A novel protein elicitor (PeSy1) from Saccharothrix yanglingensis induces plant resistance and interacts with a receptor‐like cytoplasmic kinase in Nicotiana benthamiana

Abstract Previously, we reported a rare actinomycete Saccharothrix yanglingensis Hhs.015 with strong biocontrol ability, which can colonize plant tissues and induce resistance, but the key elicitor and immune mechanisms were unclear. In this study, a novel protein elicitor screened from the genome of Hhs.015, PeSy1 (protein elicitor of S. yanglingensis 1), could induce a strong hypersensitive response (HR) and resistance in plants. The PeSy1 gene encodes an 11 kDa protein with 109 amino acids that is conserved in Saccharothrix species. PeSy1‐His recombinant protein induced early defence events such as a cellular reactive oxygen species burst, callose deposition, and the activation of defence hormone signalling pathways, which enhanced Nicotiana benthamiana resistance to Sclerotinia sclerotiorum and Phytophthora capsici, and Solanum lycopersicum resistance to Pseudomonas syringae pv. tomato DC3000. Through pull‐down and mass spectrometry, candidate proteins that interacted with PeSy1 were obtained from N. benthamiana. We confirmed the interaction between receptor‐like cytoplasmic kinase RSy1 (Response to PeSy1) and PeSy1 using co‐immunoprecipitation, bimolecular fluorescence complementation, and microscale thermophoresis. PeSy1 treatment promoted up‐regulation of marker genes in pattern‐triggered immunity. The cell death it elicited was dependent on the co‐receptors NbBAK1 and NbSOBIR1, suggesting that PeSy1 acts as a microbe‐associated molecular pattern from Hhs.015. Additionally, RSy1 positively regulated PeSy1‐induced plants resistant to S. sclerotiorum. In conclusion, our results demonstrated a novel receptor‐like cytoplasmic kinase in the plant perception of microbe‐associated molecular patterns, and the potential of PeSy1 in induced resistance provided a new strategy for biological control of actinomycetes in agricultural diseases.


| INTRODUC TI ON
As the most important producer of the ecosystem, plants have long lived in a complex environment, including a wide range of pathogenic microorganisms, herbivorous insects, and beneficial microorganisms. Some plant pathogens, such as Sclerotinia sclerotiorum, Phytophthora capsici, and Pseudomonas syringae pv. tomato (Pst), cause significant economic losses in crop production worldwide (Chen et al., 2020;Granke et al., 2012;Shuang et al., 2006). It is undeniable that fungicides and insecticides are widely used to control diseases and pests to improve crop productivity, but their abuse has caused severe pollution to the environment and other impacts in some cases (Ons et al., 2020;Zaker, 2016). For sustainable agriculture, plant immunity-based elicitors that induce plant resistance represent a fundamental approach for disease control, due to their environmental friendliness and low cost (Pieterse et al., 2014;Thakur & Sohal, 2013).
To sense and defend against potential pathogens, plants have evolved two different types of immune receptors (Jones & Dangl, 2006). The plasma membrane-localized pattern recognition receptors (PRRs) recognize the conserved motifs of pathogen-associated molecular patterns (PAMPs) or microbe-associated molecular patterns (MAMPs), thereby providing PAMP-triggered immunity (PTI). Intracellular resistance proteins, usually nucleotide-binding leucinerich repeat-containing proteins with specific recognition, monitor the presence of pathogen effector proteins resulting in effector-triggered immunity (ETI) (Bigeard et al., 2015). Recent studies have shown that the boundaries between PTI and ETI are not distinct, and their immune signals overlap and are shared (Ngou et al., 2021;Yuan et al., 2021), such as reactive oxygen species (ROS) outbreaks, callose deposition, defence hormone synthesis, and defence-related gene expression (Tsuda & Katagiri, 2010;Yu et al., 2017). ROS, including hydrogen peroxide (H 2 O 2 ) and superoxide anion (O ⋅− 2 ), have dual roles in plant biology (Miller et al., 2010). High concentrations of ROS can lead to the hypersensitive response (HR), while a basal level of ROS is closely related to the regulation of plant growth and hormone signalling (Mittler, 2017;Waszczak et al., 2018). The ROS burst also involves Ca 2+ and NO signalling, which play a crucial role in plant defence responses, inducing callose deposition on the cell wall to slow down pathogen infection (Nishimura et al., 2003;Shetty et al., 2008).
Furthermore, the precise regulation by three defence hormones, salicylic acid (SA), jasmonic acid (JA) and ethylene (ET), plays an important role in the plant fight against systemic spread of pathogens (Spoel & Dong, 2008;Vlot et al., 2021).
Elicitors have steadily emerged as a potent alternative to fungicides and insecticides in green agriculture. Numerous elicitors, such as proteins, sugars, lipids, and peptides, have been isolated and characterized from various pathogens and biocontrol microorganisms (Abdul Malik et al., 2020). PeBL1, the protein elicitor from Brevibacillus laterosporus A60, induces a typical HR in tobacco, enhancing plant resistance to tobacco mosaic virus and P. syringae pv.
tabaci . The protein elicitor OPEL of Phytophthora parasitica induces early cellular defence responses and up-regulation of PTI marker genes and SA pathway genes, and its glycosyl hydrolase (GH) domain is necessary to elicitor activity (Chang et al., 2015).
The S. sclerotiorum protein elicitor SsCut causes cell death in various plants and improves plant disease resistance by inducing the accumulation of secondary metabolites (Zhang et al., 2014). Despite the diversity of known elicitors, our understanding of how elicitors trigger plant defence responses is still limited (Pršić & Ongena, 2020).
Furthermore, BIK1 mediates several MAMP-triggered responses through interaction with PRRs such as EFR, PEPRs, and CERK1 (Lal et al., 2018;Liu et al., 2013;Zhang et al., 2010). The characterization of novel MAMPs and their corresponding targets not only contributes to the understanding of host-microbe co-evolution, but also provides a new resource for molecular disease resistance engineering (Abdul Malik et al., 2020;Chisholm et al., 2006).
Actinomycetes are a unique group within the bacterial domain, capable of producing a variety of bioactive substances that have important practical and economic uses, but little research has been done on elicitors. Saccharothrix yanglingensis Hhs.015 is an actinomycete isolated from cucumber roots that can effectively control plant pathogens through competition, secretion of antagonistic factors, and induced resistance Lu et al., 2018;Wang et al., 2019). Hhs.015 has shown great efficacy against apple Valsa canker, Sclerotinia diseases of rapeseed, and tomato leaf mould in field tests Li et al., 2016;Min et al., 2007).
Remarkably, Hhs.015 can colonize apple tissue culture seedlings and promote the related defence enzymes activity Yan et al., 2019). Protein elicitors have been screened using the whole-genome data of Hhs.015 . The protein elicitor BAR11 can induce a defence response in Arabidopsis thaliana, and molecular experiments showed that it interacts with catalase in plant cells (Zhang et al., 2018). However, the molecular mechanism of the plant immune response induced by Hhs.015 protein elicitors remains unclear.
Given the ability of Hhs.015 to induce plant resistance, we speculated that there would be a corresponding MAMP recognition mechanism in plants; therefore, we conducted an in-depth analy-

| Transient expression of protein elicitor induces cell death in N. benthamiana
As a form of programmed cell death (PCD), the HR is an important immune component during plant-pathogen interactions (Heath, 2000;Thakur & Sohal, 2013). To search for protein elicitors that induce cell death and are potentially recognized by nonhost plants, 32 genes of Hhs.015 predicted to encode secreted proteins (Table S1) were introduced into N. benthamiana by agro-infiltration.
Our results identified five predicted secreted proteins capable of inducing cell death in leaf cells ( Figure S1). Transient expression of the secreted protein gene Hhs.015_GM7245 in N. benthamiana induced intense cell death after 4 days, and necrotic plaques were clearly observed by trypan blue staining (Figure 1a), therefore we named this protein PeSy1. Previous studies have shown that the degree of cell death is positively correlated with ion leakage (Yu et al., 2012).
By measuring the electrolyte permeability of N. benthamiana transiently expressing PeSy1, it was found that PeSy1 was significantly different from the negative control green fluorescent protein (GFP), but not the positive control Bax (Figure 1b).
To further verify the location where PeSy1 works, PeSy1 without its signal peptide (SP) (PeSy1 ΔSP ) was transiently expressed in N. benthamiana. The results showed that PeSy1 ΔSP also induced cell death at 5 days post-agroinfiltration (dpa), which was slower than full-length PeSy1 (Figure 1c). N. benthamiana expressing full-length PeSy began to show cell death symptoms at 3 dpa. These results suggest that PeSy1 may play a role in both intracellular and extracellular processes.

| Characterization of PeSy1
PeSy1 is a hypothetical protein with no significant structural match other than the SP by SMART analysis (Figure 2a (e) Cell death response in N. benthamiana expressing PeSy1 and PeSy1 ΔSP . PVX-PeSy1 and PVX-PeSy1 ΔSP were transiently expressed in N. benthamiana by agroinfiltration. Photographs were taken 3 and 5 dpa. Bax was used a positive control. Green fluorescent protein (GFP) and buffer were used as a negative control. Mean and SE were calculated from three biological replicates. The statistical analyses were performed with Student's t test. Bars indicate ± SE. NS, no significant difference; ***p < 0.001. Bars are 1 cm. SP, signal peptide.
its isoelectric point is 6.70. Twenty sequences with high homology to PeSy1 were obtained from NCBI to construct the phylogenetic tree ( Figure 2b). These proteins belonged to hypothetical proteins of rare actinomycetes, which included Saccharothrix sp., Nonomuraea sp., Prediction Server showed that the structural confidence of the most matched structure of PeSy1 was 50.4%, which only had nine β-folds ( Figure 2e). These findings indicate that PeSy1 was a novel protein from S. yanglingensis.

| HR induced by recombinant protein PeSy1-His in N. benthamiana
The recombinant vector pET28a:PeSy1-His was expressed in

| PeSy1 triggers immunity responses in N. benthamiana
It is already known that HR is accompanied by plant early defence responses, such as ROS burst, callose deposition, and alterations in defence hormone pathways (Han & Hwang, 2017;Schwessinger & Ronald, 2012). To determine whether PeSy1-induced HR is associated with plant immunity responses, 5 μM PeSy1-His was injected into N. benthamiana leaves. Apparent reddish-brown ROS accumulation was observed using 3,3′-diaminobezidine (DAB) staining and pale green fluorescence of callose deposition was observed under UV light excitation after aniline blue staining in leaf cells (Figure 4a,b). As shown in Figure 4c, the relative expression of some genes related to the defence response, such as NbPR1, NbPR2, NbPR4 and NbERF1, after PeSy1 treatment significantly increased compared with the control. NbPR1 and NbPR2 are marker genes for SA-dependent immunity, and NbPR4 and NbERF1 are marker genes for JA/ET-dependent immunity. Thus, PeSy1 may function as a protein elicitor and induce plant defence responses mediated by SAand JA/ET-dependent signalling pathways.

| PeSy1 enhances plant resistance to pathogens
N. benthamiana leaves treated with 5 μM PeSy1-His or phosphatebuffered saline (PBS) control for 12 h were inoculated with S. sclerotiorum or P. capsici. As shown in Figure 5a

| PeSy1 interacts with a receptor-like cytoplasmic kinase in N. benthamiana
Protein elicitors bind to the target proteins in plants and cause the plant immune response. To further explore how PeSy1 exerts its elicitor activity, the proteins pulled down by PeSy1-His in N. benthamiana were analysed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). A total of 678 proteins were identified in each of two independent mass spectrometry tests ( Figure 6a). Considering that PeSy1 may play a role in both intracellular and extracellular spaces, 14 candidate targets were selected to identify their interaction relationships with PeSy1 (Table 1). The interaction of PeSy1 with several candidate targets was tested using a co-immunoprecipitation (Co-IP) assay ( Figure S2). One of these, a receptor-like protein kinase (Niben101Scf02819g03010.1), was designated as RSy1 (response to PeSy1). PeSy1-FLAG and RSy1-GFP were co-expressed in N. benthamiana leaves using green fluorescent protein (GFP) as a control, and a Co-IP assay in vivo was performed using anti-FLAG magnetic beads. The results showed that all genes were successfully expressed in N. benthamiana leaves and detected in the total protein extract. Immunoblot analysis indicated that RSy1-GFP was detected by the anti-GFP antibody in the PeSy1- A microscale thermophoresis (MST) assay was performed to investigate the binding capacity of the PeSy1-RSy1 association in vitro.
RSy1-GST protein was obtained by prokaryotic expression using the pGEX-6p-1 vector and concentrated to 150 μM using ultrafiltration tubes with appropriate molecular mass (Millipore) (Figure 6d). The MST results showed that PeSy1-His and RSy1-GST could form a saturated Sshaped binding curve with a dissociation constant (K d ) of 8.77 ± 3.18 μM and signal-to-noise ratio of 19.29 (Figure 6e). On the contrary, no binding curve was formed between PeSy1-His and GST. These data suggest that PeSy1 and RSy1 can bind with high affinity in vitro.
RSy1 contains 365 amino acid residues with a molecular weight of 40.95 kDa and an isoelectric point of 6.09. Analysis of functional domains showed that RSy1 contains only the serine/threonine kinase domain ( Figure S3), suggesting that RSy1 belongs to the typical RLCK family (Liang & Zhou, 2018). Further analysis of subcellular localization revealed that PeSy1 was localized in the plasma membrane and F I G U R E 4 PeSy1 triggers immunity responses in Nicotiana benthamiana. (a) N. benthamiana leaves were treated with 5 μM PeSy1-His or buffer control for 24 h. Reactive oxygen species (ROS) were observed as brick red staining after 3,3′-diaminobenzidine (DAB) reaction (bars are 1 cm) and callose was observed as pale green fluorescence under UV light after aniline blue reaction (bars are 100 μm). (b) Quantification of ROS accumulation and callose deposition in tobacco leaf tissue (n = 5) was determined by ImageJ. Mean ± SE. (c) N. benthamiana leaves were treated with 5 μM PeSy1 or buffer control for 12 h. Relative expression of salicylic acid (NbPR1, NbPR2) and jasmonic acid (NbPR4, NbCOI and NbPDF1.2) and ethylene (NbERF1) pathway-related marker genes in N. benthamiana were determined by reverse transcription-quantitative PCR analysis. NbActin was used as an internal control gene for normalization. Mean and SE were calculated from three biological replicates. The statistical analyses were performed with Student's t test. Bars indicate ± SE. **p < 0.01, ***p < 0.001. nucleus, and RSy1 was localized in the cytoplasmic membrane and cytoplasm under natural conditions (Figure 6f), while PeSy1 carrying its signal peptide was observed in the apoplast under plasmolysis conditions, which indicates that the signal peptide of PeSy1 may have an exocrine function (Figure 6f). In addition, considering the interaction between PeSy1 and RSy1, we used PeSy1-mcherry and RSy1-GFP to determine the colocation of PeSy1 and RSy1. As shown in Figure 6g, the red fluorescence of PeSy1-mCherry substantially overlapped with the green fluorescence of RSy1-GFP in the plasma membrane and cytoplasm. These results prove that PeSy1 could function in the apoplast and interact with RSy1 in the plasma membrane and cytoplasm.

| PeSy1 is considered as a MAMP
According to the above results, PeSy1 was conserved in Saccharothrix sp. and recognized by an RLK family protein of N. benthamiana.
Thus, we speculate that PeSy1 could act as a MAMP from Hhs.015.
The expression of PTI marker genes in N. benthamiana after PeSy1 treatment was examined by reverse transcription-quantitative PCR (RT-qPCR). These results show that 5 μM PeSy1 significantly up-regulated the expression of the PTI marker genes NbPti5, NbCYP71D20, NbAcre31, NbWRKY8, and NbWRKY7 (Figure 7a), indicating that PeSy1 is likely to be a MAMP.
To verify that RSy1 is involved in PTI signal transduction induced by PeSy1, the expression of RSy1 was detected by RT-qPCR after silencing RSy1 in N. benthamiana for 3 weeks. As shown in Figure 7b,c, the expression of RSy1 was significantly down-regulated, indicating that the RSy1-silenced N. benthamiana was successfully obtained.
Afterwards, 5 μM PeSy1-His was injected into the silenced leaves, and RNA was extracted 12 h later to measure the expression of PTI marker genes and PR genes. Compared with the controls, transcription of PTI marker genes NbPti5, NbCYP71D20, and NbWRKY7 was impaired, but PR-related genes were not markedly changed in RSy1-silenced plants.
One of the features of a MAMP is dependence on co-receptors.
Research has shown that NbBAK1 and NbSOBIR1 interact with most PRRs to facilitate intracellular signalling and have been shown to be critical for PTI (Heese et al., 2007;Liebrand et al., 2014). To test F I G U R E 5 Resistance of plants to pathogens after PeSy1 treatment. (a, b) Representative leaves showing Sclerotinia sclerotiorum and Phytophthora capsici infection after treatment with 5 μM PeSy1-His or phosphate-buffered saline (PBS) control. At 12 h posttreatment (hpt) the Nicotiana benthamiana leaves (n = 9) were inoculated by S. sclerotiorum or P. capsici. The lesion area at 48 h postinoculation (hpi) decreased 33.5% and 54.2%, respectively. Photographs were taken 48 hpi. Bars are 1 cm. (c-e) Plants were sprayed with 10 μM PeSy1-His or buffer control. After 24 h, tomato leaves (n = 9) were challenged by Pseudomonas syringae pv. tomato DC3000 and photographed 48 hpi. Compared with the control-treated leaves, the density of bacteria in PeSy1-treated leaves was significantly reduced. Bars are 1 cm. The experiment was conducted three times with similar results. The statistical analyses were performed with Student's t test. Bars indicate ± SE. ***p < 0.001.
whether PeSy1-induced cell death is associated with co-receptors, virus-induced gene silencing (VIGS) constructs were used to target

| DISCUSS ION
Biocontrol microorganisms can colonize the host plants, thereby inhibiting the invasion of plant pathogens and increasing crop yield (Compant et al., 2005;Haggag, 2010). In addition, the PTI response Most secreted proteins are characterized by their small molecular weight and are rich in cysteine residues (Templeton et al., 1994). PeSy1 was screened from our cell death-like HR analysis of the Hhs.015 secreted proteins ( Figure S1). Bioinformatics analysis showed that PeSy1 encodes an 11 kDa protein with four cysteine residues (Table S1), but no domain was identified. Sec and Twin Arginine Translocation (TAT) systems are widely distributed in grampositive and gram-negative bacteria to secrete proteins across the cytoplasmic membrane (Natale et al., 2008). Furthermore, different specialized secretion systems are known in gram-negative bacteria to perform more specialized functions, including the type I to type VI protein secretion pathways (Costa et al., 2015). On the contrary, gram-positive bacteria do not have specialized secretion systems, with the exception of the type VII secretion system reported in Mycobacteria and other high DNA G+C-content bacterial species (Actinobacteria) (Bottai et al., 2017). Previously, we described S. yanglingensis Hhs.015 as a gram-positive actinomycete with high genomic DNA G+C content (70.94 mol%) (Yan et al., 2012), but we did not predict the type VII system secretion system in Hhs.015 genome. Signal peptide predictionshows that PeSy1 has 24 N-terminal SP residues and belongs to the standard secretory SP transported by Sec translocation (Figure 2a).
Plant intercellular apoplastic space is a complex place where many important interactions occur, reflecting the close relationship between plants and pathogens (Mott et al., 2014). Using A. tumefaciens to mediate the transient expression of PeSy1 in N. benthamiana, we found that PeSy1 can induce cell death with or without its SP (Figure 1c). However, PeSy1 ΔSP showed HR symptoms later than full-length PeSy1 (Figure 1c), which is similar to the case of VmE02, a cell death elicitor from Valsa mali (Nie et al., 2019). Of note, PeSy1 was observed to be located on the apoplast and cytoplasm/ nucleus (Figure 6f), which indicates that it has a variety of ways to play its role. PeSy1 may function in the plant apoplast to induce the immune response in N. benthamiana, including a ROS burst, callose deposition, and SA-and JA/ET-mediated resistance pathway activation ( Figure 4). In addition, recombinant protein PeSy1-His treatment also induced resistance in plants against a variety of pathogenic fungi and bacteria ( Figure 5). Beneficial microorganisms can induce a systemic defence response, which is controlled by a signal network involving the plant hormones SA and JA/ET (Hammond-Kosack & Parker, 2003). According to different pathogens, hormone cross-talk regulates the plant defence response (Koornneef & Pieterse, 2008).
In this research, we found that the plant receptor-like protein kinase RSy1 is a target of PeSy1 using a pull-down mass spectrometry screening. The interaction between PeSy1 and RSy1 was confirmed by Co-IP, BiFC, and MST (Figure 6b,c,e). RSy1 belongs to RLCK in RLK family because it contains only one intracellular kinase domain ( Figure S3). Furthermore, PeSy1 and RSy1 showed the same intracellular localization (Figure 6g). The PTI marker genes were up-regulated after PeSy1 treatment (Figure 7a), indicating that PeSy1 is a MAMP of Hhs.015. Although the known PAMPs interact with the extracellular domain of transmembrane receptors (Dunning et al., 2007;Liu et al., 2012), there are exceptions. The β-glucan-binding proteins from Fabaceae plants are thought to be the receptor for β-glucan, a PAMP of Phytophthora, but do not feature the signalling domains found in other innate immune receptors (Mithöfer et al., 2000). Our findings showed that the transcription NbActin was used as the internal reference gene to standardize the samples. Mean and SE were calculated from three biological replicates. The statistical analyses were performed with Student's t test. *p < 0.05, **p < 0.01, ***p < 0.001.
redundancy of RSy1, we put forward a hypothesis that there is still an unknown transmembrane receptor to sense apoplastic PeSy1.
In this case, as our experiment shows, the activation of immunity depends on the co-receptors NbBAK1 and NbSOBIR1 (Figure 7d).
Subsequently, RSy1 may be recruited by PRR complexes to participate in the transmission of immune signals. More research is needed to verify this hypothesis.
We cannot exclude the possibility that PeSy1 functions as a cytoplasmic effector. Previous studies have shown that necrosis and ethylene-inducing peptide 1-like proteins (NLPs) and ceratoplatanin in microorganisms are both PAMPs and effectors (Böhm et al., 2014;Pazzagli et al., 2014). Plant RLCKs are immune regulators that are involved in the recognition and transmission of PAMP signalling in the PTI response (Lin et al., 2013). In previous studies, some RLCKs targeted by phytopathogen effectors also indicated their importance for immunity (Yamaguchi et al., 2013;Zhang et al., 2010). Here, we demonstrate that RSy1 promotes defence-    (Liu et al., 2002). All constructs were validated by sequencing by Tsingke Biotech (Beijing, China). All primers used are described in Table S1.  (Qi et al., 2016). The whole leaves were soaked in 1 mg/mL trypan blue dye solution (trypan blue dissolved in equal proportions solution of water, glycerol, lactic acid, and watersaturated phenol) and boiled for 3 min. After standing in the dark for 14 h, the leaves were decolourized with 2.5 g/mL chloral hydrate.

| Plasmid construction
The degree of leaf cell death can be characterized by ion leakage (Mittler et al., 1999). After agroinfiltration in N. benthamiana leaves for 4 days, six leaf disks (1 cm diameter) were put into a 5-mL centrifuge tube with 4 mL of double-distilled water and left to sit overnight. A conductivity meter (Mettler-Toledo) was used to measure the liquid conductivity (E1) after shaking. Then the disks were boiled in sealed tubes for 25 min and the conductivity (E2) measured again after cooling. The conductivity ratio was calculated as (E1/E2) × 100.

| Bioinformatics analysis
With regard to the analysis of Hhs.015 secretory proteins, the signal peptides of the proteome were predicted using SignalP (http:// www.cbs.dtu.dk/servi ces/Signa lP/) and proteins with transmembrane helix structure were excluded using TMHMM (http://www. cbs.dtu.dk/servi ces/TMHMM/). The conserved domains were searched with SMART (http://smart.embl-heide lberg.de/). The amino acid composition, molecular weight, and isoelectric point of proteins were analysed by Expasy (https://web.expasy.org/protp aram/) (Duvaud et al., 2021). The tertiary structure of protein was predicted by Phyre 2 (http://www.sbg.bio.ic.ac.uk/phyre 2/) and visualized by PymolWin (Kelley et al., 2015). The homologous sequences were compared using BLASTP in the NCBI database, and 20 sequences with high similarity were selected. Phylogenetic dendrograms were constructed using MEGA with the neighbour-joining method. DNAMAN was used to analyse the sequence conservation.
The PeSy1 data were submitted to the NCBI repository and can be found under GenBank accession number MZ396298.1.

| Prokaryotic expression and purification of PeSy1
The

| Pull-down and mass spectrometry analysis
The total protein from N. benthamiana leaves was extracted using native lysis buffer (50 mM Tris, 150 mM NaCl, 1 mM EDTA, 5% glycerol) (Solaribio) containing 1 mM PMSF, and 1% proteinase inhibitor cocktail following the manufacturer's instructions. PeSy1-His and total protein were first incubated for 10 min at room temperature, and then left at 4°C for 50 min. The interaction targets were enriched by pull down using His-tag magnetic agarose beads (Beaver) and protein samples were boiled for 10 min in 5× SDS loading buffer for SDS-PAGE. After treatment by in-gel digestion, interaction targets were identified using a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific).

| Co-IP assay
A. tumefaciens cells carrying pCAMBIA1302:RSy1-GFP, pCAMBIA-1302:PeSy1-FLAG, and P19 were co-injected into N. benthamiana leaves and the total protein was extracted using native lysis buffer at 60 hpa as described above. The lysed protein sample was used as the input sample and the remaining N. benthamiana proteins were incubated with anti-FLAG magnetic beads (Beaver). Buffer (20 mM Tris, 150 mM glycine, pH 7.5) was used for washing three times. To avoid nonspecific band contamination in the results, 0.1 M glycine-HCl (pH 2.8) in 1 M Tris (pH 9.0) was used for acid elution. Anti-GFP (Abmart) and anti-FLAG (Abmart) monoclonal antibody were detected by western blotting.

| BiFC assay
pSPYNE(R)173-RSy1 and pSPYCE(M)-PeSy1 vectors were transformed into A. tumefaciens and co-infiltrated into N. benthamiana leaves. Empty vector was used as the negative control. Two days after agroinfiltration, yellow fluorescent protein fluorescence was observed using an Olympus FV3000 confocal laser microscope. All the assays were repeated at least three times.

| Microscale thermophoresis assay
PeSy1-His was labelled using a protein labelling kit RED (Nanotemper Technologies) following the manufacturer's instructions and incubated with PBST (PBS with 0.05% Tween 20) before 30 min at room temperature. The labelled proteins were individually reacted with serially diluted RSy1-GST in 16 standard treated capillaries. The measurements were performed using a Monolith NT.115 (Nanotemper Technologies). The information of sample name and concentration was set by the Monolith Control system. The excitation power was set to 40% and the MST power was selected as medium. The experiment was carried out with three times biological repetition. The data were analysed by the Monolith Affinity Analysis program.

| Subcellular localization analysis
N. benthamiana leaves were imaged 2 days after agroinfiltration with pCAMBIA1302 or pICH86988 constructs using a laser confocal microscope (Olympus). GFP fluorescence was observed with an excitation wavelength of 488 nm and mCherry fluorescence was observed with an excitation wavelength of 561 nm. To analyse the fluorescence distribution after plasmolysis, cells were treated with 1.5 M NaCl for 5 min. A. tumefaciens infiltration carrying the pCAM-BIA1302 or pICH86988 empty vector was used as a control.

| VIGS in N. benthamiana
N. benthamiana leaves at the four-leaf stage were infiltrated with cultured A. tumefaciens GV3101 containing a mixture of TRV1, P19, and TRV2 constructs (OD 600 = 0.4). TRV2:GFP and TRV2:PDS (phytoene desaturase) were used as the control groups. Three weeks after agroinfiltration, the albino phenotype caused by PDS gene silencing was used as a reference (Ma et al., 2015), and the silencing efficiency of the target genes was determined by RT-qPCR at the position corresponding to the processed leaves. The primers used for RT-qPCR are listed in Table S2. All experiments were performed with at least three biological replicates.

| Statistical analysis
The data were reported as mean ± SE from three independent biological replicates and analysed with GraphPad Prism 8 software.
Significant differences between mean values were evaluated using Student's t test (p < 0.05). Plan of Northwest A&F University (S202110712293).

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.