Convergent Loss of an EDS1/PAD4 Signaling Pathway in Several Plant Lineages Reveals Coevolved Components of Plant Immunity and Drought Response

Abstract

Plant innate immunity relies on nucleotide binding leucine-rich repeat receptors (NLRs) that recognize pathogen-derived molecules and activate downstream signaling pathways. We analyzed the variation in NLR gene copy number and identified plants with a low number of NLR genes relative to sister species. We specifically focused on four plants from two distinct lineages, one monocot lineage (Alismatales) and one eudicot lineage (Lentibulariaceae). In these lineages, the loss of NLR genes coincides with loss of the well-known downstream immune signaling complex ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1)/PHYTOALEXIN DEFICIENT 4 (PAD4). We expanded our analysis across whole proteomes and found that other characterized immune genes were absent only in Lentibulariaceae and Alismatales. Additionally, we identified genes of unknown function that were convergently lost together with EDS1/PAD4 in five plant species. Gene expression analyses in Arabidopsis (Arabidopsis thaliana) and Oryza sativa revealed that several homologs of the candidates are differentially expressed during pathogen infection, drought, and abscisic acid treatment. Our analysis provides evolutionary evidence for the rewiring of plant immunity in some plant lineages, as well as the coevolution of the EDS1/PAD4 pathway and drought responses.

Figure 1.

Phylogenetic Relationship and NLR Repertoires of the Plant Species Used in This Study. (A) Boxplot showing the variation in number of NB-ARC domains across Amborellales (n = 1), monocot (n = 42), and eudicot (n = 60) genomes available on Phytozome, Ensembl Plants or CoGe. (B) Histogram of numbers of NB-ARC domains identified in angiosperm genomes. (C) Species tree of monocot and eudicot genomes of interest. Number of NLRs with all 6 characteristic NLR amino-acid motifs annotated in each species are displayed in line with the species Latin name. Number of NLRs identified by PfamScan and the plant_rgenes pipeline are in parentheses. Black/red numbers on branches indicate the number of gained/lost NLRs.

Figure 2.

Maximum likelihood phylogeny of NLRs in the 18 representative plant species and selected reference NLRs. The maximum likelihood tree is based on the alignment of the NB-ARC domains of the 18 representative species of A. trichopoda, Z. marina, S. polyrhiza, E. guineensis, A. comosus, P. equestris, O. thomaeum, Z. mays, O. sativa, A. thaliana, A. coerulea, N. nucifera, A. hypochondriacus, S. lycopersicum, F. excelsior, E. guttata, U. gibba, and G. aurea. Bootstraps >80 are indicated by a red dot; branch colors denote species. Clades as defined by bootstrap >80; TNLs and RNLs are within the blue and red sections, respectively. Inlayed sub-trees provide a zoom-in to the TNL, RNL, and CNL clades, where colored branches are indicative of Z. marina, S. polyrhiza, and G. aurea. Sub-tree within gray box is an example of a CNL expansion present in S. polyrhiza.

Figure 3.

Presence/absence analysis of known plant immunity components. Rows denote species, which are arranged as per phylogenetic relationship, with the green and purple bars indicating monocots, or, respectively, dicots. Gene names are listed at the top. Circles in columns denote the presence or absence of known components of the NLR signaling pathway. Black filled circles represent orthologs identified by reciprocal blastp. Orthologs supported by tblastn are indicated by black circles outlined in red. Absent orthologs are displayed by white circles with a red outline, and partial orthologs are shown as gray circles with a black outline. Orthology was also manually curated using Ensembl Plant gene trees or Phytozome synteny (where available).

Figure 4.

Syntenic Blocks of Genomic Loci of ASTREL Genes PAD4, ADR1, and EDS1 Between A. comosus and S. polyrhiza. Genes and their direction are represented as arrows along the loci. Orthologs between A. comosus and S. polyrhiza are indicated by orange lines. Gray triangles with numbers indicate groups of additional genes not displayed here. The focal gene is highlighted with a red outline. Subplots visualize the syntenies of the loci of PAD4 (A), ADR1 (B), and EDS1 (C).

Figure 5.

Protein Family Analysis to Identify ASTREL Genes. (A) Schematic diagram of the OrthoMCL approach to cluster protein families separately among monocot and eudicot species and then filtering for protein families present in species with EDS1. Proteins are denoted by different line drawings, colored by species of origin. Phylogenetic trees represent gene trees for each protein. The diagram at the bottom provides the number of Arabidopsis proteins that are absent only in monocots, eudicots, or in all angiosperms without EDS1. (B) Illustration of the GeneSeqToFamily method, which uses monocot and eudicot proteomes together to establish gene trees across the angiosperms. (C) Diagram summarizing the results of the two methods. Genes and numbers marked in red are those subsequently referred to as ASTREL genes. (D) Schematic of gene clustering followed by blastp and tblastn to filter ASTREL genes for presence or absence in the A. officinalis genome.

Figure 6.

Differential Gene Expression Analysis of Arabidopsis and O. sativa High Confidence ASTREL Genes Upon Biotic and Abiotic Stress. (A) Pearson hierarchical clustering of differential gene expression of ASTREL genes from Arabidopsis upon pathogen, ABA, and nicotinamide treatments. (B) Pearson hierarchical clustering of differential gene expression of ASTREL genes from O. sativa upon pathogen, drought, cold, and salt treatment.

Figure 7.

Schematic Model of Hypothetical Relationships Between ASTREL Genes and Known Biotic and Abiotic Stress Pathways in Arabidopsis. The model is based on literature review and available gene expression of potential interactions of ASTREL genes within the known Arabidopsis biotrophic pathogen disease resistance genetic pathway.