Expression Analysis of Cell Wall-Related Genes in the Plant Pathogenic Fungus Drechslera teres

Drechslera teres (D. teres) is an ascomycete, responsible for net blotch, the most serious barley disease causing an important economic impact. The cell wall is a crucial structure for the growth and development of fungi. Thus, understanding cell wall structure, composition and biosynthesis can help in designing new strategies for pest management. Despite the severity and economic impact of net blotch, this is the first study analyzing the cell wall-related genes in D. teres. We have identified key genes involved in the synthesis/remodeling of cell wall polysaccharides, namely chitin, β-(1,3)-glucan and mixed-linkage glucan synthases, as well as endo/exoglucanases and a mitogen-activated protein kinase. We have also analyzed the differential expression of these genes in D. teres spores and in the mycelium after cultivation on different media, as well as in the presence of Paraburkholderia phytofirmans strain PsJN, a plant growth-promoting bacterium (PGPB). The targeted gene expression analysis shows higher gene expression in the spores and in the mycelium with the application of PGPB. Besides analyzing key cell-wall-related genes, this study also identifies the most suitable reference genes to normalize qPCR results in D. teres, thus serving as a basis for future molecular studies on this ascomycete.


Introduction
Barley is the fourth most-produced cereal in the world behind maize, wheat and rice. This crop is used mostly for animal feed (55%-60%) and by the malt industry (up to 35%) [1]. Globally, twenty million tons of malt are produced per year, and barley is intended for beer production in breweries (https://www.planetoscope.com). However, the production of this monocotyledon may be compromised by the phytopathogen ascomycete Drechslera teres [2,3]. This filamentous fungus is responsible for net blotch, easily recognizable by the occurrence of brown necrotic lesions on leaves [4,5]. This disease negatively impacts barley physiology and development and thus causes important agronomic and economic losses. Because of the use of chemical products to control D. teres, the emergence of fungicide resistance is a matter of growing concern.
The cell wall is a crucial structure for fungal development, constituting the first physical barrier of protection and is, therefore, a target for antifungal agents [6]. Although the fungal cell wall composition

Barley Oat Meal Agar (BOA)
The medium composition is as follows: 18 g/L meal agar, 50 g/L milled leaves of barley and 17 g/L agar. This enriched medium brings all nutritive elements to the growth and spore production of D. teres. The fungus was grown for 15 days on this medium with 12 h in darkness and 12 h under blue light near UV emission at 20 • C. Onesirosan and Banttari (1969) demonstrated that spore production was greatest when the fungal cultures were exposed to this wavelength [21]. Two inoculation conditions were tested including the mycelium of the fungus containing spores (first condition) or spore suspension (second condition). For the first type of inoculation, sterile water was deposited on the surface of the fungal mycelium to facilitate its harvest with a rod. For the second inoculation condition, mycelium and spores were filtered through sterile gauze tissues. The filtrate consisted mainly of spores. The concentration of the suspension was adjusted at 10 5 spores/mL using a Malassez counting chamber (Marienfeld, Lauda-Königshofen, Germany).

Potato Dextrose Agar (PDA)
Co-culture of the fungus and PsJN was carried out on PDA (+/−) medium (39 g/L PDA, pH 4.5 ± 0.2), a medium allowing both the development of fungi and bacteria. After 15 days of growth, a rod was passed on the whole mycelium surface together with sterile water.

Bioinformatics
The maximum likelihood phylogenetic analysis of CHS (protein sequences) was obtained with several CHSs from the following fungi: D. teres, Aspergillus nidulans [22], Alternaria alternata, Botrytis cinerea [23], Blumeria graminis [24], Fusarium graminearum, Tuber melanosporum [25] and Magnaporthe grisea [9]. The tree was generated with PhyML [26] and available at http://www.phylogeny.fr. The tree was visualized with the online software iTOL (http://itol.embl.de). The CHS sequences from D. teres, A. alternata, B. graminis, F. graminearum, T. melanosporum and M. grisea were obtained by blasting the A. nidulans CHSs (National Center for Biotechnology Information (NCBI) [27]. The identification of CHS domains was carried out with Motif Scan [28]. Sequences were aligned with Clustal Omega [29]. The prediction of the transmembrane domain was performed using the online programs TMHMM (v. 2.0) [30] and Phobius [31]. Protein identifications and corresponding accession numbers from NCBI are indicated in Table 1. Table 1. List of the chitin synthase (CHS) protein accession numbers from the species used in this study.

Species of Fungi Protein Id Accession Number
Drechslera teres Aspergillus nidulans

Species of Fungi Protein Id Accession Number
Botrytis cinerea Tuber melanosporum 2.3. RNA Extraction and cDNA Synthesis D. teres mycelium was crushed with liquid nitrogen using a mortar and a pestle. Spore suspensions, stored at −80 • C, were lyophilized (Freeze Dryer Alpha 1/ 2-4 Christ) for 12 h at −55 • C, then milled using a ball mill MM400 (Retsch) for 2 min at 20 Hz. The extraction of total RNA was carried out using the RNeasy Plant Mini Kit including the DNase I on-column digestion (Qiagen, Leusden, The Netherlands). The integrity of the obtained RNA was evaluated with an Agilent Bioanalyzer (Agilent, Santa Clara, CA, USA). RNA Integrity Numbers (RINs) were >7. The RNA quality and quantity were checked using a Nanodrop ND-1000 spectrophotometer (Thermo Scientific, Villebon-sur-Yvette, France) (A 260/280 and A 260/230 ratios between 1.9 and 2.2). In the case of contamination (ratio 260/230 < 2), samples were precipitated with ammonium acetate (NH 4 OAc) and washed in ethanol [32]. Subsequently, 1 µg of extracted RNA was retro-transcribed using the Superscript II cDNA Synthesis kit (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's instructions.

Analysis of D. teres Phenotypes on Several Media
D. teres shows varying phenotypes according to the culture media used (Figure 1). On MP (−) medium, the fungus develops a white structure, similar to a feather duster (Figure 1a) as described in the literature [39]. This structure appears when the fungus is searching for nutritive resources. On this medium, spore production is not possible since their formation and subsequent germination require, in most filamentous fungi, the availability of nutrients in the culture medium, such as sugars, amino acids and inorganic salts [40].
On BOA (+), the color of the mycelium becomes black resulting from sporulation (Figure 1b,c). When exposed to wavelengths between 355 and 495 nm, followed by a dark period on rich medium, the fungus produces a great number of spores [41,42]. Spores of D. teres are cylindrical in shape with round ends, having a length from 25 to 300 µm and thickness from 7-11 µm [4,5]. Due to their septa, these spores are recognizable from other fungi (Figure 1d) [43]. The spores presenting less than two septa will not germinate and cannot penetrate plant tissues [44]. D. teres infects the plants via the spores, the reproductive structures, which are dispersed largely by the wind or rain and often over long distances.
On PDA (+/−), D. teres covers the entire surface of the culture medium within seven days, denoting a rapid growth (Figure 1e). On this medium, D. teres produces a very small number of spores ( Figure 1g) as compared to the BOA (+) medium (Figure 1b) [45]. Since the PDA (+/−) medium is suitable for the development of bacteria and fungi, a co-culture of D. teres with PsJN, was performed. When grown alone on the PDA (+/−) medium (Figure 1e), D. teres has a fluffy mycelium with some feather dusters characteristic of this fungus [46]. Under co-cultivation with PsJN, D. teres has a different phenotype than when growing alone on PDA (+/−) (Figure 1f). Indeed, the mycelium of D. teres is less fluffy and fruiting bodies are present at the periphery, probably to provide a protective barrier against PsJN. PsJN protects indeed several crops against damages caused by different abiotic or biotic stresses and promotes plant growth [17][18][19][47][48][49][50]. According to our results, this strain seems to have no antifungal effect and is, therefore, unable to prevent the development of D. teres.
The results suggest that the variability of the phenotypes observed on the different media would be accompanied by changes in the expression of cell-wall-related genes, since the fungal cell wall is a dynamic structure accommodating the different growth stages and morphologies. On BOA (+), the color of the mycelium becomes black resulting from sporulation (Figure 1b,c). When exposed to wavelengths between 355 and 495 nm, followed by a dark period on rich medium, the fungus produces a great number of spores [41,42]. Spores of D. teres are cylindrical in shape with round ends, having a length from 25 to 300 µ m and thickness from 7-11 µ m [4,5]. Due to their septa, these spores are recognizable from other fungi (Figure 1d) [43]. The spores presenting less than two

Identification of CHS Genes in D. teres and Phylogenetic Analysis
The fungal cell wall is a complex and dynamic structure that protects the cell from environmental stresses [10,51]. Given the important role played in fungal physiology, the cell wall is considered as a suitable target for antifungal drugs [6].
Chitin is the most important structural component of the fungal cell wall [9][10][11]52,53]. The enzymes catalyzing the synthesis of chitin are CHSs, which are members of glycosyltransferases from family 2 (GT2), like cellulose synthases. CHSs are able to transfer N-acetyl-D-glucosamine from an activated sugar donor (UDP-N-acetyl-D-glucosamine) to an elongating chitin chain [6,54].
In previous work, seven chitin synthase genes chsA, chsB, chsC, chsD, chsE, csmA and csmB were identified in the model organism A. nidulans encoding CHSs of Classes I, II, III, IV, V VI and VII [22]. The presence of multiple CHSs in many fungi suggests that several CHSs can be used for chitin production at different stages of the fungal life-cycle [10].
BLASTp analysis carried out using the CHS protein sequences of A. nidulans identified six CHSs in D. teres (hereafter referred to as DtCHS1, 2, 3, 4, 5 and 7 for the protein sequences and DtCHS1, 2, 3, 4, 5 and 7 for the genes) ( Table 1). The maximum likelihood phylogenetic analysis carried out using CHS full-length protein sequences from several classes of fungi, notably Dothideomycetes, Sordariomycetes and Leotiomycetes, allowed assigning a phylogenetic relatedness for the D. teres CHS with known orthologs from other species (Figure 2).
The phylogenetic analysis demonstrates the existence of six CHS classes ( Figure 2). Class I, II and IV are present in all fungi, while Classes III, V and VII are particular to filamentous fungi [57]. The number of CHS genes changes according to the species. Most fungal species contain between three and six CHS genes [58]. For example, three CHS are present in S. cerevisiae, four in Candida albicans and eight in A. nidulans [59,60]. CHS1 and CHS2 have overlapping functions in septum formation in A. nidulans [61]. In the same way, CHS1 is crucial for infection-related morphogenesis, since 90% of the chs1 M. oryzae mutants have no septum and, therefore, display severe defects in conidium morphology [9]. CHS1 is also essential for cell wall integrity in Candida albicans [59]. Class IV enzymes contribute to the synthesis of the bulk cell wall chitin [57]. In M. oryzae, CHS1 is important for virulence and plays specific roles during conidiogenesis and appressorium formation. CHS2, CHS3, CHS4 and CHS5 are essential for plant infection and CHS6 is dispensable for pathogenesis [9]. Therefore, individual CHS genes play several roles in hyphal growth, pathogenesis, conidiogenesis and appressorium development.
BLASTp analyses and motif searches reveal similarities between the six identified CHSs from D. teres and other fungal CHSs (Table 3 and Figures 3 and 4). Structurally, CHS1, CHS2 and CHS3 are more similar to each other than to other CHSs in D. teres. As in A. nidulans and other filamentous fungi, Class V CHSs in D. teres have myosin motor-like domains at the N-terminus [11,62]. Myosins are actin-dependent molecular motors and play roles in several cellular processes. More specifically, the head myosin domain binds to actin in an ATP-dependent manner and generates force by ATP hydrolysis [63].
CHSs have several conserved domains which are considered as signature sequences since they are found in all CHSs [64]. Most of the amino acids of these signature motifs were found to be essential for activity [65]. CHSs contain the conserved EDR motif and the pentapeptide QRRRW ( Figure 4) which was reported also in CHSs from the chordate Branchiostoma floridae [66]. The importance of these motifs has been studied in yeast. Thereby, in Saccharomyces cerevisiae, mutations of the QRRRW motif lead to a significant decrease in CHSs activity [58]. By comparing the amino acid sequences of the six D. teres CHSs in NCBI, DtCHS1/DtCHS2, DtCHS2/DtCHS3 and DtCHS4/DtCHS5 have similarities, with 43%, 44% and 45% identity, respectively (Figures 3 and 4). DtCHS4 and DtCHS5 are closer with respect to the similarity of amino acid sequences. This sequence similarity is also confirmed by the phylogenetic tree (Figure 2).  Table 2.
Bootstraps > 80% are represented as black circles on the branches. The bootstrap % value increases with the circle size.
CHS1 and CHS2 have overlapping functions in septum formation in A. nidulans [61]. In the same way, CHS1 is crucial for infection-related morphogenesis, since 90% of the chs1 M. oryzae mutants have no septum and, therefore, display severe defects in conidium morphology [9]. CHS1 is also essential for cell wall integrity in Candida albicans [59]. Class IV enzymes contribute to the synthesis of the bulk cell wall chitin [57]. In M. oryzae, CHS1 is important for virulence and plays specific roles during conidiogenesis and appressorium formation. CHS2, CHS3, CHS4 and CHS5 are essential for plant infection and CHS6 is dispensable for pathogenesis [9]. Therefore, individual CHS genes play several roles in hyphal growth, pathogenesis, conidiogenesis and appressorium development.
BLASTp analyses and motif searches reveal similarities between the six identified CHSs from D. teres and other fungal CHSs (Table 3 and Figures 3 and 4). Structurally, CHS1, CHS2 and CHS3 are more similar to each other than to other CHSs in D. teres. As in A. nidulans and other filamentous fungi, Class V CHSs in D. teres have myosin motor-like domains at the N-terminus [11,62]. Myosins are actin-dependent molecular motors and play roles in several cellular processes. More specifically, the head myosin domain binds to actin in an ATP-dependent manner and generates force by ATP hydrolysis [63].  Table 2. Bootstraps > 80% are represented as black circles on the branches. The bootstrap % value increases with the circle size.
Genes 2020, 11, x FOR PEER REVIEW 10 of 18      CHSs have several conserved domains which are considered as signature sequences since they are found in all CHSs [64]. Most of the amino acids of these signature motifs were found to be essential for activity [65]. CHSs contain the conserved EDR motif and the pentapeptide QRRRW ( Figure 4) which was reported also in CHSs from the chordate Branchiostoma floridae [66]. The importance of these motifs has been studied in yeast. Thereby, in Saccharomyces cerevisiae, mutations of the QRRRW motif lead to a significant decrease in CHSs activity [58]. By comparing the amino acid sequences of the six D. teres CHSs in NCBI, DtCHS1/DtCHS2, DtCHS2/DtCHS3 and DtCHS4/DtCHS5 have similarities, with 43%, 44% and 45% identity, respectively (Figures 3 and 4). DtCHS4 and DtCHS5 are closer with respect to the similarity of amino acid sequences. This sequence similarity is also confirmed by the phylogenetic tree ( Figure 2).

Expression Analysis of Cell Wall-Related Genes in D. teres
The software geNorm PLUS [67] was used to analyze gene expression stability across different samples of D. teres: spores and mycelium on BOA (+), mycelium on MP (−), and mycelium on PDA (+/−) in the presence or absence of the bacterium. The stability and transcript levels of the eleven

Expression Analysis of Cell Wall-Related Genes in D. teres
The software geNorm PLUS [67] was used to analyze gene expression stability across different samples of D. teres: spores and mycelium on BOA (+), mycelium on MP (−), and mycelium on PDA (+/−) in the presence or absence of the bacterium. The stability and transcript levels of the eleven candidate reference genes actin, ApsC, Cos4, GlkA, PfkA, PgiA, SarA, IsdA, H2B, GAPDH and RS14 were investigated with geNorm PLUS . Cos4 and PgiA were identified as the best two transcripts for use in normalization of the data (Supplementary Figure S1).
The cell wall of ascomycetes also contains β-(1,3;1,4)-glucan accounting for 10% of the glucan content in the A. nidulans cell wall [35,72]. celA encodes a putative mixed-linkage glucan synthase in A. nidulans [22]; one ortholog was found in D. teres and was used in this study.
We also considered a gene encoding a protein kinase, PTK1, which was shown to be required for conidiation, appressoria formation and pathogenicity in D. teres [36].
The hierarchical clustering of the heat map ( Figure 5) shows that the CHS genes studied group in different clusters. CHS3, CHS4, CHS5 and CHS7 show a higher expression in spores and in the mycelium grown in the presence of the bacterium. This suggests that the bacterium induces the expression of specific CHSs in the phytopathogenic fungus. This phenomenon could be explained by an effect on the cell wall of the fungus. Indeed, PGPB is able to produce different types of cell-wall-lysing enzymes including chitinases, proteases, cellulases and β-(1,3)-glucanases [73,74]. The fungus could respond to PsJN by increasing the expression of CHSs in the attempt to restore the structural integrity of its cell wall.  Some CHS genes, such as CHS1 and CHS2, showed a tendency towards decreased expression on the poor medium MP (-) (Figures 5 and 6). This finding may indicate that these genes have a role in vegetative growth and the lower expression reflects the slower mycelium growth rate in a nutrient-poor environment.   In D. teres, EngA is more expressed in the mycelium cultivated on the poor medium (MP (-) medium) (Figures 5 and 6). Likewise, the exo-β-(1,3)-glucanases ExgB is expressed at higher levels on the poor medium (MP (-) medium) (Figures 5 and 6). However, the differences are not statistically significant and only show a trend.
FksA is more expressed in spores and the mycelium with PsJN ( Figures 5 and 6). According to Figure 6, the expression of celA is significantly higher in D. teres spores, as compared to the other conditions. PTK1 has a slightly larger expression in the spores as compared to the other experimental conditions (Figures 5 and 6). This is in agreement with the reported role of PTK1 in conidiation [36].
Our results show that the expression of CHS4, CHS5 and FksA is higher in the mycelium cultivated on PDA (+/−) medium with PsJN, as compared to the growth on BOA (+), MP (−) and PDA (+/−) ( Figure 6). According to Figure 1, PsJN has an effect on the phenotype of D. teres. This could be due to a possible secretion of hydrolytic enzymes by PsJN, acting on the integrity of the fungal cell wall [73,74]. D. teres could also perceive PsJN as a stress and thus strengthen its cell wall.
This hypothesis will have to be confirmed and validated by future experiments.

Conclusions
To the best of our knowledge, this is the first study devoted to the cell wall-related genes of D. teres. We identify key genes involved in the biosynthesis/remodeling of D. teres cell wall and differentially expressed in spores and/or in the mycelium depending on the culture media used. We also identify some cell wall biosynthetic genes induced by PsJN, a plant growth-promoting bacterium. Since PsJN seems to disturb fungal growth, it is reasonable to hypothesize that it could produce cell wall degrading enzymes causing a response in D. teres at the gene level. Additionally, we propose a list of potential candidate reference genes for qPCR analysis in D. teres. Our study paves the way to follow-up studies aiming at a functional characterization of cell wall genes of this economically relevant pathogen.

Supplementary Materials:
The following is available online at http://www.mdpi.com/2073-4425/11/3/300/s1, Figure S1: Ranking of eleven candidate reference genes in D. teres according to the parameter M computed by geNorm PLUS . The increase in stability of the candidate genes is determined by a decrease in the M value.