Subfamily C7 Raf-like kinases MRK1, RAF26, and RAF39 regulate immune homeostasis and stomatal opening in Arabidopsis thaliana

Abstract

The calcium-dependent protein kinase CPK28 regulates several stress pathways in multiple plant species. Here, we aimed to discover CPK28-associated proteins in Arabidopsis thaliana. We used affinity-based proteomics and identified several potential CPK28 binding partners, including the C7 Raf-like kinases MRK1, RAF26, and RAF39. We used biochemistry, genetics, and physiological assays to gain insight into their function. We define redundant roles for these kinases in stomatal opening, immune-triggered reactive oxygen species (ROS) production, and resistance to a bacterial pathogen. We report that CPK28 associates with and trans-phosphorylates RAF26 and RAF39, and that MRK1, RAF26, and RAF39 are active kinases that localize to endomembranes. Although Raf-like kinases share some features with mitogen-activated protein kinase kinase kinases (MKKKs), we found that MRK1, RAF26, and RAF39 are unable to trans-phosphorylate any of the 10 Arabidopsis mitogen-activated protein kinase kinases (MKKs). Overall, our study suggests that C7 Raf-like kinases associate with and are phosphorylated by CPK28, function redundantly in stomatal opening and immunity, and possess substrate specificities distinct from canonical MKKKs.

Keywords: Arabidopsis thaliana; C7 Raf‐like kinase; CPK28; MRK1; RAF26; RAF39; immunity; stomata.

Fig. 1

CPK28 associates with C7 Raf‐like kinases and phosphorylates RAF26 and RAF39. (a–c) Split‐luciferase (Luc) complementation assays with FER‐nLuc or CPK28‐nLuc and cLuc‐MRK1 (a), cLuc‐RAF26 (b), and cLuc‐RAF39 (c). Total photon counts are plotted as relative light units (RLU) after co‐expression of the respective proteins in Nicotiana benthamiana. Individual values are plotted from a representative experiment (n = 12) and are significantly different from each control (Student's unpaired t‐test; P < 0.0001); error bars represent SE. These assays were repeated over 4 times each by BD over a 12‐month period with similar results; representative data are shown. (d–f) In vitro kinase assays using His₆‐MBP‐CPK28 as the kinase and catalytically inactive His₆‐MRK1^K110E (d), His₆‐RAF26^K87E (e), or His₆‐RAF39^K101E (f) as substrates. Autoradiographs (autorad) indicate incorporation of γ³²P and protein loading is indicated by poststaining with Coomassie Brilliant Blue (CBB). Asterisks indicate proteins of interest. Assays were performed more than 3 times each by MGD over a 6‐month period with similar results; representative data are shown. Cloning credits are provided in Supporting Information Table S1. All loci refer to gene names in Arabidopsis thaliana.

Fig. 2

Subfamily C7 Raf‐like kinases have a unique extended loop in the N‐lobe of the kinase domain and auto‐phosphorylate in vitro. (a) Protein sequences from the C family of Raf‐like kinases were retrieved from The Arabidopsis Information Resource and a multiple sequence alignment was generated using the Muscle algorithm in MegaX (Kumar et al., 2018). The alignment was used to generate a neighbor‐joining tree with 1000 bootstraps; the tree shown here is just the C7 subfamily. The full C‐Raf family alignment was used to analyze the consensus signature motif G‐T‐x‐x‐[W/Y]‐M‐A‐P‐E and visualized here using WebLogo (Crooks et al., 2004). The protein kinase domains are labeled in green based on the Uniprot database; the protein lengths are indicated on the far right. (b) Multiple sequence alignment of the C7 Raf‐like kinases compared to RAF36 to illustrate the unique extension identified in C7 Raf‐like kinases, which forms an extended disordered loop between the β4 and β5 sheets of the N‐lobe (predictions shown for MRK1 and RAF36). Predicted protein structures were downloaded from Alphafold2 (Jumper et al., 2021) and visualized using ChimeraX (Pettersen et al., 2021). (c–e) In vitro kinase assays indicate that His₆‐MRK1 (c), His₆‐RAF26 (d), and (e) His₆‐RAF39 are able to auto‐phosphorylate. Each assay included catalytically inactive variants as controls. Autoradiographs (autorad) indicate incorporation of γ³²P and protein loading is indicated by poststaining with Coomassie Brilliant Blue (CBB). Asterisks indicate proteins of interest. JM performed the analysis in a and b; MGD performed the kinase assays in c–e at least three times over a 6‐month period with similar results and representative data is shown. Cloning credits are provided in Supporting Information Table S1. All loci refer to gene names in Arabidopsis thaliana.

Fig. 3

MRK1, RAF26, and RAF39 localize to endomembranes. (a–f) Confocal micrographs of green fluorescent protein (GFP)‐tagged MRK1, RAF26, and RAF39 co‐expressed with either red fluorescent protein (RFP)‐tagged BRI1 (a–c) or an mCherry‐tagged endoplasmic reticulum (ER) marker protein (d–f) in Nicotiana benthamiana. Maximum projections (max) are shown in the lower panels and single‐plane sections (secant) are shown in the upper panels. Bars: (a) 20 μm; (b–f) 10 μm. These assays were repeated 3 times by AR over a 6‐month period with similar results. (g) MRK1‐GFP, RAF26‐GFP, and RAF39‐GFP were expressed in N. benthamiana, proteins extracted and a western blot using anti‐GFP antibodies was performed. MRK1‐GFP (c. 69.6 kDa), RAF26‐GFP (c. 67.7 kDa), and RAF39‐GFP (c. 69.7 kDa) migrated to their expected sizes. Coomassie Brilliant Blue (CBB) of RuBisCO indicates loading. This experiment was repeated twice with identical results by MGD. Cloning credits are provided in Supporting Information Table S1. All loci refer to gene names in Arabidopsis thaliana.

Fig. 4

MRK1, RAF26, and RAF39 regulate immune homeostasis and stomatal opening. (a–c) Reactive oxygen species (ROS) production measured in relative light units (RLUs) after treatment with 100 nM flg22 (a), 100 nM elf18 (b), or 500 nM AtPep1 (c). Values represent means ± SE (n = 6–12). Data presented in a was collected by BD; data presented in b and c was collected by MGD. These assays were repeated several times by BD, MGD, EC, and JM over multiple years. (d) Stomatal apertures before (0 min) and following exposure to 1 μM flg22 (60, 180 min). Individual values are plotted and represent ratios of stomatal width : length. The straight line represents the mean (n = 120). Lower case letters indicate statistically significant groups, determined by a one‐way analysis of variance (ANOVA) followed by Tukey's post hoc test (P < 0.005). (e) Representative micrographs of stomata before flg22 treatment, showing visibly smaller apertures in mrk1‐1 raf26‐2 raf39‐2 compared to Col‐0. Bar, 5 μm. Experiments in d and e were repeated 5 times by BD and AR over a 12‐month period; representative data collected by BD is shown. (f) Growth of Pseudomonas syringae pv tomato (Pst) isolate DC3000 3 days after spray‐inoculation. Data from three independent experimental replicates are plotted together, denoted by gray, blue, and magenta dots. Values are colony‐forming units (CFU) per leaf area (cm²) from 4–5 samples per genotype (each sample contains three leaf discs from three different infected plants). The line represents the mean (n = 14). Asterisks indicate significantly different groups, determined by a Student's unpaired t‐test (P < 0.0001). Data was collected by AR over a 12‐month period. Credits for genetic crosses and genotyping are provided in Supporting Information Table S1. All loci refer to gene names in Arabidopsis thaliana.

Fig. 5

MRK1, RAF26, and RAF39 do not trans‐phosphorylate MKKs. (a) An unrooted phylogenetic tree of the Arabidopsis thaliana MKKK, ZIK/WNK, and Raf‐like subfamilies (green) together with Homo sapiens MKKK, MLK, and Raf kinases (gray). Subfamily C7 Raf‐like kinases are indicated. A multiple sequence alignment using the full‐length sequences of all proteins in the subfamilies was performed using Clustal Omega and the resulting neighbor‐joining phylogenetic tree was visualized using iTOL (Letunic & Bork, 2021); subfamilies are collapsed at the ends of nodes. Analysis performed by JM. (b–d) In vitro kinase assays indicate that His₆‐MRK1 (b), His₆‐RAF26 (c), and His₆‐RAF39 (d) are unable to trans‐phosphorylate any of the 10 Arabidopsis MKKs N‐terminally tagged with glutathione S‐transferase (GST). Catalytically inactive MKK variants were used and are numbered as 1–10 for MKK1^K97E, MKK2^K108E, MKK3^K112E/K113E, MKK4^K108E, MKK5^K99E, MKK6^K99E, MKK7^K74E, MKK8^K82E/K83E, MKK9^K76E, MKK10^K77E, and are indicated by asterisks. Autoradiographs (autorad) indicate incorporation of γ³²P and protein loading is indicated by poststaining with Coomassie Brilliant Blue (CBB). MGD and TD performed the assays three times over a 3‐month period with similar results and representative data is shown. Cloning credits are provided in Supporting Information Table S1. All loci in b–d refer to gene names in A. thaliana.