Apoplastic immunity and its suppression by filamentous plant pathogens


Author for correspondence:

Gunther Doehlemann

Tel: +49 6421 178602

Email: doehlemann@mpi-marburg.mpg.de



Microbial plant pathogens have evolved a variety of strategies to enter plant hosts and cause disease. In particular, biotrophic pathogens, which parasitize living plant tissue, establish sophisticated interactions in which they modulate the plant's metabolism to their own good. The prime decision, whether or not a pathogen can accommodate itself in its host tissue, is made during the initial phase of infection. At this stage, the plant immune system recognizes conserved molecular patterns of the invading microbe, which initiate a set of basal immune responses. Induced plant defense proteins, toxic compounds and antimicrobial proteins encounter a broad arsenal of pathogen-derived virulence factors that aim to disarm host immunity. Crucial regulatory processes and protein–protein interactions take place in the apoplast, that is, intercellular spaces, plant cell walls and defined host–pathogen interfaces which are formed between the plant cytoplasm and the specialized infection structures of many biotrophic pathogens. This article aims to provide an insight into the most important principles and components of apoplastic plant immunity and its modulation by filamentous microbial pathogens.

I. Live and let die – the basics of interaction

In nature, plants are under constant attack by various microbes that aim to establish different kinds of interactions from friendly to fatal. Microbial plant pathogens are classified in accordance to their life style. Necrotrophs invade and kill the plant tissue, feeding on the dead tissue debris, whereas hemibiotrophs first colonize the plant tissue before inducing cell death (Horbach et al., 2011). Biotrophs, on the other hand, depend on the living plant metabolism as their nutritional source, and therefore interact intimately with the host cells to modify metabolic processes to serve their needs (Glazebrook, 2005). This mode of interaction obviously involves a prolonged and effective suppression of the host immune system. Filamentous pathogens, on which this article focuses, show a broad spectrum of infection structures to accommodate themselves inside the host plant (Fig. 1). Extracellular pathogens, such as Cladosporium fulvum, colonize the existing apoplast in between plant cells, whereas smut fungi, such as Ustilago maydis, invaginate the plant plasma membrane to form a biotrophic interface that serves as an interaction zone between pathogen and host (Joosten & de Wit, 1999; Brefort et al., 2009). Other pathogens, such as the rust fungi, form complex interaction structures, the haustoria. Here, a structure called the neckband physically isolates the apoplastic space from a distinguished interaction zone surrounding the haustorium, the extrahaustorial matrix (Catanzariti et al., 2007). Thus, biotrophic pathogens have adapted different interaction structures to establish compatibility and nutrient transfer via the apoplast.

Figure 1.

Infection structures of different fungal plant pathogens. (a) Schematic representation of the infection structures of different biotrophic fungi and the resulting apoplastic interaction zones (Blumeria graminis, Colletotrichum graminicola, Puccinia graminis, Ustilago maydis, Cladosporium fulvum). Ap, appressorium; C, conidium; CW, cell wall; EHM, extrahaustorial membrane; H, haustorium; HMC, haustorial mother cell; IB, interfacial bodies; N, nucleus; NB, neckband; NT, necrotic tissue; PM, plasma membrane; S, sporidia; SH, secondary hypha; St, stoma; US, uredospore; V, vesicle. (b) Confocal image shows the penetration of maize epidermal cells by the biotrophic smut fungus Ustilago maydis. The plant plasma membrane (PIN1-YFP, green) surrounds the invading hypha (RFP, red). Arrowhead marks the point of penetration. Bar, 5 μm. Image was modified from Doehlemann et al. (2009). (c) Scanning electron microscopy image of a maize mesophyll cell penetrated by U. maydis. Cellular matrix surrounds the penetrating hypha (asterisk). CW, plant cell wall; V, vacuole. Bar, 2.5 μm. Image kindly provided by Dr Karl-Heinz Rexer, Philipps-Universität Marburg, Germany.

To defend themselves against various kinds of pathogens, plants apply a multilayered immune system consisting of both preformed and inducible mechanisms. Although being fundamentally different, both plant and mammalian immune systems share their basal function, which is to discriminate ‘self’ from ‘nonself’. The ability to differentiate its own molecules from those of other organisms represents a first, essential line of defense of any immune system. Contact with conserved structures or molecules presented by microbial cells triggers the first level of inducible defenses. Such eliciting molecules are termed pathogen- or microbe-associated molecular patterns (PAMPs or MAMPs). PAMPs, such as the bacterial flagellin or the fungal cell wall component chitin, are recognized by specific plant receptor proteins. PAMP perception initiates a set of basal defenses, amongst which the first detectable response is an accumulation of reactive oxygen species (ROS) within minutes after signal perception (see also Section 'To burst or not to burst – reactive oxygen in plant defense'). In addition to PAMPs, the plant immune system also recognizes endogenous molecular patterns (damage-associated molecular patterns, DAMPs), for example, cell wall fragments that are released on pathogen attack or wounding.

The perception of a conserved molecule trigger activates distinct pathways of plant immunity. Such a defense response is called PAMP-triggered Immunity (PTI). During the plant immune response, the phytohormones salicylic acid (SA) and jasmonic acid (JA) play a key role in signal transduction (Dong, 1998). Generally speaking, SA and JA trigger different, antagonistic immune response pathways, depending on the nature of the pathogen attack (Glazebrook, 2005). SA-dependent responses are associated with a massive ROS accumulation and eventually lead to programmed cell death (PCD) (Morel & Dangl, 1997; Shirasu et al., 1997), which is an effective defense against biotrophic pathogens. Necrotic pathogens and herbivores, by contrast, are efficiently controlled by JA-dependent signaling, which results in the secretion of phytoalexins, such as flavonoids and terpenes, which act directly as toxins on the intruder (Gundlach et al., 1992; Memelink et al., 2001).

To circumvent the activation of defense responses, successful pathogens secrete sets of so-called effector proteins that interfere with a range of physiological processes of the host plant (de Wit et al., 2009; Stergiopoulos & de Wit, 2009). The suppression of PTI by effectors is called effector-triggered susceptibility (ETS). As a common rule, all successful (biotrophic) pathogens must overcome PTI to achieve host colonization (Nomura et al., 2005), whereas evolutionary pressure on the host plants promotes the development of strategies to prevent invasion by microbial pathogens. Consequently, plants have developed mechanisms to recognize intruding effectors and react with effector-triggered immunity (ETI), a stronger form of defense response. This has led to an evolutionary arms race between plants and pathogens, advancing new infection and defense strategies to achieve a small advantage over the opposing party (Birch et al., 2006; Jones & Dangl, 2006; Pieterse et al., 2009).

As discussed above, PTI is a rather nonspecific response that is elicited by basic structural components of the pathogen or host cell. According to the gene-for-gene hypothesis (Flor, 1971), ETI, on the other hand, is based on the highly specific interaction of a pathogen effector (in this context, also termed an avirulence (Avr) protein) and a resistance (R) protein. It is broadly accepted that only ETI can trigger strong defense responses, such as the hypersensitive response (HR), and systemic effects, such as systemic acquired resistance (SAR), whereas PTI triggers weaker, local defense, such as ROS accumulation and papilla formation (Jones & Dangl, 2006; Tsuda et al., 2009). However, this strict distinction has recently been challenged, as PAMPs have also been shown to elicit HR, and some effectors trigger rather weak ETI responses (Thomma et al., 2011). It can be concluded that not only does the nature of a triggering molecule determine the defense responses, but also the period of exposure and the perceived dosage are crucial factors. For many kinds of interaction, the apoplastic space represents an important contact area between host and intruder, and therefore equals a battlefield into which either party sends their most developed troops to conquer or defend the plant tissue. Key processes happening in this compartment during microbial infections therefore determine the fate of the interaction.

II. A kettle of color – components of apoplastic plant defense

Before entering a plant cell, pathogenic microbes need to pass the apoplast, where various defense compounds of the plant immune system form an efficient barrier. On pathogen recognition, attacked cell walls are remodeled and the apoplast is poisoned by means of antimicrobial low-molecular-weight compounds (Monaghan & Zipfel, 2012).

This first line of defense relies on apoplastic sensors to recognize foreign nonself components in order to initiate an immune response. The perception of extracellular triggers, such as PAMPs or DAMPs, during PTI is mediated by membrane-bound pattern recognition receptors (PRRs) (Felix et al., 1993; Boller, 1995; Gomez-Gomez & Boller, 2000) . PRRs are highly conserved transmembrane proteins that consist of an extracellular leucine-rich repeat (LRR) domain and an intracellular kinase domain that triggers a defense-specific mitogen-activated protein kinase (MAPK)-dependent signal cascade. A well-documented example for PTI is the perception of bacterial flagellin by the receptor FLS2, which induces defense gene activation by direct binding to the flagellin-derived peptide flg22 (Gomez-Gomez et al., 1999; Gomez-Gomez & Boller, 2000; Chinchilla et al., 2006; Takai et al., 2008), but was also found to be involved in pathogen recognition by binding the Xanthomonas protein Ax21 (Danna et al., 2011). The bacterial elongation factor EF-Tu elicits PTI in Brassicaceae through recognition of the peptide elf18 (Kunze et al., 2004). To fulfill their signaling function, the PRRs involved in the perception of these extracellular cues need to interact with arginine–aspartate (RD) kinases, such as BAK1, in a ligand-dependent fashion (Chinchilla et al., 2007; Roux et al., 2011). BAK1 was first identified as a requirement for brassinosteroid signaling via the receptor BRI1 (Li et al., 2002) and is a common component in many receptor-like protein signal transduction pathways. Similarly, FLS2 from Arabidopsis thaliana also depends on interaction with BAK1 on pattern recognition (Schulze et al., 2010; Roux et al., 2011). By contrast, the transmembrane protein CEBiP from rice binds extracellular chitin fragments via two LysM domains and triggers defense signaling without binding to BAK1 (Kaku et al., 2006). For AtCERK1, a CEBiP homolog of A. thaliana, it has been shown recently that dimerization of the receptor is induced by binding of chitin octamers, which is critical for the activation of immune responses (Liu et al., 2012). In tomato, the ethylene-inducing xylanase EIX elicits a defense response triggered by the receptor-like protein EIX2 (Boller, 1995; Ron & Avni, 2004). The essential roles of PRRs for plant immunity in initiating intracellular signal transduction has been excellently summarized in recent articles (Zipfel, 2009; Monaghan & Zipfel, 2012; Spoel & Dong, 2012).

Important components of the induced plant defense response against microbial pathogens are proteins that are produced on pathogen perception to restrict pathogen growth. These pathogenesis-related proteins (PRs) were originally defined as ‘proteins encoded by the host plant but induced only in pathological or related situations’ (Antoniw et al., 1980), and are historically classified into 17 families on the basis of their characteristics and biological activity (van Loon et al., 2006). Although recent genome and transcriptome data challenge the concept of PR families, it still serves as a helpful basis to categorize plant defense responses. Classical PRs, such as PR-1, PR-2 and PR-5, were isolated from A. thaliana apoplastic washing fluid after artificial induction of host defense by treatment with the SA analogue 2,6-dichloroisonicotinic acid (INA) (Uknes et al., 1992). The apoplastic PRs, which actually comprise all PR families except the ribonuclease-like PR-10, can be found in cell wall appositions in response to pathogen attack. In addition, xylem and guttation fluids contain PRs, suggesting their secretion and transport through the veins (van Loon et al., 2006). PRs have various antimicrobial activities. In the case of the β-1,3-glucanase PR-2 or the chitinases PR-3, PR-4 and PR-8, these include obvious enzymatic attacks on major components of fungal cell walls (van Loon et al., 2006). Most prominent PRs are members of the PR-1 family, which serve as prime markers for SA-induced defense signaling in many plant systems (Vlot et al., 2009; Nishimura and Dangl, 2010). They have been reported to be biologically active against the oomycete Phytophthora infestans, fungal pathogens such as Blumeria graminis and pathogenic bacteria (Niderman et al., 1995; Schultheiss et al., 2003; Sarowar et al., 2005). However, conclusive evidence for the function of PR-1 in pathogen interactions is still lacking and the mode of action of PR-1 proteins still remains unclear.

In maize, microarray studies during different steps of infection with the corn smut fungus Ustilago maydis revealed a strong upregulation of PRs, including chitinases and glucanases, on penetration of the epidermis (Doehlemann et al., 2008a). In addition, a massive induction of a terpene synthase (TPS6/11), involved in isoprenoid production (Köllner et al., 2008; Schmelz et al., 2011), was observed, indicating that phytoalexins also play a role in the response to the biotroph U. maydis (van der Linde et al., 2011). Phytoalexins are defined as secondary metabolites of low molecular mass that are induced by stress and exhibit antimicrobial activity (Hammerschmidt, 1999). It has been shown that the secretion of kauralexins in maize inhibits the growth of the pathogenic fungi Colletotrichum graminicola and Rhizopus microsporus (Schmelz et al., 2011), substantiating the importance of the role of phytoalexins, not only in defense against necrotrophs, but biotrophs as well. In a compatible biotrophic system, such as the U. maydis–maize interaction, the expression of PRs, as well as components of phytoalexin synthesis, is attenuated once biotrophy has been established successfully by the intracellular fungal cells (Doehlemann et al., 2008b). The U. maydis–maize example shows that induction of nonspecific, PAMP-induced defenses during the initial phase of infection cannot be avoided, even by a fully adapted pathogen. This defense, however, is efficiently suppressed by a multiplicity of factors from both interaction partners, which will be addressed in more detail in the following sections.

III. To burst or not to burst – reactive oxygen in plant defense

The accumulation of ROS is a core component of the early plant immune response. ROS can act as toxins on intruding pathogens by damaging peptides as well as nucleic acids via their oxidative capabilities (Mehdy, 1994; Kohen & Nyska, 2002; Apel & Hirt, 2004). In addition, ROS have been shown to mediate the structural reinforcement of the plant cell wall in papilla formation (Ros Barceló, 1997; Hückelhoven, 2007; Bhuiyan et al., 2009). Hydrogen peroxide (H2O2) acts as an oxidant in peroxidase-mediated lignin synthesis (Freudenberg, 1959; Davin & Lewis, 2005). Of particular importance for the induction of defense responses is the interplay of ROS with hormone signaling pathways, particularly with the SA pathway. A synergistic effect of ROS and SA is believed to activate a signal amplification loop that drives the HR and also the establishment of systemic defenses (Draper, 1997). This involves H2O2-induced SA accumulation, as well as a downregulation of ROS-scavenging systems by SA (Shirasu et al., 1997; Klessig et al., 2000). The prompt accumulation of ROS on pathogen attack is termed the oxidative burst. The oxidative burst consists of two phases, with a rapid increase in ROS production occurring in both. The first phase of the oxidative burst is initiated within minutes after initial PAMP perception and usually lasts for c. 60–180 min (Lamb & Dixon, 1997). This early burst event is initiated by secreted enzymes that are available in the apoplast in an inactive form. On initiation of ROS production, transcription of additional ROS-producing enzymes is induced (Mehdy, 1994; Daudi et al., 2012). The second phase usually begins 3–4 h after pathogen attack, if the intruder fails to establish a compatible interaction (Lamb & Dixon, 1997). Characteristic of this phase is a longer duration of ROS production with a continuous increase in ROS levels (Baker & Orlandi, 1995). This process can eventually lead to the activation of an HR and PCD in the infected tissue (Torres et al., 2006).

The rapid accumulation of ROS during the oxidative burst is mediated by the activity of two classes of enzymes: NADPH oxidases and class III heme peroxidases (Bolwell et al., 1995; O'Brien et al., 2012). NADPH oxidases produce the unstable superoxide (inline image) by single electron reduction of molecular oxygen (O2) (Sawyer et al., 1978). In the presence of water, superoxide dismutase catalyzes the reaction from superoxide to the more stable H2O2, hydroxide (OH) and O2 (Sutherland, 1991; Taylor et al., 1993; Sagi & Fluhr, 2001). Class III peroxidases, on the other hand, form H2O2 directly without the activity of the superoxide dismutase. Featuring two catalytic cycles, they can either produce H2O2 via the ferric enzyme in the oxidase cycle (Pichorner et al., 1992; Bolwell et al., 1999; Berglund et al., 2002) or act as H2O2 scavengers in the peroxidase cycle via dismutation of H2O2 to water and molecular oxygen (Passardi et al., 2005; Zamocky et al., 2012).

The role of peroxidases in plant defense has been well documented. In cotton, apoplastic accumulation of ROS by peroxidase activity was shown after inoculation with Xanthomonas campestris (Martinez et al., 1998; Delannoy et al., 2003). In rice, a secreted peroxidase was shown to be required in defense against a Xanthomonas infection (Hilaire et al., 2001). In lettuce, peroxidase activity is increased during a nonhost interaction (Bestwick et al., 1998) and, in the green bean Phaseolus vulgaris, a peroxidase enzyme has been shown to catalyze the apoplastic oxidative burst (Blee et al., 2001). In addition, silencing of the peroxidase FBP1, as well as knockdown of the peroxidase PRX33, of A. thaliana leads to increased susceptibility towards microbial pathogens (Bindschedler et al., 2006; Daudi et al., 2012). By contrast, overexpression of the pepper peroxidase CaPO2 in A. thaliana results in increased pathogen resistance (Choi et al., 2007). These findings indicate the importance of peroxidases for apoplastic immunity.

NADPH oxidases also play a crucial role during the formation of the oxidative burst. For example, silencing of the NADPH oxidases NbrbohA and NbrbohB in Nicotiana benthamiana results in susceptibility to an otherwise avirulent strain of P. infestans (Yoshioka et al., 2003). In A. thaliana, the NADPH oxidases RBOHD and RBOHF have been shown to be involved in the oxidative burst, triggered by avirulent strains of Pseudomonas syringae and Hyaloperonospora arabidopsidis (Torres et al., 2002). However, deletion mutants of the corresponding genes did not exhibit a strongly enhanced susceptibility towards a virulent P. syringae strain (Chaouch et al., 2012).

In some cases, the concerted action of both NADPH oxidases and peroxidases has been documented to be responsible for the oxidative burst (Tang & Smith, 2001; Ranieri et al., 2003). It has been suggested that apoplastic peroxidases initiate the early oxidative burst on pathogen recognition. The resulting ROS signal induces the expression and activation of NADPH oxidases, which, in turn, amplify the ROS signal (Bindschedler et al., 2006). In A. thaliana, RBOHD and RBOHF could only be activated in the presence of H2O2 (Torres et al., 2005), substantiating their role as amplifiers rather than initial generators of the oxidative burst. In barley, silencing of HvRBOHF also led to increased susceptibility to the powdery mildew Blumeria graminis f. sp. hordei (Proels et al., 2010). However, in this case, H2O2 production, as well as papilla formation at penetration sites, and, eventually, the HR were not affected by the knockdown. In addition, the silenced plants showed a wild-type-like oxidative burst after elicitation with the PAMP flg22. These observations suggest a functional redundancy among NADPH oxidases, for example, the A. thaliana genome contains 10 rboh genes (rboha–rbohj). An alternative explanation is the existence of a different ROS source, such as peroxidases. Support for a positive feedback between peroxidases and NADPH oxidases comes from maize, where treatment with abscisic acid (ABA) induces a biphasic oxidative burst (Lin et al., 2009). The second phase of this burst is dependent on the MAPK ZmMPK5, and can be induced by the addition of H2O2 alone, which is in accordance with a role of NADPH oxidases as amplifiers of an apoplastic ROS signal that could also have originated from peroxidase activity. Although the authors of this article sympathize with this model, it should be mentioned that other theories suggest pathogen-specific burst profiles, where the synthesis of ROS is mediated by either class of enzyme, depending on the nature of the pathogen attack (Grant & Loake, 2000; Apel & Hirt, 2004; Bindschedler et al., 2006). This is also in line with the idea of peroxidase-catalyzed ROS generation being mainly induced by chitosaccharide elicitors (and SA), whereas NADPH oxidase-catalyzed ROS production results from peptide elicitors (and ABA) (Kawano, 2003).

IV. How they learned to live with the bomb – disarming proteases

The induction of protease activity is a common reaction of plants responding to biotic stresses. In particular, the induction of PCD in SA-dependent defense is functionally connected with the activation of proteases (Heath, 2000). Mitochondria-dependent, apoptosis-like cell death involves plant metacaspases, which are structurally related to the mammalian caspases, and share a catalytic histidine–cysteine (His–Cys) diad. In addition to metacaspases, the vacuolar processing enzymes, as well as papain-like cysteine proteases (PLCPs), are implicated in cell death signaling. Plants possess c. 140 putative cysteine proteases that are divided into 15 families. PLCPs represent the C1 Clan CA, which consists of 30 members in A. thaliana (van der Hoorn, 2008). In particular, these proteases seem to hold key functions in plant immunity. In N. benthamiana, the PLCP Cathepsin B is necessary for the HR and for resistance to bacterial pathogens, such as Erwinia amylovora and Pseudomonas syringae (Gilroy et al., 2007). Interestingly, the activation of Cathepsin B by pathogens is associated with its allocation to the apoplast (Gilroy et al., 2007). For the A. thaliana PLCP RD19, a role in R-protein-mediated resistance against Ralstonia solanacearum has been reported (Bernoux et al., 2008). Moreover, RD21 (Responsive to Desiccation 21), which is the most abundant PLCP in A. thaliana, contributes to immunity. Consequently, rd21 mutants show increased susceptibility to the fungal necrotroph Botrytis cinerea (Shindo et al., 2012). RD21 contains an exceptional C-terminal granulin domain that shares homology with animal growth factors which are released on wounding (Bateman & Bennett, 2009). Not only in structure, but also in terms of localization, RD21 is a complex protein: it localizes to the vacuole as well as to endoplasmic reticulum (ER) bodies, which are ribosome-associated structures that occur in stressed plant leaves (Hayashi et al., 2001). In addition, RD21 can also traffic through the Golgi into lytic vacuoles (Andeme Ondzighi et al., 2008; Gu et al., 2012). However, the N. benthamiana ortholog of RD21, C14, is targeted by an effector of the oomycete pathogen P. infestans to prevent its secretion to the apoplast (see Section 'Effect or not – apoplastic virulence factors' and Bozkurt et al., 2011). This indicates that extracellular activity of this PLCP triggers plant defense, which, again, is in line with the apoplastic localization of Cathepsin B on pathogen treatment (Gilroy et al., 2007), and further supports the notion that the extracellular space is crucial for protease-activated defense signaling.

The importance of cysteine proteases in plant immunity is also reflected by the physiological impact of their plant-derived inhibitor molecules, namely the cystatins, which are defined as inhibitors of papain C1A family cysteine proteases (Martinez et al., 2009). PCD in oat that is activated by the fungal toxin victorin is sensitive to E-64, which is a specific inhibitor of cysteine proteases (Navarre & Wolpert, 1999). In A. thaliana, the cystatin AtCYS1 is required to suppress HR in response to avirulent pathogens and wounding (Belenghi et al., 2003). Pioneering work on the crucial interplay of cysteine proteases and cystatins was performed in soybean cell cultures (Solomon et al., 1999). PLCP activation by oxidative stress was found to coincide with the induction of cell death which could be inhibited by the ectopic expression of a cystatin, but not by the expression of serine protease inhibitors. These findings suggest that PCD is regulated by the protease–cystatin interplay and therefore define cystatins as important modulators of cell death (Solomon et al., 1999).

In maize, the secreted cystatin CC9 was identified as a suppressor of host immunity to the fungal biotroph U. maydis. CC9, which is strongly induced during epidermal penetration by U. maydis, was transiently silenced in maize leaves using virus-induced gene silencing. This resulted in maize resistance to U. maydis which became evident by an epidermal cell death response and the transcriptional induction of SA marker genes, such as PR-1 (van der Linde et al., 2012a). In seedling leaves of transgenic maize plants that constitutively overexpressed CC9, the cystatin was predominantly found in the apoplastic space. Remarkably, these plants did not show PR gene induction in response to SA treatment, indicating that the presence of CC9 blocks the signaling cascade (van der Linde et al., 2012a). As apoplastic inducers of SA-induced defense gene expression, a set of five PLCPs was identified. These include the C14 homolog CP1a, an aleurain-like protease (CP2) as well as a Cathepsin B-like protease. The activity of these proteases was completely inhibited by CC9, and this inhibition abolished the ability of the proteases to trigger immune responses (van der Linde et al., 2012a,b). On the one hand, this explains the function of CC9 in suppressing defense responses in maize. In addition, it provides evidence for the crucial role of extracellular PLCP activity in maize immunity. Indications for this mechanism not being restricted to maize are given by the finding that CC9 can also inhibit SA-induced cell death in N. benthamiana: infiltration of SA to N. benthamiana leaves results in the formation of necrotic lesions, which can largely be inhibited by the co-infiltration of recombinant CC9 (Fig. 2).

Figure 2.

CC9 inhibits salicylic acid (SA)-induced lesion formation in Nicotiana benthamiana. (a) Nicotiana benthamiana leaf, 2 d after infiltration with SA (2 mM) or SA (2 mM) + CC9 (0.5 mg ml−1) protein purified from Escherichia coli. Circles mark the infiltrated areas. (b) Quantification of lesion rate per infiltrated area (2 d after infiltration) shows a significant reduction of SA-induced lesion formation by CC9. Experiments were performed in three biological replicates with six independent infiltration events each. Error bars show standard error of the mean (SEM). *, = 0.0014 (compared with sample with SA only). Figure modified from van der Linde (2012).

Such findings, however, provoke a couple of questions regarding ‘how’ apoplastic protease functions in plant immunity. At present, the mechanisms of protease activation in the plant apoplast are largely unknown. PLCPs are synthesized as inactive zymogens, in which the catalytic center containing the nucleophilic cysteine residue is covered by the N-terminal prodomain (van der Hoorn, 2008). In the absence of specific triggers, PLCPs can also be associated with cystatins, which form relatively stable complexes by acting as pseudo-substrates for the proteases, and thereby block their active site (Yamada et al., 2000, 2001; Benchabane et al., 2010). The activation of cysteine proteases is often observed in SA-associated processes, such as the pathogen-dependent HR. Early examples were described for the cowpea–cowpea rust interaction, where cysteine protease activation appeared as a regulatory element during HR (D'Silva et al., 1998). Furthermore, the already mentioned PLCPs in soybean are activated within 40 min on elicitation by H2O2 (Solomon et al., 1999). A more recent example is the tomato protease RCR3, which is transcriptionally induced on infection by incompatible C. fulvum strains (Shabab et al., 2008). Beyond an induction at the gene expression level, the activation of PLCPs seems to occur at the post-translational level. In different plant systems, latent proteases that are bound in cystatin complexes have been shown to be activated by sodium dodecyl sulfate (SDS) (Yamada et al., 1998, 2001; Nissen et al., 2009; Tajima et al., 2011; Gu et al., 2012). Similarly, the weak protease activity in the apoplast of untreated maize plants can be massively increased by SDS treatment (van der Linde et al., 2012a). However, so far, it is not clear whether this activation is caused by the release of an unknown exogenous SDS-sensitive inhibitor, or whether an SDS-induced conformational change releases the active site of the proteases (Yamada et al., 1998; Gu et al., 2012). With the cleavage of inhibitory domains (N-terminal prodomain, granulin domain), the release of proteinaceous inhibitors (cystatins) and a stimulus-dependent reallocation (as described for Cathepsin B and C14), activation of PLCPs can be post-translationally modulated at different levels. This tight temporal and spatial regulation reflects the strong impact of these enzymes in stress signaling and, finally, induction of cell death. Thus, PLCPs appear to be molecular bombs that need to be cautiously handled by the plant and must not explode at the wrong place and wrong time. The actual molecular mechanisms that trigger PLCPs, however, still remain elusive. Similarly, the close connection of cysteine proteases and SA signaling requires further investigation. Being activated during SA-triggered processes, the activity of apoplastic PLCPs is, at the same time, sufficient to trigger SA downstream signaling, putting them upstream of this signaling cascade. It is therefore important to understand how PLCP activation is integrated into the signaling of SA and its hormonal antagonist in plant defense, JA.

V. Hide and seek – how to survive the apoplast

As discussed above, apoplastic processes execute and initiate the first line of inducible immune responses on pathogen attack. Therefore, it is obvious that any pathogen needs to overcome, that is, circumvent, the activation of apoplastic defenses. First, a successful pathogen needs to defend itself from degrading enzymes or toxic molecules, such as H2O2, and, second, it needs to avoid the production of signals that would lead to an induction or even amplification of defense responses. Basically, pathogens have three options to handle this challenge, namely hiding, detoxification/sequestering and inhibition.

Invading filaments of oomycetes and fungal pathogens contain a variety of PAMPs that are recognized by cognate PRRs, which, in turn, will – amongst others – induce the expression of cell wall-degrading enzymes (CWDEs), that is, PR proteins, whose activity will release even more defense elicitors. To avoid this vicious circle, microbes attempt to modify the structure and composition of their cell wall to avoid, or at least minimize, recognition by the plant. The fungal cell wall is composed mainly of α- and β-glucans, chitin and mannans, with branched β-glucan cross-linked to chitin being the core components (Latge, 2007). As chitin and β-glucans are the major PAMPs of fungal cell walls, their perception is reduced by several modifications. Chitin can be de-acetylated to chitosan, which is found in infection structures of various plant pathogens, such as, for instance, the rust fungi Uromyces fabae and Puccinia graminis f. sp. tritici, as well as the maize anthracnose fungus C. graminicola. This modification not only reduces PAMP release, but also protects the cell wall from plant chitinases (El Gueddari et al., 2002; Werner et al., 2007).

Another strategy is the masking of β-glucans by α-glucan. This was originally observed in the human pathogen Histoplasma capsulatum, where α-(1,3)-glucan blocks recognition by the β-glucan receptor (Rappleye et al., 2007). In plants, the rice blast Magnaporthe oryzae has been found to modulate its cellular surface by α-glucan apposition during pathogenic development (Fujikawa et al., 2009). Recently, the same authors have provided direct evidence that surface α-(1,3)-glucan, which cannot be degraded by plants, facilitates ‘stealth infection’ by protecting the pathogen from detection by plant innate immunity (Fujikawa et al., 2012). In line with this observation, transgenic rice plants expressing a bacterial α-(1,3)-glucanase obtained strong resistance to fungal infection (Fujikawa et al., 2012). The requirement for fungal pathogens to modulate their surface during plant invasion becomes particularly evident in mutants with defects in cell wall synthesis. In U. maydis, deletion mutants for chitin synthase V elicit an oxidative burst and cause HR-like cell death, which stops further infection (Treitschke et al., 2010). Probing U. maydis hyphae with a specific antibody revealed a strong reduction of α-(1,3)-glucan on the surface of chitin synthase V mutants compared with wild-type cells. Similarly, chitosan, which covers infective hyphae of U. maydis wild-type, was hardly detected in this mutant (Treitschke et al., 2010).

In addition to the integral carbohydrate components of the cell wall, surface proteins may also mask microbial PAMPs. An early example is CIH1p of Colletotrichum lindemuthianum, a glycoprotein that is oxidatively cross-linked specifically to biotrophic, intracellular hyphae (Perfect et al., 1998). Appropriate glycosylation of cell wall proteins is also required in U. maydis, where the deletion of the α-glucosidase Gas1, which processes N-linked glycoproteins in the ER, leads to a virulence arrest on plant penetration (Schirawski et al., 2005). Electron microscopy revealed that gas1 mutants failed to establish a normal host–parasite interface, indicating that N-glycosylation is required for appropriate cell wall assembly of infection structures. This is in line with earlier observations made in M. oryzae and C. lindemuthianum, where infectious hyphae showed a strong reduction in lectin affinity compared with hyphae formed outside the host plants (Howard et al., 1991; O'Connell, 1991). These findings all suggest that a fine-tuned modulation of the cell surface is crucial to avoid and resist basal, PAMP-triggered plant defense in the apoplast.

In addition to this avoidance strategy, plant pathogens also actively detoxify defense-induced host molecules. An important aspect is the handling of plant-derived ROS. Despite being clothed in a stealth jacket, an invading pathogen will never manage to completely avoid the initiation of defense, that is, the accumulation of ROS. Therefore, these molecules need to be detoxified by the pathogen to block the amplification of the host's defense signaling, but also to prevent damage caused by the toxic molecules themselves. The baker's yeast Saccharomyces cerevisiae regulates H2O2 stress by means of the transcription factor Yap1, and this system is also conserved in plant pathogenic fungi. The U. maydis ortholog of Yap1 was found to be required for H2O2 tolerance. As a consequence, yap1 deletion mutants fail to induce plant tumors, and H2O2 accumulates around infectious hyphae of this mutant (Molina & Kahmann, 2007). This increase in H2O2 in the absence of Yap1 could be explained by a lack of activation of the ROS detoxification system that is controlled by Yap1. Amongst the Yap1 downstream genes are two peroxidases, whose deletion causes similar defects to Yap1 itself. This suggests that the peroxidase activity of these enzymes is required to detoxify H2O2 in planta and therefore to promote disease (Molina & Kahmann, 2007). Similar findings have also been described recently for M. oryzae, where Yap1 deletion causes a loss of virulence, which is also accompanied by H2O2 accumulation and results in necrotic lesions at sites of mutant infections (Guo et al., 2011). A fascinating example of the crucial role of ROS modulation is the beneficial symbiotic interaction of the fungal endophyte Epichloë festucae with its grass host Lolium perenne. Tanaka et al. (2006) found that disruption of the fungal NADPH oxidase gene noxA basically transformed the symbiont E. festucae into a pathogen that massively colonized its host, leading to stunted plant growth. Apparently, E. festucae controls itself by maintenance of a certain ROS level at the plant–fungal interface. This case illustrates the delicate equilibrium between biotrophic microbes and their host, and reminds us that there is only a thin line between beneficial and pathogenic interactions.

VI. Effect or not – apoplastic virulence factors

From the situation described so far, it becomes apparent that, for an invading microbe, the plant apoplast is a quite inhospitable place with numerous dangers and traps that need to be handled. To prevent the deluded impression of pathogens being the victims in a plant–microbe interaction, it needs to be emphasized how these organisms treat the plant immune system to accommodate themselves in their chosen environment. Infectious hyphae of both fungi and oomycetes secrete hundreds of proteins to obtain entry and to shape the colonized plant tissue into a more comfortable place. A major group of secreted proteins are CWDEs. These enzymes are thought to aid invasion into plant cells mainly by catalyzing the hydrolysis of plant cell wall polymers, especially cellulose and pectin (Walton, 1994). As a result of the high redundancy of CWDE encoding genes in fungal genomes, there are only a few reports providing direct evidence for a contribution of these enzymes to fungal virulence. In U. maydis, for instance, simultaneous deletion of its three pectinase genes did not impair virulence (Doehlemann et al., 2008b). While U. maydis encodes a fairly reduced set of these lytic enzymes (Mueller et al., 2008), in the genome of the biotrophic powdery mildew fungus B. graminis hemicelluloses and celluloses are largely absent (Spanu et al., 2010). Nevertheless, a recent study has suggested a role for cellulases in the virulence of the hemibiotrophic pathogen M. oryzae (van Vu et al., 2012). Knockdown of glycoside hydrolase family 6 and 7 genes resulted in a weakened infection of rice because of reduced penetration efficiency caused by increased papilla formation that blocked invasion. A similar phenotype was observed on simultaneous silencing of the M. oryzae xylanase genes (Nguyen et al., 2011). As a comment, these findings are somehow surprising because M. oryzae is the prime model of a microbial pathogen that enters the plant by means of physical force through its highly differentiated and melanized penetration structure, the appressorium (Talbot, 1995).

Although CWDEs are secreted from the invading pathogen to the plant tissue and apparently interfere with the plant structure, these proteins are commonly not considered as ‘effectors’. This term is usually reserved for proteins of initially unknown purpose that do not contain conserved functional domains. The most conservative definition that has been used to define effectors is that of SSPs (small secreted proteins). In principle, this comprises every protein that is secreted by a pathogen during host infection, which is no larger than c. 300 amino acids, does not contain conserved motifs of known enzymes and, in the best case, contains some cysteines that possibly form disulfide bridges to stabilize the protein after secretion. A less restrictive and very useful definition of effectors comes from Hogenhout et al. (2009), who define effectors as ‘proteins and small molecules that alter host-cell structure and function’. So far, all genomes of plant pathogenic fungi and oomycetes, but also those of beneficial symbionts, such as mycorrhiza (like) fungi, contain hundreds of genes encoding putative effectors following this definition (Kämper et al., 2006; Martin et al., 2008; Haas et al., 2009; Duplessis et al., 2011; Zuccaro et al., 2011).

A group of effectors that is responsible for masking the pathogen from detection by the host immune system contains a LysM chitin binding motif. The effector Ecp6 from C. fulvum, for example, contains three LysM domains and sequesters chitin oligomers originating from the fungal cell wall before they are detected by plant PRRs, preventing PTI (de Jonge & Thomma, 2009; de Jonge et al., 2010). In addition, apoplastic LysM effectors have been found to contribute to the virulence of M. oryzae (Mentlak et al., 2012) and the wheat pathogen Mycosphaerella graminicola (Marshall et al., 2011). Similarly, the C. fulvum effector Avr4 also features chitin binding activity. Unlike Ecp6, Avr4 is not responsible for hiding chitin oligomers from the immune system. It binds directly to chitin in the fungal cell wall to protect it from degradation by secreted plant chitinases (van den Burg et al., 2004; van Esse et al., 2007). In addition to chitinases, glucanases can threaten the invading pathogen and hence become effector targets. Phytophthora sojae secretes the glucanase inhibitor GIP1, which inhibits the secreted soybean glucanase EgaseA before the release of elicitor-active glucan oligosaccharides from the pathogen cell wall by glucanase activity (Rose et al., 2002). In maize, β-1,3-glucanase activity is modulated by the secreted toxin Fumonisin B1, produced by Fusarium verticillioides (Sanchez-Rangel et al., 2012).

The first U. maydis effector identified as being essential for host colonization was the apoplastic effector Pep1 (Doehlemann et al., 2009) (Fig. 3). Ustilago maydis pep1 deletion mutants are immediately blocked on epidermal penetration and elicit strong defense responses, such as ROS accumulation, papilla formation and collapse of the infected cells (Doehlemann et al., 2009) (Fig. 4). Recently, the reason for the inability of the pep1 mutant to suppress host defense was identified: Pep1 effectively blocks the oxidative burst via a direct inhibition of secreted maize peroxidases (Hemetsberger et al., 2012). This finding establishes Pep1 as the first example of a fungal effector that interferes directly with the apoplastic ROS-generating system of the plant host. It also emphasizes that the activity of individual plant defense proteins and their inhibition by apoplastic effectors can decide the outcome of a biotrophic interaction.

Figure 3.

Localization of secreted effector proteins in the biotrophic interface of Ustilago maydis. (a) Confocal localization of Pit2-mCherry (red). The plant plasma membrane (PIN1-YFP, green) surrounds the invading hypha. At cell to cell passages, the effector diffuses into the apoplastic space between cells neighboring the infection (arrowheads). Bar, 10 μm. (b) Confocal localization of Pep1-mCherry (red). The effector surrounds the hypha and accumulates at cell to cell passages, resulting in a ring-shaped signal at the passage (arrowhead). Bar, 5 μm. (c) Three-dimensional rendering of a confocal stack showing a biotrophic U. maydis hypha secreting Pep1-GFP. Gray, maize cell walls; red, fungal hyphae, RFP; green, Pep1-GFP.The effector localizes at the biotrophic interface and accumulates at the tip of the hypha where it is secreted (arrowhead). Bar, 3 μm.

Figure 4.

Different effector mutants of Ustilago maydis trigger different layers of defense in maize. Maximum projections of confocal images showing U. maydis penetration events. (a) The U. maydis pep1 mutant (red, cytoplasmic RFP) fails to establish a biotrophic interface and elicits a strong defense response in penetrated epidermis cells (visualized by enhanced autofluorescence induced by UV laser (405 nm), cyan). Arrowhead, U. maydis appressorium. Bar, 15 μm. (b) The deletion of U. maydis pit2 does not affect initial penetration; however, a second layer of defense is triggered in the mesophyll, resulting in enhanced autofluorescence (arrowheads). Bar, 10 μm.

Another group of effectors also protects the pathogen from degradation by plant enzymes, particularly proteases. Given the crucial role of apoplastic proteases in plant immunity, it is not surprising that pathogen effectors target protease activity. One of the first identified protease inhibitors is the Kazal-like protein EPI1 from P. infestans, which inhibits subtilisin A activity in vitro and the subtilisin-like apoplastic serine protease P69B from tomato (Tian et al., 2004). Similarly, the related Kazal-like effector EPI10 also targets subtilisin A and P69B (Tian et al., 2005), suggesting that important components of the immune response can be targeted by multiple effectors to ensure immune suppression. In addition to serine proteases, plant cysteine proteases have been identified as prominent effector targets. Two secreted cystatin-like proteins from P. infestans inhibit the tomato cysteine proteases PIP1 and RCR3 (Tian et al., 2007; Song et al., 2009). As discussed in Section 'How they learned to live with the bomb – disarming proteases', the apoplastic papain-like protease C14 contributes to resistance against P. infestans. Therefore, its binding by the effector Avrblb2, which prevents the secretion of C14 into the apoplast, promotes P. infestans virulence (Bozkurt et al., 2011). The C. fulvum effector Avr2 also acts as a protease inhibitor in tomato and interacts with the cysteine proteases RCR3 and PIP1 (Rooney et al., 2005; van Esse et al., 2008). Interestingly, the inhibition of RCR3 does not contribute to virulence (Dixon et al., 2000), suggesting that RCR3 merely acts as a decoy for R-mediated ETI. Indeed, the guard protein Cf-2 from tomato recognizes the Avr2–RCR3 complex and elicits a defense response (Rooney et al., 2005). This guard and decoy hypothesis could reflect an additional line of defense against imperfectly adapted pathogens (van der Hoorn & Kamoun, 2008). Nevertheless, the allergen-like effector Gr-VAP1 from the nematode Globodera rostochiensis was also found to target RCR3 (Lozano-Torres et al., 2012). This finding argues for an important role of RCR3 as an effector target, whether as a decoy or as a ‘true’ target. Whatever the role of RCR3 in its interaction with effectors, the given examples of protease targeting virulence factors strongly suggest a central role of PLCPs as conserved plant immune components. Recent evidence for this idea comes from the U. maydis–maize system, where the effector Pit2 was identified as an inhibitor of apoplastic PLCPs (Mueller et al., 2013). Pit2 also localizes in the apoplast (Fig. 3) and is essential for tumor induction in maize plants, that is, pit2 deletion mutants are blocked by a plant defense response during host colonization (Doehlemann et al., 2011) (Fig. 4). Pit2 inhibits directly the proteolytic activity of the apoplastic maize PLCPs CP1, CP2 and XCP2, which were found previously to be targets of the maize cystatin CC9 (van der Linde et al., 2012a; Mueller et al., 2013). Mutational analyses of Pit2 revealed that protease inhibition, which is essential for U. maydis virulence, depends on the conserved residues of a 14-amino-acid motif that serves as the protease inhibitor domain (Mueller et al., 2013). This again substantiates the fact that apoplastic PLCPs need to be blocked by both plant and pathogen factors to efficiently suppress defense induction during the entire infection process.

VII. Back to the future – challenges and opportunities

In the recent past, enormous progress has been made in the understanding of defense-related plant signaling pathways and the functional characterization of microbial effectors, which either suppress cell death pathways or act on phytohormone signaling (Djamei et al., 2011; Dou & Zhou, 2012). By contrast, our knowledge of the early interactions of host immune mechanisms with microbial virulence factors which determine pathogen entry is rather fragmentary. The reason for our lack of knowledge on apoplastic interactions may be the functional redundancy of the contributing factors, which makes it hard to detect phenotypes in reverse genetic approaches. CWDEs may be an example of such functional redundancy. Furthermore, research on apoplastic effector–target interactions currently is limited by various experimental pitfalls. This holds true for microscopic analyses, where the spatial resolution of light microscopy is not sufficient to discriminate any distinct localization within the biotrophic interface which, in the case of U. maydis, spans only c. 200 nm. In addition, fluorescence tags, such as green fluorescent protein (GFP), show little or even no function in an apoplastic environment. In some plant systems, it has been shown that affinity or fluorescence tags are entirely cleaved off in the apoplast (van Esse et al., 2006), which makes both protein localization and the identification of interaction partners a challenging enterprise. Great potential also lies within improved bioinformatics approaches to predict pathogen effectors with apoplastic localization, as well as the corresponding defense components. The combination of functional, biochemical and structural properties of individual effector proteins with the analysis of rapidly increasing genome data may allow the identification of novel motifs and interaction sites of apoplastic proteins.

The known examples of pathogen effectors that function in the apoplast have led to the conclusion that plants depend on basic defense strategies that are likely to be highly conserved amongst species (Fig. 5). Hence, in all host plants, an invading pathogen is challenged with similar components of basal defense responses, which suggests a convergent development of suppression strategies that is represented by the primary wave of secreted core effectors. Conserved processes that have been revealed from studies on different pathosystems are the masking and sequestering of chitin, suppression of the oxidative burst and inhibition of apoplastic proteases (Fig. 5). Nevertheless, direct evidence for interactions of effectors with their host targets is still rare (Table 1). For the oxidative burst, U. maydis Pep1 is the only effector identified so far to directly target a central component of the ROS-generating machinery. However, at the molecular level, also for Pep1 the mode of peroxidase inhibition remains elusive. Given the fundamental role of Pep1 for U. maydis infection, it is very likely that other pathogens, particularly biotrophs, deploy effectors with similar functions. Oxidative burst reactions, however, can be induced by different stimuli in aboveground parts, as well as the roots, of plants, and therefore one can assume that mechanisms other than that employed by Pep1 have evolved.

Table 1. Apoplastic effectors of filamentous pathogens and their function
EffectorOrganismLife stylePhenotype/taskKnown targetCitation
Avr2 Cladosporium fulvum BiotrophCysteine protease inhibitionPIP1, RCR3Rooney et al. (2005); van Esse et al. (2008)
Avr4 C. fulvum BiotrophChitin shieldingChitinvan den Burg et al. (2004)
Avr4E C. fulvum Biotroph??Westerink et al. (2004)
Avr9 C. fulvum BiotrophCarboxypeptidase inhibitor?Van den Ackerveken et al. (1992)
Ecp1 C. fulvum BiotrophReduced sporulation in planta?Van den Ackerveken et al. (1993); Lauge et al. (1997)
Ecp2 C. fulvum BiotrophReduced host colonization?Van den Ackerveken et al. (1993); Lauge et al. (1997)
Ecp4 C. fulvum Biotroph??Lauge et al. (2000)
Ecp5 C. fulvum Biotroph??Lauge et al. (2000)
Ecp6 C. fulvum BiotrophReduced virulence/chitin sequestrationChitinBolton et al. (2008); de Jonge et al. (2010)
Ecp7 C. fulvum BiotrophReduced virulence?Bolton et al. (2008)
ChEC3 Colletotrichum higginsianum Hemibiotroph??Kleemann et al. (2012)
ChEC6 C. higginsianum Hemibiotroph??Kleemann et al. (2012)
ChEC89 C. higginsianum Hemibiotroph??Kleemann et al. (2012)
NIS1 Colletotrichum orbiculare Hemibiotroph??Yoshino et al. (2012)
Six1Fusarium oxysporum f. sp. lycopersiciHemibiotrophRequired for full virulence?Rep et al. (2004)
Six2F. oxysporum f. sp. lycopersiciHemibiotrophProbably not required for virulence?Houterman et al. (2007)
Six3F. oxysporum f. sp. lycopersiciHemibiotrophRequired for full virulence?Houterman et al. (2007)
Avr1(Six4)F. oxysporum f. sp. lycopersiciHemibiotrophSuppression of I-2 and I-3 resistance?Houterman et al. (2007)
Fumosin B1 Fusarium verticillioides HemibiotrophGlucanase inhibitionβ-1,3-GlucanasesSanchez-Rangel et al. (2012)
Avrblb2 Phytophthora infestans HemibiotrophInhibition of C14 secretion (affects apoplastic immunity)C14 proteaseBozkurt et al. (2011)
EPI1 P. infestans HemibiotrophSerine protease inhibitionSerine proteasesTian et al. (2004)
EPI10 P. infestans HemibiotrophSerine protease inhibitionSerine proteasesTian et al. (2005)
EPIC1 P. infestans HemibiotrophCysteine protease inhibitionPIP1, RCR3Tian et al. (2007); Song et al. (2009)
EPIC2B P. infestans HemibiotrophCysteine protease inhibitionPIP1, RCR3Tian et al. (2007), Song et al. (2009)
GIP1 Phytophthora sojae HemibiotrophGlucanase inhibitionβ-1,3-GlucanasesRose et al. (2002)
AvrLm1 Leptosphaeria maculans Hemibiotroph??Gout et al. (2006)
AvrLm4-7 L. maculans Hemibiotroph??Parlange et al. (2009)
Mg1LysM, Mg3LysM Mycosphaerella graminicola HemibiotrophChitin sequestration & shieldingChitinMarshall et al. (2011)
AvrP(123) Melampsora lini BiotrophContains serine protease inhibitor motif?Catanzariti et al. (2006)
BAS2–BAS4 Magnaporthe oryzae Hemibiotroph??Mosquera et al. (2009)
MC69 M. oryzae/Colletotrichum orbiculare HemibiotrophRequired for virulence?Saitoh et al. (2012)
Slp1 M. oryzae HemibiotrophChitin sequestrationChitinMentlak et al. (2012)
Nip1 Rhynchosporium secalis Hemibiotroph??Rohe et al. (1995)
Nip2 R. secalis Hemibiotroph??Rohe et al. (1995)
Nip3 R. secalis Hemibiotroph??Rohe et al. (1995)
Pep1 Ustilago maydis BiotrophPeroxidase inhibitionPOX12Doehlemann et al. (2009); Hemetsberger et al. (2012)
Pit2 U. maydis BiotrophCysteine protease inhibitionCP2, CP1A/B, XCP2 proteasesDoehlemann et al. (2011); Mueller et al. (2013)
Ave1 Verticilium dahliae HemibiotrophContributes to virulenceVe1 receptorde Jonge et al. (2012)
Figure 5.

The role of selected effectors in plant defense suppression. Schematic overview of a hypothetical plant infection by a filamentous microbial pathogen. Components of basic plant defense and interfering pathogen effectors are depicted. See text for details on the individual interactions. Abbreviations used in the figure. Microbial effectors: Avr2, Avr4, Avrblb2, Ecp6, EPIC1, GIP1, Pep1, Pit2. Plant proteases: C14, PIP1, PLCP (papain-like cysteine protease), Rcr3. ROS-generating enzymes: POX (peroxidase), RBOH (respiratory burst oxidase homolog). BAK1, Arabidopsis bri1-associated receptor kinase-1; C-genes, compatibility genes; CC9, maize cystatin ‘corn cystatin 9’; Cf-2, tomato resistance protein Cf-2; JA, jasmonic acid; MAPK, mitogen-activated protein kinases; PAMP/DAMP, pathogen/damage-associated molecular patterns; PR genes, genes encoding for PR proteins; PRR, pattern recognition receptor; ROS, reactive oxygen species; SA, salicylic acid.

Increasing knowledge from diverse systems indicates that PLCPs hold a key function in plant defense signaling. Given the major impact of their activity on apoplastic immunity, it will be a major challenge to elaborate the nature of the signals that are released by their activity. A tempting possibility would be the release of small peptides, which could then act as DAMP signals. A similar mechanism has been described for PEPs that have been found to stimulate JA signaling in A. thaliana and maize (Huffaker et al., 2006, 2011). Whatever the targets of apoplastic cysteine proteases, their striking potential to activate PR gene expression illustrates that one cannot understand apoplastic processes independently from the intracellular processes that are modulated during infection.

On top of the first line of apoplastic virulence factors are the more specialized, often translocated, effectors that interfere with specific (intra-)cellular processes of the host to promote virulence. For example, the transformation of organ primordia to plant tumors which is induced by U. maydis requires highly specialized effectors that act in an organ- or even cell-type-specific manner (Skibbe et al., 2010). By contrast, the apoplastic ‘core effectome’ interferes with central components of innate plant immunity to set the basis for the initial establishment of a compatible interaction. The identification of the key players in this battle will help us to answer some of the central questions in plant pathology, namely what makes a pathogen successful and what makes a plant become a vulnerable host.


We are grateful to Dr Stefanie Reißmann, Dr Armin Djamei and Dr Karina van der Linde for critical comments on the manuscript. Our research is funded by the Max-Planck-Institute for Terrestrial Microbiology, the German Research Foundation (DFG), the German Environmental Foundation (DBU) and the German Academic Exchange Service (DAAD).