Variation in plant Toll/Interleukin-1 receptor domain protein dependence on ENHANCED DISEASE SUSCEPTIBILITY 1

Abstract

Toll/Interleukin-1 receptor (TIR) domains are integral to immune systems across all kingdoms. In plants, TIRs are present in nucleotide-binding leucine-rich repeat (NLR) immune receptors, NLR-like, and TIR-only proteins. Although TIR-NLR and TIR signaling in plants require the ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1) protein family, TIRs persist in species that have no EDS1 members. To assess whether particular TIR groups evolved with EDS1, we searched for TIR-EDS1 co-occurrence patterns. Using a large-scale phylogenetic analysis of TIR domains from 39 algal and land plant species, we identified 4 TIR families that are shared by several plant orders. One group occurred in TIR-NLRs of eudicots and another in TIR-NLRs across eudicots and magnoliids. Two further groups were more widespread. A conserved TIR-only group co-occurred with EDS1 and members of this group elicit EDS1-dependent cell death. In contrast, a maize (Zea mays) representative of TIR proteins with tetratricopeptide repeats was also present in species without EDS1 and induced EDS1-independent cell death. Our data provide a phylogeny-based plant TIR classification and identify TIRs that appear to have evolved with and are dependent on EDS1, while others have EDS1-independent activity.

Figure 1

Land plants share four TIR groups. A, ML tree (evolutionary model WAG+F+R7) of 307 predicted TIR domain sequences representing major TIR families across plant species (full 2,317 sequence tree in Supplemental Figure S2B). Branches with BS support ≥90% are marked with black dots. Taxonomically shared TIR groups from more than one order are highlighted with colored boxes and their predominant domain architecture is depicted. Additional domains predicted in the TIR proteins are annotated as black boxes next to each TIR protein (used HMMs listed in Supplemental Table S2). Four TIR domain groups shared by at least two taxonomic groups (e.g. Rosids and Asterids in the case of TNL#2) were named after the predominant domain architecture of full-length proteins. The presence of TPRs in TNPs was deduced based on the TPR HMMs (Supplemental Table S2). The TIR-only RBA1/AtTX1 does not belong to conserved TIR-only proteins. The scale bar corresponds to number of substitutions per site. B, Counts of predicted full-length TIR proteins, proteins with taxonomically shared TIRs, ADR1/NRG1, and EDS1 family predicted in the species analyzed in this study. TNPs are not included in the counts of TNL, TN, and TIR-only proteins. TIR-only proteins are defined as sequences shorter than 400 amino acids, without other predicted PFAM domains. Sizes of circles reflect the counts. Eucalyptus grandis has a fragment of PAD4-like sequence as determined by TBLASTN searches. C, Comparison of important TIR domain motifs across the four conserved plant TIR groups. Full sets of TIR domains were taken based on phylogeny (tree in Supplemental Figure S2B). Sequence motifs were generated for each TIR group to show conservation of the catalytic glutamate, AE, and BE interfaces, as well as residues in the αD helix. Arabidopsis thaliana RPP1^WsB TIR domain was taken as reference. Chemical attributes of the important amino acids are annotated in different colors. C-JID, C-terminal jelly roll/Ig-like domain; NBARC, nucleotide-binding domain shared by APAF-1, certain R-gene products, and CED-4; RBA1, RECOGNITION OF HOPBA1. Full species names are in Supplemental Table S1.

Figure 2

Conserved TIR-only genes are upregulated during immune signaling and their expression triggers EDS1-dependent cell death in Nb. A, Comparison of untriggered and immune-triggered expression of genes corresponding to taxonomically shared TIR groups in Arabidopsis and barley (H. vulgare). Data were taken from publicly available RNA-seq experiments (Supplemental Table S4) including immune-triggered and infected samples. The significance of the association between the expression of conserved TIR-only genes and the immune-triggered status of RNA-seq samples was assessed with Fisher’s exact test. The test evaluated whether the expression of conserved TIR-only genes (transcript per million >0) is more likely in immune-triggered samples. Asterisks next to names of the conserved TIR-only genes denote the significance level: *P < 0.05, **P < 0.01, ***P < 0.001. Minima and maxima of boxplots—first and third quartiles, respectively, center line—median, whiskers extend to the minimum and maximum values but not further than 1.5 interquartile range. Data points (number given above the boxplot) with the same color correspond to one gene. For details, check the “Data availability”. Created with elements from BioRender.com. B, Heatmaps showing expression of conserved TIR-only genes in Arabidopsis with PTI or ETI. Expression data were taken from Saile et al. (2020). Triggers include P. fluorescens Pf0-1 empty vector (EV) for PTI, Pf0-1 avrRpm1, Pf0-1 avrRpt2, and Pf0-1 avrRps4 for PTI + ETI. Asterisks inside the heatmap indicate that conserved TIR-only AT1G52900 is upregulated at log₂ fold change >4 and adjusted ***P < 0.001 relative to mock at 0-h postinfiltration (Saile et al., 2020). C, Macroscopic cell death symptoms induced by Agrobacterium-mediated overexpression of conserved monocot YFP-tagged TIR-only proteins in Nb WT and the eds1a mutant. Pictures were taken 3 days after agroinfiltrations. Numbers below panels indicate necrotic/total infiltrated spots observed in three independent experiments. D, TIR-only protein accumulation in infiltrated leaves shown in (C) was tested via immunoblot. Expected sizes for YFP-tagged TIR-only proteins and free YFP as control are indicated. All tested variants of conserved TIR-only proteins are expressed in Nb WT and eds1a lines. Ponceau S staining of the membrane served as loading control. The detection was performed for two independent experiments with similar results.

Figure 3

A maize TNP induces EDS1-independent cell death in N. tabacum. A, ML tree (from IQ-TREE, evolutionary model JTT+G4) of 77 predicted TNP NBARC (Supplemental File S12; hmmsearch E < 0.01) domains representing the plant species analyzed within this study. Branches with BS support ≥90% are marked with black dots. The three conserved TNP clades are highlighted with colored boxes. Clade nomenclature was partly adapted from Zhang et al. (2017). The scale bar is number of substitutions per site. B, Macroscopic cell death symptoms induced by Agrobacterium-mediated overexpression of C-terminally YFP-tagged TNP proteins from four major clades (A) in tobacco (N. tabacum) “Samsun” WT and the RNAi:EDS1 knockdown line. Pictures were taken 5 days after agrobacteria infiltrations. Numbers below panels indicate necrotic/total infiltrated spots observed in three independent experiments. C, Overexpression of ZmTNP-IIa WT and mutant variants in the two adjacent putative catalytic glutamates (E130 and E131) or P-loop (G305A/K306A/T307A) in leaves of indicated tobacco varieties. Pictures were taken 5 days after agrobacteria infiltration. Numbers below panels indicate necrotic/total infiltrated spots observed in three independent experiments. D, ZmTNP-IIa-YFP protein accumulation in infiltrated leaves shown in (C) was tested via α-GFP IP and subsequent immunoblot. Expected sizes for YFP-tagged ZmTNP variants are indicated. Ponceau S staining of the IP input samples served as loading control. Similar results were obtained in another independent experiment.

Figure 4

TNPs are not required for plant survival but negatively influence B. cinerea disease symptoms in Nb. A, Macroscopic images of 2-week-old M. polymorpha Tak1 WT and two independent tnp CRISPR knockout lines. Scale bars = 0.1 cm. Genomic sequences of the two tnp lines are depicted in Supplemental Figure S12. B, Side-view images of 4-week-old Nb WT, two independent tnp quadruple CRISPR knockout lines (tnp-q1 and tnp-q2), and the eds1a mutant. Scale bars = 5.0 cm. Plants were grown in long-day (16-h light) conditions. Genomic sequences of the two tnp quadruple lines are depicted in Supplemental Figure S12. C, ROS burst upon several PAMP triggers in Nb WT, eds1a, eds1a pad4 sag101a sag101b (epss) and tnp quadruple mutants (tnp-q1 and tnp-q2). Values are means of log₂-transformed relative luminescence units (RLUs) after the addition of 2-μM nlp24, 200- nM flg22 or 4-mg mL⁻¹ chitin and were recorded for 60 min, n = 10–12, from three independent biological replicates. D, Total ROS produced after 60 min PAMP treatment. Values are sums of log₂-transformed RLU in (C). The letters above boxplots indicate significant differences among genotype-treatment combinations (Tukey’s HSD, α = 0.05, n = 10–12, from three independent biological replicates). E, Xcv growth assay in Nb. Plants were syringe infiltrated with Xcv 85-10 (WT) and XopQ-knockout strains (Δ xopQ) at OD₆₀₀ = 0.0005. Bacterial titers were determined at 3 and 6 dpi. Genotype-treatment combinations sharing letters above boxplots are not significantly different (Tukey’s HSD, α = 0.01, n = 12, from three independent biological replicates). Error bars represent standard error. F, Electrolyte leakage assay as a measure of XopQ-triggered cell death in Nb 3 days after Agrobacterium infiltration (OD₆₀₀ = 0.2) to express XopQ-Myc. YFP overexpression was used as negative control. Genotype-treatment combinations sharing letters above boxplots are not significantly different (Tukey’s HSD, α = 0.01, n = 18, from three independent biological replicates). G, Lesion area induced by B. cinerea strain B05.10 infection in Nb. Plants were drop-inoculated with spore suspension (5*10⁵ spores mL⁻¹) and lesion areas were measured 48 h after inoculation. Values shown are lesion areas normalized to WT. Genotypes sharing letters above boxplots are not significantly different (Tukey’s HSD, α = 0.01, n = 10–12, from five independent biological replicates). Boxplot elements in (F) and (G): first and third quartiles define maximum and minimum, respectively, center line: median, whiskers extend to the minimum and maximum values but not further than 1.5 interquartile range. H, Macroscopic images of B. cinerea-induced lesions measured in (G).