Polyphenol oxidase

Polyphenol oxidase (PPO) is often portrayed in scientific literature as a single enzyme, and one whose name is frequently used interchangeably with the enzyme names tyrosinase, polyphenolase, phenolase, catechol oxidase, cresolase and catecholase (Yoruk and Marshall, Reference Yoruk and Marshall2003). In reality, PPO is a broad term encompassing three distinct enzymes: catecholase, laccase and cresolase or tyrosinase. Cresolase and tyrosinase are the same enzyme, but they have sometimes been assigned the name cresolase in the case of plants, and tyrosinase in the case of animals and microbes (Aniszewski et al., Reference Aniszewski, Lieberei and Gulewicz2008; Kaintz et al., Reference Kaintz, Mauracher, Rompel and Christov2014). However, the first tyrosinase reported of plant origin was recently isolated from walnut (Juglans regia) leaves (Zekiri et al., Reference Zekiri2014). Laccases are a diverse group of enzymes with broad substrate specificity that catalyse the oxidation of a wide range of phenolic substrates in plants, fungi, bacteria and insects (Giardina et al., Reference Giardina2010).

PPO enzymes oxidize phenolic compounds to generate quinones. Quinones are highly reactive compounds that will further react non-enzymatically through self-polymerizing or covalently bonding and cross-linking with amino acids or proteins to form high molecular weight melanin pigments (Walker and Ferrar, Reference Walker and Ferrar1998). PPOs catalyse distinct hydroxylation and oxidation reactions: (1) the hydroxylation of monophenols to o-diphenols, and (2) the oxidation of diphenols (and other substrates) to the associated quinones in the presence of molecular oxygen. The first reaction is catalysed by cresolase or tyrosinase (also known as monophenolase or monophenol monooxygenase). The second reaction is catalysed by catecholase (also known as diphenolase or diphenol oxidase) or laccase (Marusek et al., Reference Marusek2006; Aniszewski et al., Reference Aniszewski, Lieberei and Gulewicz2008). Catecholases oxidize o-diphenols to o-diquinones, whereas laccases oxidize o- and p-diphenols, and other substrates including triphenols, to p-diquinones, other quinones, and semiquinones (Aniszewski et al., Reference Aniszewski, Lieberei and Gulewicz2008). The enzymes that catalyse only the latter reaction are often referred to broadly as catechol oxidases and their activity in a diverse range of living organisms has been extensively studied (Gerdemann et al., Reference Gerdemann, Eicken and Krebs2002b). In contrast, cresolase activity has been far less investigated. Whether cresolase is always present in organisms that produce CO is debated in the literature (Marusek et al., Reference Marusek2006).

PPOs are a highly diverse group of enzymes that differ widely in structure, amino acid sequence, function, temporal and spatial expression, and substrate specificity (Mayer, Reference Mayer2006; Cai et al., Reference Cai2013). Given the latter characteristic, measured activity may differ if one substrate is used versus another. A highly conserved trait shared by all PPOs is that they are metalloenzymes with an active site consisting of a binuclear type-3 copper centre containing two copper ions (CuA and CuB), each bound to three histidine side chains (Marusek et al., Reference Marusek2006; Aniszewski et al., Reference Aniszewski, Lieberei and Gulewicz2008). Additional common structures shared by PPOs include a signal peptide, which directs the enzyme to the thylakoid lumen of chloroplasts or to the vacuolar lumen; a highly conserved N-terminal containing the features for copper binding, substrate catalysis, and structural maintenance; a variable C-terminal domain; and a linker region between the C- and N-termini that is highly variable in size and structure (Marusek et al., Reference Marusek2006; Tran and Constabel, Reference Tran and Constabel2011; Cai et al., Reference Cai2013).

The described structure of the mature, latent PPO enzyme ranges in size from 39 to 73 kDa depending on plant species (Aniszewski et al., Reference Aniszewski, Lieberei and Gulewicz2008). Evidence suggests that proteolytic cleavage of the 15–20 kDa C-terminal is required for enzyme activation because the C-terminal physically shields the activation site (Gerdemann et al., Reference Gerdemann2002a; Marusek et al., Reference Marusek2006; Flurkey and Inlow, Reference Flurkey and Inlow2008). The mature, inactive form of PPOs is highly stable, but it shows much greater thermal lability following activation via proteolytic cleavage (van Gelder et al., Reference van Gelder, Flurkey and Wichers1997). Endogenous and exogenous proteases, such as those from pathogenic fungi, are known to activate PPOs; anionic detergents, acids and lipids have been shown to activate PPOs in vitro (van Gelder et al., Reference van Gelder, Flurkey and Wichers1997; Fuerst et al., Reference Fuerst2011, Reference Fuerst2014).

The mechanism by which PPOs catalyse reactions in plants relies on the interaction of the copper ions in the enzyme with molecular oxygen and substrates (Aniszewski et al., Reference Aniszewski, Lieberei and Gulewicz2008). Possible ways which PPO imparts anti-fungal defences to seeds may include (1) producing toxic quinones, (2) reducing nutrient bioavailability by cross-linking molecules, (3) cell wall lignification, and (4) involvement in reactive oxygen species (ROS) generation (Constabel and Barbehenn, Reference Constabel, Barbehenn and Schaller2008; Fuerst et al., Reference Fuerst2014).

PPO is induced in seeds challenged with fungal pathogens and its activity exhibits spatial and temporal variability. Following inoculation of developing wheat heads with Fusarium graminearum, maximum PPO activity, as assessed using pyrocatechol (1,2-dihydroxybenzene) as the substrate, occurred during the milk stage for resistant and susceptible cultivars, and steadily declined thereafter (Mohammadi and Kazemi, Reference Mohammadi and Kazemi2002). The level of PPO activity was three times greater in the resistant cultivars compared with the non-inoculated controls, and twice as great as the susceptible cultivars, suggesting that PPO induction in developing wheat heads could be a defensive response to pathogen attack. Moreover, seven isoforms of the PPO enzyme were detected in the extracts of wheat heads and they were differentially expressed among the cultivars and throughout grain development.

PPO was induced in dormant caryopses of wild oat isoline M73 following incubation on Fusarium avenaceum isolate F.a.1, as detected using l-DOPA (l-3,4-dihydroxyphenylalanine) as the substrate (Anderson et al., Reference Anderson2010; Fuerst et al., Reference Fuerst2011, Reference Fuerst2018). Fractionation of the extracted proteins from caryopsis leachates showed predominantly 57 kDa proteins, and to a lesser extent 36 kDa, from untreated caryopses; several lower molecular weight proteins, including 24, 25 and 36 kDa, were present in the F.a.1-treated caryopses. The 36 and 57 kDa protein sequences were highly similar to wheat PPO, but the 24 and 25 kDa proteins were most similar to oxalate oxidase and chitinase, respectively. The authors hypothesized that F.a.1 induction of latent wild oat PPO involves proteolytic cleavage, thereby yielding a lower molecular weight activated form of water-soluble PPO that readily diffuses into the seed environment (Anderson et al., Reference Anderson2010; Fuerst et al., Reference Fuerst2011). Detection of PPO activity in the leachates of wild oat caryopses following incubation on F.a.1 using l-DOPA as substrate (Fuerst et al., Reference Fuerst2011, Reference Fuerst2018), as well as in the supernatant of imbibed wheat bran and using hydroquinone monomethyl ether and 3-methylbenzothiazolin-2-one hydrazone as substrates (Jerkovic et al., Reference Jerkovic2010), suggests that PPO seed defence function is amplified by its mobility in the external seed environment. For example, PPO activity in leachates from dormant wild oat caryopses incubated on F.a.1 increased 7.5-fold compared with untreated caryopses, whereas PPO activity from whole caryopses increased only 3.4-fold after incubation on F.a.1 compared with the untreated controls (Fuerst et al., Reference Fuerst2018).

After wild oat caryopses were incubated in vitro on three Fusarium spp. isolates (F.a.1, F. culmorum-2, and F. culmorum-4), PPO activity, as measured with l-DOPA substrate, increased 106, 47 and 24%, respectively, compared with the untreated controls. PPO detection in the controls, however, indicated that PPO was also expressed constitutively. In contrast, incubation on a Pythium isolate decreased PPO activity by 26% (Fuerst et al., Reference Fuerst2011), illustrating a species-specific response. PPO induction from the three Fusarium isolates correlated with their virulence, as demonstrated by the rate at which they caused visible symptoms of seed decay, F.a.1 > F.c.2 > F.c.4 (de Luna et al., Reference de Luna2011). The hull fraction (lemma and palea) of wild oat seeds also showed induction of PPO activity following incubation on F.a.1, and at a greater level than in the caryopses, suggesting that these non-living tissues are also capable of actively mounting a biochemical defence response (Fuerst et al., Reference Fuerst2011).

The mechanism by which dead maternal seed tissue can respond in defence has not been conclusively determined. Studies have shown that genes encoding enzyme expression are active at various developmental stages of the pericarp. For example, a gene encoding a protein similar to polyphenol oxidase is expressed in seed coats of developing Arabidopsis thaliana (Pourcel et al., Reference Pourcel2005). Barley seed transcriptomics indicates that numerous genes regulate expression of protease enzymes in the pericarp and that seed developmental stages are characterized by different expression patterns. For example, distinct proteases are involved at specific phases of programmed cell death in the pericarp (Sreenivasulu et al., Reference Sreenivasulu2006). Programmed cell death of maternal seed tissues results in the degradation and redistribution of nutritional cellular materials to the developing filial tissues (Domínguez and Cejudo, Reference Domínguez and Cejudo2014). Research suggests that enzymes that serve defence roles remain stored in the non-living outer seed layers following programmed cell death and that these active enzymes are released from the tissues upon hydration (Godwin et al., Reference Godwin, Raviv and Grafi2017; Raviv et al., Reference Raviv2017a,Reference Ravivb). Additional hypotheses propose that fungal proteases activate latent PPO through proteolytic cleavage of its C-terminal peptide and that this can occur in non-living seed hulls (Fuerst et al., Reference Fuerst2014). Other seed defence enzymes may be activated in a similar manner, but the hypothesis has not been tested (Fuerst et al., Reference Fuerst2018). Given the presence of proteases in seed pericarps, it is also possible that endogenous seed proteases proteolytically activate PPO.

Research indicates that PPO is generally localized in the outer layers of seeds. For example, microdissection followed by proteomic analysis using two-dimensional gel electrophoresis (2DGE) showed that PPO activity was only present in the water-extractable protein fraction from the pericarp of wheat bran (Jerkovic et al., Reference Jerkovic2010). Fuerst et al. (2010) likewise concluded that PPO activity is predominantly present in the bran fraction of wheat. Investigation of enzyme activity in the germ aleurone layer of mature barley kernels indicated that phenol oxidase was present in the cytoplasm of germ aleurone cells (Cochrane, Reference Cochrane1994). This publication was also one of the earliest studies to suggest a possible anti-microbial defence function of phenol oxidases in seeds (Cochrane, Reference Cochrane1994). PPOs most often are initially translocated into the chloroplast thylakoid lumen (Golbeck and Cammarata, Reference Golbeck and Cammarata1981), but a vacuolar form has also been reported (Tran and Constabel, Reference Tran and Constabel2011), and other possible subcellular locations have been proposed (Mayer, Reference Mayer2006). Of course, such subcellular localization becomes less relevant when considering the outer tissues of seeds, some of which are non-living and thus lack cellular compartmentalization.

Peroxidase

The peroxidase (POD) enzyme category encompasses a vast diversity of enzymes that are found throughout living systems. Heme PODs are the most well-known PODs, appearing in the scientific literature since at least 1908 (Kastle and Porch, Reference Kastle and Porch1908). More recently, in 1996, heme-free and thiol-dependent PODs, termed peroxiredoxins, were discovered in plants (Stacy et al., Reference Stacy1996). While limited studies suggest peroxiredoxins may be involved in pathogen defence (Rouhier et al., Reference Rouhier2004; Haddad and Japelaghi, Reference Haddad and Japelaghi2015), relatively little is known about the function of these thiol PODs in seeds (Bhatt and Tripathi, Reference Bhatt and Tripathi2011). Therefore, this review will address only heme-containing plant PODs.

Heme PODs are classified into two superfamilies: the ‘animal’ PODs (found only in animals), and the ‘plant’ PODs (found in plants, fungi and prokaryotes) (Dunford, Reference Dunford1999). The ‘plant’ PODs all have a common heme group, formed from protoporphyrin IX and an Fe(III) (Banci, Reference Banci1997), and share a similar three-dimensional protein structure (Cosio and Dunand, Reference Cosio and Dunand2008), but are further divided into three separate classes based on amino acid sequence (Welinder, Reference Welinder1985, Reference Welinder1992).

Class III PODs are the plant-specific secretory PODs. Unlike class I and II PODs, class III PODs are glycoproteins, contain two Ca²⁺ ions for enhanced structural stability (Banci, Reference Banci1997), and they are secreted into plant cell walls and vacuoles (Barceló et al., Reference Barceló2003; Passardi et al., Reference Passardi2004). Found in all land plants and throughout the plant lifecycle, class III PODs are members of large multigenic families (Passardi et al., Reference Passardi2004; Cosio and Dunand, Reference Cosio and Dunand2008). They catalyse the single-electron oxidation of various hydrogen donors, such as phenolics, lignin precursors, auxin or specialized metabolites, thereby reducing H₂O₂ in the process (Barceló et al., Reference Barceló2003; Passardi et al., Reference Passardi2004). PODs generally have broad substrate specificity, with a moderate specificity for phenols (Hiraga et al., Reference Hiraga2001), and they demonstrate an unusually high degree of thermal stability (Vámos-Vigyázó and Haard, Reference Vámos-Vigyázó and Haard1981; Fujita et al., Reference Fujita1995; Barceló et al., Reference Barceló2003). Numerous isoforms exist within the class III PODs, each exhibiting a distinct amino acid sequence and subject to heterogeneous regulation of gene expression (Hiraga et al., Reference Hiraga2001). This vast isozyme diversity is likely to be responsible for the diverse physiological processes that PODs catalyse (Hiraga et al., Reference Hiraga2001; Passardi et al., Reference Passardi2005).

PODs have been detected in seeds since as early as 1973 when 14 POD isozymes were extracted from ground mature barley grains (LaBerge et al., Reference LaBerge, Kruger and Meredith1973). At that time, however, the functional role of PODs in seeds was relatively unknown, and they were viewed rather as possible genetic markers for breeding since the isozyme profiles varied widely by cultivar. PODs have since been implicated in numerous active and passive defence-related activities. Passive, or constitutive, defences include cell wall lignification and suberization via cross-linking cell wall compounds, whereas active defences include production of ROS and synthesis of diverse antimicrobial phytoalexins (Passardi et al., Reference Passardi2005; Almagro et al., Reference Almagro2009). Lignification of plant cell walls by POD inhibits penetration by fungal hyphae (Cochrane et al., Reference Cochrane, Paterson and Gould2000). Class III POD defence activity occurs in response to both abiotic and biotic stressors. These can include heavy metal exposure, physical wounding, or pathogen and herbivore attack (Passardi et al., Reference Passardi2005). Research suggests that class III PODs are constitutively produced, but that levels modulate in response to abiotic and biotic stressors (Barcelo et al., Reference Barceló2003; Almagro et al., Reference Almagro2009).

POD activity has often been shown to increase when seeds are challenged with pathogens, although the expression varies with fungal species and plant cultivar. POD was isolated from mature seeds of French bean (Phaseolus vulgaris cv. Kentucky Wonder) and in vitro tests showed it strongly inhibited the fungi Coprinus comatus and Botrytis cinerea, but weakly inhibited Mycosphaerella arachidicola and Fusarium oxysporum (Ye and Ng, Reference Ye and Ng2002). This study demonstrated not only the active pathogen defence function of POD in seeds, but also the variable specificity of the enzyme activity towards different fungal species. Some pathogen-specific differences in the defence enzyme gene expression profiles of soybean [Glycine max (L.) Merr.] seeds 35 days post-anthesis (DPA) were also apparent; 48 h after inoculation with Diaporthe phaseolorum var. meridionalis, ~2-fold greater magnitude of upregulated POD gene expression was observed compared with inoculation with Cercospora kikuchii (Upchurch and Ramirez, Reference Upchurch and Ramirez2010). Compared with the control, D. phaseolorum and C. kikuchii induced ~6 and ~12 times greater POD expression, respectively, in soybean seeds. Induction of POD activity in response to a fungal pathogen is also seen in dormant caryopses, although the degree of enzyme response differs by plant species. Following incubation of dormant wild oat and wheat caryopses on the seed-decay fungal isolate F.a.1, POD activity from the whole caryopses increased 2.4- and 3.4-fold in the wild oat and wheat, respectively, compared with the POD activity from untreated caryopses (Fuerst et al., Reference Fuerst2018).

The POD isozyme profiles from different barley and wheat cultivars change as seeds develop and according to cultivar (Kruger and LaBerge, Reference Kruger and LaBerge1974; LaBerge, Reference LaBerge1975). Mature barley kernels (24–72 DPA) exhibited a greater number of POD isozymes than immature kernels (10–19 DPA) and the overall activity was higher at 24–72 DPA than at 10–19 DPA (LaBerge, Reference LaBerge1975). The opposite trend was seen in a recent study in which hull-less barley was inoculated during anthesis with Fusarium graminearum, the major fungal pathogen causing Fusarium head blight, and the albumin and globulin soluble protein fractions from the grain were assessed at five phenological stages post-inoculation. While albumin and globulin seed proteins function primarily in nutrient storage, increasing evidence suggests they also function in defence against fungi, bacteria and insects (Terras et al., Reference Terras1992, Reference Terras1993; Marcus et al., Reference Marcus1999; Sales et al., Reference Sales2000; Freire et al., Reference Freire2015). Using 2DGE and mass spectrometry, POD was detected at increasing abundance during the milk stages from 7 to 14 days after inoculation (DAI), but decreased during the proceeding dough stages (21 to 54 DAI) to non-detectable levels (Trümper et al., Reference Trümper2016). Comparable results were seen in the POD activity of resistant and susceptible wheat heads inoculated with F. graminearum, which increased significantly during the milk stage compared with the non-inoculated controls (Mohammadi and Kazemi, Reference Mohammadi and Kazemi2002). Gel staining indicated the presence of three basic and six acidic POD isozymes and that they were differentially distributed and expressed among the four cultivars. Proteomic analysis of resistant and susceptible peanut (Arachis hypogaea L). seeds subjected to variable watering conditions and Aspergillus flavus treatments showed differential responses between the two cultivars (Wang et al., Reference Wang2010). POD expression increased in the resistant and susceptible cultivars under drought conditions + A. flavus inoculation compared with drought without inoculation, but POD was upregulated ~1.5 times more in the resistant cultivar when pathogen challenged compared with the susceptible cultivar. Using similar methods and also quantitative real-time polymerase chain reaction, POD expression in non-inoculated field-grown wheat grain was increasingly up-regulated 12 to 16 DPA, but by 21 DPA its expression drastically declined (Dong et al., Reference Dong2012). The contrasting correlation between POD expression and developmental stage seen in LaBerge (Reference LaBerge1975) compared with the latter studies may be due to several factors, including innate differences between plant cultivars, or use of more precise methods and advanced technology in the more recent studies.

In one of the earliest studies to suggest a possible anti-microbial seed defence function of POD, POD activity was observed in the cytoplasm and cell walls of germ aleurone cells of mature barley (cv. Triumph) kernels following overnight incubation of tissue in substrate solution (Cochrane, Reference Cochrane1994). In a study of 14 POD isozymes in barley (cv. Centennial) kernels, there was great variability in the distribution pattern of the individual isozymes among the grain layers; some isozymes were confined to particular layers, while others would be present throughout the grain (LaBerge, Reference LaBerge1975). Studies also show that the distribution of individual POD enzymes changes as wheat and barley grains develop and this varies by cultivar (Kruger and LaBerge, Reference Kruger and LaBerge1974; LaBerge, Reference LaBerge1975). For example, 10 days after flowering, 75% of total seed POD was located in the wheat grain pericarp, but by 40 days after flowering, this percentage had dropped to ~10% in cv. Hercules but to only ~30% in cv. Manitou (Kruger and LaBerge, Reference Kruger and LaBerge1974). Differences in proteomic profiles were also seen between cultivated wheat and wild emmer wheat (Triticum turgidum var. dicoccoides), with significantly greater expression of detoxifying and oxidative enzymes, including POD, in the wild emmer wheat (Raviv et al., Reference Raviv2017b). In their analysis of the wheat bran proteome, Jerkovic et al. (Reference Jerkovic2010) identified POD activity solely in the water-extractable protein fraction from the pericarp, and studies indicate that POD is located in the outer surface layers of wild oat (Fuerst et al., Reference Fuerst2014). Significant POD activity was detected in the glumes of wild emmer wheat (Raviv et al., Reference Raviv2017b) and in the seed coat of Sinapis alba and Anastatica hierochuntica (Raviv et al., Reference Raviv2017a).

Oxalate oxidase

The nomenclature surrounding oxalate oxidases (OxOs) needs to be clarified due to research developments and inconsistencies in nomenclature that occurred as biochemical advances in the enzymology of OxOs were refined. OxO enzymes are one of two main subgroups within the germin protein family, which itself belongs to the cupin protein superfamily (Dunwell, Reference Dunwell1998; Dunwell et al., Reference Dunwell, Purvis and Khuri2004). Given that the other subgroup contains ‘germin-like proteins’, OxOs are sometimes referred to as ‘true germins’ (Davidson et al., Reference Davidson2009). Germins comprise a group of homologous proteins that are found solely in true cereals (Lane, Reference Lane2002). They were first identified as a marker associated with germination onset in wheat embryos, hence the name germin (Thompson and Lane, Reference Thompson and Lane1980; Lane, Reference Lane2002). In older literature before it was widely determined that some cereal germins were in fact oxalate oxidases, OxOs were sometimes referred to as germin-like proteins, although that misnomer has been avoided in more recent literature. Thorough histories of germin and OxO characterization are presented in the literature (Dunwell, Reference Dunwell1998; Lane, Reference Lane2002; Dunwell et al., Reference Dunwell2008). OxOs are encoded by a large homogenous group of genes found exclusively in true cereals (Poaceae family) (Davidson et al., Reference Davidson2009). Proteins with OxO activity have been discovered in non-cereal crops including banana, sorghum and beet, but they appear to be proteins distinct from germin-OxOs (Lane, Reference Lane2000; Davidson et al., Reference Davidson2009) and are instead considered germin-like proteins. Germin-like proteins and OxOs share an average of 50% sequence identity, including the conserved germin motif, and have similar biochemical properties (Dunwell et al., Reference Dunwell, Khuri and Gane2000; Davidson et al., Reference Davidson2009). However, germin-like proteins differ from OxOs in numerous ways. In contrast to OxOs, germin-like proteins are encoded by a heterogeneous group of genes found in diverse plant species (Davidson et al., Reference Davidson2009). Germin-like proteins are also more functionally diverse, as they may exhibit activity from various enzymes, including OxO, superoxide dismutase, PPO and ADP glucose pyrophosphatase (Barman and Banerjee, Reference Barman and Banerjee2015).

All OxOs are glycoproteins consisting of six β-jellyroll monomers, each a barrel consisting of four pairs of anti-parallel beta sheets, locked in a homohexamer (Woo et al., Reference Woo1998, Reference Woo2000). This structure is responsible for the high stability and broad resistance of all OxOs to proteolysis, dehydration, heat, SDS and pH extremes (Woo et al., Reference Woo2000; Dunwell et al., Reference Dunwell2008). Aside from the conserved jellyroll β-barrel structure of all OxOs, specific OxO enzymes display a range of physical and chemical characteristics. For example, analysis of four purified OxOs from rice (Oryza sativa L.) leaves showed that they differed in molecular mass, optimum pH, stability, and responses to inhibitors and activators (Li et al., Reference Li2015). A highly purified and crystallized germin-OxO from barley showed both OxO and extracellular superoxide dismutase activity, although superoxide dismutase activity is typically not detected in the germin-OxOs (Woo et al., Reference Woo2000).

OxOs are a group of water-soluble enzymes that catalyse the two-electron oxidative decarboxylation of endogenous oxalate to hydrogen peroxide and carbon dioxide. Insoluble crystallized calcium oxalate present in plant vacuoles and cell walls is an additional OxO substrate, whose oxidation also yields free Ca²⁺ (Lane, Reference Lane1994; Dunwell et al., Reference Dunwell2008). OxO-catalysed reactions rely only on the presence of manganese (II) and unlike other apoplastic oxidase enzymes, such as the amine oxidases, OxO is independent of other external cofactors (Requena and Bornemann, Reference Requena and Bornemann1999; Woo et al., Reference Woo2000). OxO preproteins usually contain apoplastic secretory signal peptides at the N-terminal that direct the protein from the site of synthesis in the cytosol to the endoplasmic reticulum membrane for secretion, consistent with their role in cell wall function and pathogen defence (Zimmermann et al., Reference Zimmermann2006; Imai and Nakai, Reference Imai and Nakai2010; Fuerst et al., Reference Fuerst2014).

OxOs are a functionally diverse enzyme group, involved in calcium regulation, oxalate metabolism, cell wall strengthening, stress response, and pathogen defence (Dunwell et al., Reference Dunwell2008; Davidson et al., Reference Davidson2009). Moreover, OxO may in fact function by several mechanisms. Cell wall strengthening may result from peroxidative cross-linking involving the generated hydrogen peroxide, and also via lignification and papillae formation (Davidson et al., Reference Davidson2009; Kanauchi et al., Reference Kanauchi, Milet and Bamforth2009). Fortified cell walls may impede fungal hyphal penetration and be more resistant to degradation by fungal enzymes (Davidson et al., Reference Davidson2009). Hydrogen peroxide is also directly toxic to pathogens and plays a role in plant immune signalling cascades (Alvarez et al., Reference Alvarez1998). Fungal-derived oxalic acid may be degraded by plant OxOs (Kanauchi et al., Reference Kanauchi, Milet and Bamforth2009). The defence role of oxalate oxidase is illustrated in the numerous studies in which the barley or wheat oxalate oxidase gene transformed into numerous crops such as soybean, oilseed rape, sunflower and peanut confers resistance to the fungal pathogen Sclerotinia spp. (Donaldson et al., Reference Donaldson2001; Hu et al., Reference Hu2003; Livingstone et al., Reference Livingstone2005; Dong et al., Reference Dong2008). In germinating barley seedlings, OxO and POD genes were induced in response to the seed-borne fungus Pyrenophora graminea (Haegi et al., Reference Haegi2008). However, there are few studies documenting OxO induction in non-germinating seeds. Indeed, Fuerst et al. (Reference Fuerst2018) demonstrated inhibition of OxO activity in dormant wild oat seeds in response to a pathogen.

OxOs are generally constitutively expressed in multiple types of plant tissue (Davidson et al., Reference Davidson2009), but OxO activity also displays temporal and spatial variability in cereal grains (Lane, Reference Lane2000). For example, four rice OxO genes with >90% amino acid identity exhibit widely variable expression patterns: OsOXO1 is expressed in the panicles and during flowering and pollination, OsOXO3 is expressed in the roots and seeds, and OsOXO4 is expressed in healthy roots, shoots, leaves and seeds as well as during drought, cold stress and CuSO₄ stress (Carrillo et al., Reference Carrillo2009). OxOs are extracted from soluble and cell wall plant protein fractions, suggesting that they are secreted into the apoplast (Davidson et al., Reference Davidson2009). They concentrate in epidermal tissues of mature grains and developing embryos of cereals, and are also detected in mesophyll tissues (Lane, Reference Lane2000; Wu et al., Reference Wu2000). OxO was detected in the aleurone and embryo of ungerminated malting barley (Kanauchi et al., Reference Kanauchi, Milet and Bamforth2009). In contrast, OxO activity was detected in the pericarp and the intermediate fractions (testa and nucellar tissue) of wheat bran, but was not detected in the aleurone cells (Jerkovic et al., Reference Jerkovic2010).

Chitinase

Chitinases (CHIs) are glycosyl hydrolase enzymes that catalyse the hydrolytic cleavage of β-1,4-glycoside bonds within chitin, a linear homopolymer of β-1,4-linked N-acetyl-d-glucosamine that is abundant in fungal cell walls, as well as in algae, bacteria and invertebrate exoskeletons (Grover, Reference Grover2012). CHIs are ubiquitous in nature as they are produced by microbes, insects, plants and animals. CHI nomenclature and classification have changed over the decades and consistency is not apparent in the literature (Patil et al., Reference Patil, Ghormade and Deshpande2000; Dahiya et al., Reference Dahiya, Tewari and Hoondal2006; Duo-Chuan, Reference Duo-Chuan2006). N-acetylglucosaminidases, which have different cleavage patterns from CHI, are sometimes referred to as chitinolytic enzymes, although they are considered CHIs in this review (Dahiya et al., Reference Dahiya, Tewari and Hoondal2006; Seidl, Reference Seidl2008).

According to similarities in amino acid sequences of the catalytic domain, CHIs have been classified predominantly as members of families 18 and 19 within the glycosyl hydrolase (GH) superfamily, and the N-acetylglucosaminidases are found in GH20 (Seidl, Reference Seidl2008). Recent, yet very limited, research indicates that CHIs are also present in GH families 23 (Arimori et al., Reference Arimori2013) and 48 (Fujita et al., Reference Fujita2006). GH18 and GH19 are distinguished by their amino acid sequence, three-dimensional structure, signal peptide, isoelectric pH, enzyme localization and catalytic mechanism, suggesting that they have unique evolutionary origins (Duo-Chuan, Reference Duo-Chuan2006; Karlsson and Stenlid, Reference Karlsson and Stenlid2008; Hamid et al., Reference Hamid2013). GH18 CHIs have ancient evolutionary origins and are widely present in archaea, bacteria, fungi, viruses, plants and mammals (Funkhouser and Aronson, Reference Funkhouser and Aronson2007), whereas GH19 CHIs comprise almost exclusively plant CHIs, as well as limited bacterial CHIs (Adrangi and Faramarzi, Reference Adrangi and Faramarzi2013). CHIs were traditionally categorized into as many as eleven classes (Gomez et al., Reference Gomez2002), but modern molecular genetic techniques have resulted in redistribution of CHIs into seven distinct classes (Kasprzewska, Reference Kasprzewska2003; Duo-Chuan, Reference Duo-Chuan2006; Hamid et al., Reference Hamid2013). According to the CAZy classification [Carbohydrate Active Enzymes database (http://www.cazy.org)], GH18 contains classes III and V, and GH19 contains classes I, II, IV, VI and VII (Adrangi and Faramarzi, Reference Adrangi and Faramarzi2013; Lombard et al., Reference Lombard2013). CHIs within each family are further divided into the major categories of endo- and exochitinases, depending on a cleavage pattern of either randomly within the polymer or from a single terminus (Horn et al., Reference Horn2006). Plant CHIs are most frequently endochitinases (Hamid et al., Reference Hamid2013).

Given the huge diversity of CHI enzymes, the specific molecular structure and size, organismal location, substrate specificity and catalytic mechanism vary widely (Kasprzewska, Reference Kasprzewska2003). CHIs found in seeds typically range in size from ~20 to ~40 kDa (Yeboah et al., Reference Yeboah1998; Santos et al., Reference Santos2004; Chang et al., Reference Chang2014; Raviv et al., Reference Raviv2017a, Reference Ravivb). CHIs contain a peptide signal sequence at the N-terminal that directs them into the lumen of the endoplasmic reticulum (Chrispeels, Reference Chrispeels1991).

Plant CHIs, and especially those containing a carbohydrate-binding module, function primarily in pathogen defence, notably against pathogenic fungi (Jashni et al., Reference Jashni2015b). Induction of CHI activity in plants by several microbial pathogens has been reported (Metraux and Boller, Reference Metraux and Boller1986; Schlumbaum et al., Reference Schlumbaum1986; Zhu et al., Reference Zhu1994; Robert et al., Reference Robert2002). CHIs are pathogenesis-related (PR) proteins, present in four (PR-3, -4, -8 and -11) of the 17 PR families (Ebrahim et al., Reference Ebrahim, Usha and Singh2011; Sultan et al., Reference Sultan2016). As such, they are produced by plants in response to pathogen attack and participate in systemic acquired resistance, a defence system that enables plants to respond to diverse pathogens (Adrangi and Faramarzi, Reference Adrangi and Faramarzi2013). PR-3 CHIs, which are specific fungal growth inhibitors, concentrate in vacuoles and are abundant in the intermediate bran layers of wheat grains where they may prevent fungal hyphae from attacking the living aleurone layer (Jerkovic et al., Reference Jerkovic2010). A CHI isolated from seeds of the perennial legume Adenanthera pavonina was likewise localized to vacuoles within cotyledon cells (Santos et al., Reference Santos2004). Evidence shows that soluble CHI is secreted into the environment of both germinating and non-germinated seeds, possibly to aid in seed defence (Santos et al., Reference Santos2004; Jerkovic et al., Reference Jerkovic2010).

CHI has long been suggested as a seed defence enzyme (Leah et al., Reference Leah1991; Huynh et al., Reference Huynh1992; Gomes et al., Reference Gomes, Oliveira and Xavier-Filho1996). A 26 kDa CHI purified from mature barley seeds inhibited fungal growth alone and when combined with other seed proteins. For example, alone it inhibited growth of Trichoderma reesei 50%, but when paired with ribosome-inactivating protein, T. reesei was more than 95% inhibited (Leah et al., Reference Leah1991). Fusarium sporotrichioides, a barley seed rot pathogen, Rhizoctonia solani, and Botrytis cinerea were similarly inhibited by this barley seed CHI acting synergistically with other seed enzymes (Leah et al., Reference Leah1991).

CHIs are expressed constitutively in plant stems, seeds, flowers and tubers (Hamid et al., Reference Hamid2013). They are also induced locally in response to stress, including heavy metal exposure, osmotic stress and low temperature, with expression exhibiting spatial and temporal variability (Gomez et al., Reference Gomez2002; Hong and Hwang, Reference Hong and Hwang2006; Rodríguez-Serrano et al., Reference Rodríguez-Serrano2009; Su et al., Reference Su2014). Seed CHIs are present in several seed tissues, such as endosperm, aleurone, embryo or husk (Gomez et al., Reference Gomez2002). Proteomic analysis of wheat bran found CHI activity expressed in the intermediate bran layer, which includes the testa and nucellar tissue, and in the aleurone layer (Jerkovic et al., Reference Jerkovic2010). CHI expression in mature barley seeds was detected only in aleurone cells and not in the endosperm (Leah et al., Reference Leah1991). Baek et al. (Reference Baek, Han and Jo2001) discovered that four distinct CHIs extracted from rice seed were differentially located among the polished rice (endosperm), rice bran and rice hull fractions, with the greatest activity present in the hulls. Activity from a 22 kDa CHI was detected in the caryopses, glumes, lemmas and paleas of wild emmer wheat, with the strongest activity present in the lemma fraction (Raviv et al., Reference Raviv2017b). An additional 40 kDa CHI was also detected in the lemmas, illustrating the diversity of CHI found in seeds and their tissue specificity. CHI activity has also been detected in non-living seed coats of Sinapis alba, Anastatica hierochuntica and 37-year-old radish (Raviv et al., Reference Raviv2017a). Expression of a CHI transcript was strongly evident in developing soybean seeds, but virtually no expression of the transcript was detectable in the soybean leaves or stems, highlighting the tissue specificity of CHI enzymes. Two kidney bean (Phaseolus vulgaris L.) cultivars exhibited differential CHI distribution between the cotyledon, axis and seed coat, with the greatest activity present in the seed coat, although a significant difference was seen between cultivars (Ramos et al., Reference Ramos1998). For example, the seed coat and cotyledon of kidney bean cv. Maisugata had 597 and 27 units mg^–1 CHI, respectively, whereas the levels in cv. Surattowonder were 1238 and 25 units mg^–1, respectively. These studies illustrate the diversity of CHIs present in seeds and the wide inter- and intraspecific variability in localization patterns between different plant species and cultivars. Of two CHIs present in rye (Secale cereal L.) seed, RSC-a localized only to aleurone cells, while RSC-c was predominantly located in the starchy endosperm; neither was present in the testa (Taira et al., Reference Taira2001). Both rye CHIs significantly inhibited hyphal growth of soil-isolated Trichoderma sp. in vitro within 24 h, suggesting that seeds contain multiple layers of anti-fungal CHI defences.

Temporal variability of CHI expression may serve to protect developing seeds when they are most vulnerable, as seen in the increasing expression of a CHI transcript in developing soybean seeds from 13 to 38 DPA, followed by a drastic decline in expression by 48 DPA (Yeboah et al., Reference Yeboah1998). A CHI enzyme isolated from dehulled cowpea (Vigna unguiculata L. Walp) seeds was shown to inhibit growth of the phytopathogenic fungi Colletotrichum lindemuthianum in vitro (Gomes et al., Reference Gomes, Oliveira and Xavier-Filho1996), yet no CHI activity was detected in the exudates of germinating cowpea seeds, suggesting possible regulation of CHI expression according to developmental stage (Rose et al., Reference Rose2006). A CHI isolated from maize seeds strongly inhibited mycelial growth of the plant pathogenic fungi Fusarium oxysporum and Alternaria solani, and the non-pathogenic Trichoderma reesei (Huynh et al., Reference Huynh1992). However, this CHI could not inhibit the pathogens Sclerotinia sclerotiorum or Gaeumannomyces graminis, illustrating the pathogen specificity of CHI.

Plant CHIs also fulfil some roles in plant growth and development, abiotic stress response, and beneficial microbe associations that may be indirectly associated with defence (Gomez et al., Reference Gomez2002; Grover, Reference Grover2012). For example, enhancing relationships with beneficial microbes may enhance the protective capacity of the microbial consortium at the seed surface. CHIs perform these functions via hydrolysis of chitin-containing compounds including arabinogalactan and glycoproteins in plant cell walls, peptidoglycan in bacteria, and lipochitooligosaccharides, which are Nod factors produced by rhizobia (van Hengel et al., Reference van Hengel2001, Reference van Hengel, Van Kammen and De Vries2002; Dyachok et al., Reference Dyachok2002; Grover, Reference Grover2012). Plant CHIs also function in calcium storage (Masuda et al., Reference Masuda, Zhao and Mikami2015), which may increase CHI stability, or provide a calcium reservoir to utilize for seed protection mechanisms (Franceschi and Nakata, Reference Franceschi and Nakata2005).

Chitinase

Fungal CHIs have not been classified as thoroughly as plant CHIs, but analysis of 25 fungal genomes shows they belong exclusively to GH18 (Seidl, Reference Seidl2008; Hamid et al., Reference Hamid2013) and are contained within classes III and V. Phylogenetic analyses resulted in the division of GH18 fungal CHIs into subgroups A, B and C (Seidl et al., Reference Seidl2005), based on differences in the structure of their substrate-binding site and their carbohydrate-binding modules. CHI class V (exochitinase) and class III (endochitinase) are contained within subgroups A and B, respectively. Subgroup C contains a newly discovered group of CHIs not previously identified in fungi, but predicted to be related to class V exochitinases (Seidl, Reference Seidl2008; Hartl et al., Reference Hartl, Zach and Seidl-Seiboth2012). Carbohydrate-binding modules, found in CHI B and C subgroups and other glycosidases, allow enzymes to bind substrates tighter and as such, enhance enzyme efficiency (Eijsink et al., Reference Eijsink2008; Hartl et al., Reference Hartl, Zach and Seidl-Seiboth2012; Paës et al., Reference Paës, Berrin and Beaugrand2012). Fungal CHIs are typically 30–60 kDa in size, but subgroup C variants range from 120 to 200 kDa (Hartl et al., Reference Hartl, Zach and Seidl-Seiboth2012).

Fungal CHIs catalyse the same basic reaction as plant CHIs, namely the hydrolysis of chitin, a polysaccharide of β-1,4 linked N-acetylglucosamine units (Hartl et al., Reference Hartl, Zach and Seidl-Seiboth2012). Unlike plants, fungal cell walls are composed of chitin; consequently, they utilize CHI not only for degradation of exogenous chitin, but also for endogenous roles including growth, development and morphogenesis (Hartl et al., Reference Hartl, Zach and Seidl-Seiboth2012; Hamid et al., Reference Hamid2013). Fungal CHIs function in nutrient acquisition, defence, mycoparasitism and entomopathogenesis (Seidl, Reference Seidl2008; Hamid et al., Reference Hamid2013; Langner et al., Reference Langner2015). All fungi contain CHI, but the number of different CHI genes varies widely, as it depends partially on the fungal growth form (Latgé, Reference Latgé2007; Hartl et al., Reference Hartl, Zach and Seidl-Seiboth2012) and is correlated to the chitin content of fungal cell walls. For example, cell walls of yeast-like species have only 0.5 to 5% chitin and contain few CHI genes, as seen in the model yeast Schizosaccharomyces pombe which contains a single CHI gene (Karlsson and Stenlid, Reference Karlsson and Stenlid2008; Kubicek et al., Reference Kubicek2011). In contrast, filamentous fungi cell walls consist of ± 20% chitin and contain an average of 15 unique CHI-encoding genes, such as the 27 in Fusarium oxysporum (Seidl, Reference Seidl2008). Mycoparasitic or entomopathogenic fungi tend to have exceedingly high numbers of CHI genes, needing this diverse arsenal to degrade their chitin-containing hosts (Gao et al., Reference Gao2011; Kubicek et al., Reference Kubicek2011). For example, the mycoparasitic fungi Trichoderma virens and T. atroviride contain the greatest numbers of chitinolytic enzymes of any fungi, with 36 and 29, respectively (Kubicek et al., Reference Kubicek2011).

Specific knowledge about the genetic regulation and biochemical mechanisms of fungal CHIs, especially as they pertain to plant pathogenesis, remains sparse (Langner and Göhre, Reference Langner and Göhre2016), although recent fungal genome sequencing has improved our understanding (Seidl, Reference Seidl2008; Hartl et al., Reference Hartl, Zach and Seidl-Seiboth2012). The high number of unique fungal CHIs reflects the diversity of roles they fulfil in fungi, but characterizing the function of individual CHI enzymes proves challenging because the CHI genes exhibit pronounced redundancy (Langner et al., Reference Langner2015). The physiological function of individual CHIs influences how they are regulated. Housekeeping CHIs involved in cell wall maintenance are constitutively expressed, whereas those involved in pathogenesis or specific morphogenic processes are spatially and temporally regulated (Langner and Göhre, Reference Langner and Göhre2016). Studies that characterize specific fungal CHIs or analyse CHI gene expression and regulation do so in relation to mycoparasitism, entomopathogenesis, or transgenic plants expressing fungal CHI genes. Very little research exists on fungal CHIs in relation to plant pathogenesis, and virtually no research has been conducted on assessing fungal CHI as it pertains to seed decay. Preliminary studies suggest that when dormant wild oat or wheat caryopses are incubated on Fusarium isolate F.a.1, the fungal mycelia produce approximately 2.6 times greater activity levels of β-N-acetylglucosaminidase, compared with mycelia from the control F.a.1 plates that do not contain caryopses (Fuerst et al., Reference Fuerst2018). While fungi do not theoretically require CHI to degrade chitin-free seed tissue, fungal CHIs may impart a competitive advantage to fungi by aiding their attack of beneficial fungi associated with seeds or other competing fungal pathogens in the seed microbiome (Dalling et al., Reference Dalling2011). Moreover, upregulation of specific fungal CHIs may enhance fungal cell wall plasticity, thereby hastening filamentous growth which represents the first phase of pathogenic development.

Protease

Proteases catalyse the hydrolytic cleavage of peptide bonds in proteins, yielding peptides and free amino acids. Fungi produce an extensive variety of proteases, also known as peptidases, proteolytic enzymes and peptide hydrolases. Given the enormous structural and functional diversity of proteases, they are classified at several levels and into numerous groups. Proteases are divided into nine families based on the functional group present at the catalytic domain and each is denoted by a capital letter: aspartic (A), cysteine (C), glutamic (G), metallo (M), asparagine (N), mixed (P), serine (S), threonine (T), or unknown (U) (Monod et al., Reference Monod2002; Rawlings et al., Reference Rawlings, Barrett and Finn2016). Proteases are additionally classified by the hydrolytic cleavage site (endo- or exo-), their pH optima (acid, neutral or alkaline), and clans based on phylogenetic associations (Rao et al., Reference Rao1998). Serine proteases constitute the most well-studied protease family, comprising over 33% of all identified proteases (Yike, Reference Yike2011). Extracellular fungal proteases concentrate in the serine family (Chandrasekaran et al., Reference Chandrasekaran2016), which includes the common subtilisin and trypsin types (Kudryavtseva et al., Reference Kudryavtseva2013). Alkaline proteases contain either a serine or metallo- centre (Sharma et al., Reference Sharma2017). Subtilisin proteases are also sometimes referred to as alkaline proteases (Kumar and Takagi, Reference Kumar and Takagi1999).

Optimum pH, temperature, ionic strength and substrate to achieve peak activity vary within and between classification groups (Sharma et al., Reference Sharma2017). For example, an alkaline thiol protease from Botrytis cinerea shows maximum activity at pH 10–11 and ~60°C (Abidi et al., Reference Abidi, Limam and Marzouki2007), whereas peak activity for an alkaline serine protease identified in Fusarium culmorum is at pH 8.3–9.6 and 50°C (Pekkarinen et al., Reference Pekkarinen, Jones and Niku-Paavola2002). Moreover, Pekkarinen et al. (Reference Pekkarinen, Jones and Niku-Paavola2002) assayed a serine protease using five different substrates and the activity levels ranged from 0.2 to 1360 nkat (mg protein)^–1. Research on the model fungus Aspergillus nidulans shows that extracellular fungal protease production is regulated by carbon, nitrogen, sulphur, pH, and in certain circumstances, exogenous proteins (Yike, Reference Yike2011).

Proteases are important enzymes secreted by plant pathogenic fungi that function in signalling, nutrition, plant host tissue degradation, sporulation, morphogenesis, septum formation, and plant defence enzyme degradation (Abidi et al., Reference Abidi, Limam and Marzouki2007; Yike, Reference Yike2011). Fungal proteases appear to play significant roles at certain phases of plant infection, such as host cell adhesion and initial cell wall penetration and colonization (Olivieri et al., Reference Olivieri2004; Soberanes-Gutiérrez et al., Reference Soberanes-Gutiérrez2015; Chandrasekaran et al., Reference Chandrasekaran2016). Fungal proteases function in plant pathogenesis by activating or inactivating plant proteins such as plant defence enzymes, influencing autolysis, and by increasing plant plasma membrane permeability (Chandrasekaran et al., Reference Chandrasekaran2016). Fungal metalloproteases, notably Zn-metalloproteases, are known to function in plant pathogenicity and virulence by directly degrading the plant cell wall (Staats et al., Reference Staats2013). Recent studies indicate that the significance of proteases in plant pathogenesis is influenced by the specific plant–pathogen association (Dong et al., Reference Dong2014; Figueiredo et al., Reference Figueiredo, Monteiro and Sebastiana2014; Jashni et al., Reference Jashni2015a). For example, Fusarium graminearum and F. culmorum, the fungal pathogens responsible for Fusarium head blight disease, produce alkaline serine proteases in the endosperm of infected wheat and barley grains, with the highest enzyme activity produced by the most virulent fungal strains, suggesting their role in pathogenicity (Nightingale et al., Reference Nightingale1999; Pekkarinen et al., Reference Pekkarinen2003).

Numerous examples exist of fungal pathogens that escape plant defence CHIs by secreting proteases to degrade or otherwise modify plant CHI enzymes, especially those containing a carbohydrate-binding module, thereby nullifying their ability to serve a defensive role (Jashni et al., Reference Jashni2015b). This behaviour has been widely demonstrated by numerous Fusarium spp., including Fusarium verticillioides, F. oxysporum, F. graminearum, F. proliferatum and F. subglutinans that secrete Zn-metalloproteases capable of degrading three unique class IV carbohydrate-binding module-CHIs in maize (Naumann et al., Reference Naumann, Wicklow and Price2011; Slavokhotova et al., Reference Slavokhotova2014). Fusarium proteases with anti-CHI activity are also seen in F. solani f. sp. phaseoli that modifies CHI in bean to promote fungal colonization (Lange et al., Reference Lange1996); F. oxysporum f. sp. lycopersici that modifies a carbohydrate-binding module-CHI in tomato (Jashni et al., Reference Jashni2015a); and a subtilisin serine protease from F. solani f. sp. eumartii that modifies CHI present in potato tubers (Olivieri et al., Reference Olivieri2002). Interestingly, cleavage efficiency and specificity for different CHI enzymes vary widely among different Fusarium proteases (Naumann et al., Reference Naumann, Wicklow and Price2011; Jashni et al., Reference Jashni2015a). Moreover, different proteases cooperate to enhance enzyme degradation. For example, synergistic effects were evident when metallo and serine proteases acted together to degrade CHI in tomato (Jashni et al., Reference Jashni2015a) and it is hypothesized that fungi simultaneously secrete exo- and endoproteases for synergistic functioning (Girard et al., Reference Girard2013). Additional fungi display protease activity against plant CHI which ultimately leads to increased fungal virulence, including Bipolaris zeicola, Cochliobolus carbonum, Stenocarpella maydis, Botrytis cinerea and Verticillium dahliae (Naumann et al., Reference Naumann, Wicklow and Kendra2009; Naumann and Wicklow, Reference Naumann and Wicklow2010; Jashni et al., Reference Jashni2015a).

Seeds also produce inhibitors to protect against fungal proteases and these are abundant in cereal grains (Pekkarinen and Jones, Reference Pekkarinen and Jones2003). For example, wheat and barley seeds produce multiple inhibitors in response to the subtilisin- and trypsin-like proteases produced by Fusarium graminearum and F. culmorum during Fusarium head blight infection (Pekkarinen et al., Reference Pekkarinen2003). Three inhibitors were likewise isolated from dormant buckwheat (Fagopyrum esculentum) seeds that inhibited trypsin-like proteases from Fusarium oxysporum and Alternaria alternata, common phytopathogenic fungi, and suppressed their mycelial growth in vitro (Dunaevskii et al., Reference Dunaevskii1995). High concentrations of protease inhibitors were identified in the surface proteome of mature barley seeds (Sultan et al., Reference Sultan2016).

Xylanase

Xylanases are hydrolytic enzymes that degrade xylan, which is a general term to describe a variety of hemicellulosic polysaccharides of d-xylose linked by β-1,4-bridges which is abundant in cell walls of commelinid monocots (Collins et al., Reference Collins, Gerday and Feller2005; Polizeli et al., Reference Polizeli2005; Hatsch et al., Reference Hatsch2006). Xylans can be further classified according to the substitutions present along the main xylose backbone. The predominant hemicellulose of endosperm in cereal grains is arabinoxylan (McCleary et al., Reference McCleary2015). Endo-1,4-β- d-xylanase (endoxylanase) is one of multiple enzymes required for the complete degradation of xylan. It catalyses the hydrolytic cleavage of β-xylosidic bonds in xylose (Collins et al., Reference Collins, Gerday and Feller2005). Xylanases are produced by various organisms, but predominantly by microorganisms, and fungal-derived xylanases show especially high activity (Polizeli et al., Reference Polizeli2005).

Xylanases are a diverse group of enzymes, exhibiting differences in catalytic domain, optimum pH and temperature, substrate specificity, and catalytic efficiency (Paës et al., Reference Paës, Berrin and Beaugrand2012). Fungal xylanases are integral to plant pathogenesis and most fungi produce numerous different xylanase enzymes (Paës et al., Reference Paës, Berrin and Beaugrand2012). Fungal xylanases have inter- and intraspecific variability, and are regulated both spatially and temporally. The large number of unique xylanases stems from factors including genetic redundancy (Wong et al., Reference Wong, Tan and Saddler1988), a strategy that allows fungi to adapt to diverse plant substrates at different growth stages and in different environments. Each enzyme is probably fine-tuned to optimally degrade a specific substrate under certain environmental parameters (Paës et al., Reference Paës, Berrin and Beaugrand2012).

The majority of plant pathogenic fungi contain numerous genes encoding endoxylanases (Sella et al., Reference Sella2013). Fungal endoxylanases are present primarily in families 10 and 11 of the glycosyl hydrolase superfamily (Lombard et al., Reference Lombard2013), but they are also distributed in GH families 5, 8, 16, 26, 30, 43 and 62, albeit to a far lower extent (Collins et al., Reference Collins, Gerday and Feller2005; Paës et al., Reference Paës, Berrin and Beaugrand2012; Lombard et al., Reference Lombard2013). They were traditionally divided into GH10 and GH11 according to high and low molecular weight, respectively (Sella et al., Reference Sella2013), but with increasing discovery and characterization of xylanases, this classification system gave way to one based on similarities of the amino acid sequence of the catalytic domain (Lombard et al., Reference Lombard2013). Additional distinctions between GH10 and GH11 include substrate specificity and structure. GH10 xylanases have broad substrate specificity, are preferentially active on soluble substrates, can degrade linear chain xylans and those with multiple substitutions, and have a TIM-barrel fold at the active site (Beaugrand et al., Reference Beaugrand2004). In contrast, GH11 have high substrate specificity, can degrade insoluble substrates, cannot degrade xylan backbones with a high degree of substitution, and have a highly conserved β-jelly roll structure at the active site (Pollet et al., Reference Pollet, Delcour and Courtin2010; van den Brink and de Vries, Reference van den Brink and de Vries2011; Paës et al., Reference Paës, Berrin and Beaugrand2012; Sultan et al., Reference Sultan2016). The narrow substrate specificity of GH11 xylanases may be supported by their carbohydrate-binding module, which is a distinct region in xylanases and other glycosidases (Paës et al., Reference Paës, Berrin and Beaugrand2012) that enhances substrate binding and hastens cell wall disruption (Pollet et al., Reference Pollet, Delcour and Courtin2010). Carbohydrate-binding modules also significantly increase the thermostability of the xylanase enzyme (Paës et al., Reference Paës, Berrin and Beaugrand2012). In a comparative study of a GH10 and a GH11 xylanase, the GH11 xylanase was more efficient in hydrolysing wheat bran than GH10 and it had a 2-fold greater affinity for wheat bran than the GH10 xylanase (Beaugrand et al., Reference Beaugrand2004). Interestingly, in a proteomic analysis of the barley grain surface, only GH11 fungal xylanases were detected (Sultan et al., Reference Sultan2016).

While cereal grains produce endogenous xylanases, research suggests that over 90% of the xylanase activity detected on cereal grains is of microbial origin (Dornez et al., Reference Dornez2006b). Several studies show that microbial xylanases concentrate in the outer bran layers of cereal grains rather than in the endosperm, as seen in a comprehensive survey of common wheat, durum wheat, spelt, einkorn, emmer, barley, rye and oat (Gys et al., Reference Gys2004; Dornez et al., Reference Dornez2006a; Gebruers et al., Reference Gebruers2010). Moreover, cereal grains with hulls, such as oat, generally contain higher levels of fungal xylanases compared with hull-less varieties because the space between the hull and caryopsis provides a niche for microbial colonization (Noots et al., Reference Noots, Delcour and Michiels1999). For example, fungal mycelia covered an average of 59.2% of the lemma and 70.2% of the palea of barley grain (Warnock, Reference Warnock1971).

Proteomic analysis of the barley grain surface identified numerous fungal enzymes that function in plant cell wall degradation and are also required for virulence, including β-1,4-xylanase (Sultan et al., Reference Sultan2016). Numerous xylanases specifically identified in the grain washing liquids came from the following fungi: (1) the soil-borne pathogen V. dahlia (contributed two different xylanases), (2) the highly virulent cereal grain pathogen Cochliobolus sativus, and (3) the broadly specific grass pathogen Pyrenophora tritici-repentis (Sultan et al., Reference Sultan2016). Genomic analysis of F. graminearum, the causative agent of Fusarium head blight, indicates that it contains 10 xylanase-encoding genes, six of which are expressed during infection of hop and barley, and five during wheat infection (Güldener et al., Reference Güldener2006; Hatsch et al., Reference Hatsch2006; Sella et al., Reference Sella2013).

Knowledge of the mechanisms underlying the interaction between fungal xylanases and plant cells remains sparse, yet it is known that plant cells have defences to fungal xylanases by means of protein inhibitors. For example, three classes of inhibitors (XIP, TAXI and TLXI) have been identified in several cereal grains such as barley, wheat and rye (Elliott et al., Reference Elliott2003; Goesaert et al., Reference Goesaert2003; Dornez et al., Reference Dornez2010) that differ in their structure and xylanase specificity (Paës et al., Reference Paës, Berrin and Beaugrand2012). XIP-I is concentrated in the xylan-rich nucellar tissue within the intermediate layer of wheat bran, illustrating its importance in protecting the inner seed fraction from fungal attack (Jerkovic et al., Reference Jerkovic2010). The endosperm of cereal grains has also been identified as a region of concentrated xylanase inhibitors. For example, the flour produced from debranned wheat kernels showed no significant reduction in xylanase inhibitor proteins compared with wholemeal flour (Gys et al., Reference Gys2004), and strong inhibition of the xylanases produced by Aspergillus niger and Bacillus subtilis was seen in the flour fraction of various dormant European wheat varieties (Gebruers et al., Reference Gebruers2002).

Secreted during the early infection stages of cereal crops including wheat, barley and rye, endoxylanases degrade cell walls in seeds and leaves; however, the exact role they play in virulence or pathogenicity is not well known. A xylanase from Botrytis cinerea is required for virulence in grape berries, but its contribution to infection resides in its ability to induce tissue necrosis, and not in its catalytic hydrolysing activity (Brito et al., Reference Brito, Espino and González2006; Noda et al., Reference Noda, Brito and González2010). Mutated xylanases from Trichoderma reesei with enzymatic activity decreased 100- or 1000-fold elicited a defence response of hypersensitive necrosis in tomato and tobacco leaves equivalent to the wild-type xylanase, indicating that the enzymatic function of xylanases is not necessarily required for fungal virulence (Enkerli et al., Reference Enkerli, Felix and Boller1999). Numerous Fusarium graminearum isolates with a mutated xylanase gene displayed only 40% xylanase activity 4 and 7 DAI compared with the wild type (Sella et al., Reference Sella2013). However, 21 DAI of wheat spikes with these mutants, there was no significant reduction in necrosis symptoms compared with the wild type, indicating that reduction in this level of xylanase activity was not required for virulence. Fusarium graminearum xylanase in the wheat lemma tissues, which are high in arabinoxylans, resulted in accumulated hydrogen peroxide deposits under the epidermal tissues and localized tissue necrosis, suggesting that this fungal xylanase not only degraded plant cell walls, but also triggered acute cell death (Sella et al., Reference Sella2013).

The activities of microbial xylanases are strongly influenced by environmental conditions and to a lesser extent, by plant genotype; in contrast, environment minimally influences the activity of xylanase inhibitors in seeds (Dornez et al., Reference Dornez2008; Gebruers et al., Reference Gebruers2010; Sultan et al., Reference Sultan2016). In an assessment of xylanase and xylanase inhibitor activity among more than 200 cereal varieties that included winter wheat, spring wheat, durum, einkorn, emmer, spelt, barley, rye and oat, variability in xylanase activity expressed in the grain was attributed to environmental conditions (50%), genotype (11–14%), and environment–genotype interactions (34–39%) (Gebruers et al., Reference Gebruers2010). Similarly, in a comparison of surface proteomes of barley cultivars grown in different years, Sultan et al. (Reference Sultan2016) concluded that microbial xylanases associated with barley grains are strongly affected by environmental conditions, due to the significant influence of environment on the grain microbial consortia; meanwhile, expression of xylanase inhibitors was more stable and not strongly affected by environment.