The flowering time regulator FLK controls pathogen defense in Arabidopsis thaliana

Abstract

Plant disease resistance is a complex process that is maintained in an intricate balance with development. Increasing evidence indicates the importance of posttranscriptional regulation of plant defense by RNA binding proteins. In a genetic screen for suppressors of Arabidopsis (Arabidopsis thaliana) accelerated cell death 6-1 (acd6-1), a small constitutive defense mutant whose defense level is grossly in a reverse proportion to plant size, we identified an allele of the canonical flowering regulatory gene FLOWERING LOCUS K HOMOLOGY DOMAIN (FLK) encoding a putative protein with triple K homology (KH) repeats. The KH repeat is an ancient RNA binding motif found in proteins from diverse organisms. The relevance of KH-domain proteins in pathogen resistance is largely unexplored. In addition to late flowering, the flk mutants exhibited decreased resistance to the bacterial pathogen Pseudomonas syringae and increased resistance to the necrotrophic fungal pathogen Botrytis cinerea. We further found that the flk mutations compromised basal defense and defense signaling mediated by salicylic acid (SA). Mutant analysis revealed complex genetic interactions between FLK and several major SA pathway genes. RNA-seq data showed that FLK regulates expression abundance of some major defense- and development-related genes as well as alternative splicing of a number of genes. Among the genes affected by FLK is ACD6, whose transcripts had increased intron retentions influenced by the flk mutations. Thus, this study provides mechanistic support for flk suppression of acd6-1 and establishes that FLK is a multifunctional gene involved in regulating pathogen defense and development of plants.

Introduction

Plants are constantly challenged by a myriad of pathogens with various lifestyles. In response, plants can recognize many pathogen-derived molecules, such as pathogen associated molecular patterns (PAMPs) and effector proteins, and subsequently activate layers of defense responses, including PAMP-triggered immunity (PTI), effector-triggered immunity, and systemic acquired resistance (Boller and Felix, 2009; Dodds and Rathjen, 2010; Fu and Dong, 2013; Toruno et al., 2016). Depending on the type of invading pathogen, plants deploy different and sometimes conflicting downstream signaling to combat the invader. Salicylic acid (SA) is generally considered to be important for defense against biotrophic pathogens, e.g. Pseudomonas syringae. The major SA biosynthesis genes include ISOCHORISMATE SYNTHASE 1 (ICS1) and HOPW1-1-INTERACTING3 (WIN3) (Wildermuth et al., 2001; Lee et al., 2007; Rekhter et al., 2019). Additional genes, such as ACCELERATED CELL DEATH 6 (ACD6), ENHANCED DISEASE SUSCEPTIBILITY (EDS1), and PHYTOALEXIN-DEFICIENT 4 (PAD4) (Falk et al., 1999; Lu et al., 2003), also affect SA levels but their functions have not been completely understood. SA receptors NONEXPRESSER OF PR GENES 1 (NPR1), NPR3, and NPR4 perceive and transduce SA signal to the nucleus for reprogramming gene expression (Wu et al., 2012; Yan and Dong, 2014). On the other hand, jasmonic acid (JA) promotes resistance to necrotrophic pathogens, e.g. Botrytis cinerea, as well as to insects (Glazebrook, 2005; Campos et al., 2014). SA and JA act antagonistically in the defense response against certain pathogens, yet they also work synergistically under other stress conditions. Proper interplay between signaling pathways is important to ensure robust plant defense. Without a thorough understanding of how different signaling molecules activate defense against pathogens, our design of strategies to improve disease resistance of crop plants against their natural pathogens will be limited.

Defense is an energetically costly process and defense activation against some pathogens can come at the expense of plant development (Karasov et al., 2017; Ning et al., 2017). Flowering is one of the most critical developmental landmarks in the lifecycle of plants. It is well known that flowering in the model plant Arabidopsis (Arabidopsis thaliana) is tightly controlled by at least five pathways: the autonomous pathway, photoperiod (light), vernalization, hormones, and the circadian clock (Amasino, 2010). These pathways ultimately converge upon a few downstream genes, e.g. the flowering repressor gene FLOWERING LOCUS C (FLC) and flowering activator genes FLOWERING LOCUS T and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1, to determine flowering time. Growing evidence supports the involvement of flowering pathways in Arabidopsis defense, including the autonomous pathway genes FPA and FLOWERING LOCUS D (Lyons et al., 2013; Singh et al., 2013), certain light receptors (Genoud et al., 2002; Griebel and Zeier, 2008; Roden and Ingle, 2009), flowering regulatory hormones (Spoel and Dong, 2008; Bari and Jones, 2009), and the circadian clock (Lu et al., 2017). On the other hand, defense activation reciprocally affects flowering. For instance, pathogen infection (Korves and Bergelson, 2003) and changes in SA levels via exogenous SA application or genetic manipulation, i.e. using SA mutants ics1 and eds5, are associated with altered flowering time (Martinez et al., 2004; Jin et al., 2008; March-Diaz et al., 2008; Endo et al., 2009; Wada et al., 2009; Liu et al., 2012). These observations suggest crosstalk between plant defense and flowering control. Mechanisms underlying this crosstalk remain to be fully elucidated.

The FLOWERING LOCUS K HOMOLOGY DOMAIN (FLK) gene is known as a positive regulator of flowering, acting in the autonomous pathway. It encodes a putative protein with three K homology (KH) repeats (Lim et al., 2004; Mockler et al., 2004). The KH domain is an ancient RNA binding motif found in proteins from diverse organisms (Nicastro et al., 2015). The KH domain has about 70 amino acids and can be in arrays of up to 15 KH repeats. Some KH-domain proteins were shown to function in RNA metabolism, affecting pre-mRNA processing and mRNA stability. Disruptions of KH-domain proteins are associated with multiple diseases in humans (Geuens et al., 2016). There are 26 predicted KH-domain proteins in Arabidopsis (Lorkovic and Barta, 2002) but only a few of them have been characterized. Like FLK, several Arabidopsis KH-domain proteins were shown to play roles in flowering control (Cheng et al., 2003; Ripoll et al., 2009; Yan et al., 2017; Ortuno-Miquel et al., 2019; Woloszynska et al., 2019). Some KH-domain proteins were shown to be important for stress response, including resistance against a fungal pathogen (Thatcher et al., 2015) and abiotic stress (Chen et al., 2013; Guan et al., 2013). To date, none of the KH-domain proteins are shown to influence both defense and development.

To better understand gene networks involved in pathogen defense, we have utilized the Arabidopsis mutant accelerated cell death6-1 (acd6-1) in our studies. acd6-1 exhibits a constitutively high level of defense that is inversely associated with the size of the plant (Rate et al., 1999; Lu et al., 2003). This feature of acd6-1 makes it a convenient genetic tool in gauging genetic interactions among defense genes (Song et al., 2004; Ng et al., 2011; Wang et al., 2011a; Wang et al., 2014; Hamdoun et al., 2016; Zhang et al., 2019). acd6-1 has also been proven as a powerful starting material in a large-scale genetic screen for acd6-1 suppressors, allowing for the uncovering of defense genes (Lu et al., 2009; Wang et al., 2014). From this genetic screen, we isolated an allele of the FLK gene, flk-5, that suppressed acd6-1-conferred phenotypes, including SA accumulation, cell death, and dwarfism. In addition to late flowering, flk mutants showed enhanced susceptibility to the biotrophic pathogen Pseudomonas syringae but enhanced resistance to the necrotrophic pathogen Botrytis cinerea. In addition, the flk mutants exhibited compromised SA accumulation and basal defense upon P. syringae infection and/or elicitation by P. syringae derived molecules. RNA-seq analysis revealed that expression of some major defense- and development-related genes as well as alternative splicing of a number of genes were influenced by the flk mutations. We further provided evidence to show that the ACD6 gene is a potential target of FLK for alternative splicing. Together our data support that FLK is a multifunctional protein important for plant defense and development and establish mechanistic links for these roles of FLK.

Results

Identification of an FLK allele from acd6-1 suppressor screen

In order to better understand defense gene networks, we took advantage of the Arabidopsis mutant acd6-1 whose defense level is in a gross inverse association with the size of the plant (Rate et al., 1999; Lu et al., 2003). We conducted a large-scale screen of an acd6-1 mutant library generated by T-DNA insertional mutagenesis for acd6-1 suppressors, which showed partially increased plant size and were likely disrupted in genes involved in pathogen defense (Lu et al., 2009; Wang et al., 2014). Among the suppressor mutants isolated, we identified a plant displaying delayed flowering time. We backcrossed the mutant to acd6-1 two times and confirmed the association of acd6-1 suppression and late flowering phenotypes. In addition, we outcrossed the suppressor mutant to Arabidopsis Col-0 and obtained the single mutant. We further conducted TAIL-PCR (Liu et al., 1995) and located a T-DNA insertion in the second intron of the FLK gene (Figure 1A). FLK encodes a protein with three predicted KH repeats that is previously known as a flowering regulator acting in the autonomous pathway (Lim et al., 2004; Mockler et al., 2004). Because four other flk mutants were previously described (Lim et al., 2004; Mockler et al., 2004), we designated this allele as flk-5. The acd6-1 flk-5 mutant showed a partial suppression of acd6-1 in terms of plant size and cell death, and it flowered later than both the acd6-1 mutant and Col-0 (Figure 1, B–D). Such partial suppression of acd6-1-conferred phenotypes was confirmed by introducing another FLK allele, flk-1 (SALK_007750), into acd6-1 (Figure 1). In addition, we made a construct carrying the 2.3 kb FLK promoter and the full-length FLK genomic fragment that was translationally fused with the reporter gene GFP (FLK-GFP) and used it to transform flk-1. Independent homozygous FLK-GFP/flk-1 transformants demonstrated a rescue of late flowering conferred by flk-1 (Supplemental Figure S1, A and B). The FLK-GFP protein was localized to the nucleus (Supplemental Figure S1C).

Figure 1

Mutations in FLK suppress acd6-1-conferred phenotypes and lead to late flowering. A, Gene structure of FLK. Triangles indicate the positions of two FLK alleles, boxes indicate exons, and the line indicates the promoter, introns, and 3′ downstream region. B, Pictures of 25-d old plants (upper panel) and cell death (lower panel). Leaves at the fourth to seventh positions of each genotype were stained with trypan blue for cell death. C, Salicylic acid (SA) quantification. Total SA was extracted from the plants and quantified by HPLC. D, Flowering time measurement. Plants were grown in a light cycle of 16 h L/8 h D and recorded for flowering time, days after planting for the first visible appearance of inflorescence of a plant. Error bars represent mean ± Sem in (C) (n = 4) and (D) (n ≥ 18). Different letters in C and D indicate significant difference among the samples (P < 0.05; One-way ANOVA with post hoc Tukey's HSD test). These experiments were repeated two times with similar results.

FLK is a positive regulator of P. syringae resistance and a negative regulator of B. cinerea resistance

The suppression of acd6-1-conferred phenotypes by flk mutants suggests that FLK is important for pathogen defense, in particular with a role in SA regulation. To test this, we infected Col-0, flk-1, and flk-5 plants with the virulent P. syringae pv. maculicola ES4326 strain DG3 (PmaDG3). We found that both flk mutants were more susceptible than Col-0 and this susceptibility was rescued by FLK-GFP, as demonstrated in two representative rescued lines #7 and #20 (Figure 2A). The flk mutants also had lower SA accumulation and lower expression of the SA marker gene PR1 than Col-0 with PmaDG3 infection (Figure 2, B and C), supporting the SA regulatory role of FLK.

Figure 2

FLK positively regulates P. syringae resistance, salicylic acid (SA) accumulation, and PTI. A, Bacterial growth with PmaDG3 infection (n = 6). B, Salicylic acid (SA) quantification with PmaDG3 infection (n = 2). C, Expression of PR1 with PmaDG3 infection (n = 3). D, ROS burst in seedlings treated with 1 µM flg22 (left) and 1 µM elf26 (right) (n = 12). RLU: relative luminescence unit. E, Images of callose deposition. The fourth to seventh leaves of 25-d old plants were infiltrated with 1 µM flg22 for 24 h followed by callose staining, using 0.01% aniline blue. F, Root growth inhibition with 1 µM flg22 treatment. 4-d old seedlings were treated with 1 µM flg22 and measured for root length 4 d post treatment. The flg22-insensitive ecotype Ws that contains a loss-of-function mutation in the FLS2 gene was included as a control. The fold difference is the ratio of flg22-treated and water-treated root length of each genotype. Data represent the average of three independent experiments. Error bars represent mean ± Sem in A, D, and F and mean ± STDV in panel B and C. Different letters indicate significant difference among the samples of the same time point (P < 0.05; One-way ANOVA with post hoc Tukey's HSD test). These experiments were repeated two times with similar results.

Plant perception of biotrophic pathogens is often associated with activation of PTI, a basal level of defense (Zipfel et al., 2004). Flg22, a 22-aa peptide from the conserved region of the flagellin protein of P. syringae, is widely used to elicit PTI, including rapid oxidative burst, cell wall strengthening via callose deposition, and reduced plant growth. Within minutes of flg22 treatment, Col-0 and the two rescued lines activated an oxidative burst, producing a high level of ROS (Figure 2D, left). The flk mutations compromised such a ROS burst. This phenotype was further corroborated with another PAMP molecule, elf26, a 26-aa peptide from the elongation factor Tu protein (Kunze et al., 2004) (Figure 2D, right). We also found that the flk mutants displayed fewer callose depositions and less seedling growth inhibition upon flg22 treatment (Figure 2, E and F). These flk-conferred PTI phenotypes were rescued by FLK-GFP (Figure 2, D–F). Together our data suggest that FLK positively regulates P. syringae resistance through activating SA signaling and PTI.

While SA is important for P. syringae resistance, SA could antagonize other signaling pathways to suppress plant defense against necrotrophic pathogens (Glazebrook, 2005; Campos et al., 2014). To test if this is a possibility in FLK-mediated defense, we infected the plants with the necrotrophic fungal pathogen B. cinerea. Compared with Col-0 and the two rescued lines, the flk mutants were more resistant to B. cinerea (Figure 3). Thus, this result supports a negative role of FLK in B. cinerea defense, likely through SA cross-talking with other defense signaling.

Figure 3

FLK negatively regulates Botrytis resistance. A, Pictures of B. cinerea-infected plants. B, Disease symptom scoring with B. cinerea infection. The rating scale is from 1 (bottom; no lesion or small, rare lesions) to 5 (top; lesions on over 70% of a leaf). Error bars represent mean ± Sem (n = 30). Different letters indicate significant difference among the samples in the same rating group (P < 0.05; One-way ANOVA with post hoc Tukey's HSD test). These experiments were repeated two times with similar results.

RNA-seq analysis supports the roles of FLK in defense and development regulation

To elucidate how FLK is mechanistically linked to plant defense and development, we performed RNA-seq analysis. We generated transcriptome profiles of Col-0, flk-1, flk-5, acd6-1, acd6-1 flk-1, and acd6-1 flk-5, using the Illumina NovaSeq 6000 platform. Principal component analysis of the sequencing data after removing low-quality reads showed an association of gene expression profiles of the replicates in each genotype (Figure 4A). To identify FLK-affected genes, we compared four groups: (a) Col-0 vs. flk-1; (b) Col-0 vs. flk-5; (c) acd6-1 vs. acd6-1 flk-1; and (d) acd6-1 vs. acd6-1 flk-5. We found that there was a total of 8,083 differentially expressed genes (DEGs) in at least one comparison group (adjusted p-value < 0.05) (Supplemental Dataset S1). Consistent with the known role of FLK in regulating development, Gene Ontology (GO) analyses showed that the DEGs were significantly enriched in those involved in primary metabolism (GO:0009415 and GO:0010243) and development-related processes (GO:0009639 and GO:0048511) (Figure 4B). The flowering repressor gene FLC is among the development-related DEGs affected by the flk mutations. Expression of FLC was induced by flk mutations, which was validated by reverse transcription quantitative PCR (RT-qPCR) analysis (Supplemental Figure S2).

Figure 4

RNA-seq analysis reveals genes differentially affected by the flk mutations. A, Principal component analysis. B, Gene Ontology (GO) analysis of biological processes differentially affected by the flk mutations. These GO categories were significantly different in all four comparison groups: (a) Col-0 vs. flk-1; (b) Col-0 vs. flk-5; (c) acd6-1 vs. acd6-1 flk-1; and (d) acd6-1 vs. acd6-1 flk-5. C, Heatmap analysis of defense genes differentially affected by the flk mutations. D, Suppression of major defense genes by the flk mutations in acd6-1. The GO category for salicylic acid (SA) is GO:0009751 (response to SA), for jasmonic acid (JA) GO:0009753 (response to JA), for ethylene (ET) GO:0009723 (response to ET), for programmed cell death (PCD) GO:0008219 (cell death), and for PAMP-triggered immunity (PTI) GO:0007166 (cell surface receptor signaling pathway).

In addition, the DEGs affected by the flk mutations are enriched in those related to biotic and abiotic responses (GO:0009620 [response to fungus], GO:0009617 [response to bacterium], GO:0006979 [response to oxidative stress], GO:0010200 [response to chitin], GO:0009409 [response to cold], and GO:0009611 [response to wounding]) (Figure 4B). A total of 955 defense genes showed differential expression in at least one comparison group (Figure 4C and Supplemental Dataset S1). Defense genes affected by flk mutants are from multiple signal pathways involving SA, JA, ethylene (ET), programmed cell death (PCD), and/or PTI (Figure 4, C and D). Expression of most of these defense genes was suppressed by the flk mutations in the acd6-1 background. For instance, several major SA pathway genes, including the major SA biosynthesis genes (ICS1/EDS16 and WIN3/AVRPPHB SUSCEPTIBLE 3 [Wildermuth et al., 2001; Lee et al., 2007; Rekhter et al., 2019], SA regulatory genes (EDS1, PAD4, SENESCENCE-ASSOCIATED GENE101, CALMODULIN BINDING PROTEIN 60 g, and SYSTEMIC ACQUIRED RESISTANCE DEFICIENT 1) (Feys et al., 2001, 2005; Wang et al., 2009; Zhang et al., 2010c; Wang et al., 2011b), and SA receptors (NPR1 and NPR3) (Fu et al., 2012), showed lower expression in the acd6-1 flk mutants than in acd6-1 (Figure 4D and Supplemental Dataset S1). This lower expression of SA pathway genes in the acd6-1 flk mutants is consistent with the suppression of acd6-1-conferred phenotypes in these plants, lending strong support for the role of FLK in SA regulation.

In the JA pathway, we found that a number of genes involved in JA biosynthesis showed reduced expression in acd6-1 flk-1 and acd6-1 flk-5, compared with in acd6-1 (Figure 4D and Supplemental Dataset S1). These JA genes include the LIPOXYGENASE genes (Vellosillo et al., 2007; Seltmann et al., 2010), ALLENE OXIDE CYCLASE (Hofmann and Pollmann, 2008), and the 12-oxophytodienoate reductase genes (Schaller et al., 2000; Stintzi and Browse, 2000). No expressional changes were observed for the JA receptor gene CORONATINE INSENSITIVE 1, the key downstream signal transducer (transcription factor MYELOCYTOMATOSIS [MYC2]), and several MYC2 target genes, including NAC DOMAIN CONTAINING PROTEIN 19, ANAC055, and RESPONSIVE TO DESICCATION 26 (Bu et al., 2008; Zheng et al., 2012). However, we found that expression of several JASMONATE ZIM-DOMAIN genes (Vanholme et al., 2007) was suppressed while the JA signaling marker gene PLANT DEFENSIN 1.2 (PDF1.2) was highly induced in the acd6-1 flk mutants compared with in acd6-1. Expression of PDF1.2 was validated by RT-qPCR (Supplemental Figure S2). It is possible that at least some branches of JA signaling are affected by the flk mutations.

Additionally, we observed a down-regulation of key PTI-related genes in the acd6-1 flk mutants compared with in acd6-1. These PTI genes include the pattern recognition receptors (FLAGELLIN-SENSING2 and ELONGATION FACTOR-TU RECEPTOR), the coreceptor BRASSINOSTEROID INSENSITIVE 1-ASSOCIATED RECEPTOR KINASE [Boller and Felix, 2009], PTI signal transducers (MITOGEN-ACTIVATED PROTEIN KINASE 3, MITOGEN-ACTIVATED PROTEIN KINASE KINASE 5, and a number of WRKY genes [e.g. WRKY22, WRKY 29, and WRKY 33]) (Asai et al., 2002; Hsu et al., 2013), and members of the Arabidopsis Tóxicos en Levadura gene family that were shown early responses to PTI (Libault et al., 2007; Lin et al., 2008). These data strongly support the role of FLK in PTI regulation.

FLK plays a role in alternative splicing

Some KH-domain proteins were shown to function in RNA metabolism (Nicastro et al., 2015). For instance, two of the most similar human homologs of FLK, the POLY(C)-BINDING PROTEIN and NOVA ALTERNATIVE SPLICING REGULATOR 1 proteins, were shown to regulate RNA alternative splicing (Zhang et al., 2010a, 2010b). To test a potential role of FLK in RNA alternative splicing, we analyzed the differential alternative splicing (DAS) profile, using the RNA-seq data. We found that there was a total of 867 DAS genes significantly influenced by the flk mutations in at least one of these comparison groups: (a) Col-0 vs. flk-1; (b) Col-0 vs. flk-5; (c) acd6-1 vs. acd6-1 flk-1; and (d) acd6-1 vs. acd6-1 flk-5 (Supplemental Dataset S2). There was an enrichment of genes involved in development and defense in the DAS genes (Supplemental Figure S3 and Supplemental Dataset S2). Among these DAS genes, 219 also showed differential expression influenced by the flk mutations (Supplemental Dataset S3).

One of the DAS genes identified in our bioinformatics analysis is ACD6. There are six major isoforms for the ACD6 transcript, according to Arabidopsis Thaliana Reference Transcript Dataset 2 (Zhang et al., 2017) (Supplemental Figure S4A). DAS analysis indicated that the At4g14400_ID4 isoform, which includes the third intron, was significantly induced by the flk mutations in the acd6-1 background (Figure 5A). We confirmed this result by RT-PCR, which showed a clear increase in the At4g14400_ID4 isoform in the acd6-1 flk mutants compared with in acd6-1 (Supplemental Figure S4B, primer set 1). In addition, we detected an extra band with primer set 2, which flanked intron 4 of the gene, the level of which was higher in the acd6-1 flk mutants than in acd6-1 (Supplemental Figure S4B, primer set 2). The size of this band is the same as the amplicon from the genomic DNA template, suggesting the retention of intron 4 due to the flk mutations. The amplicons corresponding to intron 3 or intron 4 retention were further confirmed by sequencing. No difference in ACD6 isoforms was observed with primer sets 3 and 4 (Supplemental Figure S4B).

Figure 5

FLK regulates alternative splicing of ACD6 transcripts. A, Relative abundance of ACD6 isoforms in different genotypes from 3D RNA-seq analysis. Colors indicate six major isoforms of ACD6 transcript, according to Arabidopsis Thaliana Reference Transcript Dataset 2. B, RT-qPCR analysis of expression of intron 3 and intron 4 using intron specific primers. Error bars represent mean ± Sem (n = 3). Different letters indicate significant difference among the samples (P < 0.05; One-way ANOVA).

While it can aid the visualization of ACD6 isoforms with different primer sets, RT-PCR is less reliable in accurately quantifying the abundance of specific isoforms. Therefore, we conducted RT-qPCR to measure the levels of intron-containing isoforms using intron specific primers (Supplemental Table S1). Figure 5B showed significantly higher levels of intron 3 and intron 4 expression in the acd6-1 flk mutants than in acd6-1. Retention of intron 3 and 4 in acd6-1flk mutants should result in early translation terminations (Supplemental Figure S5A). While the wild-type ACD6 protein is predicted to have nine ankyrin repeats and five transmembrane helix, the predicted ACD6 isoforms with intron 3 or with intron 4 are truncated and only have four and nine ankyrin repeats, respectively (Supplemental Figure S5B). Such protein truncates likely lead to reduced or null function of the encoded proteins.

We further observed that ACD6 expression was much higher in the acd6-1 background than in the Col-0 background (Supplemental Figure S2 and Supplemental Dataset S1), consistent with prior reports that demonstrated acd6-1 is a hypermorphic allele (Lu et al., 2003). Interestingly, ACD6 transcript abundance was comparable in acd6-1 flk mutants and acd6-1 (Supplemental Figure S2 and Supplemental Dataset S1). Thus, increased retention of intron 3 and intron 4 will likely reduce the overall function of ACD6 in the acd6-1 flk mutants, contributing to flk suppression of acd6-1-conferred phenotypes and defense. Together, these results lend strong support for the role of FLK in targeting ACD6 for alternative splicing, accounting at least partly for FLK function in regulating defense and acd6-1-conferred phenotypes.

Multiple SA regulators contribute to the regulation of acd6-1-conferred phenotypes by FLK

In addition to the flk mutations, defects in a number of SA genes were previously reported to suppress acd6-1-conferred phenotypes (Ng et al., 2011; Wang et al., 2011a). Because our RNA-seq analysis showed that expression of several major SA genes, including two SA biosynthesis genes (ICS1 and WIN3), an SA regulatory gene (PAD4), and an SA receptor gene (NPR1), was suppressed in the acd6-1 flk mutants, compared with in acd6-1 (Figure 4D and Supplemental Data S1), we hypothesized that these SA genes act downstream of FLK to affect FLK-mediated defense and development. To test this hypothesis, we crossed the flk-1 mutant with mutants impaired in these SA genes, including ics1-1, win3-1, pad4-1, and npr1-1, in the acd6-1 background. Like acd6-1 flk-1, all the double mutants, acd6-1 ics1-1, acd6-1 win3-1, acd6-1 pad4-1, and acd6-1 npr1-1, are larger than acd6-1 and showed reduced cell death relative to acd6-1 (Figure 6, A and B and Ng et al., 2011; Wang et al., 2011a). We obtained higher-order mutant combinations and analyzed these plants for acd6-1-conferred phenotypes, including plant size, cell death, and SA level. We made inferences on the interactions of FLK and an SA gene on the basis of these phenotypes displayed by the higher-order mutants. A nonadditive suppression of these phenotypes would suggest that FLK and the SA gene act in the same pathway, whereas an additive suppression would suggest that FLK acts partially through or independently of the SA gene. Similar analyses have been successfully used to interrogate the genetic interactions among several defense genes (Song et al., 2004; Ng et al., 2011; Wang et al., 2011a; Wang et al., 2014; Hamdoun et al., 2016; Zhang et al., 2019).

Figure 6

Genetic analysis of the interactions between flk-1 and major SA mutants in acd6-1. A, Pictures of 25-d old plants. The size bar represents 1 cm and is applied to all panels. B, Cell death staining. The size bar represents 0.5 mm and is applied to all panels. C, Total salicylic acid (SA) measurement by HPLC. D, Flowering time. Plants were grown in a light cycle of 16 h L/8 h D and recorded for flowering time, days after planting for the first visible appearance of inflorescence of a plant. Error bars represent mean ± Sem in C, (n = 4) and D, (n ≥ 18). Different letters indicate significant difference among the samples (P < 0.05; One-way ANOVA with post hoc Tukey's HSD test). Data for Col-0 and flk-1 were shown in Figure 1. These experiments were repeated two times with similar results.

For plant size and cell death, we found that all triple mutants containing acd6-1 and flk-1 showed much larger plant stature and reduced cell death than the corresponding double mutants (Figure 6, A and 6B). These results suggest that FLK acts partially through each of these SA pathway genes, ICS1, PAD4, WIN3, or NPR1, in regulating acd6-1-conferred dwarfism and cell death.

We further measured the total SA level in these plants (Figure 6C). Consistent with the roles of ICS1 and WIN3 as major SA synthesis proteins, there were residual SA levels in acd6-1 ics1-1 and acd6-1 win3-1 with or without the presence of flk-1. While pad4-1 suppressed SA accumulation in acd6-1, the SA receptor mutant npr1-1 led to a high SA level in acd6-1 (Figure 6C and Ng et al., 2011). Further reduced SA levels were found in acd6-1 flk-1 pad4-1 and acd6-1 flk-1 npr1-1 compared with in the corresponding double mutants, suggesting that FLK acts partially through PAD4 and NPR1 in SA regulation.

FLK regulation of flowering is largely independent of Sa

We were interested in elucidating how FLK-mediated flowering is related to its role in SA regulation. We quantified the flowering time under the same light condition (12 h light [L]/12 h dark [D]) used for SA measurement and other defense assays. Although they had drastic differences in SA levels, the acd6-1 flk double and flk single mutants showed similar late flowering phenotypes (Figure 1D). The late flowering of acd6-1flk-1 was not affected when additional SA deficient mutants, ics1-1 or pad4-1, were introduced (Figure 6, C and D). Thus, we concluded that FLK-mediated flowering regulation is largely independent of SA under a 12 h L/12 h D cycle. In fact, Col-0 and acd6-1 flowered at a similar time although they accumulated drastically different levels of SA (Figure 1 and Wang et al., 2011a). Thus, we believe that the effect of SA on flowering time regulation is minor under our condition.

We previously showed that npr1-1 and win3-1 work additively to stimulate early flowering and suppress acd6-1-conferred phenotypes (Figure 6 and Wang et al., 2011a). Consistent with this observation, we found here that npr1-1 and win3-1 individually suppressed late flowering of flk-1. Together in the quadruple mutant acd6-1 flk-1 npr1-1 win3-1, they brought the flowering time to the wild-type level and completely suppressed acd6-1-conferred small stature, cell death, and high SA level (Figure 6). However, the early flowering of npr1-1 is likely due to a second site mutation in the background (Dong XN, personal communication). Together, these results suggest that WIN3, NPR1, and an unknown protein(s) act downstream of FLK to contribute to the regulation of flowering in addition to acd6-1-conferred phenotypes, including SA accumulation, cell death, and plant size.

Discussion

The FLK protein is known as a positive regulator of flowering time. We report here a role of FLK in pathogen defense. Our data showed that loss of function in FLK conferred enhanced disease susceptibility to the biotroph P. syringae and enhanced disease resistance to the necrotroph B. cinerea. The RNA-seq analysis supports that FLK acts through regulating expression and alternative splicing of target genes in defense and development. One such FLK target gene is ACD6. Thus, this study provides mechanistic links of FLK function in regulating acd6-1-conferred phenotypes as well as plant defense and development.

FLK encodes a putative protein with three KH motifs. The KH motif is a highly conserved RNA binding motif found in proteins from diverse organisms. KH-domain proteins were shown to affect RNA metabolism, such as pre-mRNA processing and mRNA stability (Nicastro et al., 2015). Our RNA-seq analysis revealed that many genes involved in development and defense were differentially expressed in the flk mutants compared with in control plants (Figure 4). In addition, the flk mutations affected alternative splicing of some gene transcripts, including ACD6. These observations support the biochemical function of FLK in affecting gene transcript levels as well as pre-mRNA processing.

We previously reported that acd6-1 harbored a point mutation, causing a Leu-to-Phe substitution at position 591 of the ACD6 protein in a predicted transmembrane helix (Lu et al., 2003, 2005). This mutation makes ACD6-1 a gain-of-function and hypermorphic protein. Because acd6-1 expresses high levels of ACD6-1 and other defense genes and displays drastic defense-associated phenotypes as well as extreme dwarfism, acd6-1 provides a sensitive background to detect the changes in gene expression and physiological phenotypes, which may not be otherwise detected in the Col-0 background. Indeed, our data showed that a lack of FLK expression led to increased retention of intron 3 and 4 in ACD6 transcripts in the acd6-1 background, however, there was less of such an increase in the Col-0 background (Figure 5). ACD6-1 transcripts containing intron 3 and/or intron 4 should have translation terminations before the transmembrane domain (Supplemental Figure S5), likely leading to reduced or null function of the encoded proteins. Therefore, although the abundance of ACD6-1 transcripts remains similarly high in acd6-1 and acd6-1 flk mutants (Supplemental Figure S2 and Supplemental Dataset S1), the increased amount of aberrant splicing of ACD6 transcripts in acd6-1 flk mutants could drastically reduce the function of ACD6-1, leading to a partial suppression of acd6-1-conferred phenotypes by the flk mutations.

Suppression of acd6-1-conferred phenotypes by flk could also be due to FLK regulation of other defense genes. Indeed, our RNA-seq analysis showed that FLK influenced expression of a large number of defense genes, including those involved in SA, JA, ET, PCD, and PTI. In particular, a number of major SA pathway genes involved in SA biosynthesis, SA regulation, and SA perception were down-regulated in the acd6-1 flk mutants than in acd6-1 (Figure 4D and Supplemental Data S1). Because acd6-1-conferred phenotypes are sensitized to changes in SA levels and/or signaling, such lower expression of the SA pathway genes could also contribute to the partial suppression of acd6-1-conferred dwarfism, cell death, and high SA accumulation in the flk mutant, in addition to the influence of FLK on ACD6 alternative splicing. Consistent with multiple downstream defense targets of FLK, our genetic analysis showed that FLK acts only partially through each of the SA pathway genes, ICS1, PAD4, WIN3, and/or NPR1, in influencing acd6-1-conferred plant size and cell death (Figure 6). While requiring the major SA biosynthesis genes ICS1 and WIN3 for SA accumulation, FLK-mediated SA accumulation is partially through the known SA regulator PAD4. These data further suggest that expression of these SA genes is likely regulated by additional factors in addition to FLK. This is because the full effect of these genes can only be detected when they are completely knocked out, leading to enhanced flk suppression of acd6-1-conferred phenotypes (Figure 6).

Consistent with FLK regulation of ACD6, SA regulatory genes, and those involved in additional defense signaling pathways, the flk mutations resulted in lower resistance to the biotroph PmaDG3 and suppressed SA accumulation induced by PmaDG3 infection (Figures 1 and 2). While S, A is a key signaling molecule critical for defense against biotrophs, SA could adversely affect plant resistance against necrotrophs through cross-talking with other defense pathways, such as those mediated by JA. Indeed, our data showed an increased resistance of the flk mutants to the necrotrophic pathogen B. cinerea, associating with decreased expression of JA signaling repressor genes and increased expression of the JA signaling marker gene PDF1.2 compared with expression in Col-0 (Figure 3 and Supplemental Figure S2). Thus, our data support the crosstalk between SA and JA signaling.

Previous studies showed a positive role of SA in influencing flowering time. Plants with lower SA levels, the NahG transgenic plant or the SA mutants ics1 and eds5, flowered later than Col-0 (Martinez et al., 2004). It is worth noting that the effect of SA deficiency on flowering time in these plants appeared to be more evident in a short-day photoperiodic condition of 8 h L and 16 h D. In congruent with our defense assays, which were conducted under 12 h L or 16 h L in a 24 h cycle, we quantified flowering time with the same light regime. Under these conditions, we found that Col-0 and acd6-1 had similar flowering time although there was a drastic difference in the SA levels. Thus, we believe that the effect of SA on flowering time regulation is minor with 12 h or more L in a day (Figures 1D and 6D and Wang et al., 2011a). Consistent with this note, our additional data failed to reveal a strong association of SA levels and FLK-mediated flowering regulation. For instance, these plant pairs, e.g. acd6-1 flk-1/-5 vs. flk-1/-5 and acd6-1 flk-1 vs. acd6-1 ics1-1 flk-1, had large differences in SA levels yet they all showed flk-like late flowering (Figures 1 and 6). Thus, we concluded that, FLK-mediated flowering regulation is largely independent of SA under our experimental conditions. However, we do not exclude the possibility that the effect of SA mutants on flk-conferred late flowering could be detected under different light conditions, e.g. 8 h L/16 h D cycle, or with different light intensities. Because we are interested in FLK's role in regulating development and defense, flowering time quantification under a different light condition would require additional experiments to test the defense role of FLK under the same condition, e.g. FLK's regulation of SA accumulation and pathogen infections.

Of the SA mutants we tested, only npr1-1 and win3-1 showed contributions to FLK's role in flowering regulation. Because the flowering phenotype in the npr1-1 mutant is likely due to a second site mutation, a further identification of this mutation should shed light on the molecular pathway(s) mediated by FLK in flowering control. WIN3 encodes a GH3 acyl adenylase-family enzyme that conjugates glutamate to isochorismate to produce isochorismate-9-glutamate, a precursor of SA (Rekhter et al., 2019; Torrens-Spence et al., 2019). The lack of WIN3 enzymatic activity might produce an altered metabolic profile; one or more such metabolites could contribute to altered defense and flowering in the win3-1 mutant. Our previous data suggest that these two roles of WIN3 are independent of each other (Wang et al., 2011a). WIN3 suppresses expression of the flowering activator gene FT and promotes the repressor gene FLC. This action of WIN3 is opposite to that of FLK in regulating expression of these two genes (Lim et al., 2004; Mockler et al., 2004), likely forming the basis of win3 suppression of flk in flowering control.

How does FLK execute its function in regulating various downstream pathways to affect defense and development of plants? The KH-domain proteins can function through a direct binding to and/or an association in protein complexes with RNA molecules for processing (Nicastro et al., 2015). It is possible that FLK directly binds to its target gene transcripts to influence their mRNA stability and/or alternative splicing. Alternatively, FLK could form protein complexes with other proteins to regulate pre-mRNA processing. Indeed, the FLK protein was previously shown to physically interact with several proteins, including three RNA binding proteins (PEPPER 1, HUA ENHANCER 4, and HUA1) (Cheng et al., 2003; Ripoll et al., 2009; Rodriguez-Cazorla et al., 2015), and an E3 ubiquitin ligase (HIGH EXPRESSION OF OSMOTICALLY RESPONSIVE GENES 1) (Lee et al., 2012). A recent study further showed that FLK and some of the FLK interactor proteins were copurified with another flowering regulator FPA (Parker et al., 2021), suggesting that these proteins might form protein complexes. The FLK interactor proteins, PEP1, HEN4, and HUA1, were shown to be involved in alternative splicing of certain target genes (Ripoll et al., 2009; Hornyik et al., 2010). Although not known to affect alternative splicing of gene transcripts, FPA was demonstrated to have a role in promoting 3′ end poly(A) site selection of gene transcripts in Arabidopsis (Parker et al., 2021). However, none of these FLK-associated RNA binding proteins have been shown to directly bind to their potential transcript targets. More research will be necessary to reveal direct transcriptional regulation of target genes of FLK and its associated proteins, filling this knowledge gap.

While the detailed biochemical mechanism by which FLK acts to regulate RNA metabolism remains to be elucidated, this report strongly supports the multifunctionality of FLK. It is possible that FLK multifunctionality is through its interaction with distinct protein partners to target downstream genes for their RNA levels and/or pre-mRNA processing. Such a multifunctionality of FLK could be decoupled by various downstream genes with distinct roles in defense and/or development regulation. Thus, it is critically important to uncover direct gene targets of FLK to further advance our understanding of the mechanistic actions of FLK. Some FLK pathway genes are potentially powerful molecular tools that can be used to develop novel biotechnological strategies to precisely control crop traits, maximizing plant growth while maintaining proper responses to environmental assaults.

Materials and methods

Plant materials

Arabidopsis (A. thaliana) plants were grown in growth chambers with 180 µmol m⁻² s⁻¹ photon density, 60% humidity, and 22°C, and a photoperiod of 12 h L/12 h D unless otherwise indicated. The acd6-1 flk-5 mutant was isolated from the acd6-1 suppressor screen (Lu et al., 2009) and was backcrossed to Col-0 to generate the flk-5 mutant. The flk-1 mutant (SALK_007750) was previously described (Mockler et al., 2004). The acd6-1 flk-1 mutant was generated by crossing the two single mutants and selecting the homozygous double mutant in the F₂ generation, using relevant primers. The triple mutants of acd6-1 and flk-1 in combination with SA mutants, including ics1-1, pad4-1, npr1-1, and win3-1, were created by crossing the relevant mutants together. The triple mutants acd6-1 flk-1 npr1-1 and acd6-1 flk-1 win3-1 were previously described (Wang et al., 2011a) and were used to generate the quadruple mutant acd6-1 flk-1 npr1-1 win3-1. Primers for mutant allele detection were previously described (Ng et al., 2011; Wang et al., 2011a) or are listed in Supplemental Table S1.

Generation of transgenic complementation lines

A 5.9-kb fragment consisting of 2.4 kb FLK promoter and the 3.5 kb full-length FLK genomic fragment was PCR amplified, using primers xz798s_FLKproNotIF (ggg ccc gcg gcc gcA AGT TTC CAA CAC CAA GAA GCC A) and xz801s_FLKnonstopSalI (ggg ccc gtc gac GTA ACC GTA GCC TGA GCT GTA). The fragment was cloned in the pGlobug vector (Zhang and Mount, 2009) in frame at the 3′ end with the reporter gene GFP. The pFLK-FLK-GFP-NOS fragment was subcloned into the binary vector pMLBart, using the Not I site. This construct was designated as FLK-GFP and was used to transform the flk-1 mutant, using Agrobacterium-mediated plant transformation. flk-1 was chosen because both FLK-GFP and flk-5 carried the BASTA gene, preventing herbicide selection for the transformants. Homozygous transgenic plants that are independent transformants were obtained from the T₂ generation.

Pathogen infection

Pseudomonas syringae pv. maculicola ES4326 strain DG3 (PmaDG3) was used to infect plants as previously described (Zhang et al., 2013). The fourth to seventh leaves of 25-d old plants were infiltrated with a PmaDG3 solution, using a needleless syringe. Discs of infected leaves were produced with a biopsy puncher of 4 mm diameter, homogenized in 10 mM MgSO₄, and serially diluted. The dilutions were plated on King's Broth agar media with 50 mg L⁻¹ kanamycin for bacterial growth at 30°C.

Botrytis cinerea strain BO5-10 was kindly provided by Tesfaye Mengiste (Purdue University). A B. cinerea solution of 2 × 10⁵ spores mL⁻¹ was evenly sprayed onto plants for infection. Plants were covered for disease symptom development. Disease symptoms of the fourth to seventh leaves of each plant were scored three days post infection with the 1–5 scale as described (Wang et al., 2011a): 1 = no lesion or small, rare lesions; 2 = lesions on 10%–30% of a leaf; 3 = lesions on 30%–750% of a leaf; 4 = lesions on 50%–70% of a leaf; 5 = lesions on over 70% of a leaf.

SA quantification

SA extraction and quantification were performed as previously described (Wang et al., 2011a). For pathogen-induced SA accumulation, the fourth to seventh leaves of 25-d old plants were infiltrated with 1 × 10⁷ cfu mL⁻¹ PmaDG3 and were harvested at the indicated times for total SA extraction. For non-infected plants, the whole plants were collected for total SA extraction.

Gene expression analyses

Total RNA was extracted from the fourth to seventh leaves of 25-d old plants using TRIzol reagent (Invitrogen) per the manufacturer's instructions. After the removal of genomic DNA, total RNA was reverse-transcribed, using a cDNA synthesis kit (ThermoFisher Scientific). Maxima SYBR Green/ROX qPCR Master Mix (ThermoFisher Scientific) was used for qPCR reactions. PDF2 (AT1G13320) was used as an internal control for RT-qPCR experiments (Czechowski et al., 2005). Amplicons from RT-PCR were sampled at 30 cycles. DNA electrophoresis was used to separate PCR amplicons of ACD6 isoforms. Primers for RT-qPCR and RT-PCR are listed in Supplemental Table S1.

Luminol assay

Leaf discs of 4 mm diameter were isolated from the fourth to seventh leaves of 25-d old plants and floated atop 100 µl sterile water in a 96-well plate. The plate was covered with tin foil and placed in a growth chamber overnight to measure the PAMP-elicited ROS burst and for 2 h for basal ROS measurement. A 100 µl solution containing 300 µM L-012 (Wako Chemicals USA Inc. Richmond, VA) in 10 mM MOPS buffer (pH 7.4) and 1 µM flg22 or 1 µM elf26 was added to replace sterile water in each well and luminescence was recorded immediately using a Modulus II Microplate Reader.

Cell death staining

Trypan blue staining was performed to visualize cell death as previously described (Ng et al., 2011). Briefly, the fourth to seventh leaves of 25-d old plants were harvested and boiled in lactophenol (phenol: glycerol: lactic acid: water = 1:1:1:1; v/v) containing 0.01% trypan blue for 2 min. The stained leaves were boiled in alcoholic lactophenol (95% ethanol: lactophenol = 2:1 [v/v]) for 2 min, rinsed in 50% ethanol (v/v) for 1 min, and finally cleared in 2.5 g/ml chloral hydrate for 2 min. At least six leaves of each genotype were stained and visualized with a Leica M80 stereomicroscope. Cell death images were captured with a Leica IC80 HD camera connected to the microscope.

Callose staining assay

The fourth to seventh leaves of 25-d old plants were infiltrated with 1 µM flg22 or sterile water. Leaves were harvested 24 h post infiltration and boiled in alcoholic lactophenol (2:1 95% ethanol: lactophenol [v/v]) for 2 min followed by rinsing in 50% ethanol (v/v) for 1 min. Cleared leaves were stained with aniline blue solution (0.01% aniline blue (w/v) in 150 mM phosphate buffer, pH 9.5) for 90 min in darkness. Callose deposition was visualized with a Leica M205FA fluorescence stereomicroscope and photographed with an Amscope CMOS digital camera.

Root growth inhibition assay

Sterilized seeds were plated on agar plates containing ½ MS media supplemented with 1% sucrose (w/v) (pH 5.7), stratified at 4° C for 2 days, and then transferred to a tissue culture chamber to grow for 4 days. Seedlings of relatively uniform size were transferred to a 24-well tissue culture plate with 1.5 ml 1 µM flg22, or sterile water. A minimum of four seedlings per genotype per treatment were measured for root length 4 days post treatment and imaged with a Canon camera.

Flowering time determination

Flowering time was determined by counting the number of days post planting until the appearance of the first flower bud in plants. The light cycles were either long-day (16 h L/8 h D) or short-day (12 h L/12 h D). All plants used in this study for flowering time determination showed similar germination time.

RNA-seq analysis

Total RNA was extracted from 25-d old Col-0, flk-1, flk-5, acd6-1, acd6-1 flk-1, and acd6-1 flk-5 plants. A total amount of 1 μg RNA per sample was used to generate cDNA libraries, using NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB, USA) following manufacturer's recommendations. Deep sequencing was performed using Illumina NovaSeq 6000 (Novogene Corporation Inc.). Triplicate biological samples were used for all genotypes except acd6-1 flk-5, which had duplicate samples because one sample failed to pass quality control. The samples were multiplexed and sequenced with the standard paired-end sequencing that has a read length of 150 bp per end and 20 M reads per end per sample. The raw reads in FASTQ format were filtered by removing reads containing adapters and reads of low quality and mapped to the Arabidopsis genome (TAIR 10). For global gene expression profiling, a relative expression value (Reads Per Kilobase of transcript per Million mapped reads [RPKM]) higher than 0.3 was used as the cutoff to include genes for further analyses. Each RPKM value was corrected by adding the number one and then was log₂-transformed. Differentially expressed genes were identified in four comparison groups: (a) Col-0 vs. flk-1; (b) Col-0 vs. flk-5; (c) acd6-1 vs. acd6-1 flk-1; and (d) acd6-1 vs. acd6-1 flk-5, using the R package DESeq2 with default parameters (Anders and Huber, 2010). Genes with an adjusted P-value <0.05 found by DESeq2 were assigned as differentially expressed. Gene Ontology (GO) enrichment analysis of DEGs was implemented by the clusterProfiler R package, in which gene length bias was corrected. GO analysis was also independently verified, using the online tool PANTHER (Mi et al., 2019). Graphical representation of gene expression correlation was analyzed, using heatmap.2 function. DEGs were further corroborated via an additional independent analysis, using the webtool 3D RNA-seq (Calixto et al., 2018; Guo et al., 2021).

3D RNA-seq was further used to analyze differential alternative splicing (DAS) profiles of genes affected by the flk mutations. For DAS genes, the log₂ fold change (L₂FC) of each individual transcript was compared to the weighted average of L₂FCs (weights were based on their standard deviation) of all remaining transcripts in the same gene. A gene was considered a significant DAS gene in a contrast group if any of its transcripts had a Δ Percent Spliced (ΔPS) ratio ≥ 0.1 with an adjusted P-value < 0.05 (Calixto et al., 2018).

Statistical analyses

Statistical analyses were performed using Prism 8 (GraphPad Software, LLC.). Details of experimental design, the number of replicates, and data presentation were indicated in the relevant figure legends and Method section.

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession number GSE164424.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Complementation of the flk-1 mutant by FLK-GFP.

Supplemental Figure S2. Regulation of gene expression by FLK.

Supplemental Figure S3. GO analysis of FLK regulated DAS genes.

Supplemental Figure S4. FLK regulates alternative splicing of ACD6 transcripts

Supplemental Figure S5. Retention of intron 3 and/or intron 4 likely result in premature termination of the encoded ACD6 protein.

Supplemental Table S1. Primers used in this report.

Supplemental Dataset S1. Genes differentially affected by the flk mutations.

Supplemental Dataset S2. DAS of genes affected by the flk mutations.

Supplemental Dataset S3. DAS and differential expression of genes affected by the flk mutations.

Acknowledgments

We thank the members in the Lu laboratory for their assistance in this work. We thank Drs. Gordon Simpson and Ashis Nandi for helpful discussions.

Funding

This work was partially supported by grants from National Science Foundation, NSF 1456140 to Hua Lu and 1923069 to Xiaoning Zhang and Hua Lu, UMBC Technology Catalyst Fund, and UMBC CENTRE Funding Initiative.

Author notes

These authors contributed equally to this research.

Present address: State Key Laboratory of Crop Stress Biology in Arid Areas, College of Horticulture, Key Laboratory of Horticultural Plant Biology and Germplasm Innovation in Northwest China, Ministry of Agriculture, Northwest A&F University, Yangling, Shaanxi 712100, China.

Present address: College of Horticulture, Henan Agricultural University, Zhengzhou 450002, China.

M.F. performed pathogen infections and PTI and ROS-related assays, measured flowering time and pathogen-induced SA accumulation, and assisted in the bioinformatics analysis. M.G. performed cell death staining, SA measurement, and gene expression analysis with the flk mutants and prepared samples for the RNA-seq experiments. J.S. assisted in the cloning of the FLK gene. X.Z. cloned the FLK-GFP construct and conducted plant transformation and confocal microscopy for FLK-GFP localization. L.V. did RT-PCR for detection of ACD6 isoforms and RT-qPCR for expression of FLC, PDF1.2, and ACD6 isoforms. S.K. measured flowering time. P.P. assisted M.F. in PTI assays. A.H. assisted in the bioinformatics analysis. H.L. designed experiments, conducted genetic crosses, analyzed gene alternative splicing, and wrote the manuscript with input from all authors.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/General-Instructions) is: Hua Lu ([email protected]).

Conflict of interest statement. None declared.