Generally, A. thaliana plants used in this study were grown in a phytotron under exposure to an 8/16 h light/dark cycle, at a temperature between 22°C and 24°C, relative humidity of 45% and illumination of 5500 LX. For the protoplast preparation and detection of callose deposition, all the plants are cultivated under short day chamber, with 12 h light/12 h dark, temperature 22°C, humidity 40%, and light 5500 LX.

133 A. thaliana LRR-RLKs were selected as experimental genes from the Ensembl Plants database (Howe et al., 2020) and protein domain prediction tool SMART (NORMAL mode) (Letunic et al., 2021) ( Supplementary Table S1 ).

For transient expression experiment in Arabidopsis protoplasts, 134 RLK KD fragments were amplified using wild-type cDNA as templet, which carrying 15 bp homologous sequences of both ends of the vectors at each end by specific gene primers. After digesting the vectors HBT95-FLS2 NT-GFP (previously constructed in the lab) with NcoI and KpnI, above amplified fragments were cloned into the linearized vectors by fusion cloning method, respectively, to generate 133 HBT95-FLS2 NT-RLK KD-GFP plasmids plus the positive control plasmid HBT95-FLS2 NT-FLS2 KD-GFP. For stable transgenic plants construction, the coding sequences of the 6 selected RLK KD domains were obtained by digesting HBT95-FLS2 NT-RLK KD-GFP with BamHI and PstI, and cloned into pCamBia1305-FLS2 NT-GFP (previously constructed in the lab) to generate pCamBia1305-FLS2 NT-RLK KD-GFP. These constructs were then transferred into Agrobacterium tumefaciens strain GV3101 by heat shock method and subsequently transformed into fls2 mutant plants by floral dipping (Clough and Bent, 1998).

For selection of homozygous transgenic lines, T₁ generation transgenic plants that exhibit hygromycin resistance are harvested individually to obtain seeds for the T₂ generation. Approximately 40 seeds from each line are then planted and their segregation ratios on resistance plates are used to determine whether they consist of single-copy insertions (with 3/4 of the plants being resistant). From the T₂ lines judged to be single-copy insertions, 12 resistant seedlings are randomly chosen (of which theoretically 1/3 should be homozygous lines) for transplantation and individual harvesting to obtain T₃ generation seeds. The T₂ plants corresponding to T₃ lines where all seeds survive on resistance plates are considered to be homozygous lines. Two independent homozygous transgenic lines are then selected for further characterization and analysis.

Method for extracting mesophyll protoplasts (Yoo et al., 2007) from Arabidopsis fls2 mutants is described below. Briefly, 4-week-old fls2 leaves were cut into 0.5-1 mm filaments and quickly immersed in an enzymatic hydrolysis solution (0.4% macerozyme R10 and 1.5% cellulose Onozuka R10). Under the same conditions of plant culture without light, the leaves were hydrolyzed for 4 h before an equal volume of W5 buffer (2 mM MES, pH 5.7; 154 mM NaCl; 125 mM CaCl₂; 5 mM KCl) was added, which was subsequently filtered through 350-µm nylon mesh and centrifuged at 667 rpm for 2 min. The supernatant was removed via vacuum suction and the protoplasmic cells were re-suspended by suitable W5 before placed on ice for 30 min to deposit the protoplasmic cells. After removing W5, the concentration of protoplasts was adjusted by MMG buffer (4 mM MES, pH 5.7; 15 mM MgCl₂; 0.4 M mannitol) to 2×10⁵ cells/mL to facilitate subsequent transformation experiment.

Polyethylene glycol (PEG)-mediated transfection (Yoo et al., 2007) was performed to transfer total 10 µL plasmids (35S::FLS2 NT-RLK KD-GFP+ FRK1::LUC ( Asai et al., 2002)+ UBQ10::GUS (Sheen, 2001)) into A. thaliana protoplasts before incubating (at 24°C and 2000 LX) in W5 buffer for 11-12 h, and then treated with 1 µM flg22 for another 12 h.

After treatment with flg22, the above protoplasts were harvested by centrifuging at 667 rpm for 2 min at 4°C for luciferase and β-glucuronidase (GUS) activity test (Popov et al., 2016), which method was described as follows. Firstly, 50 µL cell lysate was added to the collected protoplasts and a cell crusher was used to assist in full lysis. Then, 5 µL of the above protoplast lysate was added into a 96-well plate mixing with 100 μL luciferin before the luciferase activity was tested with a Perkin Elmer VICTOR X3 (Waltham, Massachusetts, USA). On the other hand, 2 µL protoplast lysate was added into a 96-well plate mixing with 25 µL MUG buffer (10 mM Tris-HCl, pH 8; 2 mM MgCl₂; 1 mM 4-methyl-umbelliferyl-β-D-glucuronide), which was then incubated at 37°C for 30 min before adding 225 μL 0.2 M Na₂CO₃ to terminate reaction. Finally, the GUS activity was measured by the Perkin Elmer VICTOR X3.

6 recombinant 35S::FLS2 NT-RLK KD-GFP transient expression vectors were individually co-transfected with 35S::SCAMP1-RFP expression vector into fls2 protoplasts, and the localization of fluorescent proteins was observed by a confocal microscope (OLYMPUS FV1200, Japan) after 12 hours of cultivation. Confocal images of 6 rRLK-GFP fluorescence in root cells of plate-grown fls2 transgenic plants were obtained by a confocal microscope (OLYMPUS FV1200, Japan). GFP signal was detected at 488 nm excitation (500-540 nm emission, HyD3 detector).

Leaves from 4-week-old plant were infiltrated with 1 μM flg22 or ddH₂O for 18 h, respectively, and soaked in the six-well plate with 4 mL fixative (ethanol∶ acetic acid = 3∶ 1) for 1 h. After replacing the fixative with 50% ethanol, the plate was incubated at 65°C for 15 mins for decolorization. Then, the ethanol was taken away and the samples were treated with 150 mM K₂HPO₄ (pH 9.5) solution containing 0.01% Methyl Blue for 1 h in the dark. Finally, glycerol was added to preserve the samples after blotting out the staining solution, and pictures were taken using fluorescence microscope for callose deposition analysis.

About 80 mg plant tissue was collected from 14-day-old seedlings grown on 1/2 MS medium after 1 μM flg22 or ddH₂O treatment. RNA was extracted using the EASY spin Plus Plant RNA kit (TaKaRa), which then reverse transcribed by HiScript^® III RT SuperMix for qPCR (+gDNA wiper) (Vazyme) to obtain cDNA. cDNA samples were initially normalized with ACTIN2 by real-time PCR using the ChamQ Universal SYBR qPCR Master Mix kit (Vazyme). The primers used for amplification of ACTIN2, FRK1, PR1 and WRKY33 were as follows. ACTIN2, 5’-CTTGTTCCAGCCCTCGTTTG-3’ and 5’-CAGCGATACCTGAGAACATAGTG G-3’; FRK1, 5’-TCTGAAGAATCAGC TCAAGGC-3’ and 5’-TGTTGGCTTCACATCTCTGTG-3’; PR1, 5’- ACGCAGAAC AACTAAGAGGCAAC-3’ and 5’- AGCCTT CTCGCTAACCCACAT-3’; WRKY33, 5’- GCTGCTATTGCTGGTCACTCC-3’ and 5’-T GCGTTTGAAGGTTGCTGTT-3’.

For virulent Pseudomonas syringae pv. tomato (P.s.t.) strain DC3000 infection, bacterial cultures diluted to OD₆₀₀ = 5×10^-4 with 10 mM MgCl₂ were used to infiltrate leaves of around 4-week soil-grown plants. Leaf disks from the infected areas were taken at 0 d and 3 d after infiltration to quantify the bacterial colony-forming units (cfu) on LB plates with Kanamycin and Rifampicin antibiotic selections.

70 mg leaf tissue from 15-day-old 1/2 MS medium-grown plants was collected after 1 μM flg22 treatment for 5 min and ground into powder by liquid nitrogen. The samples were homogenized in extraction buffer (10% 1 M Tris-HCl, pH8; 0.1% SDS; 2% β-mercaptoethanol; 0.1% 1 M DTT; 1% PMSF), and the supernatant of each sample was obtained by centrifuging at max speed for 5 min. After adding 4× SDS loading buffer (60 mM Tris-Cl, pH6.8; 2% SDS; 5% β-mercaptoethanol; 10% glycerol; 0.01% bromophenol blue) to a final concentration of 1×, the samples were boiled at 95°C for 5 min and stored at -20°C.

The above protein samples were separated by 10% SDS-PAGE. For MAPK phosphorylation detection, western blotting was carried out using an antibody specific for p42/44-MAPK (Cell Signalling Technology, Danvers, MA, USA) to determine whether the phosphorylation of MPK3 and MPK6 was affected.

For the sake of discovering some common features of kinase domain in the immune-related rRLKs, multiple sequence alignment was executed by Clustal Omega (v: 1.2.4) (Sievers and Higgins, 2018; Madeira et al., 2024).

Furthermore, to verify whether these crucial amino acids were also conserved across species, 13 LRR-RLKs in other species that have been confirmed to be immune-related were screened from NCBI database, and their KD regions were predicted by SMART (Letunic et al., 2021) ( Supplementary Table S2 ). The kinase domain of FLS2 in Arabidopsis were multiply aligned with the kinase domain of the 13 RLKs in other species (denoted as KD-F-O below, that is KD-FLS2-Other species) using Clustal Omega (Sievers and Higgins, 2018).

Specifically, we focused on the alignment of three groups (immune-positive group “KD-p” which includes 109 KDs of RLKs identified from our protoplast screening that do not activate PTI, immune-negative group “KD-n” encompassing 24 KDs from RLKs that were identified to activate PTI to varying degrees in the protoplast screening and other species group “KD-F-O” consisting of 16 published KDs of RLKs from other species known to be involved in immunity), and compared their conserved amino acid sites based on Clustal Omega and its visual tool Jalview (v: 2.11.3.3) (Waterhouse et al., 2009): (I) Initially, highly conserved residues within the KD-p group were identified, serving as potential characteristic sites of immune-associated RLKs. (II) Subsequently, these candidate residues were cross-referenced against highly conserved residues within the KD-n group to eliminate common fundamental characteristics present in KDs across all LRR-RLKs. This step was crucial to pinpoint sites with a high likelihood of being related to immune functions. (III) Finally, these putative immune-related sites were compared with corresponding sites in the KD-F-O group to assess whether these key sites exhibited conserved traits across different species.

Utilizing a recombinant RLK approach, 133 HBT-FLS2 NT-RLK KD-GFP transient expression vectors ( Supplementary Table S1 ) were successfully constructed. By means of the Arabidopsis protoplast transient expression system, whether these RLK KDs participated in plant immune signal transduction was screened and identified ( Figure 1A ). Among them, a total of 24 RLK kinase domains after flg22 treatment could proficiently mediate PTI pathway signal transduction and effectively trigger the activation of the downstream reporter FRK1, with HBT-FLS2 NT-FLS2 KD-GFP serving as the positive control and HBT-FLS2 NT-GFP as the negative control ( Supplementary Figure S1 , Figure 1B ). Of the 24 RLKs, the functions of 16 (Wang et al., 2001; Caño-Delgado et al., 2004; Huffaker and Ryan, 2007; Tsuwamoto et al., 2008; Kinoshita et al., 2010; Chang et al., 2013; Xiao et al., 2015; de Oliveira et al., 2016; Huang et al., 2016; Patharkar et al., 2017; Wang et al., 2017; Hu et al., 2018; Isner et al., 2018; He et al., 2021; Huang et al., 2023; Chang et al., 2024) have been reported to be either related to immunity or development.

Notably, 6 RLK KDs derived from At1g17750 (PEPR2) (Huffaker and Ryan, 2007), At2g13790 (SERK4) (de Oliveira et al., 2016), At2g01950 (BRL2) (Wang et al., 2001), At4g28490 (HAESA) (Patharkar et al., 2017), At3g47570 and At5g63930 (PSYR3) (Ogawa-Ohnishi et al., 2022) demonstrated comparable or even superior functionality to the WT FLS2 KD in activating the PTI signaling pathway ( Figure 1B ). Although reports have identified the homologous gene HaOr7 in sunflower, which encodes a RLK protein that confers resistance to Orobanche cumana race F (Duriez et al., 2019), there have been no reports to date regarding whether AT3G47570 is involved in plant immunity in Arabidopsis.

In order to ascertain whether the above identified RLKs do function in immune response, 6 rRLKs-GFP including FLS2 NT-At1g55610 (BRL1) (Caño-Delgado et al., 2004) KD-GFP, FLS2 NT-At2g37050 (SIMP1) (He et al., 2021) KD-GFP, FLS2 NT-At3g47570 KD-GFP, FLS2 NT-At2g28990 KD-GFP, FLS2 NT-At5g63930 KD-GFP, FLS2 NT-At2g01950 KD-GFP) that can respond to PTI elicitor signals and activate downstream immune responses in protoplasts to varying degrees were selected to further construct fls2 OE rRLK-GFP transgenic plants in this study. Firstly, as plant signal receptors, these rRLKs can effectively transmit signals from the extracellular environment to the intracellular milieu only if they are correctly localized within the cell. Therefore, we initially examined the GFP fluorescence localization of the 6 rRLKs-GFP both in protoplasts and in each selected transgenic homozygous line. As demonstrated in Figure 2 , all the tested rRLK fusion proteins co-localized with the plasma membrane marker SCAMP1-RFP in protoplasts. Moreover, they were normally expressed in the fls2 background and appropriately localized to the plasma membrane as anticipated ( Supplementary Figure S2 ).

In our examination of the transgenic plants expressing the pre-screened rLRR-RLKs, it was imperative to ensure that the presence of these recombinant proteins did not impinge on normal plant functions. Consistent with the elicitor induction required for the occurrence of PTI, the morphological evaluation showed that these homozygous transgenic lines revealed no notable deviations from the Col-0 and fls2 control groups throughout their entire life cycle ( Figure 3 ). Correct localization of these rRLKs confirmed their potential to function as intended, and crucially, their expression did not automatically provoke an autoimmunity response in the plants. Therefore, we concluded that the engineered PTI receptors did not negatively impact plant vitality or reproductive processes.

Continuing our analysis, we delved into the PTI signaling capacity of these rRLKs and discovered diversity in their subsequent signaling pathways. Through an array of resistance characterization assays including callose deposition, defense marker gene expression and resistance against P.s.t. DC3000, we sought to delineate the impact of these 6 rRLKs on the PTI pathway in Arabidopsis. Unlike fls2 with no response to the induction of flg22, the level of callose deposition in the transgenic background recovered to varying degrees after flg22 treatment, and some lines almost recovered to levels similar to or even higher than that of Col-0 ( Figure 4 ). Consistent with the transient screening results, all rRLK transgenic lines could respond to the induction of flg22 and produce high levels of FRK1 expression ( Figure 5A ), while they had discrepant expression levels of PR1 or WRKY33 under flg22 induction ( Figures 5B, C ). Notably, the KD kinase domains from At1g55610, At2g37050, and At2g01950 or from At5g63930 appear to facilitate more robust signal transduction than WT FLS2 in the FRK1 or PR1-dependent pathway, respectively, ultimately conferring heightened resistance against the virulent bacterial pathogen P.s.t. DC3000 in plants ( Figure 5D ).

Findings demonstrate that all 6 kinase domains (from At1g55610, At2g37050, At3g47570, At2g28990, At5g63930, At2g01950), when fused with FLS2 NT, could mediate transmission of the PTI elicitor signal, triggering immune-related physiological changes in transgenic plants. These alterations, however, differed among the transgenic lines in terms of callose deposition and flg22-induced expression of defense genes. This suggests that, while the rRLKs may utilize the same FLS2 NT for signal reception, their downstream signaling pathways diverge, potentially orchestrating differing defense responses via distinct routes. Western blot analyses were performed to tentatively examine the underlying mechanisms of these enhanced disease-resistant signals in rRLKs ( Figure 6 ). Results revealed that flg22 induction led to varied enhancements in MPK3 and MPK6 phosphorylation, indicating a MAPK cascade-dependent pathway in mediating plant immune responses across these kinase domains.

Building upon the insights gleaned from the clustering of immune-related rRLKs, we integrate bioinformatics analyses to spotlight the shared conserved amino acid residues critical to the KD function in plant immunity. The alignments showcase 57 highly conserved sites across 24 KDs that are positively linked to immune signaling (referred to as KD-p). Moreover, these 57 sites encompass 16 sites that are also deeply conserved within a broader set of 109 KDs that do not activate PTI (labeled as KD-n), indicating their foundational role in the structural integrity and function of LRR-RLKs in general. Remarkably, a subset of 9 sites out of these 57, a group marked by absolute conservation within KD-p but not within KD-n, emerges as particularly promising candidates for immune-specific functionality. These sites are also found to be conserved across an array of LRR-RLKs in different species (known as KD-FLS2-Other or KD-F-O), implicating them as critical determinants in the activation of immune signaling across the plant kingdom ( Figure 7 , Supplementary Table S3 ).

This assessment, drawing connections between the protoplast pre-screening experiment and broad phylogenetic data, reinforces the notion that key conserved sites within rLRR-RLKs are not just vestiges of protein architecture, but are actively engaged in sensing and defending against external biotic stress. By highlighting these sites’ preservations in related species’ RLKs—spanning EFR in Brassica (Zhang et al., 2018), FLS2 variants in grapevine (Trdá et al., 2014), tomato (Robatzek et al., 2007; Hind et al., 2016), and other Solanaceae members (Wang et al., 2016; Wang et al., 2018; Xu et al., 2018), and beyond (Soderlund et al., 2009) ( Supplementary Table S2 ), the analysis underscores the evolutionary importance of these domains in plant immunity.

For a long time, it was believed that plants lacked an immune system. However, with the development of botany and in-depth research into the molecular mechanisms underlying plant disease resistance, it has become clear that, similar to animals, plants have evolved a unique multi-layered innate immune system to respond to pathogen infections in their environment (Jones and Dangl, 2006; Spoel and Dong, 2012; Cheng et al., 2019). The most fundamental component of the plant immune response is pattern recognition, which includes the specific recognition of the conserved pathogen PAMPs by PRRs on the surface of plant cell membrane, and the detection of pathogen-secreted avirulence proteins by intracellular immune receptor-resistance (R) proteins (Boller and Felix, 2009; Lolle et al., 2020). Among these components, PTI serves as the first line of defense for plants against pathogens. The key to triggering this response lies in the recognition of specific PAMPs by PRRs. To date, multiple PRRs have been identified in plants using molecular biology, proteomics, and cell biology methods. Most of these PRRs belong to the LRR-RLK class and play crucial roles in various signal transduction pathways. These PRRs can specifically recognize corresponding ligands to activate the intracellular kinase domains, thereby transmitting extracellular signals into the cell through either self-phosphorylation or co-phosphorylation with their co-receptors, and then dynamically regulate plant growth, development and immune response (Clark et al., 1997; Gómez-Gómez and Boller, 2000; Fisher and Turner, 2007; Robatzek et al., 2007; Kinoshita et al., 2010; Furukawa et al., 2014; Sun et al., 2017). Despite the large number of RLKs encoded by the genomes of various species, the functions of only a small fraction of them have been elucidated, primarily because many of their corresponding ligands remain unknown. Fortunately, early studies revealed that recombinant RLK (rRLK) can still successfully mediate the plant immune responses when the extracellular domain and transmembrane domain (NT) of one RLK are recombined with the kinase domain (KD) of another RLK (Albert et al., 2010; Greeff et al., 2012). Based on this finding, it is possible to assess whether an unknown RLK is involved in a biological process by recombining the NT of a known ligand RLK with the KD of the unknown RLK (Brutus et al., 2010).

Leveraging the Arabidopsis protoplasts PTI activation transient reporting system, the aforementioned recombinant RLK strategy (FLS2 NT + tested RLK KD) was employed to systematically screen 133 LRR-RLKs from the Arabidopsis genome. This comprehensive screen was designed to identify candidates that may play a role in the PTI pathway, and found 18.0% of the tested RLKs were able to complete PTI immune signaling and activate downstream reporter gene expression in Arabidopsis protoplasts after flg22 induction. In addition to identifying several RLKs that have been reported to participate in plant immune responses, such as At2g01950 (Wang et al., 2001), At1g17750 (Huffaker and Ryan, 2007), At5g01540 (LecRKA4.1) (Singh et al., 2012), At1g09970 (RLK7) (Hou et al., 2014), At2g13790 (de Oliveira et al., 2016), At4g28490 (Patharkar et al., 2017), and At5g65710 (HSL2) (Wang et al., 2017), it was also found that some RLKs that were not previously known to mediate PTI immune responses in protoplasts. It is well-established that proper subcellular localization is essential for protein function. RLKs act as plant receptor proteins situated on the plasma membrane, where they receive extracellular ligand signals and activate intracellular signaling pathways via their kinase domains (Hunter, 1995; Shiu and Bleecker, 2001). By constructing transgenic lines expressing 6 rRLKs (FLS2 NT-At1g55610 KD, FLS2 NT-At2g37050 KD, FLS2 NT-At3g47570 KD, FLS2 NT-At2g28990 KD, FLS2 NT-At5g63930 KD, FLS2 NT-At2g01950 KD) and conducting disease resistance characterization analysis, it was revealed that the recombination of RLK did not affect its correct subcellular localization, and further confirmed the involvement of these RLKs in plant disease resistance. These results indicate that the transient screening system has high credibility, and utilizing recombination strategy with the help of such system can promote our understanding of unknown RLK functions.

Research on identified immune RLKs suggests that the patterns of recognition to pathogen invasion signal and the downstream signal transduction pathways triggered by different RLKs are remarkably similar (Bjornson et al., 2021; Huang et al., 2023). However, the reasons behind this commonality remain unclear. In this study, the NT domains of each RLK were artificially unified. That is, on the basis of excluding differences in ligand recognition or even recognition sensitivity, the same extracellular signal was transmitted into the cell through different KD domains. Although the degree of immune activation differed, there were still some commonalities. Through sequence analysis, it was discovered that certain conserved motifs exist within the identified immune-related RLKs, which are hypothesized to be part of the reason for the commonality in downstream immune responses mediated by different RLKs. Additionally, based on the standardized plant RLK bioinformatics analysis database (Yin et al., 2024), the discovery of these common features may assist researchers in the expeditious and efficacious identification of pertinent candidates involved in plant immunity among the numerous putative RLKs in the Arabidopsis genome, and even in other species. Once the target RLK is identified, its upstream specific ligand can be screened, its downstream immune-related interaction proteins can be identified, and the specific molecular mechanism of its involvement in PTI signaling pathway can be analyzed. The elucidation of the roles of these RLKs represents a considerable leap forward, establishing a foundation for deeper investigation into the processes governing plant immunity.

Studies to date have shown that expression of RLK can enhance the disease resistance of plants upon induction, while exhibiting no negative effects on the plant growth and development under normal growth conditions. The implication of this work could extend to identifying potential targets for improving crop resilience against pathogens by understanding and potentially enhancing the signaling pathways leading to immune responses.