Effector-triggered defence against apoplastic fungal pathogens

1Breeding agricultural crops for resistance against pathogens is essential to secure global food production. Despite efforts to control crop diseases, pathogens are estimated to account for losses of 15% of global food production. It is suggested that losses would be almost twice as much without disease control measures, such as crop resistance breeding. There are now opportunities to improve the effectiveness of breeding crops for resistance against damaging pathogens by exploiting new molecular and genetic insights to improve understanding of the defence system of crop plants against pathogens. In this opinion, we focus on the resistance of crops against foliar fungal pathogens that exploit the host apoplast for retrieval of nutrients. Some of these pathogens are globally widespread and of considerable economic importance. They include pathogens that penetrate the host leaf cuticle and then exploit a niche underneath it (e.g., Pyrenopeziza brassicae, oilseed rape light leaf spot; Venturia inaequalis, apple scab; and Rhynchosporium commune, barley leaf blotch, global losses of approximately US$3.5 billion per year). Others enter leaves through stomata, then grow between host mesophyll cells (e.g., Cladosporium fulvum, tomato leaf mould; Leptosphaeria maculans, oilseed rape phoma stem canker, global losses of approximately US$1 billion per year; and Zymoseptoria tritici, wheat septoria leaf blotch, with a global loss of approximately US$5 billion per year) (Table 1, Figure 1). These apoplastic pathogens are all ascomycetes and many of them are dothideomycetes.

2The plant immune system was originally defined as including PTI and ETI. After arrival of a potential pathogen, PTI is rapidly activated by extracellular host recognition of PAMPs by pattern recognition receptors (PRRs); most PRRs are receptor-like kinases with an extracellular receptor domain, a transmembrane domain, and typically a cytoplasmic kinase domain. Pathogens produce effectors to suppress PTI and cause effector-triggered susceptibility (ETS). In resistant hosts, ETI is initiated by intracellular host recognition of pathogen effectors. The term ‘PAMPs’ was originally used to describe invariant molecules present in many microbial pathogens that are essential for their survival. PAMPs that fit this definition include flagellin and chitin, which are present in bacteria and fungi, nematodes or insects, respectively. It was proposed to reduce the stringency of this original definition by considering conserved effectors as not significantly different from PAMPs, without knowledge of their essential functions. PTI has sometimes been defined as PRR-triggered immunity. In addition to their original role in PAMP recognition, it was proposed that PRRs directly or indirectly recognise effectors produced by apoplastic pathogens. However, it is not universally accepted that resistance against apoplastic pathogens should be considered as PTI; some consider it as ETI. Furthermore, this new definition of PTI does not account for the fact that some R genes encode proteins that are similar to PRRs but interact with extracellular effectors. To explain interactions between apoplastic fungal pathogens and their host plants and reconcile these opposing views, it is proposed that this resistance is referred to as ETD.

3Given that this resistance does not generally eliminate these apoplastic pathogens (in the cases of P. brassicae and Z. tritici, the pathogen is subsequently able to complete its sexual life cycle on the resistant host ), ‘defence’ is a more appropriate term than ‘immunity’, which is generally associated with death of the pathogen (Table 2). The response is effector triggered and mediated by R gene-encoded proteins, thus justifying ETD as a relevant term. Whereas PTI operates against both apoplastic and cell-penetrating filamentous pathogens (referred to as ‘haustorial pathogens’), the mode of R gene-mediated resistance against them differs between the two types of pathogen (Box 1).

4In contrast to ETI against haustorial pathogens, ETD responses against apoplastic pathogens are relatively slow and not associated with a fast, orchestrated hypersensitive host cell death response (HR) (Figure 2A–F). Effectors of apoplastic pathogens are recognised at the cell surface, whereas those of haustorial pathogens are recognised within the cell cytoplasm. The receptors that initiate ETI or ETD differ not only in their domain structures but also in their interactions with other proteins and signalling partners. R gene-mediated ETD has some similarities to PTI, in that both are initiated at the cell surface (Box 1).

5ETD operates against apoplastic pathogens that are adapted to colonise the intercellular matrix, relying on effectors secreted into the apoplast and not forming haustoria (Table 1, Figure 1E1–F2). Conversely, ETI works against obligate biotrophic filamentous pathogens that produce haustoria (Figure 2E,F) to retrieve nutrients from living plant cells (e.g., Blumeria graminis, barley powdery mildew; Bremia lacucae, lettuce downy mildew; and Puccinia striiformis, wheat yellow rust), and some hemibiotrophic oomycetes (e.g., Phytophthora infestans, potato late blight; Figure 1I–J2) and fungi (e.g., Magnaporthe grisea, rice blast).

6ETD has been studied with plant hosts of several apoplastic foliar fungal pathogens. Those pathogens listed in Table 1 are all ascomycetes and, except for P. brassicae and R. commune, they are dothideomycetes. Although some dothideomycetes are saprophytes, many are epiphytes, endophytes, or pathogens of plants. Apoplastic pathogens, such as C. fulvum, L. maculans, Z. tritici, P. brassicae, V. inaequalis, and R. commune, do not trigger an effective host defence response during the early stages of invasion (Table 1, Figure 2A–C) but slowly colonise the apoplast to adapt to the constitutive antimicrobial compounds present. For example, Z. tritici grows seemingly undetected inside its wheat (Triticum aestivum) host for 10–13 days after inoculation.

7Apoplastic pathogens overcome constitutive and induced basal plant defences (PTI) using both offensive and defensive strategies. Among the offensive strategies are: (i) direct penetration of cuticles, using cutinases, by subcuticular pathogens such as P. brassicae; and (ii) entry into leaves through stomatal pores by intercellular pathogens such as L. maculans (Figure 1H1–2). Furthermore, they also constitutively produce enzymes such as tomatinase and proteases. For example, the antifungal activity of the tomato (Solanum lycopersicum) glycoalkaloid α-tomatine in its apoplast is overcome by C. fulvum detoxifying it into the less fungitoxic compounds tomatidine and oligosaccharide lycotetraose. The lycotetraose released might even serve as a carbon source for the pathogen. A C. fulvum strain deficient in GH10 tomatinase was less virulent, particularly during later stages of colonisation around 10 days post-inoculation (dpi). At this later stage, the biomass of wild type C. fulvum increased dramatically. Another offensive mechanism is the destruction by cleavage of plant constitutive or induced class IV chitinases that can hydrolyse fungal cell walls.

8Among the defensive strategies, protection against constitutive apoplastic tomato chitinases is achieved by C. fulvum secreting into the apoplast the Avr4 effector that binds to chitin in its cell walls to protect them from hydrolysis. Homologues of Avr4 occur in several dothideomycetes, making Avr4 a common effector gene. It is conceivable that protection of fungal cell walls by Avr4 against chitinases is incomplete and that some chito-oligosaccharides are still released. Chito-oligosaccharide-triggered basal plant defence is avoided by some apoplastic pathogens because they produce chitin-binding effectors that sequester these oligosaccharides. Cladosporium fulvum produces a LysM domain-containing effector protein, Ecp6, that sequesters chito-oligosaccharides, preventing them from inducing chitin-triggered immunity. Zymoseptoria tritici also secretes different types of LysM effector during the endophytic phase of plant colonisation by the pathogen, either to protect itself against plant chitinases or to sequester chitin-oligosaccharides to prevent PTI. LysM effectors have also been found in other types of pathogen, including M. grisea. Ecp6 occurs even more widely than Avr4. Thus, effectors of apoplastic fungi target basal defence components occurring in the apoplast.

9The numerous proteases expressed by Z. tritici during the endophytic phase may compromise host defence enzymes, such as chitinases and β-glucanases, and amino acids released might serve as pathogen nutrients. Basal wheat defence is avoided by Z. tritici through production of a specific set of adapted plant cell wall-degrading enzymes (CWDEs). Host-adapted CWDEs are produced during the early endophytic phase of colonisation. During the subsequent necrotrophic growth phase, Z. tritici produces a different set of CWDEs to permit extensive colonisation of the host. However, these CWDEs are initially expressed at low concentrations during the endophytic growth phase to avoid recognition by the plant immune system.

10By contrast, basal host defences against cell-penetrating, haustoria-forming pathogens are suppressed by effectors that are active in the host cytoplasm; these effectors can induce an intracellular ETI resistance response (Figure 2D–F). Furthermore, a similar combination of strategies, but with entry associated with hyphopodia or penetration of stomatal pores, is used by host-specific necrotrophic fungal pathogens that deliver host-selective toxins (HSTs) inside plant cells to induce ETS (Figure 2G–I).

11In contrast to the fast timing of ETI in the host cytoplasm (Figure 2D–F), or ETS (Figure 2G–H; e.g., Phaeosphaeria nodorum, Figure 1K,L), triggered by haustoria or biotrophic interfacial complex (BIC)-forming biotrophic hemibiotrophic, or necrotrophic filamentous pathogens, respectively (Table 2), triggering of extracellular ETD (Figure 2A–C) by apoplastic fungal pathogens is generally slow. This is because the maximum expression of their effector genes occurs only after an initial phase of endophytic growth and the effectors often need to be processed to exert their virulence function or to be recognised by RLPs (Table 1). Although expression of the Avr4 effector gene has been detected as early as 1 day after stomatal entry by C. fulvum, maximum expression occurs at 4–6 dpi; a similar pattern is observed for Avr9. Both Avr4 and Avr9 effectors of C. fulvum contain a functional glycosylation site (NSS for Avr9 and NLS for Avr4). Glycosylation of both effectors may affect their ability to induce ETD and their recognition by RLPs.

12Similarly, ETD defence responses by oilseed rape (Brassica napus) against L. maculans colonisation are slow because expression of the AvrLm1, AvrLm6, AvrLm4-7, and AvrLm11 effectors of L. maculans (Figure 1G1–H4) is limited at 3 dpi, reaching a maximum at 7 dpi. Few pathogen hyphae are present at 3 dpi but by 7 dpi the mycelial mass has increased. A microscopic reaction involving cell death occurs 8 days after leaf inoculation of B. napus cv. DarmorMX (carrying the resistance gene Rlm6) with ascospores of a L. maculans isolate carrying the corresponding AvrLm6 gene. However, there has been subcellular evidence of necrosis occurring by 6 days after infiltration of oilseed rape cotyledons with L. maculans conidia. An AvrLm1-dependent increase in expression of the pathogenesis-related (PR) gene PR1 was detected at 5 days after infiltration with conidia, at the same time as an increase in salicylic acid production. These host defence responses against L. maculans do not eliminate the pathogen, which can be re-isolated from resistant host tissue and cultured.

13ETD against Z. tritici is also expressed only after the pathogen has entered wheat leaves through stomata and grown endophytically between mesophyll cells in leaves for the first week after inoculation. ETD operates in the resistant host at approximately 10 dpi when the rapid switch from endophytic to necrotrophic growth occurs in the susceptible host, but this ETD is not associated with host cell death.

14Plant defence against apoplastic pathogens that occupy a subcuticular niche is also generally slower than defence against haustoria or BIC-forming filamentous pathogens. Although the effect of R gene-mediated resistance could be detected as early as 3 dpi of barley (Hordeum vulgare) leaves with R. commune (Figure 1C1–D3), cell death only occurred at 21 dpi. In the presence of the resistance gene Rrs1, both subcuticular growth and sporulation of an avirulent isolate were impeded. Expression of PR genes PR1, PR5, PR9, and PR10, all of which encode extracellular proteins, is induced as early as 1 dpi in the Rrs1-containing genotype that is responsive to the corresponding effector NIP1. In the susceptible host, induction of these PR genes occurs 2 days later.

15Expression of oilseed rape resistance against P. brassicae (Figure 1A1–B3) is also slow, operating from 13 to 36 dpi to prevent a 300-fold increase in pathogen biomass. The operation of resistance, associated with the PBR2 gene, triggers a collapse of epidermal cells or ‘necrotic flecking’ accompanied by little or no asexual sporulation. However, this resistance does not interfere with subsequent sexual sporulation on senescent leaves 36 dpi.

16These slow ETD responses are in contrast to ETI induced by haustoria or BIC-forming filamentous pathogens. Expression of effector genes in these pathogens evokes a fast HR, generally resulting in death of the host cells and the pathogen. Effector genes of P. infestans, for instance, reach a maximum expression 2 dpi of potato (Solanum tuberosum). HSTs, which are effectors of host-specific necrotrophic fungal pathogens, are even expressed constitutively (Table 2). Effector recognition is equally fast, triggering host cell death within 1–2 dpi of wheat by P. nodorum or oats (Avena sativa) by Cochliobolus victoriae (Table 2).

17ETD involves host recognition of apoplastic effectors in intercellular spaces by cell surface RLPs that are integral plant membrane proteins containing an extracellular leucine-rich repeat (eLRR) domain and a short cytoplasmic tail without a signalling motif (Figure 2A). Recognition of apoplastic effectors and triggering of RLP-mediated defence is described best in the tomato–C. fulvum pathosystem (Table 1). Avr2 is a cysteine protease inhibitor that binds to and inhibits the plant protease Rcr3. The corresponding RLP, Cf-2, acts as a guard to survey the modulation of Rcr3. Cf-2 probably recognises the Avr2-modulated Rcr3 protein because Avr2 mutants with reduced Rcr3 binding are impaired in their ability to trigger Cf-2-mediated host cell death. Three RLPs, Cf-2, Cf-4, and Cf-9, were all shown to interact with the receptor-like kinase (RLK) SOBIR1/EVR for downstream signalling and defence. It is suggested that SOBIR1 functions specifically in receptor complexes with RLPs involved in RLP-mediated ETD and RLP-mediated plant development (CLV2); SOBIR1 does not engage with BAK1 to trigger PTI (Box 1). However, it should be noted that complexing of SOBIR1 with RLPs has usually been studied in the absence of their corresponding ligands and it is expected that ligand-triggered RLP-mediated defence or development responses require new complex associations or dissociations to amplify defence or development signalling. It has also been suggested that SOBIR1 is involved in stabilisation and trafficking of RLPs.

18RLPs are also involved in oilseed rape resistance against L. maculans and apple (Malus domestica) resistance against V. inaequalis (Table 1). AvrLm1 encodes a secreted mature protein of 21 kDa with a single cysteine residue. AvrLm1 genetically interacts with LepR3, the first R gene cloned from B. napus to encode an RLP. One of four R gene paralogues, HcrVf2, is necessary and sufficient for apple resistance against V. inaequalis strains that carry the cognate effector gene AvrRvi6. It is also likely that NIP1, an effector of the barley leaf blotch pathogen R. commune, is recognised at the cell wall because it binds to a plasma membrane protein and stimulates H+-ATPase activity. Resistance to NIP1-producing strains of R. commune is governed by the Rrs1 gene. Although this opinion focuses on apoplastic fungal pathogens of leaves, ETD responses against vascular pathogens, such as Verticillium dahliae, are also mediated by RLPs that require SOBIR1, such as Ve1.

19By contrast, pathogen effectors that are targeted to the host cytoplasm are directly or indirectly recognised by nucleotide binding site (NBS) LRR receptors (NLRs) located in the cytoplasm (Figure 2D,G). These recognition events usually trigger a fast HR or programmed cell death that leads to resistance against biotrophic pathogens, such as Blumeria graminis on barley, Bremia lactucae on lettuce (Lactuca serriola), and Puccinia striiformis on wheat, and some hemibiotrophic pathogens, such as P. infestans on potato and Magnaporthe grisea on rice (Oryza sativa) (ETI), or susceptibility to some host-specific necrotrophic pathogens (ETS) (Table 2). However, it should be noted that it is still not clear whether the HR is a cause or consequence of resistance, because induction of a HR and resistance can sometimes be separated.

20There are now unprecedented opportunities to exploit new genomic information about crop hosts and pathogens and new knowledge about the operation of host defence against pathogen attack in plant breeding to produce crops with more durable resistance against damaging pathogens. For example, if resistance against all apoplastic pathogens is mediated by genes encoding RLPs, for crops attacked by apoplastic pathogens, it should be possible to screen their genomes, especially those regions identified as containing loci for resistance against these pathogens, specifically for R genes that encode RLPs. The genes identified can then be considered as candidate ETD resistance genes. Such methods have already been used to identify ETI resistance genes encoding cytoplasmic NLRs. To identify the most useful RLPs as candidate R gene targets for breeding, they could be functionally analysed for responsiveness to corresponding effectors of apoplastic fungal pathogens that are essential for virulence. Breeders could subsequently design molecular markers to combine several functional R genes encoding RLPs in elite crop cultivars to increase the durability of the cultivar resistance against apoplastic pathogens.

21To preserve valuable sources of resistance and avoid catastrophic epidemics associated with changes in virulence spectra within pathogen populations that render ineffective single R genes widely deployed over large areas, there is also a need for schemes to guide the deployment of different R genes in space and time by farmers selecting which cultivars to grow in their fields. This requires knowledge of the R genes deployed in different commercial cultivars and a web-based scheme to guide farmers by grouping cultivars according to the R genes they contain. Such schemes are currently being operated in France and Australia for management of L. maculans. If effective R genes operating against apoplastic pathogens and the virulence spectra of their attacking pathogens can be identified more rapidly and deployed more carefully, they will make a substantial contribution to sustainable crop protection and improved food security.