The receptor kinase FER is a RALF-regulated scaffold controlling plant immune signaling

See allHide authors and affiliations

Science  20 Jan 2017:
Vol. 355, Issue 6322, pp. 287-289
DOI: 10.1126/science.aal2541

Small peptides allow rapid responses

RALFs (rapid alkalinization factors), a family of small peptides in plants, are produced in response to rapidly changing conditions. Stegmann et al. studied the agility and diversity built into this signaling network. Some RALFs, such as RALF23 and its relative RALF33, are activated by proteolytic cleavage. Others, such as RALF32, are not. RALF23 and RALF33 are called into play after a pathogen triggers immune responses. RALF32, on the other hand, regulates seedling growth. All three of these RALFs use the same receptor kinase, which can interact with other signaling components. Thus, plant responses can be fine-tuned by rapid release of peptides.

Science, this issue p. 287


In plants, perception of invading pathogens involves cell-surface immune receptor kinases. Here, we report that the Arabidopsis SITE-1 PROTEASE (S1P) cleaves endogenous RAPID ALKALINIZATION FACTOR (RALF) propeptides to inhibit plant immunity. This inhibition is mediated by the malectin-like receptor kinase FERONIA (FER), which otherwise facilitates the ligand-induced complex formation of the immune receptor kinases EF-TU RECEPTOR (EFR) and FLAGELLIN-SENSING 2 (FLS2) with their co-receptor BRASSINOSTEROID INSENSITIVE 1–ASSOCIATED KINASE 1 (BAK1) to initiate immune signaling. We show that FER acts as a RALF-regulated scaffold that modulates receptor kinase complex assembly. A similar scaffolding mechanism may underlie FER function in other signaling pathways.

Plant immune pattern recognition receptors (PRRs) are often receptor kinases (1). The Arabidopsis thaliana (hereafter, Arabidopsis) receptor kinases FLAGELLIN-SENSING 2 (FLS2) and EF-TU RECEPTOR (EFR) bind bacterial flagellin (or the epitope flg22) and EF-Tu (or the epitopes elf18/elf24), respectively, and form ligand-induced complexes with their co-receptor BRASSINOSTEROID INSENSITIVE 1–ASSOCIATED KINASE 1 (BAK1) (1).

To decipher the negative regulation of plant PRR-mediated immune signaling, we screened for modifier of bak1-5 (mob) mutants that regain immune responses in the immunodeficient mutant background bak1-5 (2). Here, we report the characterization of MOB6.

bak1-5 is impaired in the production of reactive oxygen species (ROS) upon flg22 or elf18 treatment (3). We identified the recessive bak1-5 mob6 mutant on the basis of restoration of this response (fig. S1A). To map the mob6 locus, we sequenced bulked F2 segregants from a backcross between bak1-5 mob6 and bak1-5 combined with phenotypic and genetic analysis of an F2 population from a cross between bak1-5 mob6 and Col-0 [see the supplementary materials (SM)]. We found a homozygous missense mutation [where Ser replaces Pro612 (P612S)] in At5g19660 encoding the subtilase SITE-1 PROTEASE (S1P)/SBT6.1 (fig. S1, B and C). Allelism tests and transcomplementation assays confirmed that the mob6 phenotype is caused by a mutation in S1P (fig. S1, D and E). We thus renamed mob6 as s1p-6.

We characterized the effect of mutant S1P alleles on immune signaling. s1p-3 and s1p-6 single mutants produced more ROS in response to elf18, flg22, and chitin (fig. S1, D and E; fig. S2A; and fig. S3) and exhibited increased expression of the immune marker genes FRK1 and PHI1 upon elf18 treatment (fig. S2B). Also, s1p-3 and s1p-6 mutants were more resistant to the hypovirulent bacterium Pseudomonas syringae pv. tomato (Pto) DC3000 coronatine-minus (COR) strain (fig. S2C). Thus, S1P is a negative regulator of immunity.

Similar to its human ortholog, the Arabidopsis subtilase S1P processes substrates with the canonical cleavage site RxxL/RxLx (where R is arginine, L is leucine, and x is any amino acid) (4). The endogenous peptide RAPID ALKALINIZATION FACTOR 23 (RALF23; At3g16570)—an established S1P substrate in Arabidopsis (5)—is a major hub in a flg22-regulated transcriptional network (6). Consistent with a potential role of RALF23 in immunity, elf18 treatment or inoculation with the type III–deficient strain Pto DC3000 hrcC rapidly increased the processing of the propeptide PRORALF23 (fig. S4, A and B), in an S1P-dependent manner (fig. S4C).

RALF23 treatment led to a dose-dependent inhibition of elf18-induced ROS production [median inhibitory concentration (IC50) ~ 200 nM] (Fig. 1A and fig. S5) and resistance to Pto DC3000 (Fig. 1B). Furthermore, RALF23 overexpression (7) inhibited elf18-triggered ROS production and increased susceptibility to Pto DC3000 COR (Fig. 1, C and D) and to the fungus Plectosphaerella cucumerina (7). Conversely, loss of RALF23 (fig.S6A) led to increased elf18-triggered ROS production (Fig. 1E) and resistance to Pto DC3000 COR (Fig. 1F). RALF23 similarly inhibited flg22-triggered ROS production (fig. S7). Furthermore, treatment with RALF23 suppressed the heightened elf18-induced ROS production in s1p-6 mutants (fig. S8), which suggests that the negative regulatory function of S1P is executed by the processing of RALF23 or related peptides. For example, the closely related RALF33 (At4g15800) peptide (figs. S9A and S10A) (8, 9) could also inhibit elf18-induced ROS production (figs. S6B; S9, B and C; and S10B). We conclude that RALF23, as well as RALF33, negatively regulates immunity.

Fig. 1 The endogenous peptide RALF23 negatively regulates immunity.

(A) ROS production in Col-0 leaf discs treated with 100 nM elf18, 1 μM RALF23, or cotreatment [all in 2 mM 2-(N-morpholino)ethanesulfonic acid–potassium hydroxide (MES-KOH), pH 5.8]. Values are means of total photon counts over 30 min ± SE, n = 16. Letters indicate significance in one-way analysis of variance (ANOVA) (a and b, P < 0.001; a to c, P < 0.001; b and c, P < 0.05). Kinetics are shown in fig. S18A. (B) Colony-forming units (cfu) of Pto DC3000 after syringe inoculation in leaves pretreated with mock treatment, 1 μM elf18, 1 μM RALF23, or cotreatment (all in 2 mM MES-KOH, pH 5.8) for 24 hours, determined 2 days after inoculation. Values are means ± SD, n = 4 (one-way ANOVA; P < 0.05). (C and E) ROS production after elicitation with 100 nM elf18 or water. Values are means of total photon counts over 30 min ± SE. Letters indicate significance in one-way ANOVA (a and b, P < 0.001; a to c, P < 0.001; b and c, P < 0.05). Kinetics are shown in fig. S18, B and C. (D and F) Number of Pto DC3000 COR bacteria determined 3 days after surface inoculation. Values are means ± SD, n = 4 (one-way ANOVA; P < 0.05). Similar results were obtained in three independent experiments.

The Arabidopsis genome encodes about 35 RALF peptides (8, 9). Only 11 RALFs (including RALF23 and RALF33) display a S1P cleavage site (fig. S10A). Cotreatment with RALF34 (containing a predicted S1P cleavage site) inhibited elf18-induced ROS burst to the same extent as RALF23 and RALF33, whereas RALF32 (lacking a predicted S1P cleavage site) did not (fig. S10B). This suggests that S1P-cleaved RALFs inhibit immunity. RALF23, RALF33, and the more divergent RALF32 triggered seedling growth inhibition (fig. S10C) in a way similar to RALF1 (9). However, RALF32 does not affect elf18-triggered ROS production (fig. S10B). All RALF peptides previously tested induced alkalinization (8, 10), suggesting that immune inhibition is not a general property of RALFs and is independent of the alkalinization activity, which is consistent with our bioassays involving RALF peptides being performed under buffered conditions (SM).

The Arabidopsis malectin-like receptor kinase FERONIA (FER; At3g51550) was recently identified as a receptor for RALF1 (9). RALF1, RALF23, and RALF33 are closely related (fig. S9A and S10A) and have overlapping gene expression patterns with FER (fig. S11) (8, 9). fer-2 and fer-4 mutants were insensitive to the inhibitory effect of RALF23 or RALF33 peptide on elf18-induced ROS production (Fig. 2A and fig. S12A), which was complemented in a fer-4/FER–GREEN FLUORESCENT PROTEIN (fer-4/FER-GFP) line (Fig. 2A). Furthermore, RALF23 overexpression in fer-2 had no effect on elf18-induced ROS production (fig. S12B). The genetic dependence of FER in RALF23-, RALF33-, and RALF32-triggered growth inhibition (fig. S10C) suggested that FER may bind additional RALF peptides.

Fig. 2 RALF23-mediated inhibition of immunity is FER dependent.

(A) ROS production after treatment with 100 nM elf18 alone or cotreated with 1 μM RALF23 (all in 2 mM MES-KOH, pH 5.8). Values are means of total photon counts over 40 min ± SE, n = 16. Letters indicate significance in one-way ANOVA (a and b, P < 0.001; a to c, P < 0.05; b and c, P < 0.01). Kinetics are shown in fig. S18D. (B) In vitro binding assay with maltose-binding protein (MBP)–ectoFER or MBP-ectoEFR purified from E. coli. Pull-down assay for immunoprecipitation (IP) was done with neutravidin beads; Western blots (WB) were probed with antibody against MBP (α-MBP). (C) In vitro binding assay with ectoFER–V5-6xHis (ectoFER tagged with V5 antibody epitope and hexahistidine) produced in insect cells. Pull-down was performed with streptavidin beads; Western blots were probed with antibody against V5 (α-V5). (D) Quantitative binding analysis using synthetic RALF23 peptide and ectoFER–V5-6xHis using microscale thermophoresis. Similar results were obtained in three independent experiments, except for (C), where the assays were performed twice with identical results.

Biotinylated RALF23 bound to recombinant FER ectodomain (ectoFER) but not the unrelated EFR receptor ectodomain (ectoEFR) (Fig. 2B). Similarly, ectoFER expressed and purified from insect cells bound to biotin-RALF23 but not biotin-elf24 (Fig. 2C), with dissociation constant (Kd) values ~300 nM and ~600 to 900 nM, as revealed by microscale thermophoresis and isothermal titration calorimetry, respectively (Fig. 2D and fig. S13), which is consistent with values reported for other ligand-receptor kinase pairs (11, 12) and the IC50 for RALF23-mediated inhibition of elf18-triggered ROS production (fig. S5). Thus, in addition to RALF1 (9), FER is also a receptor for RALF23.

FER is enriched in detergent-resistant membranes after flg22 treatment (13). The fer-2 and fer-4 mutants were hyposensitive to elf18, flg22, and chitin (Figs. 2A and 3A and figs. S12 and S14) and were more susceptible to Pto DC3000 COR (Fig. 3B), indicating that FER can also positively regulate immunity. FER weakly associates with both FLS2 and BAK1, with the latter being strongly enhanced upon flg22 treatment (Fig. 3C). Flg22-induced FLS2-BAK1 complex formation was reduced in fer-4 and restored in fer-4/FER-GFP (Fig. 3D). Cotreatment with RALF23 reduced ligand-induced FLS2/EFR-BAK1 complex formation (Fig. 3, E and F). RALF23 overexpression had a similar effect on flg22-induced FLS2-BAK1 complex formation (fig. S15). RALF23 or the loss of FER did not affect accumulation of FLS2, EFR, or BAK1 (Fig. 3, D to F, and fig. S15). Thus, ligand-induced complex formation between FLS2/EFR and their co-receptor BAK1 is promoted by FER and inhibited by RALF23.

Fig. 3 FER is a RALF-regulated scaffold for immune receptor complexes.

(A) ROS production after elicitation with 100 nM elf18 or water. Values are means of total photon counts over 30 min ± SE, n = 8. Letters indicate significance in one-way ANOVA (a and b, P < 0.05; b and c, P < 0.05; b to d, P < 0.05; a to c P < 0.001; a to d, P < 0.001). Kinetics are shown in fig. S18E. (B) Number of Pto DC3000 COR bacteria 3 days after surface inoculation. Values are means ± SD, n = 4 (one-way ANOVA; P < 0.05). (C to F) Coimmunoprecipitation of (C) FER-GFP from fer-4/FER-GFP; (D) FLS2 from Col-0, fer-4, or fer-4/FER-GFP; (E) FLS2-GFP from Col-0/FLS2-GFP; or (F) EFR-GFP from efr/EFR-GFP seedlings treated with or without the indicated peptides (100 nM flg22, 100 nM efl18, 1 μM RALF23, or water) for 10 min. Western blots were probed with antibodies α-GFP, α-BAK1, and α-FLS2. CBB, Coomassie brilliant blue. Similar results were obtained in three independent experiments.

Our data suggest that FER acts as a scaffold to regulate immune receptor–complex formation. FER may reside in plasma membrane microdomains as part of preformed signaling “platforms,” together with receptors and co-receptors. Whether FER-mediated regulation intersects with other regulators of FLS2/EFR-BAK1 complex formation, such as BIR2 and IOS1 (14, 15), will be an interesting topic for future investigation. Two-thirds of RALF proteins lack a predicted S1P cleavage site and are devoid of a propeptide region (fig. S10A). Treatment with one of these, RALF17, induced ROS production and acted additively to elf18 (fig. S16, A and B). RALF17 pretreatment was also sufficient to induce resistance to Pto DC3000 (fig. S16C). As RALF17-induced ROS production is dependent on FER (fig. S16B), the activity is not caused by a possible contamination with synthetic peptides (e.g., flg22 or elf18) used in our laboratory.

PRORALF23 is processed by S1P within minutes of elicitor perception (fig. S4), which suggests a negative-feedback mechanism to inhibit the scaffolding function of FER and to dampen immune signaling. Fungal pathogens secrete peptides with homology to RALF23 (16, 17); these fungal RALF orthologs may suppress immunity by inhibiting the formation of active receptor complexes.

FER has emerged as a regulator of many biological processes, ranging from fertilization to inhibition of cell elongation and growth (18). Many of these processes involve receptor kinases and co-receptors (19). This suggests that FER may have a similar scaffolding function within other receptor kinase complexes. Different, but overlapping, expression patterns of FER and RALF genes (fig. S11) (9, 20) suggest that a variety of FER-RALF modules may regulate diverse receptor kinase complexes during growth, development, or environmental sensing.

Supplementary Materials

Materials and Methods

Figs. S1 to S18

Tables S1 and S2

References (2130)

References and Notes

  1. Acknowledgments: This research was funded by the Gatsby Charitable Foundation (C.Z.), the European Research Council (grant “PHOSPHinnATE” to C.Z.), the Austrian Academy of Science through the Gregor Mendel Institute (Y.B.), the Deutsche Forschungsgemeinschaft (Fellowship STE 2448/1 to M.S.), the European Molecular Biology Organization and the U.K. Biotechnology and Biological Sciences Research Council (Fellowships ALTF 459-2011 and BB/M013499 to J.M.), and the Erasmus Mundus program (H.R.). We thank L. Stransfeld, the John Innes Centre horticultural service and The Sainsbury Laboratory tissue culture service for technical assistance, and all members of the Zipfel laboratory for comments. We thank N. Keinath, M. Sussman, P. Vera, S. Wolf, S. Robatzek, and H.-M. Wu for providing biomaterials; the VBCF Protein Technologies Facility for excellent assistance; and T. Clausen for access to his isothermal titration calorimetry platform. M.S., J.M., E.S.-L., Y.B., and C.Z. designed and conceived the experiments and analyzed the data. M.S., J.M., E.S.-L., H.R., A.L., and Y.B. performed the experiments. N.H. provided unpublished constructs. M.S., J.M., and C.Z. wrote the manuscript with input from all authors. Supplement contains additional data.
View Abstract

Stay Connected to Science

Navigate This Article