Investigating chitin deacetylation and chitosan hydrolysis during vegetative growth in Magnaporthe oryzae

Summary Chitin deacetylation results in the formation of chitosan, a polymer of β1,4‐linked glucosamine. Chitosan is known to have important functions in the cell walls of a number of fungal species, but its role during hyphal growth has not yet been investigated. In this study, we have characterized the role of chitin deacetylation during vegetative hyphal growth in the filamentous phytopathogen Magnaporthe oryzae. We found that chitosan localizes to the septa and lateral cell walls of vegetative hyphae and identified 2 chitin deacetylases expressed during vegetative growth—CDA1 and CDA4. Deletion strains and fluorescent protein fusions demonstrated that CDA1 is necessary for chitin deacetylation in the septa and lateral cell walls of mature hyphae in colony interiors, whereas CDA4 deacetylates chitin in the hyphae at colony margins. However, although the Δcda1 strain was more resistant to cell wall hydrolysis, growth and pathogenic development were otherwise unaffected in the deletion strains. The role of chitosan hydrolysis was also investigated. A single gene encoding a putative chitosanase (CSN) was discovered in M. oryzae and found to be expressed during vegetative growth. However, chitosan localization, vegetative growth, and pathogenic development were unaffected in a CSN deletion strain, rendering the role of this enzyme unclear.


| INTRODUCTION
Chitin, a polymer of β-1,4-linked N-acetylglucosamine, is unique amongst the major polysaccharide components of the fungal cell wall in that it is able to be chemically modified by deacetylation. Chitin deacetylation is catalyzed by a family of carbohydrate esterase enzymes known as chitin deacetylases (CDAs), belonging to the Carbohydrate Esterase 4 family, according to Carbohydrate-Active EnZymes database (CAZy) classification. The product of this deacetylation is an often heterogenous polymer of glucosamine and N-acetylglucosamine, known as chitosan. CDAs are broadly conserved, known to occur not only in all fungi (Ruiz-Herrera & Ortiz-Castellanos, 2010) but also in insects, where they have a number of developmental roles, including in tracheal development and moulting (Arakane et al., 2009;Luschnig, Batz, Armbruster, & Krasnow, 2006;Wang, Jayaram, et al., 2006b;Xi, Pan, Ye, Yu, & Zhang, 2014). Considering the relative chemical and physical properties of chitin and chitosan, we hypothesized that chitin deacetylation could also have key roles in fungal development. Chitin forms highly crystalline, rigid microfibrils in the cell wall (Rinaudo, 2006), but its deacetylation results in the creation of primary amine groups, with a pKa of~6.5 (Wang, Chen, et al., 2006a). As a result, chitosan is polycationic at physiological pH and therefore soluble and potentially more flexible than chitin. Here, comparisons could be made with the demethylesterification of homogalacturonan (HG) in plant cell walls by pectin methylesterase enzymes. Such demethylesterification has a critical role in the regulation of the mechanical properties of the cell wall and therefore in plant development (reviewed in (Wolf, Mouille, & Pelloux, 2009). HG is secreted in a methylesterified form, only to be demethylesterified in muro by pectin methylesterase enzymes, the effects of which are twofold. First, demethylesterification creates anionic carboxyl groups that can be cross-linked by Ca 2+ , which increases wall rigidity. On the other hand, demethylesterified HG is also more susceptible to hydrolysing enzymes such as endopolygalacturonase (Wolf et al., 2009). Determining whether chitin deacetylation has analogous roles in fungal cell walls could therefore be a valuable line of investigation. It may yield novel insights into the Fungal CDAs have currently only been characterized in a very limited number of species. In Saccharomyces cerevisiae and Cryptococcus neoformans, chitosan appears to have a structural role in ascospores and vegetative cells, respectively (Baker, Specht, Donlin, & Lodge, 2007;Christodoulidou, Bouriotis, & Thireos, 1996). Chitosan is also a component of the spore wall in Ashbya gossypii. In this case, however, deletion of the sole CDA in this species resulted in a complete loss of sporulation (Lickfeld & Schmitz, 2012), suggesting a possible developmental role for chitin deacetylation. Most recently, we discovered a novel role for chitosan in appressorium development in Magnaporthe oryzae. Here, chitosan does not play a morphogenic or structural role but is required for germling adhesion and surface perception (Geoghegan & Gurr, 2016). Taken together, these data suggest that chitosan may have multiple, discrete roles, depending on the species and cell type in question. It is therefore of interest to examine the possible involvement of chitin deacetylation in other developmental processes. Indeed, the role of chitin deacetylation has not hitherto been investigated in hyphal growth. The process of filamentous hyphal growth in fungi involves several morphological changes, including hyphal branching and hyphal tip extension, which could require localized alterations in cell wall flexibility. We hypothesized that these changes to the cell wall could involve the deacetylation of chitin. In order to test this hypothesis, we investigated the role of chitin deacetylation during hyphal growth in M. oryzae. In addition, we also examined the extent to which the hydrolysis of chitosan is also involved in cell wall remodelling in M. oryzae. There is currently very little functional data regarding the role of chitosanases in fungi. Deletion of the chitosanase CsnB in Aspergillus fumigatus had no effect on vegetative growth or sporulation, although it was required for growth on media containing chitosan as a carbon source (Beck, Broniszewska, Schwienbacher, & Ebel, 2014). Similarly, silencing of CSN1 in Fusarium solani also had no effect on mycelial growth or sporulation, although reduced chitosanase activity was recorded (Liu, Zhang, Li, & Bao, 2010). However, because chitin deacetylation has not been investigated in either of these fungi, these studies lack any frame of reference. Indeed, it is difficult to determine the role of a chitosanase without first having prior knowledge of when and where chitosan is synthesized during the life cycle of the fungus. Thus, we sought a more complete understanding of synthesis and turnover in fungi by coupling the characterization of chitosan during vegetative growth with the recently described role of chitin deacetylation during appressorium development in M. oryzae (Geoghegan & Gurr, 2016).

| Chitosan is a component of the cell wall in vegetative hyphae
In order to first determine whether chitosan is a component of the cell wall in the vegetative hyphae of M. oryzae, two approaches were used.
First, mycelium resulting from growth in liquid complete medium (CM) was stained with the anti-chitosan antibody mAbG7 (Schubert et al., 2010). However, this revealed only weak staining of the hyphae ( Figure S1), which nevertheless suggested the presence of chitosan in the cell walls of vegetative hyphae. In the second approach, mycelium was stained with the recently developed chitosan-specific fluorescent probe OGA488 (Mravec et al., 2014). In this case, much stronger staining of hyphae was observed (Figure 1a), which is perhaps attributable to the superior penetrability of the OGA488 probe compared with the antibody. Septa in particular demonstrated very strong fluorescence. Intriguingly, this fluorescence was often restricted to the outer edge of the septa, thereby giving rise to an "annular" pattern of fluorescence. Unfortunately, the organization and distribution of hyphae are not easily apparent upon microscopic observation of mycelial pellets, making it difficult to draw any conclusions about the developmental regulation of chitin deacetylation during vegetative growth.
In an attempt to shed light on this question, staining with OGA488 was also performed of hyphae, which had been grown over a glass coverslip, which had previously been overlaid with solid CM. In this case, staining was not observed in hyphae growing at the outermost edge of the colony but was typically first observed 1.5-2 mm from this point. Here, strong labelling of septa was again observed, as well as weaker labelling of lateral walls, and occasionally of hyphal tips ( Figure 1b). However, the extreme hydrophobicity of hyphae grown on solid medium prevented the labelling of all but the outer~5 mm of the colony, as washing buffer and OGA488 failed to penetrate beyond the surface hyphae.
Chitosan is known to be a component of the cell wall in the ascospores of A. gossypii (Lickfeld & Schmitz, 2012) and S. cerevisiae (Christodoulidou, Briza, Ellinger, & Bouriotis, 1999;Christodoulidou et al., 1996). To determine whether this was also the case in the conidia of M. oryzae, staining was performed with both eosin Y and OGA488.
Although no staining was observed with OGA488 (not shown), weak fluorescence was observed in the cell wall of conidia labelled with eosin Y (Figure 1c). Considering the highly specific nature of the interaction between OGA488 and chitosan (Mravec et al., 2014), this suggests that chitosan may not be a component of the cell wall in conidia. The weak staining observed with eosin Y may be a result of a nonspecific electrostatic interaction with another cationic cell wall component.

| Deletion of CDAs operating during vegetative growth
Having established that chitin is deacetylated during hyphal growth, the next objective was to determine which CDAs were responsible.
The genome of M. oryzae was previously found to contain 10 putative CDA genes (Geoghegan & Gurr, 2016). In order to find which of these genes were expressed during vegetative growth, we performed quantitative reverse transcription polymerase chain reaction analysis on RNA extracted from mycelia grown in liquid CM. Relative transcript abundances for all 10 CDA genes revealed that CDA1 was by far the most highly expressed CDA in this tissue, followed by CDA4 ( Figure 1d). These data align well with a published transcriptomics dataset (Soanes, Chakrabarti, Paszkiewicz, Dawe, & Talbot, 2012).
From the transcriptional analysis outlined above, it was clear that two CDA genes were responsible for chitin deacetylation during vegetative growth: CDA1 and CDA4.
Cda1 is a predicted secreted protein with a polysaccharide deacetylase domain (Pfam 01522) and a C-terminal chitin binding domain (Pfam 00187; Figure S2). Cda4 is also a predicted secreted protein, with a polysaccharide deacetylase domain (Pfam 01522), but also a putative transmembrane domain at its C-terminus ( Figure   S2). In order to characterize the role of chitin deacetylation by these two proteins, targeted deletion of CDA1 and CDA4 was performed.
In addition, CDA5 was also deleted, both alone and in combination with CDA4. Like Cda4, Cda5 also has a putative single transmembrane domain at its C-terminus, and so it was considered likely that it may demonstrate redundancy with Cda4. Successful replacement of the target genes by antibiotic resistance markers was confirmed by polymerase chain reaction (PCR) and Southern blotting ( Figures   S3 and S4).

| Cell wall composition is altered in the CDA deletion strains
Following successful gene replacement as described above, chitin deacetylation in the resulting deletion strains was characterized. First, mycelial pellets resulting from 72-hr growth in CM were stained with OGA488. Although hyphae of the wild-type (WT) strain demonstrated strong labelling of septa and lateral walls as discussed previously between the hyphae of the WT and Δcda1 strains, suggesting that the reduction in chitin deacetylation had not resulted in a concomitant increase in chitin content in Δcda1. It was also apparent from the CFW staining that septum and hyphal morphology were unchanged in the Δcda1 strain. Hyphae grown on solid medium were also stained with OGA488. In this case, a dramatic reduction in labelling of hyphae 1.5-2 mm from the colony margins was observed in the Δcda4 and ΔΔcda4/cda5 strains (Figure 3c), whereas staining of these hyphae was unaffected in Δcda1 ( Figure 2b). Staining was not entirely abolished in ΔΔcda4/cda5; septa in particular continued to show some labelling, perhaps due to the action of Cda1. Nevertheless, these data suggest that Cda1 and Cda4 appear to have developmentally distinct   whereas Cda4 (and perhaps to a lesser extent Cda5) deacetylates chitin in hyphae at colony margins. Conidia of the deletion strains were also stained with eosin Y. However, no change in labelling was observed in either deletion strain ( Figure 3b and not shown), nor was there any clear change in conidiation (not shown).

| Localization of Cda1 and Cda4
Cda1 and Cda4 appear to have distinct roles in chitin deacetylation during vegetative growth. In order to further examine the roles of these enzymes, the respective deletion strains were complemented by expressing mCherry-tagged Cda1 or Cda4.
In strains expressing CDA1:mCherry, weak fluorescence was  Thines, & Steinberg, 2006). We hypothesized that Cda4:mCherry may therefore colocalize with a chitin synthase, which would provide a valuable insight into possible coordinated transport of these proteins and the mechanism of chitin deacetylation. There are seven chitin synthases in M. oryzae, which could possibly colocalize with Cda4: mCherry (Kong et al., 2012). However, taking into consideration previous work, Chs1 appeared to be a likely candidate; Chs1:eGFP had already been found to localize to hyphal tips and to developing conidia (Kong et al., 2012). To see if Chs1:eGFP colocalizes with Chs4:mCherry, first, the Chs1:eGFP construct was cloned in precisely the same way as described previously (Kong et al., 2012) and transformed into strains expressing CDA4:mCherry. However, although Chs1:eGFP and Cda4:mCherry were often expressed in the same hyphae, colocalization of the fusion proteins was not observed.
Chs1:eGFP fluorescence localized to the cell periphery and Spitzenkörper at hyphal tips, in contrast to Cda4:mCherry ( Figure   S7). In addition, although Chs1:eGFP also demonstrated intracellular localization, this only colocalized with Cda4:mCherry in the larger intracellular bodies, most likely vacuoles.
2.5 | Vegetative growth and pathogenic development are largely unaffected in the CDA deletion strains To determine the role of chitin deacetylation during vegetative growth, growth of the deletion strains was evaluated under a range of different stress conditions on solid medium. No significant differences in radial growth were observed between the WT and Δcda1 strains after 10day growth (Student's t test, p < .05; Figure S5), suggesting that chitin deacetylation by Cda1 is not required for optimal radial growth under the conditions tested. In addition, no consistent differences in colony appearance, conidiation or hyphal morphology were observed between the WT and Δcda1 strains (Figures 2f and S5). In order to further investigate the properties of the cell wall in the Δcda1 strain, a protoplast release assay was performed. Here, the mycelium resulting from 72-hr growth in CM was harvested and incubated with cell wall lysing enzymes (a mixture of chitinases, glucanases, and proteases).
Under these conditions, the Δcda1 released fewer protoplasts than either the WT or complemented strain (Δcda1/CDA1:mCherry; Figure 2g), suggesting that the cell wall of the Δcda1 strain is more resistant to enzymatic lysis.
Vegetative growth was also largely unaffected in the Δcda4 and ΔΔcda4/cda5 strains ( Figure S8), with overall colony appearance, conidiation, and hyphal morphology identical to the WT strain ( Figure   S8). Growth upon media containing caffeine resulted in the largest growth reductions in the deletion strains, although even here, the reduction was only about 15%. This suggests that chitin deacetylation in hyphae at colony margins is also not required for optimal growth under the conditions tested.
Pathogenic development was also evaluated in the CDA deletion strains. Conidia of deletion strains germinated on a hydrophobic glass surface developed appressoria with identical morphology to WT ( Figures S5 and S9). Pathogenicity was also unaffected in the cda strains; rice leaves inoculated with conidia of the deletion strains demonstrated similar lesion density to the WT strain ( Figures S5 and   S9). CDA1, CDA4, and CDA5 are therefore not required for pathogenic development in M. oryzae, which is as expected, considering their apparent specificity to the vegetative growth phase.

| Ectopic expression of CDA1:mCherry
The deacetylation of chitin had been hypothesized to be required for increasing cell wall flexibility during morphogenic events. However, the deletion of CDA1 or CDA4 did not result in any growth or morphological differences. A separate approach was therefore sought, in order to further investigate a possible link between chitin deacetylation and morphogenesis. We therefore decided to ectopically express CDA1: Such fluorescence also appeared to be restricted to the mature hyphae and was not observed in those growing at colony margins. Hyphal morphology was unaffected by ectopic expression of Cda1:mCherry as was overall colony appearance (not shown). In conidia germinated on a hydrophobic glass surface, fluorescence was observed in the cell walls of the conidia, germ tubes, and appressoria, although intracellular localization (most likely vacuolar) was also apparent (Figure 4b). During invasive growth, cell wall-localized fluorescence was also observed in invasive hyphae (Figure 4c), but morphology was, once again, unaffected. Although germling morphology was identical to the WT in pEF1:CDA1mCherry strains, staining with OGA488 demonstrated that ectopic chitin deacetylation had indeed occurred in the conidial wall ( Figure 4d).  (Table S2) suggests that the M. oryzae chitosanase is unique with regard to these repeat sequences.

|
However, a BLAST search using these repeat sequences as a query revealed a number of proteins with similar types of repeat sequences, including the cell wall protein BibA from Streptococcus agalactiae (Santi et al., 2007), procyclic acidic repeat proteins from Trypanosoma brucei (Mowatt & Clayton, 1987), and the periplasmic protein TonB from Eschericia coli (Postle & Good, 1983).
Alignment of chitosanase sequences from a number of fungi ( Figure S9) shows that the key catalytic residues (aspartate and glu- were an exception, with four, as discovered previously (Cheng et al., 2006). Thus, the distribution and number of chitosanases do not mirror that of the CDAs.

| Chitosanase is expressed in vegetative hyphae but is not required for normal vegetative growth or development
Chitin deacetylation is known to occur at multiple different stages of the life cycle in M. oryzae (Geoghegan & Gurr, 2016). To determine whether the expression of CSN is coincident with chitin deacetylation, a promoter fusion was created. The upstream sequence of CSN was cloned upstream of three copies of YFPvenus, and the entire construct transformed into WT M. oryzae. In the resulting transformants, fluorescence was observed most strongly in vegetative hyphae, particularly those at the colony margins ( Figure 5a). Additionally, fluorescence was observed in appressoria postpenetration (~48 hpi; Figure 5B), but neither in invasive hyphae nor during appressorium development (not shown). Therefore, the expression of CSN is not coincident with chitin deacetylation in M. oryzae.
We also attempted to determine the localization of the chitosanase protein, by creating a C-terminal eGFP fusion. However, Based on the promoter analysis, it appears that CSN has a role in vegetative growth. In order to characterize this role, a deletion strain for CSN was created. The entire coding sequence of CSN was replaced with a gene encoding resistance to bialaphos, and successful deletion was confirmed by PCR and Southern blotting ( Figure S12). Two independent deletion lines (named 9 and 46) were chosen for characterization.
Δcsn strains were first stained for the presence of chitosan, to determine whether deletion of CSN had resulted in alterations in either the amount or distribution of chitosan. However, staining of vegetative hyphae with OGA488, germlings with the anti-chitosan antibody mAbG7 (Schubert et al., 2010), or conidia with eosin Y did not reveal any differences in the patterns or intensity of staining in Δcsn, compared with the WT strain (Figures 5b and S13). The morphology of hyphae, germlings, and conidia was also unaffected in the Δcsn strain. Radial growth of the Δcsn strain was assessed on solid media under a range of stress conditions. As with the CDA deletion strains assessed in this study, radial growth was largely unaffected in the Δcsn strain, although some small yet statistically significant changes were detected under certain conditions, including SDS and H 2 O 2 ( Figure S14).
In A. fumigatus, deletion of CsnB resulted in reduced growth on media containing chitosan (Beck et al., 2014). To see if this was also true of the Δcsn strains in M. oryzae, growth assays with chitosan as a carbon source were performed on solid and in liquid medium. On solid medium containing 0.5% chitosan (w/v), equal growth was observed in all strains after a 14-day incubation (Figure 5d). In some fungi, a "halo" indicative of substrate degradation is also observed around fungal colonies grown on chitosan (Palma-Guerrero, Jansson, Salinas, & Lopez-Llorca, 2008), but this was not observed for M. oryzae.
In liquid medium, chitosan was clearly able to be used as a carbon source, because biomass was much higher in flasks containing 0.33% chitosan (w/v; and 16.6 mM sodium acetate) than those containing sodium acetate alone. However, the biomass of the Δcsn strains was identical to the WT strain ( Figure S13).

Cell wall hydrolases have previously been implicated in autolysis
under starvation conditions. To determine whether Csn is required for autolysis, equal concentrations of conidia were inoculated into flasks containing minimal media and incubated for up to 30 days, with biomass measurements taken every 4 days after an initial 5-day incubation.
Biomass increased up to 17 dpi, then decreased between 21 and 30 dpi, suggesting the onset of autolysis ( Figure S13). However, the biomass of the WT and Δcsn strains was similar across the duration of the experiment, suggesting that chitosanase is not required for autolysis.

| Pathogenicity and pathogenic development are unaffected by deletion of CSN
The cell walls of appressoria and invasive hyphae appear to have a particularly high chitosan content (Geoghegan & Gurr, 2016). In order to determine whether chitosanase is required for morphogenesis in these has roles beyond those already described for pathogenic development in this fungus (Geoghegan & Gurr, 2016). Two CDAs (Cda1 and Cda4) were found to be necessary for chitin deacetylation in vegetative hyphae. Interestingly, they appear to be involved in two distinct modes of chitin deacetylation, in discrete zones of colonies. for these enzymes (Hartl, Zach, & Seidl-Seiboth, 2012). The increased resistance of the cell wall to enzymatic degradation in Δcda1 could perhaps be attributed to a reduction in cell wall permeability. Chitin is known to be a highly crystalline polysaccharide and therefore has low permeability. The deacetylation of chitin by Cda1 could cause loss of crystallinity and therefore increased permeability, allowing hydrolytic enzymes better access to substrates in the cell wall. This could relate to the biological role of chitin deacetylation at septa, where both chitosan and Cda1:mCherry were found to localize most strongly.

| Chitin deacetylation in mature hyphae
Septa are known to be sites of protein secretion in hyphae (Hayakawa, Ishikawa, Shoji, Nakano, & Kitamoto, 2011), and so efficient secretion from septa may be dependent upon high cell wall permeability here.
Such permeability may be imparted by chitin deacetylation by Cda1.
Chitosan has also been shown to localize to hyphal septa in Neurospora crassa (Mravec et al., 2014), suggesting that this may be a conserved feature of the septal cell wall in filamentous fungi.
An additional role for chitin deacetylation by Cda1 in mature hyphae could be as part of a carbon-recycling mechanism. The cell wall is a major carbon source, and so its deconstruction could serve to redistribute this carbon from the aged parts of the fungal colony to the growing edge. Chitin deacetylation could either serve as a direct source of carbon in the form of acetate (released by the deacetylation reaction) or be the first step to making the cell wall of old hyphae more susceptible to degradation by endogenous chitinases and glucanases.
However, no evidence for a role in carbon recycling was found in the present study. First, Cda1 appears to be highly expressed even after a relatively short incubation in nutrient-rich CM, at which point carbon recycling from the cell wall would probably not be necessary. Second, under conditions of carbon or nitrogen starvation on solid medium, no reduction in radial growth was observed. Yet further investigation is required before this hypothesis can be dismissed; more sensitive methods may be required to study the efficiency of carbon recycling from the cell wall in the Δcda1 mutant.

| Chitin deacetylation in the subperipheral zone
Chitosan was also found in the cell walls of vegetative hyphae located in the subperipheral zone of colonies. Here, chitosan localized not only to the septa and lateral walls of hyphae but occasionally also to hyphal tips, although not exclusively so. Chitin deacetylation in these cells did not require Cda1 but did require Cda4 (and perhaps to a lesser extent Cda5), two predicted secreted proteins with putative single transmembrane domains. Although the staining of chitosan in these hyphae was reduced in the Δcda4 and ΔΔcda4/cda5 strains, the loss of chitosan had little impact on radial growth or hyphal morphology. This suggests that chitin deacetylation is not required for cell wall integrity or morphogenesis in these hyphae. However, we cannot rule out compensatory changes in cell wall composition, which may have masked the effects of chitosan loss. The deletion strains did appear to be more sensitive to caffeine, but even here, the reduction in growth was marginal. Additionally, because caffeine has pleiotropic effects on cells (Levin, 2011), the increased sensitivity of the deletion strains to this chemical does not necessarily provide any particular insight into the role played by Cda4 and Cda5. However, numerous studies have found that mutations in components of the cell wall integrity signalling pathway or in genes with roles in cell wall biogenesis are also more sensitive to caffeine (Levin, 2011). The change in cell wall composition caused by deletion of CDA4 and CDA5 most likely causes the increased caffeine sensitivity, but the mechanism behind this is unknown. between the CDA and a chitin synthase. Such membrane association has previously been found to be necessary for chitin deacetylation by Cda2 in C. neoformans (Gilbert, Baker, Specht, & Lodge, 2012).

| Hydrolysis of chitosan in M. oryzae
Chitosan is clearly present in the cell wall of a number of distinct cell types in the M. oryzae life cycle, with apparently disparate roles. As part of this study, we also examined the role played by chitosanase, to gain insight into the relationship between chitin deacetylation and the hydrolysis of chitosan.
Unexpectedly, the expression of CSN did not coincide with the deacetylation of chitin over the life cycle of M. oryzae. Expression was highest during vegetative growth but was very low, or even absent, during appressorium development and in planta growth. This is despite the fact that significant concentrations of chitosan known to be present in the cell wall during these processes (Geoghegan & Gurr, 2016). The hydrolysis of chitosan may therefore presumably not always a requirement during development. Yet, although there is only a single chitosanase, this does not necessarily mean that there is only a single enzyme with chitosanase activity. Fungal chitosan is typically only 70%-90% deacetylated (Chatterjee & Guha, 2014;Munoz, Valencia, Valderruten, Ruiz-Durantez, & Zuluaga, 2015;Pochanavanich & Suntornsuk, 2002;Yen & Mau, 2007), meaning that some GlcNAc residues remain in the polymer. Both chitosanases and chitinases are able to hydrolyze a mixed GlcN-GlcNAc linkage (Hartl et al., 2012), and so chitinases are able to hydrolyze chitosan to some degree, depending on the degree of deacetylation. Thus, from the perspective of cell wall remodelling by cell wall hydrolases, it is difficult to make a clear distinction between chitin and chitosan. Any remodelling of chitosan required during cellular morphogenesis may just as well be performed by chitinases as by chitosanase. Such redundancy between chitinases and chitosanases could also explain a number of other observations, including the lack of chitosanases in some fungi and the normal growth of the M. oryzae Δcsn strain on chitosan. This could also help to explain the lack of any growth or morphogenic defects observed upon CSN deletion or knockdown in M. oryzae, A. fumigatus (Beck et al., 2014), F. solani (Liu et al., 2010), or N. crassa (Maddi, Dettman, Fu, Seiler, & Free, 2012). The presence of as yet uncharacterized enzymes with chitosanase activity is also a possibility.
There are also other possible reasons for the lack of chitosanase expression in certain developmental stages. For example, it may reflect differences in the mode of chitin deacetylation: We hypothesize that chitin deacetylation can occur either postsynthesis or coincident with chitin synthesis. There may therefore be different requirements for chitosan hydrolysis, dependent upon the mode of chitin deacetylation.
Last, the inherent properties of chitosan as a hydrophilic, flexible polysaccharide could also mean that hydrolysis is not required to render it pliable in the same way as chitin, for example. Clearly, there are many questions that require further investigation in this area. In

| Fungal strains and growth conditions
The WT rice pathogenic M. oryzae strain Guy11 and mutant strains were cultured at 24°C with a 14-hr light 10-hr dark cycle. Strain maintenance and composition of media were essentially as described by Talbot et al. (Talbot, Ebbole, & Hamer, 1993 Table S1). These fragments were joined to the resistance gene by overlapping PCR, using primer pairs 5/4 and 3/8 for CDA1, 9/2 and 1/12 for CDA4, 13/4 and 3/16 for CDA5, and 17/2 and 1/20 for CSN. The final, complete construct was made by overlapping PCR, amplifying the products from the previous reactions using primer pairs 5/8 for CDA1, 9/12 for CDA4, and 13/16 for CDA5. The complete CSN deletion cassette could not be amplified with the requisite purity and so was used in two overlapping fragments (see Figure S8). The final PCR products were used directly for DNAmediated protoplast transformation of WT Guy11 strain following protocols described by Talbot et al. (Talbot et al., 1993). Putative transformants were selected on minimal medium supplemented with 300 μg/ml hygromycin B (Calbiochem, Merck, Darmstadt, Germany) or defined complex medium supplemented with 60 μg/ml bialaphos (Goldbio, St. Louis, MO, USA). Deletion of the target gene was confirmed by both PCR and Southern blot analysis, as described in Samalova et al. (Samalova et al., 2013).
Double CDA deletion strains were generated in the Δcda4 background strain. The coding sequence of CDA5 was replaced with a hygromycin resistance cassette. The deletion construct was made as described above. The CDA5 deletion cassette was transformed directly into protoplasts of the Δcda4 strain, to generate the ΔΔcda4/cda5 mutant. Putative transformants were selected on minimal medium supplemented with 300 μg/ml hygromycin B. Deletion strains were confirmed as above.

| Cloning of fluorescently tagged CDAs, CSN, and promoter fusion
Standard molecular techniques (Ausubel et al., 1999) were used to prepare the complementation constructs with fluorescently tagged CDAs and CSN. A set of transformation vectors based on pUCAP was generated as described in Samalova et al. (Samalova, Meyer, Gurr, & Fricker, 2014). The vectors contain polyadenylation signal pATrpC and either bialophos or hygromycin resistance marker that was cloned into recreated SalI sites using primer pairs 1/2 or 3/4, respectively (see Table S1 and Figure S15). For PCR amplification of CDA1, CDA4, CHS1 (MGG_01868), or CSN, primer pairs 21/22, 23/24, 27/28, or 25/26 were used, respectively. Genomic DNA from the WT strain Guy11 was used as a template and amplified using Herculase DNA polymerase (Agilent). This resulted in amplification of the coding sequence of the genes (without stop codons), together with up to 2 kb of native promoter sequence, depending on proximity of neighbouring genes.
For the mCherry fusions, the PCR products were cloned into the AscI sites of the vector described above ( Figure S15), creating C-terminal mCherry fusions. For eGFP fusions, mCherry was replaced with eGFP, and the PCR products cloned into the AscI-SbfI sites.
For ectopic expression of CDA1:mCherry, the promoter sequence of EF1α (MGG_03641) was amplified using primer pair 37/38 (Table S1). The coding sequence of CDA1 was amplified using primer pair 39/22. The resulting DNA fragments were joined by overlapping PCR using primer pair 37/22, and the final construct cloned into the AscI site of the vector ( Figure S15).

| Confocal imaging
For imaging of infection structures, conidia (2.5 × 10 5 ml −1 ) of Guy11 and fluorescently tagged strains were collected from 10-day-old plates and inoculated in 50-μl droplets onto hydrophobic glass coverslips or onion peels as described in Samalova et al. (Samalova et al., 2014) and incubated for specified times in the growth chamber. For imaging of mycelial growth, conidia were incubated into liquid CM and incubated in the dark at 24°C for~72 hr, with shaking at 150 rpm. The resulting mycelial pellets could then be removed for imaging. Alternatively, a sterile glass coverslip was overlayed with CM and placed at the edge of a colony growing on solid medium. The fungus was allowed to grow over the coverslip for 24-72 hr, and then the coverslip could be removed for imaging.
For viewing mCherry fluorescence, the samples were viewed using the C-Apochromat 40×/1.2 water corrected objective lens of a Zeiss LSM 510 Meta confocal microscope at 543-nm excitation from the HeNe laser and emission collected with an LP585 filter. For dual imaging of OGA488 and mCherry or eGFP and mCherry, samples were viewed using a C-Apochromat 40×/1.2 water corrected objective lens of a Zeiss LSM 510 Meta confocal microscope. Dual excitation at 488 nm and 543 nm was provided by Argon and Helium-Neon lasers, respectively. Emitted light was collected with BP500 to 530-and LP585-nm filters.
CFW staining was performed as follows: Mycelial pellets were washed briefly with dH 2 O, then incubated on ice with 0.05% CFW for 20 min. Samples were washed 2-3 times with dH 2 O and viewed using the Zeiss LSM510 microscope as above, with 405-nm excitation and emission collected with an LP420 filter.
OGA488 staining was performed essentially as described in Geoghegan and Gurr (2016). Samples (mycelial pellets and coverslips that were overgrown with hyphae or had previously been inoculated with conidia) were washed briefly with 25 mM MES (pH 5.6) and incu- Staining with the monoclonal anti-chitosan antibody mAbG7, a generous gift from Stefan Schillberg (Schubert et al., 2010), was performed as described previously (Geoghegan & Gurr, 2016).
Eosin Y staining was performed as described by Baker et al. (2007).
Briefly, conidia were harvested, pelleted by centrifugation at 2,850 g in a Beckman Coulter Allegra X-15R, and then resuspended in McIlvaine buffer (pH 6). Conidial concentration was adjusted to 1 × 10 6 conidia per milliliter. To 500 μl of conidial suspension, 10-μl eosin Y (5 mg/ml) was added and incubated on ice for 30 min. Conidia were washed by centrifugation of the labelled spore suspension at 16,000 g for 5 min, removal of the supernatant, and resuspension in 1-ml McIlvaine buffer.
This was repeated 3 times, except after the last wash, conidia were resuspended in 100 ml of McIlvaine buffer. Labelled conidia could then be viewed with an Olympus BX50 microscope, using a NIBA filter.

| Pathogenicity and infection-related morphogenesis assays
Conidial germination and appressorium development were assessed at 1, 8, 16, or 24 hpi by following germling differentiation on hydrophobic glass coverslips (Gerhard Menzel, Glasbearbeitungswerk GmbH & Co., Braunschweig, Germany). Conidia (2.5 × 10 5 ml −1 ) of the Guy11 and mutant strains were inoculated in 50-μl droplets onto hydrophobic glass coverslips and incubated in the growth chamber for the specified time. Samples were viewed under an Olympus BX50 microscope, and~500 germlings in three independent experiments counted per strain/time point.