Many biotrophic fungal and oomycete pathogens share a common infection process involving the formation of haustoria, which penetrate host cell walls and form a close association with plant membranes. Recent studies have identified a class of pathogenicity effector proteins from these pathogens that is transferred into host cells from haustoria during infection. This insight stemmed from the identification of avirulence (Avr) proteins from these pathogens that are recognized by intracellular host resistance (R) proteins. Oomycete effectors contain a conserved translocation motif that directs their uptake into host cells independently of the pathogen, and is shared with the human malaria pathogen. Genome sequence information indicates that oomycetes may express several hundred such host-translocated effectors. Elucidating the transport mechanism of fungal and oomycete effectors and their roles in disease offers new opportunities to understand how these pathogens are able to manipulate host cells to establish a parasitic relationship and to develop new disease-control measures.
Plant pathogens must negotiate the complex multilayered defence systems that their hosts have evolved to prevent infection. These include preformed barriers such as the waxy cuticle and antimicrobial compounds, as well as defences that are induced by pathogen recognition (Jones & Takemoto, 2004; Chisholm et al., 2006). These inducible responses are triggered by the plant innate immunity system, which is now thought to be comprised of two levels (Jones & Dangl, 2006). The first involves recognition of conserved structural components of pathogens such as chitin or flagellin, which are collectively known as pathogen-associated molecular patterns (PAMPs). Recognition of these factors by cell-surface receptors leads to PAMP-triggered immunity (PTI), which is effective at preventing infection by nonadapted pathogens and probably underlies nonhost-resistance mechanisms. Bacterial pathogens of plants are known to overcome these defences through the use of effector proteins, which are delivered into host cells by the bacterial Type III secretion system. Many of these effector proteins have been shown to interfere directly with the PTI signalling process or downstream responses (Chisholm et al., 2006; Jones & Dangl, 2006). However, many of these effectors are recognized by a second layer of the plant defense system, which involves intracellular receptors that are the products of the classically defined resistance (R) genes of the gene-for-gene system. These receptors recognize the products of pathogen avirulence (Avr) genes, leading to rapid activation of defence mechanisms, such as increased ion fluxes, an extracellular oxidative burst and a localized cell death termed the hypersensitive response (HR), which is thought to limit the spread of the pathogen from the infection site (Flor, 1971). This layer of defense has been termed effector-triggered immunity (ETI), and involves direct or indirect recognition of pathogen-effector proteins by plant R proteins. Recent advances in the study of biotrophic oomycetes and fungi indicate that this general picture of pathogen effector/host immunity interactions also holds true for these eukaryotic pathogens (Ellis et al., 2007; Kamoun, 2007; Birch et al., 2008; Tyler, 2009). Here we review recent work uncovering the role of host-translocated proteins of fungi and oomycetes in host–pathogen interactions.
Many biotrophic and hemi-biotrophic pathogens share a common infection process that involves the formation of haustoria within living plant cells. During infection, the pathogen penetrates the cell wall and invaginates the plasma membrane of a host cell where it forms this specialized feeding structure (Fig. 1). The haustorium appears to play an essential role in nutrient acquisition and there is evidence to suggest that haustoria are involved in the redirection of host metabolism and the suppression of host defenses (Voegele & Mendgen, 2003). Haustoria are surrounded by the haustorial wall and remain separated from the host-cell cytoplasm by a region known as the extrahaustorial matrix and the extrahaustorial membrane, which is derived from the invaginated host plasma membrane (Fig. 1). Haustoria formation induces substantial re-organization of the host-cell cytoskeleton, nuclear DNA and endomembrane system (Kobayashi et al., 1994; Heath, 1997). Thus, it is likely that significant information exchange between the host and the pathogen occurs at this site and is important for the establishment of a successful parasitic relationship. Indeed it is now clear that haustoria-forming pathogens deliver numerous effector proteins into host cells across this interface. This understanding has emerged from the successful cloning, over the last few years, of a number of Avr genes from these pathogens. These Avr genes encode small secreted proteins that are recognized by cytoplasmic R proteins inside the host cell.
Avirulence proteins from biotrophic fungi and oomycetes
Avirulence proteins from bacteria and extracellular fungi such as Cladosporium fulvum and Rhyncosporium secalis have been well known for many years (Espinoza & Alfano, 2004; Stergiopolous & DeWit, 2009). Bacterial Avr proteins are part of the suite of Type III secreted effectors that are delivered directly into host cells through a specialized secretory apparatus (the Type III secretory system) where they act to influence host transcription or target specific host proteins for degradation. By contrast, extracellular fungi, which grow only in the intercellular space and do not penetrate host cells, produce secreted Avr proteins that are recognized and perform pathogenicity functions in the apoplast of infected plants. Until recently, identification of Avr genes from haustoria-forming biotrophic fungi and oomycetes has lagged behind studies of bacterial Avr proteins, largely because of the difficulties with culturing and transforming these pathogens. However, technological advances have made these organisms much more tractable, and genome-sequencing efforts have led to rapid insights and the emergence of a new paradigm. While Avr proteins from these biotrophic pathogens are secreted through the standard endomembrane pathway, similarly to the extracellular fungal pathogens, it has now been shown that the secreted proteins are subsequently delivered into the host cell by an as yet unknown mechanism.
Among the biotrophic fungi, the basidiomycete flax rust fungus, Melampsora lini, has so far yielded the most information regarding host-translocated Avr proteins (Lawrence et al., 2007). Four Avr gene families have been identified: AvrL567, AvrM, AvrP123 and AvrP4 (Dodds et al., 2004; Catanzariti et al., 2006). All of these encode small secreted proteins that are expressed in haustoria. Indeed, during R gene-dependent resistance to rust fungi, the HR is first observed in plant cells containing emerging haustoria (Kobayashi et al., 1994; Heath, 1997). Transient expression of these proteins in flax plants of different genotypes induced R gene-dependent leaf necrosis, confirming their avirulence function. Importantly, this response was observed when the proteins were expressed without their signal peptides, indicating that recognition occurs inside the plant cytoplasm. This is consistent with the location of the corresponding R proteins, and in fact the AvrL567 and AvrM proteins have been shown to interact directly with the cytoplasmic L6 and M resistance proteins, respectively (Dodds et al., 2006; P. N. Dodds, unpublished).
Two Avr genes, Avr-a10 and Avr-k1, recognized by Mla10 and Mlk resistance genes respectively, have been isolated from the ascomycete pathogen Blumeria graminis pv. hordei (Bgh), which causes powdery mildew infections on barley (Ridout et al., 2006). In contrast to the rust proteins, these Avr proteins are not predicted to contain a secretion signal peptide, but nevertheless are recognized by intracellular barley resistance proteins when transiently expressed in planta. This suggests that they may be secreted from the fungus by a nonendomembrane pathway. These genes are closely related to each other and are part of a large diversified family with at least 30 homologs present in Bgh.
Other fungi that do not produce haustoria nevertheless show evidence for the intracellular delivery of at least some Avr or effector proteins. For instance, the AVR-Pita protein of the rice blast fungus Magnaporthe oryzae encodes a secreted protein that is recognized inside plant cells by direct interaction with the Pi-ta resistance protein (Jia et al., 2000). Likewise, secreted AvrPiz-t is recognized by an intracellular R protein (Li et al., 2009). M. oryzae invasive hyphae propagate within plant cells and, like haustoria, are surrounded by an extra-invasive hyphal membrane contiguous with the host plasma membrane, so it appears likely that this pathogen also delivers secreted proteins across the plant membrane (Kankanala et al., 2007). Recently, Mosquera et al. (2009) identified a set of fungal secreted proteins that are upregulated during biotrophic growth. Several of these effector candidates co-localize with AVR-Pita to a distinct region of the extra-invasive hyphal space, known as the biotrophic interfacial complex. It is suggested that this structure serves as an assembly area for secreted effectors before they are transported into host cells.
Likewise, the maize smut fungus, Ustilago maydis, also makes close contact with living host cells via intracellular hyphae that invaginate the host cell membrane. Over 400 secreted proteins were predicted in the U. maydis genome, approx. 20% of which occur in 12 clusters of 3–26 genes containing one to five different gene families (Kämper et al., 2006). Complete deletions of 5 of the 12 clusters had an observable effect on pathogenicity, indicating an important role in infection. Moreover, several U. maydis genes for secreted proteins have homology to proteins with intracellular regulatory functions, such as RING fingers and F-box proteins that are involved in protein ubiquitination, RNA-binding proteins, Rho/Rac/Cdc42-like GTPases and EF-hand Ca2+-binding proteins (Mueller et al., 2008). The predicted functions of these secreted proteins in intracellular processes suggest that they perform these functions after translocation into plant cells. Functional studies are still needed to confirm both the biological function and the possible translocation inside plant cells of these putative effectors. Several secreted-in-xylem (SIX) proteins have been identified from Fusarium oxysporum f. sp. lycopersici (Fol) through their abundance in xylem sap of infected tomato plants (Rep et al., 2004; Houterman et al., 2008). Although these are mostly thought to act extracellularly, Avr2 is recognized by the I-2 Fol resistance protein intracellularly (Houterman et al., 2009), so at least one Fol Avr protein enters host cells.
Four avirulence genes were originally isolated from oomycete pathogens essentially by map-based cloning: Avr1b-1 from Phytophthora sojae, Avr3a from Phytophthorainfestans, and ATR13 and ATR1NdWsB from Hyaloperonospora arabidopsidis, the downy mildew of Arabidopsis (Allen et al., 2004; Shan et al., 2004; Armstrong et al., 2005; Rehmany et al., 2005). Like the fungal Avr genes, these all encode small secreted proteins, and transient expression in the host cytoplasm leads to R-gene-dependent cell death. Again this indicates that recognition occurs inside plant cells, implying delivery of these pathogen proteins across the plant membrane during infection. Subsequently, genome sequence interrogation based on conserved features of these Avr proteins, particularly the presence of an RxLR motif (discussed later), has led to the identification of further Avr genes: Avr1a, Avr3a and Avr4/6 from P. sojae (Dou et al., 2008a; Qutob et al., 2009), and Avr4 and AvrBlb1 from P. infestans (van Poppel et al., 2008; Vleeshouwers et al., 2008).
Delivery of effectors into host cells
The critical question now posed is how do these fungal and oomycete proteins cross the plant plasma membrane to enter the host cytoplasm? Here, work on the oomycete systems has provided the most clues so far (Birch et al., 2008; Tyler, 2009). Each of the oomycete Avr proteins contains a short sequence motif just downstream of the signal peptide that consists of the consensus amino acid sequence RxLR (where x is any amino acid), and in the case of the Phytophthora proteins, this is followed shortly afterwards (5–21 amino acids) by a dEER motif. Additional amino acids flanking the core conserved residues of the motifs are also required for host cell uptake, suggesting that the local protein structure influences the presentation of the uptake motif (Bhattacharjee et al., 2006; Dou et al., 2008b). Whisson et al. (2007) showed that the RxLR motif was not required for secretion of Avr3a from the haustorium of P. infestans, but was necessary for transfer from haustoria into infected potato cells. Furthermore, an Avr3a–β-glucuronidase fusion protein could be delivered into host cells from transformed P. infestans, which was dependent on the RxLR–dEER motif. This suggests that oomycete Avr proteins enter the host cell in a two-step process, involving signal peptide-mediated secretion followed by host cell translocation mediated by the RxLR motif. One important question is whether the transfer process depends on a specialized pathogen secretory structure, analogous to bacterial Type III secretion, or whether transfer relies on the host plant cell transport machinery. Strong support for the second hypothesis comes from a recent study of P. sojae Avr1b by Dou et al. (2008b), who demonstrated that the RXLR–dEER motif mediates delivery of Avr1b into plant cells in the absence of the pathogen. This motif was required for recognition of Avr1b transiently expressed in soybean cells as a secreted protein, but not when it was expressed intracellularly. In addition, fusion to the Avr1b RXLR–dEER domain allowed purified green fluorescent protein (GFP) to enter soybean root cells from solution. These results suggest that uptake of RxLR effectors into host cells is not dependent on a pathogen-derived transport mechanism.
The RxLR motif is related to a host-targeting signal found in proteins secreted by the human malaria parasite (Plasmodium falciparum), which invades red blood cells and is enveloped by a parasitophorous vacuole membrane derived from the invagination of the host cell membrane. During infection, proteins secreted by P. falciparum enter the parasitophorous vacuole, and those containing an N-terminal PEXEL (plasmodium export element) motif with the conserved sequence RxLxE/D/Q, are further transported across the parasitophorous vacuolar membrane into the erythrocyte cell (Hiller et al., 2004; Marti et al., 2004). Bhattacharjee et al. (2006) showed that the RxLR oomycete motif can also function as a host-targeting signal in P. falciparum, providing strong evidence that these eukaryotic microbes share a conserved translocation mechanism for effector proteins. Dou et al. (2008b) recently provided evidence for the reciprocal situation, that the plasmodium PEXEL motif can direct translocation of Avr1b into plant cells. Again, as for the RxLR motif, this transfer was independent of the presence of the pathogen. The PEXEL motif can also substitute for the RxLR motif in Avr3a in transformed P. infestans, providing further evidence for the interchangeability of these motifs (Grouffaud et al., 2008). This raises some interesting questions about their respective transport systems. Pathogen-independent transfer suggests that RxLR proteins utilize an intrinsic plant protein-uptake process, while the strict conservation of this motif suggests this is recognition-driven, possibly involving a host receptor. However, red blood cells are terminally differentiated structures that completely lack a nucleus and other subcellular structures. Plasmodium infection includes the formation of novel subcellular structures within the erythrocytes, such as Maurer's clefts, which are intrinsic to protein translocation and are constructed from pathogen-encoded proteins. In fact, mutations in several Plasmodium genes have been shown to disrupt delivery of effectors to the erythrocyte cell surface (Maier et al., 2008), indicating that they form part of the transport mechanism. Recently it was shown that the PEXEL motif of several malarial effectors is subject to proteolytic cleavage immediately after the central leucine residue, and an N-acetyl group is added to the mature protein (Chang et al., 2008; Boddey et al., 2009). PEXEL cleavage and modification occurs within the parasite endomembrane system, implicating pathogen-encoded processing enzymes. Recognition of the cleaved and acetylated N-terminus may be the critical step for directing these proteins into host cells. Thus, trafficking of malarial effectors is largely a pathogen-driven process and it is unclear whether the mechanism is conserved with oomycete effector delivery. It will be important to determine whether similar cleavage of oomycete effectors, and of synthetic PEXEL-containing reporter proteins, also occurs during translocation into plant cells.
Effector proteins from rust fungi do not contain a clearly conserved RxLR motif, but are also directed into the host cell during infection. Kemen et al. (2005) detected (by immunocytochemistry) the bean rust secreted protein, RTP1, in the nucleus of infected host cells, confirming that these pathogens also deliver proteins to the host. Using a similar method, we have recently detected AvrM in the cytoplasm of flax cells infected with flax rust (P. Gan, unpublished data). Other flax rust Avr proteins are secreted from haustoria and are recognized in plant cells, suggesting a two-step process, namely secretion followed by translocation, as for oomycete effectors. Again, uptake appears to be independent of the pathogen because the secreted forms of these proteins still trigger defense responses when transiently expressed in resistant host cells (Catanzariti et al., 2006). In the case of AvrM, these resistance responses are inhibited by the addition of a C-terminal HDEL endoplasmic reticulum retention signal, which would hold the AvrM protein within the plant endomembrane system. The closely related, but inactive, HDDL tag allows full AvrM avirulence activity. This suggests that the secreted form of the AvrM protein re-enters the host cytoplasm in the absence of the rust fungus. Likewise, transient expression of the secreted AvrM and AvrL567 proteins fused to GFP leads to intracellular accumulation of GFP, and this is dependent on sequences in the N-terminal region of these proteins (M. Rafiqi, unpublished data). Thus, rust effectors also appear to utilize a plant-derived transport mechanism, but sequence comparisons have not identified clearly conserved motifs common to these proteins. The N-terminal region of AvrM is rich in the positively charged amino acids arginine and lysine, which are common to several cell-penetrating peptides (CPPs) that facilitate protein transport across membranes. A number of proteins from distant phylogenetic groups, such as the Tat protein of the human immunodeficiency virus type-1 (Frankel & Pabo, 1988) and the antennapedia protein of Drosophila melanogaster (Derossi et al., 1996), have been found to possess cell-internalization properties. Mutation analyses identified short CPP sequences within these proteins that mediate translocation into eukaryotic cells. Several CPPs defined in mammalian systems have also been shown to function in plants (Chugh & Eudes, 2008). Although these show little sequence homology with each other, all are rich in positively charged arginine or lysine residues.
The apparent divergence between the oomycete and rust transport signals, especially given the strong conservation within the oomycetes and the lack of conservation even between rust effectors, may suggest that different transport mechanisms are involved. However, it is possible that convergent evolution has led to targeting of similar host transport pathways by both classes of pathogens without leaving a clearly recognizable sequence relationship. We still know little of the uptake pathway of these proteins. Ultrastructure studies have detected tubular extensions of the extrahaustorial membrane with associated budding vesicles that reach into the host cell cytoplasm and can form close contact with the host endoplasmic reticulum and dictyosomes (Mims et al., 2002). This suggests that endocytosis may be a key process in the uptake of effector proteins.
In contrast to flax rust and oomycete Avr genes, the barley powdery mildew genes do not encode proteins with N-terminal secretion signals and it has been proposed that an alternative secretion pathway may exist for these proteins (Ridout et al., 2006). Thus, it remains possible that these ascomycete pathogens do encode a specific host translocation mechanism analogous to the bacterial Type III secretion. Recent work in the basidiomycete Cryptococcus neoformans, a human pathogen, has highlighted an alternative effector secretion pathway in fungi (Rodrigues et al., 2008; Panepinto et al., 2009). Some C. neoformans virulence factors, both with and without N-terminal secretion signals, are found within extracellular vesicles, known as exosomes. In animals these membrane vesicles can be released into the extracellular environment by exocytic fusion of multivesicular bodies with the cell surface and can allow direct cell-to-cell transfer of molecules (Fevrier & Raposo, 2004).
Role of host-translocated effectors in fungal and oomycete infection
Specific delivery of oomycete and fungal effectors into plant cells during infection implies that they perform functions in this environment that are important for pathogen virulence and counterbalance the risk of triggering resistance through recognition by host R proteins. Much is now known about the specific roles of bacterial Type III effectors, some of which act as transcription factors while others target host proteins for degradation or phosphorylation (Espinoza & Alfano, 2004). However, as yet little is known about the roles of putative effector proteins from fungi and oomycetes, and elucidating the biological functions of these proteins will be crucial in understanding the infection process.
Most of the Avr and effector proteins identified to date have no homology to any proteins of known function, so there are few clues to their roles in infection. However, there is evidence for positive roles of some of these proteins in infection, particularly in suppressing PTI. For instance, expression of the P. infestans protein Avr3a in Nicotiana benthamiana inhibits the cell-death response induced by the INF1 elicitor, also from P. infestans (Bos et al., 2006). No effect was observed on cell death induced by the PiNPP1 and CRN2 elicitors, suggesting a specific role in suppressing the INF1 response. In addition, ATR1 and ATR13 from H. arabidopsidis suppress basal resistance in Arabidopsis (Sohn et al., 2007). This was shown by expressing these proteins in the bacterial pathogen Pseudomonas syringae pv. tomato fused to a Type III secretion signal to deliver the proteins into host cells. Bacterial strains expressing these proteins triggered defense responses on plants carrying the corresponding R genes, but showed enhanced virulence on susceptible lines. ATR13 also suppressed callose deposition induced by bacterial PAMPs, again suggesting a role in PTI inhibition. Dou et al. (2008a) showed that Avr1b of P. sojae could suppress cell death induced by the mouse protein BAX1, which is a positive regulator of apoptosis. Avr1b suppressed BAX1-induced cell death in soybean and N. benthamiana, as well as in yeast, suggesting that it may target a component of programmed cell-death pathways that is common to diverse organisms. Interestingly, the majority of oomycete effectors seem to belong to a large superfamily that shares one or more of a series of conserved sequence features: the W, Y and L motifs (Jiang et al., 2008). Avr1b contains adjacent W and Y motifs that are required for its activity in suppressing BAX1-induced necrosis, and three other putative oomycete effectors with W–Y motifs also possess this activity (Dou et al., 2008a). Thus, cell-death suppression may be a common function of many effectors and it is probable that there is considerable redundancy in the effector repertoire for this function. Interestingly, the ipiO/Avr-blb1 family contains an RGD cell-adhesion motif overlapping the RxLR motif, and binds an 80-kDa membrane-bound receptor via the RGD motif (Senchou et al., 2004; Vleeshouwers et al., 2008). IpiO can disrupt connections between the plasma membrane and the cell wall that are mediated by lectin receptor kinases, which may be pathogenicity targets for ipiO. This raises interesting questions of whether there are pathogenicity targets of this protein both outside and inside the cell, or whether the 80-kDa binding protein is involved in the mechanism of effector uptake, or both. It is possible that specific binding to a receptor may help to bring the effector into a position where the RxLR dEER motif enables uptake.
The B. graminis Avra10 and Avrk1 proteins also have an effector function in promoting pathogen infection (Ridout et al., 2006). Transient expression of these proteins in barley epidermal cells increases the frequency with which B. graminis is able to successfully establish haustoria. It is not clear whether this function involves suppression of defence responses, or some other positive effect on haustoria growth or nutrition. No clear evidence for effector roles of rust Avr proteins is available. However, all of these proteins are highly polymorphic and have been subject to diversifying selection as a result of the need to escape host R protein recognition (Dodds et al., 2004; Catanzariti et al., 2006). The accumulation of amino acid sequence variation to evade recognition, as opposed to gene loss or inactivation, suggests that there is positive selection to maintain a pathogenicity-related function of these proteins. The bean rust protein RTP1 is localized to host nuclei during infection, suggesting that it may function to influence host gene expression (Kemen et al., 2005). Among fungal translocated effectors, only two have predicted functions based on sequence homology. AvrP123 from flax rust is related to the Kazal family of serine protease inhibitors (Catanzariti et al., 2006), and AVR-Pita of M. oryzae is a metalloprotease (Orbach et al., 2000). However, the presumed host targets of these proteins are not known. AvrPita is not required for virulence on rice, but this may be a result of functional redundancy because it belongs to a small gene family in M. oryzae (Khang et al., 2008).
The field of biotrophic pathogen effector biology was only spawned in the last 5 years, but is moving very rapidly. The application of genome sequencing has led to the identification of large repertoires of translocated effectors from oomycete pathogens. Similar genome-sequencing efforts are also generating important data on fungal secretomes, although definition of the transport signals used by fungal effectors remains a high priority to allow more generalized effector prediction. The other exciting challenges that remain are to understand the nature and components of the transport process and how these host-translocated effectors function in promoting infection.
Work in the authors’ laboratories is supported by Australian Research Council grant DP0771374 and the Grains Research and Development Corporation.