Abstract
Plant and animal nucleotide-binding domain and leucine-rich repeat-containing (NLR) proteins often function in pairs to mediate innate immunity to pathogens. However, the degree to which NLR proteins form signaling networks beyond genetically linked pairs is poorly understood. In this study, we discovered that a large NLR immune signaling network with a complex genetic architecture confers immunity to oomycetes, bacteria, viruses, nematodes, and insects. The network emerged over 100 million years ago from a linked NLR pair that diversified into up to one half of the NLR of asterid plants. We propose that this NLR network increases robustness of immune signaling to counteract rapidly evolving plant pathogens.
Plants and animals rely on nucleotide-binding domain and leucine-rich repeat-containing (NLR) proteins to activate immune responses to invading pathogens (1-3). NLR are among the most diverse and rapidly evolving protein families in plants (4, 5). They are modular proteins that broadly fall into two classes based on their N-terminal domain, which is either a Toll-interleukin 1 receptor (TIR) or a coiled coil (CC) domain (6). Most plant disease resistance genes encode NLR receptors that detect effector proteins secreted by pathogens either by directly binding them or indirectly via effector-targeted host proteins (3, 7). An emerging model is that “sensor” NLR dedicated to detecting pathogen effectors require “helper” NLR to initiate immune signaling resulting in a hypersensitive cell death response that restricts pathogen invasion (8-12). Although paired NLR have been described across flowering plants, the degree to which plant NLR have evolved to form higher order networks is poorly known.
The Solanaceae form one of the most species-rich plant families that includes major agricultural crops, such as potato, tomato and pepper (13). Solanaceae genomes harbor hundreds of NLR-type genes, over 20 of which have been demonstrated to confer resistance to infection by diverse and destructive pathogens and pests, including the Irish potato famine agent Phytophthora infestans (14, 15). As part of a study performed in Nicotiana benthamiana to identify genetic components required for resistance to P. infestans conferred by the potato NLR-type gene Rpiblb2 (16, 17), we discovered that another NLR protein, NRC4 (NLR required for cell death 4) (9, 18), is required for Rpi-blb2 function (Fig. 1). Silencing of NRC4 compromised Rpi-blb2 resistance to P. infestans (Fig. 1A) and hypersensitive cell death to the P. infestans AVRblb2 effector (Fig. 1B) (17). This phenotype was rescued by a silencing-resilient synthetic NRC4 gene (Fig. 1C-D, Fig. S1A-B). NRC4-silencing did not affect Rpi-blb2 accumulation (Fig. S1C). Mutation in the ATP-binding p-loop motif of both Rpi-blb2 and NRC4 abolished their activities (Fig. S2). Thus a strict sensor/helper model where only one NLR requires ATP binding (19, 20) is too simple to explain the interaction between the Rpi-blb2/NRC4 pair.
NRC4 defines a distinct clade within the NRC family (Fig. S3A)(18). Of the 9 NRC genes in N. benthamiana, four were expressed to significant levels in leaves but only NRC4 transcript levels were reduced in NRC4-silenced plants (Fig. S1D, Fig. S3B). Among the expressed genes, NRC2 and NRC3 are required for bacterial resistance mediated by the NLR protein Prf in N. benthamiana (9, 21) but were not essential for Rpi-blb2 functions in our silencing experiments (Fig. 1A-B). In contrast, NRC4 was not essential for Prf-mediated cell death and resistance to the bacterial pathogen Pseudomonas syringae (Fig. 1B; Fig. S4).
Phylogenetic analyses of the complete repertoire of NLR proteins from the solanaceous plants tomato, potato, pepper and N. benthamiana revealed that the NRC family groups with the Rpiblb2 and Prf clades in a well-supported superclade (Fig. S5). Interestingly, this superclade includes additional well-known NLR, such as Rx (22, 23), Bs2 (24), R8 (25), Sw5b (26), R1 (27) and Mi-1.2 (28), which confer resistance to diverse plant pathogens and pests (Fig. S5; Table S1). This prompted us to test the extent to which NRC proteins are involved in immune responses mediated by these phylogenetically related disease resistance proteins.
Silencing of NRC2 and NRC3 only affected Prf and did not alter the hypersensitive cell death mediated by 13 other NLR proteins (Fig. 2). In contrast, silencing of NRC4 compromised the hypersensitive cell death mediated by Mi-1.2 (28), an Rpi-blb2 ortholog that provides resistance to nematodes and insects; CNL-11990D474V (18), an autoactive mutant of a CNL (NLR with a N-terminal coiled-coil domain) of unknown function, and R1 (27), an NLR that confers resistance to P. infestans (Fig. 2, Fig. S6A). Further, NRC4 silencing abolished R1-mediated disease resistance to P. infestans and the phenotype was rescued by a silencing-resilient synthetic NRC4 gene (Fig. S6B-D).
Given that the three expressed NRC proteins share extensive sequence similarity (Fig. S7), we hypothesized that NRC2, NRC3 and NRC4 are functionally redundant for additional NLR in the “NRC” superclade (Fig. 2). To test our hypothesis, we simultaneously silenced the three NRC genes and discovered that silencing compromised hypersensitive cell death mediated by Sw5b, R8, Rx and Bs2 in addition to the 5 NLR mentioned above (Fig. 2, Fig. S8, Fig. S9). In contrast, the triple NRC silencing did not affect hypersensitive cell death mediated by the 5 tested NLR that map outside the NRC superclade (Fig. 2) and did not abolish resistance to P. infestans conferred by two of these NLR proteins (Fig. S10).
We validated NRC2, NRC3 and NRC4 redundancy by complementation in the triple silencing background with silencing-resilient synthetic NRC (Fig. S11). This confirmed that the three NRC proteins display specificity to Rpi-blb2 and Prf but have redundant functions in Rx, Bs2, R8 and Sw5b mediated hypersensitive cell death (Fig. S11).
To further validate that NRC2, NRC3 and NRC4 redundantly contribute to immunity, we examined the resistance mediated by Rx to Potato virus X (PVX) (22, 23) in plants silenced for single, double or triple combinations of NRC genes (Fig. S12). Rx-mediated resistance to PVX was only abolished in the triple silencing background resulting in systemic spread and accumulation of the virus (Fig. S12, Fig. S13). Remarkably, silencing-resilient synthetic NRC2, NRC3 and NRC4 individually complemented the loss of resistance to PVX in triple NRC-silenced plants confirming their functional redundancy (Fig. S14). This and previous results indicate that the three NRC proteins display varying degrees of redundancy and specificity towards the 9 NLR revealing a complex immune signaling network (Fig. S15).
Our observation that NRC proteins and their NLR mates are phylogenetically related (Fig. S5) prompted us to reconstruct the evolutionary history of the NRC superclade. Higher order phylogenetic analyses of complete NLR repertoires from representative plant taxa revealed that the NRC superclade is missing in rosids but present in all examined caryophyllales (sugar beet) and asterids (kiwifruit, coffee, monkey flower, ash tree and Solanaceae species) (Fig. S16, Fig. 3A-B, Fig. S17-20). Interestingly, sugar beet and kiwifruit, the early branching species, have only a single protein that groups with the NRC family, along with 2 and 4 NLR that cluster with the NRC-dependent NLR (Fig. 3A-B, Fig. S20). The dramatic expansion of the NRC superclade started prior to the divergence of Gentianales (coffee) from other asterids about 110-100 million years ago (29, 30) to account for over one half of all NLR in some of the species (Fig. 3B). We conclude that the NRC superclade evolved from an ancestral pair of genetically linked NLR genes, as in sugar beet, to duplicate and expand throughout the genomes of asterid species into a complex genetic network that confers immunity to a diversity of plant pathogens (Fig. 3C-D, Fig. S21).
Thus, NLR pairs can evolve into a signalling network with a complex architecture. However, NLR evolution must be constrained by their mode of action. The selective pressures shaping the evolution of NLR pairs that operate by negative regulation would limit their expansion due to the genetic load caused by autoimmunity (Fig. S22). The NRC family appears to function through a mechanism that accommodates evolutionary plasticity beyond genetically linked pairs of NLR.
Genetic redundancy increases robustness of signaling networks (31-33). The NRC network may therefore augment the plant capacity to counteract rapidly evolving pathogens. Multiple NRC would further enhance evolvability of sensor NLR, i.e. their capacity to undergo rapid adaptive evolution. Harnessing the processes that underpin NLR network structure and function would open up new approaches for developing disease resistant crops.
Acknowledgments
We thank Oliver Furzer, Jonathan Jones, John Rathjen, Brian Staskawicz, Geert Smant, Sebastian Schornack, Frank Takken, Vivianne Vleeshouwers and Cyril Zipfel for providing materials and technical supports. We are grateful to Lida Derevnina, Yasin Dagdas, Benjamin Petre, Erin Zess and Esther van der Knaap for helpful suggestions. This project was funded by the Gatsby Charitable Foundation, Biotechnology and Biological Sciences Research Council (BBSRC), and European Research Council (ERC).