Epigenetic regulation of development and pathogenesis in fungal plant pathogens

Summary

Evidently, epigenetics is at forefront in explaining the mechanisms underlying the success of human pathogens and in the identification of pathogen‐induced modifications within host plants. However, there is a lack of studies highlighting the role of epigenetics in the modulation of the growth and pathogenicity of fungal plant pathogens. In this review, we attempt to highlight and discuss the role of epigenetics in the regulation of the growth and pathogenicity of fungal phytopathogens using Magnaporthe oryzae, a devastating fungal plant pathogen, as a model system. With the perspective of wide application in the understanding of the development, pathogenesis and control of other fungal pathogens, we attempt to provide a synthesized view of the epigenetic studies conducted on M. oryzae to date. First, we discuss the mechanisms of epigenetic modifications in M. oryzae and their impact on fungal development and pathogenicity. Second, we highlight the unexplored epigenetic mechanisms and areas of research that should be considered in the near future to construct a holistic view of epigenetic functioning in M. oryzae and other fungal plant pathogens. Importantly, the development of a complete understanding of the modulation of epigenetic regulation in fungal pathogens can help in the identification of target points to combat fungal pathogenesis.

Introduction

In the early 1940s, Conrad Waddington coined the term ‘epigenetics’, which was followed by an unprecedented growth in research. In the past, the term ‘epigenetics’ has been used to describe the process through which phenotypes arise from genotypes during development. In modern science, ‘epigenetics’ is defined as follows: ‘molecular processes around DNA that regulate genome activity independent of DNA sequence and are mitotically stable’ (Skinner et al., 2010). In the main, evolution and phenotypic variability is explained on the basis of underlying genetic changes in DNA sequences (Laland et al., 2011). However, the low frequency of beneficial mutations fails to explain the large phenotypic variability (Laland et al., 2011; Nei, 1988; Nei and Nozawa, 2011). Interestingly, the decoupled frequencies of genotypic mutations and phenotypic variations (Jablonka and Raz, 2009; Pigliucci, 2007; Skinner, 2015) can be explained on the basis of epigenetics. Instances of the loss or retention of epigenetic modifications in subsequent generations can throw light onto the trend of epigenetic evolution of any organism and explain the phenotypic variability (Billmyre et al., 2013; Bird, 2002; Hu et al., 2013; Ikeda et al., 2013; Iyer et al., 2011; Nakayashiki et al., 2001). In addition, the epigenetic processes influenced by environmental factors (Skinner, 2002) can affect the molecular basis of inheritance, indicating a plausible combinatorial effect of epigenetics and genetic variation in evolution.

Hitherto, epigenetics has strongly contributed to the understanding of the epigenetic basis of human diseases (such as cancer development and control, toxic pathology, cardiovascular and neurogenetic diseases, and pharmaco‐epigenomics) (Costa et al., 2009; Herceg et al., 2013; Liang et al., 2013) and the virulence of diverse human pathogens (Bierne et al., 2012; Silmon de Monerri and Kim, 2014). In plants, epigenetics is being widely used to understand pathogen‐induced modifications in the host that eventually affect plant immunity (Ding and Wang, 2015; Kazan and Lyons, 2014). In fungi, non‐pathogenic models, such as the budding yeast Saccharomyces cerevisiae, fission yeast Schizosaccharomyces pombe and filamentous fungus Neurospora crassa, have been extensively used to unravel epigenetic regulation (Allshire and Ekwall, 2015; Aramayo and Selker, 2013; Grunstein and Gasser, 2013). Studies on the aforementioned models have revealed three prevalent epigenetic mechanisms of regulation in fungi: DNA methylation, histone modifications and the RNA silencing system. Interestingly, there are reports of the absence or presence of alternative mechanisms of epigenetic regulation in fungal models. For instance, N. crassa exhibits RNA interference (RNAi)‐based silencing systems, such as MSUD (meiotic silencing by unpaired DNA) and quelling (homology‐dependent gene silencing mechanism) (Aramayo and Selker, 2013), which have not been reported in the case of S. cerevisiae. Likewise, although heterochromatin assembly via small RNAs (sRNAs) has been studied extensively in Sc. pombe (Allshire and Ekwall, 2015), the role of RNAi in heterochromatin formation could not be found in N. crassa. The diversity of epigenetic mechanisms in fungi shows the intriguing aspect of epigenetic regulation that can be undertaken to identify similar, alternative or divergent mechanisms of regulation. However, despite the advances in epigenetic studies of non‐phytopathogenic fungi, an understanding of the role of epigenetics in the modulation of plant–fungal pathogenesis is still in its infancy (Fig. 1). The evident disparity in the available knowledge between non‐pathogenic and phytopathogenic fungal models invokes the need for a deeper understanding of the epigenetics of fungal pathogenesis.

Figure 1.

Discrepancy in the number of publications for epigenetics in general, epigenetics of microbes and epigenetics of fungal plant pathogens. The number of publications in the field of epigenetics was retrieved from PubMed using ‘epigenetic*’ as the search term. The number of publications in microbes was retrieved using ‘epigenetic* and (pathogen* or microbe* or fungal or fungi or bacteria* or virus)’. The number of publications in fungal plant pathogens was retrieved using the search term, ‘epigenetic* and (fungi* or fungal*) and plant and pathogen’ (only 21 hits).

Considering the above knowledge gap, we aim to shed light on the epigenetic basis of development and pathogenesis of fungal plant pathogens using Magnaporthe oryzae, a fungal pathogen causing rice blast disease, as a model system. Ranked as the most devastating fungal crop pathogen, M. oryzae [formerly Magnaporthe grisea; syn. Pyricularia oryzae; please refer to Couch and Kohn (2002) and Zhang et al. (2011, 2016) for an ongoing discussion on the nomenclature of M. oryzae] is responsible for huge economic losses worldwide. With estimated losses of 30% of the total rice harvest, M. oryzae poses a major threat to global food security (Dean et al., 2012; Maciel et al., 2014; Martin‐Urdiroz et al., 2015; Wilson and Talbot, 2009). More importantly, the spread and diversification of the fungal pathogen in a diverse host range, from rice to wheat, barley, oats, rye grass and millet, are creating serious problems in various parts of the world (Maciel et al., 2014). However, it is not only the alarming magnitude of loss caused by M. oryzae which underlines the importance of studying this pathogen, but the availability of genomic sequences of both the pathogen and the host (rice), which makes it an experimentally tractable and ideal model system to obtain an understanding of plant–fungus interactions. At several instances, epigenetics has been shown to play an important role in the success of plant pathogen persistence in diverse host niches (Angers et al., 2010; Bierne et al., 2012; Castonguay and Angers, 2012). Although relatively scarce in the case of M. oryzae, epigenetic studies have started to reveal crucial information related to pathogenicity and other infection‐related processes. Given the importance of M. oryzae as a fungal plant pathogen and the potential use of epigenetics, the understanding of M. oryzae‐associated plant pathogenicity may be instrumental in: (i) delineating the epigenetic regulation of pathogenicity in other related fungal pathogens; (ii) identifying the primary epigenetic changes that can potentially act as markers of pathogenicity for several other fungal pathogens; (iii) facilitating the understanding of host–pathogen interaction as the genomic sequences of both the fungus (M. oryzae) and its primary host (rice) are available, allowing the genetic and experimental tractability of this pathosystem; and (iv) allowing the development of prevention and control strategies for fungal pathogens.

In this review, we aim to emphasize the role of epigenetics in the regulation of fungal development and pathogenesis (with M. oryzae as a model) with further elaboration of undiscovered aspects. Thus, first, we discuss the prevalent epigenetic mechanisms in M. oryzae, including mechanisms that have not yet been reported (MSUD, quelling, sRNA‐directed heterochromatin formation) in M. oryzae and other fungal plant pathogens. Brief discussions of the undiscovered mechanisms aim to result in the prospective discovery of similar or alternative mechanisms of epigenetic regulation in the near future. Second, we also identify several other areas of epigenetic regulation that are unexplored and can potentially aid in the understanding of the persistence, spread and success of M. oryzae or other fungal pathogens. Overall, we bring into focus the evident gap in the research of epigenetics of fungal plant pathogens and highlight the epigenetic modifications which can be used as targets to combat fungal plant diseases in general.

Mechanisms of Epigenetic Modifications in M. Oryzae

The importance of epigenetic mechanisms in eukaryotes cannot be denied as they underpin the nature of gene expression and development. Factors that mediate epigenetic phenomena include sRNA, DNA methylation and histone modifications (Agger et al., 2008; Fu and He, 2012; Rothbart and Strahl, 2014). A growing body of recent evidence (Aramayo and Selker, 2013; Jeon et al., 2015; Pham et al., 2015; Soyer et al., 2014; Woloshuk and Shim, 2013) has implicated the involvement of such epigenetic factors in diverse aspects of fungal biology, including pathogenicity. Here, we discuss the chief epigenetic mechanisms reported in M. oryzae, further interpreting their role in the understanding of evolution, fungal development and pathogenicity, with an aim to relate them to other fungal plant pathogens. Furthermore, we also briefly discuss the epigenetic mechanisms that have not yet been reported in M. oryzae or other fungal pathogens, but are open to further research in understanding the epigenetic regulation of fungal plant pathogens. The noted epigenetic modifications in M. oryzae and other fungal pathogen species are as follows: (i) DNA methylation; (ii) modifications of the histone protein; and (iii) sRNA.

DNA methylation

DNA methylation is an epigenetic mechanism that controls gene expression via the chemical modification of DNA. By altering the affinity of DNA‐binding proteins (transcriptional machinery) to DNA, or recruiting proteins involved in gene repression, DNA methylation regulates the expression of target genes. DNA methylation is known to underlie biological processes, such as genetic imprinting, X‐chromosome inactivation, cell differentiation and gene silencing (Bond and Baulcombe, 2015; Reik and Lewis, 2005). Although prokaryotes exhibit adenine methylation (N6‐methyladenine, 6mA), eukaryotes, including M. oryzae, undergo cytosine methylation (C5‐methyl‐cytosine, 5mC). DNA methyltransferases (DNMTs) are special enzymes that catalyse the methylation of cytosine bases in DNA (Fig. 2). Amongst the two categories of DNMTs, DNMT1 enzymes are involved in the maintenance of the methylated state and DNMT3, the de novo DNMT, is involved in the development of the methylation pattern. Notably, DNA methylation is preferentially associated with transposons and silenced DNA (Dang et al., 2013; Jeon et al., 2015; Montanini et al., 2014), and the methylation pattern created by these enzymes is dynamic and subject to reversibility on environmental and physiological changes. Phylogenetic analysis of the DNMT1 and DNMT2/DNMT3 groups of enzymes revealed an independent origin of the two groups, which could be traced back to different bacterial precursors (Iyer et al., 2011; Jurkowski and Jeltsch, 2011). However, the phylogenetic relationship of DNMT2 and DNMT3 in the evolutionary tree is still unclear.

Figure 2.

A summarized representation of the epigenetic research in Magnaporthe oryzae performed to date. Small red circle, methyl group; green polygon, acetyl group. Orange boxes (black border) represent the known epigenetic mechanisms in M. oryzae. Green boxes indicate research which has been reported successfully in M. oryzae. Violet boxes represent the research areas which have been delineated in other fungal pathogens, but not in M. oryzae, and hold the potential to unravel the epigenetic regulation of M. oryzae. Full arrows represent the established associations, whereas dotted arrows represent possible relationships between two components. Question marks indicate interesting areas that should be examined in the field of M. oryzae epigenetics. In the red boxes, we highlight the conclusive research trajectory that could be undertaken to understand the epigenetic regulation of fungal plant pathogens. DNMT, DNA methyltransferase; esRNA, endogenous short RNA; H3K9me3, trimethylation of lysine residue9 in histone3; H3 and H4, histone3 and histone4; HAT, histone acetyltransferase; HDAC, histone deacetylase; HDM, histone demethylase; HMT, histone methyltransferase; mC, cytosine methylation; me1, me2, me3, degree of methylation (mono, di and tri); MAGGY, Magnaporthe gypsy‐like element; sRNA, small RNA; TE, transposable element; TFs, transcription factors.

The significance of DNA methylation in governing the pathogenicity of M. oryzae has not been proven to date (Ikeda et al., 2013; Jeon et al., 2015; Nakayashiki et al., 2001). Although the DNA methylation status of a long terminal repeat retrotransposon (LTR), Magnaporthe gypsy‐like element (MAGGY), differs among various Magnaporthe grisea subgroups (fungal isolates derived from rice, common millets, foxtail millet and torpedo grass) (Nakayashiki et al., 1999a), its inactivation could not be attributed to DNA methylation in M. grisea (Nakayashiki et al., 2001). Analyses of the F1 progeny derived through the cross of methylation‐proficient (Br‐48) and methylation‐deficient (GFSI1‐7‐2) isolates revealed the role of MoDMT1 (DNMT), an orthologue of N. crassa Dim‐2, in the methylation of MAGGY elements (Ikeda et al., 2013). However, phenotypic analyses of the MoDMT1 mutant revealed no significant role of MoDMT1 in the regulation of the development and pathogenicity of M. oryzae (Ikeda et al., 2013). Furthermore, a trend of a loss in DNA methylation was observed in 60 field isolates of M. oryzae. The rate of transposition and the average copy number of MAGGY elements in the genome of methylation‐deficient versus methylation‐proficient isolates (five methylation‐proficient and methylation‐deficient isolates) did not differ significantly, indicating that DNA methylation via MoDMT1 is not crucial in the silencing of the MAGGY elements (Ikeda et al., 2013). However, retention of the entire MoDMT1 open reading frame (ORF) despite a high rate of mutations (Ikeda et al., 2013) indicates that DNA methylation may not be completely indispensable. In addition, the existing effect of DNA methylation at the transcriptional level (Ikeda et al., 2013) indicates the possible involvement of a downstream process which functions in combination with MoDMT1 to silence MAGGY elements.

On the other hand, genome‐wide DNA methylation analysis of M. oryzae at different growth stages revealed several methylation sites in the genome (Jeon et al., 2015). Bisulfite sequencing of genomic DNA extracted from mycelia, conidia and appressoria of M. oryzae revealed several methylation sites in the genome with an average methylation level of individual mC sites between 20% and 30% across samples and chromosomes. Interestingly, the mC sites in mycelia were clustered into densely methylated domains around transposable element (TE)‐rich and gene‐poor regions, indicating that the TE regions are targeted for methylation in the fungi. Furthermore, compared with mycelia, methylated domains were found to occur to a relatively lesser extent in conidial and appressorial samples (Jeon et al., 2015). The aforementioned findings indicate the dynamic nature of DNA methylation in the pathogen, which is dependent on the fungal isolate or strain type, growth conditions and stage, such as mycelia, conidia and appressoria. An exhaustive analysis of the fungal methylome across different isolates and life cycle stages could be useful to determine the evolutionary role of DNA methylation in fungal pathogenicity with applications in the design of molecular tools to combat fungal infections.

Relics of repeat‐induced point (RIP) mutation

RIP mutation (Selker, 1990; Selker et al., 1987), a mechanism of gene silencing that belongs to the homology‐based genome defence system, is closely associated with DNA methylation (Aramayo and Selker, 2013). Here, repeat or duplicate sequences in the genome, whether arising from native or foreign sources, are subjected to transition mutations (G:C to A:T) in the haploid or sexual phase of the fungal life cycle (Galagan and Selker, 2004). Although indirectly, repeated sequences have been shown to affect the DNA methylation status of the genome via RIP (Freitag et al., 2002; Singer et al., 1995). Neurospora crassa has been extensively studied in relation to RIP, and its genome‐wide analysis reveals that most of the methylated regions are relics of transposons, which are inactivated via RIP (Galagan and Selker, 2004; Selker et al., 2003). Nakayashiki et al. (1999b) demonstrated that the MAGGY transposons in M. grisea isolates were silenced via a RIP‐like process. Canonically, RIP occurs preferentially at 5′—C—phosphate—G—3′ (CpA) dinucleotides, and RIP'd sequences are mostly associated with cytosine methylation during the sexual stages of the fungal life cycle. However, the sexual stage in M. grisea has been reported only under laboratory conditions (Ikeda et al., 2002). Thus, the existence of a RIP‐like mechanism in M. grisea was investigated by the introduction of MAGGY retrotransposons and hygromycin B phosphotransferase into the fungus, followed by crossing. Analysis of the F1 progeny revealed frequent transition alternations (A/Tp)Cp(A/T) in the sequence after crossing (Ikeda et al., 2002), suggesting the existence of RIP‐like mechanisms in M. grisea. In addition, a survey of the a novel Ty3/Gypsy retrotransposon, Pyret in six field isolates of M. grisea revealed the existence of RIP‐like transitions, which could be correlated with their fertility (Ikeda et al., 2002), pointing to the possible existence of a sexual cycle in the fungus under natural conditions. The identification of RIP or a RIP‐like process in Aspergillus fumigatus, Fusarium graminearum, Leptospharia maculans and M. grisea (Cuomo et al., 2007; Galagan et al., 2005; Ikeda et al., 2002; Rouxel et al., 2011), and its indirect involvement in DNA methylation, suggests its potential role in DNA methylation, which should be uncovered to understand the de novo methylation or maintenance of DNA methylation in fungal phytopathogens.

Histone modifications and variants

Post‐translational modifications (PTMs) involve modification of the highly conserved core protein, called ‘histone’. Histones are basic proteins (H1, H2A, H2B, H3, H4), which form an octameric complex, over which DNA winds to form the nucleosome (Fig. 2). Repeated nucleosome units package eukaryotic DNA into higher order chromatin fibres. Epigenetic marks created on the histone proteins via the covalent addition of different chemical groups to certain residues of the histone constitute the epigenetic code. Singly, or in combination, the epigenetic code forms a histone language, which is further interpreted at the cellular level via the proteins (Rothbart and Strahl, 2014; Strahl and Allis, 2000). The noted PTMs of histones in M. oryzae include: (i) methylation; and (ii) acetylation.

Histone methylation and demethylation

The methylation and acetylation level of the histones is responsible for the compact packing of the chromatin structure and regulation of the availability of ‘unwound’ chromatin to the transcriptional machinery for further gene expression (Fig. 2). Methylation occurs at different lysine and arginine residues of histone proteins to create an epigenetic code, which is read differently by the cell (Rothbart and Strahl, 2014). The methylation pattern thus created on the histone proteins regulates the epigenetic landscape of a given cell.

Specific lysine residues are methylated in the N‐terminal tail of histone H3 and H4. For instance, lysine residues K4, K9, K27 and K36, lying on the N‐terminal tail, and K79, within the histone core H3, together with K20 within the H4 tail, are usually methylated. The degree and specificity of lysine methylations [mono (me1), di (me2) or tri (me3)] are responsible for transcriptional activation and repression of the genes (Miao and Natarajan, 2005; Sims et al., 2003). Generally, trimethylation of the H3K4 residue is associated with gene activation, whereas trimethylation of the H3K9 residue represents the repressed state of chromatin (Bok et al., 2009; Briggs et al., 2001; Pham et al., 2015). In addition to the methylation of lysine residues, arginine residues are also subjected to methylation at H3R2, H3R8, H3R17, H3R26 and H4R3 sites. Frequently, the methylated arginine residues are located near other modified histone residues (H3K4me3, H3K9, H3K27 and DNA methylation), indicating plausible crosstalk between them (Migliori et al., 2010).

Histone methyl transferases (HMTs) are a group of enzymes responsible for the methylation of lysine and arginine residues on histone proteins. With the exception of one [the Dot1 family of histone lysine methyltransferases (KMTs)], all HMTs belong to the SET (Su(var)3‐9, enhancer of zeste [E(Z)] and Trithorax) domain of the protein family, found to occur in both prokaryotes and eukaryotes (Alvarez‐Venegas, 2014). Historically, histone methylation was considered to be irreversible (Byvoet et al., 1972; Duerre and Lee, 1974); however, the discovery of a histone demethylase (HDM), lysine histone demethylase (LSD1), demonstrated the reversibility of the methylation process (Shi et al., 2004). LSD1 has been shown to specifically demethylate H3K4 via an oxidative reaction, which is ascribed to active transcription. The association of LSD1 with CoREST (restin corepressor) and other proteins creates an enzymatic complex that is responsible for lysine demethylation. It is noteworthy that LSD1 is highly substrate specific and can distinguish H3 peptides with the same methylation on different lysine residues (Shi et al., 2004). The discovery of LSD1 and PADI4 (petidylarginine deiminase 4) led to the identification of other classes of protein with demethylase activity, named Jumonji C (JmjC) domain‐harbouring enzymes, which catalyse the lysine demethylation of histones through an oxidative reaction that requires iron [Fe(II)] and α‐ketoglutarate (αKG) as cofactors (Tsukada et al., 2006; Yamane et al., 2006). The LSD1‐CoREST (lysine‐specific demethylase 1; REST corepressor) model has been successfully used to understand the regulation of chromatin and nucleosomal accessibility to enzymatic complexes (Pilotto et al., 2015). Such studies are useful in delineating the binding interactions of nucleosome tails, chromatin modifiers, transcription factors and DNA interaction sites, which can further help to decode the epigenetic codes that are read differently under different combinations of interactions.

In M. oryzae, histone methylation has been associated with gene activation as well as repression. A study reported the substrate‐induced activation of a cellulase gene, mediated through histone methylation, in M. oryzae (Vu et al., 2013). The cellulase gene MoCel7C is up‐regulated on contact with the host leaf surface, demonstrating the recognition ability of the pathogen on contact with the host surface. MoSET1, an orthologue of SET1 found in S. cerevisiae, encodes histone KMTs, and is responsible for H3K4 methylation in M. oryzae at the MoCel7C locus (Fig. 2). In the presence of exogenous cellulosic substrate [2% carboxymethylcellulose (CMC)], H3K4 methylation is triggered at the MoCel7C locus with further activation of the MoCel7C gene. In line with the role of methylation in gene activation, it was observed that MoCel7C gene activation was drastically reduced in the deletion mutant of MoSET1 in M. oryzae. Interestingly, in the absence of substrate CMC, the expression of MoCel7C was increased in the MoSET1 deletion mutant, thus implying a role for H3K4 methylation in gene suppression under non‐inducing conditions. The antagonistic action (both activation and suppression) of histone methylation can be explained by chromatin remodelling, which renders the promoter accessible or inaccessible to the transcriptional machinery under different environmental conditions. Likewise, another finding successfully demonstrated the role of MoSET1 and other KMTs (MoKMT1, MoSET1, MoKMT3, MoKMT6 and MoKMT2H) in the promotion of the pathogenicity of M. oryzae (Pham et al., 2015) (Fig. 2). Disruption of MoSET1, which is essential for the infection process, severely impaired conidiation and appressorium formation in M. oryzae, rendering the strain ineffective in causing disease in the host plant. The study by Pham et al. (2015) demonstrated that methylation of H3K4me is an important epigenetic mark for infection‐related gene expression in M. oryzae. Furthermore, the above‐mentioned study supports the association of H3K4 methylation directly with gene activation and indirectly with gene repression. Through RNA‐sequencing (RNA‐seq) and chromatin immunoprecipitation‐sequencing (ChIP‐seq) analysis, Pham et al. (2015) observed that MoSET1 plays a direct role in gene activation, whereas gene repression is mediated through indirect effects in M. oryzae. This is in accordance with the finding of Vu et al. (2013) where, again, MoSET1 was involved in gene suppression under non‐inducing conditions. The switch in the activity of histone methylation from activation to repression under different environmental conditions supports the link between external factors and the dynamics of histone methylation. Conclusively, it will be advantageous to identify the external factors/conditions that work in association with HDM and HMT to regulate the pathogenicity of fungal pathogens. In addition, an understanding of the molecular pathway of the epigenetic process could be exploited to tackle notorious fungal plant pathogens.

Histone acetylation and deacetylation

Histone acetylation is one of the best‐characterized dynamic modifications regulated by histone acetyltransferases (HATs) and histone deacetyltransferases (HDACs) (Fuks et al., 2000). Protein complexes composed of HATs and acetyl‐CoA cofactor acetylate multiple lysine residues, thereby triggering active transcription (Roth et al., 2001). However, HDACs antagonize the action of HATs by removing acetylation from the amino group of lysine residues. This process restores the positive charge of lysine, thereby causing the underlying DNA sequences to coil and wind up tightly to the histone and become inaccessible (Bannister and Kouzarides, 2011). This is in line with the observation of HDACs functioning as transcriptional repressors. HATS and HDACs can be categorized into several classes, and readers are referred to a review dedicated to the classification of HATs and HDACs (Jeon et al., 2014). In addition, there are hidden Markov model‐based online webtools available for the grouping of histone modifying enzymes (HMEs) based on their domain structure and other criteria, e.g. ‘dbHiMo: a web‐based genomics platform for histone‐modifying enzymes’; this can be used in the classification of HMEs (Choi et al., 2015).

In addition to HMTs and HDMs, HDACs have also been shown to have an impact on the pathogenicity of M. oryzae. Treatment of M. oryzae with HDAC inhibitors of the Rpd3/Hda1 family of lysine deacetylases and trichostatin A resulted in the inhibition of appressorium formation and decreased pathogenicity, respectively, indicating the significant role of HDACs in the asexual differentiation of M. oryzae (Izawa et al., 2009). Homologues (in yeast and mammalian cells) of the transducin β‐like gene TIG1 found in M. oryzae have been identified as part of the conserved histone deacetylase (HDAC) transcriptional corepressor complex (Fig. 2). A study showed that TIG1 is essential for the pathogenicity and conidiogenesis of M. oryzae (Ding et al., 2010). Deletion of TIG1 rendered the mutant hypersensitive to oxidative stress, which was incapable of forming invasive hyphae. Further, the purified protein complexes associated with TIG1 resulted in the identification of two HDACs, which were homologous to parts of the Set3 complex in yeast. This study highlighted the importance of TIG1, a component of the HDAC complex, in the formation of infection structures of M. oryzae, and the importance of chromatin modification during plant infection. Recently, the role of HDAC has been demonstrated in the phototropic induction of asexual development and pathogenesis in M. oryzae (Deng et al., 2015). Twilight (TWL), a circadian clock‐regulated protein in M. oryzae, shuttles between the cytosol and nucleus during dark and light cycles, and peaks at twilight (Deng et al., 2015). The translocation is mediated via transcription factor Tbf5, probably in association with GCN5 (HAT), functioning downstream of the TWL–HDAC transcription relay. The expression of TWL is essential for asexual development and pathogenesis, and facilitates pathogen attack on the host system. Collectively, the above‐mentioned studies indicate the direct correlation of histone acetylation and deacetylation with the pathogenicity of the fungal pathogen, thereby emphasizing the need for the investigation of the phenomena in relation to pathogen physiology, adaptation and success.

sRNA

sRNA is another epigenetic mechanism which functions at the post‐transcriptional level, thereby regulating the cellular and developmental processes of biological systems. Gene silencing regulated via sRNA is also referred to as post‐transcriptional gene silencing (PTGS) and functions in association with the components of RNAi, such as Dicer (double‐strand RNA endonuclease), Argonaute (sRNA‐binding protein) and RNA‐dependent RNA polymerases (RdRPs) (Nicolas and Ruiz‐Vazquez, 2013). Dicer cleaves the double‐stranded RNA (dsRNA) into 20–30‐bp‐long small interfering RNAs (siRNAs), which are loaded onto the RNA‐induced silencing complex (RISC), which triggers the degradation or translational inhibition of the target mRNA. Most of the sRNA‐mediated epigenetic modifications have been reported in non‐pathogenic fungi, and it is beyond the scope of this review to discuss in detail the epigenetic studies in other fungal model systems [reviewed in Aramayo and Selker (2013), Billmyre et al. (2013), Buhler et al. (2007), Djupedal and Ekwall (2009) and Volpe and Martienssen (2011)].Therefore, we briefly summarize the known sRNA‐mediated epigenetic regulation systems in other fungal models and further discuss them in relation to M. oryzae. Generally, in fungal systems, three kinds of RNAi‐dependent gene silencing mechanism are known: (i) heterochromatin formation; (ii) quelling; and (iii) meiotic silencing by unpaired DNA (MSUD).

Heterochromatin formation

Heterochromatin is a large part of the genome that is packed tightly into highly condensed regions and is important for the maintenance of genomic integrity and the silencing of genes within or close to the heterochromatic regions (position effect variegation and spreading of silencing) (Aramayo and Selker, 2013; Ebert et al., 2006). Together with the factors associated with RNAi, sRNA targets homologous DNA sequences and modifies chromatin to aid heterochromatin assembly and silence the target genes (Djupedal and Ekwall, 2009). Amongst the fungal models, heterochromatin assembly via sRNAs has been extensively studied in Sc. pombe (Aramayo and Selker, 2013; Billmyre et al., 2013; Volpe et al., 2002), where it has been shown that sRNA, together with the components of RNAi, is essential for heterochromatin formation at centromeres (Reinhart and Bartel, 2002; Volpe et al., 2002). Studies have demonstrated the involvement of H3K9me in the triggering of gene silencing through the recruitment of Heterochromatin Protein1 (HP1) or yeast homologue SWI6 (Bannister et al., 2001; Chicas et al., 2004; Nakayama et al., 2001). However, S. cerevisiae lacks the classical RNAi components, and studies involving genetic mutants of RdRP and other RNAi‐related genes in N. crassa could not reveal any role of RNAi in heterochromatin formation (Camblong et al., 2007; Chicas et al., 2004; Freitag et al., 2004). These findings indicate the existence of an alternative RNA‐based regulatory mechanism that governs epigenetic inheritance.

In Sc. pombe, an RNA‐based regulatory mechanism involving exosomes has been identified, which works in parallel with the RNAi pathway (Buhler et al., 2007). Exosomes are protein complexes that are known to degrade aberrant RNA and aid in heterochromatin formation. RNAs which are shuttled between the cells via exosomes are called endogenous short RNAs (esRNAs) (Lotvall and Valadi, 2007; Schorey et al., 2015). Deep sequencing of M. oryzae revealed a spectrum of esRNAs, which were accumulated in the vegetative and infection‐specialized tissues (appressoria) (Nicolas and Ruiz‐Vazquez, 2013; Nicolas et al., 2013; Nunes et al., 2011). Although the vegetative tissues exhibited the accumulation of sRNAs of 28–35 nucleotides mapping to intergenic and repetitive elements, the appressoria exhibited accumulation mapping to tRNA‐derived fragments (tRFs). However, there is no evidence of the involvement of the RNAi machinery in the production of tRFs, which needs to be validated further before considering them as bona fide esRNAs (Nicolas and Ruiz‐Vazquez, 2013). Furthermore, it would be interesting to study the role of tRFs in the epigenetic regulation of M. oryzae during the infection process.

Kadotani et al. (2004) identified two Dicer‐like genes, MoDcl1 and MoDcl2, in M. oryzae, which are orthologous to dcl‐1 and dcl‐2 of N. crassa (Catalanotto et al., 2004). Unlike N. crassa, in which dcl‐1 and dcl‐2 exhibit functional redundancy for RNA silencing (Catalanotto et al., 2004), MoDcl1 and MoDcl2 of M. oryzae show functional diversification, with MoDcl2 being involved in RNA silencing independent of MoDcl1 (Kadotani et al., 2008). Further, Kadotani et al. (2008) found that MoDcl1 produces siRNAs when overexpressed in the MoDcl2 knockout mutant, suggesting the existence of a MoDcl2‐independent RNAi mechanism. However, the direct or indirect role of these Dicer‐like proteins in the epigenetic regulation of M. oryzae has not been elucidated to date. In addition, sRNA transcriptome profiling of M. oryzae under different physiological stresses (such as nitrogen starvation, carbon starvation, minimal media and paraquat‐induced oxidative stress) has revealed that the sRNA profile of the pathogen is altered depending on the physiological stress and host conditions (Raman et al., 2013). Genetic analysis of the sRNA biosynthetic mutants of Dicer‐like genes and RdRP genes revealed an increase in the transcription of subsets of genes, including genes involved in the virulence of M. oryzae (Raman et al., 2013). In Phytophthora infestans, sRNA‐mediated silencing of TE‐derived sequences led to the bidirectional silencing of nearby sequences as a result of heterochromatin formation at the affected locus (Judelson and Tani, 2007). A clear correlation of sRNA in the epigenetic regulation of M. oryzae growth or pathogenicity has not been confirmed to date. However, the current reports urge the need for the understanding of sRNA‐mediated epigenetic control in M. oryzae for a clearer picture of pathogen growth, pathogenicity and host–pathogen relations.

Quelling

Discovered in N. crassa, quelling is a homology‐dependent gene silencing mechanism that silences the native sequence homologous to transformed DNA (Romano and Macino, 1992) in vegetative tissues. It is an RNAi‐dependent silencing mechanism induced by aberrant RNA (Catalanotto et al., 2004). Quelling is known to affect heterochromatin formation and gene regulation (Chang et al., 2012; Nicolas et al., 2013). In N. crassa, QDE‐1, QDE‐2 and QDE‐3 genes encode for RdRP, Argonaute‐like protein and Rec‐Q‐like DNA helicase, respectively, implicated in the production of Qde‐2‐associated RNAs (qiRNAs), which inhibit protein translation during DNA damage (Billmyre et al., 2013). QDE‐1 plays an important role in the generation of aberrant RNA, which is important for RNAi‐mediated silencing. However, to the best of our knowledge, quelling in M. oryzae has not been reported to date, indicating the existence and prevalence of alternative mechanisms of RNA‐based silencing in these systems.

MSUD

MSUD is another RNAi‐dependent silencing mechanism discovered in N. crassa. It is known to occur at the sexual developmental stage when the unpaired DNA initiates silencing of the unpaired DNA itself and its homologues (also known as ‘trans’‐silencing) (Aramayo and Metzenberg, 1996; Shiu et al., 2001). MSUD signals trigger the production of aberrant RNA specific to the unpaired DNA, which is converted to dsRNA by a group of proteins, namely SAD‐1 (an RdRP), SAD‐2 (scaffold protein) and SAD‐3 (RNA/DNA helicase) (Chang et al., 2012; Shiu et al., 2001). Further, the dsRNA is cleaved via Dcl1 into 20–25‐nucleotide sRNAs, which next bind to suppressor of meiotic silencing (SMS‐2; an Argonaute homologue) and QIP (an exonuclease), resulting in post‐transcriptional silencing of the homologous genes (Aramayo and Selker, 2013). The above‐mentioned RNAi machinery is essential for heterochromatin formation and functioning of the centromere and telomere (Billmyre et al., 2013). Quelling and MSUD function with different machinery, except for dcl‐1 and QIP, which indicates a shared ancestry that has diverged over time. In addition to N. crassa, MSUD has been reported in Fusarium oxysporum (Chen et al., 2014) and members of the Basidiomycota (Hu et al., 2013). However, MSUD does not seem to be prevalent in fungal plant pathogens. This is understandable in the case of M. oryzae, where the dominant prevalence of the sexual phase has not been reported under natural conditions.

Unexplored Aspects of Epigenetics in M. Oryzae and Other Fungal Plant Pathogens

Fungal model organisms, such as Aspergillus nidulans and N. crassa, have been studied extensively at the epigenetic level (Aramayo and Selker, 2013; Foss et al., 1998; Freitag et al., 2002; Liu et al., 2012; Manzanares‐Miralles et al., 2016; Selker et al., 2003), but the role of epigenetics in the regulation of the pathogenicity and development of fungal plant pathogens has not been fully explored to date. Despite being limited in numbers, epigenetic studies in M. oryzae and a few other fungal pathogens highlight the crucial role of epigenetic modifications in the success of fungal plant pathogens. Reports on the role of epigenetics in the regulation of fungal virulence through effector proteins are examples that highlight the unexplored areas of epigenetic regulation that warrant further investigation (Qutob et al., 2013; Soyer et al., 2014). Here, we highlight the important aspects of epigenetic modifications that are still unexplored, but hold promising potential for the understanding and combating of fungal plant pathogens (Fig. 2).

Epigenetic switch in the control of effectors

Certain fungal avirulence (Avr) genes encode virulence factors, which are called effector proteins (Stergiopoulos and de Wit, 2009). Pathogen effector proteins aid pathogen infection, but can readily become avirulent under host immune surveillance and recognition. Avr genes undergo mutations or multiple translocations to escape from immunity and become virulent again (Chuma et al., 2011; Gijzen et al., 2014; Kang et al., 2001; Kasuga and Gijzen, 2013; Qutob et al., 2013). Alternatively, epigenetic switches can also allow the reuse or recycling of effectors for virulence (Fig. 2). Gijzen et al. (2014) have proposed a model, displaying how changes in epigenetic balance can possibly affect transposon activity or Avr gene expression, culminating in changes in virulence. H3K9me3 was specifically shown to be involved in the epigenetic regulation of effector gene expression in Leptosphaeria maculans, a fungal pathogen of the phylum Ascomycota (Soyer et al., 2014). Changes in the environment of the pathogen from axenic culture to the primary infection site triggered epigenetic modifications that led to a change in expression of effector genes present in the AT‐isochore (adenine–thymine isochore) regions. RNAi silencing of HP1 and defective in DNA methylation (DIM‐5), which are required for heterochromatin assembly and maintenance in L. maculans, led to the overexpression of SSP‐encoding (small secreted protein‐encoding) genes within AT‐rich isochores in the silenced transformants during growth in axenic culture. Evidently, the changes in environment could trigger an epigenetic switch that expressed or repressed pathogenesis via chromatin remodelling.

Modulation of secondary metabolite synthesis

In several fungal pathogens, including M. oryzae, secondary metabolites, such as melanin, have been shown to contribute significantly to fungal pathogenicity (Scharf et al., 2014). Essentially, melanin is a derivative of a phenolic precursor, and is a dark‐coloured pigment that helps to create the high turgor pressure required for fungal penetration into the plant cell wall (Eisenman and Casadevall, 2012; Soanes et al., 2012). Melanin‐deficient M. oryzae mutants produce non‐pigmented appressoria that are unable to penetrate the epidermal cells of the leaf cuticle (Howard and Valent, 1996). Given the significant role of secondary metabolites in the establishment of the pathogenicity of fungal plant pathogens, it is imperative to study the epigenetic modulation of secondary metabolite synthesis and its impact on the success of the pathogen. However, to our knowledge, epigenetic control of secondary metabolites in M. oryzae is still an untouched topic (Fig. 2). Conversely, a number of studies on different fungal systems (A. nidulans, N. crassa) have shown that fungal secondary metabolite synthesis is highly dependent on chromatin modifications and histone methylation, acetylation and deacetylation as a result of the cluster localization of genes on DNA (Gacek and Strauss, 2012; Nutzmann et al., 2011; Reyes‐Dominguez et al., 2010). In A. nidulans, H4 acetylation and acetylation of histone H3 at promoter regions of the secondary metabolite clusters, regulated by GcnE‐containing complexes, have been shown to be crucial for the transcriptional activation of the secondary metabolite gene cluster (Nutzmann et al., 2011; Reyes‐Dominguez et al., 2010). Likewise, mycotoxins, such as aflatoxins, fumonisins and trichothecenes, are other types of secondary metabolite that are produced by crop‐invading fungi (Woloshuk and Shim, 2013). Interestingly, barring DNA methylation, current evidence suggests the involvement of histone modification in the epigenetic regulation of mycotoxin biosynthesis (Liu et al., 2012). Fusarium verticillioides, a fungal pathogen, produces class B fumonisin (FB) mycotoxin which is encoded by a set of clustered and co‐transcribed genes, called FUM genes. ChIP analysis has revealed that histone acetylation‐mediated epigenetic modification regulates the expression of FUM genes (Visentin et al., 2012). Loss of Aflr (a transcription factor) expression (LAEA) is the strongest evidence indicating the epigenetic regulation of mycotoxin biosynthesis. In Aspergillus, it has been shown that the loss of LAEA function represses the production of secondary metabolites, but conidiation and fungal growth are not affected (Bok and Keller, 2004; Georgianna et al., 2010). However, overexpression of LAEA leads to the synthesis of other secondary metabolites (Bok and Keller, 2004; Bok et al., 2006; Georgianna et al., 2010). In addition, a clear link between LaeA and H3K9 methylation/demethylation with the transcription of sterigmatocystin biosynthesis genes has been established (Reyes‐Dominguez et al., 2012) in F. graminearum. Likewise, the mevalonate biosynthetic pathway gives rise to isoprenoids that are precursors of ergosterol, a secondary metabolite. MoAcat1 and MoAcat2 are the homologues of genes encoding acetoacetyl‐CoA in M. oryzae, which is an important catalytic enzyme of the mevalonate pathway. A recent study has shown that the mitochondrial localization of MoAcat1 and the cytoplasmic localization of MoAcat2 are essential for the virulence and vegetative growth (plus virulence) of M. oryzae (Zhong et al., 2015). An understanding of the epigenetic control of secondary metabolite production in M. oryzae could be of huge advantage in the development of preventative strategies of the pathogen. For instance, epigenetic regulation of the secondary metabolite pathway could be effective in disease control. Thus, chromatin‐controlled transcriptional activation of secondary metabolite clusters is of huge importance in the case of fungal pathogens, and needs to be studied further in more detail.

Epigenetic basis of evolution, host–pathogen interactions and epidemic epidemiology

It is proposed that epigenetics can act as a compelling driving force for adaptive evolution in prokaryotes as well as eukaryotes. Generally, the target recognition domain (TRD) of DNMTs recognizes specific DNA sequence patterns for methylation. In the case of bacteria, it has been reported that TRDs move between non‐orthologous genes and sometimes within genes, thereby altering the methylation pattern of the entire genome, eventually causing global changes in gene expression (Furuta and Kobayashi, 2012a, 2012b). It is therefore suggested that, instead of the diverse genome, it is the diverse methylome which is the real target of adaptive evolution in bacterial pathogens. However, fungal genomes, unlike their plant and animal counterparts, are not extensively methylated. Existing reports suggest that M. oryzae possesses DNA methylation for a small portion of the cytosine sites, with low methylation levels (Ikeda et al., 2013; Jeon et al., 2015). Therefore, the possible action of evolutionary selection forces and the eventual role of the methylome in driving adaptive evolution appear to be ambiguous. Alternatively, we suggest that the determination of the pattern of histone methylation in different geographical fungal isolates could help to trace the evolutionary history and spread of fungal pathogens.

Changes in the effector proteins of fungal pathogens aid in the evasion of host immunity. Although this phenomenon is regulated by sequence alterations or loss‐of‐virulence genes, an alternative transgenerational epigenetic mechanism also plays an important role in evading host defence mechanisms. In fungal pathogens, epigenetic variation can control virulence through the epigenetic regulation of TEs, which play a pivotal role in phenotypic diversification and host–pathogen co‐evolution (Kasuga and Gijzen, 2013). Epigenetic mechanisms are strongly implied in host–pathogen interactions, demonstrating the usefulness of epigenetic epidemiology for future research on infectious diseases (Gomez‐Diaz et al., 2012). Although the molecular basis of epigenetics and variation is being delineated, the role of epigenetics in host–pathogen interactions has yet to be explored.

Conclusion and Future Directions

Time and again, epigenetics has been shown to play an important role in the success of plant pathogen persistence in diverse host niches. Although the molecular basis of epigenetics and variation is being slowly delineated in pathogens, the comparative epigenetics of pathogens remains largely unexplored. Epigenetic mechanisms are strongly implied in pathogen success, persistence and host–pathogen interactions, thus indicating the usefulness of epigenetics in laying the foundation of epigenetic epidemiology for future research on infectious diseases. Insights into epigenetic changes in the pathogen during infection and intra‐ and intercellular proliferation may be informative in combating plant diseases. Moreover, comparing the epigenetic changes under different growth stages, forms and conditions between different types of plant pathogen may highlight specific trends of epigenetic changes in plant pathogens. Briefly, we have attempted to delineate the epigenetic changes in M. oryzae with prospective implications on other fungal plant pathogens, and have also identified the unexplored areas requiring further work in M. oryzae itself. However, the number of studies specific to fungal plant pathogens is seriously limited for the effective synthesis and understanding of the process, such that there is a need for further research in the field of fungal pathogen control, evolution and association with hosts. Studies on the loss or retention of epigenetic modifications can throw light onto the fungal life cycle and its evolutionary success. Further, the identification of conserved epigenetic modifications in diverse pathogens located in different geographical regions can be informative in understanding successful pathogenicity or epigenetic trends during pathogenesis.

Conflict of Interest

The authors declare no conflicts of interest regarding the publication of this review paper.