XCP1 is a caspase that proteolyzes Pathogenesis-related protein 1 to produce the cytokine CAPE9 for systemic immunity in Arabidopsis


 Proteolytic activation of cytokines regulates immunity in diverse organisms. In animals, cysteine-dependent aspartate-specific proteases (caspases) play central roles in cytokine maturation. Although the proteolytic production of peptide cytokines is also essential for plant immunity, evidence for a plant caspase is still lacking. In this study, we discovered that proteolysis of a caspase-like substrate motif “CNYD” within Pathogenesis-related protein 1 (AtPR1) in Arabidopsis generates an immunomodulatory cytokine (AtCAPE9). Salicylic acid enhances CNYD-targeted protease activity and the proteolytic release of AtCAPE9 from AtPR1 in Arabidopsis. We show that this process involves a caspase, identified as Xylem cysteine peptidase 1 (XCP1). XCP1 exhibits a calcium-modulated pH-activity profile and a comparable activity to human caspases. XCP1 is required to induce systemic immunity triggered by pathogen-associated molecular patterns. This work reveals XCP1 as the first known plant caspase, which produces the cytokine AtCAPE9 from the canonical salicylic acid signaling marker PR1 to activate systemic immunity.


Introduction
Proteases, which hydrolyze proteins into shorter proteins, peptides or amino acids, are involved not only in protein turnover but also in the regulation of diverse physiological events 1 . For instance, proteases regulate immunity by generating peptides from host or pathogen precursors; these peptides activate and orchestrate defense responses to defeat biological threats 2 . In the animal kingdom, members of the cysteine-dependent aspartate-speci c protease (caspase) family serve as central mediators in the initiation and execution of apoptosis, as well as the activation of in ammation via proteolytic maturation of cytokines 3 . Caspases belong to a cysteine protease family with high speci city; they recognize a motif of at least four amino acids that ends in and cleaves immediately after aspartate (xxxD↓x) 4 . Although plants are expected to have caspases 5 , no sequence homologue of caspases has been directly identi ed 6 , and no plant protease has been found to cleave a precursor after aspartate to produce a mature cytokine directly. Currently, the only known functional analogs of animal caspases in plants are metacaspases (a type of cysteine protease that cleaves lysine/arginine) and phytaspases (a type of serine proteases that cleaves aspartate) 7,8 . Although caspase-like activity has been reported in plants 9 , and plant immune responses can be initiated by the maturation of plant immunomodulatory peptide cytokines 10 , the underlying caspases are not clear.
Among the known immunomodulatory peptides in plants, a tomato wound-induced peptide is produced by cleavage event that is after a caspase-like substrate motif, "CNYD↓", within the proprotein "Pathogenesis-related protein 1" (PR1) 11 . PR1 belongs to the cysteine-rich secretory protein, antigen 5, and pathogenesis-related 1 protein (CAP) superfamily 12 , thus the rst identi ed mature peptide was named CAP-derived peptide 1 (CAPE1) 13 . Tomato CAPE1 (SlCAPE1) induces antipathogen and minor antiherbivore responses, without signi cantly triggering programmed cell death. SlCAPE1, and the CNYD domain positioned N-terminal to SlCAPE1, are highly conserved within PR1 across diverse plant species. PR1 is the most common marker for salicylic acid-regulated plant immunity and its secretion is critical for the activation of systemic acquired resistance (SAR) in Arabidopsis 14 . Nevertheless, how PR1 regulates SAR is poorly understood. In Arabidopsis, CAPE9 (AtCAPE9) is a putative CAPE that is also derived from PR1 (AtPR1), and treatment with synthetic AtCAPE9 induces antipathogen activity in Arabidopsis 13 . However, endogenous AtCAPE9 has not yet been detected, and whether AtPR1 is proteolytically processed to regulate plant systemic immunity is unknown. To uncover a potential role for AtPR1 proteolysis in the activation of SAR, we aimed to identify AtCAPE9 and its corresponding protease.

Results
AtCAPE9 is generated from aspartate-speci c proteolysis of AtPR1 and enhances immunity in Arabidopsis To decipher the function of AtCAPE9 in regulating Arabidopsis immunity, we treated two groups of Arabidopsis plants with water or an aqueous solution of synthetic AtCAPE9 (PRGNYVNEKPY). Compared to plants treated with water, those treated with AtCAPE9 displayed an increased level of salicylic acid (SA) (3.3-fold) and reduced infection upon inoculation with Pseudomonas syringae pv. tomato DC3000 (Pst DC3000) 11 (72.5-fold in this study) (Fig. 1a). We performed LC-MS/MS to identify AtCAPE9 in SA-treated leaves and detected endogenous AtCAPE9, whose MS/MS spectrum closely matched that of synthetic AtCAPE9 (Fig. 1b). Thus, AtCAPE9 is an endogenous peptide cytokine that can regulate SA levels and enhances pathogen resistance.
To investigate AtPR1 processing is depending on the aspartate of CNYD motif that is N-terminal to the AtCAPE9 sequence in AtPR1, we expressed a series of modi ed AtPR1 sequences fused to enhanced yellow uorescent protein (eYFP) in Arabidopsis. Speci cally, we examined native (N, CNYDP) or alaninesubstituted versions (D150A: CNYAP or P151A: CNYDA) of AtPR1-eYFP (Fig. 1c). We detected a ~44.7 kDa protein with all three constructs, corresponding to intact (uncleaved) AtPR1-eYFP. In contrast, the putative AtCAPE9-eYFP cleavage product (~27.0 kDa) was produced from AtPR1-eYFP of the native and P151A mutant, but not from the D150A mutant. These data suggest that the aspartate in the potential caspase substrate domain CNYD is important for AtPR1 cleavage.
SA increases AtCAPE9 level, AtPR1-eYFP cleavage and CNYD-targeted protease (CNYDase) activity in Arabidopsis We hypothesized that AtCAPE9, like AtPR1, is involved in SA-triggered immunity in Arabidopsis. To determine if SA treatment induces the production of AtCAPE9, we used LC-MS/MS to quantify the level of endogenous AtCAPE9 in plants with or without SA treatment. Brie y, LC-MS/MS was operated in selected reaction monitoring (SRM) mode targeting the fragmentation transitions of AtCAPE9. We found that the level of AtCAPE9 was ~6-fold higher in SA-treated plants compared to the untreated controls (Fig. 2a).
Moreover, the levels of AtCAPE9-eYFP produced from AtPR1-eYFP were higher in transgenic plants treated with SA or the SA functional analogue 2,6-dichloroisonicotinic acid (INA) compared to mocktreated controls (Fig. 2b).
To monitor CNYDase activity, we incubated a uorogenic protease substrate Ac-CNYD-AMC ( Fig. S1) with plant extract. Substrate cleavage was elevated in extracts from SA-treated and INA-treated plants, compared to extracts from mock-treated or wounded plants (Fig. 2c), suggesting that CNYDase activity increases upon SA and INA treatment. Together, these data suggest that SA treatment enhances CNYDase activity to generate increased level of AtCAPE9 from AtPR1 in Arabidopsis.
Identi cation of a putative caspase targeting the CNYD motif in Arabidopsis Next, we used the CNYDase assay to further examine whether the conserved CNYD motif in AtPR1 is essential for proteolytic release of AtCAPE9. Indeed, we observed substantially reduced cleavage of uorogenic substrates with a disrupted CNYD motif (Ac-CNAD-AMC, Ac-ANAD-AMC) compared to the canonical Ac-CNYD-AMC substrate (Fig. 3a). In addition, cleavage of the Ac-CNYD-AMC substrate was signi cantly enhanced by CaCl 2 , but inhibited by ZnCl 2 and the metal chelator EDTA, as compared to the control ( Fig. 3b; left panel). The Ca 2+ -enhanced CNYDase activity was strongly suppressed by a general cysteine protease inhibitor (E-64) or a biotinylated aldehyde tetrapeptide CNYD (biotin-CNYD-CHO; Fig. S1, Fig. 3b; right panel). The suppression by E-64 suggested that the CNYDase is a cysteine protease that catalyzes a cysteine-dependent proteolytic reaction.
The biotin-CNYD-CHO inhibitor was designed to probe and covalently modify the cysteine in the active site of CNYDases, and displayed dose-dependent inhibition of CNYDase activity with the Ac-CNYD-AMC substrate ( Fig. 3c; left panel). We incubated biotin-CNYD-CHO with wild-type (WT) Arabidopsis extract, and detected biotin-CNYD-CHO-labelled proteins with streptavidin-HRP by western blotting. Intriguingly, we observed prominent labeling of a potential 35 kDa CNYDase and two minor protein bands at 64 and 24 kDa ( Fig. 3c; right panel). Together, our results so far point to a protease that is mainly 35 kDa, Ca 2+activated cysteine-dependent aspartate-speci c CNYDase (caspase) in Arabidopsis.
XCP1 is a CNYDase enzyme speci c for CAPE production from AtPR1 To discover the enzyme speci c for CAPE production (designated ESCAPE), we investigated two cysteine protease families in Arabidopsis, the papain-like cysteine proteases (PLCPs) and metacaspases (MCs), which have caspase-like functions in regulating plant immunity and programmed cell death 15,16 . The MC family members are not aspartate-speci c and therefore were excluded as candidates. Among the 31 Arabidopsis PLCPs, we identi ed ESCAPE candidates that had: (1) a molecular weight of 34.00-37.00 kDa (untruncated, pro-form) and 23.00-25.00 kDa (truncated, processed form) (Table S1), and (2) an expression pattern that correlates with AtPR1 expression during leaf development 17 (Fig. S2). The only PLCP that t both criteria was Xylem cysteine peptidase 1 (XCP1; At4g35350), which is a member of PLCP subfamily III (Fig. S3a). In addition, the ~35 kDa untruncated form of XCP1 is signi cantly more abundant than the ~23 kDa truncated form in 35S:XCP1 plants 18 .
To examine the potential CNYDase activity of XCP1, we used the xcp1 T-DNA insertion mutant (SALK_084789), which we con rmed by genotyping and gene expression ( Fig. S3b-d). We observed signi cantly lower CNYDase activity in lysate from the xcp1 mutant compared to from WT and other PLCP mutants (Fig. S4). Moreover, biotin-CNYD-CHO did not label any proteins in xcp1 mutant extract, unlike WT (Fig. 4a). Together these data suggest that XCP1 directly recognizes the CNYD motif and catalyzes a proteolytic reaction on the aspartate.
To further investigate the activity of XCP1, we expressed and a nity puri ed an XCP1-His fusion protein from tobacco leaves. The expected molecular weights of the untruncated and truncated XCP1-His proteins are 37.6 kDa and 24.6 kDa, respectively ( Figure S5a), and we detected puri ed His-tagged proteins of 45 kDa, 32-37 kDa and 25 kDa by immunoblotting ( Figure S5b). The 45 kDa protein has been suggested to be a preprotein form of XCP1, which can be observed when the protein is overexpressed 18 .
Among the different forms of the XCP1-His protein that were expressed and puri ed from tobacco, the 34-36 kDa and 25 kDa forms could be labeled by biotin-CNYD-CHO ( Figure S5b), suggesting that these forms are active CNYDases. Puri ed XCP1-His displayed proteolytic activity with the Ac-CNYD-AMC substrate, but not with the Ac-CNAD-AMC nor Ac-ANAD-AMC substrates, and this activity was e ciently inhibited by co-incubation with biotin-CNYD-CHO (Fig. 4b). Moreover, the CNYDase activity of puri ed XCP1-His was enhanced by CaCl 2 and inhibited by ZnCl 2 and E-64, and slightly reduced by EDTA as compared to the control (Fig. 4c), similar to our observations with WT Arabidopsis lysate. We found that puri ed XCP1-His exhibited the highest CNYDase activity at pH 6.0 in the presence of excess CaCl 2 , and at pH 5.0 in the absence of excess CaCl 2 (Fig. 4d). In addition, puri ed XCP1-His (1 µg) was substantially more active at 22 °C as compared to 32 °C and 37 °C (Fig. S6). At 22 °C, puri ed XCP1-His (0.2 µg) cleaved Ac-CNYD-AMC with a K m of ~26.5 µM and V max of ~251.2 RFU/min (Fig. 4e). To determine whether the cysteine (C161) in the putative active site of XCP1 is important for CNYDase activity, we expressed and puri ed an alanine-substituted (C161A) XCP1-His mutant. Indeed, the CNYDase activity of (C161A) XCP1-His mutant was dramatically lower than native XCP1-His (Fig. S7). Together, these data suggest that XCP1 is a caspase in Arabidopsis.
Importantly, we observed increased production of AtCAPE9-eYFP (~27.0 kDa) from AtPR1-eYFP in WT Arabidopsis lysate supplemented with Ni-NTA-puri ed proteins from tobacco expressing P19 plus XCP1-His versus P19 alone (Fig. 4f). To determine if XCP1-His can interact with and process AtPR1-eYFP directly, we immobilized AtPR1-eYFP on Protein A/G beads with an anti-PR1 antibody, followed by incubation Ni-NTA-puri ed proteins from tobacco without or with expressing XCP1-His. Incubation with puri ed XCP1-His increased the release of AtCAPE9-eYFP from immobilized AtPR1-eYFP compared to the control (Fig. 4g). These data suggest that XCP1 directly processes AtPR1 into AtCAPE9. Therefore, we suggest a functional name for XCP1 as ESCAPE.

XCP1/ESCAPE regulates local and systemic immunity in Arabidopsis
Our data so far suggest that XCP1 is ESCAPE that can generate the immune elicitor AtCAPE9 from AtPR1, so we further investigated how this enzyme regulates CNYDase activity and disease resistance of Arabidopsis. Brie y, we used homozygous wild-type XCP1/ESCAPE (ESCAPE w/w ), and mutated XCP1/ESCAPE (ESCAPE m/m ) and heterozygous mutated XCP1/ESCAPE (ESCAPE w/m ) plants from F2 generation of WT and xcp1 (escape) crossed plant (WT x escape) lines. We found that the reduction of the CNYDase activity and pathogen resistance in ESCAPE m/m lines as compared to ESCAPE w/w and ESCAPE w/m lines, and the pathogen infection of ESCAPE m/m lines was almost as severe as loss of AtPR1 in the plants (atpr1 mutant) (Fig. 5a, b).
To show that ESCAPE is involved in g22-triggered systemic immunity, we treated plants locally with g22 and monitored CNYDase activity and disease resistance in untreated leaves. Speci cally, we employed ESCAPE w/w , and ESCAPE m/m F2 lines from WT x escape and a complement line obtained by expressing ESCAPE-His in the escape mutant (escape/ESCAPE-His). We locally in ltrated g22, and observed reduced CNYDase activity and diminished resistance to Pst DC3000 in the untreated leaves of the ESCAPE m/m line compared to both ESCAPE w/w and escape/ESCAPE-His lines (Fig. 5d).

Discussion
In this study, we identi ed endogenous AtCAPE9 and showed its function in activating SA biosynthesis and an immune response. We show for the rst time that AtCAPE9 is derived from the cleavage of the putative caspase substrate domain CNYD in AtPR1, in SA-treated Arabidopsis leaves. The aspartate residue (D150) of the CNYD domain that is immediately N-terminal to the AtCAPE9 domain is crucial for the cleavage of AtPR1-eYFP to generate AtCAPE9-eYFP in plants, indicating that AtCAPE9 production requires an aspartate-speci c protease. We further demonstrated that SA treatment enhanced the endogenous production of AtCAPE9, the proteolytic cleavage of AtPR1-eYFP to produce AtCAPE9-eYFP, and CNYDase activity in plants. We found that CNYDase activity and thus AtCAPE9 production were not induced by wounding in Arabidopsis. However, our observation that SA both induces and is induced by AtCAPE9 suggests the existence of a positive-feedback loop to amplify SA-mediated defense responses in plants.
We found that cleavage of CNYD substrates is mediated by recognition of the CNYx motif. CNYDase activity in plant extract was speci cally suppressed by the cysteine protease inhibitor E-64 and the substrate-speci c inhibitor biotin-CNYD-CHO, but not by the serine protease inhibitor PMSF, consistent with a caspase activity. We successfully identi ed an Arabidopsis PLCP member XCP1 as enzyme speci c for CAPE production (ESCAPE). XCP1 was identi ed as ESCAPE based on our discovery of the sizes of CNYDase and its overlapping gene expression pattern with AtPR1 in plant leaves. XCP1/ESCAPE and its paralog XCP2 have been reported to aid in the formation of tracheary elements (TEs) prior to the macro-autolysis followed by vacuole and protoplast disruption 19 . These two proteases are translocated to the vacuole during the formation of TEs and released to the xylem after the breakdown of the protoplast. The XCPs loss-of-function plants xcp1xcp2 show a delay of micro-autolysis, but do not exhibit a noticeable change in growth and developmental phenotypes, including TEs, as compared to wild-type plants. Although XCPs are not essential in growth and development, XCP1/ESCAPE has been shown to participate in basal defense functions 20 . The XCP1/ESCAPE protein and its cysteine protease activity is also a target of the pathogen effector Avr2 from Cladosporium fulvum, to suppress plant resistance.
Together with the defense function and identi cation of XCP1/ESCAPE in maize xylem sap 21 , it was suggested that XCP1/ESCAPE serves defense functions that are expressed and propagated in plant xylem 22 .
Here, we demonstrate that XCP1/ESCAPE plays a major role in the CNYDase activity in plants, given that protein extracts from xcp1 (escape) mutants were not tagged with biotin-CNYD-CHO and showed a signi cant loss of CNYDase activity. Puri ed XCP1-His (ESCAPE-His) was highly speci c for the CNYD domain and its CNYDase activity was enhanced by Ca 2+ but repressed by Zn 2+ and E-64. Importantly, we provide strong evidence that puri ed ESCAPE-His directly cleaves AtPR1-eYFP to release AtCAPE9-eYFP.
XCP1/ESCAPE activity was temperature sensitive, displaying optimal CNYDase activity at 22 °C, which is also optimal for growth of Arabidopsis. is thought to be crucial for SAR 14 and XCP1/ESCAPE has been suggested to be an apoplastic and vacuolar localized protein 19 , our data suggest that PR1 is processed by XCP1/ESCAPE in the apoplast to induce SAR. The cysteine protease activity of XCP1/ESCAPE can be targeted by the effector Avr2, which may disrupt its CNYDase activity and, in turn, the production of AtCAPE9 and the induction of SA. The temperature sensitivity of XCP1/ESCAPE activity might help to explain why pathogen-induced SA-defense responses are temperature-vulnerable 24 ; lower XCP1/ESCAPE activity at higher temperatures is expected to yield less AtCAPE9 and thus diminished SA biosynthesis. Compared to the activation and secretion of immunomodulatory cytokines triggered by caspases in animals, plant PR1 acts like a pro-cytokine that is activated by XCP1/ESCAPE to produce the SAR signal AtCAPE9. As PR1 is conserved in humans 25 , these ndings might stimulate the search for a similar mechanism in mediating human immunity.

Methods
Plant materials and growth conditions  Table S2. All mutants and complement lines were genotyped for zygosity using allele-speci c primers shown in Table   S3.

DNA cloning and transformation
The cDNA sequences of Arabidopsis were produced via reverse transcription and the DNA fragments of target genes AtPR1 and XCP1 were ampli ed by the speci c primer using Perfectread Pfu polymerase and the nal product of XCP1 (ESCAPE) was added with the C-terminal 6x Histidine tag (HisTag) for puri cation purposes. The DNA products were puri ed using Gel/PCR DNA fragment extraction kit. The primers used for cloning are listed in Table S3.
To generate the clones carrying the sequences of interest, the ampli ed DNA fragment was cloned into the pCR8 entry vector using pCR™8/GW/TOPO™ TA Cloning Kit (Invitrogen). The pCR8-XCP1-His clone was transferred into a gateway vector pMDC32 carrying a dual 35S promoter. The native and two alaninesubstituted (N, D150A and P151A) AtPR1 clones were transferred into pK7YWG2 carrying a single 35S promoter, using a method outlined in our previous study 26 . The C161A mutated XCP1-His clone was obtained by cloning of pMDC32-XCP1-His using a pair of point mutation primers. The mutated XCP1-His clone was inserted into the PCR8 entry clone and then transfer to the pMDC32.
To observe the proteolytic processing of AtPR1, the N, D150A, and P151A AtPR1-eYFP clones were separately transformed to WT Arabidopsis (Col-0) plants to produce a stable transgenic line (T3 generation lines were used in this study). To express the XCP1-His protein, the 2x35S::XCP1-His was transformed into tobacco plants for transient expression and the proteins extracted from tobacco leaves were puri ed by Ni-NTA column. The ESCAPE-His complemented escapes mutant (escape/ESCAPE-His) lines were obtained from the T3 generation of the xcp1 (escape) mutant transformed with native promoter driven XCP1-His (ESCAPE-His). The expression of ESCAPE in the WT, escape and escape/ESCAPE-His plants was quanti ed by RT-PCR (Fig. S8b). All the binary vectors were transformed into Agrobacterium tumefaciens strain GV3101 using electroporation 27 . Arabidopsis thaliana (Col-0) and Nicotiana benthamiana plants were transformed by the oral dipping method and agroin ltration using Agrobacterium tumefaciens-mediated transformation system, respectively 28 .
Peptide or phytohormone treatment The bacterial populations were calculated from leaf discs after 5 days of inoculation, represented as Log CFU per leaf disc (cm 2 ).

Endogenous peptide isolation and detection
Nine-week-old Arabidopsis plants treated with or without 1 mM SA with 0.0015% Silwet L-77 solution were collected and individually ground into powder under liquid nitrogen using a homogenizer (Nissei ACE Homogenizer AM-5). Frozen leaf powder (50 g) was dissolved in 150 ml of 1% TFA and homogenized to leaf juice using a blender for 2 minutes. The leaf juice was ltered through four layers of cheesecloth and one layer of Miracloth (Calbiochem). The ltrated leaf juice was then centrifuged at 10,000 × g for 20 minutes at 4 °C. The supernatant was adjusted to pH 4.5 with 10 N NaOH and centrifuged at 10,000 × g for 20 minutes at 4 °C. Then the supernatant was re-adjusted to pH 2.5 using TFA and 50 μg tryptic β-casein peptides were added to the supernatant as an internal control for peptide abundance normalization. To avoid the trypsin residue reacting with the endogenous proteins or peptides, the tryptic β-casein peptides were acidi ed by TFA and puri ed using C18 Sep-pak (20 mg) cartridge (Waters). To purify the supernatant, the customized C18 Sep-pak (20 g) cartridge was used and rst equilibrated by 60 ml 0.1% TFA. The supernatant was loaded into the Sep-Pak cartridge, washed with 100 ml 0.1% TFA and eluted by 150 ml of 60% methanol in 0.1% TFA. The eluted solution was vacuumevaporated to remove methanol using a vacuum centrifugation concentrator (miVac Duo Concentrator, Genevac) to dryness 13 . The dried crude extract was dissolved in 1 ml of 0.1% TFA, centrifuged at 10,000 × g for 10 minutes at 4 °C and ltered through a 0.45 μm lter (Millipore) before peptide fractionation. The ltrated peptide extract was injected into a Superdex Peptide 10/300 column (GE Healthcare) and eluted by 0.5 ml/min of 0.1% TFA with 1 fraction/min for collecting the peptide fractions and evaporated to dryness by a vacuum centrifugal concentrator. Each fraction was puri ed by C 18 ZipTip (Merck Millipore) for LC-MS/MS analysis. To identify the AtCAPE9 from endogenous peptides, the LC-MS/MS operated in data-dependent acquisition (DDA) mode was used. To quantify the AtCAPE9 abundance in Arabidopsis, the LC-MS/MS operated in selected reaction monitoring (SRM) mode targeting on the fragmentation transitions of AtCAPE9 was applied.

Phytohormone Extraction
After peptide treatment, the metabolites were extracted from leaf tissues for phytohormone quantitation.
The extraction procedure was modi ed from a previously published protocol 29 . The leaf tissues (about 0.4 g fresh weight) were ground into a powder under liquid nitrogen and transferred to a 50 ml screw-cap tube. The frozen leaf powder was dissolved in 4 ml extraction solvent and d 6 -SA (2 ng to 0.4 g leaf tissue) added as internal standards. The samples were extracted by rotating at a speed of 100 rpm at 4 °C for 30 min and then 8 ml dichloromethane was added to each sample and shaken at 100 rpm at 4 °C for 30 min. The samples were centrifuged at 13,000 × g at 4 °C for 5 min, and two phases were formed. The lower phase was transferred carefully into a new tube and evaporated to dryness by a vacuum centrifugal concentrator. The dried samples were dissolved in 300 μl methanol, mixed well and centrifuged at 10,000 × g at 4 °C for 5 min and then the supernatant was transferred to the sample vial for targeted quantitation analysis using LC-MS/MS.

Targeted Peptide and Phytohormone Quantitation using LC-MS/MS
For targeted peptide quantitation, a linear ion trap-orbitrap mass spectrometer (Orbitrap Elite, Thermo Fisher Scienti c) coupled online with a nanoUHPLC system (nanoACQUITY UPLC, Waters) was used. The nanoUHPLC method was followed by our previous study 13,30 . The mass spectrometer was operated in the positive ion mode and set to one full FT-MS scan (m/z 400-1600) with 60,000 resolution and switched to one ion trap analysis in selected reaction monitoring (SRM) mode. For SRM targeted on AtCAPE9, the doubly charged AtCAPE9 precursor ion (m/z 668.84) was selected for fragmentation and product ions m/z of 1058.52, 529.76 and 930.44 were monitored. The relative abundances of AtCAPE9 in untreated and SA-treated samples were estimated by combining SRM peak areas of product ions. To quantify the abundance of AtCAPE9, one doubly charged tryptic β-casein peptide (m/z 1031.42) was selected for fragmentation and product ions m/z of 1105.44, 1361.61 and 747.34 were monitored. The normalized abundance of AtCAPE9 was calculated by the peak area of fragment ions and normalized with the abundance of a selected tryptic β-casein peptide.
For phytohormone quantitation, a linear ion trap-orbitrap mass spectrometer (Orbitrap Elite, Thermo Fisher Scienti c) coupled online with a UHPLC system (ACQUITY UPLC, Waters) was used. The phytohormones were separated by an HSS T3 column (Waters) using gradients of 0.5-25% ACN at 0-2 min, 25-75% ACN at 2-7 minutes and 75-9.5% ACN at 7-7.5 minutes. The mass spectrometer was operated in the negative ion mode and set to one full FT-MS scan (m/z 100-600) with 60,000 resolution and switched to two FT-MS product ion scans (in 30,000 resolution) for two precursors: m/z of 137.02 (for SA), 141.05 (for d 6  Assay for protease activity To test the protease activity against different substrates, metals ion, and inhibitors, the substrate reaction buffer 50 mM MOPS with 0.1% CHAPS for pH 6.0 was used. To investigate the speci city of the protease, three synthetic peptides (CNYD, CNAD, and ANAD) tagged by 7-amino-4-methylcoumarin (AMC) at the Cterminus were purchased from Mission Biotech. These substrates were incubated with Arabidopsis protein extract or the puri ed XCP1-His protein in pH 6.0 reaction buffer, then the protease activity was Immunoblotting and immunoprecipitation for AtPR1-eYFP processing For the analysis of AtPR1-eYFP processing, the samples of Arabidopsis WT, D150A, and P151A of AtPR1-eYFP transgenic plants were frozen in liquid nitrogen and ground into ne powder, then dissolved in the plant extraction buffer (50 mM sodium acetate with 200 mM NaCl and 3 mM DTT at pH 5.0) rotating 50 rpm for 30 min at 4 °C. Protein extract was centrifuged 15 minutes at 16,000 x g at 4 °C and the supernatant was ltered by 70 µM Nylon cell strainer (Falcon). The protein concentration was measured and taking 50 µg proteins dissolved by a loading buffer with 5x sample buffer (250 mM Tris at pH 6.8 with 50% glycerol, 5% SDS, and 0.02% bromophenol blue) and β-mercaptoethanol for heating at 95 °C for 5 min. The proteins were separated by 12.5% SDS-PAGE gels (1.0 mm) and transferred onto PVDF membranes by wet western blot. Membranes were rst blocked with a 5% milk in Tris buffer saline-Tween 20 (TBST) buffer and then incubated with anti-GFP antibody (mouse, 1:1000; #11814460001, Roche) for 4 °C overnight. For detecting the recombinant proteins tagged with C-terminal 6XHis, the proteins in the transferred membrane were incubated with anti-His antibody (mouse, 1:5000; #PPT-66005-1, Biotools) at room temperature for 1 h. Protein complexes were labeled with the secondary antibody (mouse, 1:5000; # 61-6520, Thermo Fisher). For measuring the biotin-CNYD-CHO-bounded proteins, the western blot was performed by streptavidin-HRP (1:5000; #016-030-084, Jackson ImmunoResearch). All HRP-conjugated proteins were detected on the membrane with enhanced chemiluminescence (ECL) reagent kit (#193508, Biotools) reacting to the HRPs.
For the direct interaction of AtPR1-eYFP and XCP1-His, the immunoprecipitation of AtPR1-eYFP was performed by anti-AtPR1 antibody (Rabbit, #AS10687, Agrisera) immobilized on the Pierce™ Protein A/G Magnetic Beads (#88803, Thermo Fisher) and the immobilized AtPR1-eYFP proteins were incubated with the Ni-NTA puri ed proteins from the tobacco overexpressed P19 or P19 plus XCP1-His gene in pH 6.0 reaction buffer for 1 h. The elution of the intact AtPR1-eYFP and the proteolytic fragment of AtCAPE9-eYFP using IgG Elution Buffer, pH 2.0 (#21028, Thermo Fisher) from the Protein A/G Magnetic Beads was detected by anti-GFP antibody (mouse, 1:1000; #11814460001, Roche).   Values are means ± SD of biological replicates (n=3). Each replicate was obtained from pooled tissues of three plants. P values were calculated by unpaired t-test (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Biotinylated proteins were detected by western blotting with streptavidin-HRP and Coomassie Blue staining shows total protein loaded (right). Characterization of the enzyme speci c for CAPE production (ESCAPE) by CNYDase activity and interaction with AtPR1-eYFP. (a) Western blot of biotinylated proteins from Arabidopsis wild-type (WT) and xcp1 extracts with or without addition of biotin-CNYD-CHO. The biotinylated proteins were detected with streptavidin-HRP. (b) Proteolytic activity of puri ed XCP1-His on three uorogenic protease substrates (CNYD, CNAD, and ANAD) with or without adding biotin-CNYD-CHO. Proteolytic activity of puri ed XCP1-His was measured by RFU of the cleaved uorophore after 10 h incubation with the substrate. (c) CNYDase activity of puri ed XCP1-His supplemented with ZnCl2, MgCl2, CaCl2, E-64, PMSF, or EDTA for 1 h before substrate incubation. RFU of the cleaved uorophore was measured after 10 h