|Home | About | Journals | Submit | Contact Us | Français|
NEP1-like proteins (NLPs) are secreted proteins from fungi, oomycetes and bacteria, triggering immune responses and cell death in dicotyledonous plants. It has been unclear for a long time, whether NLPs are toxins or triggers of plant immunity. In a recent study we report that NLPs are toxins that exert cytolytic activity on dicotyledonous plants. Mutational analysis revealed a causal link between membrane damaging, cell death inducing and virulence promoting properties of NLPs. Interestingly, also induction of immune responses by NLPs required the same protein fold, providing evidence for damage-induced immunity in plants. Structural similarity to pore forming toxins from marine invertebrates allows the proposal of a model for the mode of NLP interaction with the host's membrane.
The family of NEP1-like proteins (NLPs) comprises polypeptides that are secreted by fungi, oomycetes and bacteria. The founding member of this family was isolated in 1995 from the culture medium of Fusarium oxysporum based on its capability to trigger both ethylene production and necrosis in numerous dicotyledonous plants, therefore named NEP1 (necrosis and ethylene inducing peptide 1).1 Follow-up studies identified a number of oomycete and bacterial NLPs, such as PaNIE from Pythium aphanidermatum (NLPPya),2 NPP1 from Phytophthora parasitica (NLPPp),3 PsojNIP from Phytophthora sojae (NLPPs)4 or NIP from Pectobacterium carotovorum subsp. carotovorum (NLPPcc)5 that elicit diverse defense reactions in dicots. Thus, NLPs were commonly referred to as elicitiors of plant immune responses. Intriguingly, it has been repeatedly reported that monocots are not affected by NLPs, although NLPs have been identified from pathogens known to infect exclusively monocots.6
An unclear issue in the past was how NLPs stimulate defense responses in plants. These responses share certain characteristics with PAMP-triggered immunity (PTI). PTI is defined by a receptor-mediated recognition of highly conserved microbial epitopes that are indispensible for microbial life, but that are absent in the host.7 PTI gives rise to the basal immunity of susceptible host plants and determines immunity in non-host plants. NLPs mediate the activation of MAPKs, induction of ion fluxes, the production of phytoalexins and reactive oxygen species, callose deposition and the induction of defense-related genes. These responses resemble to a great extent those triggered by the well-studied PAMP flg22.1–4,8,9 Furthermore, NLPs show a broad taxonomical distribution with high conservation.9–11 However, NLPs also clearly differ from true PAMPs. NLPs are transiently expressed proteins5,9,12 and an elicitor-active minimal motif, as it is found for many PAMPs, could not be identified.3,13 Additionally, genuine PAMPs trigger immunity-associated hypersensitive cell death (hypersensitive response, HR) only in exceptional cases, whereas NLPs caused lesion formation in all dicotyledonous plants tested. Moreover, the NLP-induced cell-death differs genetically from a HR.9
We have recently shown that NLPs are cytolytic toxins.12 Crystallization of NLPPya revealed a structural similarity to actinoporins, which are pore forming toxins from marine invertebrates (see next paragraph). Comparable to actinoporins, a membrane disrupting activity could be demonstrated for NLPs. This activity was restricted to plasma membrane vesicles from dicot plants. In contrast, plasma membranes derived from monocots turned out to be resistant towards NLP application. Structure function analysis revealed that key residues for the cytolytic activity are either important for the coordination of a divalent cation within a surface exposed NLP groove or are involved in the formation of a putative membrane interacting loop (see the next paragraph). Notably, only those mutants that retained the cytolytic activity were capable of restoring the virulence of a Pectobacterium carotovorum strain that lacked its own NLP gene and that was previously shown to be less virulent on potato.5 Thus, our findings implicate that (1) NLPs are virulence factors that (2) are conserved across kingdom borders and that (3) exert their function by mediating membrane disruption, thereby facilitating plant cell death during infection. Having shown that NLPs are toxins, one of the most fascinating questions concerns their capability to alert the plant immune system. We figured out that such an ability of NLPs depends on its membrane disintegrating and hence cell death promoting activity, suggesting that plant cells recognize the toxin's action but not the molecule itself. Thus, plants seem to be capable of sensing the cellular changes induced by membrane damage, such as the release of host-derived endogenous elicitors or changes in ion homeostasis. By stimulating plant defenses through interference with host cell integrity, NLPs resemble triggers of toxin-induced immunity in animals, revealing an additional conceptual similarity in eukaryotic innate immunity.
The crystal structure of NLPPya revealed a single-domain molecule (Fig. 1A) consisting of a central β-sandwich encompassed by three helices, forming a flat surface at the top. In contrast, an uneven surface is established at the base of the polypeptide mainly by three broad loops (L1, L2, L3). Above these loops, a cavity with a highly negative charge carrying a divalent cation is conspicuous. A search for structurally similar proteins using the DALI program (http:///www.ebi.ac.uk/dali/)14 resulted in fungal lectins (XCL from Xerocomus chrysenteron, ABL from Agaricus bisporus)15,16 and actinoporins produced by sea anemones [Equinatoxin II (EqtII) from Actinia equina, Sticholysin (StnII) from Stichodactyla helianthus].17–19 Additionally, upon visual inspection the membrane-interacting domains of Perfringolysin (domain 4, D4) and PKCα (C2-domain) have been found to display a similar topology to NLPPya (Fig. 1A). All these proteins are known to interact with specific structures at the membrane surface. While Perfringolysin attaches to membranes in a sterol dependent manner, lectins and actinoporins are targeting carbohydrates and sphingomyelin, respectively, via a groove at the bottom of the sand wich.15,19,20 Likewise, also the groove of NLPPya would be well suited to target specific components exposed at the surface of membranes. On condition that a divalent cation represents at least one of the physiological ligands, this feature is reminiscent of C2-like β barrel domains which act as calcium-dependent membrane targeting modules in a diverse group of extrinsic membrane proteins.21 Here, a membrane binding mechanism has been proposed in which one calcium ion directly bridges the C2-domain to a specific phospholipid-headgroup.22
Actinoporins are eukaryotic poreforming cytolytic toxins that use a N-terminal helix for host cell penetration. This region is absent in the lectins ABL and XCL, which could explain the lack of cytolytic activity. Based upon combined 3D structural data of StnII, Mancheno et al. proposed a model for the formation of a tetrameric transmembrane pore.19 The individual protomers of a membrane attached tetramer maintain their β sandwich core intact but undergo striking conformational changes in their flexible N-terminal regions, each of which extends to a longer helix that penetrates into the bilayer. Strikingly, NLPPya clearly carries a 29 residue N-terminal extension including a short helix (α1), which has been shown to be essential for the cell death inducing activity of NLPPp.3
Taking these data into account, a plausible model for a membrane attachment of NLPPya can be generated (Fig. 1B). The mainly hydrophobic keel-like structure consisting of W155, P156 and L157 may penetrate into the lipid-core of the membrane similar to the hydrophobic regions of actinoporins and Perfringolysin, whereas the exposed side chains of K54, R64, K92, R102, H103, H128, K132 and K152 constitute a positively charged belt that might interact with the polar headgroups of membrane phospholipids or, alternatively, yet to be identified membrane structures.
Previously published online: www.landesbioscience.com/journals/psb/article/9669