The Arabidopsis ZED1-Related Kinase Genomic Cluster Is Specifically Required for Effector-Triggered Immunity

CRISPR/Cas9-mediated deletion of an Arabidopsis gene cluster encoding eight kinases supports their immunity-specific roles in sensing pathogenic effectors.

Dear Editor,ZED1-related kinases (ZRKs) associate with the nucleotide binding, Leu-rich repeat (NLR) protein HOPZ-ACTIVATED RESISTANCE1 (ZAR1) to mediate effector-triggered immunity (ETI) against at least three distinct families of pathogenic effector proteins. However, it is unknown whether ZRKs specifically function in ETI or whether they also have additional roles in immunity and/or development. Eight ZRKs are clustered in the Arabidopsis (Arabidopsis thaliana) genome, including the three members with known roles in ETI. Here, we show that an ∼14-kb CRISPR-mediated deletion of the Arabidopsis ZRK genomic cluster specifically affects ETI, with no apparent defects in pattern-recognition-receptor–triggered immunity (PTI) or development.Phytopathogens deliver effector proteins into plant cells that suppress PTI and promote the infection process (Jones and Dangl, 2006). In turn, plants have evolved NLRs that recognize effectors, leading to an ETI response. This recognition often occurs indirectly, whereby NLRs monitor host “sensor” proteins for effector-induced perturbations (Khan et al., 2016). In the absence of their respective NLRs, some of these sensors are effector virulence targets that modulate immunity and development, while others appear to be decoys that mimic virulence targets, with ETI-specific roles (van der Hoorn and Kamoun, 2008; Khan et al., 2018).The ZAR1 NLR recognizes at least six type-III secreted effector (T3SE) families from bacterial phytopathogens. This remarkable immunodiversity appears to be conveyed through associations with members of the receptor-like cytoplasmic kinase XII-2 (RLCK XII-2) family, which all display characteristics of atypical kinases (Lewis et al., 2013; Roux et al., 2014). The ZAR1-mediated ETI responses against the Pseudomonas syringae T3SEs HopZ1a and HopF1r (formerly HopF2a) require ZED1 and ZRK3, whereas recognition of the Xanthomonas campestris T3SE AvrAC requires ZRK1/RKS1 (Lewis et al., 2013; Wang et al., 2015; Seto et al., 2017). ZRKs currently have no ascribed functions outside of ZAR1-associated ETI responses and are therefore considered decoy sensors or adaptors (Lewis et al., 2013; Wang et al., 2015; Khan et al., 2018). However, functional redundancy may exist among members of the ZRK family, masking phenotypes of individual mutants beyond gene-for-gene–type ETI responses (Lewis et al., 2013). We therefore utilized the CRISPR/Cas9 system to knock out the Arabidopsis genomic region containing eight of the 13 members of the RLCK XII-2, including all ZRK genes known to contribute to ETI, to investigate any non-ETI roles of ZRKs. The 14-kb ZRK gene cluster in Arabidopsis Col-0 plants includes ZRK1, ZRK2, ZRK3, ZRK4, ZED1, ZRK6, ZRK7, and ZRK10. A CRISPR/Cas9-mediated deletion of 13.3 kb was accomplished by designing guide RNAs flanking the ends of the ZRK gene cluster, which would result in a double-stranded break on both sides of the ZRK cluster, leaving only the 5′ end 63 nucleotides (21 amino acids) of ZRK10 and the 3′ end 118 nucleotides (39 amino acids) of ZRK7 (Fig. 1A). We obtained a T1 individual (zrk_1.11) homozygous for the deletion, as well as a T1 individual heterozygous for the mutation (zrk_1.10; Fig. 1B), from which we obtained homozygous T2 (zrk_2.11) and T3 (zrk_3.10) plants, respectively. Sequencing results from zrk_2.11 confirmed that the expected region had been deleted (Supplemental Fig. S1). Plants homozygous for the ZRK gene cluster deletion were morphologically indistinguishable from wild-type Col-0 plants, as well as zar1-1 plants (Fig. 1C). In addition, zrk plant fresh weight did not significantly differ from Col-0 plants (Supplemental Fig. S2), indicating that the ZRK cluster does not play a major role in vegetative plant development.Open in a separate windowFigure 1.Deletion of the ZRK gene cluster results in loss of ZRK-mediated ETI and does not significantly alter vegetative growth. A, Representation of ZRK gene cluster before (top) and after (bottom) CRISPR/Cas9-mediated deletion depicting guide RNAs and primers used for genotyping (see Supplemental Methods S1). ZRK KO primers (magenta) were used to confirm the deletion of the ZRK cluster, while ZRK3 primers (green) were used to check if the ZRK cluster was still present in T1 individuals. B, PCR genotyping for deletion of ZRK gene cluster. Amplification of product by ZRK3 F + R primers indicates lack of deletion; amplification by ZRK KO F + R indicates deletion has occurred. Examples for wild type (WT), heterozygous (HT; zrk_1.10), and homozygous for the deletion (HM KO; zrk_1.11) are shown. T1 lines (zrk_1.10 and zrk_1.11) are compared to wild-type Col-0. C, Uninfected morphology of homozygous zrk KO plants (zrk_3.10 or zrk_2.11) compared to Col-0 and zar1-1 plants. Bar = 1 cm. D, Phenotypes of zrk_2.11 plants 7 d after being sprayed with PtoDC3000(hopZ1a; left) or PtoDC3000(hopF1r; right) relative to wild-type Col-0 and zar1-1 plants. Plant immunity and disease image-based quantification of disease symptoms is presented in Supplemental Figure S3A (Laflamme et al., 2016).Next, we wanted to confirm that the deletion of the ZRK gene cluster compromised ZRK-mediated ETI responses. We sprayed the zrk_2.11 line with PtoDC3000(hopZ1a) or PtoDC3000(hopF1r), as both T3SEs require a ZRK as well as the NLR ZAR1 for their recognition in Arabidopsis (Lewis et al., 2013; Seto et al., 2017). We observed that the zrk_2.11 line was susceptible to both PtoDC3000(hopZ1a) and PtoDC3000(hopF1r), and this susceptibility was to the same level as zar1-1 plants as quantified by plant immunity and disease image-based quantification (Fig. 1D, Supplemental Fig. S3, A and C; Laflamme et al., 2016). We observed a similar phenotype for the zrk_3.10 line, confirming that the ZRK cluster deletion compromised ZRK-mediated ETI responses (Supplemental Fig. S3, B and C). Furthermore, the ZAR1-mediated ETI responses against the P. syringae T3SEs HopBA1a, HopX1i, and HopO1c were also lost in zrk_2.11, demonstrating the ZRK-dependence of these ETI responses (Supplemental Fig. S4; Laflamme et al., 2020). To ensure that the ZRK gene cluster deletion specifically impacted ZRK-related ETI responses, the zrk_3.10 and zrk_2.11 lines were also sprayed with PtoDC3000(avrRpt2), an ETI elicitor that does not require a ZRK or ZAR1 for its recognition (Mackey et al., 2003). zrk_3.10 and zrk_2.11 plants remained resistant to PtoDC3000(avrRpt2), indicating that the ZRK gene cluster deletion specifically impacts ZRK-mediated ETI responses (Supplemental Fig. S3C). In addition, growth of virulent PtoDC3000 on the zrk_3.10 and zrk_2.11 lines was unchanged compared to wild-type Col-0 plants, indicating that the ZRKs within this cluster likely do not represent virulence targets (Supplemental Fig. S5).We then examined whether knocking out the ZRK gene cluster impacted PTI. We first measured induction of peroxidase (POX) enzyme activity, as POX enzymes are produced in response to PTI (Mott et al., 2018). After treatment with the PTI elicitor flg22, addition of the POX substrate 5-aminosalicylic acid produces a brown end-product in the presence of active POX enzymes, which is quantified by reading at an optical density of 550 nm (OD₅₅₀; Mott et al., 2018). Twenty h after leaf discs were treated with flg22, zrk_3.10 and zrk_2.11 plants showed the same level of PTI-associated POX activity as wild-type Col-0 plants (Fig. 2A). To further examine the role of the ZRK gene cluster in PTI, we quantified the growth of PtoDC3000ΔhrcC, which is defective in T3SE secretion and is sensitive to altered host PTI responses under high humidity conditions such as those used in our growth assays (Guo et al., 2009; Xin et al., 2016). Growth of PtoDC3000ΔhrcC on zrk_3.10 and zrk_2.11 plants was not significantly different compared to wild-type Col-0 plants (Fig. 2B). In addition, we monitored reactive oxygen species (ROS) production and found that zrk_3.10 and zrk_2.11 plants did not show a significant difference in the ROS burst observed in wild-type Col-0 plants (Fig. 2, C and D). Finally, we treated seedlings with flg22, and found that growth of zrk_3.10 and zrk_2.11 seedlings was inhibited by the same amount as in wild-type Col-0 seedlings, indicative of a similar induction of PTI responses (Fig. 2, E and F; Gómez-Gómez et al., 1999). Together, these results indicate that the ZRK gene cluster does not play a significant role in Arabidopsis PTI responses.Open in a separate windowFigure 2.Deletion of the ZRK gene cluster does not alter pattern-recognition-receptor–triggered immune responses. A, Response to the PTI elicitor flg22 measured by POX activity. Activity from leaf discs was quantified 20 h after treatment with 1 μm of flg22 at a measurement of OD₅₅₀ (n = 6; Mott et al., 2018). B, Bacterial growth of the T3SS-compromised PtoDC3000ΔhrcC on zrk KO plants (zrk_3.10 and zrk_2.11) relative to wild-type Columbia-0 (wild-type Col-0) and zar1-1 plants 3-d postinoculation. Plants were domed for the duration of the experiment (n = 8). C, Response of Col-0, zrk KO plants (zrk_3.10 and zrk_2.11), and fls2 to the PTI elicitor flg22 measured using luminol-based detection of ROS over a time course of 60 min, with relative light units measured every 2 min (n = 12). D, Boxplots of total relative light units over a period of 30 min from treatments in C (n = 12). E, Growth inhibition of seedlings 7 d after treatment with 1 μm of flg22. F, Seedling growth inhibition was quantified by measuring fresh weight of flg22-treated seedlings as a percentage of water-treated controls (n = 4). Error bars in A, B, C, D, and F, represent se. Lowercase letters represent significantly different statistical groups by Tukey’s honest significant difference test (P < 0.05). Experiments were replicated three times with similar results.Overall, our results support an ETI-specific role for ZRKs in Arabidopsis, acting as sensors of the ZAR1 NLR. Structural insights have revealed important residues required for ZAR1-ZRK1 complex formation, and these are conserved across the RLCK XII-2 family, which includes ZRKs outside the genomic cluster (Supplemental Fig. S6; Lewis et al., 2013; Wang et al., 2019). This suggests that the ZRKs outside this genomic cluster may also play a similar role as ZAR1 sensors. As such, the ZRK family would have evolved to mimic and/or interact with the numerous kinase virulence targets of pathogenic effectors, thereby expanding the surveillance potential of ZAR1.Supplemental DataThe following supplemental materials are available.

• Supplemental Figure S1. Sequencing confirmation of the ZRK gene cluster deletion.
• Supplemental Figure S2. Fresh weight of zrk knockout (KO) plants (zrk_3.10 and zrk_2.11) relative to wild-type Columbia-0 (wild-type Col-0) and zar1-1 plants.
• Supplemental Figure S3. ZRK gene cluster deletion specifically compromises ZRK-dependent ETI responses.
• Supplemental Figure S4. ZRK gene cluster compromises the ZAR1-dependent ETI responses against HopBA1a, HopO1c, and HopX1i.
• Supplemental Figure S5. Bacterial growth of the virulent PtoDC3000 strain on zrk KO plants (zrk_3.10 and zrk_2.11) relative to wild-type Col-0 (wild-type Col-0) plants 0- and 3-d post-inoculation via syringe infiltration.
• Supplemental Figure S6. Multiple sequence alignment of RLCK XII-2 family shows high conservation of putative ZAR1-interacting residues.
• Supplemental Methods S1. Generation and characterization of ZRK cluster deletion lines.