Introduction

The cell wall is the outermost part of fungal cells which is in constant contact with the environment. It primarily consists of polysaccharides (mainly glucans and chitin), which account for ~90 % of the walls dry weight (Latgé 2007). Both rigidity and plasticity of the cell wall are essential in maintaining normal growth and survival of the fungus. In recent years, cell wall-remodeling enzymes attracted the interest of basic and applied researchers as their role in biotechnological applications, e.g., biofuel production from plant or fungal biomass broadens (Eijsink et al. 2008).

Chitin is the most abundant amino polysaccharide and after cellulose the second most abundant biopolymer in the biosphere (Hours and Gortari 2013). It is composed of β-1-4 linked N-acetylglucosamine (GlcNAc) residues forming a linear homopolymer. As an essential scaffold material chitin is found in the exoskeleton of Ecdysozoa and in fungal cell walls (Bowman and Free 2006; Merzendorfer 2003), where it provides mechanic stability. Chitin remodeling is carried out by an interplay of membrane-associated chitin synthases (CHS; EC 2.4.1.16), which catalyze the de novo synthesis of chitin from UDP-GlcNAc (Roncero 2002; Lenardon et al. 2010), and chitin hydrolyzing enzymes (Latgé 2007). Chitinolytic enzymes include lytic polysaccharide monooxygensases (LPMOs) of the auxiliary activity 10 family (AA10; EC N/A), N-acetyl-glucosaminidases (EC 3.2.1.52), and chitinases (EC 3.2.1.14; for a classification of carbohydrate active enzymes, see www.cazy.org; Lombard et al. 2014).

Fungal chitinases exclusively belong to the glycosyl hydrolase (GH) 18 family, which is further sub-classified in three groups and five classes (I–V) according to their domain architecture and sequence homology (Henrissat 1991; Patil et al. 2000). Although GH18 chitinases share a common catalytic mechanism, they can either have endo- or exo-acting character, i.e., they hydrolyze at random positions in the polymer or degrade it from one terminus, respectively (Horn et al. 2006).

The number of chitinases encoded in fungal genomes is highly variable and ranges from only one GH18 family member in the yeast Schizosaccharomyces pombe to more than 30 in mycoparasitic Trichoderma spp. (Gruber and Seidl-Seiboth 2011). Most of our current knowledge about the physiological function of fungal chitinases has been gained from studies on model species of the phylum Ascomycota, e.g., Saccharomyces cerevisiae, Candida albicans, Neurospora crassa, Ashbya gossipii, and Trichoderma harzianum (Kuranda and Robbins 1991; Dünkler et al. 2005; Tzelepis et al. 2012; Dünkler et al. 2008; Carsolio et al. 1994), but insights into their role in Basidiomycota are still scarce. In this review we summarize the physiological roles of fungal chitinases and highlight recent findings in Ustilago maydis as a representative of plant pathogenic basidiomycetes. We also address questions that go beyond the current knowledge and concern possible roles of fungal chitinases in plant/fungus interactions.

Chitin synthesis

To maintain plasticity, fungal cell walls need continuous remodeling, which involves synthesis, cross-linking, and degradation of cell wall polysaccharides (reviewed in Teparić and Mrša 2013). The enzymatic repertoire which maintains these processes is conserved across the fungal kingdom (reviewed in Ruiz-Herrera et al. 2002). Additionally, a regulatory pathway for cell wall maintenance in response to stress, the cell wall integrity pathway, is well-described in S. cerevisiae (reviewed in Levin 2011).

De novo chitin synthesis is catalyzed by membrane-associated chitin synthases (reviewed in Roncero 2002 and Lenardon et al. 2010). The number of chitin synthase genes in fungi ranges from only one member in S. pombe to eight in the filamentous ascomycetes Aspergillus fumigatus and A. nidulans or the basidiomycetes Cryptococcus neoformans and U. maydis (Lenardon et al. 2010).

Several chitin synthases have already been functionally characterized. They act in various processes such as septum formation, cell division, polar growth, stress response, or pathogenic development (Ichinomiya et al. 2005; Silverman et al. 1988; Kim et al. 2009; Weber et al. 2006; Treitschke et al. 2010; Cui et al. 2013). Interestingly, many chitin synthases are of zymogenic nature and localize to the plasma membrane at sites of polar growth where they are further processed into their active form (Starr et al. 2012; Valdivia and Schekman 2003; Bowen et al. 1992; Durán et al. 1975). The delivery to these sites involves the microtubule and actin cytoskeleton, motor proteins like kinesin and myosin, secretory vesicles or so called chitosomes, and in some cases an N-terminal myosin-like motor domain (Sietsma et al. 1996; Takeshita et al. 2005; Schuster et al. 2012; Egan et al. 2012). Once associated with the plasma membrane, chitin synthases catalyze the vectorial synthesis of the chitin chains from activated UDP-GlcNAc monomers, releasing the nascent polymer to the extracellular space. There, single chitin chains associate with each other through intramolecular interactions forming chitin microfibrils and covalent crosslinking to cell wall glucans occurs (Cabib and Arroyo 2013; Cabib et al. 2007; Teparić and Mrša 2013). This crosslinking is maintained by transglycosylases or by intrinsic transglycosylation activity of chitinases, as described for Chit42 and Chit33 from T. harzianum or ChiB1 from A. fumigatus as well as for bacterial chitinases (Zakariassen et al. 2011; Hartl et al. 2012).

The hierarchy of chitin degradation

The breakdown of polymeric chitin in the fungal cell wall is mediated by a number of chitinolytic enzymes, including LPMOs, chitinases, and N-acetylglucosaminidases, which work in concert for efficient, but tightly regulated degradation (Fig. 1). LPMOs and endo-acting GH18 chitinases insert strand breaks at random positions within the chitin chains. The exact mechanism how LPMOs insert strand breaks is not understood, but accumulating evidence shows that it involves binding of a copper ion by the imidazole and the peptide-backbone amino group of a conserved N-terminal histidine and the imidazole of a further downstream histidine residue (Nakagawa et al. 2015; Vaaje-Kolstad et al. 2005). The catalytic mechanism also involves electron donors such as ascorbic acid or reduced glutathione (Vaaje-Kolstad et al. 2010). In the current model, the copper ion enables transport of an electron to molecular oxygen, thereby creating a superoxo-intermediate which then initiates cleavage of the polymer. By contrast, the catalytic mechanism of chitinases is well-characterized. To insert strand breaks, GH18 chitinases employ a neighboring group participation mechanism, in which the N-acetamido carbonyl oxygen serves as nucleophile (Terwisscha van Scheltinga et al. 1995). A conserved glutamate in the catalytic center (DxxDxDxE) serves as general acid/base and protonates the leaving group, thereby creating an oxazolium intermediate which, after break down, leads to hydrolysis of the polymer (van Aalten et al. 2001). During hydrolysis the anomeric configuration of the substrate is retained.

Fig. 1
figure 1

The hierarchical machinery of chitin degradation. In fungal cell walls polymeric chitin (purple hexagons) is densely packed into higher ordered microfibrils. Lytic polysaccharide monooxygenases (LPMOs) and endo-acting GH18 chitinases insert strand breaks at random positions within the polymer, thereby generating reducing (RE) and non-reducing ends (NRE) and loosening up the packed substrate. LPMOs use an oxidative mechanism, while GH18 chitinases use a neighboring group participation mechanism. Exo-acting GH18 chitinases subsequently cleave chito-oligosaccharides (CHOS), predominantly (GlcNAc)2, from either the reducing or the non-reducing ends also using a neighboring group participation mechanism. These CHOS serve as substrate for N-acetylglucosaminidases of the GH20 family which cleave GlcNAc monomers from the non-reducing end

Full size image

Cleavage of polymeric chitin by LMPOs or endo-acting chitinases generates free reducing and non-reducing ends. Such free ends serve as targets for processive GH18 chitinases, which employ the same catalytic mechanism as endo-acting chitinases (Vaaje-Kolstad et al. 2010, 2012). Notably, the increased processivity depends on the architecture of the binding cleft, which is “tunnel-shaped”, as opposed to a shallow and rather “open” architecture in non-processive, endo-acting chitinases (Horn et al. 2006; Vaaje-Kolstad et al. 2013). Upon cleavage, processive chitinases release short chito-oligosaccharides (CHOS), mainly (GlcNAc)2 and (GlcNAc)3, from either the reducing or non-reducing end without dissociating from the substrate, leading to the complete decomposition of the strand (Horn et al. 2006; Zakariassen et al. 2009).

CHOS, mainly (GlcNAc)2, are the preferred substrate for N-acetylglucosaminidases of the GH20 family which catalyze the further degradation to GlcNAc from the non-reducing end (Langner et al. 2015; Chen et al. 2015). Similar to chitinases they use a substrate assisted, retaining mechanism (Drouillard et al. 1997; Jones and Kosman 1980) in which a conserved glutamate of the catalytic diad (D–E) serves as general acid/base. After protonation the aspartate might stabilize the oxazolium intermediate during hydrolysis (Tews et al. 1996; Williams et al. 2002).

Remarkably, the chitinolytic machinery shares similarities with the enzymatic repertoire which is used for the breakdown of other complex biopolymers such as cellulose (Beeson et al. 2015; Igarashi et al. 2011; Phillips et al. 2011). Therefore mechanistic insights into the mode of action, localization, and regulation of chitinolytic enzymes might also contribute to the understanding of these other enzymatic systems. In the future, this knowledge may even be applied to drive forward industrial applications, like, e.g., the efficient degradation of plant and fungal biomass for biofuel production (Eijsink et al. 2008).

Physiological functions of fungal chitinases

The biological functions of chitinases are very diverse and include roles in yeast and filament morphogenesis, autolysis, acquisition of chitin for nutritional purposes, and mycoparasitism (Kuranda and Robbins 1991; Shin et al. 2009; Yamazaki et al. 2007; Leake and Read 1990; Gruber et al. 2011). In general, the number of chitinase genes is lower in yeasts as compared to filamentous fungi. S. pombe possesses just one, S. cerevisiae only two chitinases (Karlsson and Stenlid 2008). By contrast, filamentous fungi can contain expanded chitinase gene families. The filamentous ascomycetes A. nidulans and A. fumigatus have 19 and 20 genes encoding GH18 enzymes, respectively, and huge gene families are described for mycoparasitic species like T. virens with 36 chitinase genes (Gruber and Seidl-Seiboth 2011). Due to a high degree of functional redundancy within the GH18 family, a major challenge in chitinase research is to assign specific physiological functions to single members of this family. Often it is necessary to investigate multiple deletion mutants to observe phenotypic alterations. Notably, dimorphic fungi, like the ascomycete C. albicans (four chitinases; Dünkler et al. 2005) or the plant pathogenic basidiomycete U. maydis (three chitinases; Kämper et al. 2006; Koepke et al. 2011) often encode rather small numbers of chitinase genes which facilitates functional characterization during both, yeast-like and filamentous growth.

The role of chitinases during yeast-like growth

First knowledge about the biological function of chitinases was gained in early studies in the model yeast S. cerevisiae. The chitinase ScCts1p is essential for chitin degradation during cell division and deletion of the CTS1 gene leads to a defect in cell separation and the formation of multicellular aggregates (Kuranda and Robbins 1991). During cell cycle ScCts1p dynamically localizes to the mother-daughter neck, where it contributes to the degradation of the primary septum (Fig. 2a; Colman-Lerner et al. 2001). Similarly, the ScCts1p homolog of the yeast Kluyveromyces lactis (Colussi et al. 2005) and CaCht3p of the opportunistic pathogen C. albicans are required for cytokinesis. In complementation studies Dünkler et al. (2005) showed, that CaCht3p indeed is the functional homolog of ScCts1p. Similar phenotypic observations were made in S. pombe. However, here the primary septum is mainly composed of glucan, which is degraded by the glucanases SpAgn1 and SpEng1 (Martin-Cuadrado 2003; García et al. 2005). Consequently, deletion of the glucanase genes leads to defects in cell separation, reminiscent of chitinase deletions in other yeasts. By contrast, the deletion of single chitinase genes in the dimorphic smut U. maydis had no effect on yeast cell morphogenesis or cell separation (Koepke et al. 2011). Only double deletions of the chitinase genes cts1 and cts2 cause severe defects in cell separation suggesting redundant functions in this growth stage. Both chitinases contribute to the degradation of remnant chitin in an extracellular division zone between the primary and secondary septum, whereas septum formation remains unaffected (Langner et al. 2015). Notably, leaderless UmCts1 dynamically localizes to the fragmentation zone in a cell cycle dependent, daughter cell-specific manner during late stages of cell division (Fig. 2b, c), while UmCts2 contains a signal peptide for secretion (Langner et al. 2015). Intriguingly, these results indicate that functional redundancy even exists between UmCts1 and UmCts2, two chitinases which employ different secretory mechanisms.

Fig. 2
figure 2

Spatio-temporal regulation of chitinases in S. cerevisiae and U. maydis. a In S. cerevisiae CTS1 is expressed in the mother and the budding daughter cell during early stages of cell division. As the daughter cell continues budding CTS1 expression becomes downregulated. In late mitotic stages CTS1 expression is restricted to daughter cells by the concerted action of ScCbk1p and ScAce2p, both constituents of the RAM network. The ScCts1p protein eventually localizes to the daughter cell site of the primary septum (red arrowhead) facilitating cell separation. b In U. maydis, the chitinase UmCts1 is equally distributed in both, the mother and daughter cell early in cytokinesis. After insertion of the primary septum (red arrowhead) by the mother cell UmCts1 dynamically localize to the division zone between mother and daughter cell. This relocalization is exclusively maintained by the daughter cell. Upon insertion of the secondary septum (blue arrowhead) by the daughter cell, UmCts1 accumulates in the fully established division zone to degrade remnant chitin and facilitate cell separation. c Co-localization of UmCts1:eGfp (green) and cell membrane stained with the lipophilic dye FM4-64 (red) in dividing U. maydis cells. After accumulation of membrane material on the mother side of the division zone (red arrowhead) UmCts1 relocalization starts within the daughter cell. By the insertion of the secondary septum and the daughter cell membrane (blue arrowhead) Cts1 is delimited within the division zone

Full size image

During sporulation, the second S. cerevisiae chitinase, ScCts2p, is involved in asci formation (Giaever et al. 2002). Similarly, the chitinase AgCts2 in A. gossypii plays an important role in spore formation (Dünkler et al. 2005, 2008). The sporulation defect caused by the deletion of cts2 can be complemented by inserting CTS2 from S. cerevisiae or CHT4 from C. albicans, indicating a comparable function in the sporulation process (Dünkler et al. 2008). By contrast, in U. maydis teliospore formation is independent of chitinase activity, since quadruple mutants lacking all chitinase activity are able to produce pigmented teliospores (Langner et al. 2015).

A different, cell cycle unrelated function of chitinases has been reported in the opportunistic basidiomycete C. neoformans. Deletion mutants of the whole GH18 gene family comprising four members revealed none of the above described phenotypes. Instead, the mutants display reduced tolerance against the cell wall stress inducing agent SDS and defects in mating suggesting a role of these chitinases in cell wall maintenance during yeast-like growth (Baker et al. 2009). In C. albicans cell wall stress resistance is enhanced during biofilm formation. An extracellular matrix composed of polysaccharides and N-acetylglucosamine is produced to protect the cells against antimicrobial compounds (Mathé and van Dijck 2013). It is conceivable that chitinases might also be involved in the conversion of chitin to contribute sugars to this matrix. However, the role of chitinases in tolerance against cell wall stress seems not to be conserved in other model systems. In line with this, a comprehensive genetic analysis in U. maydis could not reveal a function of chitinases under various cell wall stress inducing conditions (Langner et al. 2015).

Taken together, chitinases play diverse, but important roles in cell wall remodeling during different stages of the fungal life cycle influencing cell wall integrity, cell separation, mating, and stress resistance. However, the physiological relevance of these functions remains to be elucidated.

The role of chitinases during filamentous growth

Functional characterization of chitinases in filamentous fungi is often difficult because of expanded gene families and functional redundancy. Nevertheless, a number of studies addressed this issue and functions of chitinases in hyphal growth, autolysis, and mycoparasitism were described (reviewed in Hartl et al. 2012; Adams 2004). During hyphal growth extension of the filament is supposed to involve controlled cell wall remodeling. Consistently, in N. crassa the deletion of the chitinase gene chit-1 leads to a reduction in hyphal growth suggesting that cell wall-associated NgChit-1 has a plasticizing function (Tzelepis et al. 2012; Maddi et al. 2009). Although in A. nidulans AnChiA localizes to polar growth sites of germ tubes and filaments, pointing at an involvement in polar growth, no function in filamentous growth could be observed in the deletion mutant, likely due to redundancy within the chitinase family (Yamazaki et al. 2008). By contrast, ChiB single deletions mutants have drastically reduced chitinolytic activity and show enhanced aggregation in the stationary phase, indicating a role during autolytic degradation of the cell wall (Yamazaki et al. 2007).

In the phytopathogenic basidiomycete U. maydis, mating of compatible partners on the plant surface induces filamentous growth (Feldbrügge et al. 2004). The filament represents the virulent form which is able to form appressoria and penetrate its host plant Zea mays. During this stage of the life cycle septa are inserted at the basal pole of the filament leading to the formation of empty sections. Interestingly, deletion mutants of cts1 tend to aggregate at their basal poles leading to fast sedimentation in liquid culture, which indicates a role in autolytic chitin degradation at the empty sections of filaments (Koepke et al. 2011). This assumption was substantiated by co-localization studies revealing specific localization patterns of UmCts1 at the cell wall and at septa of empty sections (Langner et al. 2015). Despite that basal localization, UmCts1 shows a gradual increase towards the growing pole which suggests additional functions in polar growth, but deletion mutants did not show altered filament formation. Noteworthy, UmCts2 which acts redundantly in yeast cells does not contribute to the overall chitinolytic potential in filaments, suggesting specific downregulation of this chitinase at the passage from saprotrophic to biotrophic growth (Langner et al. 2015).

In sum, in filamentous growing fungi chitinases seem to have only minor effects on the establishment of unipolar growth, possible due to an intimate regulation of chitin synthesis and degradation. Nevertheless, chitinases are important in autolytic processes that might facilitate growth under non-optimal conditions, such as starvation. Furthermore, chitinases might also be involved in the defense against competing fungi, and thus could contribute to overall fitness in complex environments such as soil or plant microbiomes.

Regulation of chitinases

Depending on their physiological function fungal chitinases differ in their regulation. Housekeeping chitinases are thought to be involved in continuous remodeling of the cell wall and are therefore constitutively expressed. By contrast, specialized chitinases which function in degrading exogenous chitin or during specific growth processes, e.g., cell division/separation or autolysis, require spatio-temporal regulation of chitinase activity.

A comprehensive gene expression analysis of twelve GH18 chitinases in N. crassa showed selective induction under varying conditions, while phenotypes were only detected in two of the single deletion mutants (Tzelepis et al. 2012). gh18-10 is constitutively expressed under the tested growth conditions and its deletion results in reduced growth as expected for a housekeeping chitinase. Similarly, the chitinase genes gh18-1 and gh18-2 are constitutively expressed albeit at much lower level, also indicating general functions in cell wall remodeling during hyphal growth, but without any apparent deletion phenotype. In this case mutants with multiple deletions might be necessary to reveal underlying functions. chit-1 expression is upregulated in media containing colloidal chitin as the major carbon source, pointing towards a function in nutritional chitin acquisition in addition to its role in hyphal growth. Four chitinase genes are specifically induced during competitive interaction with Fusarium sporotrichoides, although the biological relevance of this data remains elusive. In mycoparasitic fungi such as Clonostachys rosea and T. atroviride it was shown that the chitinases ech42 and ech37 are induced in the presence of exogenous chitin, N-acetylglucosamine (NAG), and fungal cell walls (Mach et al. 1999; Seidl et al. 2005; de las Mercedes Dana et al. 2001), and a direct role in mycoparasitism was suggested for TaEch42 (Mamarabadi et al. 2008; Carsolio et al. 1999). Interestingly the regulation of ech42 in Trichoderma also depends on the GH20-gene nag1 (Brunner et al. 2003) indicating a feedback-loop in which the degradation of chitin leads to the production of NAG, which might be taken up and trigger further upregulation of chitinolytic enzymes. In addition, expression of these chitinases is regulated through carbon catabolite-repression in the presence of glucose (Lorito et al. 1996; Mamarabadi et al. 2009).

Another mechanism for spatio-temporal regulation of chitinases has been described in S. cerevisiae. As mentioned above ScCts1p is essential for cell separation after cytokinesis (Kuranda and Robbins 1991), by degrading chitin of the primary septum. Therefore Cts1 expression gets activated specifically in daughter cells during late M-phase and the protein subsequently dynamically localizes to the daughter site of the septum (Fig. 2a; Colman-Lerner et al. 2001). This tight regulation requires the concerted action of the FEAR (CDC14 early anaphase release), mitotic exit network (MEN), and regulation of Ace2p and morphogenesis (RAM) networks. During mitosis activation of FEAR and MEN leads to the release of the phosphatase Cdc14 from the nucleolus/nucleus into the cytoplasm where it counteracts cyclin-dependent kinase (Cdk1) activity to initiate cytokinesis (reviewed in Stegmeier and Amon 2004; Wurzenberger and Gerlich 2011). Furthermore, Cdc14 dephosphorylates the Cbk1p–Mob2p complex and the transcription factor Ace2p to promote their translocation into the daughter cell nucleus (for details see Sanchez-Diaz et al. 2012; Brace et al. 2010; Mancini Lombardi et al. 2013). Here the Cbk1p-Mob2p complex activates Ace2p to initiate the transcription of downstream genes, including the chitinase gene CTS1 and the glucanase gene ENG1 (Baladrón et al. 2002; Colman-Lerner et al. 2001; Nelson 2003). Thus, the integration of different regulatory networks enables cell cycle dependent, daughter cell-specific regulation of cell wall degrading enzymes.

In U. maydis two chitinases, UmCts1 and UmCts2, act redundantly in cell separation. Remarkably, the third GH18 chitinase, UmCts3, is not expressed and hence does not contribute to chitinolytic activity during yeast-like or filamentous growth suggesting a specialized function, e.g., in competition with other fungi or nutrient acquisition, which yet remains to be identified (Langner et al. 2015). The major chitinolytic activity in saprotrophic cultures is attributed to UmCts1, which is expressed during both yeast-like and filamentous growth. During cell division UmCts1 dynamically localizes to the division zone in a cell cycle dependent, daughter cell-specific manner (Fig. 2b, c; Langner et al. 2015). However, the exact mechanism which regulates UmCts1 relocalization remains elusive. Furthermore, it is yet unclear if cts1 expression is differentially regulated in daughter cells and if a regulatory network reminiscent of the RAM network exists. During filamentous growth which represents the initial phase of pathogenic development of U. maydis, extracellular UmCts1 activity is increased (Koepke et al. 2011). This increased activity is controlled by a transcriptional upregulation of cts1 expression eight to twelve hours post filament induction (Heimel et al. 2010). Intriguingly, UmCts2 does not contribute to the chitinolytic potential of U. maydis during filamentous growth. The loss of UmCts2 activity in filaments is accompanied by a transcriptional downregulation in artificially induced filaments in axenic culture or on hydrophobic surfaces resembling the situation on the host leaf (Langner et al. 2015; Lanver et al. 2014). To date the functional relevance of this downregulation of cts2 expression for polarized growth or infection of the host remains elusive. It is conceivable that UmCts2 secretion into the apoplastic space during plant-pathogen interaction might be perceived by the host, either directly or indirectly through the production of CHOS, and could therefore elicit defense responses leading to immunity.

Secretion of chitinolytic enzymes

To act on the cell wall, chitinolytic enzymes essentially have to be secreted. In eukaryotes, most of them therefore contain a signal peptide which targets the respective protein to the endoplasmic reticulum (ER) for conventional secretion (Seidl et al. 2005; Langner et al. 2015). During the passage through the ER and the Golgi apparatus they often undergo post-translational modifications, like glycosylation or proteolytic cleavage (Tanaka et al. 2005; Gentzsch and Tanner 1997; Baladrón et al. 2002; Lange et al. 1996). Such processed proteins are then packed into secretory vesicles and transported to the plasma membrane for secretion. Interestingly however, some cell wall-associated proteins do not contain classical secretion peptides (Teparić et al. 2007; Chaffin 2008; Koepke et al. 2011) indicative of an unconventional, yet unknown secretion pathway (Steringer et al. 2015; Shoji et al. 2014). For example, the leaderless chitinase Cts1 of U. maydis avoids the classic route for secretion (Stock et al. 2012; Sarkari et al. 2014). Similarly, ScCts2p of S. cerevisiae and its orthologue CaCht4p of C. albicans do not contain classical secretion signals, but are still present in the cell wall (Teparić et al. 2007; Chaffin 2008). ER avoidance could occur for two reasons: One possibility is that post-translational modifications might render the enzyme inactive. Predicted N-glycosylation sites in the leaderless chitinases ScCts2p, CaCht4p, and UmCts1 support this idea. Secondly, chitinases might interfere with the glycosylation machinery by binding to GlcNAc-containing sugars in the ER. Such endo-ß-N-acetylglucosaminidase- or peptide-N-glycosidase-activities were shown for bacterial enzymes (Murata et al. 2005; Larsen et al. 2010; Frederiksen et al. 2013). UmCts1 can also bind and process small CHOS (Langner et al. 2015) and might therefore interfere with the core sugar residues of N-glycosylated proteins in the ER similar to the bacterial enzymes. Thus, it is conceivable that a conserved alternative secretion pathway exists for secretion of chitinolytic enzymes to prevent the cell from severe damage due to false glycosylation and misfolding of proteins.

The role of chitin and chitinases in plant-fungus interactions

Phytopathogenic fungi take advantage of their host plant for completion of the life cycle. However, plants possess various defense mechanisms to counteract fungal infection. The first layer of the plant immune system involves recognition of microbe/pathogen-associated molecular pattern (MAMPs/PAMPs) to trigger basal defense responses (reviewed in Zipfel 2014; Sánchez-Vallet et al. 2014; Ben Khaled et al. 2015). Perception of MAMPs is mediated by highly conserved, extracellular transmembrane receptor proteins, called pattern-recognition receptors (PRRs) and receptor-like proteins (RLPs), which activate defense responses leading to MAMP-triggered immunity (Fig. 3; reviewed in Shinya et al. 2015; Tanaka et al. 2013). However, pathogens have evolved several mechanisms to evade the recognition and subsequent defense responses of the host in an evolutionary process often referred to as “molecular arms race” (Fig. 3; reviewed in Oliveira-Garcia and Valent 2015; Stergiopoulos and de Wit 2009).

Fig. 3
figure 3

Chitin triggered immunity in plants and evasion strategies of fungal pathogens. After entry into the host plant, the fungal cell wall is attacked by host secreted chitinases, which liberate soluble chito-oligosaccharides (CHOS) into the apoplastic space. CHOS with a chain length of n ≥ 8 are bound by the plant LysM containing receptor-like protein (RLP) CEBIP. Subsequently, the chitin elicited receptor kinase 1 (CERK1) gets recruited which in turn leads to phosphorylation of its intracellular kinase domain. Activated CERK1 subsequently phosphorylates the Rac1 specific guanine exchange factor 1 (Rac1GEF1), which in turn activates a MAPK cascade that triggers expression of defense-genes, including hydrolytic enzymes. To protect the cell wall and to avoid recognition by the host, fungal pathogens secrete effector proteins to the apoplast. One such effector is Avr4 which binds to chitin and protects the cell wall from host-derived chitinases. Avr4 in turn can be perceived by the plant receptor CF-4 inducing defense-gene expression and hypersensitive response (HR). Therefore, fungi employ alternative strategies to evade host immunity. Fungus-derived proteases are secreted into the apoplastic space to degrade host chitinases and thereby counteracting the hydrolytic attack. To avoid CHOS recognition fungi can also secrete CHOS-binding effectors, like Ecp6. Ecp6 binds to CHOS with ultra-high affinity, thereby outcompeting binding by plant receptors and preventing recognition

Full size image

Chitin or soluble CHOS are potent MAMPs that reveal the presence of fungal structures to the plant (Liu et al. 2012, 2014; Kombrink et al. 2011). CHOS binding by plant receptors is mediated through extracellular lysine motifs (LysM) containing PRRs and RLPs. The first immune receptor described in chitin reception was the chitin elicitor-binding protein (OsCEBIP) in rice (Kaku et al. 2006). Although CEBIP lacks a cytoplasmic kinase domain, it is directly involved in chitin-triggered immunity (Hayafune et al. 2014; Kouzai et al. 2013). OsCEBIP requires long CHOS, such as chito-octamers, to activate immune responses, most probably through homodimerization. Subsequently, the OsCEBIP homodimer recruits the chitin elicitor receptor kinase 1 (OsCERK1) which triggers phosphorylation of its intracellular kinase domain (Hayafune et al. 2014; Liu et al. 2012). Activated OsCERK1 phosphorylates the Rac1 guanine nucleotide exchange factor 1 (OsRacGEF1) which leads to the activation of MAPK cascades through the small GTPase Rac1 (Fig. 3; Akamatsu et al. 2013; Shimizu et al. 2010). A similar pathway was also proposed in Arabidopsis thaliana which involves interaction of AtCERK1 with the LysM RLK AtLYK5 (Cao et al. 2014). As a result, defense responses are triggered which include the production of reactive oxygen species (ROS) and the expression of defense-related proteins, such as hydrolytic enzymes (Miya et al. 2007; Su et al. 2015).

To protect their cell wall from host chitinases or to prevent CHOS-mediated receptor activation, fungi have evolved multiple strategies. In the leaf mold Cladosporium fulvum expression of the apoplastic effector CfAvr4 enhances virulence (van Esse et al. 2007). CfAvr4 is a chitin-binding lectin that is secreted during infection and protects the integrity of the fungal cell wall against plant chitinases (van den Burg et al. 2006). However, this protective mechanism is only functional in host plants which do not carry CF-4, a membrane-anchored LRR-RLP that recognizes Avr4 and subsequently triggers a hypersensitive response (HR; Fig. 3; Stergiopoulos et al. 2010). However, many successful pathogens, e.g., F. oxysporum or Verticillium dahliae do not encode homologs of avr4 in their genomes. Nevertheless, most of them are still able to suppress degradation of their cell wall by plant chitinases, suggesting alternative strategies to counteract chitinase activity.

In line with this, F. oxysporum as well as the maize pathogens F. verticillioides and Bipolaris zeicola employ such a second strategy which involves the secretion of proteases acting against antifungal plant chitinases (Jashni et al. 2015; Naumann and Price 2012; Naumann et al. 2009). These proteins, including metallo- and serine-proteases, specifically cleave plant chitinases, reducing their antifungal activity, and thus, promote virulence of the pathogen (Fig. 3; Jashni et al. 2015).

Alternatively, fungi can evade recognition by the host through the secretion of apoplastic LysM effectors, like MoSlp1 of the rice blast fungus Magnaporthe oryza or CfEcp6 of C. fulvum (Mentlak et al. 2012; de Jonge et al. 2010). CfEcp6 can bind to soluble CHOS with ultra-high affinity through intrachain dimerization of two LysM domains (Sánchez-Vallet et al. 2013). Thereby CfEcp6 is able to out-compete binding of CHOS to plant PRRs and hence suppresses chitin-triggered immunity (Fig. 3).

The above mentioned strategies to mediate and avoid chitin recognition exemplify the extent of co-evolution of virulence and defense mechanisms in plant/fungus interactions. Although huge progress has been made in the last decades, our knowledge of immunity induced by apoplastic molecules is still fragmented. How did fungal pathogens adapt to the perception and induction of immune responses by the host? Can pathogen-derived chitinases, which remodel the cell wall, also act as elicitors either through direct interaction with plant receptors/R-proteins or indirectly by releasing CHOS to the apoplast? How do phytopathogenic fungi employ their chitinases in competition with other fungi of the microbiome while protecting their own cell wall?

Concluding remarks and outlook

Investigation of plant infectious fungal model species has shed light on virulence and defense mechanisms related to chitin. However, little is known about the role of fungal chitinases during interaction with the host plant. Recently we showed that two chitinases, UmCts1 and UmCts2, are involved in cell wall remodeling in the corn smut U. maydis (Langner et al. 2015). Both enzymes differ in their binding/cleavage-site specificity, suggesting that UmCts1 might have a higher affinity against short CHOS, whereas UmCts2 might release longer soluble fragments. Interestingly, cts2 transcription is repressed during early infection stages, but the biological importance of this downregulation is still unknown. It is conceivable that UmCts2 activity needs to be repressed to prevent release of CHOS acting as MAMPs in planta. Alternatively, apoplastic UmCts2 itself might be recognized by plant receptors, and thus, trigger immunity. Conversely, it can be speculated that fungal apoplastic chitinolytic enzymes might degrade CHOS as a counter-defense strategy, although experimental evidence for this is still lacking. In this context it would be interesting to investigate the precise role of fungal chitinases during intimate contact of the pathogen and its host plant.

There is still a shortage in model pathosystems to investigate chitinase function and chitin perception by the host at the same time during biotrophic interactions. A promising model system might be the biotrophic smut fungus Thecaphora thlaspeos which infects Arabis spp. in nature. Future progress in this field will show which roles chitin and/or chitinase perception play during biotrophic interactions and to which extend co-adaptation of the pathogen and the host plant enables invasion and colonization of the host. Such insights will support our understanding of beneficial versus detrimental interactions and could support resistance breeding in the future.