Phylogenetic and structural classification of chitin synthases (chs) and -deacetylases (cda) in Trichoderma atroviride

Blast analysis using known chitin synthase sequences from Aspergillus, Neurospora and Saccharomyces identified eight genes in Trichoderma atroviride (IDs and references, S4 Table), of which seven can be directly assigned to the proposed classes I-VII [35, 52, 53]. The chitin synthases were designated CHS1-7 (Fig 1A and S1 Fig), with CHS1-3 and CHS6 representing chitin synthases with the most simple structure with only a chitin_synth_I (PF01644) or a chitin_synth_II motif (PF03142), respectively [35, 54]. CHS4 is the homolog of ScChs3p, which accounts for the bulk chitin synthesis in yeast [31, 54, 55], and contains a cytb5-like domain in addition to a chitin_synth_2 motif (Fig 1A). The related CHS5 and CHS7 additionally harbor a Dek domain at the C-terminus and a myosin motor head domain (MMD), but only CHS5 retained the functional ATP binding and switch I and II motives [38] (Fig 1A and S1 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Phylogenetic and structural classification of chitin synthases (chs) and -deacetylases (cda) in Trichoderma atroviride.

Structure/function prediction of the (A) eight chitin synthases (CHS1-8) and (B) six chitin deacetylases (CDA1-6). On the left side a condensed phylogenetic tree (dendrogram) is shown for both enzyme groups. (C) Phylogeny and functional relation of all six identified chitin deacetylases from T. atroviride in comparison to other organisms. Three functional groups are highlighted: light blue, virulence/autolysis related chitin deacetylases; light green/green, vegetative and conidial development-related transmembrane CDAs group I/ II; yellow, CDAs putatively involved in self-/nonself recognition. Bootstraps above 80 are not indicated. The bar marker indicates genetic distance. Gene name abbreviations are described in S1 Fig. A complete phylogeny of CHS1-8 in filamentous fungi is depicted in the S1 Fig.

https://doi.org/10.1371/journal.ppat.1008320.g001

The eighth chitin synthase (CHS8) falls out of the classification (with a QRKRW motif) and is present in a separate group together with the already identified chitin synthase FgCHS8 from Fusarium graminearum. FgCHS8 was shown to be important for virulence and fungal cell wall sensitivity to environmental stress [56]. This new fungal class of chitin synthases was also identified in a genome-wide analysis [53, 57, 58] where it was grouped into the proposed CV and ESV subclasses, which share high homology to chloroviral chitin synthases [59, 60]. Our analysis provides evidence, that they are most closely related to Class VI enzymes (S1 Fig).

In addition, blasting two known conserved chitin deacetylase motives from S. cerevisiae identified six chitin deacetylases in T. atroviride (Fig 1B). Subsequent phylogenetic analysis of the chitin deacetylases, using sequences of the saprotrophic and phytoparasitic ascomycetes and mucorales as comparators, revealed that CDA1 and CDA5 cluster together with deacetylases related to virulence [43, 61, 62]. They are most closely related to the only two S. cerevisiae chitin deacetylases (ScCda1 and ScCda2, [63, 64]; Fig 1B and 1C) due to the presence of a signal peptide and conserved type I chitin binding domains (ChitBI motifs/ CBM18) at their N- and C-termini, suggesting enhanced affinity to chitin polymers. However, ScCda1 and ScCda2 group to a separate clade with proteins involved in vegetative development and conidial maturation [63–65]. Trichoderma CDA2 is present next to this clade in a smaller separate group with chitin deacetylases that are cell wall associated and expressed during vegetative development, but lack the ChitBI motif (e.g. in Magnaporthe oryzae [43]). In contrast to the Zygomycete species Amylomyces/Mucor rouxii or Cryptococcus neoformans [39, 65, 66] in T. atroviride only CDA2 possesses also signature characteristics of membrane proteins (L4 –Y23; TMHMM server 2.0; http://www.cbs.dtu.dk/services/TMHMM/ [67]). Trichoderma CDA4 and CDA6 are highly homologous (76% identity/83% similarity) and only TaCDA3 and TrCDA4 were also found in addition in this fourth separate branch (Fig 1C). CDA3 diverges the most from all other CDAs mainly because it harbors a very degenerate chitin deacetylase domain that even lacks the conserved, catalytic aspartic acid and histidine residues (Fig 1B).

The genomic environment of chs and cda reflects their specific roles

Identifying conserved gene clusters is a useful bioinformatics approach to get first indications on putative functionalities, since it can help identifying regulators and co-factors for a certain pathway [68–70]. Hence, we performed a synteny analysis of chitin synthase and deacetylase genes in T. atroviride. A comparison of the genetic neighborhood with the saprophyte T. reesei provided further insights into the conservation of the clusters in related Trichoderma species with different life styles. We found that the organization of the chs4 gene cluster identified in Trichoderma spp. strongly resembles that of the different Aspergillus species that were analyzed by Pacheco-Arjona et al. [69], (Fig 2A). chs4 shows a chromosomal head-to-head arrangement with csa1 (encoding the putative activator of class IV chitin synthases [29]), a gene organization typical for fungal secondary metabolism gene clusters or a coordinated transcriptional regulation of functionally related genes. chs2 and chs3 were also found on the same contig in T. atroviride and the same chromosome (VI) in T. reesei, but they are about 500 kb apart, so that co-regulation or shared regulatory elements are unlikely (T. atroviride: chs2: (contig_29) 659,292–663,807 chs3: (contig29) 1,153,752–1,157,194). While chitin synthases are involved in building up the cell wall, glycoside hydrolases, such as glucanases and chitinases, act antagonistically, by degrading the cell wall to allow cell growth. Indeed, a glucanase was identified upstream of chs4 directly after csa1. Moreover, also the gene encoding NAG1 (a secreted N-acetylglucosaminidase) that is responsible for mobilization of GlcNAc from chitobiose [71] is located only 55 kb upstream from chs4 in Trichoderma spp. Interestingly, the chitin deacetylase encoding gene cda5, was found in close vicinity to nag2 (the second, but membrane bound N-acetylglucosaminidase) [71] on another scaffold in T. atroviride (contig 23: cda5: 1,914,660–1,916,233; nag2: 1,943,411–1,945,301) and two genes coding for chitosanases (CHO2 and CHO4) are located next to it but are both missing in T. reesei. A closer investigation of the genomic neighborhood of chs5 showed that it most probably shares a bi-directional promoter with chs7 since they are in a head-to-head arrangement, which was observed also in other fungi (Fig 2B and [38]). Furthermore, a serine/threonine kinase gene is found in close proximity of chs5, and a histidine kinase and -phosphatase down- and upstream, respectively, which might play crucial roles in activation of the myosin head of the chitin synthases as proposed by Pacheco-Arjona et al., [69]. chs6 is located only 55 kb apart from chs7 and contains additional regulatory and accessory genes in close vicinity (Fig 2B). chs8 is located in a head to tail arrangement with cda1 and a UDP-N-acetylglucosamine 6-dehydrogenase gene (UNGD) is located further upstream, which might also imply co-regulation, even though via separate promoters (Kappel and Gruber, manuscript in preparation). A similar arrangement was found in all fungi harboring this specific kind of chitin (hyaluronan) synthase [57]. Therefore, the chromosomal arrangements hint at a close transcriptional connection and imply a tight post-translational regulation of degradation and build-up of the cell wall in Trichoderma spp. Further experiments with the additionally identified genes need to be carried out to verify the proposed putative functions from the cluster analysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Genomic environment of chs and cda reflects their specific roles.

(A, B) A cluster analysis was performed according to the chromosomal location of chitin synthase genes and their conservation among T. atroviride (Ta) and T. reesei (Tr) annotated at the JGI database for Trichoderma atroviride v2.0 (https://genome.jgi.doe.gov/Triat2/Triat2.home.html) and Trichoderma reesei v2.0 (https://genome.jgi.doe.gov/Trire2/Trire2.home.html). Conservation of exons (lilac), UTRs (untranslated regions, turquoise) and CNS (conserved non-coding sequences, red) is shown in the above panel of each diagram. Gene location on the chromosomes (T. reesei, [146]) and contigs (T. atroviride) is indicated in k = 1,000 nucleotides. Gene (white arrows) size and position are to scale and numbers correspond to gene description in S1 and S2 Tables. Chitin synthase genes and the activator are highlighted in red. (A) Gene cluster of chitin synthase (chs4) and the putative chitin synthase activator (csa1). (B) chs5-chs7-chs6 gene cluster. (C-D) Transcript levels of chitin synthase (left panel) and chitin deacetylase genes (right panel) during vegetative growth and conidiation. Transcript levels of chitin synthase genes (chs1-8) were related to chs1 and chitin deacetylase genes (cda1-6) to cda2 expression at 16 h of germination for (C) and 24 h of growth on PDA for (D). Normalized expression of chs1 and cda2 corresponds to ‘1’ as indicated (see also S2 Fig). (C) Expression analysis during germination after 0 h, 4 h (isotropic growth), 8 h (formation of germination tube) and 16 h (first vegetative hyphae). Pictures of corresponding germination stages are shown on the right-hand side; scale bar = 20μm. (D) Expression analysis during hyphal development from mycelium grown on PDA for 24 h and 48 h and asexual development at different maturation stages with (w) white/ nascent, (y) yellow/immature and (g) green/mature conidia; scale bar = 1 cm.

https://doi.org/10.1371/journal.ppat.1008320.g002

Transcriptional analysis during hyphal development (Fig 2C and 2D) showed that the chitin synthases and—deacetylases can be grouped into those with low and very low expression such as chs8, cda1, cda3 and cda5, intermediate expression (chs1, chs4, chs6, chs7, cda2, cda4) and high expression (chs2, chs3, chs5, cda6). During generation of asexual conidia (from early white to yellow and mature, green spores (Fig 2D) the expression of the chitin synthases decreased slowly and only chs1, chs2, chs3, chs5 and chs7 were still expressed at intermediate levels in mature/green conidia (Fig 2D). Transcription of cda2 and cda6 was found in early stages of conidial maturation, but decreased also during late maturation. Interestingly, cda3 transcription, that was very weak during all stages of hyphal growth, increased strongly during conidial maturation and might therefore be important in spore formation. High expression of chs5 and chs7 throughout hyphal development, vegetative growth and also conidial maturation was found (Fig 2C and 2D). Interestingly, cda1 and chs8 were only weakly expressed during vegetative development and also cda5 expression was nearly undetectable, indicating involvement in other possibly stress related processes. These results outline the importance of chitin synthases chs3, chs5 and chs7 and cda2, cda6 during vegetative development and cda3, as a crucial player in conidial maturation.

Differential expression of chs and cda upon environmental stresses

Perturbation of fungal cell wall synthesis triggers a compensatory response to ensure cell wall integrity (CWI) with, among others, increased chitin synthesis. These stress responses are transduced via the CWI pathway. Perturbances that negatively affect cell wall composition and cell integrity such as osmotic shock and high salt levels (HOG), reactive oxygen species (ROS), or cell wall synthesis inhibitors (CWSD) are first sensed by integral membrane proteins and relayed to transcription factors. This leads to up- and downregulation of genes encoding cell wall synthesizing and modifying enzymes in a highly adaptive and effective manner [72, 73]. It has been shown that chitin synthesis and chitosan conversion is also influenced by external environmental stress factors in e.g. Aspergillus spp., C. albicans and M. oryzae [74–77]. To see if transcriptional regulation of chitin synthases and deacetylases also plays a critical role in T. atroviride, we investigated fold-changes of expression due to changed culture conditions simulating different environmental stresses. Interestingly, the expression of chs1, chs3 and chs4, as well as cda2, which all were expressed throughout hyphal development, was not up-regulated upon any of the tested stress condition. Only chs6 and chs8 expression was increased up to 14-fold upon osmotic stress (Fig 3A). Remarkably, a hypothetical HSP88, family HSP110/HSP70 encoding gene, is located in a head to tail arrangement downstream of chs6 (Fig 2B). Among the deacetylases cda3, cda1 (up to 32-fold) and, to a lower extent, cda4, cda5 and cda6 (up to 21-fold) responded to HOG stress. A moderate response in transcription was observed towards cell wall disturbing agents, were especially chs2, but also chs5 and chs7 as well as cda1, cda3 and cda4 showed the strongest response with up to 6-fold increased transcript levels (Fig 3A). chs2 levels also increased moderately (5-fold) when more harsh treatments such as 40% glucose and SDS were applied, which eventually resulted in hyphal lysis (S3 Fig). cda4 (19-fold) and cda6 (10-fold), but also cda1 and cda3 levels (5.5-fold) were increased under these conditions (Fig 3A). During the mycoparasitic attack fungal hosts use defense strategies such as generation of reactive oxygen species [78]. Interestingly, ROS created from incubation with H₂O₂ resulted in very strong upregulation of most chitin synthases (48-fold, exceptions chs1, chs3, chs4) and chitin deacetylases (145-fold, exception cda2). In silico analysis of all chitin synthase promoters using known motives for transcription factor binding sites from S. cerevisiae corroborated our results from the transcriptional approach. An enrichment of stress related elements in the promoter region of the strongest stress responders from the expression analysis chs2, chs5, chs6, chs7 and chs8, in comparison to the other chitin synthases was confirmed (Fig 3B, S3 Table). By contrast, developmental elements were rather found in chs3, chs5 and chs7. The occurrence of oxidative stress elements was especially high in chs2 and chs8, whereas CWI-stress related elements were enriched in the MMD-domain chs5 and chs7 as well as in chs2 and the other two division I chitin synthases. Thus, our findings provide evidence that especially chs2 and chs6 as well as chitin deacetylase transcription, with the notable exception of cda2, is highly adaptive and that T. atroviride is well prepared for and capable of fast reaction to a changing environment and defense reactions by the host during the mycoparasitic attack.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Differential expression of chs and cda upon environmental stresses.

(A) Transcription analysis of chitin synthase (left panel) and chitin deacetylase (right panel) genes using various stressors. Comparison of growth under control conditions (glc1%, 1% glucose, ctrl) and with different stressors. On the left side the specific growth condition is indicated and on the right side a stress group is assigned: CWI (cell wall integrity disturbing stressors); CWSD (cell wall synthesis disturbing agents); HOG (high osmolarity glycerol) and ROS (reactive oxygen species). Detailed growth conditions are described in S3 Fig. The clustering results of transcription levels (values calculated by ΔΔC_T method) are represented in ‘jet’-style color mapping. (B) Graphical representation of the 1,200 bp upstream promoter regions of the eight chitin synthases and the identified stress element binding regions. Color code boxes indicate various stress groups whose height corresponds to the score level. CWI, HOG, ROS see (A); DEV, developmental regulation; OX, oxygen stress related; STRESS, general stress element; TATA, TATA-box or CAAT-box. Other detected and enriched motifs are marked with an asterisk. For more details on the promoter response elements see S3 Table.

https://doi.org/10.1371/journal.ppat.1008320.g003

Critical roles of chs and cda in growth and development are emphasized in single knockout mutants

To study the role of the genes identified, we generated corresponding knockout lines. We assumed that null mutants of non-redundant cell wall remodeling genes could repress growth or even be lethal given their importance for fungal development and survival. Nonetheless, we were able to generate single deletion strains of all chitin synthase and chitin deacetylase genes, except for cda4. The deletion of cda4 proved difficult even after several rounds of transformation (ectopic integration detected > 20). Whether this is due to the gene being essential or if the gene is located at an inaccessible position on the chromosome remains elusive.

Colony extensions of all deletion strains and their susceptibility to stress were assayed on solid medium (Fig 4A and 4B). Significant differences were observed after 3 to 7 days of growth between the WT and mutant strains, suggesting half of the chitin synthases are indispensable for optimal growth under the tested conditions. Especially the MMD-chitin synthase mutants Δchs5 and Δchs7 were severely compromised in development with irregular colony appearance and thin mycelial network formation. In addition, little or no conidial production with irregular size was observed (Fig 4C and 4D). These severe defects in group V and VII CHS have also been reported for other fungi and they play a crucial role in virulence in plant pathogens [38, 79]. Δchs1 and Δchs2 show a considerably reduced growth rate on PDA. But unlike chs5 and chs7 KO strains, division I synthase mutants form a dense hyphal matrix suggesting aberrant branching frequency or hyphal fusion in this area [80]. Δchs1, Δchs2, had strongly decreased amounts of spores and Δchs1, Δchs2, Δchs5 and Δchs7 produced mainly whitish to yellow spores, even when grown with 1.2 M sorbitol as osmotic stabilizer (S4B Fig), indicating a critical role of chitin biosynthesis in conidial maturation (Fig 4C and 4D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Critical roles of chs and cda in growth and development are emphasized in single knockout mutants.

(A) Growth of generated deletion mutants and the T. atroviride wild type strain (WT) on PDA after 72 h. (B) Colony extension in percent of the parental strain (wt). Growth on PDA only and on plates containing 1.2 M sorbitol (+sorbit), 200 μg/ml calcofluor white (+CFW) or 150 μg/ml congo red (+CR) was compared. (C, D) Generation of asexual conidia by all deletion mutants, see also S4A Fig; (C) spores/cm² and (D) average conidial size. (E-F) Compensatory transcription analysis of the transglycosylase genes crh1a, crh1b and crh2 in the Δchs (E) and Δcda (F) strains compared to expression in the WT strain after 25 h and 48 h of growth on PDA (tef1 used as housekeeping gene). Mean +/- SEM are shown and p>0.005 are indicated with *, upregulated; °, downregulated. (B-F) Data was generated from at least two independent experiments.

https://doi.org/10.1371/journal.ppat.1008320.g004

When strains were exposed to the cell wall intercalating dyes Calcofluor white (CFW) and Congo red (CR) to determine changes in the cell wall composition, the parental strain already showed reduced growth rates with about 60% inhibition. Interestingly, unlike other mutant strains tested in our experiments, group V and VII CHS cannot compensate for the cell wall intercalator (~95% inhibition; Fig 4B). Furthermore, all three division I chitin synthase knockout strains have severe growth defects (85–90% inhibition) on the cell wall dyes (Fig 4B). These findings corroborate our hypotheses drawn from the transcriptional profiles that these chs play fundamental functional roles in germination and vegetative development (Fig 2C and 2D). Interestingly though, the slower hyphal elongation of Δchs1, Δchs2, Δchs4 and Δchs6 was almost completely recovered in the presence of 1.2 M sorbitol, but only slightly in the Δchs5 and Δchs7 strains. Note that the parental strain was also very sensitive to sorbitol on solid medium (~ 50% reduction) whereas in liquid medium this effect is negligible (S3 Fig, [81]).

In S. cerevisiae ScChs3 chitin synthase activity is influenced by activators such as the ScChs4/ScSkt5 and other proteins such as ScChs5, ScChs6 and ScChs7 [29, 82, 83]. ScChs5/ScChs6 are involved in the polarized transport of ScChs3 in specialized vesicles (the so called exomer complex [84–86]), where ScChs5 seems to serve as the major adaptor for ScChs3 and ScChs7 controls export of ScChs3p from the endoplasmic reticulum (ER). Among these (additional) regulators, the Chs7 chaperone seems to be critical to mediate ScChs3 (TaCHS4) activity, which has been confirmed recently in N. crassa and F. oxysporum [87, 88]. We therefore identified the homologs of the two most important proteins in T. atroviride (pID 179314, 158601) and designated them Chitin Synthase Export chaperone, CSE7 (ScChs7/NcCSE7) as suggested by Rico-Ramirez et al., [87], and CSE5 (ScChs5) Chitin Synthase Exomer adaptor. The deletion of each of the genes affected growth on PDA (~ 30% reduced) and showed an obvious defect when grown on CFW (50–60% reduced compared to WT). The Δcse5 and Δcse7 built a highly layered, laminar mycelium with abundantly branched mycelial mats stacked upon each other leading to a very irregular colony diameter (Fig 4).

Among the chitin deacetylase knockouts, Δcda2 and Δcda3 were suppressed at an early stage of growth (0–36 h, S4A Fig), and the radial colony had a dense network formation phenotype similar to single Δchs1/ Δchs2. Both strains also produced significantly reduced amounts of spores (Fig 4C and 4D). Δcda6 showed a concrete green ring at the colony center with crystalline spores (Fig 4 and S4A Fig), and minor effects, when treated with CFW or CR, suggesting minor changes in polysaccharide architecture. All complemented strains (with the WT gene reintroduced at a random site or gene replacement) showed the WT phenotype and conidiation pattern (S5 Fig).

In order to find additional evidence that the lack of chitin synthases leads to divergence of cell wall structuring and therefore affects cell wall remodeling, polysaccharide assembly and linkage, we were interested in the three most important chitin-glucan transglycosylases in Trichoderma atroviride. In the course of our work we identified crh1a, crh1b and crh2, that are highly homologous to ScCrh1 and ScCrh2, which catalyze the cross-linking of chitin and β-1,6-glucan [45, 46]. In the chitin synthase mutants they were significantly down-regulated after 48 h of development, especially in Δchs5 and Δchs7 (Fig 4E). During this later stage of maturation, hyphal fusion normally produces a dense mycelial network, which induces high expression of transglycosylases. The significant down-regulation of crh1a in all investigated Δchs is an additional hint towards co-regulation and a strong evidence for altered cell wall morphology in chs mutants (Fig 4E). In contrast to Δchs, the transglycosylases were found strongly upregulated in the cda deletion strains, hinting at a compensatory action to decreased cell wall flexibility in these mutants [89, 90] (Fig 4F).

In conclusion, functional analysis of mutant strains strongly suggests that at least 5 chitin synthases and 3 chitin deacetylases are critical for vegetative development including conidial maturation. Our findings highlight the importance of chitin, and most probably chitosan, for cell wall integrity during asexual development and in a hostile environment.

Single knockout mutants show severe morphological changes and aberrant chitin deposition in patches with reduced chitin levels

Deletion of chitin synthases and—deacetylases affects cell wall composition and distribution of cell wall components (Fig 4; [27, 52, 65]) and as a result hyphal morphology is equally compromised. Microscopic analysis of mutant strain morphology using the inverted agar method [91] showed an aberrant hyphal development, with a weak and thin mycelial network and fewer branches in Δchs1 in comparison to the T. atroviride wild type. To identify growth characteristics that might be linked to aberrant chitin synthesis, epifluorescence microscopy was performed using the fluorochrome CFW, which stains β-1,4-linked polysaccharides such as chitin. In the wild type, CFW fluorescence was strongest at the septa and tips, where chitin is actively synthesized and weaker staining occurred at the lateral walls. Both, Δchs1 and Δchs2 had considerably thinner hyphae and branching started very late, and more distant from the tip than in the parental strain (Fig 5A and 5B, S4C Fig). In addition, Δchs2 showed increased CFW staining indicating higher deposition of chitin at the tip and lateral walls. Δchs4, Δchs6 and Δchs8 did not show pronounced morphological aberrations, but only a slight increase in hyphal tip staining (Fig 5A). Phenotypic characterization of T. atroviride Δchs5 and Δchs7 revealed the most severe defects. They showed short lateral hyperbranching alternated with long, thin filamentous hyphae without branches (S4C Fig) and grew irregularly, with strongly inhibited hyphal elongation and chitin accumulated in patches at the lateral walls (Fig 5A and 5B). Remarkably, Δchs5/ Δchs7 showed an elevated frequency in septation together with severe constrictions, possibly caused by a disproportionate distribution of chitin between the septa and the lateral walls. In addition, these mutants seemed inhibited in apical growth and often extensive isotropic tip swelling (increased sensitivity to CFW suggesting a cell wall weakening) was observed (Fig 5B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Single knockout mutants show severe morphological changes and aberrant chitin deposition in patches with reduced chitin levels.

(A) Chitin deposition and tip morphology in leading hyphae of the peripheral zone of each colony was visualized by CFW staining in epifluorescence microcopy using the inverted agar method for live-cell imaging [91]; scale bar = 20μm. (B) CFW staining highlights aberrant chitin deposition and septal development in selected deletion mutants; scale bar = 10μm. (C) Altered cell wall chitin deposition in chs and cda deletion strains compared to the WT was investigated by semiquantitative fluorescence microscopy using CFW and CR as stains. A defined area of the tip apex and subapex (shown in B) of leading hyphae was analyzed by densitometry using the ImageJ software platform (http://rsb.info.nih.gov/ij/). Mean +/- SEM from n = 25–60 hyphae are shown.

https://doi.org/10.1371/journal.ppat.1008320.g005

Microscopic examination of all cda mutants showed pronounced staining of CFW at their apices in comparison to the lateral walls suggesting aberrant chitin accumulation, resulting in a brighter Spitzenkörper staining in all cda mutant strains. Quantification of the ratio of the relative dye fluorescence between tip apex and subapex with CFW and CR (Fig 5C), which more specifically binds to chitin than CFW, clearly showed that significantly less dye was incorporated into the cell walls of all deletion mutants. Only Δchs8 showed CR staining comparable to the WT in the hyphal tips but staining was also reduced in the cell wall in the sub apical zone (Fig 5C). This change in the cell wall deposition pattern of tip apex to sub apex was also observed in all of the three class I chs as well as chs6 knockout and the Δcda. The highest ratio (2.2) was observed in the cda2 knockout strain for CFW (Fig 5B and 5C). Baker et al., also reported a strong accumulation of chitin and very low levels of chitosan in the chitin deacetylase deletion strains in C. neoformans [65].

Moreover, we adapted the biochemical assay for cell wall chitin determination described in this publication ([27, 65]; see Materials and methods) to get further insights in the impact on chitin synthesis in the cell wall of the Trichoderma mutants. The analysis of the vegetative hyphae of all mutants confirmed that a strong reduction of chitin in the cell wall was caused by the absence of most of the chitin synthases, cse5 and cda6, with the lowest amounts in Δchs5 and Δchs7 (48 and 37%), while deletion of the other deacetylases, or cse7 resulted only in a 84–93% (Table 1). Deletion of chs2, chs6 and cda5 did not impact chitin levels under the tested growth condition.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Relative GlcNAc content in the cell wall of mutant strains compared to the wild type.

https://doi.org/10.1371/journal.ppat.1008320.t001

Microscopic analysis of Δcda2, Δcda3 and Δcda6, Δcse5 and Δcse7 showed increased branching with dichotomous branching indicative of a dysregulated apical dominance and also more lateral branches (Fig 5A). Δcda5 revealed a loss of hyphal avoidance, and accumulation of lateral and cytoplasmic staining. In summary, in most of the Δchs (predominately Δchs2/5/7), and Δcda, patches of bright fluorescence were observed at the lateral wall of hyphae. This altered distribution of cell wall components seems to interfere drastically with polarized growth and branching patterns. Our results indicate that aberrant cell wall synthesis activities caused by absence of nearly any of the synthesizing enzymes lead to reduced chitin levels and altered distribution of chitin and putatively other cell wall polymers in the mutant strains, which strongly affects their hyphal development.

Mycoparasitism depends on coordinated expression of chitin synthases, -deacetylases and chitosanases

Mycoparasitism by Trichoderma has been studied already extensively and many critical events have been elucidated but what remains still unknown is how exactly mycoparasitic species remodel their own cell walls to circumvent the defense mechanisms of phytopathogenic fungi that themselves comprise an aggressive plethora of chitinolytic enzymes during the invasive process. We therefore first tested the expression of chitin and chitosan metabolic enzymes, when T. atroviride was confronted with itself or the fungal hosts S. sclerotiorum, B. cinerea and R. solani. Expression analysis revealed exclusively high induction of chs8 and cda1 expression under mycoparasitic conditions during confrontation with all hosts (Fig 6A). The concerted activation of both enzymes is possibly due to their genomic arrangement in a gene cluster driven by a shared regulatory element (Kappel and Gruber, manuscript in preparation). Interestingly, chs1, chs4, chs5 and chs7 were also moderately upregulated during the mycoparasitic attack next to their critical role in vegetative growth. The upregulation of chs5/ chs7 during mycoparasitism (Fig 6A) highlights the importance of the MMD containing synthases in any pathogenic event, either phyto- or mycoparasitic. Interestingly, cda3, cda4 and cda6 were induced in self-recognition of the fungus (control condition: Ta), indicating a general function in cell wall remodeling (synthesis and repair), and primary growth. Transcription of the stress responders chs2/6 and the genes for vegetative development chs3/cda2 was not significantly altered during mycoparasitism (Fig 6A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Mycoparasitism depends on coordinated expression of chitin synthases,–deacetylases and chitosanases.

Differential transcription analysis of chitin synthases, chitin deacetylases and chitosanases during the mycoparasitic attack of T. atroviride against three different hosts. Expression of all target genes was analyzed using the ΔΔC_T method. T. atroviride against itself (Ta), R. solani (Rs), S. sclerotiorum (Ssc) or B. cinerea (Bc). The WT grown alone on a plate (control condition) was arbitrarily set to ‘1’. Samples were collected at contact and after contact (4 mm overgrowth). The gene sar1 was used as the housekeeping gene. Data were generated from two independent experiments and three technical replicates. Mean +/- SEM and p>0.005 are indicated with *, upregulated; °, downregulated. (A) Transcription analysis in confrontation assays with fungal hosts, chs1-8, chitin synthase; cda1-6, chitin deacetylase. The clustering results are represented in ‘jet’-style color mapping as described in Fig 3A. (B) Relative expression of all identified chitosanase genes (cho1-cho6) during mycoparasitism.

https://doi.org/10.1371/journal.ppat.1008320.g006

As discussed, cell wall turnover, chitin crosslinking to glucan, as well as its conversion to chitosan might be critical for the antagonistic action of Trichoderma towards its prey. Equally important seems the role of chitosanases, which hydrolyze chitosan and chitooligosaccharides in host interaction, lysis and cell wall remodeling, an area currently of particular attention [47]. We identified 6 chitosanases in T. atroviride. CHO1 and CHO2 show high homology to CHO from Aspergillus spp. (EU302818.1, [92, 93]), that possess endo-chitinolytic properties and CHO3 is phylogenetically related to the chitosanase found in F. oxysporum (EGU78186.1, [94]), that is involved in virulence (S6A Fig). Expression analysis of all identified chitosanases during mycoparasitism revealed their importance especially during late stages, when T. atroviride invades the host cells. Particularly when confronted with R. solani, transcription of cho1- cho3 and cho5 was significantly upregulated (~ 60-fold) and might therefore be important for mobilization of chitosan from the host cell wall (Fig 6B and S6B Fig). A significant but lower induction of expression of these enzymes was also observed for the ‘after contact’ condition with S. sclerotiorum and B. cinerea, two ascomycetic hosts. Interestingly, non-antagonistic growth conditions by confrontation of T. atroviride with itself (control condition) also resulted in moderate transcriptional upregulation of cho1 to cho3 (Fig 6B), which is a strong indication for the presence of chitosan also in the CW of T. atroviride.

Cell wall chitin and chitosan affect the mycoparasitic attack and resistance to hosts

In order to assess the mycoparasitic activity of mutant strains we monitored their capacity to overgrow S. sclerotiorum, B. cinerea and R. solani. We therefore investigated the mycoparasitic interaction in dual-plate confrontation assays, in which T. atroviride strains (antagonists) were inoculated at a defined distance from one of the three hosts. After recognition, Trichoderma spp. hyphae attach to and coil around the host hyphae and hook-like appressoria penetrate the cell wall, which is followed by growth of the antagonist inside the host [95]. All cda knockout lines demonstrated severe defects in overgrowth of S. sclerotiorum (Fig 7A) and B. cinerea and less pronounced with R. solani (S7A Fig). This behavior of T. atroviride to parasitize closely related phytopathogenic Ascomycota [96, 97] is considered to be the major trait that sets Trichoderma apart from the other mycoparasitic Hypocreaceae that mainly attack Basidiomycota. Interestingly, while the parental strain was able to entirely overgrow S. sclerotiorum within 5 days (overgrowth is marked with a dotted line), the Δcda mutants completely stalled growth already during the early phase of interaction, i.e. growth towards the host fungus and establishment of direct contact (Fig 7A). A distinct contact zone was particularly seen in Δcda1 confrontations, which was also supported by microscopic analysis (Fig 7B). However, most cda deletion strains were able to overgrow S. sclerotiorum after prolonged incubation (14 days), suggesting that the deletion of cdas results in reduced chasing, coiling, and perhaps inability to invade the host. A strong defect in mycoparasitism was also observed for most of the chs mutants, with Δchs5 and Δchs1 being the most severely compromised KO strains. Again, mutants are delayed at contact and penetration of the hosts. A submerged growth towards the host was also observed indicating a loss of osmotic stability in such mutants. Thus, at contact, the tested Δchs1/2/5/7 are almost avirulent against S. sclerotiorum (Fig 7A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Cell wall chitin and chitosan affect the mycoparasitic attack and resistance to hosts.

(A) Dual confrontation assays: A 5 x 30 mm slice of fully overgrown PDA plates of the indicated T. atroviride WT and mutant strains (T) was placed on the left side of a PDA plate 5 cm apart from a slice of the host S. sclerotiorum (H, right side) and incubated for 5 days. Arrows indicate the zonal overgrowth (dotted line) by T. atroviride over S. sclerotiorum. (B) Microscopic analysis of the confrontation zone of. T. atroviride WT and selected mutant strains against S. sclerotiorum using CFW staining in the inverted agar method for live-cell imaging [91]; arrows indicate hyphal attachment; L, hyphal leakage; r, hyphal retreat; *, increased hyphal branching; scale bar indicated. (C) Plates from dual confrontation assay in (A) were photographed from the bottom to show the brownish contact zone, in cda mutants compared to the wild type. (D) Relative expression of mycoparasitic indicator genes, prb1, nag1, ech42 in the WT (wt, striped columns) and cda deletion (12356) mutants against S. sclerotiorum. Samples were collected at contact and after contact (4 mm overgrowth). WT alone (control condition) was arbitrarily set to ‘1’, (wt, striped columns). Expression data was normalized to the housekeeping gene sar1. Data were generated from two independent experiments and three technical replicates. Mean +/- SEM and p>0.005 are indicated with *, upregulated; °, downregulated. (E) Effects of antifungal metabolites and secreted enzymes from the host S. sclerotiorum on growth of T. atroviride WT and mutant strains. S. sclerotiorum was cultured on top of a diffusible cellophane membrane covering PDA plates and after 48 h of growth the cellophane disc was removed and plates containing the secreted active compounds were inoculated with an agar plug of Trichoderma WT and mutant strains. The percent growth inhibition by S. sclerotiorum secreted active extract was calculated from comparison with the growth of the respective strains on fresh PDA. A representative of three independent experiments is shown. (F) The susceptibility to oxidative stress of Δchs and Δcda strains was demonstrated on a solid medium containing H₂O₂. Colony extension in percent of the parental strain on minimal medium with (right panel) or without (left panel) the addition of 0.0075% H₂O₂ and growth for 72 h was determined. WT (white), chitin synthase mutants (grey), chitin deacetylase mutants (light grey) were compared. Mean +/- SEM from two independent experiments is shown.

https://doi.org/10.1371/journal.ppat.1008320.g007

To obtain a more detailed picture of the contact zone, microscopic analysis was performed on selected chs and cda deletion mutants that macroscopically showed the most prominent phenotype in the confrontation assay (Fig 7B). While the wild type attached and lysed the host S. sclerotiorum (H) the investigated mutants were unable to fully attack the host. Δchs5 and Δchs7 did not invade the host upon contact. Furthermore, Δchs1 and the myosin motor domain mutants Δchs5 and Δchs7, but also Δcda1 showed massive hyphal leakage (examples of deflated cell contents are indicated, Fig 7B(L)). The Δchs1 strain responded to host stress with extensive clamydospore development and in both, Δchs1 and Δcda1, intra-hyphal growth caused by the fungicidal effect on hyphae was observed (S7B Fig). Δcda1 was blocked in growth or retracted (Fig 7B(r)) upon contact with the host, and often balloon-like structures were visible. In Δchs8 an exorbitant and irregular apical branching was observed (Fig 7B(*)). No direct attachment could be observed, but lysis of the host.

Trichoderma inhibits or kills a host by parasitizing on its hyphae through cell wall degrading enzymes like chitinases (e.g. ech42), proteases (e.g. prb1) and glucanases [95]. Interestingly, although cda mutant strains seemed to be strongly compromised in host invasion, it seemed that they increased their production of lytic enzymes as a compensatory mechanism, which was visible by a browning of the growth medium most probably from generated oxidation products (Fig 7C). Transcriptional profiling of the mycoparasitic indicator genes supports this hypothesis by showing significantly higher prb1 and ech42 transcript levels in Δcda1-6 than in the WT (Fig 7D). Expression of ech42 was strongly induced at contact with S. sclerotiorum (∼13-fold) as well as prb1 expression in the Δcda2 and Δcda3 mutants (~5-fold). Expression of nag1 (N-acetyl-glucosaminidase), and after contact also prb1, were strongly downregulated in Δcda5 clones, indicative for a synergistic regulation of cda5 and these enzymes. Importantly, the absence of cda3, that was found massively induced during overgrowth of S. sclerotiorum (Fig 6A), resulted in a sustained induction of all indicator genes (Fig 7D), confirming its critical role in mycoparasitism.

Next we wanted to distinguish between an increased production of lytic enzymes by the chitin deacetylase mutants and their susceptibility to metabolites or active compounds, which are produced by the host and secreted into the growth medium. First all mutant strains and the wild type were grown on PDA plates covered with a cellophane disc to investigate the effect of secreted enzymes (and smaller diffusible factors) on the growth of the host (details are described in Materials and methods). S. sclerotiorum was placed on the plates after removal of the T. atroviride strains and growth was monitored.

As expected, S. sclerotiorum was highly susceptible to diffusible factors produced by the WT. Interestingly, growth was completely abolished on plates from mutant strain diffusible active compounds with the exception of Δcda6 (58.5% +/-6.9), Δcda5 (44.7% +/-5.9%) and Δchs8 (26.7% +/- 5.1), were growth of S. sclerotiorum was better compared to growth on the WT exudates (4.2% +/-1.9, and S7C Fig). Δchs3 (2.3% +/-0.4) had comparable growth to the WT. These results corroborate our findings and indicate that a perturbance of the cell wall composition results in elevated secretion of diffusible factors as defense strategy. Since cellophane allows passage of enzymes and small molecules an increase of secondary metabolite production is also very likely. In a second test series the susceptibility of mutant strains towards S. sclerotiorum secreted secondary metabolites and enzymes was tested. WT growth was reduced to 66% by the presence of exudates of S. sclerotiorum compared to growth on fresh PDA and growth of Δchs strains was further decreased to 30–45% (Fig 7E). Strikingly, the Δcda strains showed a pronounced growth reduction by 80 to nearly 100 percent. Thus, the deletion of the chitin synthases and deacetylases might affect cell wall composition in general leading to increased porosity with a negative impact on their resistance to extracellular active compounds produced by the host.

We further tested the sensitivity of the mutant strains towards H₂O₂ to simulate the presence of ROS (Fig 7F). In the past years the generation of reactive oxygen species in plant-pathogens, such as B. cinerea, M. oryzae, Claviceps purpurae and S. sclerotiorum, was discovered as conserved mechanism in virulence, infectious structure- and sclerotia formation and it has been speculated that ROS also serves as defense strategy during a mycoparasitic attack, similar to plant defense systems [78, 98–101]. Interestingly, growth of the chitin deacetylase knockout strains was most severely affected on plates containing H₂O₂ but also all Δchs mutants showed decreased growth rates in comparison to the control condition (Fig 7F). Since most of the cda genes and chs2, chs5—chs8 were also highly induced during ROS treatment in the wild type (Fig 3A) we conclude that chitin and especially chitosan play a critical role in scavenging ROS, which is putatively produced by the host as defense mechanism, and might contribute to the reduced virulence observed in the confrontation assays.

Strains and cultivation conditions

T. atroviride IMI206040 (teleomorph: Hypocrea atroviridis, (http://genome.jgi-psf.org/Triat2/Triat2.home.html) maintained on potato dextrose agar (PDA, BD, Franklin Lakes, USA), incubated at 28 °C with a 12h/12h light/dark cycle. Chemicals were obtained from Roth (Karlsruhe, Germany), Sigma (Sigma Aldrich, St. Louis, MO) and enzymes and kits were obtained from Thermo Fisher (Thermo Fisher Scientific, Waltham, MA USA) and Bio-Rad (Bio-Rad Laboratories, Hercules, CA, USA). For assessing the growth rate of WT and mutant strains mycelial agar plugs were placed on PDA, with or without hygromycin B (200 μg/ml) and incubated for 72 h. In order to test susceptibility towards CW disturbing agents and HOG on solid medium, PDA was supplemented with 200 μg/ml Calcofluor white (Fluorescent brightener 28, CFW), 150 μg/ml Congo red (CR) or 1.2M sorbitol. Sensitivity to oxidative stress provoked by hydrogen peroxide (0.0075%) was tested on SM medium [124] (2% agar agar, 2% glucose). Radial growth was measured and captured every 24 h. Conidial maturation, spore size and concentration were investigated after 7 days on PDA. Conidia were harvested in a defined volume of 0.9% NaCl /0.05% Tween-80 solution and counted with a Tecan reader (Spark multimode microplate reader, Tecan Group Ltd., Männedorf, Switzerland) in a cell counting chamber (Cell Chip, NanoEnTeK, Tecan). Conidia from different maturation stages were prepared as described in [70] for RNA preparation. For shake flask cultivations SM with 0.05% peptone and 1% glucose was inoculated with 1×10⁶ conidia/ml and cultivated at 28 °C and 250 rpm. To investigate the stress induced transcription replacement studies [70] were performed by transferring pre-cultures (WT) from liquid medium grown for 24 h to medium with different stressors: 10% and 40% glucose, 10% glycerol, 0.7 M NaCl, 1.2 M sorbitol, 0.025% SDS, 0.75% H₂O₂, 150 μg/ml CR, 200 μg/ml CFW, 15 μg/ml Caspofungin (CF). After 1 h and 24 h mycelia were harvested for RNA extraction and biomass determination, respectively.

Generation of knockout strains—Genetic manipulation of T. atroviride

Knockout lines were established using the split marker technique [125] with an E. coli hph- phosphotransferase marker cassette. The hph gene marker cassette is under control of the Trichoderma pki1 promoter and cbh2 [126] terminator, derived from plasmid pan7-1 [127] and cloned in pBluescript II KS/SK- backbone (unpublished). Primers (F1, F2, F5 and F6; S4 Table), were designed to amplify around 1.5 kb fragments up-und downstream of the ORF of the target gene. Primer combinations HY/F4 and F3/YG amplify half of the 5´or the 3´ regions of the hph gene plus 1.5 kb of the target gene 5´or 3´flanking region, respectively [125]. Purified single PCR-fragments were subjected to double joint PCR reaction [128] with primer combination F1 and YG for the 5´ part, and HY and F6 for the 3´part, respectively. All PCR reactions were performed using the Phusion High-Fidelity DNA Polymerase (Thermo Fisher). Cassettes were fused in one single reaction and purified (to avoid unspecific integration), using the GeneJET Gel Extraction Kit (Thermo Fisher). Purified cassettes in equimolar amounts were used for protoplast [129] and spore mediated fungal transformation, which was adapted for T. atroviride from a protocol described for T. reesei in [130]. Briefly, freshly harvested spores were incubated in YPD (1% (w/ v) glucose) at 30°C, 300 rpm for 6 hours for conidial swelling, then carefully washed with decreasing amounts of 1.1 M ice cold sorbitol until a final volume of 75 μl per reaction. Split marker inserts (3–6 μg) were added and subjected to electroporation by use of the Biorad MicroPulser Electroporator (Bio-Rad) with 1.8 kV for ca. 5.5 ms. Regeneration followed with YPD/ 1.1 M sorbitol over night at 28°C before plating on PDA with 200 μg/ml hygromycin B. Positive transformants were confirmed by PCR. A set of PCR amplifications (KO_Fw/KO_Rv; combination with YG and HY, S4 Table) was used to discriminate homologous recombination from ectopically inserted constructs, parental genotypes or heterokaryons (S4D Fig). Candidates were purified employing single spore isolation until mitotic stability. As mock control, for all experiments the empty hph plasmid was transformed into the WT.

Complementation strains (ReChs or ReCda) were generated by reintroducing the WT sequence flanked by ca. 1,600 bp native promoter and terminator sequences into the genome of the mutant by random integration or gene replacements by co-transformation with plasmid p3SR2 [131]. Transformants were selected on acetamide containing medium and purified by single spore isolation (S5A and S5C Fig). A list of all mutant lines generated in this study is presented in S6 Table.

Isolation and purification of DNA

A rapid DNA purification protocol [132] and an isolation protocol for highly purified genomic DNA were used [133].

Dual confrontation assays

Confrontation assays of T. atroviride WT, chs and cda deletion strains against Rhizoctonia solani, Botrytis cinerea, and Sclerotina sclerotiorium were performed on PDA as described in [104] with approximately 5 cm distance of parasite and host. Plates were incubated for five days in circadian illumination and photographed. As controls, T. atroviride confrontation with itself or non-challenged T. atroviride were used. For RNA extraction confrontation assays were performed on PDA covered with sterile cellophane discs. T. atroviride mycelia were harvested before contact (BC), and after contact (AC) and immediately frozen in liquid nitrogen.

Growth inhibition by metabolites and secreted enzymes

Susceptibility to secreted secondary metabolites and proteins of T. atroviride and mutant strains was tested in plate assays according to [134] with some modifications. S. sclerotiorum was inoculated on PDA plates covered with a cellophane disc for 72 h in the dark at 25°C, allowing the produced metabolites to diffuse through the membrane into the agar. Note, that cellophane not only allows the passage of low-molecular-weight metabolites, but also active enzymes [135]. The cellophane disc was removed when the mycelium covered about 3/4 of the plate and an agar plug of T. atroviride wild type or mutant strains was placed in the middle of the plate to assess the mutants putatively altered susceptibility. The growth diameter was recorded after 52 h and growth inhibition was calculated in % compared to growth of the corresponding WT or mutant strains on fresh PDA. The same procedure was followed to investigate the secretion of metabolites and enzymes of T. atroviride WT and mutant strains and their potential to inhibit S. sclerotiorum growth. The antifungal activity against S. sclerotiorum was evaluated by comparing growth to fresh PDA plates in %. Data was generated from two independent experiments in both set-ups.

RNA isolation and cDNA synthesis

For RNA extraction ca. 100 mg of mycelium was ground to a fine powder with glass beads and a bead mill (2x 60s at 6 m/s; Bead Ruptor 24 Elite Bead Mill Homogenizer, Omni International, VWR). Total RNA from liquid shaking culture and from spores was isolated using the guanidinium thiocyanate method [136]. When RNA was extracted from mycelium grown on plates the GeneJET Plant RNA Purification Kit was used. Isolated RNAs were treated with DNAse I, and purified with the GeneJET RNA Cleanup and Concentration Micro Kit. cDNAs were generated with the Revert Aid H-minus cDNA synthesis kit, using 1 or 5 μg RNA. All kits and enzymes for RNA purification and cDNA synthesis were purchased from Thermo Fisher.

Gene expression analysis

qRT-PCR reactions were performed in a Biorad iCycler iQ (Bio-Rad) and a Rotor-Gene 6000 (QUIAGEN, Venlo, Netherlands) as described previously [137]. Primers and protein IDs from the DOE Joint Genome Institute database are listed in S5 Table. Relative gene transcript levels were quantified and normalized to the corresponding signals of sar1 (differentiation between high sequence homologies of T. atroviride and hosts in confrontation assays), and tef1 [138, 139]. The fold change relative to the control conditions was calculated using the ΔΔC_T method [140] with REST software [141]. All samples were analyzed from at least two independent experiments with three technical replicates.

Hyphal GlcNAc content assay

To measure the chitin content in the cell wall of wild type and mutant strains the protocol described by [27] was adapted for Trichoderma atroviride. Strains were grown on PDA plates covered with a cellophane disc until a diameter of approximately 2–2.5 cm was reached. 40–50 mg mycelium was harvested into tared 2-ml microcentrifuge tubes containing 2 large and 100 μl small glass beads (2.85–3.45 and 0.75–1.0 diameter, respectively) and dried overnight to determine dry weight (typically 5–7 mg). The cell walls were extracted with 1 ml 6% KOH at 80°C for 90 min. Samples were centrifuged at 14,000 rpm for 20 min, and the supernatants were used to indirectly determine dry biomass via a protein content assay with Bradford reagent (Biorad). A linear correlation of protein content to biomass had been established with defined amounts of dry weight, previously. Pellets were washed 3x in 1 ml 1xPBS, and 1x with 1 ml of Mc Ilvaine’s buffer (0.2 M Na2HPO4, 0.1 M citric acid, pH 6.0) and frozen at 20°C or directly processed further. 5μl of Chitinase from Trichoderma viride lyophilized powder, ≥600 units/g solid enzyme mix (5 mg/ml in PBS, Sigma) were added to 200μl of Mc Ilvaine´s buffer to hydrolyze chitin to GlcNAc. Samples were incubated for 20–24 h at 37°C. For colorimetric determination the protocol given by [27] was followed.

Microscopic analysis

Strains grown on PDA for 18 h were applied in a droplet of 30μl CFW stain (20μM, Sigma #F3543) with the inverted agar method [91] and imaged with an inverted Zeiss Axio Observer Z1 (Zeiss, Oberkochen, Germany) with differential interference contrast optics and 405nm excitation/ 430–470nm emission. Germlings were investigated on an Olympus CX33 (Olympus, Hamburg, Germany) after 0, 4, 8, and 16 h of germination in liquid PDB at 28°C, 300 rpm. Visualization and semi-quantification of Trichoderma cell wall chitin was performed using the following cell wall stains: Calcofluor White (CFW) M2R at a final concentration of 2 μM to non-specifically label β-1,4-glucans including chitin and Congo Red (CR, Sigma #C6767) at a final concentration of 50 μM to very specifically label α- and β-chitin [115]. Fluorescence signals were assessed using CFW and CR. All samples were incubated for 5 min before imaging. In order to compare hyphae from parental and mutant strains, all samples were measured by densitometry using the MacBiophotonics ImageJ work package available at (https://www.macbiophotonics.ca/software.htm) as described in [115]. Defined rectangular areas were measured in the Spitzenkörper and in the lateral cell walls of the subapical zone. On average 25–60 analysis were performed per hyphae.

Transcriptional response element detection

Literature search for stress response elements for different cellular processes in fungi found four main groups whith specific elements: CWI (CDRE, RML1, crzA, Crz1p, CRZ1-sn2/4, CRZ1-2, CreA), DEV (ARE, AREanalog, Are2/Nit2, BrlA), HOG (Sko1p, AtfA-1) and ROS (Hap1, Yap1, AtfA-2,CRE-CRZ-Mn). Additional fungi jaspar matrixes, most due to yeast studies were used. Scanning of the chs1-8 promoters showed specific elements for groups of general stress (TEC1, HSF1, STP4) and oxidative stress (ROX1, SKN7, STB5, AFT2). For more details see S3 Table. For scanning of sequences against patterns the Regulatory Sequence Analysis Tools (RSAT) http://rsat-tagc.univ-mrs.fr/rsat/ was used. For the eight chs promoters (1,200 bp upstream) sequences against our own and jaspar patterns only hits with a score < 5 were discarded. Finally, overlapping motives and hits on the reverse strand were joined. Similar enriched motives in groups of chs promoters (e.g. chs5,7) were detected using RSAT. Additionally, the 50 most enriched motifs for the four condition groups CWI, ROS, HOG, FUS (here using T. atroviride and related species orthologues) were detected with Meme http://meme-suite.org/. The eight chs promoters from T. atroviride were then scanned with the 200 motives using RSAT. Further, a correlation to known jaspar motives was calculated for functional hints by RSAT.

Bioinformatics and statistics

For transcriptional analyses, the hierarchical clustering algorithm Hierarchical Clustering Explorer 3 (HCE3) [142] with average linkage and Euclidean distance measure was applied and values were quantitatively illustrated using the grey scale scheme in Fig 2 and color scheme in Figs 3 and 6. The colors were selected based on the jet color mapping for the results of the function y = a*1/e^b*^x+c, where a, b, c are constants that were chosen so that value differences are easy to compare visually and x represents the relative expression value that was calculated with the REST software [141]. The values of a, b, c and the according formulas are freely available at https://github.com/mamut-m/expression-colormap. Statistical analysis for qRT-PCR was performed with REST software with a pair wise fixed reallocation randomization test. For all other experiments were statistical analysis was needed Student´s t-test was used, assuming unequal variance of groups. At least two biological and three technical replicates were used for statistical analysis. Phylogenetic analyses were performed in Mega7 [143], using Neighbor Joining, a distance algorithmic method. Stability of clades was evaluated by 1,000 bootstrap rearrangements. Structure/function prediction was performed using InterProScan [144], signalP 5.0 [145], and transmembrane domain prediction tool TMHMM 2.0 [67].