2.1 The fungal life cycle

As mentioned above, most PM fungi grow epiphytically on their respective host plants. Only single PM species of the genera Leveillula and Phyllactinia are an exception, as they propagate (L. taurica) or form haustoria (P. guttata) endophytically in the leaf mesophyll tissue after entering through stomata (Boesewinkel, 1980). In natural environments, PM conidiospores (mitotic, asexual spores) are mostly distributed by wind or animals. Under laboratory conditions, inoculations are performed by brushing, leaf-to-leaf transfer or dusting of spores from infected material onto healthy plants (Micali et al., 2008). Once situated on a plant leaf or stem, the PM spore develops a short germ tube (Figure 2), and approximately six hours post inoculation (hpi), the appressorium, a thickened infection structure, forms at the tip of this hypha. At least in the case of Bgh the appressorium builds up high pressure in order to breach the plant cuticle and cell wall (Pryce-Jones et al., 1999). Unlike in many other plant-pathogenic fungi, cell wall-degrading enzymes seem to play a minor role in host cell invasion, as Bgh has a comparatively low number of genes encoding such carbohydrate-active enzymes (CAZymes; Spanu et al., 2010). After successful cell wall penetration, the fungus enters the host cell without disrupting the host plasma membrane and the haustorium, a specialized hyphal feeding structure with protrusions for surface enlargement, is formed (12–14 hpi; Figure 2). Haustorium development involves the formation of the extrahaustorial membrane (EHM), which separates plant and fungal structures. The haustorium represents the major interaction site between the fungus and the host plant, and it is supposed to be the hub for effector secretion and nutrient uptake (reviewed in O'Connell and Panstruga, 2006). Supposedly once the haustorium is established, the fungus gains the nutrients necessary for its epiphytic growth. This becomes visible as secondary hyphae forming the PM colony (from ca. 24–48 hpi onwards; Figure 2). Further, the secondary hyphae form new appressoria and penetrate nearby cells. The cycle concludes by the formation of conidiophores, specialized hyphae giving rise to new conidiospores (3–7 days post inoculation (dpi); Figure 2). Sporulation of the two preferentially studied Arabidopsis-infecting PMs, Gc UCSC1 and Go, becomes macroscopically visible at 7 to 10 dpi (see Figure 6).

Figure 1.

Graphical overview of the topics discussed in this article.

EHM, extrahaustorial membrane; EHMx, extrahaustorial matrix; ER, endoplasmic reticulum; JA, jasmonic acid; MAPK(KK/KKK), mitogen-activated protein kinase (kinase kinase/kinase kinase kinase); MVBs, multivesicular bodies; PRRs, pattern recognition receptors; RPW8, RESISTANCE TO POWDERY MILDEW 8 protein; ROS, reactive oxygen species; SA, salicylic acid; TGN, trans-Golgi network. See text for further explanations.

In temperate climates, PM fungi have to overwinter periods during which the host plant is either not present (annual plants) or defoliates (perennial plants). To cope with such conditions, the fungal pathogen can engage in sexual reproduction based on two compatible mating types. This process gives rise to endurable ascospores (meiospores) enclosed in asci, emerging from fruiting bodies (cleistothecia or chasmothecia). These structures form in the mesophyll and are visible as black-brownish spots on leaves in fall. The ascospores mature within the ascus and are able to persist for longer periods outside the host plant. Meta-analysis of genomic data, however, suggests that at least in the case of grass PMs sexual reproduction is a comparatively rare event (Hacquard et al., 2013; Wicker et al., 2013). To our knowledge, the formation of fruiting bodies has not been demonstrated on Arabidopsis.

Figure 2.

Asexual life cycle of G. orontii in association with Arabidopsis.

The central part of the figure illustrates schematically the key steps of the life cycle, while the micrographs show the actual fungal infection structures. The confocal laser scanning micrographs were obtained from transgenic Col-0 plants stably expressing yellow cameleon inoculated with Go. Fungal infection structures were stained with FM4-64 (shown in red) while green fluorescence is representative of cytosolic yellow cameleon fluorescence. Bars: 20 µm.

2.2 The haustorium

The PM haustorium is the only fungal structure that resides within the plant, namely inside plant epidermal cells (with the exception of L. taurica and P. guttata; see section 2.1). As mentioned above, this structure likely represents the main interaction site between the plant and the fungus (O'Connell and Panstruga, 2006). However, four layers separate the haustorial cytoplasm from the plant cytoplasm: the hausto rial plasma membrane, the fungal cell wall, the extrahaustorial matrix (EHMx), and the EHM. The EHM is a plant-derived membrane surrounding the haustorium. Despite the continuity of the EHM with the host plasma membrane, its composition is, however, distinct from the latter (Koh et al., 2005; O'Connell and Panstruga, 2006; Micali et al., 2011). The EHM attaches to the hausto rial neck, the contact site of the haustorium and the plant cell wall, which separates the EHMx from the apoplast (Gil and Gay, 1977). The EHMx forms the transition zone between plant and fungus and is supposed to enable both nutrient uptake and effector delivery (Bushnell, 1972).

Mature Go haustoria are typically ca. 16 µm wide and 10 µm long elliptic bodies with finger-like projections coiled around the main body (Figure 3A–B; Micali et al., 2011 ). They contain a single nucleus and numerous mitochondria. In addition, the hausto rial cytoplasm and the EHMx comprise a high number of vesicles, potentially due to fusion of multi-vesicular bodies (MVBs) with the plasma membrane resulting in the release of cargo vesicles into the EHMx (exosomes; Figure 3A). On the plant side, the endoplasmic reticulum (ER) and plant MVBs locate close to the EHM (Micali et al., 2011).

Mature haustoria are often fully or partially encapsulated by encasements and cell wall appositions enclosing the EHM and EHMx, even during the compatible interaction between Go and Arabidopsis (Figure 3C). In fact, 20–55 % of Go hausto ria are encased to different degrees. These encapsulations depend on the age of the haustorium and contain β-1,3-polyglucans (e.g. callose), xyloglucans, rhamnogalacturonans, and arabinogalactan proteins. Deposition starts at the haustorial neck and gradually encloses the maturing haustorium (Meyer et al., 2009; Micali et al., 2011). Compared with papillae (see section 3.3), which represent multi-layered focal cell wall reinforcements (Naumann et al., 2013), encasements seem to comprise a uniform single layer surrounding the hausto ria (Micali et al., 2011). Although they typically contain callose, the formation of these encapsulations is independent from the pathogen-induced callose synthase GLUCAN SYNTHASE-LIKE 5/POWDERY MILDEW RESISTANT 4 (GSL5/PMR4: At4g03550), suggesting that in the absence of the enzyme other cell wall polymers replace the β-1,3-polyglucan (Meyer et al., 2009). The hypothesis that encasements indicate incomplete adaption of Go to Arabidopsis is supported by the fact that they are absent in interactions with Gc (Koh et al., 2005; Meyer et al., 2009). Moreover, haustoria of Bgh are encapsulated in leaves of the nonhost plant Arabidopsis, but not in leaves of its host plant barley, indicating that Bgh effectively suppresses the encasement of haustoria in a suitable host (Meyer et al., 2009).

Figure 3.

The PM haustorium.

The fungal haustorium forms within cells of the leaf epidermis after penetration. A. Scheme of a PM haustorium (grey) separated from the plant cytoplasm by fungal haustorial membrane (fHM), fungal cell wall (fCW), extrahaustorial matrix (EHMx) and extrahaustorial membrane (EHM). The inset depicts the proposed exocytosis of fungal multivesicular bodies (fMVBs) B. Wheat germ agglutinin staining of chitin in an isolated haustorium of Go. The confocal laser scanning micrograph shows a mature haustorium body (HB) with numerous haustorial lobes (L). C. Partial callose encasement of an isolated Go haustorium. The electron-opaque EHM (arrowheads) surrounds the haustorium (H) but not the callose-containing encasement (E). Bars: B 5 µm; C 2 µm. Panels B and C reproduced with permission from (Micali et al., 2011) (Copyright by John Wiley & Sons (Cellular Microbiology)).

3.1 MAMP-triggered responses

The first barriers PM pathogens encounter during infection are the cuticle and epicuticular waxes overlying the plant cell wall (Malinovsky et al., 2014). As mentioned above (see section 2.1), PMs presumably employ mainly hydrostatic pressure to penetrate this preformed perimeter of epidermal cells. Accordingly, the plant can sense the pathogen in several ways. Firstly, the pressure exerted on the plant cell might activate plant mechanosensors (Bhat et al., 2005; Ellinger and Voigt, 2014a). Secondly, damage-associated molecular patterns (DAMPs) released by the breakdown of the plant cell wall, or microbe-associated molecular patterns (MAMPs) released by the fungus, can be detected by pattern recognition receptors (PRRs) and activate immune signaling (Boiler and Felix, 2009).

The carbohydrate polymer chitin is a major constituent of fungal cell walls and when exogenously applied to Arabidopsis activates MAMP-triggered immune responses. Chitin is perceived by the membrane-localized PRRs CHITIN ELICITOR RECEPTOR KINASE 1 (CERK1 : At3g21630) (Miya et al., 2007) and the LYSIN MOTIF RECEPTOR-LIKE KINASEs 4/5 (LYK4/5: At2g23770/ At2g33580) (Cao et al., 2014). Application of MAMPs, including chitin, leads to the accumulation of defense-related proteins and the deposition of callose at seemingly random locations in treated tissues (Gómez-Gómez et al., 1999; Luna et al., 2011; Underwood and Somerville, 2013). This phenomenon is similar to the localized formation of papillae (see section 3.3) and suggests that MAMP-induced PRR activation alone can trigger the establishment of papilla-like structures. The adapted PM Gc shows increased sporulation on cerk1 mutants in comparison to wild type plants, which suggests that signaling through CERK1 contributes to basal resistance to the PM disease (Wan et al., 2008). It is presently unknown whether Iyk4 or Iyk5 mutants are more susceptible to PM as well. Presumably, plants can perceive further PM-derived MAMPs. However, the only other known molecule from a PM pathogen that activates defense gene expression and decreases fungal growth in various cereals after application is a soluble carbohydrate elicitor isolated from conidia of the wheat PM pathogen, Blumeria graminis f.sp. tritici (Bgt) (Schweizer et al., 2000).

3.2 Papilla formation

Plant cell responses to the early detection events mentioned above include polarization of cellular organelles and rearrangement of cytoskeletal elements (microtubules and actin filaments) below the attack site (Schmelzer, 2002; Hückelhoven and Panstruga, 2011). Underneath the attempted penetration site defense-related proteins focally accumulate (Assaad et al., 2004; Bhat et al., 2005; Kwon et al., 2008b; Meyer et al., 2009; Kwaaitaal et al., 2010). Furthermore, the plasma membrane is altered locally and gains lipid raft-like properties (Bhat et al., 2005). Both virulent and non-virulent PM fungi induce the formation of a small, dome-like structure called papilla below the incipient fungal appressorium (Collins et al., 2003; Assaad et al., 2004; Koh et al., 2005). Among other components such as membranous vesicles, the papilla contains callose (Figure 4), silicon, reactive oxygen species (ROS) and phenolic compounds. The resulting structure is believed to reinforce the cell wall to prevent fungal invasion (Zeyen et al., 2002). This hypothesis is supported by a correlation between the timing of papilla formation and PM resistance. Mutation of the target membrane SOLUBLE N-ETHYLMALEMIDE-SENSITIVE FACTOR ATTACHMENT PROTEIN RECEPTOR (t-SNARE) PENETRATION (PEN)1/SYNTAXIN OF PLANTS (SYP)121 (At3g11820) and the vesicle-associated SNAREs VESICLE-ASSOCIATED MEMBRANE PROTEINS (VAMPs)721/722 (At1g04750/At2g33120) delays papilla formation and increases Bgh penetration success (Assaad et al., 2004; Kwon et al., 2008b; Böhlenius et al., 2010). The ubiquitin ligase ARABIDOPSIS TOXICOS EN LAVADURA 31 (ATL31 : At5g27420), which is a regulator of responses to changes in the cellular carbon/nitrogen ratio, interacts with PEN1 in co-immunoprecipitation experiments. Transient expression assays in Nicotiana benthamiana leaves inoculated with Bgh show accumulation of ubiquitination activity deficient ATL31^C143S-GFP around papillae and in vesicle-like structures near papillae, while fluorescence is undetectable for ATL31-GFP. Furthermore, overexpression of ATL31 in Arabidopsis pen1–1 mutants results in enhanced penetration resistance to Bgh and faster formation of papillae (Maekawa et al., 2014). The same observation was made for RAB GTPASE HOMOLOG 4c (RABA4c: At5g47960) overexpression lines (Ellinger et al., 2014), suggesting that only the timely formation of papillae can restrict invasion by PM fungi.

Figure 4.

Nanoscale resolution of callose polymer fibrils in pathogen-induced cell wall papillae.

Three-week-old Arabidopsis wild type and pathogen-resistant PMR4-GFP overexpressing lines (P35S::PMR4-GFP) were inoculated with the adapted powdery mildew Gc. Localization microscopy (dSTORM: direct stochastical optical reconstruction microscopy) of aniline blue-stained callose polymer fibrils in pathogen-induced papillae at sites of attempted fungal penetration at 12 hpi in rosette leaves. Scale bars = 2 µm. Unpublished micrograph, courtesy of Christian A. Voigt and Dennis Eggert.

3.3 Callose deposition

Accumulation of the β-1,3 polyglucan callose, as the major constituent of papillae, is a generic response to pathogen challenge (Ellinger and Voigt, 2014b). Together with the β-1,4 polyglucan cellulose, callose generates a three-dimensional network of ∼250 nm fibrils, which can provide protection against cell wall hydrolysis by fungal enzymes (Figure 4; Eggert et al., 2014). Callose could not be unambiguously linked to resistance for a long time, and work on a mutant of the GSL5/PMR4 gene initially even suggested that loss of papillary callose reduces sporulation of adapted PMs (Jacobs et al., 2003; Nishimura et al., 2003). However, the inhibition of post-invasive fungal growth in pmr4 mutants relies on hyper-induced salicylic acid (SA) responses upon PM attack. When the pmr4 knockout mutation is combined with a further mutation that leads to a loss of SA biosynthesis or signaling the increased resistance is compromised (Nishimura et al., 2003). Except for the loss of callose, papillae of pmr4 mutant plants have a similar appearance as papillae of wild type plants (Nishimura et al., 2003). While loss of callose in the pmr4 mutant has only limited impact on penetration resistance (Jacobs et al., 2003; Ellinger et al., 2013), increased callose deposition after PM attack caused by PMR4 overexpression results in full penetration resistance to both Gc and Bgh. The latter effect seems to correlate with structural differences of papillae in PMR4 overexpression lines compared to the wild type, as the transgenic lines show larger cores of callose-dense deposits, whereas wild type papillae display a more diffuse structure (Naumann et al., 2013). Together these findings indicate that additional papillary components support the contribution of callose to prevent fungal penetration (Ellinger et al., 2013). The PMR4-GFP fusion protein focally accumulates at the PM attack site and its presence coincides with the occurrence of callose deposits. The callose accumulations in the PMR4-GFP overexpression line are not only enlarged, but also deposited in a layer facing the fungus on top of the cellulose microfibrilar network (Eggert et al., 2014). The increase in the proportion of callose presumably protects the cellulose component of papillae from enzymatic digestion (Eggert et al., 2014). In contrast to the increased post-penetration resistance in pmr4 mutants, the increase in resistance caused by PMR4-GFP overexpression is independent from SA- or jasmonic acid (JA)-mediated defense (Ellinger et al., 2013).

Similar to what was reported for barley (Böhlenius et al., 2010), ADP ribosylation factor-GTP exchange factor (ARF-GEF)-mediated vesicle trafficking is essential for callose accumulation in papillae in Arabidopsis (see section 7; Nielsen et al., 2012). This ARF-GEF-dependence indicates that either PMR4 accumulation at fungal attack sites, the delivery of callose precursors, and/or the callose deposition process itself involves vesiclemediated transport processes. PMR4 interacts with and acts as an effector of the small GTPase of the Ras (rat sarcoma) superfamily, RABA4c (Ellinger et al., 2014). RABA4c expression is transiently upregulated prior to callose deposition in response to biotic stress. Knockouts of RABA4c exhibit a delayed increase of callose synthase activity, slightly reduced numbers of callose deposits, and slightly increased Gc penetration rates. By contrast, overexpression of RABA4c results in full penetration resistance to Gc and hastens and increases callose deposition. Both effects depend on the presence of PMR4 and RABA4c GTPase activity. The RABA4c localization to membranes is independent on the prenylation of its C-terminal CaaX motif (‘C’ cysteine, ‘a’ aliphatic amino acid, ‘X’ variable amino acid). A C-terminal RABA4c-mCitrine fusion lacking this lipid modification still localizes to membranes, though solely when PMR4 is present, which supports the finding that both proteins physically interact in planta (Ellinger et al., 2014). Rab (Ras-related in brain) GTPases play major roles in virtually all vesicle trafficking processes in eukaryotic cells. To what extent PMR4 localization to the plasma membrane or to focal accumulation sites depends on RABA4c-dependent vesicle trafficking pathways is unknown.

3.4 Extracellular deposition of proteins into papillae

The discovery that components of a SNARE protein complex are involved in penetration resistance suggests that these proteins directly control vesicle fusion at the PM attack site (Collins et al., 2003; Kwon et al., 2008a; Kwon et al., 2008b). After vesicle fusion and cargo release, SNARE proteins are usually recycled and stay on the cytosolic side of the plasma membrane (Kwon et al., 2008a). Surprisingly, in case of the focal accumulation of SNARE proteins at attempted fungal entry sites this is not the case. Instead, fluorescent fusions of PEN1, SOLUBLE N-ETHYLMALEIMIDE-SENSITIVE FACTOR ADAPTOR PROTEIN (SNAP)33 (At5g61210; a t-SNARE) and the ATP-binding cassette (ABC) transporter PENS (At1g59870) accumulate within papillae and haustorial encasements and therefore end up in the extracellular (apoplastic) space. Within cell wall appositions, GFP-PEN1 co-localizes with the lipophilic fluorescent tracer of endosomes, FM4-64, indicating that membrane material co-accumulates with these proteins in papillae and haustorial encasements (Meyer et al., 2009; Nielsen et al., 2012). As demonstrated by electron microscopy and co-localization with the Rab-like GTPase MVB marker ARA6/RABF1-GFP (At3g54840), MVBs focally accumulate at pathogen attack sites (see Figure 7). It is therefore conceivable that MVBs contribute to extracellular deposition of otherwise intracellularly localized proteins (An et al., 2006; Meyer et al., 2009; Nielsen et al., 2012). According to this hypothesis, vesicles containing PEN1, SNAP33 and PENS may sort and incorporate into the lumen of MVBs after endocytosis. These MVBs might subsequently fuse with the plasma membrane, which could explain extracellular protein delivery (Meyer et al., 2009). The extent and significance of the extracellular deposition of otherwise intracellular proteins for PM resistance is currently unknown.

3.5 Silicon-mediated resistance

Silicon (Si) contributes to PM resistance of cereals and various other plant species (Fauteux et al., 2005). Accordantly, watering Arabidopsis plants with silicon results in a lower PM disease incidence although Arabidopsis lacks dedicated Si transporters (Ghanmi et al., 2004). Accumulation of insoluble Si at PM attack sites led to the hypothesis of Si acting as a simple physical barrier (Bélanger et al., 2002). However, not in every case presence of insoluble Si correlates with increased resistance to fungal penetration. Consequently, a physiological or biochemical role in mediating cellular resistance has been postulated (Bélanger et al., 2003). While Si fertilization alone has a minor effect on transcript abundance, Gc inoculation of Si fertilized plants versus Gc inoculation of non-Si supplemented plants attenuated the magnitude of PM-induced down-regulation of genes by more than 25 % (Fauteux et al., 2006). As many of these PM-repressed genes are related to primary metabolism, the Si-mediated reduced downregulation might indicate stress alleviation. Consequently, Si feeding potentially facilitates a more efficient response to PM infection (Fauteux et al., 2006). This hypothesis is further corroborated by transgenic Arabidopsis plants stably expressing the wheat Si transporter TaLsi1. These plants have increased Si levels and concomitantly further enhanced PM resistance in the presence of Si compared to wild type plants (Vivancos et al., 2015).

6.1 EDR genes

Transcriptional activation of PATHOGENESIS-RELATED (PR) genes is one hallmark of induced defense (reviewed in Loake and Grant, 2007). During a genetic screen aimed to identify novel elements of plant defense, three mutants with enhanced disease resistance (edr1 (At1g08720), edr2 (At4g19040) and edr3 (At3g60190)) that do not express PR1 (At2g14610) upon inoculation with Gc were isolated (Frye and Innes, 1998; Frye et al., 2001; Tang and Innes, 2002; Tang et al., 2005a; Tang et al., 2005b; Tang et al., 2006). Interestingly, all three mutants show characteristics of “late-acting” resistance (i.e., at 5 to 8 dpi), which is associated with accelerated mesophyll cell death leading to macroscopic patches of lesions and either drastically reduced or absent sporulation. Genetic epistasis analysis revealed that edr-mediated resistance is SA-dependent and JA-independent (Frye and Innes, 1998; Tang et al., 2005b, 2006).

EDR1 encodes a mitogen-activated protein kinase kinase kinase (MAPKKK) that negatively regulates plant disease resistance (Frye et al., 2001). The edr1 mutant displays enhanced cell death during infection with the adapted PM pathogen Gc and in response to drought stress (Frye et al., 2001; Tang et al., 2005b; Tang et al., 2005a). Cell death associated with edr1 resistance requires the E3 ubiquitin ligases ATL1 (At1g04360) and KEEP ON GOING (KEG: At5g13530). Both E3 proteins are inhibited by interaction with EDR1, and the cell death phenotypes associated with edr1 are suppressed upon their depletion, indicating that EDR1 acts as a negative regulator of programmed cell death (Serrano et al., 2014). KEG possibly recruits EDR1 to the trans-Golgi network (TGN) and in turn EDR1 regulates E3 ligase activity of KEG to further suppress cell death (see section 7; Gu and Innes, 2011; Liu and Stone, 2013). Overexpression of ATL1 causes extensive cell death, which depends on its E3 ligase activity. Strikingly, knockdown of ATL1 expression does not only interfere with edr1-mediated cell death, but causes hypersusceptibility to PM infection, demonstrating that ATL1 is a positive regulator of pathogen-induced cell death (Serrano et al., 2014). A further link of EDR1 to suppression of cell death is provided by its inhibitory interaction with mitogen-activated protein kinase kinase (MKK)4 (At1g51660) and MKK5 (At3g21220) that are part of the MAPK cascade fine-tuning plant immunity (see section 8.1; Zhao et al., 2014).

EDR2 encodes a mitochondrial protein with a pleckstrin homology domain and a steroidogenic acute regulatory protein-related lipid transfer (START) motif. Both EDR1 and EDR2 function in a common genetic pathway as evidenced by the edr1 edr2 double mutant, showing resistance phenotypes that are indistinguishable from the respective single mutants (Tang et al., 2005b). In addition, edr1 and edr2 both display enhanced senescence in response to ET. Interestingly, mutations in the aminotrans-ferase AGD2-LIKE DEFENSE RESPONSE PROTEIN1 (ALD1: At2g13810) suppress edr2-mediated phenotypes including PM resistance, programmed cell death and ET-induced senescence, but not the edr1 edr2 double mutant phenotype (Nie et al., 2011). This raises the question how EDR1 and EDR2 activities are coordinated during the regulation of defense, cell death and ET-induced senescence.

Different from EDR1 and EDR2, EDR3 seems to function in a separate pathway, since edr3 does not display an early senescence phenotype. EDR3 encodes a dynamin-like protein localized partially to mitochondria. Despite the absence of a constitutive cell death phenotype in Arabidopsis, the mammalian counterpart of EDR3 plays a role in regulating mitochondrial dynamics associated with programmed cell death (Tang et al., 2006).

Recently, a fourth EDR gene, EDR4 (At5g05190), with unknown protein function and preferential localization of the gene product at the plasma membrane and endosomal compartments, has been isolated. Like previously identified EDRs, EDR4 is involved in negative regulation of SA-dependent PM resistance (Wu et al., 2015). EDR4 functions in the same pathway as EDR1 and EDR2 and interacts with EDR1, recruiting it to fungal penetration sites. The shared phenotypic features of edr mutants suggest a general link between SA-mediated resistance, mitochondrial function and programmed cell death (Ausubel, 2005).

6.2 PMR genes

In a genetic screen with the aim to identify susceptibility factors involved in interactions between Arabidopsis and the PM pathogen Gc, six powdery mildew resistant mutants, pmr1 to pmr6, were isolated. Four of the corresponding genes, namely PMR2 (At1g11310), PMR4/GSL5 (see section 2.2, 3.3 and 5), PMR5 (At5g58600) and PMR6 (At3g54920), have been cloned and to some extent functionally characterized (Vogel and Somerville, 2000; Vogel et al., 2002; Jacobs et al., 2003; Nishimura et al., 2003; Vogel et al., 2004; Consonni et al., 2006). The pmr2 mutant is defective in MILDEW RESISTANCE LOCUS O (MLO)2 (At1g11310), which encodes an integral membrane protein of unknown function (see section 6.3). PMR5 belongs to a large plant-specific gene family of unknown function and PMR6 encodes a glycosyl-phosphatidyl-inositol (GPI)-anchored pectate Iyase-like protein (Vogel et al., 2002; Jacobs et al., 2003; Nishimura et al., 2003; Vogel et al., 2004; Consonni et al., 2006).

The latter pmr mutants, pmr5 and pmr6, are believed to impact cell wall integrity, further stressing the contribution of the cell wall to PM resistance. The Arabidopsis pmr5 mutant exhibits resistance to the adapted PM fungi Gc and Go, and enrichment of pectin as well as reduced pectin modification occurs in the cell walls of pmr5 plants (Vogel et al., 2004). In addition, PMR5 contributes to PEN2-mediated pre-invasion resistance to the non-adapted fungus Magnaporthe oryzae. The pen2 pmr5 double mutant shows enhanced penetration success of M. oryzae (Maeda et al., 2009), indicating that PMR5 is involved in host and nonhost resistance and emphasizing the importance of cell wall integrity for both types of resistance. PMR6 localizes at the plant cell wall, where it might degrade pectin. In line with this assumption, the pmr6 mutant displays increased pectin and uronic acid contents. Like pmr5, the pmr6 mutant is resistant to Gc and Go, which is in both cases independent of SA, ET and JA signaling (Vogel et al., 2002). The pmr5 pmr6 double mutant shows increased resistance compared to the respective single mutants, suggesting that the two genes may function separately during plant defense. Furthermore, PMR5 and PMR6 are involved in the regulation of ploidy in mesophyll cells underlying the fungal feeding sites (see section 9.4; Chandran et al., 2013).

6.3 MLO genes

Arabidopsis MLO susceptibility genes were isolated and characterized based on their sequence similarity to barley Mlo (Consonni et al., 2006) and identified as the pmr2 mutant in the above-mentioned forward genetic screen (Vogel and Somerville, 2000). According to phylogenetic analyses, there are 15 MLO genes distributed into five clades in Arabidopsis, of which MLO2, MLO6 (At1g61560) and MLO12 (At2g39200) belong to the same clade (Devoto et al., 2003; Acevedo-Garcia et al., 2014). mlo2 mutants display reduced penetration success and less sporulation after infection with the adapted PM fungus Go (Consonni et al., 2006). Interestingly, MLO2 controls penetration success of PM fungi together with MLO6 and MLO12. While the mlo6 and mlo12 single and double mutants do not show any resistance phenotype, they gradually increase resistance of mlo2 if combined in double and triple mutant combinations, with the mlo2 mlo6 mlo12 triple mutant being fully resistant (Figure 6; Consonni et al., 2006). MLO genes encode evolutionary ancient integral membrane proteins with seven transmembrane domains and unknown biochemical activity (Devote et al., 2003; Kusch et al., 2016). Besides Arabidopsis and barley, mutation of closely related MLO genes in tomato, pea and further plants render these host species resistant to PM infection, indicating a similar function of the respective proteins (Bai et al., 2008; Humphry et al., 2011).

Figure 6.

Macroscopic infection phenotypes of Col-0 and the mlo2 mlo6 mlo12 mutant.

Five-week-old wild type (Col-0) and mlo2 mlo6 mlo12 plants (in Col-0 genetic background) were inoculated with Go and photographs were taken one week after inoculation.

Similar to NHR, mlo2-mediated PM resistance does not depend on major phytohormone signaling pathways such as those relying on JA, ET or SA (Consonni et al., 2006). By contrast, all three PEN genes are required for mlo2-mediated resistance to PM (Consonni et al., 2006). These findings suggest that mlo-mediated resistance and NHR may share overlapping pathways in plant defense (Humphry et al., 2006). Besides the PEN proteins, CYP79B2 (At4g39950) and CYP79B3 (At2g22330), two cytochrome monooxygenases that catalyze the entry step towards the production of diverse indolic metabolites, including the Arabidopsis-specific phytoalexin camalexin and indole glucosinolates, are required for mlo2-mediated resistance. In contrast to CYP79B2 and CYP79B3, another cytochrome P450 monooxy-genase, PAD3 (At3g26830), which catalyzes the final step in camalexin biosynthesis, only plays a minor role in mlo2-mediated resistance (Consonni et al., 2010).

8.1 Salicylic acid-mediated resistance

As for other biotrophic interactions, SA-mediated defense signaling plays a pivotal role in Arabidopsis defense against adapted PM fungi: SA-dependent gene expression and immune responses increase in Arabidopsis leaves upon infection with PMs and contribute to restriction of colony expansion and reproduction of the fungi (Zimmerli et al., 2004; Chandran et al., 2009). Consequently, many mutants with defects in SA biosynthesis, accumulation and signaling exhibit enhanced susceptibility (hypersusceptibility) to Go and Gc (Dewdney et al., 2000; Chandran et al., 2009; Zhang et al., 2015a). Likewise, interference with SA accumulation by transgenic expression of NahG, a Pseudomonas salicylate hydroxylase that degrades SA, increases susceptibility against Gc isolates (Ederli et al., 2015; Zhang et al., 2015a). Despite these genetic indications for an involvement of SA in defense against PM infection, there is currently only limited direct evidence for increased SA levels during PM infection (Fabro et al., 2008).

As the final steps of SA biosynthesis in the leaf take place in chloroplasts (Strawn et al., 2007), export of the hormone is required for elevated cytosolic and nuclear SA levels. Consequently, loss-of-function mutation of DP-E2F-like 1 (DEL1 : At3g48160), a transcriptional repressor of the gene encoding the plastidic SA exporter EDS5/SALICYLIC ACID INDUCTION DEFICIENT (SID)1 (At4g39030), results in enhanced SA-dependent resistance against Go (Chandran et al., 2014). This phenotype correlates with elevated basal SA levels and increased transcript abundance of SA-responsive genes in the dell mutant (Chandran et al., 2014). Strikingly, DEL1 also promotes cell proliferation by repressing genes involved in the induction of endoreduplication (see section 9.4; Vlieghe et al.; Lammens et al., 2008). Together these findings suggest that DEL1-mediated control of SA levels regulates the balance between growth and immunity in developing leaves. The translation of distinct SA levels into specific defense responses occurs by the action of NPR proteins in Arabidopsis (reviewed in Pajerowska-Mukhtar et al., 2013; Seyfferth and Tsuda, 2014; Yan and Dong, 2014). The outcome of SA-mediated signaling depends on subcellular SA levels and the abundance of active NPR1 in the nucleus. Its paralogs, the SA receptors NPR3 (At5g45110) and NPR4 (At4g 19660), co-operatively fine-tune NPR1 degradation by their competitive SA concentration-dependent interaction with nuclear NPR1. Moderately elevated SA levels, as present during systemic acquired resistance (SAR), allow the accumulation of active NPR1 protein in the nucleus. Subsequent interaction of NPR1 with TGA transcription factors (binding to a TGACG nucleotide motif) promotes gene expression of SA-responsive genes and induces SA-mediated defense. The relevance of NPR1 for the Arabidopsis-PM interaction is substantiated by identification of NPR1 as a protein-protein interaction network hub during Go infection (Jiang et al., 2016). This NPR1 interaction network includes the TGA-interacting GLUTAREDOXIN 480 (GRX480/ROXY19: At1g28480) involved in regulating SA/JA antagonism (Zander et al., 2012) and several TGA transcription factors (TGA1 : At5g65210, TGA2: At5g06950, TGA3: At1g22070, TGA7: At1g77920) that can further regulate the expression of defense-related genes. Genes that show SA-dependent transcript accumulation during PM infection encode proteins involved in redox regulation, vacuolar transport, secretion, and signaling-relevant processes such as Ca²⁺ homeostasis and SA/JA cross talk (Chandran et al., 2009).

ROP GTPases (Rho (RAS homologue) of plants) are molecular switches and key regulators of immunity (Kawano et al., 2014). Mutation of the Arabidopsis ROP-GAPs ROPGAP1 (At5g22400) and ROPGAP4 (At3g11490), trapping their (yet unidentified) target ROPs in the active state, results in enhanced susceptibility to E. cruciferarum (Hoefle et al., 2011 ; Huesmann et al., 2011). Accordantly, expression of an inactive (dominant negative) ROP6 (rop6^dn; At4g35020) variant, unable to interact with downstream effectors, results in reduced penetration by Go. This correlates with an increased transcript abundance of SA-responsive genes, such as PR1, and elevated SA-mediated defense responses. However, Go resistance of the rop6^dn transgenic plants is uncoupled from SA signaling (Poraty-Gavra et al., 2013). Nevertheless, these results, together with previous findings in barley (Hoefle et al., 2011; Scheler et al., 2016), suggest a positive role of active ROPs in mediating susceptibility to adapted PMs.

A number of proteins whose deficiency leads to enhanced PM resistance, such as EDR1 to 4, LESION INITIATION 2 (LIN2: At1g03475), and GSL5/PMR4 are associated with the repression of SA-mediated defense. This is indicated by requirement of SA for the increased disease resistance phenotypes of the respective mutants (see section 6; Vorwerk et al., 2007; Zhang et al., 2007; Wawrzynska et al., 2010; Guo et al., 2013; Wu et al., 2015). The MAPKKK EDR1, for example, negatively regulates SA-dependent defense responses and cell death. In consequence, edr1 mutants are constitutively primed for SA-inducible defense which might occur via the regulation of the MAPKs MPK3 (At3g45640) and MPK6 (At2g43790) (Beckers et al., 2009). The role of EDR1 in the control of SA signaling probably relies on its interaction with MKK4 and MKK5, the upstream MAPKKs activating MPK3 and MPK6 (Zhao et al., 2014). EDR1 negatively affects MKK4 and MKK5 levels, presumably resulting in repression of the MKK4/ MKK5-MPK3/MPK6 cascade involved in the induction of SA signaling (Zhao et al., 2014; Wu et al., 2015). Mutations in MKK4, MKK5 or MPK3 (but not in MPK6) suppress the edr1 phenotype, indicating a requirement of these kinases for edr1-mediated PM resistance. The same holds true for edr4, which is in line with the need of EDR4 for the relocation of EDR1 to the PM penetration site (see section 6.1 ; Wu et al., 2015). Strikingly, overexpression of MKK4 or MKK5 causes edr1-like resistance and PM-induced cell death, pointing to an additional role for the MKK4/MKK5-MPK3/MPK6 kinase cascade in SA-induced cell death, parallel to its contribution to SA-regulated gene expression (Zhao et al., 2014).

NPR3-mediated NPR1 degradation at high SA levels promotes the onset of cell death. However, while no PM phenotype has been reported for npr3 (Liu et al., 2005), npr1 and npr4 mutant plants are more susceptible to Gc (Reuber et al., 1998; Liu et al., 2005; Humphry et al., 2010). Thus, it remains elusive to which extent NPR3-mediated cell death contributes to defense against PMs. In addition to NPR3, also NPR4 adds to pathogen-triggered cell death (Pajerowska-Mukhtar et al., 2013; Kumar, 2014; Yan and Dong, 2014). A role for SA-mediated cell death responses in PM defense is supported by correlation of enhanced SA signaling with increased PM-induced cell death in several of the above mentioned resistant mutants (Guo et al., 2013).

Mutants impaired in autophagy further corroborate a role of cell death in PM resistance, as some of them display early leaf senescence and spontaneous cell death, which in several cases coincides with increased PM resistance (Yoshimoto et al., 2009; Wang et al., 2011b; Wang et al., 2011a). Autophagy targets organelles and cytosolic proteins for vacuolar/lysosome-mediated degradation (Liu and Bassham, 2012). The mutant of AUTOPHAGY-RELATED 2 (ATG2: At3g19190), impaired in the early steps of autophagosome biogenesis, exhibits severe PM-induced cell death and increased resistance when challenged with Gc, while susceptibility towards Go is unaltered (Yoshimoto et al., 2009; Wang et al., 2011b; Wang et al., 2011a). PM resistance of atg2 plants depends on SA signaling, while cell death is partially independent of SA. In conclusion, autophagy contributes to suppression of cell death and defense response to PM fungi; however, the mechanisms by which autophagy controls these processes are yet unknown (Wang et al., 2011a; Wang et al., 2011b).

8.2 Contribution of JA and ET signaling to PM resistance

SA-mediated defense appears to act mainly against biotrophic pathogens, while the JA and ET pathways are preferentially linked to resistance against necrotrophic parasites (Thomma et al., 2001 ; Glazebrook, 2005). Induced systemic resistance (ISR) confers JA/ET-mediated protection of shoot tissues via root-to-shoot signaling. ISR is initiated by interactions with beneficial microbes such as arbuscular mycorrhiza or plant growth-promoting rhizobacteria in the root and has proven effective against necrotrophs and herbivores (reviewed in Pieterse et al., 2012). Despite the fact that PM fungi are obligate biotrophs, root colonization with the putative plant growth-promoting basidiomycete Piriformospora indica reduces Go conidiation in a JA signaling-dependent manner (Stein et al., 2008). This finding suggests that besides SA also JA contributes to resistance against PM fungi (Figure 8). Accordingly, Bgh inoculation of Arabidopsis induces expression of genes that are controlled by the JA/ET signaling pathways (Zimmerli et al., 2004). By contrast, although endogenous JA levels are enhanced during the formation of haustoria by Gc, this does not result in transcript accumulation of JA/ET-responsive genes (Reuber et al., 1998; Nishimura et al., 2003; Zimmerli et al., 2004; Glazebrook, 2005; Fabro et al., 2008). Nevertheless, constitutive or ectopic activation of the JA/ET pathway due to elevated JA/ ET levels in the mutant of CELLULOSE SYNTHASE 3/CONSTITUTIVE EXPRESSION OF VSP 1 (CESA3/CEV1: At5g05170) or treatment of Col-0 with methyl-JA enhances resistance against Gc. This effect depends on the JA receptor component CORONATINE-INSENSITIVE PROTEIN 1 (COI1 : At2g39940; Ellis et al., 2002a; Ellis et al., 2002b; Zimmerli et al., 2004). In conclusion, the findings suggest that, although elicitation of JA/ET-mediated defense signaling seems to be restricted to incompatible PM-host interactions, JA/ET-induced defense responses are effective against virulent PM fungi if stimulated constitutively, artificially or systemically, despite the biotrophic nature of the interaction. Consequently, JA/ET signaling must either be suppressed or failed to be elicited by adapted PMs during a successful infection (Figure 8; reviewed in Antico et al., 2012).

Figure 8.

Integration of phytohormone signaling in defense against PM fungi.

Although only incompatible PM-host interactions elicit JA/ET-mediated defense, JA/ET-induced defense responses are effective against virulent PM fungi if stimulated constitutively (cesA/cev1), artificially (JA treatment) or systemically (Piriformospora indica root colonization). These findings suggest that virulent fungi suppress JA/ET signaling during compatible interactions. This suppression might involve the antagonistic action of SA signaling. Solid lines indicate experimentally supported impacts, while dashed lines indicate speculative connections.

9.1 WRKY transcription factors contribute to defense regulation

Among the genes that show altered transcript accumulation in response to Go, members of the plant-specific WRKY (single amino acid letter code for tryptophan-arginine-lysine-tyrosine) TFs represent the most prominent TF family (Chandran et al., 2009). WRKY TFs are key regulators of pathogen-triggered changes in gene expression that act as transcriptional activators or repressors in various homo- and heterodimer combinations. They function up- and downstream of hormone signaling pathways, are involved in the antagonistic control of SA and JA/ET signaling pathways and can be regulated by MAPKs (reviewed in Bakshi and Oelmüller, 2014; Buscaill and Rivas, 2014; Caarls et al., 2015). The involvement of WRKYs in defense of Arabidopsis against PMs is indicated by transcriptional changes of WRKY-encoding genes in response to Go and an enrichment of WRKY-targeted W-box cis-regulatory elements in promoters of genes differentially transcribed upon PM challenge (Chandran et al., 2009). Furthermore, expression of WRKY TFs is enhanced in PM resistant edr1 plants relative to the wild type in response to Gc, and genes whose promoters contain W-boxes are likewise enriched in this dataset. As PM resistance in edr1 plants depends on the MPKK4/5-MPK3/6 cascade, the expression of WRKY TFs might be regulated via this pathway (Christiansen et al., 2011).

In barley, the MLA10 NB-LRR interacts with the Bgh AVR_A10 (avirulence A10) effector and induces transcriptional changes by inhibition of HvWRKY1 and HvWRKY2. Both TFs supposedly act as transcriptional repressors of genes involved in basal defense and effector-triggered immunity (ETI; Shen et al., 2007). Similarly, the closely related Arabidopsis TFs WRKY18 (At4g31800) and WRKY40 (At1g80840), whose transcription is rapidly induced during PM infection, negatively regulate defense against Go. Together these results indicate functional conservation of the defense-repressive role of this WRKY sub-family (Shen et al., 2007). Altered pathogen-induced transcriptional reprogramming in the Go-resistant wrky18 wrky40 double mutant corroborates the negative impact of WRKY18/40 on defense-related gene expression (Pandey et al., 2010; Schön et al., 2013). Chromatin immunoprecipitation (ChIP) experiments revealed binding of WRKY40 to W-box containing promoter regions of EDS1, the AP2 (apetala 2)-type TF gene REDOX RESPONSIVE TRANSCRIPTION FACTOR 1 (RRTF1: At4g34410) and to JASMONATE-ZIM-DOMAIN PROTEIN 8 (JAZ8: At1g30135), a member of the JA-signaling repressor gene family (Pandey et al., 2010). Thus, WRKY18/40 TFs seem to repress the transcription of positive defense regulators such as EDS1, and positively modulate JA-signaling (Pandey et al., 2010; Schön et al., 2013). Although the regulatory role of HvWRKY1/2 and WRKY18/40 is conserved between barley and Arabidopsis, an Arabidopsis R protein interfering with WRKY18/40 function remains to be identified. The conservation of MLA1 functionality and induction of MLA-dependent defense gene expression in response to Bgh might indicate the existence of a respective MLA analog in Arabidopsis (see section 4; Maekawa et al., 2012). In contrast to WRKY18 and WRKY40, WRKY70 (At3g56400) contributes to resistance of Arabidopsis to Gc and inactivation of the respective gene results in increased susceptibility to this pathogen (Li et al., 2006). WRKY70 overexpression coincides with a partially NPR1-dependent suppression of JA responsive genes, indicating a role of this TF in the control of SA/JA crosstalk (Li et al., 2006; Caarls et al., 2015).

9.2 Hormone signaling-induced transcriptional reprogramming during defense

SA signaling contributes to gene expression during the Arabidopsis-PM interaction (see section 8.1). This involves Go-induced expression of genes related to Ca²⁺ signaling and genes coding for redox regulators that contribute to NPR1 activation (Chandran et al., 2009). Transcript accumulation of SA-responsive TFs, SA biosynthesis genes (ISOCHORISMATE SYNTHASE 1 (ICS1)ISID2: At1g74710) and SA-responsive pathogenesis-related genes such as PR1 emphasizes the predominant contribution of SA signaling to PM-induced gene expression. In line with this notion, SA signaling-dependent resistance of the wrky18 wrky40 double mutant correlates with massive Go-induced transcriptional reprogramming (Pandey et al., 2010; Schön et al., 2013). The edr1 mutation, which enhances SA-dependent PM resistance, affects accumulation of defense-related transcripts in response to Gc, including transcripts encoding WRKY and AP2/ET-response element binding factor (ERF) TFs (Christiansen et al., 2011). Furthermore, genes encoding proteins associated with ROS production and the endomembrane system are induced in infected edr1 plants. PM-induced enrichment of the latter together with transcripts associated with secretion suggests that the secretory pathway may play an important role in edr1-mediated immunity (Christiansen et al., 2011). This assumption is in agreement with the relocalization of EDR1 from the ER to the plant-fungal interface during Gc infection (Christiansen et al., 2011; Wu et al., 2015).

TFs of the AP2/ERF family particularly regulate genes related to JA/ET signaling. Besides WRKY TFs, AP2/ERF TFs and transcripts associated with AP2/ERF response elements (GCC-boxes) are over-represented amongst genes upregulated during PM infection in the edr1 mutant (Christiansen et al., 2011). Arabidopsis ERF6 (At4g17490) and ERF104 (At5g61600) are phosphorylated by MPK6 and/or MPK3, indicating a regulation of these ERFs by defense-related MAPK cascades (Bethke et al., 2009; Meng and Zhang, 2013; Tsuda and Somssich, 2015). Furthermore, ERF1 (At3g23240) and ERF2 (At5g47220) transcripts accumulate upon Go infection (Chandran et al., 2009). A role of ERF1 in defense against PMs is in line with the finding that ERF1 overexpression results in enhanced resistance to Go (Gu et al., 2002; Chandran et al., 2009). OCTADECANOID-RESPONSIVE ARABIDOPSIS AP2/ ERF 59 (ORA59: At1g06160), a master regulator of ERF-controlled JA/ET signaling, has been identified, together with other ET/JA-responsive genes, as differentially regulated in SA-deficient mutants versus wild type plants upon Go infection. This suggests that ORA59 modulates the crosstalk between SA and JA/ET signaling during PM-induced defense responses (Chandran et al., 2009). In agreement with this finding, ORA59 was identified as a major target for SA antagonism (Zander et al., 2012; Van der Does et al., 2013; Zander et al., 2014; reviewed in Caarls et al., 2015). A dominant allele of SIGNAL RESPONSIVE1 (SR1: At2g22300), encoding a calmodulin-binding TF that regulates ET-induced senescence by binding to the ETHYLENE INSENSITIVE3 (EIN3: At3g20770) promoter, has been identified as a gain-of-function suppressor of edr2-mediated resistance to Gc (Nie et al., 2012). SR1 binds to the promoter of NON-RACE-SPECIFIC DISEASE RESISTANCE1 (NDR1: At5g06320). NDR1, a membrane-associated protein, contributes to plant immunity mediated by several coiled-coil NB-LRRs and SAR (Century et al., 1997; Shapiro and Zhang, 2001 ; Zhang and Shapiro, 2002). A sr1 null mutant conditions resistance to Gc, while an additional mutation in ndr1 suppresses this phenotype. Together these data suggest that SR1 plays a criticial role in PM resistance, possibly by regulating EIN3 and NDR1 expression (Nie et al., 2012).

Integrative analysis of protein interaction networks and transcriptomics during Go infection revealed a negative correlation between development-related and defense-related genes/proteins (Jiang et al., 2016). While many defense-related genes are induced after Go infection, the majority of genes linked to development are down regulated. Interestingly, auxin-related genes are overrepresented among nodes connecting the defense and development sub-networks. Together these findings emphasize that defense is triggered at the expense of developmental programs and that regulation of this trade-off involves auxin (Jiang et al., 2016).

9.3 Attuned transcriptional regulation coordinates defense

Despite its compatible nature, the Arabidopsis-Gc interaction elicits expression of genes related to NHR (Chandran et al., 2010). Remarkably, in compatible interactions, neither activation of NHR genes nor of SA-induced defense are sufficient to confer resistance (Chandran et al., 2009; Chandran et al., 2010). The P E N1/SNAP33/VAMP722 and the PEN2/PEN3 pathways are important determinants of NHR and are additionally required for mlo2-mediated immunity (see section 4 and 6.3). In line with their pivotal role in defense, PEN1, PEN2, PEN3, SNAP33 and MLO2 share a substantial amount of coexpressed genes, with the majority of transcripts accumulating in response to biotic stresses and MAMP treatment (Humphry et al., 2010). Notably, many transcripts of this regulon accumulate in Bgh-inoculated Arabidopsis plants ectopically expressing the barley MLA1 R protein. This implies a substantial overlap of MLA1-dependent transcriptional regulation and basal resistance against PM fungi (Humphry et al., 2010; Maekawa et al., 2012).

Consistent with the role of PEN2 and PEN3 in indole glucosinolate biosynthesis and secretion, many of the coexpressed genes are associated with the glucosinolate pathway. One example is the gene encoding MYELOBLASTOSIS (MYB)51 TF (At1g18570), a major regulator of defense-related expression of glucosinolate biosynthesis genes (Humphry et al., 2010). Transcriptomic evaluation of the PM resistant wrky18 wrky40 double mutant revealed that WRKY18 and WRKY40 suppress crucial biosynthesis genes of the indolic phytoalexin camalexin (Pandey et al., 2010). Accordingly, increased expression of camalexin and indole glucosinolate biosynthesis genes after pathogen challenge in wrky18 wrky40 plants correlates with the enhanced accumulation of the phytoalexin camalexin and 4MI3G (4-methoxyindol-3-ylmethyl-glucosinolate), an indole glucosinolate intermediate relevant for PM resistance, in this mutant (Pandey et al., 2010; Schön et al., 2013). Loss of function of the crucial 4MI3G biosynthesis gene CYP81F2 suppresses wrky18 wrky40-mediated inhibition of host cell entry, indicating that the indolic metabolite is required for penetration resistance of the double mutant against Go (Schön et al., 2013).

The group of genes coexpresed with PEN3 further showed a significant overrepresentation of components involved in Ca²⁺ signaling such as calcium/calmodulin-dependent protein kinases (CCaMKs; Humphry et al., 2010; Campe et al., 2015). Consistently, components of the Ca²⁺ homeostasis and signaling machinery such as CAM9/CALMODULIN-LIKE (CML)9 (At3g51920) and calreticulin-encoding genes are rapidly induced after Go inoculation, and CML38 (At1g76650) is constitutively upregulated in the highly resistant wrky18 wrky40 mutant (Chandran et al, 2009; Chandran et al., 2010, Pandey et al., 2010).

In conclusion, coexpression of genes encoding proteins involved in secondary metabolite biosynthesis (such as PEN2 or cytochrome P450s like CYP83B1 (At4g31500)) together with genes whose products mediate exocytosis/extrusion (SNAREs, exocyst subunits and ABC transporters like PEN3) suggests that production and secretion of antimicrobial compounds is transcriptionally attuned. Additional coregulation of receptor-like kinase genes, transcripts of Ca²⁺ signaling components and the heterotrimeric G-protein β and g subunits AGB1 (At4g34460) and AGG1 (At3g63420) indicates that also recognition and signaling components are coexpressed with key components of antifungal defense (Humphry et al., 2010; Lorek et al., 2013). Enhanced host cell entry by Go and E. pisi of Gβ-deficient mutants emphasizes the importance of second messenger signaling via heterotrimeric G-protein components for PM resistance (Lorek et al., 2013). Finally, identification of defense-related TFs in the regulon indicates that recognition of microbes, response initiation, and defense execution are transcriptionally coordinated to enable an efficient immune output.

9.4 Host transcriptional changes indicate an adaption to the accommodation of the biotrophic pathogen

Besides defense-related transcriptional changes induced by microbe recognition, host gene expression is potentially impacted by the action of PM effectors to promote fungal accommodation. Consequently, adaptation of the host metabolism to the presence of the pathogen has been reported (reviewed in Wildermuth, 2010). Laser microdissection-assisted site-specific profiling of transcript abundance during late Go infection stages (5 dpi) suggests a suppression of photosynthesis and points to a carbon source-to-sink transition in PM-infected cells (Figure 9; Chandran et al., 2009; Chandran et al., 2010). The PM-triggered induction of genes associated with sugar metabolism and hexose transporters associated with sink organs further reinforces this idea. The respective proteins might contribute to the availability of carbohydrates at infection sites, and consistently their elevated expression levels may reflect an increased demand for hexoses by the fungus (Fabro et al., 2008; Chandran et al., 2009; Chandran et al., 2010). Increased transcript abundance of genes related to respiration, including glycolysis, the tricarboxylic acid cycle, and the mitochondrial electron transport chain further strengthen the notion of an adaptation to the elevated energy consumption by the infected tissue (Fabro et al., 2008; Chandran et al., 2010).

Adjustment of the plant host metabolism to support the growth of the biotrophic pathogen is consistent with an increased ploidy level of the mesophyll cells underlying infected epidermal cells at later stages of compatible interactions (Figure 9; Chandran et al., 2010; Chandran et al., 2013). This correlates with the accumulation of PLANT UBX DOMAIN-CONTAINING PROTEIN 2 (PUX2: At2g01650) transcripts after Go infection (Chandran et al., 2009). Strikingly, the onset of PUX2 induction at 5 dpi overlaps with the occurrence of endoreduplication in mesophyll cells, and corresponds with fungal growth and reproduction (Chandran et al., 2013). The resulting polyploidy might compensate for the increased metabolic activity resulting from the nutritional demands of the fungus. This is supported by decreased spore formation co-inciding with reduced basal ploidy in pux2, and thus identifies endoreduplication as a potential determinant of susceptibility to PM (Chandran et al., 2010; Chandran et al., 2013). The presence of UBX (ubiquitin regulatory X) and PUB (peptide:N-glycanase/UBA or UBX-containing proteins) domains in PUX2 suggests that, like other proteins with similar domain structures, it might act as a regulatory cofactor of CELL DIVISION CONTROL PROTEIN 48 (CDC48: At3g09840). Indeed, this AAA-ATPase (ATPase associated with diverse cellular activities) interacts with PUX2 in vitro (Rancour et al., 2004). As CDC48 complexes contribute to cell cycle progression, its interaction with PUX2 might regulate cell ploidy (Rancour et al., 2004; Madsen et al., 2009; Yamanaka et al., 2012; Gallois et al., 2013). MYB3R4 (At5g11510), a cell cycle control-associated MYB3R TF activating G2/M progression, is locally induced 5 dpi with Go. As genome duplication is a controlled process that occurs during mitosis, it is conceivable that MYB3R4 is required for PM-induced polyploidy, which is supported by the phenotype of the myb3r4 mutant (Haga et al., 2007; Chandran et al., 2010; Chandran et al., 2013). Similar to pux2, myb3r4 mutants exhibit reduced PM conidiophore formation. The negative impact on basal cell ploidy levels and increased resistance of pux2 is further phenocopied by pmr6. By contrast, reduced fungal reproduction associated with pmr5 does not impact basal ploidy levels but correlates with a suppression of the PM-induced increase in ploidy (Chandran et al., 2013). Analysis of pmr5 microarray data reveals an enrichment of MYB3R TF binding elements among cell cycle regulation-related genes showing altered expression in the mutant. This suggests that PMR5 acts upstream of a MYB3R TF to control PM-induced ploidy (Chandran et al., 2013).

Figure 9.

Proposed model of endoreduplication in PM pathogenesis.

The scheme depicts regulators and mechanisms involved in the control of endoreduplication and consequences of PM-induced mesophyll polyploidy. A PM (grey)-colonized leaf epidermal cell and an underlying mesophyll cell are shown. Grey arrows indicate the proposed translocation of components between cells. Solid lines indicate proven regulatory impacts and dashed lines indicate speculative regulatory impacts. M = mitosis.

A critical role of elevated ploidy for fungal virulence is further strengthened by the identification of the TEOSINTE BRANCHED/ CYCLOIDEA/PROLIFERATING CELL FACTOR (TCP)13 (At3g02150), TCP14 (At3g47620) and TCP15 (At1g69690) basic helix-loop-helix (bHLH) TFs as common targets of several PM, oomycete and bacterial effectors (Weßling et al., 2014). TCP14 and TCP15 repress endoreduplication by directly regulating the expression of cell-cycle genes (Peng et al., 2015). Mutation of TCP13, TCP14, and to a lesser extent of TCP15, results in increased susceptibility towards Go (Weßling et al., 2014). A link to ubiquitin-mediated regulation of ploidy is provided by the ubiquitin receptors DA1 (At1g19270; “Dà” is Chinese for “large”), DA-RELATE D (DAR)1, and DAR2 (At2g39830), which interact with and modulate the stability of TCP14/15 to regulate endoreduplication (Peng et al., 2015). Remarkably, DEL1, known to repress genes required for the onset of endoreduplication (Vlieghe et al., 2005; Lammens et al., 2008), does not impact the Go-induced increase in mesophyll ploidy when mutated or overexpressed (Chandran et al., 2014). Instead, microarray analyses of del1 plants revealed an induction of basal defense gene expression compared to wild type (see section 8.1). The identification of effector targets involved in adaptation of the host metabolism (Weßling et al., 2014) marks one of the first steps towards elucidation of pathogen-induced reprogramming of the plant transcriptome. Further characterization of the mechanisms by which the fungus enforces adjustment of the plant cellular program to promote its intracellular accommodation will be an important aspect of future research.