistent with the IR spectrum of 9a, since the C=O absorption at ∼1760 cm−1 is overall more prominent than the same absorption of the database spectrum for N-acetyl-glutamate dimethylester, reflecting the presence of an additional ketonic C=O bond in pyruvate compared to acetate (Fig. 2F). Therefore, we propose that compound 9 represents N-pyruvoyl-glutamate (Pyr-Glu). Notably, in the course of our study, Pyr-Glu has been identified as side product in the ICS1- and PBS3-dependent biosynthesis of SA (21, 22), confirming our analysis. The same studies also identified isochorismoyl-glutamate (IC-Glu) as an intermediate in the isochorismate-derived biosynthesis of SA. The mass spectrum of 10a is consistent with a derivatized IC-Glu structure (Fig. 2G). On one hand, a m/z 379 ion represents a plausible M+-H2O ion fragment whose formation is favored because water loss of the six-membered ring in the not discernible M+ ion of derivatized (trimethylated) IC-Glu (m/z 397) would lead to a relatively stable benzolic fragment. On the other hand, the occurrence of the fragment pattern m/z 202, 174, 142, and 110, and other striking similarities with the mass spectrum of 9a in the medium and lower mass regions indicate the presence of the enolpyruvoylglutamyl part of the presumed IC-Glu (Fig. 2G). Therefore, our spectroscopic data also suggest the pathogen-induced, ICS1-dependent formation of both Pyr-Glu and IC-Glu as by-products and intermediates, respectively, of isochorismate-derived SA biosynthesis.

SA functions as a precursor for SA-Mal and SA-Asp but not for the identified meta-substituted SA- and BA-derivatives; however, plants convert the respective formyl-into the carboxy-substituted compounds

Just as SA, the newly described meta-substituted SA- and BA-derivatives are generated in an ICS1-dependent manner (Fig. 1). To gain further insight into the formation of these compounds in planta, we tested whether SA would function as their biosynthetic precursor. To do so, we co-infiltrated Arabidopsis leaves with 0.5 mM D4-SA and Psm and examined the possible formation of labeled compounds by GC-MS analyses 48 h later (Fig. 3, A–D). In theory, the meta-substituted compounds should lose one deuterium at the aromatic ring during their formation from D4-SA to produce D3-labeled variants, while D4-labeled variants of SA-Mal and SA-Asp should be formed (Fig. 3A). When ion chromatograms corresponding to masses of M+ ions or of the main aromatic fragments at m/z 120 (SA-Mal) and m/z 121 (SA-Asp) were compared with the corresponding (m + 4)-chromatograms, the in planta formation of both D4-SA-Mal and D4-SA-Asp was evident in addition to their unlabeled counterparts (Figs. 2, C–E, and 3B), indicating that SA functions as a biosynthetic precursor for SA-Mal and SA-Asp. However, overlays of chromatograms corresponding to the masses of M+ ions and the respective (m + 3)-masses only showed accumulation of the natural compounds, while deuterium-labeled versions were not generated (Fig. 3, C and D). This shows that SA does not function as a precursor for the biosynthesis of the meta-substituted aromatics 1 to 6.Figure 3 In contrast to SA-Mal and SA-Asp, the six identified meta-substituted SA- and BA-derivatives are not biosynthesized from SA; however, the formyl-derivatives are oxidized to the respective carboxy compounds in planta.A–D, while exogenously supplemented SA is directly converted to SA-Mal (and SA-Asp, Fig. 2E) in leaves, it is not metabolized to meta-substituted SA and BA derivatives. A, theoretical deuteration patterns of SA-Mal, meta-substituted SAs and meta-substituted BAs when the metabolic conversion of D4-SA into these substances is assumed. This corresponds to shifts of 4 (D4-SA-Mal) and 3 (D3-labeled meta-substituted compounds) mass units compared to the unlabeled substances, respectively. B, exogenously supplemented D4-SA is metabolized to D4-SA-Mal. The depicted ion chromatograms correspond to m/z values of the M+ ions of SA-Mal (blue) and D4-SA-Mal (red) (Fig. 2, A and C). Chromatograms from extracts from leaves co-infiltrated with 0.5 mM SA and Psm (left) or 0.5 mM D4-SA and Psm (right) are shown. Leaves were harvested 48 h after infiltration. B, exogenously supplemented D4-SA is not converted to deuterium-labeled meta-substituted SA (C) or BA (D) derivatives in planta. The depicted ion chromatograms correspond to m/z values of the M+ ions (Fig. 1) of the meta-substituted compounds (blue) and of assumed, D3-labeled compounds (orange). Chromatograms from extracts from leaves co-infiltrated with 0.5 mM SA and Psm (left panels) or 0.5 mM D4-SA and Psm (right panels) are shown. Leaves were harvested 48 h after infiltration. E, feeding of Arabidopsis leaves with 3-formyl-BA and 5-formyl-SA provokes in planta formation of 3-carboxy-BA and 5-carboxy-SA, respectively, while exogenously supplemented SA is converted to SA-Mal and SA-Asp. Solutions of 0.5 mM of SA, 3-FBA, 5-FSA, or 3-CMBA or water as a control were infiltrated into three leaves of individual plants. The treated leaves were harvested 4 h later, and the levels of the three meta-substituted SA-derivatives (left column), the three meta-substituted BA-derivatives (central column), and SA, SA-Mal, and SA-Asp (right column) were analyzed by GC-MS. Bars represent the mean ± SD of three biological replicate samples, each consisting of six leaves from two plants. Individual data points of biological replicates are super-imposed on the bar graphs (small circles). The presence of asterisks above bars indicates significant differences between substance- and water-control treatment (∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05; two-tailed t test).

To further examine the biosynthesis of the meta-substituted SA- and BA-derivatives, we infiltrated the leaves of plants with 0.5 mM solutions of SA, 3-FBA, 5-FSA, or 3-CMBA and assayed metabolite contents 4 h later (Fig. 3E). As expected, exogenous application of each of the compounds resulted in increases in their contents in the treated leaves. Remarkably, the application of 3-FBA and 5-FSA resulted in a strong accumulation of 3-CBA and 5-CSA, respectively, indicating efficient conversion of the formyl-substituted BA- and SA-derivatives into their carboxy-substituted counterparts in planta. This indicates the presence of oxidoreductase-like enzyme activities in the leaves that catalyze the oxidation of the BA- and SA-aldehydes into the respective carboxylic acids (Fig. 3E). By contrast, 3-CMBA was not converted to any of the other meta-substituted compounds. However, exogenous SA increased the levels of both SA-Mal and SA-Asp, corroborating that these conjugates are synthesized from SA. Interestingly, the application of 3-FBA, 5-FSA, and 3-CMBA also modestly increased the leaf levels of SA, which might indicate that treatments with these compounds stimulate SA biosynthesis to a small extent (Fig. 3E).

While isochorismate-derived metabolism is negatively regulated by NPR1, SA conjugates, and meta-substituted SA/BA-derivatives are oppositely affected by PBS3

To investigate the regulatory principles by which the ten identified ICS1-dependent metabolites accumulate (Figs. 1 and 2), we compared their Psm-induced generation at 48 h after inoculation in the leaves of the Col-0 wildtype, the three SA biosynthetic mutants sid2, eds5, and pbs3 (18, 19, 23, 45), the NahG transgenic line that rapidly degrades SA due to expression of an SA hydroxylase (46), and in the npr1 mutant defective in SA perception and downstream responses (47) (Fig. 4). Remarkably, all six meta-substituted SA/BA-derivatives showed a compromised Psm-triggered accumulation in sid2 and eds5, while they were strongly over-produced upon pathogen inoculation in pbs3 and npr1 (Fig. 4, A and B). This indicates that ICS1 and the putative isochorismate-transporter EDS5 are required for their biosynthesis, while PBS3 and NPR1 strongly dampen their accumulation. Further, NahG plants that promptly degrade biosynthesized SA to catechol (48) show a significant accumulation of meta-substituted SA/BA-derivatives, corroborating our conclusion that these compounds are not derived from SA (Fig. 4, A and B). Compared to the wildtype, NahG showed enhanced accumulation of the formyl- (5-FSA, 3-FBA) and carboxymethyl- (5-CMSA, 3-CMBA) compounds, and attenuated accumulation of the meta-carboxylated substances (5-CSA, 3-CBA).Figure 4 meta-Substituted SA/BA-derivatives and SA-conjugates show similar accumulation characteristics in the SA pathway mutants sid2, eds5, and npr1 but are oppositely regulated by PBS3.A and B, meta-Substituted SA- (A) and BA-derivatives (B) show compromised pathogen-induced accumulation in sid2 and eds5 mutants but not in SA degrading NahG plants, and they over-accumulate in both pathogen-inoculated pbs3 and npr1. A, levels of 5-FSA, 5-CSA, and 5-CMSA in the leaves of Col-0, sid2, eds5, NahG, pbs3, and npr1 plants 48 h after inoculation with Psm (dark blue) or mock-control infiltration (light blue). Metabolite levels are given in ng per gram fresh weight (FW). B, levels of 3-FBA, 3-CBA, and 3-CMBA (details as in A). C, SA-Mal and SA-Asp exhibit diminished accumulation in all the SA biosynthetic mutants and in SA catabolizing NahG plants, and they over-accumulate in npr1. Metabolite levels (ng per gram FW) in Psm- (dark red) and mock-inoculated (light red) leaves at 48 h post treatment (hpt) are given. D, Pyr-Glu and IC-Glu are biosynthesised in dependence of ICS1/SID2, EDS5, and PBS3, accumulate in NahG and over-accumulate in npr1. Metabolite levels (ng per gram FW) in Psm- (dark red) and mock-inoculated (light red) leaves at 48 hpt are given. For (A–D), the bars represent the mean ± SD of four biological replicates from different plants, each replicate consisting of six leaves from two plants (nd: not detected). Individual data points of biological replicates are super-imposed on the bar graphs (small circles). Different letters above the bars indicate statistically significant differences (p < 0.05, Kruskal-Wallis H test). Please note that the data for the Col-0 wildtype are depicted twice in the figures with different scaling of the y-axes to better illustrate comparisons with moderately accumulating (left panels) and over-accumulating (right panels) lines.

By comparison, the Psm-triggered accumulation of the SA conjugates SA-Mal and SA-Asp was compromised in all of the SA biosynthetic mutants as well as in the NahG line, confirming their derivation from SA and illustrating that the whole SA biosynthetic pathway is required for their biosynthesis (Fig. 4C). Moreover, Pyr-Glu and IC-Glu, the by-product and intermediate of SA biosynthesis, respectively, exhibited compromised accumulation in the SA biosynthetic mutants sid2, eds5 and pbs3, but were generated in a wildtype-like manner in the NahG transgenics (Fig. 4D).

Meta-substituted SA/BA-derivatives show a more moderate accumulation pattern than other ICS1-derived metabolites in P. syringae-inoculated leaves

To compare the kinetics and degree of pathogen-induced accumulation of the different ICS1-dependent substances in Arabidopsis, we Psm-inoculated or mock-treated leaves of Col-0 plants, harvested the treated leaves at 6, 10, 24, and 48 h after the treatments, and assessed their metabolite contents via GC-MS-based analyses (Fig. 5). Remarkably, SA showed the fastest Psm-induced increase of all of the examined metabolites (Fig. 5A). It significantly accumulated already at 6 h post-inoculation (hpi) with Psm, and its levels gradually increased to a maximum at 24 hpi and then decreased again at 48 hpi. By contrast, the SA glucose conjugates SAG and SGE accumulated later than 6 hpi, with a first significant pathogen-induced increase at 10 hpi. The accumulation pattern of the glucose ester SGE thereby paralleled the pattern for free SA, with a distinct maximum accumulation at 24 hpi. By comparison, the glucoside SAG showed a steady and strong increase until 48 hpi (Fig. 5A). Quantitatively, SAG showed the most prominent increase after inoculation (>50 μg g−1 FW at 48 hpi), followed by SGE (>10 μg g−1 FW at 24 hpi) and SA (∼2.5 μg g−1 FW at 24 hpi), which is consistent with the notion that SAG represents the most prominent SA derivative in Arabidopsis (9).Figure 5 Meta-substituted SA/BA-derivatives accumulate later and to more moderate levels in P. syringae-inoculated leaves than other detected ICS1-dependent metabolites.A–C, levels of different ICS1-dependent metabolites at 6, 10, 24, and 48 h post-treatment in mock-infiltrated (mock) and Psm-inoculated (Psm) leaves of Arabidopsis Col-0 plants. Three rosette leaves per plant were treated. Metabolite contents were determined by GC-MS analysis and are given in μg per gram FW. Bars represent means ± SD of three to four biological replicates. One replicate sample consisted of six leaves from two plants. Individual data points of biological replicates are super-imposed on the bar graphs (small circles). The presence of asterisks above the lines indicates whether significant differences between mock- and Psm-treatment exist for the particular timepoint (∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05; ns: no significant differences; two-tailed t test). A, SA, SA-ß-glucoside (SAG), and SA glucose ester (SGE). B, meta-substituted SA/BA-derivatives: 5-FSA, 5-CSA, 5-CMSA, 3-FBA, 3-CBA, and 3-CMBA. C, SA-Mal, SA-Asp, Pyr-Glu, and IC-Glu.

The temporal accumulation patterns of Pyr-Glu and IC-Glu were similar to those of SAG, but their quantitative levels were much more moderate (by far lower than 1 μg g−1 FW; Fig. 5C). Moreover, SA-Mal and SA-Asp did not accumulate before 24 hpi and showed the most prominent pathogen-induced increases at 48 hpi, indicating that both SA conjugates start to accumulate later than the glucose derivatives SAG and SGE (Fig. 5C). At 48 hpi, the absolute levels of SA-Mal and SA-Asp were on a similar scale than those of free SA (∼1–2 μg g−1 FW). Finally, the six meta-substituted substances 5-FSA, 5-CSA, 5-CMSA, 3-FBA, 3-CBA, and 3-CMBA showed a similar pattern of accumulation, with increased levels of these compounds detected not earlier than 48 hpi (Figs. 5B, and S5A). Moreover, the absolute levels of each of the six metabolites were consistently lower than 0.5 μg g−1 FW. Together, this indicates that the accumulation of the six meta-substituted SA/BA-derivatives in the leaves starts at relatively late times after P. syringae-inoculation and results in comparatively moderate quantitative increases of these substances.

In addition to Col-0 leaves inoculated by the compatible, virulent Psm strain, the six meta-substituted SA/BAs and the other investigated four SA-related metabolites also accumulated in leaves inoculated with an avirulent, hypersensitive-response-inducing Psm strain expressing the AvrRpm1 avirulence gene (Psm avrRpm1; (34)) (Fig. S5). Notably, five of the six meta-substituted SA/BAs started to accumulate earlier in the incompatible Psm avrRpm1-Col-0 interaction than in the compatible Psm-Col-0 interaction (Fig. S5A).

Administration assays suggest that meta-substituted SA/BA-derivatives and SA-Mal have modest activity in the induction of acquired resistance and PR1 gene expression

We next asked whether the in-planta-detected meta-substituted SA/BA-derivatives and SA-Mal would exert an impact on plant immunity and defense-related gene expression. It is well established that plants exogenously supplemented with SA acquire resistance to a series of compatible pathogens and increase expression of the SA marker gene PATHOGENESIS-RELATED1 (PR1) (46, 49, 50). Consistently, infiltration of a 0.5 mM solution of SA into the leaves of Arabidopsis Col-0 plants robustly increased the resistance of these leaves to subsequent infection by Psm in different experiments, and bacterial growth after inoculation was reduced by about one order of magnitude (Fig. 6, A, B, and E). Moreover, the ICS1-defective sid2 mutant allowed a significantly higher bacterial growth in control plants than Col-0, indicating that the inability to induce SA biosynthesis in the mutant compromises plant basal resistance (Fig. 6, A, B, and E; (18, 19)). The treatment of sid2 with exogenous SA resulted in a strong acquired resistance response, by which bacterial growth in the course of pathogen challenge was attenuated by nearly two orders of magnitude (Fig. 6, A, B, and E). As a consequence, SA-treated sid2 acquired similar (Fig. 6, A and B) or only modestly lower (Fig. 6E) levels of resistance than SA-induced Col-0 plants.Figure 6 Exogenously applied, meta-substituted SA/BA-derivatives and SA-malate have moderate impact on plant immunity and pathogenesis-related gene expression.A and B, assessment of resistance induction by exogenous application of SA and different SA- and BA-derivatives. A, application of SA, 3-FBA, 3-CBA, 3-CMBA, 5-FSA, or 5-CSA. B, application of SA, 2,3-dihydroxybenzoic acid (2,3-DHBA), or 2,5-dihydroxybenzoic acid (2,5-DHBA). A and B, three leaves of individual Arabidopsis wild-type (Col-0) or sid2 mutant plants were infiltrated with a 0.5 mM substance solution. Control plants (mock) were infiltrated with water containing 0.1% ethanol, the medium in which the applied substances were dissolved. The three pre-treated leaves were challenge-inoculated 4 h later with a suspension of bioluminescent P. syringae pv. maculicola ES4326 (Psm) strain expressing the luxCDABE operon from Photorhabdus luminescens (Psm lux; (92)) by syringe-infiltration (OD600 = 0.001). As a measure of plant susceptibility, the numbers of bacteria were determined at 60 h post-inoculation (hpi) in inoculated leaves by luminescence quantification and expressed as relative light units (rlu) per cm2 leaf area (93). The data values were log10-transformed for graphical display. Bars indicate the mean ± SD of at least 15 leaf replicates. Different letters denote significant differences (p < 0.05) according to ANOVA and post hoc Tukey HSD test. Moreover, the results of two-tailed t-tests between the mock- and the treatment-samples of one genotype are indicated by the presence/absence of asterisks above the bars of the treatment samples (∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05; absence of asterisks: no significant differences). C and D, assessment of inducible expression of the defence gene PATHOGENESIS-RELATED1 (PR1) by exogenous SA and different SA- and BA-derivatives. C, application of SA, 3-FSA, 3-BSA, 3-CMBA, 5-FSA, or 5-CSA. D, application of SA, 2,3-DHBA, or 2,5-DHBA. C and D, three leaves of individual Col-0 or sid2 plants were infiltrated with a 0.5 mM substance solution or with water containing 0.1% EtOH as a control treatment (mock). The treated leaves were harvested 4 h later, and transcript levels of PR1 were determined by RT-qPCR using POLYPYRIMIDINE TRACT-BINDING PROTEIN 1 (PTB1) as the reference gene (Fig. S7). Bars represent means ± SD of gene transcript levels calculated from the values of the three biological replicates, which themselves represented the mean of two technical replicates. The data values were log2-transformed for graphical display. The presence of asterisks above the bars indicates whether significant differences between mock- and substance-treatments exist for a particular genotype (∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05; absence of asterisks: no significant differences; two-tailed t test). E, assessment of resistance induction of SA-malate and related compounds. Three leaves of individual Col-0 or sid2 mutant plants were infiltrated with 0.5 mM solutions of SA, SA-Mal, malate, or fumarate. A mock-control infiltration was performed with water containing 0.1% EtOH. The three pre-treated leaves were challenge-inoculated 4 h later with Psm lux, and bacterial numbers were assessed after 60 h as described above. The values were log10-transformed for graphical display. Bars indicate the mean ± SD of at least 24 leaf replicates. Statistical analyses as described for (A and B). F, storage of SA-malate solutions over longer time periods resulted in degradation to SA, malate, and fumarate. A solution of 100 ng μl−1 SA-Mal in MeOH/H2O (80:20) was freshly prepared, and 10 μl thereof was analyzed by GC-MS the same day (left) or 3 months later (right). The assessed relative contents of SA-Mal, SA, malate, and fumarate (related to the amount of employed SA-Mal) are given. G, SA-Malate exogenously supplemented to leaves is hydrolyzed to SA. A solution of 0.5 mM SA-Mal or water containing 0.1% EtOH (control) was infiltrated into three leaves of individual Col-0 plants, and the treated leaves were harvested 48 h later. The contents of SA-Mal and SA in leaves were assessed and expressed in nmol g−1 FW. A–G, individual data points of biological (F, technical) replicates are super-imposed on the bar graphs (small circles).

When plants were treated with 0.5 mM solutions of the five meta-substituted SA/BA derivatives for which authentic substances were available, we observed much weaker immune responses than for SA administration (Fig. 6A). Col-0 plants lacked resistance induction towards Psm after treatment with 3-FBA, 3-CBA, 3-CMBA, and 5-FSA, and only 5-CSA induced a modest elevation of pathogen resistance in Col-0. Exogenous 5-CSA also showed a modest positive effect on the resistance of sid2, and the same was true for 3-CBA and 5-FSA applications (Fig. 6A). Comparatively, we also examined the ability of resistance induction for two other known SA derivatives, 2,3-dihydroxybenzoic acid (2,3-DHBA), and 2,5-DHBA, which occur in Arabidopsis leaves primarily as glycosylated variants (30, 32). Similar to most of the here-identified meta-substituted SA/BA-derivatives, a 0.5 mM solution of 2,5-DHBA showed no effect on plant resistance to Psm infection (Fig. 6B). However, exogenously applied 2,3-DHBA increased immunity to a moderate extent in both Col-0 and sid2, but again, this response was significantly weaker than the immune responses towards the same dose of SA (Fig. 6B).

Next, we assessed the transcript levels of PR1 in the leaves of mock-treated control plants and of plants treated with SA and the above-described SA-related substances (Fig. 6, C and D). Remarkably, the basal transcript levels of PR1 in the control plants were significantly lower in sid2 than in Col-0 (Fig. 6, C and D), which correlated with the basal endogenous contents of SA and the degree of basal resistance in these plants (Figs. 6, A, B, and E; Fig. 7A). As expected, exogenous SA triggered a strong elevation of the transcript levels of PR1 in the treated leaves of both Col-0 and sid2 (Fig. 6, C and D). By contrast, none of the five applied meta-substituted SA/BA-derivatives was able to increase PR1 transcript levels in the Col-0 wildtype. Upon 3-CMBA and 3-CBA application, the basal PR1 levels were even significantly decreased in Col-0. For 3-CMBA-treated plants, this negative effect on PR1 expression correlated with a (statistically non-significant) tendency of increased susceptibility towards Psm infection (Fig. 6A). Notably, 3-FBA, 5-FSA, and 5-CSA moderately increased the low basal levels of PR1 transcripts of sid2, while 3-CMBA and 3-CBA exerted no significant effects in the mutant (Fig. 6C). Together with the resistance data, this indicates that certain meta-substituted SA/BA-derivatives (in particular 5-CSA and 5-FSA) can moderately elevate immune responses in the SA-deficient sid2 plants that exhibit strongly reduced basal defenses. In this context, we found that application of 2,3-DHBA, but not 2,5-DHBA, markedly elevated PR1 transcript levels (Fig. 6D), which again paralleled the effects of the two DHBAs on acquired resistance (Fig. 6B).Figure 7 While an ICS1-dependent but PBS3-independent biosynthetic branch leads to the Psm-induced accumulation of free salicylic acid at later infection stages, N-hydroxypipecolic acid over-accumulates to similar levels in sid2, eds5, and pbs3.A–C, levels of different defense-related metabolites at 6, 10, 24, and 48 h post-treatment (hpt) in mock-infiltrated (mock) and Psm-inoculated (Psm) leaves of Arabidopsis Col-0, sid2, eds5, and pbs3 plants. Three rosette leaves per plant were treated. Metabolite contents were determined by GC-MS analysis of trimethylsilylated analytes and are given in μg per gram FW. Bars represent means ± SD of four biological replicates. One replicate sample consisted of six leaves from two plants. Individual data points of biological replicates are super-imposed on the bar graphs (small circles). Different letters denote significant differences (p < 0.05, Kruskal-Wallis H test). A, levels of SA, SAG, and SGE. B, levels of NHP and NHP-glucoside (NHPG). C, levels of camalexin. See also Fig. S6.

Finally, we infiltrated 0.5 mM solutions of chemically synthesized SA-Mal to Col-0 and sid2 plant and scored resistance to Psm challenge in the treated leaves (Fig. 6E). Significant induction of resistance was observed in both Col-0 and sid2 plants upon SA-Mal application, but this effect was modest compared to the effect of exogenous SA. Moreover, we found that in solution, SA-Mal is slowly decomposed to SA, malate, and fumarate. In a freshly prepared solution that underwent the analytical workup process and subsequent GC-MS analysis on the same day, about 97% of the initially dissolved SA-Mal, but also 3% of free SA and fumarate were detectable in the analysis (Fig. 6F). After 3 months of storage of the solution at 4 °C, about 90% of the SA-Mal was decomposed to free SA, and significant amounts of fumarate and malate were also detected in the solution (Fig. 6F). Therefore, we also tested whether 0.5 mM solutions of fumarate or malate would affect pathogen resistance, and found a small resistance increase for malate but not for fumarate application (Fig. 6E). Moreover, when infiltrating fresh SA-Mal solution into Arabidopsis Col-0 or sid2 leaves, about 50% of the SA-Mal was hydrolyzed to free SA in the leaves after 2 days (Fig. 6G). Considering this efficient in planta release of SA from SA-Mal, and the strong immune-stimulating activity of SA, we consider it as likely that the moderate resistance effects observed for exogenous SA-Mal application were mainly caused by the liberated SA.

Together, our administration assays suggest that the meta-substituted SA/BA-derivatives exert, when compared to SA, only modest immune-stimulatory activity in Arabidopsis, which was most evident in the SA-deficient sid2 mutant. Out of the five tested substances, 5-CSA and 5-FSA showed the most pronounced stimulatory effects. SA-Mal application also resulted in a significant resistance effect, but this is supposedly a result of a release of SA from SA-Mal in planta. In addition, our results indicate that 2,3-DHBA but not 2,5-DHBA has moderate immune-activating activity.

An ICS1-derived but PBS3-independent accumulation of unconjugated SA occurs in the later stages of the Arabidopsis-Pseudomonas interaction

According to a recently established model, the isochorismate pathway-derived, stress-inducible biosynthesis of SA in Arabidopsis proceeds via the ICS1-catalysed conversion of chorismate to isochorismate in the plastid, a translocation of isochorismate into the cytosol by the MATE transporter EDS5, and conjugation of isochorismate to Glu by cytosolic PBS3. The resulting intermediate IC-Glu would then be transformed to SA via the release of Pyr-Glu (21, 22). Knockout mutants of the isochorismate pathway genes ICS1, EDS5, and PBS3 should therefore be severely compromised in the biotic stress-inducible accumulation of SA and its conjugates. Contrasting this assumption, our metabolite analyses indicated a significant accumulation of unconjugated SA in the leaf samples of pbs3 knockout plants. At 48 h post-inoculation with Psm, pbs3 even over-accumulated SA, and this tendency was confirmed by two inherently different methodological procedures that used GC-MS-based and LC-qTOF-MS-based analyses, respectively (Figs. 7A, and S2B). This finding prompted us to include the main SA biosynthetic mutants sid2 (ics1), eds5, and pbs3 in our time-course analysis, in which metabolite accumulation was assessed in the leaves of Psm-inoculated and mock-treated plants at 6, 10, 24 and 48 hpi (Fig. 7).

This analysis showed that sid2 mutants failed to accumulate free SA in response to Psm over the whole time period between 6 and 48 hpi (Fig. 7A). In addition, the amounts of the two main conjugates, SAG and SGE, which were synthesized in the Col-0 wildtype to high levels from 10 hpi onwards, only faintly increased in sid2 at 24 and 48 hpi to levels that were well below the basal contents of the unstressed wildtype (Fig. 7A). For a quantitative comparison, we approximated the sum of SA, SAG, and SGE as the levels of “total SA”. The increases of total SA in sid2 following Psm inoculation only amounted to ∼0.5% of the corresponding increases in Col-0 at 48 hpi, the latest investigated timepoint for which the pathogen responses were highest. This indicates that virtually the whole metabolic flow towards biosynthesis of SA and its main derivatives under biotic stress proceeds via ICS1. The biotic-stress-induced biosynthesis of SA was also greatly reduced in eds5, for which marginal Psm-induced accumulation of SA, SAG, and SGE was detected at 24 and 48 hpi but not at earlier stages (Fig. 7A). For eds5, the increases of the total SA levels at 48 hpi amounted to ∼1.5% of the respective increases in the wildtype, indicating that ICS1-mediated SA biosynthesis is almost fully carried forward by EDS5. Although the Psm-inducible accumulation of free SA was heavily compromised in pbs3 during the early stages of infection (6 and 10 hpi), SA started to accumulate in this mutant at 10 hpi, reached wild-type-like accumulation levels at 24 hpi, and even over-accumulated SA at 48 hpi. However, as for sid2 and eds5, pbs3 only weakly accumulated the glucose conjugates SAG and SGE over the whole time period (Fig. 7A). In addition, as described above, the same stringent dependency on ICS1, EDS5, and PBS3 was observed for the two other SA conjugates, SA-Mal and SA-Asp (Fig. 4C). According to the above-applied sample calculation, pbs3 accumulated about 5.5% of the “total SA” accumulating in Col-0 at 48 hpi, and the main part of this overall moderate increase was due to the accumulation of free, unconjugated SA (Fig. 7A). In summary, this indicates that in addition to a major branch of the isochorismate pathway that includes ICS1, EDS5, and PBS3, a second metabolic branch exists that proceeds independent of PBS3, becomes relevant at later stages during the plant-pathogen encounter, and almost exclusively results in the production of unconjugated SA. It is noteworthy that the described tendencies were congruently detected by both the GC-MS- and LC-qTOF-MS-based analytic procedures (Figs. 7A, and S2B).

The sid2, eds5, and pbs3 mutants show similar immune-related and metabolic characteristics

We and others have previously described several immune-related characteristics of the ICS1-defective sid2 mutant that were explained by its SA-deficient metabolic phenotype. First, sid2 shows an attenuated accumulation of the SAR inducer N-hydroxypipecolic acid (NHP), its biosynthetic precursor pipecolic acid (Pip), and the phytoalexin camalexin in the early (10 hpi) interaction stages with Psm (12). Second, sid2 over-accumulates NHP, its glucose ester NHPGE, and camalexin at later infection stages (24 and 48 hpi) (10, 18, 26, 50). Third, sid2 shows strongly attenuated accumulation of the NHP N-O-glucoside NHPG (26, 50). And finally, sid2 shows a significantly attenuated but not fully compromised SAR response, a tendency that was observed for both the biological induction of SAR following a localized Psm inoculation (12, 49, 50), and for chemical induction of SAR by pre-treatment of plants with exogenous NHP (10, 11). We wondered whether the isochorismate pathway-related metabolic differences that we observed for sid2, eds5, and, particularly, pbs3 (Figs. 4 and 7A), would result in an alteration of these immune characteristics or whether the mutants would behave similarly. Therefore, the manifestations of these characteristics were directly compared in the three mutants. First, attenuated NHP, Pip, and camalexin accumulation at 10 hpi was observed for both sid2, eds5, and pbs3 (Figs. 7, B and C, and S6). Second, over-accumulation of NHP, NHPGE, and camalexin at later infection stages was observed in all of the mutants, with a somewhat smaller NHP over-accumulation in pbs3 than in sid2 or eds5 (Figs. 7, B and C, and S6). Third, a strongly attenuated accumulation of NHPG following Psm inoculation was detectable in each of the mutants (Fig. 7B). Further, naïve (previously unstressed) sid2, eds5, and pbs3 plants showed a very similar extent of compromised basal resistance toward bacterial challenge (Fig. 8A; water-treated plants). Moreover, treatments with exogenous NHP resulted in a similar degree of SAR towards Psm infection in the three mutant types, which was, however, significantly weaker than the SAR response in the wildtype (Fig. 8A). In addition, the biological induction of SAR in upper rosette leaves following an inducing inoculation of lower leaves with Psm was attenuated to similar degrees in the sid2, eds5, and pbs3 mutants (Fig. 8B).Figure 8 The establishment of systemic acquired resistance by exogenous NHP or as a consequence of an inducing pathogen inoculation is compromised to similar degrees in sid2, eds5, and pbs3.A, SAR induced by exogenous NHP treatment is attenuated in the SA biosynthetic mutants sid2, eds5, and pbs3. The soil substrate of individually cultivated Arabidopsis Col-0, sid2, eds5, or pbs3 plants was supplemented with 10 ml of a 1 mM NHP solution or 10 ml of water. One day later, three rosette leaves were inoculated with Psm lux. Bacterial numbers in the challenged leaves were assessed at 60 dpi (see Fig. 6). Bars indicate the mean ± SD of at least 15 replicate leaf samples. B, P. syringae-triggered SAR is diminished in all the SA biosynthetic mutants to a similar extent. To assess SAR establishment, three lower (1°) leaves per plant (Col-0, sid2, eds5, or pbs3) were either inoculated with Psm or mock-infiltrated with 10 mM MgCl2. Two days later, three upper (2°) leaves were challenge-inoculated with Psm lux, and bacterial numbers in the 2° leaves scored 60 h after the challenge-inoculation. Bars show the mean ± SD of at least 18 leaf replicates. A and B, individual data points of biological replicates are super-imposed on the bar graphs (small circles). Different letters denote significant differences (p < 0.05, ANOVA and post hoc Tukey HSD test).

Together, these results indicate that the proposed second PBS3-independent branch of ICS1-dependent SA biosynthesis that leads to the accumulation of unconjugated SA at later infection stages barely affects plant basal immunity or SAR. Moreover, it is reasonable to assume that the described metabolic disturbances, concerning the biosynthesis of NHP and camalexin, and the glucosylation of NHP are caused by the most common feature of the three mutants - their failure to exert the predominant first (ICS1-, EDS5-, and PBS3-dependent) branch of isochorismate-derived SA biosynthesis.

Discussion

The conversion of the key shikimate pathway intermediate chorismate to its isomer isochorismate is catalyzed by isochorismate synthase. Arabidopsis possesses two ICS isoforms, ICS1 and ICS2. While ICS1 is stress-inducible and required for more than 99% of the inducible biosynthesis of SA in P. syringae-inoculated leaves (Fig. 7A), ICS2 is constitutively expressed and does not contribute to stress-induced SA accumulation (5, 18, 19). Some bacterial ICS are bifunctional enzymes and possess, in addition to their chorismate-to-isochorismate-converting properties, an isochorismate pyruvate lyase (IPL) activity that directly produces SA from the originating isochorismate (20). However, both of the Arabidopsis ICSs were characterized as monofunctional isochorismate synthases that lack IPL activity and thus generate isochorismate as a reaction product ((4, 6); Fig. 9).Figure 9 Proposed scheme of biotic stress-inducible isochorismate synthase-associated metabolism in Arabidopsis. The biotic stress-inducible ICS isoform ICS1 catalyzes chorismate to isochorismate conversion in the chloroplast. The MATE transporter EDS5 presumably exports isochorismate out of the chloroplast. Both ICS1 and EDS5 are necessary for the accumulation of ICS-pathway products. Upper part: The cytosolic acyl acid amido synthetase PBS3 mediates the major part (∼95%) of SA biosynthesis via IC-Glu formation, which eliminates Pyr-Glu to form SA. SA can be in turn converted to glucose derivatives (SAG, SGE), other conjugates (SA-Mal, SA-Asp), methyl salicylate, and hydroxylated to dihydroxybenzoic acids. In addition, a moderate PBS3-independent path (5%) of ICS1/EDS5-dependent SA biosynthesis (that primarily produces unconjugated SA; green background) exists. Bottom part: isochorismate can also re-arrange to isoprephenate, which might either undergo dehydration to form 3-carboxyphenylpyruvate or be oxidised to 3-carboxy-4-hydroxyphenylpyruvate. The latter two compounds can be transaminated to meta-carboxy-Phe and meta-carboxy-Tyr, respectively, and are also plausible precursors for the detected meta-substituted benzoic acid (3-FBA, 3-CBA, 3-CMBA) and salicylic acid derivatives (5-FSA, 5-CSA, 5-CMSA). The formation of the formyl-SA/BA derivatives from phenylpyruvate-intermediates might by mediated by membrane-derived lipid hydroperoxides (LOOH). Arabidopsis leaves possess oxidase activity to convert the formyl-into carboxy-SA/BA derivatives. Decarboxylase reactions could convert the phenylpyruvate-intermediates into the carboxymethyl-SA/BAs. The compounds highlighted with red (fail to accumulate in pbs3 mutants) and blue (over-accumulate in pbs3, presumably due to redirection of metabolic flow) backgrounds have been analytically detected in this study. The accumulation of most of the ICS pathway products is negatively regulated by NPR1. Compounds highlighted in grey or compounds without background have not been analysed. The green arrows symbolize well-established conversions in planta, dashed arrows symbolize hypothetic conversion that are, nevertheless, plausible because of literature data. Please refer to the main text for further details.

Biosynthesis of biotic stress-inducible, meta-substituted formyl-, carboxy-, and carboxymethyl-benzoic acid and -salicylic acid derivatives in Arabidopsis

In the present study, we have conducted comparative metabolite analyses of wild-type plants and the ICS1-defective sid2 mutant to study isochorismate-derived metabolism in Arabidopsis leaves. Based on mass spectrometric and IR spectroscopic identification, we report the occurrence of the three 5-substituted SA derivatives 5-FSA, 5-CSA, 5-CMSA, and the three analogous BA derivatives 3-FBA, 3-CBA, and 3-CMBA in Arabidopsis (Fig. 1). These newly identified plant metabolites, which carry formyl-, carboxy-, and methylcarboxy moieties at the meta-positions relative to the carboxy groups of the common SA or BA core units, accumulate in leaves upon inoculation with compatible and incompatible P. syringae strains and were collectively designated as meta-substituted SA- and BA-derivatives (Figure 1, Figure 4, Figure 5 and 9; Fig. S5). The compromised accumulation of these compounds in the sid2 mutant indicates that the pathogen-inducible generation of isochorismate via ICS1 is necessary for their biosynthesis (Figs. 1A, and 4, A and B). Moreover, their formation requires the function of the chloroplast envelope-resident MATE transporter EDS5 (Fig. 4, A and B), which is presumed to mediate isochorismate transport out of plastids (21). Therefore, isochorismate exported from plastids might be required for the production of these substances. Likewise, the stress-inducible biosynthesis of SA strictly depends on both ICS1 and EDS5 (Fig. 7A). Biochemical and structural characteristics therefore appeared in line with a potential biosynthetic precursor function of SA for the formation of the six identified meta-substituted aromatics. However, exogenously administered SA or deuterium-labeled SA were not converted to any of the meta-substituted SA- and BA-derivatives in plants, irrespective of whether their leaves were inoculated with P. syringae or not (Fig. 3). Moreover, NahG overexpressing plants that, due to the SA hydroxylase activity introduced by the NahG transgene (46), rapidly degrade generated SA into catechol did not show compromised accumulation of the meta-substituted derivatives (Fig. 4, A and B). These findings indicate that SA does not serve as a biosynthetic precursor for the six meta-substituted aromatic acids (Fig. 9).

Relatively few meta-substituted aromatic natural products have been identified so far (51). Whereas the formation of 3-formyl-Tyr was described in the gamma-proteobacterium Pseudoalteromonas tunicate (52), meta-Tyr has been identified as an allelopathic constituent of root exudates of the grass species Festuca rubrum (53). Notably, the occurrence of four meta-substituted aromatic amino acids, meta-carboxy-Phe, meta-carboxy-Tyr, (3-carboxyphenyl)-glycine, and (3-carboxy-4-hydroxyphenyl)-glycine was reported in some plant species of the Brassicaceae, Resedaceae, Cucurbitaceae, and Iridaceae families (54). Incorporation of radiolabels into meta-carboxy-Phe and meta-carboxy-Tyr upon feeding with 14C- and 3H-labeled shikimate suggested that the meta-substituted aromatic amino acids are synthesized via the shikimate pathway (55). A closer analysis of the observed labeling patterns supported a biosynthesis from isochorismate and the thereof derived, putative intermediate isoprephenate. Subsequently produced 3-carboxyphenylpyruvate and 3-carboxy-4-hydroxyphenylpyruvate were proposed as specific biosynthetic precursors for the formation of meta-carboxy-Phe and meta-carboxy-Tyr, respectively (55). Consistently, Wu and colleagues (2024) recently provided biochemical evidence that the Arabidopsis aminotransferase REVERSAL OF SAV3 PHENOTYPE 1 (VAS1) catalyzes the transfer of amino groups from aromatic amino acids to the oxoacids 3-carboxyphenylpyruvate and 3-carboxy-4-hydroxyphenylpyruvate to generate meta-carboxy-Phe and meta-carboxy-Tyr, respectively (56). By employing an ics1ics2 double mutant, this study further corroborated that the meta-carboxy aromatic amino acids are derived from isochorismate, and demonstrated that the VAS1-catalyzed reaction contributes to the homeostasis of aromatic amino acids (56).

Based on the present results and the above-outlined previous reports on isochorismate-derived metabolism, we propose a plausible biosynthetic scheme for the formation of the six meta-substituted SA/BA derivatives identified in our study (Fig. 9). Isochorismate that accumulates as a consequence of stress-inducible ICS1-activity might be converted to isoprephenate. This rearrangement is analogous to the well-documented chorismate mutase reaction that converts chorismate to prephenate (57, 58), and might be catalyzed by a putative isochorismate mutase. Interestingly, an enzymatic activity that converted isochorismate to isoprephenate has been previously identified in protein extracts of woodland tobacco (Nicotiana sylvestris) (59). On one hand, isoprephenate could be dehydrated to 3-carboxyphenylpyruvate by a dehydratase in analogy to the preserved prephenate dehydratase-catalyzed reaction involved in plant Phe biosynthesis (58, 60). 3-carboxyphenylpyruvate could then serve as the precursor for the biosynthesis of the three meta-substituted benzoic acid derivatives, 3-formyl-BA, 3-carboxy-BA, and 3-carboxymethyl-BA (Fig. 9). On the other hand, isoprephenate might be enzymatically oxidized to 3-carboxy-4-hydroxyphenylpyruvate via a dehydrogenase-catalyzed reaction in which aromatization occurs by retention of the phenolic hydroxyl group. A comparable aromatization reaction represents the conversion of arogenate to Tyr catalyzed by arogenate dehydrogenase (61). 3-carboxy-4-hydroxyphenyl-pyruvate represents a plausible precursor for the biosynthesis of the three meta-substituted salicylic acid derivatives 5-formyl-SA, 5-carboxy-SA, and 5-carboxymethyl-SA (Fig. 9).

Chemically, the conversion of isochorismate to isoprephenate constitutes a Claisen-type of [3,3]-sigmatropic rearrangement with an energetically favored transition state (51). Consistently, a conversion of isochorismate to 3-carboxyphenylpyruvate, presumably via isoprephenate as the intermediate, has been observed to proceed non-enzymatically at temperatures of 100 °C as a main route of thermal isochorismate degradation (62, 63). Whether a spontaneous conversion of isochorismate to carboxyphenylpyruvate occurs under milder biological conditions is unclear, but a spontaneous, non-enzymatic formation of carboxyphenylpyruvate from isochorismate cannot be ruled out. However, because of the necessity of an electron acceptor in the proposed oxidation step of isoprephenate, the formation of 3-carboxy-4-hydroxyphenylpyruvate is unlikely to occur in a non-enzymatic manner (Fig. 9).

We propose that the two putative intermediates 3-carboxyphenylpyruvate and 3-carboxy-4-hydroxyphenylpyruvate are further converted to the identified meta-substituted BA- and SA-derivatives, respectively. As phenylpyruvate derivatives, 3-carboxyphenylpyruvate and 3-carboxy-4-hydroxyphenylpyruvate are convertible to the phenylacetate variants, 3-CMBA and 5-CMSA, respectively, via oxidative decarboxylation reactions (Fig. 9). Notably, an analogous biochemical conversion of phenylpyruvate to phenylacetate takes place in yeast as part of the Ehrlich pathway, a reaction sequence that mediates Phe catabolism. Herein, phenylpyruvate is decarboxylated by phenylpyruvate decarboxylase to phenylacetaldehyde which is in turn oxidized to phenylacetate by an aldehyde dehydrogenase (64, 65). A similar two-step conversion of phenylpyruvate to phenylacetate was identified in the soil bacterium Achromobacter eurydice (66).

The peroxidation of membrane lipids is a common metabolic process in plant tissue suffering from biotic or abiotic stress (67). In the course of the interaction of Arabidopsis leaves with compatible or hypersensitive response-inducing P. syringae strains, lipid hydroperoxides (LOOH) and other lipid peroxidation products are generated non-enzymatically following the production of reactive oxygen species (68, 69). Interestingly, chemical conversion of phenylpyruvate to benzaldehyde and phenylacetate was observed in the presence of LOOH derived from linolenic acid (70). Analogously, it is conceivable that lipid peroxidation products mediate an oxidative cleavage of 3-carboxyphenylpyruvate and 3-carboxy-4-hydroxyphenylpyruvate to the benzaldehydes 3-FBA and 5-FSA, respectively (Fig. 9). Our feeding experiments show that exogenously added 3-FBA and 5-FSA are rapidly oxidized to the respective carboxy-derivatives 3-CBA and 5-CSA in the leaf tissue, strongly suggesting that an aldehyde dehydrogenase activity is present in Arabidopsis that mediates these conversions (Fig. 3E). Because non-inoculated, uninduced plants were employed in this feeding assay, this presumed aldehyde dehydrogenase might be constitutively expressed in leaves. Interestingly, compared to the wildtype, the SA-deficient NahG plants over-accumulate the formyl variants 3-FBA and 5-FSA in response to P. syringae inoculation, while they accumulate reduced amounts of the carboxy-derivatives 3-CBA and 5-CSA (Fig. 4, A and B). This might indicate that the activity or expression of the proposed aldehyde dehydrogenase is modulated by SA.

Thus, a combination of enzymatic and non-enzymatical conversion steps might be involved in the synthesis of the six identified meta-substituted SA/BA-derivatives from isochorismate (Fig. 9). Theoretically, it is possible that during the infection process, bacterial enzymes catalyze specific steps of the proposed pathway in combination with plant enzymes. Such a scenario is supposed to imply a positive correlation between bacterial numbers in leaves and the pathway products. In particular, our finding that incompatible bacterial inoculation, which is associated with a limited bacterial propagation due to efficient activation of plant defense (71), induces the accumulation of meta-substituted SA/BA-derivatives to a similar extent than compatible bacterial infection argues against this possibility. The earlier production of meta-substituted SA/BA-derivatives in the incompatible Psm avrRpm1-Col-0 interaction compared to the compatible Psm-Col-0 interaction is reminiscent of the accumulation patterns of different plant-derived metabolites, such as indolics or tocopherols (69, 72), and rather suggests that meta-SA/BAs are generated by an active plant response that is elicited following pathogen inoculation (Fig. S5A).

Under pathogen-inoculation conditions, a major portion of the ICS1-generated isochorismate is used for the PBS3-mediated biosynthesis of SA and the SA-derived derivatives such as SAG, SGE, SA-Mal, and SA-Asp (Fig. 5). For example, out of the total levels of measured SA derivatives, about 2% SA, 90% SAG, 3% SGE, 2% of SA-Mal, and 1% SA-Asp accumulated in the Col-0 wildtype at 48 h post Psm inoculation, while the contribution of the meta-substituted SA/BA-derivatives together was less than 2% (Fig. 5). In the pbs3 mutant, however, the metabolic route towards SA is largely blocked (Figs. 4C, and 7A). A redirection of the isochorismate-derived metabolic flow toward the production of the meta-substituted SA/BA-derivatives would explain our findings that the latter compounds strongly over-accumulate in P. syringae-inoculated pbs3 mutant plants (Fig. 4, A and B). Our model also proposes 3-carboxyphenylpyruvate and 3-carboxy-4-hydroxyphenylpyruvate as central pathway intermediates that are derived from isochorismate (Fig. 9). This would explain why a chemically identical set of BA and SA derivatives carrying formyl-, carboxy-, and carboxymethyl groups at the meta-positions of their aromatic rings are generated in a strikingly concerted mechanistic and temporal manner in response to pathogen attack (Figs. 4, A and B, and 5). Whether the two previously described 3-carboxyphenylpyruvate- and 3-carboxy-4-hydroxyphenyl-pyruvate-derived metabolites meta-carboxy-Phe and meta-carboxy-Tyr (56), respectively, accumulate upon biotic stress remains to be determined. A previous microarray study indicates a modest increase of transcripts of the VAS1 gene in Psm-infected Col-0 leaves (Fig. S8; (73)).

SA-Mal and SA-Asp are biosynthesized from SA and require PBS3 for their formation

Another hitherto undescribed SA derivative that we found to occur in Arabidopsis leaves is the malate ester of SA, salicyloyl-malate (SA-Mal) (Fig. 2, A–D). Just as the amide salicyloyl-aspartate (SA-Asp) that was previously described as a plant natural product (37, 38, 39, 40, 41), SA-Mal accumulated upon P. syringae inoculation in strict dependency of the SA biosynthetic genes ICS1, EDS5, and PBS3 (Fig. 4C). Moreover, the SA-degrading NahG plants were unable to accumulate both SA-Mal and SA-Asp (Fig. 4C). In addition, exogenously administered (deuterium-labeled) SA was readily converted to the (deuterium-labeled) SA-Mal and SA-Asp variants in P. syringae inoculation leaves (Figs. 2, B and E, and 3). These findings demonstrate that both SA-Mal and SA-Asp are synthesized from SA as a direct or indirect biosynthetic precursor.

A prominent malate ester generated in the vegetative tissue of Arabidopsis is the UV-B protective metabolite sinapoyl-malate, which is synthesized from the glucose ester sinapoyl-1-glucose by a serine carboxypeptidase-like (SCPL) acyltransferase (74, 75). Whether SA-Mal is synthesized by similar biochemical principles remains to be determined. Parallel P. syringae-induced accumulation patterns of SGE and SA-Mal, though, do not argue against the possibility that SGE could function as an SA-Mal precursor ((11); Figs. 4, and S2).

The acyl acid amido synthetase GH3.5 was previously implicated in the formation of SA-Asp. GH3 enzymes generally catalyze the ATP-dependent adenylation of the carboxylic group of organic acid and a subsequent displacement of AMP from this adenylated intermediate by the amine group of an L-amino acid, so that an amide conjugate is formed as a product (76). A role for GH3.5 in the biosynthesis of SA-Asp was supported by in vitro assays, since the purified GH3.5 protein showed adenylation activity towards SA, indole acetic acid (IAA), and several other acid substrates, and furthermore catalyzed the formation of SA-Asp from SA and Asp (40, 41, 42). Also, over-expression of GH3.5 in Arabidopsis resulted in increased accumulation of SA-Asp. However, since gh3.5 knockout plants did not exhibit reduced SA-Asp levels (40), the biosynthesis of SA-Asp in Arabidopsis seems not yet fully understood. Nevertheless, our analyses illustrate that the pathogen-inducible generation of SA-Asp proceeds via isochorismate pathway-derived SA formation and via similar kinetical and regulatory principles than the generation of SA-Mal (Figs. 4C, 5C, and 9; Figs S2 and S4).

Exogenous application of 5-carboxy-SA and SA-Mal has a moderate impact on plant immunity

To test possible immune-stimulating activities of the here-identified SA-related compounds, we applied commercially available (5-FSA, 5-CSA, 3-FBA, 3-CBA, 3-CMBA) or chemically synthesized (SA-Mal) substances in the same concentrations than SA (0.5 mM) to Arabidopsis plants and assessed the induction PR1 gene expression and pathogen resistance to P. syringae (Fig. 6). Our results indicate that none of the compounds exhibits immune-inducing properties to a similar magnitude than SA. Out of the applied meta-substituted SA/BA derivatives, only 5-CSA exhibited a modest resistance-enhancing activity in wild-type plants, while 5-FSA, 5-CSA, 3-FBA, and 3-CBA, caused marginal effects on resistance- and/or PR1-induction in sid2 plants. These results, together with our findings that they accumulate to low absolute levels and relatively late after pathogen inoculation (Fig. 5), suggest a minor role for meta-substituted SA/BA derivatives in inducible plant immunity.

In accordance with earlier results (30), we found that the 3-substituted SA derivative 2,3-DHBA exerts moderate resistance- and PR1-inducing activities that were, however, markedly lower than the respective activities of SA. By contrast, the 5-substituted isomer 2,5-DHBA (gentisic acid), did not exert any immune-enhancing effect (Fig. 6, B and C). The weak or non-existent immune-activating properties of 5-FSA, 5-CSA, and 2,5-DHBA illustrate that a 5-substitution in the aromatic ring of SA can severely compromise the biological activity of the SA molecule. However, this is not necessarily true for every 5-substitution, because exogenous application of the synthetic SA derivative 5-chloro-SA resulted in marked elevation of PR1 transcript levels and resistance-induction in plants (77). Consistently, in vitro binding assays showed that the SA receptor NPR1, which mediates SA-triggered immune responses (78), tightly interacts with 5-chloro-SA in addition to SA (15). By contrast, benzoic acids that lack the 2-hydroxy group at the ortho-position of the benzene ring do not bind to NPR1 and hardly exhibit immune-stimulating activity (15, 77). The 2-hydroxy group as a key structural prerequisite for the immune-activating properties of salicylates is absent in 3-FBA, 3-CBA, and 3-CMBA, which offers an explanation for the observation that the three exogenously applied meta-substituted BAs had virtually no impact on immune responses (Fig. 6, A and C).

We assessed a marked resistance enhancement in plants treated with 0.5 mM SA-Mal, but the resistance effect was lower than for SA treatments (Fig. 6E). When infiltrated into leaves, a substantial portion of the applied SA-Mal was relatively rapidly hydrolyzed to SA, which might indicate the presence of a respective esterase activity in the plant tissue (Fig. 6G). Therefore, we cannot exclude that the observed resistance effects observed in response to exogenously applied SA-Mal solutions were due to the release of free, signal-active SA. Modest resistance-activation effects were also reported when SA-Asp was fed exogenously to plants (37, 40). However, prerequisites for convincing conclusions about the biological functions of SA-Mal and SA-Asp are the elucidation of their biosynthesis and the functional characterization of appropriate gene mutant lines that specifically lack the accumulation of these SA conjugates.

A moderate PBS3-independent path to isochorismate-derived SA operates in Arabidopsis leaves at later stages of bacterial infection

Bacterial IPLs eliminate the enolpyruvyl side chain of their substrate isochorismate to generate SA and pyruvate, but attempts to identify plant enzymes with IPL activities were unsuccessful in the past (20, 79). In 2019, two parallel studies reported on a plant-specific circuit strategy for isochorismate to SA conversion via PBS3-catalysed conjugation of L-Glu to isochorismate. SA is then synthesized from the resulting intermediate IC-Glu by elimination of Pyr-Glu (21, 22). In combination with detailed in vitro biochemical characterization of PBS3, both studies primarily relied on the phenotypic and metabolic characterization of different Arabidopsis autoimmune mutant lines that over-accumulated, in a PBS3-dependent manner, IC-Glu, Pyr-Glu, and SA in the absence of biotic stress (21, 22). The significance of PBS3 for basal and effector-triggered immunity to P. syringae was established in preceding studies, which also found that accumulation of SA and SAG, and PR1 gene expression following microbial inoculation was significantly reduced in pbs3 mutants (45, 80, 81). Our study confirms the importance of PBS3-mediated SA generation and the previously reported biochemistry via IC-Glu formation and subsequent release of Pyr-Glu in a well-defined plant-pathogen interaction context. Our time-course analysis shows that rises of IC-Glu and Pyr-Glu are detectable at 10 h post Psm inoculation and then become gradually larger until 48 hpi (Fig. 5C).

A direct comparison of the levels of totally accumulating SA (i.e., the sum of SA and SA conjugates) showed that the increases of total SA in sid2, eds5, and pbs3 following Psm inoculation amounted to ∼0.5%, ∼1.5% and ∼5.5% of the corresponding increases in the wildtype Col-0 at 48 hpi (Fig. 7A). This suggests that almost the whole pathogen-induced biosynthesis of SA proceeds via ICS1 and, almost as stringently, via EDS5. The residual, faint amounts of SA accumulating in sid2 leaves, which are below the basal levels of the wildtype (Fig. 7A), might be generated by the activity of the second Arabidopsis isochorismate synthase isoform, ICS2, or via the PAL pathway (5, 17). Nevertheless, these data illustrate that blockage of the isochorismate pathway in the leaves of the sid2 and eds5 mutants does not lead to a significant compensation of SA biosynthesis via alternative routes (Fig. 7A), such as the PAL pathway. Moreover, increased PAL activity upon pathogen detection has been functionally associated with early cell wall-based defenses and lignification events (82).

Notably, ∼5% of this isochorismate-derived SA accumulates in a PBS3-independent manner (Fig. 9). Just as eds5, pbs3 mutants are strongly compromised in the accumulation of the SA conjugates SAG, SGE, SA-Mal and SA-Asp. However, pbs3 only shows compromised accumulation of unconjugated SA at earlier infection phases (6 and 10 hpi), but accumulates free SA to wildtype-like levels (24 hpi) or to even higher amounts (48 hpi) at later interaction periods (Figs. 4C, and 7A). This suggests the existence of a moderate PBS3-independent branch to SA that operates later in the plant-pathogen interaction phase and mainly produces unconjugated SA. Our data are consistent with previous studies that, depending on the chosen pathogen strain and sampling time, either detected attenuated and not entirely compromised P. syringae-induced accumulation of free SA, or over-accumulation of free SA in pbs3 mutants. In all of the cases, however, the accumulation of SAG was strongly impaired (45, 80, 81). The PBS3-independent accumulation of free SA might be explained by a non-enzymatic [1,5]-sigmatropic rearrangement of isochorismate to SA and pyruvate rather than by a PAL-associated replenishment pathway for SA (see argumentation above; (21, 51, 62)). A spontaneous chemical conversion of isochorismate to SA is expected to proceed by a lower reaction rate than an enzyme-assisted conversion via PBS3, and therefore could gain significance in later periods after a pathogen-induced ICS1 activation and a resulting isochorismate formation. This would explain the observed time-dependent accumulation characteristics of free SA in pbs3 (Fig. 7A). The apparently compromised SA conjugation in pbs3 might then be related to spatial or temporal inaccessibility of the spontaneously generated SA to conjugating enzymes.

Our data indicate that sid2, eds5, and pbs3 plants share several immune-related features and metabolic characteristics that were previously described for the sid2 mutants (10, 12, 18, 19, 49, 50). All of the mutants show a comparable attenuation of basal immunity to P. syringae. In addition, SAR induced biologically in plants via pathogen inoculation or chemically via exogenous treatment with the natural SAR hormone NHP is compromised to similar degrees in all of the mutant types (Fig. 8). With respect to the metabolic irregularities, the three mutant lines show reduced accumulation of the defense metabolites NHP, Pip and camalexin during the early Psm-Arabidopsis interaction phase (10 hpi), over-accumulation of NHP and its glucose ester NHPGE at later periods (from 24 hpi onward), and attenuated accumulation of the NHP glucoside NHPG throughout the studied plant-pathogen interaction (Figs. 7, B and C, and S6). These findings suggest that the defects in the predominant PBS3-dependent branch of SA biosynthesis are causative for the above-described immune defects and metabolic disturbances, while the PBS3-independent branch that results in a later accumulation of free SA does not markedly affect immune responses. This is consistent with previous conclusions that SA primes NHP and camalexin biosynthesis, particularly during earlier plant-pathogen interaction phases, and prevents an excess accumulation of NHP during later infection stages by promoting NHP to NHPG conversion (11, 12, 26, 50). In addition, the enhanced susceptibility phenotype of the immune-compromised lines sid2, eds5, pbs3, npr1 or NahG might contribute to some of the observed over-accumulation patterns of metabolites at later time points (especially 48 hpi), because higher bacterial numbers in the tissue of susceptible plants likely exert enhanced response stimulation than in the tissue of the more resistant wild-type (Figs. 4 and 7). Nevertheless, as discussed here, regulatory principles and the re-direction of plant metabolic pathways in biosynthetic mutants exert strong influences on the production of metabolites in the plant lines under investigation.

Metabolic accumulation patterns argue against a previously proposed involvement of EDS5 in the biosynthesis of NHP and associated transport processes

The biosynthesis of the SAR-inducing defense metabolite NHP proceeds in three enzymatic steps out of L-Lys (9). The first two steps convert L-Lys to Pip via the aminotransferase ALD1 and the reductase SARD4 (44, 83, 84). The subcellular localization of ALD1-and SARD4-fusion proteins thereby suggests a plastidial biosynthesis of Pip (85, 86). In a third step, the flavin-dependent monooxygenase FMO1 catalyses the N-oxidation of Pip to NHP (10), but the subcellular localization of FMO1 is unclear yet. However, considering the common occurrence of cytosolic or endoplasmic reticulum-located monooxygenation reactions (87), a transport of Pip out of the plastid across the plastid envelope might be necessary for the completion of NHP biosynthesis. A preliminary study by Rekhter and colleagues, which had been posted several years ago as a pre-print to bioRxiv but, to our knowledge, not been published so far in a peer-reviewed journal, proposed a model in which the chloroplast envelope-resident transporter EDS5 was involved in the export of Pip out of the plastid and in NHP biosynthesis. This rather speculative assumption was primarily based on reduced NHPG levels detected in eds5 mutants upon UV light exposure (88), but some reviews already integrated this proposition into NHP biosynthesis-related models (e.g., (89)). However, the strong accumulation of NHP in Psm-inoculated eds5 mutants clearly exclude that EDS5 is necessary for NHP biosynthesis or involved in a possibly related transport of Pip out of plastids (Fig. 7B). The uniform accumulation characteristics of NHP and NHPG in P. syringae-inoculated sid2, eds5, and pbs3 plants outlined above rather indicate that SA functions in the promotion of the early biosynthesis of NHP and the containment of NHP accumulation at later stages by supporting NHP to NHPG conversion (Fig. 7B).

The different branches of the isochorismate pathway are negatively regulated by NPR1

Another striking metabolic feature we observed in this study was the over-accumulation of every of the ten characterized ICS1-dependent Arabidopsis metabolites (1–10) in the npr1 mutant after Psm inoculation (Figs. 4, and S2). Likewise, SA and SGE over-accumulated in npr1 in response to bacterial infection, whereas SAG did not (Fig. S2; (11)). In addition, the basal levels of 2,5-DHBA were found to be decreased in the npr1 mutant (90). This illustrates that the SA receptor NPR1 has a profound impact on the Arabidopsis isochorismate pathway. The concerted negative impact of functional NPR1 on the meta-substituted SA- and BA-derivatives, SA, SA conjugates (SGE, SA-Mal, SA-Asp), and SA biosynthetic intermediates (IC-Glu, Pyr-Glu) indicates that NPR1 negatively regulates the isochorismate pathway at an early step upstream of isochorismate production, which could lead to an over-accumulation of isochorismate in npr1 mutants, and consequently, enhanced production of isochorismate-derived compounds under pathogen infection conditions (Fig. 9). This hypothesis is consistent with the over-production of isochorismate in a snc2npr1 double mutant that exhibits snc2-triggered activation of autoimmune pathways (21), and with the observed negative impact of functional NPR1 on ICS1 expression (90). Additionally, NPR1 positively affects the strongly pathogen-inducible expression of UGT76B1, which encodes a glucosyltransferase that simultaneously glucosylates SA and NHP to their glucosides SAG and NHPG under biotic stress conditions (11, 26, 90). Quantitatively, SAG is the main accumulating SA derivative in Arabidopsis (Fig. 5; (9)), and an attenuated UGT76B1 expression in npr1 might thus redirect the SA metabolic pathway towards accumulation of free SA and its conjugates SGE, SA-Mal, and SA-Asp (Fig. 9). Similarly, the expression of the SA-5-hydroxylase gene S5H was promoted by NPR1 (90), which might also contribute to the over-accumulation of SA and SA conjugates in npr1 mutant plants due to metabolic backlog. Together, our study suggests an important function of NPR1 in the homeostasis of a whole set of differently regulated isochorismate pathway-derived metabolites in pathogen-attacked plants.

Experimental procedures

Seed material and plant cultivation

Arabidopsis (A. thaliana) plants were cultivated individually in pots containing a mixture of soil (Substrat BP3; Klasmann-Deilmann), vermiculite, and sand (8:1:1) in a growth chamber with a 10-h-day (9 AM to 7 PM; 100 μmol m−2 s−1 photon flux density)/14-h-night cycle and a relative humidity of 60%. Day and night temperatures were 21 °C and 18 °C, respectively (10). The following Arabidopsis gene knock-out lines were employed in this study: sid2 (sid2-1; (18)), eds5 (eds5-1) (91), pbs3 (pbs3-2; Salk_018225; (45)), npr1 [npr1-3; Nottingham Arabidopsis Stock Centre (NASC) ID: N3802]. Moreover, the NahG overexpressing line (NahG) was used (46). All lines are in the Col-0 (N1092) background.

Bacterial inoculation and treatment assays for the determination of disease resistance, metabolite analyses, and defence gene expression

P. syringae pv. maculicola strain ES4326 (Psm), Psm expressing the luxCDABE operon from Photorhabdus luminescens (Psm lux), and Psm expressing the AvrRpm1 avirulence gene (Psm avrRpm1) were used for bacterial inoculation assays associated with metabolite analyses and bacterial growth assessments, respectively. The bacterial strains were cultivated at 28 °C in King’s B medium with the appropriate antibiotics as previously described (10, 34, 92, 93). For plant inoculation, bacterial suspensions from overnight cultures were washed and diluted with 10 mM MgCl2 to final optical densities at 600 nm (OD600) of 0.005 (Psm and Psm AvrRpm1) and 0.001 (Psm lux). Bacterial suspensions were infiltrated into the abaxial sides of Arabidopsis rosette leaves with needleless syringes between 10 AM and 11 AM.

For the determination of metabolite contents in Arabidopsis leaves following bacterial attack, suspensions of Psm (OD600 = 0.005) were infiltrated into three rosette leaves of 5-week-old plants. As a mock-control treatment, a 10 mM MgCl2 solution was infiltrated instead. The treated leaves were harvested at different times (6, 10, 24 or 48 h post treatment), leaf fresh weights (FW) were determined, and harvested leaves shock-frozen in liquid nitrogen. Each replicate sample consisted of six leaves from two different plants. Four to five replicate samples were usually analyzed in each experiment.

For assessments of resistance to bacterial infection, purified suspensions of the bioluminescent Psm lux (OD600 = 0.001) were syringe-infiltrated into three Arabidopsis rosette leaves (87, 88). Bacterial numbers were determined 2.5 days post-inoculation by measuring the luminescence of leaf discs punched out of inoculated leaves (one disc per inoculated leaf) with a Sirius FB12 luminometer (Berthold Detection Systems, http://www.titertek-berthold.com). As a measure of disease susceptibility, bacterial numbers were assessed as relative light units (rlu) per cm2 leaf area. At least 15 replicate leaf samples were assayed for each genotype.

To assess the induction of resistance and defense gene expression following treatments with SA (Sigma, S5922), meta-substituted SA/BA derivatives [3-FBA (Acros, 296560010), 3-CBA (Alfa Aesar, A14445), 3-CMBA (Fluorochem, 446214), 5-FSA (Sigma, F17601), 5-CSA (Sigma, 797057)], other SA derivatives [2,3-DHBA (Acros, A0350739), 2,5-DHBA (Fluka, 537960), SA-Mal (chemically synthesized, see below)], malate (Roth, 3034.1), and fumarate (Fluka, 47900), 0.5 M stock solutions in ethanol were freshly prepared, and diluted with water to a working concentration of 0.5 mM. The 0.5 mM solutions were then infiltrated into three rosette leaves of a given Arabidopsis plant. As a mock-control treatment, an aqueous solution containing 0.1% ethanol was infiltrated. For the assessment of defence gene expression, six leaves from two plants were harvested 4 h after the treatment to obtain one replicate sample. To assess resistance induction, the treated leaves were inoculated with Psm lux 4 h after the pre-treatment, and bacterial growth determined via bacterial luminescence as described above.

To examine SAR establishment following exogenous N-hydroxypipecolic acid application, 10 ml of a freshly prepared 1 mM NHP solution or 10 ml of water (control treatment) was pipetted onto the soil of the individually cultivated plants (10). Challenge inoculations with Psm lux were performed 1 day later as described above. To assess SAR following a bacterial inducer inoculation, three lower rosette leaves of a given plant were inoculated with Psm (OD600 = 0.005) or mock-infiltrated with 10 mM MgCl2. Two days later, three upper leaves were challenge-inoculated with Psm lux (OD600 = 0.001), and the numbers of Psm lux assessed 2.5 days later as described above (93).

Chemical synthesis of salicyloyl-malate (SA-Mal)

SA-Mal (racemic O-salicyloyl malic acid) was obtained by a three-step chemical synthesis from malic acid and O-benzyl salicylic acid. The synthesis and spectroscopic characterization of SA-Mal are outlined in detail in Fig. S3.

Metabolite analysis by GC-MS or GC-FTIR via vapor phase extraction and analyte derivatization by methylation

The initial comparative metabolite analyses and the quantitative determination of the levels of the thereby identified meta-substituted SA/BA derivatives (5-FSA, 5-CSA, 5-CMSA, 3-BSA, 3-BSA, 3-CMBA), SA-Mal, SA-Asp, Pyr-Glu, and IC-Glu were performed via a vapor-phase extraction-based work up of leaf extracts according to Schmelz and colleagues (43), coupled with subsequent GC-MS or GC-FTIR analysis of the resulting derivatized samples (10, 44). 150 to 200 mg of shock-frozen leaf material (pooled from six leaves) was ground to a fine powder with a pre-chilled ball mill and immediately extracted with 600 μl of H2O:1-propanol:HCl (1:2:0.005; v/v/v) that was pre-heated to 70 °C and contained 100 ng of deuterium-labeled SA (D6-SA; Sigma-Aldrich, S16796) and dihydrojasmonic acid (TCI, D3225) as internal standards for the routine procedure [The D6-SA standard was, however, omitted from the extraction buffer when in planta labeling experiments with deuterated SA were analyzed. Please also note that D6-SA rapidly exchanges the two acidic deuterium atoms by hydrogens in protic solvents and thus in principle prevails as D4-SA in solution (Fig. S4).] After vortexing for 15 s, 1 ml methylene chloride was added, the suspension vigorously mixed by further vortexing for 30 s, and centrifuged at 14,000g for 1 min to facilitate phase separation. The lower organic phase was removed, dried over ∼10 mg of Na2SO4, and incubated for 5 min at room temperature with 4 μl of 2 M trimethylsilyl-diazomethane in hexane (Sigma-Aldrich), which converts analytes containing carboxylic acid groups into the corresponding methyl esters. The methylation reaction was stopped by the addition of an excess of acetic acid (4 μl of a 2 M solution in hexane), and the sample was subsequently subjected to a vapor phase extraction procedure using a volatile collector trap packed with Porapak-Q absorbent (VCT-1/4X3-POR-Q; Analytical Research Systems) (43). For that purpose, the sample was heated to 70 °C under a steady stream of nitrogen until complete evaporation of the solvent. The temperature was then increased to 200 °C for 2 min. The absorbed metabolites were eluted from the collector column with 1 ml methylene chloride. The collection was concentrated to 30 μl under a nitrogen stream and transferred to GC-vials. For gas chromatographic separation, 4 μl of the sample was injected into a 7890A GC (Agilent Technologies) equipped with a ZB5 MS capillary column (Zebron). GC injector temperature was set to 250 °C, a constant flow of helium (1.2 ml/min) was applied and the following temperature program was used: 50 °C/3 min with 8 °C/min to 240 °C, with 20 °C/min to 320 °C/3 min.

Mass spectra were recorded with a 5975C mass spectrometric detector (Agilent Technologies) in the electron ionization mode at 70 eV. The GC-MS data was evaluated using MSD ChemStation software version E.02.01.1177 (Agilent Technologies). For quantitative analysis, peaks of analytes and internal standards from selected ion chromatograms were integrated: 5-FSA (m/z 148), 5-CSA (m/z 178), 5-CMSA (m/z 133), 3-FBA (m/z 133), 3-CBA (m/z 163), 3-CMBA (m/z 177), SA-Mal (m/z 120), SA-Asp (m/z 121), Pyr-Glu (m/z 142), IC-Glu (m/z 142), D4-SA (internal standard; m/z 124), dihydrojasmonic acid (internal standard; m/z 156). Experimentally determined correction factors reflecting the ratios of areas of internal standard to analyte were considered for the quantification of 5-FSA, 5-CSA, 3-FBA, 3-CBA, 3-CMBA, and SA-Mal. Due to the unavailability of authentic compounds for 5-CMSA, SA-Asp, Pyr-Glu, and IC-Glu, correction factors were estimated considering the proportion of the selected m/z-value on the total ions of the mass spectra. All calculated values were related to the fresh weight (FW) of the leaf samples.

GC-FTIR spectra were acquired using a Hewlett-Packard 6890 Series GC coupled with an IRD3 infrared detector manufactured by ASAP Analytical (Analytical Solutions and Providers), as detailed previously (44). The GC-specific settings and the column were the same as described above, except that the flow rate of the helium carrier gas was 2 ml min−1. Infrared spectra were recorded from 4000 to 600 cm−1 with a resolution of 16 cm−1 and a scan rate of 8 scans per second. The IRD method parameters were as follows: Resolution = 16; Apodisation = Triangle; Phase correction = Mertz; Zero-Fill = 1; Co-Add = 2. The temperature of the transfer line and flow cell were both set to 250 °C, and nitrogen was used as sweep gas. The FTIR data were analyzed with the software Essential FTIR© (v3.10.037; Operant LLC).

Metabolite analysis by GC-MS via analyte derivatization by trimethylsilylation

The quantitative assessment of the leaf levels of SA, SAG, SGE, NHP, NHPG, NHPGE, and camalexin was performed by a GC-MS-based analysis of trimethylsilylated analytes as described in detail in previous studies (10, 11). The absolute metabolite levels were related to the leaf FW.

Metabolite analysis by LC-q-TOF-MS

The harvested, shock-frozen leaves were grounded to a fine powder with a ball-mill in a 2 ml Eppendorf tube. 600 μl of MeOH:H2O (80:20, v/v) was added, the sample was thoroughly vortexted and extracted for a further 10 min on a rotatory shaker at 4 °C. After centrifugation at 14000 rpm, the supernatant was carefully removed and the pellet extracted again with 600 μl of MeOH:H2O (80:20, v/v). The solvent of the combined extracts was evaporated at 30 °C under vacuum using a ScanSpeed vacuum centrifuge (Labogene ApS). The dry residue was redissolved in 100 μl of LC-MS-grade MeOH:H2O (80:20, v/v) and the solution filtered through a nylon centrifugal filter (VWR, 516-0233) at 14000 rpm. 5 μl of the filtrate was injected into a 1260 Infinity II Prime LC coupled to a 6540-quadrupole time-of-flight mass spectrometer, which was equipped with a dual electrospray ionization (ESI) source (Agilent Technologies). For chromatographic separation, an InfinityLab Poroshell 120 EC-C18 column (3.0 × 100 mm i.d., 2.7 μm particle size) was used. The mobile phase solvents were HPLC-grade water (solvent A) and acetonitrile (solvent B), both supplemented with formic acid (0.1% v/v). A gradient program was applied for the separation of the analytes, in which the percentage of solvent B was linearly changed as follows: 0 min, 5%; 1 min, 5%; 6 min, 25%; 12 min, 25%; 16 min, 50%; 20 min, 75%; 23 min, 95%; 26 min, 95%; 28 min, 5%; 34 min, 5%. The flow rate and column temperature were set to 0.7 ml min−1 and 40 °C, respectively. The MS analysis was performed in the negative ionization mode. Nitrogen was used as the nebulizer and drying gas and was set to 40 psi and 10 l/min, respectively. The drying gas temperature was set to 325 °C. The instrument ion optic voltages were as follows: fragmentor 130 V, skimmer 65 V and octopole RF 750 V. The capillary and nozzle voltages were 3500 V and 300 V, respectively. The data was acquired in the MS1 mode between m/z 50 and m/z 1700 and processed using MassHunter software (Agilent). For relative quantification, the peak areas of specific extracted ion chromatograms ([M-H]− ions or other main fragment ions of the analytes were integrated and related to the sample FW (Fig. S2).

Determination of PR1 transcript levels by RT-qPCR analysis

Leaf transcript levels of PR1 were determined by RT-qPCR analysis using 50 mg of frozen and ground leaf tissue. The RNA isolation, cDNA synthesis, and RT-qPCR steps have been previously outlined in detail (49, 83). The POLYPYRIMIDINE TRACT-BINDING PROTEIN 1 (PTB1) gene was used as a reference gene. Fig. S7 lists the primers used for RT-qPCR analysis. Two technical replicates were assessed for one sample, and their means were taken to obtain expression values for one biological replicate. The PR1 transcript levels were expressed relative to the mean value of the Col-0-mock-control sample (Fig. 6, C and D).

Statistical procedures

The numbers of biological replicates for each experiment are always indicated in the figure legends. Numerical values of the luminescence data (rlu cm−1) of the bacterial growth assays were log10-transformed and subject to ANOVA with post hoc Tukey’s HSD test (p < 0.05 for each data subset; (10)). For metabolite and RT-qPCR-derived gene expression results, non-transformed numerical values were analyzed by ANOVA with post hoc Tukey’s HSD test (p < 0.05) or by a non-parametric one-way ANOVA according to Kruskal-Wallis with stepwise step-down comparisons (p < 0.05). ANOVA analyses were performed with the SPSS statistical software (version 26; IBM Corporation). Additionally, for pairwise comparisons between control- and treatment samples, a two-tailed Student’s t test was performed using Microsoft Excel. The depicted results were confirmed in at least one other independent experiment.

Data availability

All data are contained within the manuscript.

Supporting information

This article contains supporting information (73, 94, 95, 96, 97).

Conflict of interest

The authors declare that they have no conflict of interest with the contents of this article.

Supporting information

Supporting information

Author contributions

N. S., M. H., S. A., L. A., E. R., and J. Z. performed the metabolite analyses by GC-MS and the associated plant experiments; E. R. and J. Z. conducted the LC-MS analyses; N. S. and S. A. performed the bacterial assays; N. S. conducted gene expression analysis; A. A. and T. J. J. M. performed the chemical synthesis of SA-Mal and the associated conceptualization and documentation; N. S. and M. H. contributed to the conceptualization of the research; J. Z. secured funding, conceived the research, performed final metabolite analysis and data evaluation and wrote the manuscript.

Funding and additional information

This work was funded by the 10.13039/501100001659 Deutsche Forschungsgemeinschaft (DFG; 10.13039/501100001659 German Research Foundation ) via the 10.13039/501100001659 DFG grant ZE467/6-2 , the graduate program IRTG 1525, and within Germany’s Excellence Strategy (EXC 2048/1 – project ID 390686111 ).
==== Refs
References

1 Herrmann K.M. Weaver L.M. The shikimate pathway Annu. Rev. Plant Physiol. Plant Mol. Biol. 50 1999 473 503 15012217
2 Tzin V. Galili G. The biosynthetic pathways for shikimate and aromatic amino acids in Arabidopsis thaliana Arabidopsis Book 8 2010 e0132 22303258
3 Widhalm J.R. Dudareva N. A familiar ring to it: biosynthesis of plant benzoic acids Mol. Plant 8 2015 83 97 25578274
4 Strawn M.A. Marr S.K. Inoue K. Inada N. Zubieta C. Wildermuth M.C. Arabidopsis isochorismate synthase functional in pathogen-induced salicylate biosynthesis exhibits properties consistent with a role in diverse stress responses J. Biol. Chem. 282 2007 5919 5933 17190832
5 Garcion C. Lohmann A. Lamodière E. Catinot J. Buchala A. Doermann P. Characterization and biological function of the ISOCHORISMATE SYNTHASE2 gene of Arabidopsis Plant Physiol. 147 2008 1279 1287 18451262
6 Macaulay K.M. Heath G.A. Ciulli A. Murphy A.M. Abell C. Carr J.P. The biochemical properties of the two Arabidopsis thaliana isochorismate synthases Biochem. J. 474 2017 1579 1590 28356402
7 Gross J. Cho W.K. Lezhneva L. Falk J. Krupinska K. Shinozaki K. A plant locus essential for phylloquinone (vitamin K1) biosynthesis originated from a fusion of four eubacterial genes J. Biol. Chem. 281 2006 17189 17196 16617180
8 Spoel S.H. Dong X. Salicylic acid in plant immunity and beyond Plant Cell 36 2024 1451 1464 38163634
9 Zeier J. Metabolic regulation of systemic acquired resistance Curr. Opin. Plant Biol. 62 2021 102050
10 Hartmann M. Zeier T. Bernsdorff F. Reichel-Deland V. Kim D. Hohmann M. Flavin monooxygenase-generated N-hydroxypipecolic acid is a critical element of plant systemic immunity Cell 173 2018 456 469 29576453
11 Yildiz I. Mantz M. Hartmann M. Zeier T. Kessel J. Thurow C. The mobile SAR signal N-hydroxypipecolic acid induces NPR1-dependent gene expression and immune priming Plant Physiol. 186 2021 1679 1705 33871649
12 Löwe M. Jürgens K. Zeier T. Hartmann M. Gruner K. Müller S. N-hydroxypipecolic acid primes plants for enhanced microbial pattern-induced responses Front. Plant Sci. 14 2023 1217771
13 Shulaev V. Leon J. Raskin I. Is salicylic acid a translocated signal of systemic acquired resistance in tobacco? Plant Cell 7 1995 1691 1701 12242358
14 Mou Z. Fan W. Dong X. Inducers of plant systemic acquired resistance regulate NPR1 function through redox changes Cell 113 2003 935 944 12837250
15 Wu Y. Zhang D. Chu J.Y. Boyle P. Wang Y. Brindle I.D. The Arabidopsis NPR1 protein is a receptor for the plant defence hormone salicylic acid Cell Rep. 1 2012 639 647 22813739
16 Ding Y. Sun T. Ao K. Peng Y. Zhang Y. Li X. Opposite roles of salicylic acid receptors NPR1 and NPR3/NPR4 in transcriptional regulation of plant immunity Cell 173 2018 1454 1467 29656896
17 Lefevere H. Bauters L. Gheysen G. Salicylic acid biosynthesis in plants Front. Plant Sci. 11 2020 338 32362901
18 Nawrath C. Métraux J.P. Salicylic acid induction-deficient mutants of Arabidopsis express PR-2 and PR-5 and accumulate high levels of camalexin after pathogen inoculation Plant Cell 11 1999 1393 1404 10449575
19 Wildermuth M.C. Dewdney J. Wu G. Ausubel F.M. Isochorismate synthase is required to synthesise salicylic acid for plant defence Nature 414 2001 562 565 11734859
20 Kerbarh O. Ciulli A. Howard N.I. Abell C. Salicylate biosynthesis: overexpression, purification, and characterization of Irp9, a bifunctional salicylate synthase from Yersinia enterocolitica J. Bacteriol. 187 2005 5061 5066 16030197
21 Rekhter D. Lüdke D. Ding Y. Feussner K. Zienkiewicz K. Lipka V. Isochorismate-derived biosynthesis of the plant stress hormone salicylic acid Science 365 2019 498 502 31371615
22 Torrens-Spence M.P. Bobokalonova A. Carballo V. Glinkerman C.M. Pluskal T. Shen A. PBS3 and EPS1 complete salicylic acid biosynthesis from isochorismate in Arabidopsis Mol. Plant 12 2019 1577 1586 31760159
23 Nawrath C. Heck S. Parinthawong N. Métraux J.P. EDS5, an essential component of salicylic acid-dependent signaling for disease resistance in Arabidopsis, is a member of the MATE transporter family Plant Cell 14 2002 275 286 11826312
24 Serrano M. Wang B. Aryal B. Garcion C. Abou-Mansour E. Heck S. Export of salicylic acid from the chloroplast requires the multidrug and toxin extrusion-like transporter EDS5 Plant Physiol. 162 2013 1815 1821 23757404
25 Noutoshi Y. Okazaki M. Kida T. Nishina Y. Morishita Y. Ogawa T. Novel plant immune-priming compounds identified via high-throughput chemical screening target salicylic acid glucosyltransferases in Arabidopsis Plant Cell 24 2012 3795 3804 22960909
26 Bauer S. Mekonnen D.W. Hartmann M. Yildiz I. Janowski R. Lange B. UGT76B1, a promiscuous hub of small molecule-based immune signaling, glucosylates N-hydroxypipecolic acid, and balances plant immunity Plant Cell 33 2021 714 734 33955482
27 Dean J.V. Delaney S.P. Metabolism of salicylic acid in wild-type, ugt74f1 and ugt74f2 glucosyltransferase mutants of Arabidopsis thaliana Physiol. Plant 132 2008 417 425 18248508
28 Zhang K. Halitschke R. Yin C. Liu C.-J. Gan S.-S. Salicylic acid 3-hydroxylase regulates Arabidopsis leaf longevity by mediating salicylic acid catabolism Proc. Natl. Acad. Sci. U. S. A. 110 2013 14807 14812 23959884
29 Zhang Y. Zhao L. Zhao J. Li Y. Wang J. Guo R. S5H/DMR6 encodes a salicylic acid 5-hydroxylase that fine-tunes salicylic acid homeostasis Plant Physiol. 175 2017 1082 1093 28899963
30 Bartsch M. Bednarek P. Vivancos P.D. Schneider B. von Roepenack-Lahaye E. Foyer C.H. Accumulation of isochorismate-derived 2,3-dihydroxybenzoic 3-O-beta-D-xyloside in Arabidopsis resistance to pathogens and ageing of leaves J. Biol. Chem. 285 2010 25654 25665 20538606
31 Li X. Svedin E. Mo H. Atwell S. Dilkes B.P. Chapple C. Exploiting natural variation of secondary metabolism identifies a gene controlling the glycosylation diversity of dihydroxybenzoic acids in Arabidopsis thaliana Genetics 198 2014 1267 1276 25173843
32 Huang X.X. Zhu G.Q. Liu Q. Chen L. Li Y.J. Hou B.K. Modulation of plant salicylic acid-associated immune responses via glycosylation of dihydroxybenzoic acids Plant Physiol. 176 2018 3103 3119 29483147
33 Chen F. D'Auria J.C. Tholl D. Ross J.R. Gershenzon J. Noel J.P. An Arabidopsis thaliana gene for methylsalicylate biosynthesis, identified by a biochemical genomics approach, has a role in defence Plant J. 36 2003 577 588 14617060
34 Attaran E. Zeier T.E. Griebel T. Zeier J. Methyl salicylate production and jasmonate signaling are not essential for systemic acquired resistance in Arabidopsis Plant Cell 21 2009 954 971 19329558
35 Effmert U. Saschenbrecker S. Ross J. Negre F. Fraser C.M. Noel J.P. Floral benzenoid carboxyl methyltransferases: from in vitro to in planta function Phytochemistry 66 2005 1211 1230 15946712
36 Park S.-W. Kaimoyo E. Kumar D. Mosher S. Klessig D.F. Methyl salicylate is a critical mobile signal for plant systemic acquired resistance Science 318 2007 113 116 17916738
37 Steffan H. Ziegler A. Rapp A. N-Salicyloyl-Asparaginsäure: Eine neue phenolische Verbindung aus Reben Vitis 86 1988 79 86
38 Bourne D.J. Barrow K.D. Milborrow B.V. Salicyloylaspartate as an endogenous component in the leaves of Phaseolus vulgaris Phytochemistry 30 1991 4041 4044
39 Zhang Z. Li Q. Li Z. Staswick P.E. Wang M. Zhu Y. Dual regulation role of GH3.5 in salicylic acid and auxin signaling during Arabidopsis-Pseudomonas syringae interaction Plant Physiol. 145 2007 450 464 17704230
40 Chen Y. Shen H. Wang M. Li Q. He Z. Salicyloyl-aspartate synthesised by the acetyl-amido synthetase GH3.5 is a potential activator of plant immunity in Arabidopsis Acta Biochim. Biophys. Sin. 45 2013 827e836
41 Westfall C.S. Sherp A.M. Zubieta C. Alvarez S. Schraft E. Marcellin R. Arabidopsis thaliana GH3.5 acyl acid amido synthetase mediates metabolic crosstalk in auxin and salicylic acid homeostasis Proc. Natl. Acad. Sci. U. S. A. 113 2016 13917 13922 27849615
42 Mackelprang R. Okrent R.A. Wildermuth M.C. Preference of Arabidopsis thaliana GH3.5 acyl amido synthetase for growth versus defence hormone acyl substrates is dictated by concentration of amino acid substrate aspartate Phytochemistry 143 2017 19 28 28743075
43 Schmelz E.A. Engelberth J. Tumlinson J.H. Block A. Alborn H.T. The use of vapor phase extraction in metabolic profiling of phytohormones and other metabolites Plant J. 39 2004 790 808 15315639
44 Hartmann M. Kim D. Bernsdorff F. Ajami-Rashidi Z. Scholten N. Schreiber S. Biochemical principles and functional aspects of pipecolic acid biosynthesis in plant immunity Plant Physiol. 174 2017 124 153 28330936
45 Nobuta K. Okrent R.A. Stoutemyer M. Rodibaugh N. Kempema L. Wildermuth M.C. The GH3 acyl adenylase family member PBS3 regulates salicylic acid-dependent defence responses in Arabidopsis Plant Physiol. 144 2007 1144 1156 17468220
46 Lawton K. Weymann K. Friedrich L. Vernooij B. Uknes S. Ryals J. Systemic acquired resistance in Arabidopsis requires salicylic acid but not ethylene Mol. Plant Microbe Interact. 8 1995 863 870 8664495
47 Cao H. Glazebrook J. Clarke J.D. Volko S. Dong X. The Arabidopsis NPR1 gene that controls systemic acquired resistance encodes a novel protein containing ankyrin repeats Cell 88 1997 57 63 9019406
48 Gaffney T. Friedrich L. Vernooij B. Negrotto D. Nye G. Uknes S. Requirement of salicylic acid for the induction of systemic acquired resistance Science 261 1993 754 756 17757215
49 Bernsdorff F. Döring A.-C. Gruner K. Schuck S. Bräutigam A. Zeier J. Pipecolic acid orchestrates plant systemic acquired resistance and defence priming via salicylic acid-dependent and -independent pathways Plant Cell 28 2016 102 129 26672068
50 Yildiz I. Gross M. Moser D. Petzsch P. Köhrer K. Zeier J. N-hydroxypipecolic acid induces systemic acquired resistance and transcriptional reprogramming via TGA transcription factors Plant Cell Environ. 46 2023 1900 1920 36790086
51 Hubrich F. Müller M. Andexer J.N. Chorismate- and isochorismate converting enzymes: versatile catalysts acting on an important metabolic node Chem. Commun. 57 2021 2441
52 Blasiak L.C. Clardy J. Discovery of 3-formyl-tyrosine metabolites from Pseudoalteromonas tunicata through heterologous expression J. Am. Chem. Soc. 132 2010 926 927 20041686
53 Bertin C. Weston L.A. Huang T. Jander G. Owens T. Meinwald J. Grass roots chemistry: meta-tyrosine, an herbicidal nonprotein amino acid Proc. Natl. Acad. Sci. U. S. A. 104 2007 16964 16969 17940026
54 Larsen P.O. Meta-carboxy-substituted aromatic amino acids in plant metabolism Biochim. Biophys. Acta 141 1967 27 46 6051583
55 Larsen P.O. Onderka D.K. Floss H.G. Biosynthesis of phenylalanine, tyrosine, 3-(3-carboxyphenyl)alanine and 3-(3-carboxy-4-hydroxyphenyl)alanine in higher plants Biochim. Biophys. Acta 381 1975 397 408 1120151
56 Wu J. Chen Y. Huang Y. Hao B. Dai S. Zhao L. The cytosolic aminotransferase VAS1 coordinates aromatic amino acid biosynthesis and metabolism Sci. Adv. 10 2024 eadk0738
57 Westfall C.S. Xu A. Jez J.M. Structural evolution of differential amino acid effector regulation in plant chorismate mutases J. Biol. Chem. 289 2014 28619 28628 25160622
58 Qian Y. Lynch J.H. Guo L. Rhodes D. Morgan J.A. Dudareva N. Completion of the cytosolic post-chorismate phenylalanine biosynthetic pathway in plants Nat. Commun. 10 2019 15 30604768
59 Zamir L.O. Nikolakakis A. Bonner C.A. Jensen R.A. Evidence for enzymatic formation of isoprephenate from isochorismate Bioorg. Med. Chem. Lett. 3 1993 1441 1446
60 El-Azaz J. de la Torre F. Ávila C. Cánovas F.M. Identification of a small protein domain present in all plant lineages that confers high prephenate dehydratase activity Plant J. 87 2016 215 229 27125254
61 Rippert P. Matringe M. Purification and kinetic analysis of the two recombinant arogenate dehydrogenase isoforms of Arabidopsis thaliana Eur. J. Biochem. 269 2002 4753 4761 12354106
62 Young I.G. Batterham T.J. Gibson F. The isolation, identification and properties of isochorismic acid. an intermediate in the biosynthesis of 2,3-dihydroxybenzoic acid Biochim. Biophys. Acta 26 1969 389 400
63 DeClue M.S. Baldridge K.K. Kast P. Hilvert D. Experimental and computational investigation of the uncatalyzed rearrangement and elimination reactions of isochorismate J. Am. Chem. Soc. 128 2006 2043 2051 16464106
64 Vuralhan Z. Morais M.A. Tai S.L. Piper M.D. Pronk J.T. Identification and characterization of phenylpyruvate decarboxylase genes in Saccharomyces cerevisiae Appl. Environ. Microbiol. 69 2003 4534 4541 12902239
65 Kneen M.M. Stan R. Yep A. Tyler R.P. Saehuan C. McLeish M.J. Characterization of a thiamin diphosphate-dependent phenylpyruvate decarboxylase from Saccharomyces cerevisiae FEBS J. 278 2011 1842 1853 21501384
66 Asakawa T. Wada H. Yamano T. Enzymatic conversion of phenylpyruvate to phenylacetate Biochim. Biophys. Acta 170 1968 375 391 4303395
67 Alché J.D. A concise appraisal of lipid oxidation and lipoxidation in higher plants Redox Biol. 23 2019 101136
68 Zoeller M. Stingl N. Krischke M. Fekete A. Waller F. Berger S. Lipid profiling of the Arabidopsis hypersensitive response reveals specific lipid peroxidation and fragmentation processes: biogenesis of pimelic and azelaic acid Plant Physiol. 160 2012 365 378 22822212
69 Stahl E. Hartmann M. Scholten N. Zeier J. A role for tocopherol biosynthesis in Arabidopsis thaliana basal immunity to bacterial infection Plant Physiol. 181 2019 1008 1028 31515446
70 Hidalgo F.J. Zamora R. Formation of phenylacetic acid and benzaldehyde by degradation of phenylalanine in the presence of lipid hydroperoxides: New routes in the amino acid degradation pathways initiated by lipid oxidation products Food Chem. X 2 2019 100037
71 Debener T. Lehnackers H. Arnold M. Dangl J.L. Identification and molecular mapping of a single Arabidopsis thaliana locus determining resistance to a phytopathogenic Pseudomonas syringae isolate Plant J. 1 1991 289 302 21736648
72 Stahl E. Bellwon P. Huber S. Schlaeppi K. Bernsdorff F. Vallat-Michel A. Regulatory and functional aspects of indolic metabolism in plant systemic acquired resistance Mol. Plant 9 2016 662 681 26802249
73 Nafisi M. Goregaoker S. Botanga C.J. Glawischnig E. Olsen C.E. Halkier B.A. Arabidopsis cytochrome P450 monooxygenase 71A13 catalyzes the conversion of indole-3-acetaldoxime in camalexin synthesis Plant Cell 19 2007 2039 2052 17573535
74 Lorenzen M. Racicot V. Strack D. Chapple C. Sinapic acid ester metabolism in wild type and a sinapoylglucose-accumulating mutant of Arabidopsis Plant Physiol. 112 1996 1625 1630 8972602
75 Li X. Bergelson J. Chapple C. The Arabidopsis accession Pna-10 is a naturally occurring sng1 deletion mutant Mol. Plant 3 2010 91 100 19969522
76 Staswick P.E. Serban B. Rowe M. Tiryaki I. Maldonado M.T. Maldonado M.C. Characterization of an Arabidopsis enzyme family that conjugates amino acids to indole-3-acetic acid Plant Cell 17 2005 616 627 15659623
77 Conrath U. Chen Z. Ricigliano J.R. Klessig D.F. Two inducers of plant defence responses, 2,6-dichloroisonicotinec acid and salicylic acid, inhibit catalase activity in tobacco Proc. Natl. Acad. Sci. U. S. A. 92 1995 7143 7147 11607566
78 Delaney T.P. Friedrich L. Ryals J.A. Arabidopsis signal transduction mutant defective in chemically and biologically induced disease resistance Proc. Natl. Acad. Sci. U. S. A. 92 1995 6602 6606 11607555
79 Zhou Y. Memelink J. Linthorst H.J.M. An E. coli biosensor for screening of cDNA libraries for isochorismate pyruvate lyase-encoding cDNAs Mol. Genet. Genomics 293 2018 1181 1190 29796759
80 Jagadeeswaran G. Raina S. Acharya B.R. Maqbool S.B. Mosher S.L. Appel H.M. Arabidopsis GH3-like defence gene 1 is required for accumulation of salicylic acid, activation of defence responses and resistance to Pseudomonas syringae Plant J. 51 2007 234 246 17521413
81 Lee M.W. Lu H. Jung H.W. Greenberg J.T. A key role for the Arabidopsis WIN3 protein in disease resistance triggered by Pseudomonas syringae that secrete AvrRpt2 Mol. Plant Microbe Interact. 20 2007 1192 1200 17918621
82 Chezem W.R. Memon A. Li F.S. Weng J.K. Clay N.K. SG2-Type R2R3-MYB transcription factor MYB15 controls defense-induced lignification and basal immunity in Arabidopsis Plant Cell 29 2017 1907 1926 28733420
83 Návarová H. Bernsdorff F. Döring A.-C. Zeier J. Pipecolic acid, an endogenous mediator of defence amplification and priming, is a critical regulator of inducible plant immunity Plant Cell 24 2012 5123 5141 23221596
84 Ding P. Rekhter D. Ding Y. Feussner K. Busta L. Haroth S. Characterization of a pipecolic acid biosynthesis pathway required for systemic acquired resistance Plant Cell 28 2016 2603 2615 27758894
85 Sharma S. Shinde S. Verslues P.E. Functional characterization of an ornithine cyclodeaminase-like protein of Arabidopsis thaliana BMC Plant Biol. 13 2013 182 24237637
86 Cecchini N.M. Jung H.W. Engle N.L. Tschaplinski T.J. Greenberg J.T. ALD1 regulates basal immune components and early inducible defence responses in Arabidopsis Mol. Plant Microbe Interact. 28 2015 455 466 25372120
87 Li J. Kristiansen K.A. Hansen B.G. Halkier B.A. Cellular and subcellular localization of flavin-monooxygenases involved in glucosinolate biosynthesis J. Exp. Bot. 62 2010 1337 1346 21078824
88 [preprint] Rekhter D. Mohnike L. Feussner K. Zienkiewicz K. Zhang Y. Feussner I. The plastidial exporter enhanced disease susceptibility 5 is required for the biosynthesis of N-hydroxy pipecolic acid bioRxiv 2019 10.1101/630723
89 Ding P. Ding Y. Stories of salicylic acid: a plant defence hormone Trends Plant Sci. 25 2020 549 565 32407695
90 Liu Y. Sun T. Sun Y. Zhang Y. Radojičić A. Ding Y. Diverse roles of the salicylic acid receptors NPR1 and NPR3/NPR4 in plant immunity Plant Cell 32 2020 4002 4016 33037144
91 Rogers E.E. Ausubel F.M. Arabidopsis enhanced disease susceptibility mutants exhibit enhanced susceptibility to several bacterial pathogens and alterations in PR-1 gene expression Plant Cell 9 1997 305 316 9090877
92 Fan J. Crooks C. Lamb C. High-throughput quantitative luminescence assay of the growth in planta of Pseudomonas syringae chromosomally tagged with Photorhabdus luminescens luxCDABE Plant J. 53 2008 393 399 17971037
93 Gruner K. Zeier T. Aretz C. Zeier J. A critical role for Arabidopsis MILDEW RESISTANCE LOCUS O2 in systemic acquired resistance Plant J. 94 2018 1064 1082 29660188
94 Fujisawa T. Kurosawa M. Katagi T. Uptake and transformation of testicide metabolites by duckweed (Lemna gibba) J. Agric. Food Chem. 54 2006 6286 6293 16910721
95 Huang S.-T. Hsei I. Chen C. Synthesis and anticancer evaluation of bis(benzimidazoles), bis(benzoxazoles), and benzothiazoles Bioorg. Med. Chem. 14 2006 6106 6119 16714116
96 Krell C.M. Seebach D. Preparation of free and of specifically protected oligo[β-malic acids] for enzymatic degradation studies Eur. J. Org. Chem. 2000 2000 1207 1218
97 Lee S.S. Li Z.-H. Lee D.H. Kim D.H. (2R,3S)- and (2S,3R)-2-Benzyl-3,4-epoxybutanoic acid as highly efficient and fast acting pseudomechanism-based inactivators for carboxypeptidase A: design, asymmetric synthesis and inhibitory kinetics J. Chem. Soc. Perkin Trans. 1 1995 2877 2882
