The structural basis of effector recognition and defence signalling in plant disease resistance and susceptibility

Abstract

Listen

Translate article

In plants, cytoplasmic immune receptors known as resistance proteins contain nucleotidebinding and leucine-rich repeat domains and resemble mammalian NOD (nucleotidebinding and oligomerization domain)–like receptors (NLRs). During infection, recognition of the corresponding pathogen effectors by plant NLRs activates effector-triggered immunity (ETI). The ETI response is usually associated with programmed cell death (the hypersensitive response (HR)) and defense gene induction. Many plant NLRs carry an Nterminal Toll/interleukin-1 receptor (TIR) domain that functions in signaling. The TIR domains of some plant NLRs have been shown to trigger HR in an effector-independent manner when overexpressed in plant.In certain cases, plant immunity requires the presence of two NLRs. In Arabidopsis thaliana, both NLRs RPS4 and RRS1 are required to confer resistance against three distinct pathogens. The RPS4 TIR domain and the RRS1 TIR domain interact with each other and also self-associate. Crystal structures reveal that the two TIR domains share a conserved interface for heterodimerization and homodimerization. Disruption of the TIRTIR interface with mutations in the full-length context of RPS4 and RRS1 prevents effector-triggered HR. The RPS4 TIR domain is auto-active and induces effectorindependent HR, which is inhibited by the presence of the RRS1 TIR domain. Mutations in the conserved interface prevent the autoactivity of the RPS4 TIR domain, while mutations in the RRS1 TIR domain interface abolish its inhibition of the RPS4 TIR domain autoactivity. These results suggest that heterodimerization of the TIR domains in the context of full-length RPS4 and RRS1 is required to maintain the RPS4-RRS1 complex in a proper state for effector recognition. The ensuing conformational change after effector recognition may release the RPS4 TIR domain from interacting with the RRS1 TIR domain and allow the RPS4 TIR domain to homodimerize and initiate the signalling. In this thesis, the RRS1 TIR domain was purified and crystallized, and its crystal structure was determined to help design mutations to affect its heterodimeriztion with the RPS4 TIR domain. These results are described in chapters 2 and 3.Flax-rust fungus (Melampsora lini) secretes the effector protein AvrL567 into flax (Linum usitatissimum) cells during infection. There are 12 genetic variants of AvrL567 and some can be recognized by flax NLRs L5, L6 or L7 to trigger ETI. Currently, the function of AvrL567 during flax rust infection is unknown. AvrL567-A was identified to physically interact with the flax cytokinin oxidase LuCKX1.1. Compared with the wide-type plants, the transgenic flax plants that overexpress AvrL567-A adopt a dwarfed phenotype with larger root systems, indicating imbalanced cytokinin levels in vivo. Cytokinins are a class of plant hormones that regulate cell division, plant development and plant immunity. Cytokinin oxidases catalyse the irreversible degradation of cytokinins. Pathogen infections have been shown to manipulate plant cytokinin levels to promote pathogen virulence. In this thesis, the LuCKX1.1 protein was purified and crystallized, and its crystal structure was determined. End-point assays were performed to investigate the substrate specificities of LuCKX1.1. Kinetic assays revealed that the enzyme activity of LuCKX1.1 increases in the presence of AvrL567-A. Utilizing the structure of LuCKX1.1 and previously determined AvrL567-A and AvrL567-D structures, docking studies were performed to investigate the interaction interface and suggested that AvrL567 A binds in the vicinity of the opening of LuCKX1.1 substrate channel to modulate the enzyme activity. These results are described in chapters 4 and 5. Plant NLRs detect pathogen effectors either by physical interaction or by monitoring intracellular perturbations caused by effectors. The LRR domain of plant NLRs represents the most polymorphic domain and is suggested to confer effector recognition specificity. At present, no structure of the LRR domain from a plant NLR has been reported yet. This is a likely result of reported difficulties in obtaining high-quality proteins of full-length plant NLRs and their LRR domains. To overcome the barriers in the purification of plant NLR LRR domains, a LRR hybrid strategy involving fusion of LRR modules from internalin A, a bacterial LRR-containing protein, was investigated and found to greatly improve the expression of plant NLR LRR domains in the E. coli system. In the study, the LRR domains from flax NLRs L6 and L8 were investigated using the LRR hybrid strategy. Soluble hybrid proteins of the two NLRs have been obtained. The LRR hybrids of L6 that contained 26 and 12 predicted LRR modules expressed poorly and were purified with poor purity and with a tendency to aggregate. The LRR hybrids of L6 and L8 that contained 6 and 8 predicted LRR modules, respectively, expressed well and were purified in good solubility and with much less impurities. The L6 and L8 LRR hybrids predominantly formed oligomers with a small fraction of the proteins migrating as expected for monomers. One single-residue mutation in the L8 LRR hybrid was identified to force the protein into monomeric species. These results are described in chapter 6.