Journal of Experimental Botany, Vol. 72, No. 7 pp. 2727–2740, 2021 doi:10.1093/jxb/eraa610 Advance Access Publication 21 January 2021 RESEARCH PAPER A genome-wide association study reveals cytokinin as a major component in the root defense responses against Ralstonia solanacearum Alejandro Alonso-Díaz1, Santosh B. Satbhai2,3,*, , Roger de Pedro-Jové1, Hannah M. Berry4,5 Christian Göschl2, Cristiana T. Argueso4, , Ondrej Novak6, Wolfgang Busch2,3, , Marc Valls1,7, and Núria S. Coll1,†, 1 Centre for Research in Agricultural Genomics (CSIC-IRTA-UAB-UB), Bellaterra, Barcelona, Spain 2 Gregor Mendel Institute (GMI), Austrian Academy of Sciences, Vienna Biocenter (VBC), Dr Bohr-Gasse 3, Vienna 1030, Austria 3 Salk Institute For Biological Studies, Plant Molecular and Cellular Biology Laboratory, 10010 N Torrey Pines Rd, La Jolla, CA 92037, USA 4 Department of Bioagricultural Sciences and Pest Management, Colorado State University, Fort Collins, CO 80523, USA 5 Cell and Molecular Biology, Colorado State University, Fort Collins, CO 80523, USA 6 Laboratory of Growth Regulators, Olomouc, The Czech Republic 7 Genetics Department, University of Barcelona, Barcelona, Spain *Present address: Department of Biological Sciences, Indian Institute of Science Education and Research (IISER) Mohali, Punjab, India. † Correspondence: nuria.sanchez-coll@cragenomica.es Received 9 October 2020; Editorial decision 21 December 2020; Accepted 19 January 2021 Editor: Steven Spoel, University of Edinburgh, UK Abstract Bacterial wilt caused by the soil-borne pathogen Ralstonia solancearum is economically devastating, with no ef- fective methods to fight the disease. This pathogen invades plants through their roots and colonizes their xylem, clogging the vasculature and causing rapid wilting. Key to preventing colonization are the early defense responses triggered in the host’s root upon infection, which remain mostly unknown. Here, we have taken advantage of a high-throughput in vitro infection system to screen natural variability associated with the root growth inhibition phenotype caused by R. solanacearum in Arabidopsis during the first hours of infection. To analyze the genetic de- terminants of this trait, we have performed a genome-wide association study, identifying allelic variation at several loci related to cytokinin metabolism, including genes responsible for biosynthesis and degradation of cytokinin. Further, our data clearly demonstrate that cytokinin signaling is induced early during the infection process and cytokinin contributes to immunity against R. solanacearum. This study highlights a new role for cytokinin in root immunity, paving the way for future research that will help in understanding the mechanisms underpinning root defenses. Keywords: Bacterial wilt, cytokinin, defense, GWAS, hormones, immune system, Ralstonia solanacearum, root, salicylic acid. © The Author(s) 2021. Published by Oxford University Press on behalf of the Society for Experimental Biology. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 2728 | Alonso-Díaz et al. Introduction The majority of studies analyzing the role of cytokinin in plant defense have been performed using foliar pathogens Plant hormones are extremely important for the regulation (Clarke et al., 1998a; Pogány et al., 2004; Choi et al., 2010; of the plant defense against pathogens (Pieterse et al., 2012). Argueso et al., 2012; Naseem et al., 2012; Pieterse et al., 2012; Several studies have shown that the accumulation of salicylic Gupta et al., 2020). In contrast, the role of cytokinin in root acid (SA) induces plant defense against biotrophic pathogens defenses remains mostly unexplored. (Shigenaga and Argueso, 2016), whereas jasmonic acid (JA) and Ralstonia solanacearum is a natural soil-borne bacterial vas- ethylene are essential against necrotrophs. The crosstalk be- cular pathogen that infects many plant species, including tween JA and abscisic acid (ABA) induces plant defense against Arabidopsis, and it is the causative agent of bacterial wilt, a herbivores and insects (Pieterse et al., 2012). The synergistic disease of devastating economic impact worldwide. Ralstonia or antagonistic interaction between the different hormone solanacearum invades plants through the roots, moving centri- signaling pathways enables the plant to fine-tune defense re- petally until it reaches the xylem. Xylem colonization allows sponses to the pathogen that are effective while minimizing movement of the bacteria up into the stem, causing a rapid damage or yield penalties (Pieterse et al., 2009). and permanent obstruction of the vasculature (Hayward, 1991; Cytokinin is a plant hormone traditionally associated with Planas-Marquès et al., 2020). Many genetic tools are available plant growth and development (Mok, 1994; Sa et al., 2001; to study this pathogen, since it has been widely used as a model Wybouw and De Rybel, 2019) with an emerging role in plant species for plant–pathogen interactions in the last decades immunity. Cytokinin has been shown to participate in defense (Mansfield et al., 2012; Coll and Valls, 2013). against various plant pathogens, including fungi (Argueso Transcriptomic analyses of Arabidopsis roots infected with et al., 2012; Gupta et al., 2020), bacteria (Choi et al., 2010; the bacterial pathogen R. solanacearum show expression of Naseem et al., 2012; Pieterse et al., 2012). and viruses (Clarke cytokinin biosynthetic genes at early time points [6 hours post- et al., 1998a; Pogány et al., 2004). Furthermore, application of infection (hpi)] (Zhao et al., 2019), while cytokinin-degrading exogenous cytokinin results in an increase of callose produc- genes are expressed at later stages of infection (after 72 h) in tion in Arabidopsis thaliana (henceforth, Arabidopsis) infected Medigaco truncatula roots infected with R. solanacearum (Moreau with Pseudomonas syringae or treated with the flagellin-derived et al., 2014). Interestingly, Arabidopsis plants lacking the redun- defense elicitor flg22 (Choi et al., 2010). Tight regulation of dant negative regulator of cytokinin signaling ARR6 showed cytokinin levels is essential to determine the precise signaling increased resistance to the fungal pathogen Plectosphaerella outcome. Treatments with low concentrations of exogenous cucumerina but increased susceptibility to R. solanacearum cytokinin result in greater susceptibility to infection with the (Bacete et al., 2020). However, these phenotypes did not occur oomycete Hyaloperonospora arabidopsidis in Arabidopsis, and as a direct result of the interactions between cytokinin signaling also to infection with Blumeria graminis in wheat. In contrast, and classical hormone-based defense pathways. treatments with higher levels of cytokinin increase resistance In previous work, we set up an in vitro system to study of plants to these and other pathogens (Argueso et al., 2012; Arabidopsis early root phenotypes caused by R. solanacearum Babosha, 2009; Gupta et al., 2020). infection: root growth inhibition, root hair formation, and root Importantly, cytokinin signaling in immunity is greatly tip cell death (Lu et al., 2018). This robust method revealed intertwined with SA signaling. It has been observed that in- genetic determinants of the interaction both from the bacterial creased resistance to H. arabidopsidis induced by cytokinin virulence and plant defense sides at very early stages of infec- treatment is mediated by SA accumulation and the activation tion, which were masked in classic pathogenicity assays. Here, of SA-dependent defense genes (Choi et al., 2010; Argueso we have taken advantage of the high-throughput potential of et al., 2012). Mechanistically, it has been shown that the this in vitro system to screen the natural variation of the root ARR2 (ARABIDOPSIS RESPONSE REGULATOR 2), a growth inhibition phenotype across 430 Arabidopsis accessions major transcription factor of the cytokinin signal transduction representative of the worldwide genetic variation of this spe- pathway, physically interacts with TGA3, a transcription factor cies and determine the gene(s) responsible for this trait using from the SA signaling pathway (Choi et al., 2010). The inter- genome-wide association (GWA) mapping. action of these two transcription factors is regulated by NPR1 Thanks to the large number of Arabidopsis accessions that (NON-EXPRESSOR OF PATHOGENESIS-RELATED have been sequenced and genotyped, this model plant has great PROTEINS 1), leading to changes in PR1 expression and potential for genome-wide association study (GWAS) analyses plant immune status (Choi et al., 2010). More recently, it has (Atwell et al., 2010). Previous studies using GWAS and natural been shown that cytokinin treatment of tomato leaves induces genetic variation have detected genetic variants associated with resistance against fungi in an SA-dependent manner (Gupta resistance to abiotic stress (Bac-Molenaar et al., 2015; Kalladan et al., 2020). Although sparse, current evidence indicates that et al., 2017; Satbhai et al., 2017; Li et al., 2019), root devel- the crosstalk between cytokinin and SA signaling pathways is opment (Meijón et al., 2014), or flowering time (Aranzana very important for plant immune responses (Choi et al., 2010; et al., 2005). GWAS has also been shown to be a very powerful Argueso et al., 2012). tool to unravel genomic regions associated with the natural GWAS shows cytokinin’s role in root defense against R. solanacearum | 2729 variation of disease resistance of various plants against different desktop computer running Ubuntu Linux allowed the simultaneous con- pathogens, for example Arabidopsis against Pseudomonas syringae trol of the scanners (Slovak et al., 2014). Scans were performed with a resolution of 1200×1200 dpi, resulting in an image size of 6000×6000 (Aranzana et al., 2005; Atwell et al., 2010; Iakovidis et al., 2016), pixels (36 MP) for each of our 12×12 cm agar plates. To enhance image and Xanthomonas campestris (Huard-Chauveau et al., 2013) quality, scanning was performed in a dark room and with the scanner lid and Botrytis cinerea (Corwin et al., 2016; Thoen et al., 2017), open. or Glycine max against Fusarium virguliforme (Wen et al., 2014), among others. Importantly, a GWAS has been recently used Genome-wide association mapping to study the temperature-dependent genetic variation that We measured median and mean total root length values of 430 underscores resistance of Arabidopsis against R. solanacearum Arabidopsis accessions after R. solanacearun infection using BRAT (n=2) (Aoun et al., 2017). Finally, GWAS has also highlighted the to conduct GWA using an accelerated mixed model (EMMAX) (Kang importance of hormonal crosstalk between SA and ABA in the et al., 2010) followed by EMMA (Kang et al., 2008) for the most signifi- JA pathway involved in defense in Arabidopsis (Proietti et al., cant associations among all accessions studied The GWA was performed 2018). Taking advantage of GWAS, we have identified cyto- on a cluster, with algorithms identical to those used in the GWAPP Web interface (Seren et al., 2012). Single nucleotide polymorphisms (SNPs) kinin signaling as an important component in the root growth with minor allele counts ≥10 were considered. The significance of SNP inhibition phenotype caused by R. solanacearum in Arabidopsis, associations was determined at a 5% false discovery rate (FDR) threshold contributing to root defenses against the pathogen. computed by the Benjamini–Hochberg–Yekutieli method (Benjamini and Hochberg, 1995). Materials and methods Broad-sense heritability calculation Plant material All individuals that were measured were used to calculate the broad-sense heritability (H2=VG/VP), which is defined as the proportion of pheno- A collection of 430 A. thaliana ecotypes (Supplementary Table S1) pro- typic variation (VP) due to genetic variation (VG) (estimated from the vided by the Molecular Plant Biology Stock from the Gregor Mendel between-line phenotypic variance). Institute (Vienna, Austria) was used for GWAS. The Arabidopsis mu- tant lines used in this study have been previously described in Kiba et al. (2013) (cyp735a1 and cyp735a2) and Caesar et al. (2011) (ahk2, Gene Ontology analysis ahk3, and ahk4/cre1). Transgenic line TCSn::GFP has been described The GO-finder website (https://go.princeton.edu/) was used for Gene in Zürcher et al. (2013). The eds16 TCSn::GFP line was obtained by Ontology (GO) analysis. Genes solely ‘inferred from electronic annota- crossing the eds16 mutant (Dewdney et al., 2000) to the TCSn::GFP tion associations’ were excluded from the analysis. transgenic line and screening F2 plants for the presence of TCSn::GFP by selection on Murashige and Skoog (MS) plates supplemented with BASTA and for eds16 using PCR primers that can detect the eds16 mu- In vitro inoculation assays tation (eds16 Fwd, CCTGAGAGACTATTCCAAAGGAC; eds16 Rev, Seeds were surface sterilized with a solution containing 30% bleach and ACTCTGAAGATGGGTCACTTCC). Homozygous seeds were used 0.02% Triton X-100 for 10 min, washed five times with Milli-Q water, in all the assays. and sown (20 seeds per plate) on agar plates containing MS (Duchefa Biochemie B.V.) and 0.8% agar (Becton, Dickinson and Co.). Sown plates were stratified at 4 °C in the dark for 2 days. Plates were then transferred Plant and bacterial growth conditions for GWAS to chambers and grown vertically for 7 d under constant conditions of Seeds were surface sterilized for 2 h in open 1.5 ml Eppendorf tubes in 21–22 °C, 60% humidity, and a 16 h light/8 h dark photoperiod with a a sealed box containing chlorine gas generated from 125 ml of 10% w/v light intensity of 120 µmol m–2 s–1 during the light period. sodium hypochlorite and 3.5 ml of 37% hydrochloric acid. For strati- Ralstonia solanacearum GMI1000 was grown at 28 °C in solid or liquid fication, sterile seeds were kept at 4 °C for 72 h in the dark. After that, rich B medium (0.1% yeast extract, 1% bacto pectone, and 0.1% casamino seeds were put on agar plates containing MS (Duchefa Biochemie B.V., acids) (Becton, Dickinson and Company). For inoculation, R. solanacearum Haarlem, The Netherlands) and 0.8% agar (Becton, Dickinson and Co., GMI1000 was collected by centrifugation (1500 rcf, 5 min) from over- Franklin Lakes, NJ, USA). The placement of the seeds was guided by night liquid cultures at 28 °C, resuspended with water, and adjusted to a a printout of a seed-planting grid schematic (Supplementary Fig. S1) final OD600 of 0.001 corresponding to 106 colony-forming units (CFU) placed below the plate. Each plate contained two accessions with six seeds ml–1. Arabidopsis seedlings grown on plates as detailed above were in- per accession. To account for positional effects within and between the oculated with 5 µl of the bacterial solution, which was applied 1 cm Petri dishes, we plated 12 seeds for each accession over two plates in a above the root tip, as described previously (Digonnet et al., 2012). Plates permutated block design. Plates were positioned in racks that oriented were then sealed with micropore tape (3M Deutschland GmbH, Neuss, the plates in a vertical position to a growth chamber constantly kept at Germany) and transferred to a controlled growth chamber at 25 °C, 60% 21 °C and a 16 h light/8 h dark cycle, with a light intensity of 120 µmol humidity, and a 12 h light/12 h dark photoperiod with a light intensity m–2 s–1 during the light period. Plants were inoculated as described in the of 120 µmol m–2 s–1 during the light period. section below ‘In vitro inoculation assays’. For the analysis of root growth inhibition and root hair formation, pictures were taken 48–72 hpi with an Olympus DP71 stereomicroscope (Olympus, Center Valley, PA, USA) at ×11.5. To analyze green fluores- Image acquisition cent protein (GFP) root expression, roots from seedlings grown on plates Root images were obtained using CCD flatbed scanners (EPSON were collected 48 hpi and photographed with a Leica DM6 epifluores- Perfection V600 Photo, Seiko Epson Co., Nagano, Japan). The BRAT cence microscope (Leica, Wetzlar, Germany). In order to quantify GFP (Busch-lab Root Analysis Toolchain) image acquisition tool on a standard fluorescence, the Leica Application Suite X (LAS X) software was used. 2730 | Alonso-Díaz et al. A 0.1 cm section of the maturation zone was selected and GFP intensity Publication of Quantitative Real-Time PCR Experiments (MIQE) was quantified as relative units and presented as the average of all roots guidelines (Bustin et al., 2009). measured. Three independent biological replicates were performed and, for each replica, 24 (Fig. 2) or 10 (Fig. 5) roots per condition were used. Cytokinin analysis (LC-MS/MS) Arabidopsis plants were grown in pots with sand for 5 weeks in a con- Exogenous cytokinin and salicylic acid application trolled chamber at 22 °C, 60% humidity, and an 8 h light/16 h dark For R. solanacearum in vitro root inoculation assays that included ex- photoperiod, with a light intensity of 120 µmol m–2 s–1 during the light ogenous application of hormones, 7-day-old seedlings were transferred period. The sand in the pots was drilled three times with a wooden stick from MS agar plates to fresh MS agar plates supplemented with different and immediately irrigated with bacterial solution of overnight-grown hormone concentrations (25 nM and 50 nM kinetin; 1.5 µM and 7.5 µM R. solanacearum adjusted to OD600=0.1 with distilled water (50 ml of bac- SA) from Duchefa Biochemie. Roots were inoculated 24 h later as de- terial solution per plant). Trays with plant pots infected were transferred scribed above. to a chamber at 27 °C, 60% humidity, and 12 h light/1 2h dark. Then, at 4–7 days post-inoculation (dpi), inoculated roots were washed with dis- tilled water and dried with filter paper. After that, the root samples were Pathogenicity assays weighed and stored at –80 °C. Four biological replicates with 20 mg each Ralstonia solanacearum pathogenicity tests were carried out using the soil- were used each time (0, 4, and 7 dpi). Cytokinin levels were measured as drench method (Monteiro et al., 2012). Briefly, Arabidopsis was grown described previously (Poitout et al., 2018). for 4 weeks in Jiffy pots (Jiffy Group, Lorain, OH, USA) in a controlled chamber at 22 °C, 60% humidity, and an 8 h light/16 h dark photoperiod. Jiffys were drilled three times with a wooden stick and immediately sub- merged for 30 min into a solution of overnight-grown R. solanacearum Results adjusted to OD600=0.1 corresponding to 108 CFU ml–1 with distilled water (35 ml of bacterial solution per plant). Inoculated plants were trans- Genome-wide association mapping reveals several loci ferred to trays containing a thin layer of soil drenched with the same associated with cytokinin metabolism in Arabidopsis R. solanacearum solution and kept in a chamber at 27 °C, 60% humidity, roots infected with R. solanacearum and 12 h light/12 h dark. Plant wilting symptoms were recorded every day and expressed according to a disease index scale (0, no wilting; 1, 25% Infection of Arabidopsis roots with R. solanacearum GMI1000 in wilted leaves; 2, 50%; 3, 75%; and 4, death) (Supplementary Fig. S5). At vitro results in root growth inhibition. We previously observed least 30 plants per condition were used in each assay, and at least three replicates were performed for every experiment. natural variation of this phenotype across a small population of Arabidopsis accessions (Lu et al., 2018).To identify loci respon- sible for this natural variation, we performed a GWAS using a Quantitative reverse transcription–PCR (RT–qPCR) collection of 430 Arabidopsis accessions representative of the Roots were collected from R. solanacearum-infected or water-treated worldwide genetic variation of this species (Supplementary Fig. Arabidopsis plants at 0, 24, and 48 hpi. Briefly, roots from ~40 seedlings S1; Supplementary Table S1). Arabidopsis seeds were sown on were cut and pooled. Roots were rapidly washed in water and dried be- fore freezing in liquid nitrogen. Samples were stored at –80 °C. RNA was agar plates following the scheme presented in Supplementary extracted using the Maxwell 16 LEV Plant RNA Kit (Promega, Australia) Fig. S1 to ensure randomization. After 7 d, seedlings were in- according to the manufacturer´s recommendations. RNAs were treated oculated 1 cm over the root tip with a 5 µl droplet of a 106 with DNase-free RNase (Promega, Australia) and the concentration CFU ml–1 suspension of R. solanacearum GMI1000. Images of measured with an ND-8000 Nanodrop. cDNA was synthesized from seedlings were then acquired using scanners every day for 5 2 µg of RNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, USA) according to the manufacturer’s instruc- d to measure root length and to monitor the impact on root tions. According to the SYBR Green PCR mix instructions (Roche, growth caused by R. solanacearum infection in vitro. Differences Switzerland), 2.5 µl of cDNA were used in a final reaction volume of in root length between accessions were monitored after infec- 10 µl in the LightCycler 480 System (Roche, Switzerland). Melting tion and subsequently analyzed. curves and relative quantification of target genes were determined using To identify sequence variation in genomic regions associated the software LightCycler V1.5 (Roche, Switzerland). The amplification program was set to an initial step of 10 min at 95 °C followed by 45 with the variation of the root growth inhibition phenotype cycles using 95 °C for 10 s, 60 °C for 30 s, and 72 °C for 30 s. All samples caused by the R. solanacearum root infection, we conducted were run in triplicate for each biological replicate, and the target gene GWA mapping using the Arabidopsis 250K SNP chip data was normalized to the endogenous control gene Arabidopsis tubulin β-1 (Horton et al., 2012) with a mixed model correcting for popu- chain (At1g75780). To visualize the data, we calculated the fold change lation structure (Seren et al., 2012) and the root growth data of each biological replicate in 24 h and 48 h samples by normalizing to the ΔCt of time point 0 hpi of the mock and infected samples separately. described in Supplementary Table S2. Because we were inter- The statistical analysis of the normalized data was performed using the ested in the root growth responses upon R. solanacearum root ‘rstatix’ R package (ver. 0.6.0). To test for differences in gene expression infection, we focused our analysis on root growth rates. The between mock and infected samples, the normalized data were tested for broad-sense heritability (H2) of these traits ranged from 10% normality and homogeneity of variances. If these two requirements were to 55% with an average of 36% (Supplementary Table S3). fulfilled, the parametric t-test was performed for each time point to com- pare between mock and infected samples. All primer sequences used were We observed 20 unique SNPs significantly associated with obtained from previous publications and are listed in Supplementary the root growth responses to R. solanacearum infection using Table S6. qPCR analysis conforms to the Minimum Information for a 5% Benjamini–Hochberg threshold (Supplementary Table GWAS shows cytokinin’s role in root defense against R. solanacearum | 2731 S4). The most significant of these associations (SNP 15401974, significant P-value and was found in traits relating to two dif- chromosome 5; P-value 1.64×10–9; FDR 5.6×10–5) was found ferent days of the time course, we concluded that it might for two root growth rate measurements: the mean of the rela- be important in explaining the root growth phenotypic vari- tive root growth rate between day 2 and day 3 (Fig. 1A); and ation between accessions. While this SNP is located within the median of the relative root growth rate between day 4 the 5 kb upstream region of multiple genes (At5g38450 and and day 5 (Fig. 1B). Because this SNP displayed the most At5g38460) (Supplementary Fig. S2), the highest level of Fig. 1. GWA analysis reveals association of cytokinin metabolism genes with root growth inhibition caused by R. solanacearum infection of Arabidopsis roots. (A–D) Manhattan plots of GWA results for root growth traits. Different colors represent different chromosomes. The horizontal dashed lines correspond to a nominal P<0.05 significance threshold after Benjamini–Hochberg correction. Solid red boxes highlight the SNPs with the highest P-values in: (A) mean relative root growth rate between day 2 and day 3 (P-value 2.42E-07); (B) median root growth rate between day 4 and day 5 (P-value 1.64E-09); (C) mean relative root growth rate between day 2 and day 3 (P-value 8.89E-07); and (C) median root growth rate between day 2 and day 3 (P-value 6.55E-07). Fold change values of the quantitative PCR analysis of (E) CYP735A1, (F) CKX2, and (G) CKX4 using TUBULIN as control. Asterisks indicate statistically significant differences in a paired Student’s t-test (*P<0.05) between 48 hpi and normalizing by the values of the time point control (0 h hpi). 2732 | Alonso-Díaz et al. linkage disequilibrium in any gene of this region with the to GFP (TCSn::GFP) (Zürcher et al., 2013). Arabidopsis seed- top SNP can be observed for an SNP in the At5g38450 gene lings stably expressing TCSn::GFP were grown vertically on (Pearson coefficient of correlation r=0.39).This gene encodes a MS medium during 7 d and then roots were inoculated with cytokinin hydroxylase (CYP735A1) that catalyzes the biosyn- R. solanacearum (see the Materials and methods). Infection re- thesis of the cytokinin trans-zeatin (Takei et al., 2004). Another sulted in a strong induction of GFP expression driven by the analysis guided our focus towards the cytokinin pathway: when cytokinin signaling reporter TCSn in the vasculature of the conducting a GO enrichment analysis of genes in 10 kb prox- root maturation zone (Fig. 2B, C). The intensity of the GFP imity to SNPs associated with root growth rate upon infec- induction caused by R. solanacearum infection at 48 hpi was tion (EMMAX P-value <10–6) (Supplementary Table S4), we four times higher than in the water control, which clearly in- found the process cytokinin catabolism to be significantly en- dicated that cytokinin signaling is engaged in root responses to riched (P-value 0.00025; FDR 4.86%; Supplementary Table R. solanacearum invasion (Fig. 2D). S5). These included two additional genes associated with cyto- kinin metabolism, the cytokinin oxidases At2g19500 (CKX2) Plants affected in cytokinin biosynthesis and and AT4G29740 (CKX4). CKX2 is upstream of an SNP sig- perception display enhanced susceptibility towards nificantly associated with mean relative root growth rate be- R. solanacearum tween day 2 and day 3 (SNP 8436350; chromosome 2; P-value 8.89×10–7; FDR 0.015) (Fig. 1C) and CKX4 is upstream of If cytokinin levels and cytokinin signaling are important for an SNP marginally associated with median root growth rate root defense responses against R. solanacearum, it would be ex- between day 2 and day 3 (SNP 14577216; chromosome 4; pected that impairment of cytokinin biosynthesis results in en- P-value 6.55×10–7; FDR 0.101) (Fig. 1D). Both genes code hanced susceptibility to the pathogen. To address this question, for proteins that catalyze the degradation of cytokinins (Mok we performed pathogenicity assays on knockout mutants of the and Mok, 2001). cytokinin biosynthetic enzymes CYP735A1 and CYP735A2, Next, we analyzed the level of expression of CYP735A1, which do not display any apparent phenotype (Kiba et al., CKX2, and CKX4 in Arabidopsis Col-0 roots. This ecotype 2013). For this, 4-week-old Arabidopsis plants were inoculated was selected for further analysis because it has been widely with R. solanacearum GMI1000 by soil drenching, and symp- used for pathogenicity assays using R. solanacearum, many gen- toms were evaluated over time following a disease index scale etic resources are available, and it is susceptible to the widely (Lu et al., 2018). Both cyp735a1 and cyp735a2 showed earlier available GMI1000 strain, with a clearly observable root in- wilting disease symptoms and were dramatically more suscep- hibition phenotype that appears at early stages of infection tible to R. solanacearum than wild-type plants (Fig. 3A). This (Lu et al., 2018; Supplementary Fig. S3). Quantitative PCR clearly indicates that cytokinin biosynthesis is involved in de- was used to compare plants infected with R. solanacearum and fense responses against R. solanacearum. mock-treated plants at 0, 24, and 48 h post-treatment. Ralstonia Based on this, we hypothesized that cytokinin perception solanacearum infection consistently induced expression of would be equally important for immune responses against two of these three genes (CYP735A1 and CKX2) at 48 hpi R. solanacearum. To test this idea, we performed pathogenicity (Fig. 1E–G), indicating a potential involvement of cytokinin assays on knockout mutants of the sensor histidine kinases signaling in plant root defenses against this bacterial pathogen. AHK2, AHK3, and CRE1/AHK4, which act as cytokinin re- ceptors (Ueguchi et al., 2001; Higuchi et al., 2004). All three Early R. solanacearum infection induces cytokinin cytokinin receptor mutants ahk2, ahk3, and cre1/ahk4, which signaling in Arabidopsis roots. grow normally on soil (Ueguchi et al., 2001; Higuchi et al., 2004), displayed enhanced susceptibility to R. solanacearum Next, we analyzed whether R. solanacearum infection resulted infection (Fig. 3B), indicating that perception of cytokinin in an increase of cytokinin content in Arabidopsis Col-0 roots. is an important component of defense responses during R For this, we measured the levels the cytokinins trans-zeatin, solanacearum infection. cis-zeatin, and isopentenyladenine, as well as total cytokinins using LC-MS/MS. We could observe a significant increase in Exogenous cytokinin application partially reverts trans- and cis-zeatin, as well as total cytokinins after infection R. solanacearum-induced early root phenotypes (Fig. 2A–D). We could only detect significant increases at later stages of infection (4 and 7 dpi), probably due to the sensitivity Our next goal was to determine whether exogenous cyto- constraints of the measurement method. kinin application could counteract the effects caused by In order to more specifically investigate the early effects of R. solanacearum infection using the in vitro early root pheno- R. solanacearum infection on root cytokinin signaling, we took types as a measurable output (root growth inhibition and root advantage of a more sensitive approach by analyzing expres- hair production) (Lu et al., 2018). For this, 7-day-old seed- sion of the synthetic Arabidopsis cytokinin reporter TWO lings grown in vitro were transferred to fresh MS medium COMPONENT SIGNALING SENSOR new (TCSn) fused supplemented with different concentrations of the natural GWAS shows cytokinin’s role in root defense against R. solanacearum | 2733 Fig. 2. Early R. solanacearum infection induces cytokinin signaling in Arabidopsis Col-0 roots. Four-week-old Col-0 plants were inoculated with R. solanacearum, and (A) total cytokinin, (B) trans-zeatin, (C) cis-zeatin, and (D) isopentenyladenine concentrations were analyzed in inoculated root tissues at the indicated times (0, 4, and 7 dpi) by LC-MS/MS using four biological replicates. Error bars correspond to SEs. Asterisks indicate statistically significant differences between 4 and 7 dpi and the control (0 dpi) in a paired Student’s t-test (***P <0.001). (E–G) Six-day-old seedlings stably expressing the cytokinin signaling marker TCSn::GFP were inoculated with R. solanacearum or water, and roots at 48 hpi were observed under an epifluorescence microscope. TCSn::GFP signal in (E) whole roots and (F) the root maturation zone. (G) Quantification of the fluorescence intensity on the maturation zone (F) corresponds to the average GFP intensity from 24 individual roots per condition, calculated using the LAS X software. The experiment was repeated three times with similar results. Error bars correspond to SEs. Asterisks indicate statistically significant differences between 48 hpi and the control (0 hpi) in a paired Student’s t-test (**P<0.01). 2734 | Alonso-Díaz et al. Fig. 3. Cytokinin biosynthesis and perception are important for the plant response against R. solanacearum root infection. Four-week-old plants were soil-drench inoculated with R. solanacearum and symptoms were measured over time using a disease index on a scale of 1 to 4 (0=no wilting, 1=25% wilted leaves, 2=50%, 3=75%, and 4=death). (A) Wild-type Col-0 and trans-zeatin biosynthesis mutants cyp735a1 and cyp735a2. (B) Wild-type Col-0 and cytokinin sensor histidine kinase mutant genes (ahk2, ahk3, and ahk4/cre1). Each experiment was repeated at least three times obtaining similar results, using 24 plants per experiment. Error bars correspond to SEs. Asterisks indicate statistically significant differences between the wild type and different mutant lines in a paired Student’s t-test (*P<0.05, **P<0.01, ***P<0.001). Fig. 4. Exogenous cytokinin application partially reverts R. solanacearum-induced early root phenotypes. Seven-day-old Col-0 seedlings were grown for 24 h in MS medium supplemented with kinetin (0, 25, and 50 nM) and were inoculated with R. solanacearum or water. (A) An image of representative plants was obtained using a stereoscope 72 h after infection or mock treatment. (B) Root growth was measured at the indicated time points. (C) Images of representative roots were obtained using a stereoscope at the indicated time points after infection. Error bars correspond to SEs. Asterisks indicate statistically significant differences between 24, 48, and 72 hpi in a paired Student’s t-test (**P<0.01, ***P<0.001). cytokinin kinetin (0, 25, and 50 nM). After 24 h, roots were inoculated seedlings kept growing, although to a lesser extent pin-inoculated with R. solanacearum 1 cm above the root tip, than non-infected roots. In addition, root hair production re- and root growth inhibition and root hair production were sulting from R. solanacearum infection was also inhibited by monitored over time. Interestingly, kinetin supplementation the kinetin pre-treatment (Fig. 4C). This effect was more pro- (both 25 nM and 50 nM concentrations) resulted in partial nounced when using a higher kinetin dose (50 mM), where no reversion of the root growth inhibition phenotype caused by root hairs were observed. In contrast, the lower dose resulted in R. solanacearum in vitro (Fig. 4A, B). Whereas untreated in- delayed but visible root hair production. We can thus conclude oculated seedlings stopped growing at 24 hpi, kinetin-treated that kinetin pre-treatment at concentrations between 25 nM GWAS shows cytokinin’s role in root defense against R. solanacearum | 2735 and 50 nM can alleviate the early root in vitro phenotypes which is impaired in SA biosynthesis upon pathogen challenge caused by R. solanacearum infection, without causing toxicity (Dewdney et al., 2000). TCSn::GFP expression after infection to the plants. This indicates that cytokinin contributes to the is reduced when SA signaling is suppressed, compared with onset of early responses that take place upon R. solanacearum the wild type. This can be observed at the whole-root level infection in the root. (Fig. 5B) and in a magnified area of the root maturation zone (Fig. 5C). Fluorescence quantification shows that in eds16 Salicylic acid contributes to cytokinin signaling in mutant plants, TCSn::GFP expression is 25% lower than in a Arabidopsis roots in response to R. solanacearum TCSn::GFP wild-type background (Fig. 5D). This indicates infection that SA signaling affects cytokinin signaling in response to R. solanacearum infection, pointing towards a potential cyto- A crosstalk between cytokinins and SA has been previously kinin–SA crosstalk occurring in roots in response to infection shown to regulate plant defenses against pathogens infecting with soil-borne pathogens. Further research in this area will leaves, such as Pseudomonas syringae (Choi et al., 2010), clarify the possibility of a cytokinin–SA crosstalk during re- Hyalopenospora arabidopsidis (Argueso et al., 2012), Botrytis sponses to pathogens in roots. cinerea, and Oidium neolycopersici (Gupta et al., 2020). However, whether a crosstalk between these hormones in root defenses takes place has not been determined. Previous R. solanacearum pathogenicity tests have not detected differences in suscepti- Discussion bility between wild-type and SA-deficient plants (sid2 mutant Role of cytokinin in the interaction between or NahG transgenic lines, carrying an SA-degrading enzyme) R. solanacearum and Arabidopsis (Hirsch et al., 2002; Hernández-Blanco et al., 2007; Hanemian et al., 2016). However, exogenously applied SA had a clear ef- In this study, we have taken advantage of GWAS to understand fect on the root phenotypes induced by R. solanacearum infec- the genetic nature of the root phenotypic changes induced tion in vitro. Seven-day-old seedlings were transferred for 24 h by R. solanacearum on Arabidopsis roots during early stages of to MS medium supplemented with different SA concentra- infection. GWAS has been previously used to understand the tions (0, 1, 5, and 7.5 µM).We observed that SA concentrations basis of resistance against R. solanacearum in Arabidopsis under >1 µM (5 µM and 7.5 µM) caused root growth inhibition of different temperatures and inoculation conditions (Aoun et al., up to 50% of untreated seedlings (Supplementary Fig. S4A), as 2017). The study by Aoun and colleagues used wilt disease reported in previous studies (Pasternak et al., 2019). Therefore, index rates over time as a trait to underscore temperature- we performed R. solanacearum in vitro inoculation assays only dependent genetic diversity. At lower temperatures (27 °C) the on seedlings pre-treated with 1 µM SA, which did not cause main resistance quantitative trait locus (QTL) identified was any obvious effect on root growth before inoculation. Seven- RPS4/RRS1-R, a plant immune receptor pair with a very day-old seedlings were transferred to MS medium supple- well-known role in resistance of Arabidopsis to R. solanacearum mented with 1 µM SA and, 24 h later, roots were inoculated (Deslandes et al., 2002; Le Roux et al., 2015; Sarris et al., 2015). with R. solanacearum and monitored over time for root growth In addition to that, this study revealed a new potential sus- inhibition and root hair production. Exogenous application ceptibility gene at higher temperatures (30 ºC), strictosidine of 1 µM SA partially reverted the root inhibitory phenotype synthase-like 4 (SSL4), which encodes a protein with structural caused by R. solanacerum infection (Fig. 5A). On the other hand, similarities to animal proteins involved in immunity (Aoun 1 µM SA did not have any significant effect on root hair pro- et al., 2017). This study highlights the power of GWAS in re- duction (Supplementary Fig. S4b). Root hair production was vealing new potential sources of resistance to be engineered partly inhibited only when higher SA concentrations (5, 7.5, into crops. and 1 µM) were exogenously supplied prior to R. solanacearum Our study focuses on the same Arabidopsis–R. solanacearum infection (Supplementary Fig. S2b). Considering that these pathosystem but from a different angle. At the very early concentrations affect root growth under normal conditions stages of infection (2–3 dpi), R. solanacearum infection re- (Supplementary Fig. S2A), root hair inhibition may be a pleio- sults in quick root growth inhibition, root hair formation, tropic growth/development phenotype derived from SA and root meristem cell death, which can be easily observed toxicity rather than the result of SA modulation of defense and screened in in vitro inoculation assays. We detected nat- responses to R. solanacearum. ural variation associated with these phenotypes among a small To ascertain whether SA contributes to the cytokinin subset of accessions representative of Arabidopsis diversity (Lu signaling involved in the response of Arabidopsis Col-0 roots et al., 2018). Based on that observation and on the fact that to R. solanacearum root infection, we tested if impairing SA the initial stages of plant colonization by R. solanacearum are signaling would result in a decrease of cyokinin signaling out- poorly understood, despite being important for establishment puts. For this, we quantified expression of TCSn::GFP in of the bacteria inside the plant, we took advantage of GWAS transgenic lines in a wild-type or an eds16 mutant background, to analyze the genetic diversity associated with one of these 2736 | Alonso-Díaz et al. Fig. 5. SA contributes to cytokinin signaling in roots in response to R. solanacearum infection. (A) Seven-day-old Col-0 wild-type seedlings were grown for 24 h in MS medium supplemented with SA (0 µM and 1 µM) and were then inoculated with R. solanacearum. Root growth was measured. Error bars correspond to SEs. Asterisks indicate statistically significant differences between 24, 48, and 72 hpi in a paired Student’s t-test (**P<0.01). (B–D) Six-day- old seedlings stably expressing the cytokinin signaling marker TCSn::GFP in the Col-0 and eds16 background were inoculated with R. solanacearum, and at 48 hpi roots were observed under an epifluorescence microscope. TCSn::GFP signal in (B) whole roots and (C) the root maturation zone. This experiment was performed twice using 10 plants per genotype in each experiment. (D) Quantification of the fluorescence intensity on the maturation zone (C) corresponds to the average GFP intensity from 10 individual roots per condition calculated using the LAS X software. The experiment was repeated three times with similar results. Error bars correspond to the SE. Asterisks indicate statistically significant differences between 48 hpi and the control (0 hpi) in a paired Student’s t-test (**P<0.01). traits: root growth inhibition. Root hair formation and root metabolism of cytokinin: CYP735A1 in biosynthesis (Takei meristem cell death were not included in GWAS because the et al., 2004), and CKX2 and CKX4 in cytokinin degradation technology at hand did not allow precise measurement of (Mok and Mok, 2001). Gene expression analysis by qPCR these traits. showed that the expression of these genes in Col-0 roots was Using GWAS, we screened root growth inhibition at different consistently induced by R. solanacearum at 48 hpi (Fig. 1E, F). time points after infection on a large number of Arabidopsis Although the genes involved in cytokinin degradation, CKX2 accessions and focused on three candidate loci in the close and CKX4, have not been investigated further in this work, proximity of SNPs that are significantly associated with this they might participate in modulating the increased cytokinin phenotype (Fig. 1A–C). These three genes are involved in the levels in response to R. solanacearum infection. GWAS shows cytokinin’s role in root defense against R. solanacearum | 2737 Our data are in line with previous data underscoring a poten- this, we carried out pathogenicity tests, comparing a variety tial role for cytokinins in plant defense against R. solanacearum. of mutants with defects in cytokinin perception (ahk2, ahk3, RNA sequencing results show induction of genes involved in and ahk4/cre1) and biosynthesis (cyp735a1 and cyp735a2). Our cytokinin synthesis (CYP735A2, LOG2, and LOG6), deg- data clearly showed that both cytokinin synthesis and percep- radation (CKX2, CKX3, and CKX5), and response regulation tion participate in defense against R. solanacearum, as defects in (ARR3, ARR4, ARR5, ARR7, and ARR16) in Arabidopsis either pathway result in enhanced susceptibility towards the Col-0 roots at early time points after R. solanacearum infec- pathogen (Fig. 3). tion (Zhao et al., 2019). Moreover, genes controlling cyto- Furthermore, application of low concentrations of the cyto- kinin metabolism (LOG and CKX), signaling (ARR genes). kinin kinetin partially reversed the phenotypes caused by and perception (CRE1) have been shown to be differentially R. solanacearum infection in Arabidopsis roots (Fig. 4). A plaus- expressed in roots of the susceptible A17 Medicago truncatula ible explanation could be that exogenous cytokinin applica- genotype after infection with R. solanacearum (Moreau et al., tion induces the expression of defense-related genes in the 2014). Importantly, Arabidopsis plants deficient in ARR6 show root, as we have shown here (Fig. 1F, G) similar to what has altered cell wall composition and are more susceptible to in- been previously reported for leaves (Rashotte et al., 2003; Choi fection with R. solanacearum (Bacete et al., 2020). Interestingly, et al., 2010; Argueso et al., 2012). In fact, high doses of cyto- Aoun et al. (2017) found two cytokinin-related genes among kinin were shown to induce resistance in Arabidopsis against their top SNPs obtained upon infection with R. solanacearum at the oomycete H. arabidopsidis (Argueso et al., 2012), against high temperatures (30 °C): the signal receptor AHK3 and the P. syringae in Arabidopsis (Choi et al., 2010), or even against cytokinin response factor CRF2. In agreement with this, we virus replication in Phaseolus vulgaris (Clarke et al., 1998b). In have found that ahk3 knockout mutants are more susceptible our root system, low doses of cytokinin were sufficient to par- to R. solanacerum than the wild-type control (Fig. 3B). tially prevent the root phenotypes caused by R. solanacearum. Coupled to these increases in cytokinin-regulated gene ex- We did not use high doses of cytokinin because they have been pression, we observed an activation of cytokinin signaling in shown to inhibit primary root (To et al., 2004; Argyros et al., the root 48 hpi with R. solanacearum, as evidenced by expres- 2008) growth. sion of the reporter TCSn::GFP (Fig. 2E–G). Together, these data indicate that R. solanacearum triggers cytokinin production The impact of SA on the cytokinin signaling involved in in the root by means of induction of gene expression of cyto- the response of Arabidopsis roots to R. solanacearum kinin biosynthetic genes, which is accompanied by activation infection of cytokinin signaling; in parallel, cytokinin degradation genes are up-regulated, to ensure a timely response and a tight regu- Previous research, performed mostly in shoot tissue, showed lation of cytokinin levels in the plant as has been described in that the role of cytokinins in plant immunity is deeply re- the literature (Rashotte et al., 2003; Brenner et al., 2012). lated to SA signaling, with a clear crosstalk between the two Furthermore, we could observe an increase of cytokinin signaling pathways taking place (Choi et al., 2010; Argueso levels in the root after infection (Fig. 2A–D), albeit at later et al., 2012; Gupta et al., 2020). Choi et al. demonstrated that stages of infection, since at early stages reliable detection was the SA-dependent TGA3 transcription factor binds to the re- challenging. When assessing which cytokinin forms were most sponse regulator ARR2, which is modulated by cytokinin abundant, we could detect that R. solanacearum resulted in an signaling, to generate a complex that binds to the PR1 pro- increase in both trans- and cis-zeatin levels, whereas the levels of moter and promotes defense against P. syringae (Choi et al., isopentenyladenine did not show significant changes. The fact 2010). Also, it has been shown that cytokinin regulates plant that trans-zeatin was among the most abundant forms was not immunity against the oomycete H. arabidopsidis through the surprising, since it is one of the most active forms of cytokinin elevation of defense responses that are dependent on SA in plants (Sakakibara, 2006). In contrast, cis-zeatin has always (Argueso et al., 2012). been regarded as an isomer with lower activity in plants. In Here, we addressed whether SA had any impact on cyto- fact, the study of cis-zeatin in the context of plant–pathogen kinin signaling induced as part of the defense responses against interactions has only been addressed in the Nicotiana tabacum– root-invading pathogens and root immunity. Although the Pseudomonas syringae pathosystem, where the exogenous SA-deficient plants show the same level of susceptibility to addition of this isomer promotes the resistance of the plant R. solanacearum as the wild type (Hirsch et al., 2002; Hernández- against the pathogen (Groβkinsky et al., 2011). Our data in- Blanco et al., 2007; Hanemian et al., 2016), previous reports in- dicate that cytokinin may play a role in root defense against dicate that SA may participate in defense against this pathogen. R. solanacearum, with cis- and trans-zeatin as two potentially im- For instance, RRS1-R-mediated defense in Arabidopsis eco- portant cytokinin forms for this defense function. type Niederzenz-1 is orchestrated by SA (Deslandes et al., Taking advantage of the genetic resources available for 2002). Additionally, SA partly contributes to the enhanced Arabidopsis, we tested whether cytokinin synthesis and/or tolerance to R. solanacearum observed in the Arabidopsis mu- perception participated in defense against R. solanacearum. For tant wat1 (Walls are Thin1) (Denancé et al., 2013). Further, SA 2738 | Alonso-Díaz et al. participates in defense of other plant species to R. solanacearum, Fig. S5. Disease index scale of Arabidopsis plants after such as in tomato-resistant varieties Hawaii 7996 and CRA 66 R. solanacearum infection. (Baichoo and Jaufeerally-Fakim, 2016), and in tobacco (Lowe- Table S1. Ecotypes used in this study. Power et al., 2016; Liu et al., 2017). Table S2. Root growth data. It has been previously reported that treatments with SA or Table S3. Heritability rates. its analog BTH [benzo(1,2,3)thiadiazole-7-carbothioic acid Table S.4 Top SNPs. S-methyl ester] are potent activators of plant defenses against Table S5. Gene Ontology (GO) analysis. various pathogens both in leaves (Achuo et al., 2002; Herman Table S6. Primers used for quantitative PCR. et al., 2008;War et al., 2011; Azami-Sardooei et al., 2013; Bektas and Eulgem, 2015; Kouzai et al., 2018) and in roots (Attard et al., 2010; Chuberre et al., 2018). We found that exogenous Acknowledgements SA application results in a partial reversion of the in vitro root We are grateful to Hitoshi Sakakibara (RIKEN Center, Yokohama, phenotypes caused by R. solanacearum infection (Fig. 5A), Japan) for cyp735a1 and cyp735a2, and to Klaus Harter (Center for Plant similar to that we observed after cytokinin treatment (Fig. 4A, Molecular Biology, Tübingen, Germany) for ahk2, ahk3, and ahk4/cre1. B). Our results indicate that SA might also contribute to root We thank Saul Lema and Marc Planas-Marquès for inspiring discussions defenses against R. solanacearum at early stages of infection. on the topic. We thank Ling Zhang (Salk Institute) for the calculation Although evidence is still limited, our results point towards of broad-sense heritability estimate, and Martí Bernardo (CRAG) for his help with data visualization and statistics. This work was supported the existence of an SA–cytokinins crosstalk in Arabidopsis roots by the Spanish Ministry of Economy and Competitiveness grant nos after infection, since R. solanacearum-triggered expression of RyC 2014-16158, AGL2016-78002-R, and PID2019-108595RB-I00/ the cytokinin marker TCSn::GFP is significantly reduced by AEI/10.13039/501100011033 to NSC, and through the ‘Severo Ochoa SA depletion in the eds16 mutant when compared with wild- Programme for Centres of Excellence in R&D’ (SEV-2015-0533). We type plants (Fig. 5B–D). These data demonstrate that in the also acknowledge financial support from the CERCA Programme/ context of root infection, SA levels affect cytokinin signaling Generalitat de Catalunya, from the Austrian Academy of Sciences through and, in turn, cytokinin signaling could modulate SA levels, the Gregor Mendel Institute (WB and SBS) and from funding for CTA although evidence proving the effects of cytokinin on SA and HMB by the United States Department of Agriculture (USDA) grant signaling is still limited in this system. This indicates that cross- no. COL00781. We acknowledge support of the publication fee by the talk between the cytokinin and SA pathways in response to CSIC Open Access Publication Support Initiative through its Unit of pathogens could also take place in response to R. solanacearum Information Resources for Research (URICI). in roots and might affect defense response outcomes. Whether cytokinins, SA, and their crosstalk have a more general role in immunity against root-invading pathogens will be interesting Author contributions to explore in the future. AA-D designed the research, performed the research, analyzed and in- Together, our data demonstrate that cytokinins participate terpreted data, and wrote the manuscript. SBS performed the research, in defense against R. solanacearum and are involved in the early analyzed data, and edited the manuscript. RdP-J performed the research root phenotypes caused by the pathogen at early stages of in- and analyzed data. HMB performed the research, analyzed data, and fection.While it is known that cytokinin plays a very important edited the manuscript. CG performed the research and analyzed data. role in defense against bacteria, fungi, or viruses (Clarke et al., CTA designed the research, interpreted data, and edited the manuscript. ON performed the research and analyzed data.WB designed the research, 1998a; Pogány et al., 2004; Choi et al., 2010; Großkinsky et al., interpreted data, and edited the manuscript. MV designed the research, 2011; Argueso et al., 2012), our findings highlight a novel role analyzed and interpreted data, and edited the manuscript. NSC designed for cytokinin in root immunity. Defenses in the root remain the research, analyzed and interpreted data, and wrote the manuscript. vastly unexplored and our study adds evidence indicating that pathogen perception in the root activates cytokinin metab- olism and signaling, which modulates plant immunity contrib- Data availability uting to plant defense. The data that support the findings of this study are available from the corresponding author, NSC, upon reasonable request. Supplementary data The following supplementary data are available at JXB online. References Fig.S1.Genome-wide association study experimental scheme. Achuo AE, Audenaert K, Meziane H, Höfte M. 2002. The SA-dependent Fig. S2. Zoom-in of SNP positions. defense pathway is active against different pathogens in tomato and tobacco. Mededelingen (Rijksuniversiteit te Gent. Fakulteit van de Landbouwkundige Fig. S3. Ecotype distribution based on root growth data after en Toegepaste Biologische Wetenschappen) 67, 149–157. infection with R. solanacearum. Aoun N, Tauleigne L, Lonjon F, Deslandes L, Vailleau F, Roux F, Fig. S4. Role of SA in R. solanacearum infection. Berthomé R. 2017. Quantitative disease resistance under elevated GWAS shows cytokinin’s role in root defense against R. solanacearum | 2739 temperature: genetic basis of new resistance mechanisms to Ralstonia Coll NS, Valls M. 2013. Current knowledge on the Ralstonia solanacearum solanacearum. Frontiers in Plant Science 8, 1387. type III secretion system. Microbial Biotechnology 6, 614–620. Aranzana MJ, Kim S, Zhao K, et al. 2005. Genome-wide association Corwin JA, Copeland D, Feusier J, Subedy A, Eshbaugh R, Palmer C, mapping in Arabidopsis identifies previously known flowering time and Maloof J, Kliebenstein DJ. 2016. The quantitative basis of the Arabidopsis pathogen resistance genes. PLoS Genetics 1, e60. innate immune system to endemic pathogens depends on pathogen gen- Argueso CT, Ferreira FJ, Epple P, To JP, Hutchison CE, Schaller GE, etics. PLoS Genetics 12, e1005789. Dangl JL, Kieber JJ. 2012. Two-component elements mediate interactions Denancé N, Ranocha P, Oria N, et al. 2013. Arabidopsis wat1 (walls between cytokinin and salicylic acid in plant immunity. PLoS Genetics 8, are thin1)-mediated resistance to the bacterial vascular pathogen, Ralstonia e1002448. solanacearum, is accompanied by cross-regulation of salicylic acid and Argyros RD, Mathews DE, Chiang YH, Palmer CM, Thibault DM, tryptophan metabolism. The Plant Journal 73, 225–239. Etheridge N, Argyros DA, Mason MG, Kieber JJ, Schaller GE. 2008. Deslandes L, Olivier J, Theulieres F, Hirsch J, Feng DX, Bittner- Type B response regulators of Arabidopsis play key roles in cytokinin Eddy P, Beynon J, Marco Y. 2002. Resistance to Ralstonia solanacearum signaling and plant development. The Plant Cell 20, 2102–2116. in Arabidopsis thaliana is conferred by the recessive RRS1-R gene, a Attard A, Gourgues M, Callemeyn-Torre N, Keller H. 2010. The imme- member of a novel family of resistance genes. Proceedings of the National diate activation of defense responses in Arabidopsis roots is not sufficient to Academy of Sciences, USA 99, 2404–2409. prevent Phytophthora parasitica infection. New Phytologist 187, 449–460. Dewdney J, Reuber TL, Wildermuth MC, Devoto A, Cui J, Stutius LM, Atwell S, Huang YS, Vilhjálmsson BJ, et al. 2010. Genome-wide asso- Drummond EP, Ausubel FM. 2000. Three unique mutants of Arabidopsis ciation study of 107 phenotypes in Arabidopsis thaliana inbred lines. Nature identify eds loci required for limiting growth of a biotrophic fungal pathogen. 465, 627–631. The Plant Journal 24, 205–218. Azami-Sardooei Z, Seifi HS, De Vleesschauwer D, Höfte M. 2013. Digonnet C, Martinez Y, Denance N, Chasseray M, Dabos P, Benzothiadiazole (BTH)-induced resistance against Botrytis cinerea is in- Ranocha P, Marco Y, Jauneau A, Goffner D. 2012. Deciphering the versely correlated with vegetative and generative growth in bean and cu- route of Ralstonia solanacearum colonization in Arabidopsis thaliana roots cumber, but not in tomato. Australasian Plant Pathology 42, 485–490. during a compatible interaction: focus at the plant cell wall. Planta 236, 1419–1431. Babosha AV. 2009. Regulation of resistance and susceptibility in wheat– powdery mildew pathosystem with exogenous cytokinins. Journal of Plant Groβkinsky DK, Novak O, Roitsch T, et al. 2011. Cytokinins mediate Physiology 166, 1892–1903. resistance against Pseudomonas syringae in tobacco through increased antimicrobial phytoalexin synthesis independent of salicylic acid signaling. Bacete L, Mélida H, López G, Dabos P, Tremousaygue D, Denancé N, Plant Physiology 157, 815–830. Miedes E, Bulone V, Goffner D, Molina A. 2020. Arabidopsis response regulator 6 (ARR6) modulates plant cell-wall composition and disease re- Gupta R, Pizarro L, Leibman-Markus M, Marash I, Bar M. 2020. sistance. Molecular Plant-Microbe Interactions 33, 767–780. Cytokinin response induces immunity and fungal pathogen resistance, and modulates trafficking of the PRR LeEIX2 in tomato. Molecular Plant Bac-Molenaar JA, Fradin EF, Becker FF, Rienstra JA, van der Schoot J, Pathology 21, 1287–1306. Vreugdenhil D, Keurentjes JJ. 2015. Genome-wide association mapping of fertility reduction upon heat stress reveals developmental stage-specific Hanemian M, Barlet X, Sorin C, et al. 2016. Arabidopsis CLAVATA1 and QTLs in Arabidopsis thaliana. The Plant Cell 27, 1857–1874. CLAVATA2 receptors contribute to Ralstonia solanacearum pathogenicity through a miR169-dependent pathway. New Phytologist 211, 502–515. Baichoo Z, Jaufeerally-Fakim Y. 2016. Ralstonia solanacearum upregulates marker genes of the salicylic acid and ethylene signaling pathways but not Hayward AC. 1991. Biology and epidemiology of bacterial wilt caused by those of the jasmonic acid pathway in leaflets of Solanum lines during early Pseudomonas solanacearum. Annual Review of Phytopathology 29, 65–87. stage of infection. European Journal of Plant Pathology 147, 615–625. Herman MA, Davidson JK, Smart CD. 2008. Induction of plant defense Bektas Y, Eulgem T. 2015. Synthetic plant defense elicitors. Frontiers in gene expression by plant activators and Pseudomonas syringae pv. tomato Plant Science 5, 804. in greenhouse-grown tomatoes. Phytopathology 98, 1226–1232. Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate: a Hernández-Blanco C, Feng DX, Hu J, et al. 2007. Impairment of cellu- practical and powerful approach to multiple testing. Journal of the Royal lose synthases required for Arabidopsis secondary cell wall formation en- Statistical Society: Series B (Methodological) 57, 289–300. hances disease resistance. The Plant Cell 19, 890–903. Brenner WG, Ramireddy E, Heyl A, Schmülling T. 2012. Gene regula- Higuchi M, Pischke MS, Mähönen AP, et al. 2004. In planta functions tion by cytokinin in Arabidopsis. Frontiers in Plant Science 3, 8. of the Arabidopsis cytokinin receptor family. Proceedings of the National Academy of Sciences, USA 101, 8821–8826. Bustin SA, Benes V, Garson JA, et al. 2009. The MIQE guidelines: min- imum information for publication of quantitative real-time PCR experiments. Hirsch J, Deslandes L, Feng DX, Balagué C, Marco Y. 2002. Delayed Clinical Chemistry 55, 611–622. symptom development in ein2-1, an arabidopsis ethylene-insensitive mu- tant, in response to bacterial wilt caused by Ralstonia solanacearum. Caesar K, Thamm AM, Witthöft J, Elgass K, Huppenberger P, Phytopathology 92, 1142–1148. Grefen C, Horak J, Harter K. 2011. Evidence for the localization of the Arabidopsis cytokinin receptors AHK3 and AHK4 in the endoplasmic re- Horton MW, Hancock AM, Huang YS, et al. 2012. Genome-wide pat- ticulum. Journal of Experimental Botany 62, 5571–5580. terns of genetic variation in worldwide Arabidopsis thaliana accessions from the RegMap panel. Nature Genetics 44, 212–216. Choi J, Huh SU, Kojima M, Sakakibara H, Paek KH, Hwang I. 2010. The cytokinin-activated transcription factor ARR2 promotes plant im- Huard-Chauveau C, Perchepied L, Debieu M, Rivas S, Kroj T, Kars I, munity via TGA3/NPR1-dependent salicylic acid signaling in Arabidopsis. Bergelson J, Roux F, Roby D. 2013. An atypical kinase under balancing Developmental Cell 19, 284–295. selection confers broad-spectrum disease resistance in Arabidopsis. PLoS Genetics 9, e1003766. Chuberre C, Plancot B, Driouich A, Moore JP, Bardor M, Gügi B, Vicré M. 2018. Plant immunity is compartmentalized and specialized in Iakovidis M, Teixeira PJ, Exposito-Alonso M, et al. 2016. Effector- roots. Frontiers in Plant Science 9, 1692. triggered immune response in Arabidopsis thaliana is a quantitative trait. Genetics 204, 337–353. Clarke SF, Burritt DJ, Jameson PE, Guy PL. 1998a. Influence of plant hor- mones on virus replication and pathogenesis-related proteins in Phaseolus Kalladan R, Lasky JR, Chang TZ, Sharma S, Juenger TE, Verslues PE. vulgaris L. infected with white clover mosaic potexvirus. Physiological and 2017. Natural variation identifies genes affecting drought-induced abscisic Molecular Plant Pathology 53, 195–207. acid accumulation in Arabidopsis thaliana. Proceedings of the National Academy of Sciences, USA 114, 11536–11541. Clarke SF, Burritt, DJ, McKenzie MJ, Jameson PE, Guy PL. 1998b. Influence of white clover mosaic potexvirus infection on the endogenous Kang HM, Sul JH, Service SK, Zaitlen NA, Kong SY, Freimer NB, cytokinin content of bean. Plant Physiology 120, 547–552. Sabatti C, Eskin E. 2010. Variance component model to account for 2740 | Alonso-Díaz et al. sample structure in genome-wide association studies. Nature Genetics 42, by the pathogen Ralstonia solanacearum in resistant tomato. Journal of 348–354. Experimental Botany 71, 2157–2171. Kang HM, Zaitlen NA, Wade CM, Kirby A, Heckerman D, Daly MJ, Pogány M, Koehl J, Heiser I, Elstner EF, Barna B. 2004. Juvenility Eskin E. 2008. Efficient control of population structure in model organism of tobacco induced by cytokinin gene introduction decreases suscepti- association mapping. Genetics 178, 1709–1723. bility to Tobacco necrosis virus and confers tolerance to oxidative stress. Kiba T, Takei K, Kojima M, Sakakibara H. 2013. Side-chain modification Physiological and Molecular Plant Pathology 65, 39–47. of cytokinins controls shoot growth in Arabidopsis. Developmental Cell 27, Poitout A, Crabos A, Petřík I, Novák O, Krouk G, Lacombe B, Ruffel S. 452–461. 2018. Responses to systemic nitrogen signaling in Arabidopsis roots involve Kouzai Y, Noutoshi Y, Inoue K, Shimizu M, Onda Y, Mochida K. 2018. trans-zeatin in shoots. The Plant Cell 30, 1243–1257. Benzothiadiazole, a plant defense inducer, negatively regulates sheath blight Proietti S, Caarls L, Coolen S, Van Pelt JA, Van Wees SCM, resistance in Brachypodium distachyon. Scientific Reports 8, 17358. Pieterse CMJ. 2018. Genome-wide association study reveals novel players Le Roux C, Huet G, Jauneau A, et al. 2015. A receptor pair with an in- in defense hormone crosstalk in Arabidopsis. Plant, Cell & Environment 41, tegrated decoy converts pathogen disabling of transcription factors to im- 2342–2356. munity. Cell 161, 1074–1088. Rashotte AM, Carson SDB, To JPC, Kieber JJ. 2003. Expression pro- Li B, Sun L, Huang J, Göschl C, Shi W, Chory J, Busch W. 2019. filing of cytokinin action in Arabidopsis. Plant Physiology 132, 1998–2011. GSNOR provides plant tolerance to iron toxicity via preventing iron- the author dependent nitrosative and oxidative cytotoxicity. Nature Communications Sa G, Mi M, He-chun Y, Ben-ye L, Guo-feng L, Kang C. 2001. Effects 10, 3896. of ipt gene expression on the physiological and chemical characteristics of Liu Q, Liu Y, Tang Y, Chen J, Ding W. 2017. Overexpression of NtWRKY50 Artemisia annua L. Plant Science 160, 691–698. increases resistance to Ralstonia solanacearum and alters salicylic acid and Sakakibara H. 2006. Cytokinins: activity, biosynthesis, and translocation. jasmonic acid production in tobacco. Frontiers in Plant Science 8, 1–12. Annual Review of Plant Biology 57, 431–449. Lowe-Power TM, Jacobs JM, Ailloud F, Fochs B, Prior P, Allen C. Sarris PF, Duxbury Z, Huh SU, et al. 2015. A plant immune receptor de- 2016. Degradation of the plant defense signal salicylic acid protects tects pathogen effectors that target WRKY transcription factors. Cell 161, Ralstonia solanacearum from toxicity and enhances virulence on tobacco. 1089–1100. MBio 7, 1–12. Satbhai SB, Setzer C, Freynschlag F, Slovak R, Kerdaffrec E, Lu H, Lema A S, Planas-Marquès M, Alonso-Díaz A, Valls M, Coll NS. Busch W. 2017. Natural allelic variation of FRO2 modulates Arabidopsis 2018. Type III secretion-dependent and -independent phenotypes caused root growth under iron deficiency. Nature Communications 8, 15603. by Ralstonia solanacearum in Arabidopsis roots. Molecular Plant-Microbe Seren, Ü., Vilhjálmsson, B. J., Horton, M. W., Meng, D., Forai, P., Interactions 31, 175–184. Huang, Y. S., Nordborg, M. 2012. GWAPP: a web application for genome- Mansfield J, Genin S, Magori S, et al. 2012. Top 10 plant pathogenic wide association mapping in Arabidopsis. The Plant Cell 24, 4793–4805. bacteria in molecular plant pathology. Molecular Plant Pathology 13, Shigenaga AM, Argueso CT. 2016. No hormone to rule them all: inter- 614–629. actions of plant hormones during the responses of plants to pathogens. Meijón M, Satbhai SB, Tsuchimatsu T, Busch W. 2014. Genome-wide Seminars in Cell & Developmental Biology 56, 174–189. association study using cellular traits identifies a new regulator of root devel- Slovak R, Göschl C, Su X, Shimotani K, Shiina T, Busch W. 2014. opment in Arabidopsis. Nature Genetics 46, 77–81. A scalable open-source pipeline for large-scale root phenotyping of Mok DW, Mok MC. 2001. Cytokinin metabolism and action. Annual Review Arabidopsis. The Plant Cell 26, 2390–2403. of Plant Physiology and Plant Molecular Biology 52, 89–118. Takei K, Yamaya T, Sakakibara H. 2004. Arabidopsis CYP735A1 and Mok MC. 1994. Cytokinins and plant development—an overview. In: CYP735A2 encode cytokinin hydroxylases that catalyse the biosynthesis of Mok DWS, Mok M, eds. Cytokinins—chemistry, activity, and function. Boca trans-zeatin. Journal of Biological Chemistry 279, 41866–41872. Raton, FL: CRC Press, 155–166. Thoen MP, Davila Olivas NH, Kloth KJ, et al. 2017. Genetic architecture Monteiro F, Genin S, van Dijk I, Valls M. 2012. A luminescent reporter of plant stress resistance: multi-trait genome-wide association mapping. evidences active expression of Ralstonia solanacearum type III secretion New Phytologist 213, 1346–1362. system genes throughout plant infection. Microbiology (Reading) 158, To JP, Haberer G, Ferreira FJ, Deruère J, Mason MG, Schaller GE, 2107–2116. Alonso JM, Ecker JR, Kieber JJ. 2004. Type-A Arabidopsis response Moreau S, Fromentin J, Vailleau F, Vernié T, Huguet S, Balzergue S, regulators are partially redundant negative regulators of cytokinin signaling. Frugier F, Gamas P, Jardinaud MF. 2014. The symbiotic transcription The Plant Cell 16, 658–671. factor MtEFD and cytokinins are positively acting in the Medicago truncatula and Ralstonia solanacearum pathogenic interaction. New Phytologist 201, Ueguchi C, Koizumi H, Suzuki T, Mizuno T. 2001. Novel family of sensor 1343–1357. histidine kinase genes in Arabidopsis thaliana. Plant & Cell Physiology 42, 231–235. Naseem M, Philippi N, Hussain A, Wangorsch G, Ahmed N, Dandekar T. 2012. Integrated systems view on networking by hormones War, A. R., Paulraj, M. G., War, M. Y., & Ignacimuthu, S. 2011. Role in Arabidopsis immunity reveals multiple crosstalk for cytokinin. The Plant of salicylic acid in induction of plant defense system in chickpea (Cicer Cell 24, 1793–1814. arietinum). Plant Signaling and Behavior 6, 1787–1792. Pasternak T, Groot EP, Kazantsev FV, Teale W, Omelyanchuk N, Wen Z, Tan R, Yuan J, et al. 2014. Genome-wide association mapping Kovrizhnykh V, Palme K, Mironova VV. 2019. Salicylic acid affects root of quantitative resistance to sudden death syndrome in soybean. BMC meristem patterning via auxin distribution in a concentration-dependent Genomics 91, 574–581. manner. Plant Physiology 180, 1725–1739. Wybouw B, De Rybel B. 2019. Cytokinin—a developing story. Trends in Pieterse CM, Leon-Reyes A, Van der Ent S, Van Wees SC. 2009. Plant Science 24, 177–185. Networking by small-molecule hormones in plant immunity. Nature Chemical Zhao C, Wang H, Lu Y, et al. 2019. Deep sequencing reveals early repro- Biology 5, 308–316. gramming of arabidopsis root transcriptomes upon Ralstonia solanacearum Pieterse CMJ, Leon-Reyes A, Zamioudis C, Van Wees SCM, infection. Molecular Plant-Microbe Interactions 32, 813–827. Van der Does D. 2012. Hormonal modulation of plant immunity. Annual Zürcher E, Tavor-Deslex D, Lituiev D, Enkerli K, Tarr PT, Müller B. Review of Cell and Developmental Biology 28, 489–521. 2013. A robust and sensitive synthetic sensor to monitor the transcriptional Planas-Marquès M, Kressin JP, Kashyap A, Panthee DR, Louws FJ, output of the cytokinin signaling network in planta. Plant Physiology 161, Coll NS, Valls M. 2020. Four bottlenecks restrict colonization and invasion 1066–1075.