A root-specific NLR network mediates immune signaling of resistance genes against plant parasitic nematodes

Abstract

Plant nucleotide-binding domain and leucine-rich repeat immune receptors (NLRs) confer disease resistance to many foliar and root parasites. However, the extent to which NLR-mediated immunity is differentially regulated between plant organs is poorly known. Here, we show that a large cluster of tomato (Solanum lycopersicum) genes, encoding the cyst and root-knot nematode disease resistance proteins Hero and MeR1 as well as the NLR helper NLR required for cell death 6 (NRC6), is nearly exclusively expressed in the roots. This root-specific gene cluster emerged in Solanum species about 21 million years ago through gene duplication of the ancient asterid NRC network. NLR sensors in this gene cluster function exclusively through NRC6 helpers to trigger hypersensitive cell death. These findings indicate that the NRC6 gene cluster has sub-functionalized from the larger NRC network to specialize in mediating resistance against root pathogens, including cyst and root-knot nematodes. We propose that some NLR gene clusters and networks may have evolved organ-specific gene expression as an adaptation to particular parasites and to reduce the risk of autoimmunity.

Figure 1.

Tomato NLRs form clusters of phylogenetically related or unrelated NLR genes. A) Schematic representation of the computational pipeline employed for predicting clustering of sensor-helper NLRs. The genetic distances among tomato (S. lycopersicum) NLR genes and their phylogenetic relationships were assessed using a custom Python script available at https://github.com/slt666666/gene-cluster-matrix. B) NLR matrix integrating gene distance and phylogenetic relationships. The color scale bar on the right indicates distance between genes in kilo bases (kb). Genetically linked NLR genes (distance < 50 kb) are highlighted in the matrix, with colors representing gene distance and ordering based on phylogeny. The branches of major phylogenetic NLR clades (Wu et al. 2017; Kourelis et al. 2021) are color-coded, and NRC helpers (NRC-H) and NRC sensors (NRC-S) are depicted . The NRC sensor clade is further subdivided into Rx-type (Rx) or Solanaceous domain (SD)-containing NRC sensors. An interactive HTML version of the matrix containing detailed distance information on all NLR genes is provided as Supplementary Data S2. Coiled-Coil-type (CC), other CC-type (CC-other), G10-type CC (CC_G10), RESISTANCE TO POWDERY MILDEW 8 (RPW8)-type (CC_R), Toll/interleukin-1 receptor-type (TIR), Nucleotide-binding adaptor shared by APAF-1, certain R gene products, and CED-4 (NB-ARC), Leucine-rich repeat (LRR). C) Phylogenetic tree with black lines connecting NRC helper and NRC sensor NLRs encoded in gene clusters containing phylogenetically unrelated NLRs (mixed-clade gene clusters). Branches of the tree are color-coded according to B to indicate the phylogenetic clade. The scale bar represents evolutionary distance, measured as substitutions per site. Gene distance information, alignments, and phylogenetic tree files are available as Supplementary Data Set 1.

Figure 2.

Hero and NRC6 form distinct phylogenetic subclades within the NRC helper and SD-type NRC sensor clades. A) The phylogenetic tree of NLRs was constructed based on the Nucleotide-binding adaptor shared by APAF-1, certain R gene products, and CED-4 (NB-ARC) domain from a large NLR sequence database (Supplementary Data Set 2). A total of 20.292 NLR sequences was aligned using MAFFT and the phylogeny was inferred using FastTree to determine the NRC network clades of NRC helpers, Rx-type NRC sensors and Solanaceous domain (SD)-type NRC sensors (top left). Sub-trees for the NRC helper (bottom left) and SD-type NRC sensor (top right) subclades were generated, containing well-defined phylogenetic clades including the Hero resistance protein and NRC6, respectively. The HCN clade is further divided into a tomato sub-clade (gray) which contains branches defined by proteins encoded in the NRC6 and HCN containing mixed-clade gene cluster (bottom right). The scale bar represents evolutionary distance, measured as substitutions per site. Coild-coil (CC), Leucine-rich repeat (LRR). B) Schematic depiction of the NRC6 and HCN encoding mixed-clade gene clusters of S. lycopersicum Heinz 1706 and LA1792, which contains an introgressed cluster from S. pimpinellifolium LA121 (Ellis and Maxon Smith 1971; Ernst et al. 2002). The HCNs encoded in this gene cluster define the branches of the tomato sub-clade shown in (A, bottom right). Gene IDs for the Heinz reference annotation are provided. Pseudogenes containing premature stop codons (Ernst et al. 2002) are indicated by an asterisk. SD-type NRC sensors are depicted in dark blue, NRC helpers in red, and non-NLRs in gray. Sequences, alignments, and phylogenetic tree file for A are available as Supplementary Data Set 2 or https://github.com/amiralito/Hero (Toghani et al. 2025). The LA1792 HCN sequences and cluster annotation is available under https://zenodo.org/records/10376142 (Lüdke et al. 2023).

Figure 3.

HCNs require NRC6 helper NLRs to trigger a hypersensitive cell death in Nicotiana benthamiana. A) The schematic representation illustrates the genetic complementation scheme employed throughout this experiment. Agrobacteria carrying NRC6a, NRC6b, or the specified HCN expression constructs were infiltrated into N. benthamiana leaves to express wildtype or autoactive HCNs and NRC6 helper NRCs (HCN^MHD or NRC6^MHD, respectively) in the indicated combinations. B) Representative cell death phenotypes in N. benthamiana leaves induced by HCNs when co-expressed with NRC6a or NRC6b, photographed at 7 days post-infiltration (dpi) with agrobacteria. A p19 silencing construct was co-expressed in every infiltration. Cell death was visually scored and statistically analyzed. Data points are depicted as dots, with each biological replicate represented by a different color. The central circle for each cell death category proportionally represents the total number of data points for each treatment. Scoring is shown only for wildtype and autoactive HCNs, co-expressed with wildtype NRC6a or NRC6b helpers. The scale bar represents 1 cm. Each biological replicate (Rep) consists of 2 leaves from 3 different plants each. Complete quantification and statistical analysis are presented in Supplementary Fig. S5.

Figure 4.

NRC2, NRC3 and/or NRC4-dependent disease resistance proteins do not signal through NRC6. The schematic representation illustrates the genetic complementation scheme employed throughout this experiment. Agrobacteria carrying expression constructs for sensors, helpers, or avirulence (AVR) effector genes were infiltrated in indicated combinations into the leaves of N. benthamiana nrc2/3/4 mutant plants to trigger cell death. For Mi-1.2, CNL11990, Rx, and Sw5b, autoactive sensor mutants were expressed instead of a corresponding AVR effector gene (indicated by an asterisk). Tomato NRC3 or NRC4a was co-expressed as a positive control for complementation, respectively (Wu et al. 2017). HCN-H was used as a positive control for NRC6b-dependent cell death. Leaf images were photographed at 7 days post-infiltration (dpi) with Agrobacteria, a p19 gene silencing construct was co-expressed in each infiltration. The scale bar represents 1 cm. Quantification and statistical analysis are presented in Supplementary Fig. S7.

Figure 5.

Hero and the HCNs do not signal through the other 9 tomato NRC helpers. The schematic representation illustrates the genetic complementation scheme employed throughout this experiment. Agrobacteria carrying autoactive HCNs (HCN^MHD) and the indicated wildtype tomato NRC helper expression constructs were co-infiltrated into the leaves of N. benthamiana nrc2/3/4 mutant plants. As a positive control, Rx was co-expressed with all tomato NRC helpers. Leaves were photographed at 7 days post-infiltration (dpi) with Agrobacteria, a p19 silencing construct was co-expressed for each infiltration. The scale bar represents 1 cm. Quantification and statistical analysis are presented in Supplementary Fig. S8.

Figure 6.

The root-knot nematode resistance gene MeR1 is a HCN that requires NRC6 to induce a hypersensitive cell death. A) A phylogenetic sub-tree was generated with MeR1 and tomato and potato NLRs defined as HCNs, from data presented in Fig. 2A. Re-alignment was performed using MAFFT, the phylogeny was inferred using FastTree. The scale bar represents evolutionary distance, measured as substitutions per site. B) Agrobacteria harboring wildtype or autoactive MeR1, or NRC helpers, were infiltrated in the indicated combinations into leaves of N. benthamiana wildtype (left panel) or nrc2/3/4 mutant (right panel) plants for co-expression. Leaves were photographed at 7 days post-infiltration (dpi) with Agrobacteria, a p19 silencing construct was co-expressed for each infiltration. The scale bar represents 1 cm. Quantification and statistical analyses are presented in Supplementary Fig. S10. Sequences, alignments and the phylogenetic tree file are available as Supplementary Data Set 3.

Figure 7.

HCNs and NRC6 helper genes are nearly exclusively expressed in tomato roots. A) Histograms depicting the log2-fold change frequency, comparing the expression levels of all tomato genes, all NLRs, Coiled-Coil-type (CC), and Toll/interleukin-1 receptor-type (TIR) phylogenetic NLR subclades between roots and leaves of 2-week-old unchallenged tomato plants of the Hero introgression line LA1792 (Ellis and Maxon Smith 1971; Ernst et al. 2002). The dotted magenta line and numbers below labels indicate the mean log2-fold change, the skewness is indicated in brackets. Normalized counts for the expression of cluster-encoded HCN and NRC6 genes B), and the expression of all tomato NRC clade genes C) in the roots and leaves of 2-week-old unchallenged LA1792 tomato plants. The log2-fold change and adjusted P-value are indicated for each gene. Pseudogenes are indicated by an asterisk. HCN-G and HCN-E pseudogenes exhibit negligible expression. Numbers for each gene are presented in Supplementary Table S1. The adjusted P-values (adjP) were calculated using DESeq2 with the Wald test and Benjamini-Hochberg correction for multiple testing. Complete data for log2-fold change values and normalized counts of all tomato genes and NLR subsets are available as Supplementary Datas S5 and S6.

Figure 8.

Schematic overview of organ-specific expression patterns in the NRC immune receptor network. Functional connections within the NRC network, involving known NRC sensors (NRC-S) and helpers (NRC-H), are represented based on the work of Wu et al. 2017, along with the results presented in this study. Nodes in the network are shaded to reflect their expression levels in either leaves (top) or roots (bottom), as indicated by the color scale. Expression levels are based on the mean normalized reads for each gene, derived from RNA-seq data (Supplementary Data S6) obtained from 2-week-old unchallenged tomato plants of the Hero introgression line LA1792 (Ellis and Maxon Smith 1971; Ernst et al. 2002). For NRC sensor genes from other plant species, expression data for the closest tomato homolog is displayed (Supplementary Fig. S15).

Figure 9.

Key steps in the evolution of the NRC network. The ancestral NRC pair is thought to have emerged more than 125 million years ago (MYA) (Sakai et al. 2024). We propose that, unlike the ancient NRC0 gene cluster, the NRC6 and HCN gene cluster evolved approximately 17 to 21 MYA, prior to the diversification of the Solanum genus but after the lamiid expansion of the NRC network (Wu et al. 2017; Goh et al. 2024; Sakai et al. 2024). NRC6 and the HCNs formed a gene cluster through genetic duplication and diversification events during the evolution of Solanum species. Leaf and root symbols are used to indicate the expression pattern of the respective NLRs. NRC helpers (NRC-H), NRC sensors (NRC-S).