CYCLIN-DEPENDENT KINASE8 differentially regulates plant immunity to fungal pathogens through kinase-dependent and -independent functions in Arabidopsis

Abstract

CYCLIN-DEPENDENT KINASE8 (CDK8) is a widely studied component of eukaryotic Mediator complexes. However, the biological and molecular functions of plant CDK8 are not well understood. Here, we provide evidence for regulatory functions of Arabidopsis thaliana CDK8 in defense and demonstrate its functional and molecular interactions with other Mediator and non-Mediator subunits. The cdk8 mutant exhibits enhanced resistance to Botrytis cinerea but susceptibility to Alternaria brassicicola. The contributions of CDK8 to the transcriptional activation of defensin gene PDF1.2 and its interaction with MEDIATOR COMPLEX SUBUNIT25 (MED25) implicate CDK8 in jasmonate-mediated defense. Moreover, CDK8 associates with the promoter of AGMATINE COUMAROYLTRANSFERASE to promote its transcription and regulate the biosynthesis of the defense-active secondary metabolites hydroxycinnamic acid amides. CDK8 also interacts with the transcription factor WAX INDUCER1, implying its additional role in cuticle development. In addition, overlapping functions of CDK8 with MED12 and MED13 and interactions between CDK8 and C-type cyclins suggest the conserved configuration of the plant Mediator kinase module. In summary, while CDK8's positive transcriptional regulation of target genes and its phosphorylation activities underpin its defense functions, the impaired defense responses in the mutant are masked by its altered cuticle, resulting in specific resistance to B. cinerea.

Figure 1.

The cdk8 Mutant Displays Increased Susceptibility to A. brassicicola and P. syringae but Resistance to B. cinerea. (A) Genomic structure of the CDK8 gene and positions of the T-DNA insertion in cdk8 mutant alleles. UTR, untranslated region. (B) Verification of cdk8 T-DNA insertion mutants by genomic PCR. LP, T-DNA left border genomic primer; RP, T-DNA right border primer. (C) RT-qPCR data showing loss of CDK8 expression in cdk8 mutants. Relative transcript levels were normalized with Arabidopsis ACT2. The normalized expression level of the wild type was set to 1. Error bars indicate se (n = 3). Two independent biological replicates were performed. Significance between the mean values was analyzed statistically (Student’s t test, **P < 0.01). (D) A. brassicicola disease symptoms (top panel) and disease lesion size (bottom panel) in the wild type and the cdk8 mutant at 4 DAI. (E) and (F) Trypan blue staining of A. brassicicola-inoculated leaves of the wild type (E) and the cdk8 mutant (F) 48 h after inoculation. Bar = 4 mm. (G) and (H) Closeups of the disease lesion areas from (E) and (F). Bar = 1 mm. (I) B. cinerea disease symptoms (top panel) and disease lesion size (bottom panel) in the wild type and the cdk8 mutant at 4 DAI. (J) and (K) Trypan blue staining of the wild type (J) and the cdk8 mutant (K) 36 h after inoculation with B. cinerea. Bar = 4 mm. (L) and (M) Closeups of the disease lesion areas from (J) and (K). Bar = 1 mm. (N) Bacterial growth in wild-type and cdk8 mutant plants. Leaves of 5-week-old plants were infiltrated with bacterial suspension (OD₆₀₀ = 0.0005). At 3 DAI, leaf discs were collected and bacterial growth was quantified and expressed in colony-forming units (cfu). FW, fresh weight. The disease assay was performed by drop-inoculation of B. cinerea on leaves of soil-grown plants or A. brassicicola on detached leaves. The average lesion sizes are mean values ± se from four independent replicates (n = 40). A minimum of 10 leaves for each genotype were used for each biological replicate, and the disease assay was repeated at least four times with similar results. All of the data were statistically analyzed. Asterisks indicate significant differences (Student’s t test, *P < 0.05, **P < 0.01).

Figure 2.

CDK8-Dependent Expression of ERF and Plant Defensin Genes. (A) Induced expression of genes encoding the TFs ERF1, ORA59, and ERF2 is reduced in the cdk8 mutant. (B) Induced expression of defensin genes is dependent on CDK8. Total RNAs were extracted from 5-week-old plants grown in soil before and after B. cinerea inoculation. Relative transcript levels were normalized with Arabidopsis ACT2. The normalized expression level of the wild type at 0 h was set to 1. Error bars indicate se (n = 3). Three independent biological replicates were performed. Significance of differences between the mean values was analyzed statistically (Student’s t test, *P < 0.05, **P < 0.01).

Figure 3.

Interactions between CDK8 and MED25 Promote the Expression of PDF1.2. (A) Reduced MeJA-induced expression of PDF1.2 in cdk8 and med25 mutants. Gene expression was determined by RT-qPCR with the Arabidopsis ACT2 used for normalization. The expression level of PDF1.2/ACT2 in the mock-treated wild type was set at 1. Error bars indicate se (n = 3). (B) CDK8 and MED25 are required for ERF1/ORA59-mediated PDF1.2 activation. The activity of the reporter gene construct PDF1.2 pro-GUS was normalized to the full-length LUC construct used as an internal control. Relative GUS:LUC activity ratios (fold change) are mean values from three independent biological replicates (n = 6). Error bars indicate se. (C) Root growth responses of cdk8 to MeJA. Wild-type and cdk8 seeds were germinated on half-strength Murashige and Skoog medium. Root lengths were measured at 7 d after transferring to medium containing MeJA at the indicated concentrations. Results are mean values ± se from two independent replicates (n = 50). (D) Interaction between CDK8 and MED25 revealed by split-luciferase complementation assay. Equal amounts of purified plasmids were transiently coexpressed in Arabidopsis protoplasts with substrate luciferin. The LUC activities are mean values from three biological replicates, and error bars indicate se (n = 3). (E) CDK8 and MED25 interact in Co-IP assays in Arabidopsis protoplasts. MED25-MYC, but not ERF4-MYC, was present in CDK8-HA-precipitated complex. The experiment was repeated three times. (F) A. brassicicola disease symptoms and disease lesion size in the wild type, cdk8, med25, and cdk8 med25 double mutant. (G) B. cinerea disease symptoms and disease lesion size in the wild type, cdk8, med25, and cdk8 med25 double mutant. The disease assays and lesion size measurements were performed as described in Figure 1 and were repeated at least three times. The lesion sizes are mean values ± se from at least 20 disease lesions. The lesion sizes and photographs are from 4 DAI. All of the data were statistically analyzed. Asterisks indicate significant differences (Student’s t test, *P < 0.05, **P < 0.01). [See online article for color version of this figure.]

Figure 4.

CDK8 Associates with the Regulatory Regions of the PDF1.2 Gene. (A) Schematic showing the positions of PDF1.2 gene primers used for ChIP-qPCR. (B) CDK8 is required for MeJA-induced RNAP II recruitment to the PDF1.2 promoter. ChIP-qPCR results are shown with PDF1.2 gene-specific primers. Chromatin was extracted from wild-type and cdk8 seedlings 1 h after treatment with 100 μM MeJA and then precipitated with anti-RPB2 antibody (Abcam) or only IgG (negative control with no antibody [No Ab]). The RNAP II recruitment at Arabidopsis ACT2 was used as a control because its expression is independent of CDK8. The ChIP-qPCR data show that RNAP II recruitment to PDF1.2 500-bp, TATA box, and coding and terminator regions decreased compared with the wild type. (C) CDK8 associates with the PDF1.2 promoter. Chromatin was extracted from 5-week-old wild-type and cdk8;35S:CDK8-MYC transgenic plants and then precipitated with anti-MYC antibody (Abcam) or only IgG (No Ab). (D) Recruitment of CDK8 to the PDF1.2 promoter is significantly enhanced by B. cinerea infection. Chromatin was extracted from 5-week-old cdk8;35S:CDK8-MYC #7 transgenic plants that were mock-inoculated or B. cinerea spray-inoculated. ChIP was performed with anti-MYC antibody (Abcam) or only IgG (No Ab). The CDK8 recruitment to 500-bp upstream, TATA box, and coding and terminator regions of PDF1.2 was determined by quantitative PCR using primers at different positions of the PDF1.2 gene as shown in (A). Error bars in (B) to (D) indicate se (n = 3). Two biological replicates were performed with similar results for each ChIP-qPCR experiment. The significance of differences in mean values is marked by asterisks (Student’s t test, *P < 0.05, **P < 0.01).

Figure 5.

CDK8 Is Required for the Accumulation of Arabidopsis HCAAs. (A) Expression of the AACT1 gene in response to B. cinerea and A. brassicicola. RNA was extracted from mock-, B. cinerea-, or A. brassicicola-inoculated leaves. The experiment was repeated three times. Error bars indicate se (n = 3). (B) Detection of HCAAs from rosette leaves of 5-week-old Arabidopsis wild-type and cdk8 plants at 48 h after inoculation with A. brassicicola. The data are mean values ± se from two biological replicates (n = 4). FerAgm, feruloylagmatine; CouAgm, p-coumaroylagmatine; FerPtr, feruloylputrescine; CouPtr, p-coumaroylputrescine. (C) CDK8 is specifically recruited to a 500-bp region upstream of the AACT1 promoter region. (D) Recruitment of CDK8 to the AACT1 gene is enhanced by B. cinerea infection. In (C) and (D), chromatin was extracted from 5-week-old cdk8;35S:CDK8-MYC #7 transgenic plants that were mock- or B. cinerea-inoculated. Error bars indicate se (n = 3). ChIP-qPCR experiments were repeated two times with similar results. No Ab, negative control with no antibody. The data were statistically analyzed. Asterisks indicate significant differences (Student’s t test, *P < 0.05, **P < 0.01).

Figure 6.

The cdk8 Mutant Has Increased Cuticle Permeability and a Reduced Cuticular Layer. (A) Leaves of the cdk8 mutant (right) showing the glossy phenotype. (B) Enhanced cuticle permeability of cdk8 leaves (right) revealed by toluidine blue staining. (C) Transmission electron microscopic images of wild-type and cdk8 mutant leaf epidermal cells. The white arrows mark cutin. Bar = 200 nm. (D) The cdk8 mutant shows enhanced water loss. Water loss was expressed as the percentage of initial fresh weight. Values are averages from 20 leaves for each of three independent experiments. Values shown are means ± se.

Figure 7.

CDK8 and MED25 Both Interact with the TF WIN1. (A) CDK8 and MED25 interact with WIN1 in a split-luciferase complementation assay. The LUC activities are mean values from three biological replicates, and error bars indicate se (n = 3). (B) WIN1 and CDK8/MED25 interact in a Co-IP assay. (C) The expression of WIN1 in two independent transgenic WIN1 overexpression lines. (D) B. cinerea disease symptoms and disease lesion size in the wild type and WIN1 overexpression lines at 4 dpi. (E) A. brassicicola disease symptoms and disease lesion size in the wild type and WIN1 overexpression lines at 4 dpi. (F) Expression of the PDF1.2 gene in the wild type and WIN1 overexpression lines after mock or B. cinerea inoculation. Two independent WIN1 overexpression lines (WIN1 OE #2 and WIN1 OE #8) were used in these studies. The significance of differences between mean lesion sizes was analyzed. Values shown are means ± se. Asterisks indicate significant differences (Student’s t test, *P < 0.05, **P < 0.01).

Figure 8.

CDK8 Functions in Kinase-Dependent and Independent Manners. (A) Amino acid sequence comparison between Arabidopsis and human CDK8 proteins. The conserved aspartic acid at amino acid position 176 (176D) of CDK8, which is critical for its kinase activity, is marked by the red asterisk. (B) In vitro kinase assay showing that substitution of 176D to alanine nullifies the kinase activity of Arabidopsis CDK8. GST-CDK8 recombinant protein displayed autophosphorylation as well as phosphorylation of substrate MBP. However, GST-CDK8^D176A recombinant protein lost both autophosphorylation and MBP phosphorylation activities. The experiment was repeated two times with similar results. CBB, Coomassie Brilliant Blue. (C) Immunocomplex kinase assays showing the phosphorylation of histones by CDK8. Kinase assay was performed using protein extracts from cdk8;35S:CDK8-MYC #7 and cdk8;35S:CDK8^D176A-MYC #26 transgenic seedlings. (D) Ectopic expression of CDK8-MYC or CDK8^D176A-MYC restores wild-type levels of B. cinerea responses in the cdk8 mutant. (E) CDK8-MYC but not CDK8^D176A-MYC plants restore wild-type responses to A. brassicicola. In (D) and (E), the disease assays were performed as described in Figure 1, and disease lesions were recorded at 4 DAI. Data are mean values ± se (n = 20), and statistical analysis was conducted to determine the significant differences of values (Student’s t test, **P < 0.01).

Figure 9.

The Kinase Activity of CDK8 Is Required for Target Gene Expression. (A) to (D) PDF1.2 (A), AACT1 (B), CER1 (C), and CER6 (D) gene expression levels in wild-type, cdk8;35S:CDK8-MYC, and cdk8;35S:CDK8^D176A-MYC transgenic plants. Lines 7 and 24 are two independent transgenic lines expressing 35S:CDK8-MYC, whereas lines 26 and 28 are 35S:CDK8^D176A-MYC lines. Error bars represent se (n = 3). (E) The kinase-dead CDK8 mutation does not affect the CDK8-MED25 interaction in split-luciferase complementation assays. Error bars represent se (n = 3). (F) RNAP II recruitment to the PDF1.2 promoter is dependent on the kinase activity of CDK8. Reduced RNAP II recruitment is shown in cdk8;35S:CDK8^D176A-MYC transgenic plants compared with cdk8;35S:CDK8-MYC plants after B. cinerea infection. Error bars indicate se (n = 3). ChIP-qPCR experiments were repeated two times with similar results. No Ab, negative control with no antibody. Statistical analysis was conducted to determine differences in mean values. Asterisks indicate significant differences (Student’s t test, *P < 0.05, **P < 0.01).

Figure 10.

Mediator Kinase Module Components MED12 and MED13 Mediate Defense Responses. (A) Growth phenotypes of 4-week-old wild-type, cdk8, med12, and med13 mutant plants. (B) B. cinerea disease symptoms and disease lesion sizes in the wild type and Mediator kinase module mutants. (C) A. brassicicola disease symptoms and disease lesion sizes in the wild type and Mediator kinase module mutants. (D) Enhanced cuticle permeability of the wild type and Mediator kinase module mutants revealed by toluidine blue staining. (E) Relative expression of defense and cuticular wax genes in the wild type and Mediator kinase module mutants. The disease assays in (B) and (C) were conducted as described in Figure 1. The lesion sizes are mean values ± se from at least 20 disease lesions, and the data were analyzed with Student’s t test (*P < 0.05, **P < 0.01).

Figure 11.

Arabidopsis CDK8 Interacts with Two CycCs. (A) Strong interactions between CDK8 and two CycCs in split-luciferase complementation assays. Data are mean values ± se from three independent experiments. Error bars indicate se (n = 3). (B) CDK8 interacts with CycCa and CycCb in Co-IP assays. The experiment was repeated at least three times. (C) Reduced expression of CycCa and CycCb genes in the cycCab mutant allele. The cycCab allele (SALK_039400C) carries a T-DNA insertion between the CycCa and CycCb genes disrupting the expression of both genes. CycCa and CycCb are tightly linked and transcribed in the same orientation. (D) A. brassicicola disease symptoms and disease lesion sizes in wild-type, cdk8, and cycCab mutant plants. (E) B. cinerea disease symptoms and disease lesion sizes in wild-type, cdk8, and cycCab mutant plants. The disease assays in (C) and (D) were conducted as described in Figure 1. The lesion sizes are mean values ± se from at least 20 disease lesions. Asterisks indicate significant differences (Student’s t test, *P < 0.05, **P < 0.01).