APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Oct. 2011, p. 7217–7226 0099-2240/11/$12.00 doi:10.1128/AEM.06027-11 Copyright © 2011, American Society for Microbiology. All Rights Reserved.
Vol. 77, No. 20
Differential Regulation of Orthologous Chitinase Genes in Mycoparasitic Trichoderma Species䌤† Sabine Gruber, Christian P. Kubicek, and Verena Seidl-Seiboth* Research Area Gene Technology and Applied Biochemistry, Institute of Chemical Engineering, Vienna University of Technology, Vienna, Austria Received 29 June 2011/Accepted 12 August 2011
Mycoparasitic Trichoderma species have expanded numbers of fungal subgroup C chitinases that contain multiple carbohydrate binding modules and could thus be important for fungal cell wall degradation during the mycoparasitic attack. In this study, we analyzed the gene regulation of subgroup C chitinases in the mycoparasite Trichoderma virens. In addition to regulation by nutritional stimuli, we found complex expression patterns in different parts of the fungal colony, and also, the mode of cultivation strongly influenced subgroup C chitinase transcript levels. Thus, the regulation of these genes is governed by a combination of colonyinternal and -external signals. Our results showed completely different expression profiles of subgroup C chitinase genes in T. virens than in a previous study with T. atroviride, although both fungi are potent mycoparasites. Only a few subgroup C chitinase orthologues were found in T. atroviride and T. virens, and even those showed substantially divergent gene expression patterns. Microscopic analysis revealed morphogenetic differences between T. atroviride and T. virens, which could be connected to differential subgroup C chitinase gene expression. The biological function of fungal subgroup C chitinases therefore might not be as clear-cut as previously anticipated. They could have pleiotropic roles and might be involved in both degradation of exogenous chitinous carbon sources, including other fungal cell walls, and recycling of their own cell walls during hyphal development and colony formation. chitinases of T. atroviride and T. virens (TAC2/TVC2, TAC4/ TVC4, TAC5/TVC5, TAC6/TVC6, and TAC7/TVC7) (5). This suggests the emergence of paralogues by gene duplications and a strong evolutionary pressure on this group of proteins. Other families of glycoside hydrolases, e.g., cellulases in T. reesei, are largely coregulated on the transcriptional level (4), and a similar picture was found for sgC chitinase gene expression in T. atroviride (5). All of these genes were induced during mycoparasitism of the fungal prey Botrytis cinerea, but none by Rhizoctonia solani, although both fungi can be well parasitized by T. atroviride. Correspondingly, all of them were expressed during growth on B. cinerea cell walls, but not on R. solani or T. atroviride cell walls. In contrast to cellulases, which are solely involved in substrate degradation for nutritional purposes, chitinases have multiple functions in fungal biology. They are involved in exogenous chitin degradation or, in the case of mycoparasites in a more aggressive form thereof, cell wall degradation during the mycoparasitic attack. However, chitinases are also important for the fungus during its life cycle by being involved in cell wall remodeling and recycling and hyphal branching and fusion events. Thus, the functions of chitinases range from morphogenetic roles in fungal colony development to exogenous chitin degradation for nutritional purposes. It is not yet understood how these functions are distributed among or within the different chitinase subgroups. The aim of this study was to analyze the potential coregulation of chitinases in T. virens. They form the largest glycoside hydrolase family in mycoparasitic Trichoderma spp., and T. virens has an especially rich arsenal of them. In this study, we focused on the 15 sgC chitinases of T. virens. Our results show that sgC chitinase transcription in T. virens not only is triggered by nutritional stimuli, but also strongly varies within different
Several species of the fungal genus Trichoderma (sexual form, Hypocrea) are mycoparasites and are thus able to antagonize, parasitize, and kill other fungi. An integral part of the mycoparasitic attack is degradation of the prey’s cell wall by chitinases, glucanases, and proteases (1). Genome analysis revealed that the numbers of chitinase-encoding genes are strongly expanded in mycoparasitic Trichoderma spp. in comparison to other filamentous fungi. Fungal chitinases belong to glycoside hydrolase family 18 (GH 18) and can be further subdivided into 3 different subgroups, A, B, and C (10, 11). Subgroup A chitinases in general contain no carbohydrate binding modules (CBMs), but subgroup B chitinases frequently have a CBM at the C-terminal end. All subgroup C (sgC) chitinases have multiple CBMs and have either two CBM 18 (chitin binding) domains or one CBM 18 and two CBM 50 (LysM) domains N terminal of their GH 18 module (5). CBMs enable efficient adherence of the enzyme to insoluble substrates and processive catalytic cleavage of the substrate (2). The additional chitinases in mycoparasitic Trichoderma spp. are members of subgroups B and C. While the saprotrophic species Trichoderma reesei has only 4 sgC chitinases, T. atroviride has 9 and T. virens has 15, suggesting the potential involvement of these chitinases in mycoparasitism (7). Surprisingly, only five orthologue pairs can be found in the sgC
* Corresponding author. Mailing address: Research Area Gene Technology and Applied Biochemistry, Institute of Chemical Engineering, Vienna University of Technology, Gumpendorfer Strasse 1a/ 166-5, 1060 Vienna, Austria. Phone: 43-1-58801-166554. Fax: 43-158801-17299. E-mail:
[email protected]. † Supplemental material for this article may be found at http://aem .asm.org/. 䌤 Published ahead of print on 19 August 2011. 7217
7218
GRUBER ET AL.
APPL. ENVIRON. MICROBIOL. TABLE 1. Overview of T. virens subgroup C chitinasesa
a Chitinases with orthologues in T. atroviride are highlighted in boldface. Protein IDs from the JGI database are given. Modular structure: 2 CBM18 (white boxes) and 1 GH 18 (black box) or 2 CBM50 (gray boxes), 1 CBM18 (white box), and 1 GH 18 (black box) (5).
types of hyphae and modes of cultivation. This indicates that colony-internal signals due to hyphal differentiation and interaction are significant parameters for the regulation sgC chitinase gene expression. Our results show that the two main functional aspects of chitinases—morphological development and exogenous chitin degradation—cannot be viewed separately but are strongly linked on the regulatory level and thus probably also on the functional level. MATERIALS AND METHODS Strains and cultivation conditions. T. virens (teleomorph, H. virens) strain Gv29-8 (http://genome.jgi-psf.org/TriviGv29_8_2/TriviGv29_8_2.home .html) (DDBJ/EMBL/GenBank accession number ABDF00000000) and T. atroviride IMI206040 (http://genome.jgi-psf.org/Triat2/Triat2.home.html) (accession no. ABDG00000000) were used in this study and were maintained on potato dextrose agar (PDA) plates (BD Difco, Franklin Lakes, NJ). Confrontation assays and liquid cultures were performed as described previously (5). Briefly, the plant pathogens R. solani and B. cinerea were used as fungal hosts for confrontation assays on PDA. Growth in liquid cultures was performed in static cultures in petri dishes in order to enable the fungus to efficiently adhere to the insoluble substrates predominantly used, such as cell walls and chitin. T. virens was inoculated in Vogel’s minimal medium (VMM) (15) and T. atroviride in minimal medium (MM) (12), both including 0.05% peptone for efficient germination. For cell wall stress experiments, calcofluor white (F3543-1G; SigmaAldrich, St. Louis, MO) and Congo red (L456540; Merck, Darmstadt, Germany) were added after autoclaving of the medium to the final concentrations given in Results. All cultures were inoculated in darkness at 25°C, and mycelia were harvested at different time points, as indicated in Results. At least two independent experiments were performed for each growth condition, with biological replicates (i.e., agar plates or static cultures in petri dishes, respectively) for each
time point. Gene expression data from one representative experiment are shown for the different time points and cultivation conditions. RNA isolation and RT-PCR. Mycelia were ground to a fine powder under liquid nitrogen. Total RNA was isolated as previously described (5). Note that for cultivations on agar plates, no cellophane overlays were used, but total RNA was extracted directly from the agar. All isolated RNAs were treated with DNase I (Fermentas, St Leon-Rot, Germany) and purified using the RNeasy MinElute cleanup kit (Qiagen, Hilden, Germany). Reverse transcription (RT)-PCR was performed as described previously using 25 cycles per reaction (5) with the gene-specific primers listed in Table 1. Bioinformatics analyses. For comparison of gene and protein sequences and synteny analysis, the tools available at the JGI genome website (http://www.jgi .doe.gov/) and the NCBI BLAST tools (http://blast.ncbi.nlm.nih.gov/Blast.cgi) were used. The results of the RT-PCRs were visualized by agarose gel electrophoresis and analyzed semiquantitatively by densitometry using the software ImageJ (http: ///rsb.info.nih.gov/ij). The expression levels of sgC chitinase genes were quantified and normalized to the corresponding signals of tef1 (encoding translation elongation factor 1␣; JGI database protein identifier [ID] 83874), which was used as a control gene in our studies. The normalized values were subsequently used for multiple clustering. For semiquantitative illustration, the Hierarchical Clustering Explorer 3 (HCE3) tool was used (13). This program supports five different linkage methods: average, average group, complete, single, and one-by-one linkage. In order to find meaningful groups of conditions under which sgC chitinases are upregulated, average group linkage was used to identify sgC genes similarly expressed under various cultivation conditions. For all analyses, Euclidean Distance Measure was used. Hierarchical-clustering results are displayed as grayscale mosaics attached to dendrograms. Further, the k means clustering algorithm tool was used to generate exactly k different clusters with the greatest possible distinction. Grouped cultivation conditions were visualized as a grayscale mosaic according to levels of expression. The grayscale color code corresponds to levels of expres-
VOL. 77, 2011
CHITINASE GENE REGULATION IN TRICHODERMA
TABLE 2. Summary of the properties of T. atroviride sgC chitinasesa
7219
TABLE 3. Comparison of the orthologous sgC chitinase pairs in T. virens, T. atroviride, and T. reesei % Identity Compared proteins or genes
Proteinsa TVC2 vs. TVC4 vs. TVC5 vs. TVC6 vs. TVC7 vs. TVC7 vs.
TAC2 TAC4 TAC5 TAC6 TAC7 CHI18-8
Genesb tvc2 vs. tac2 tac6 vs. tvc6 tvc7 vs. tac7 a
For the modular structure, see Table 1, note a.
sion, starting from no expression (white boxes) to maximal expression levels detected in this study (black boxes). Microscopic analyses. For morphological characterization, liquid cultures supplemented with carbon sources (chitin, cell walls, and glucose) were prepared. Two hundred fifty microliters of each culture was placed on microscope slides, incubated for 25 and 42 h, and imaged with an inverted T300 microscope (Nikon, Tokyo, Japan). An M420 Photomacroscope (Wild-Leica, Solms, Germany) was used for imaging the mycelial mat in liquid static cultures. Images were captured with a Nikon DXM1200F digital camera and digitally processed using Photoshop CS3 (Adobe, San Jose, CA).
RESULTS Comparison of the orthologous pairs of sgC chitinases between T. atroviride and T. virens. T. atroviride and T. virens have 9 and 15, respectively, members of sgC chitinases but share only five orthologues (Tables 1 and 2) (5). We were therefore interested in finding parallels between the orthologous pairs in T. atroviride and T. virens in order to elucidate the inducing stimuli and regulation of sgC chitinases in more detail. SgC chitinases in Trichoderma have two types of modular architecture (Table 1) (5), and of the orthologue pairs in T. atroviride and T. virens, only TAC7/TVC7 belong to the clade of sgC chitinases that have two LysM motifs and one CBM 18 N terminal of the GH 18 module. TAC2/TVC2, TAC4/TVC4, TAC5/TVC5, and TAC6/TVC6 contain only two CBM 18 modules N terminal of the GH 18 module. Interestingly, only for TAC7/TVC7 has an orthologue also been detected in T. reesei, whereas T. reesei has no sgC chitinases of the type with no LysM module. T. reesei and T. virens, which are taxonomically more closely related to each other than to T. atroviride, have two further orthologue pairs, TRCHI18-1/TVC1 and TRCHI18-9/TVC13, for which no corresponding orthologue was detected in T. atroviride. Thus, there is no correlation between the modular architectures and orthologous pairs of T. atroviride and T. virens. A closer comparison of the sequence similarities between sgC chitinases of T. atroviride and T. virens showed that TAC2/ TVC2 and TAC7/TVC7 are strongly conserved on the protein level, while the other orthologous pairs, TAC4/TVC4, TAC5/ TVC5, and TAC6/TVC6, show only a moderate level of conservation (Table 3). It should be noted that in these cases, the CBMs and GH 18 module show a higher degree of conserva-
% Alignment coverage
97 97 95 98 100 99 95 50 99
Amino acid
DNA
93 39 67 69 93 88
% Positive (aa)
95 53 79 79 96 94 96 75 93
a Protein IDs: TVC7, 53606; CHI18-8, 108346; TAC7, 247300; TVC2, 191688 (updated); TAC2, 348134; TAC4, 348132; TVC4, 81573; TVC5, 112098; TAC5, 348128; TAC6, 53627; TVC6, 348129. b The DNA genomic sequences of the complete genes, including 200 bp of the up- and downstream regions, were used.
tion than the parts of the proteins outside these defined regions. For the strongly conserved protein pairs TAC2/TVC2 and TAC7/TVC7, even at the DNA level, 93 to 96% identities were observed (Table 3). An alignment of 1,000 bp of the up- and downstream regulatory regions of the respective genes showed that for tac7-tvc7, 340 bp upstream of the ATG are strongly conserved, and that after a gap of ca. 80 bp, the following ⬎600 bp are also almost identical between the two fungi. The conserved region reaches into the gene upstream of the tac7-tvc7 locus, encoding a protein with a concanavalin A-like lectin/ glucanase IPR008985 domain (T. virens protein ID 53596). Further, approximately 160 bp downstream of the tac7-tvc7 locus are strongly conserved. For tac2-tvc2, the situation is similar, with 160 bp with high similarity upstream of the ATG, followed by a nonhomologous gap of ca. 160 bp and then a continuous conservation of ⬎700 bp reaching into the neighboring gene, encoding a LysM protein (T. virens protein ID 28703). We also analyzed whether the orthologue pairs are located in syntenic regions in T. atroviride and T. virens. Interestingly, this was not the case. All five of them are located in small, nonsyntenic clusters of 2 to 5 genes within large syntenic regions (Fig. 1). It was also conspicuous that all five chitinases were close to the ends of scaffolds in T. virens. TVC5 and TVC2 were located near the beginning and end, respectively, of scaffold 6 in the genome database, and TVC4, TVC6, and TVC7 were found on relatively short scaffolds. Gene expression of sgC chitinases in T. virens. We focused on three different topics in the investigation of sgC chitinase gene expression in T. virens: (i) mycoparasitism-related growth conditions (confrontation assays and growth on cell walls), (ii) consumption of chitinous carbon sources under nonmycoparasitic conditions (growth on chitin and starvation), and (iii) hyphal development and colony formation. T. atroviride sgC chitinase genes were found to be inducible only by a small set of the screened growth conditions in our previous study (Fig. 2a). We expected a similar on/off situation for these genes in T.
FIG. 1. Synteny of the genomic regions containing sgC chitinase orthologues in T. atroviride and T. virens, generated with the synteny tool in the JGI database. Three hundred fifty kilobases up- and downstream of the respective chitinase genes in T. atroviride located on the indicated contig were compared to the T. virens genome. The scaffolds covering the respective regions are shown in color-coded bars. The regions containing the chitinases indicated above the respective diagrams are circled. 7220
FIG. 2. Gene expression profiles of sgC chitinases in T. virens and T. atroviride. The grayscale ranges from white (no expression) to black (maximum expression). CA, confrontation assays; CW, cell walls; BC, before contact; C, contact; AC, after contact. (a) Gene expression profiles of sgC chitinases in T. atroviride. In order to facilitate the comparison between T. atroviride and T. virens, data from reference 5 that are directly comparable with data presented in this study were converted to a grayscale mosaic. Note that none of the T. atroviride sgC chitinase genes was inducible by R. solani or T. atroviride. (b to d) Gene expression profiles of sgC chitinases in T. virens. (b) k means clustering (k ⫽ 3) of growth conditions. min, minimum; max, maximum. (c) Clustering of genes using average group linkage. (d) Phylogenetic tree of sgC chitinases of T. reesei, T. atroviride, and T. virens (data from reference 5) in which the overall expression rates of the respective T. virens and T. atroviride genes are indicated as follows: ⫹⫹⫹, strong expression under many growth conditions; ⫹⫹, strong to moderate expression under many growth conditions; ⫹, moderate to weak expression under a few growth conditions; ⫺, no expression in our data set. Note that the genes of orthologue pairs are not necessarily expressed under the same growth conditions. The two subclades containing a strongly active and an inactive T. virens sgC chitinase are encircled.
7221
7222
GRUBER ET AL.
virens, but surprisingly, we obtained complex expression patterns under most growth conditions. The expression profiles of the 15 T. virens sgC chitinase genes from 35 different growth conditions and time points were analyzed by RT-PCR, measured by densitometry, and converted into a grayscale mosaic (see Fig. S1 in the supplemental material). Cluster analysis of the results (see Materials and Methods for details) is shown in Fig. 2b and c. When we used k means clustering to partition the growth conditions into three groups (Fig. 2b; k ⫽ 3), all confrontation assays, with a defined set of four expressed chitinase genes, grouped together in cluster II, but clusters I and III contained strongly mixed growth conditions with fragmented expression patterns. This showed that in T. virens, nutritional stimuli did not appear to be the main regulators of sgC chitinase gene expression under most tested growth conditions. Dendrograms obtained by cluster analysis of the growth conditions were too divided for interpretation (data not shown). Clustering of genes (Fig. 2c), using the average group linkage algorithm, resulted in clades of chitinase genes with similar expression patterns. The results obtained from the expression profiles are evaluated in detail below, but overall, it can be clearly stated that T. virens sgC chitinase genes have completely different expression patterns than T. atroviride sgC genes. Not even for the orthologue pairs were correlations observed, although, as stated above, two of the genes, namely, tvc2 and tvc7, are 96 and 93% identical to those of T. atroviride at the DNA level. Under most of the tested conditions, expression of tvc4 was detected, and the genes tvc1, tvc8, tvc13, and tvc15 were not found to be expressed at all. Figure 2d shows a schematic representation of the expression patterns of T. virens sgC chitinases in their phylogenetic context. TVC1 and TVC13 have apparent orthologues in T. reesei, but their genes are not expressed; however, in both cases, a T. virens chitinase in the same subclade is strongly expressed. TVC8 is closely related to TVC14, but again, only one of them is expressed. For TVC15, no protein model could be developed, and it is therefore not included in the tree. This indicated that tvc1, tvc8, and tvc13 are probably remnants from gene duplications and are thus inactive pseudogenes. Mycoparasitism. SgC chitinase gene expression during mycoparasitism was tested in confrontation assays with the ascomycete B. cinerea and the basidiomycete R. solani. As controls, T. virens was (i) confronted with itself and (ii) grown alone on a plate without a host. For comparison with other growth conditions described below, it is important to note that by analogy to previous experiments with T. atroviride (5), only ca. 5 mm of the hyphal area that was growing toward the prey fungus was harvested (see Fig. S2a in the supplemental material). Expression patterns in T. virens were substantially different from those previously reported for T. atroviride (5). In T. atroviride, all sgC chitinase genes were induced during confrontation with B. cinerea but none with R. solani or in the control. In T. virens, a defined subset of sgC chitinase genes (tvc2, tvc3, tvc4, and tvc10) was induced in all confrontation assays, including in the control confrontations of T. virens against itself, but, interestingly, not when T. virens was grown alone on an agar plate under the same conditions, even including the asymmetrical placement of the colony (Fig. 2). Growth on cell walls of B. cinerea, R. solani, and T. virens led to expression patterns that were completely different from
APPL. ENVIRON. MICROBIOL.
those in the confrontation assays (Fig. 2). Eleven out of the 15 T. virens sgC chitinase genes (tvc2, tvc3, tvc4, tvc5, tvc6, tvc7, tvc9, tvc10, tvc11, tvc12, and tvc14) were found to be induced by B. cinerea cell walls. Since the other four genes (tvc1, tvc8, tvc13, and tvc15) were not expressed under any of the tested growth conditions, it can be said that “all” sgC chitinase genes were induced by B. cinerea cell walls, similar to T. atroviride. However, in T. virens, different subsets of sgC genes were also induced by R. solani cell walls (tvc4, tvc6, tvc7, and tvc10) and T. virens cell walls (tvc2, tvc3, tvc4, tvc7, and tvc9). The gene tvc10 was found to be induced under mycoparasitism-related conditions, but not when grown on glucose or cell walls of T. virens itself as a control. Chitin and starvation. T. virens sgC gene expression was analyzed with colloidal chitin and the nonpretreated and thus less accessible practical-grade chitin (which also may contain protein impurities) as carbon sources (Fig. 2). In contrast to growth on cell walls, where strong mycelial development was observed, growth on chitin was poor (see Fig. S3 in the supplemental material). Clear differences in expression sets between colloidal chitin and crude chitin were observed (tvc2, tvc3, tvc4, tvc5, tvc6, tvc7, and tvc10 for colloidal chitin versus tvc4, tvc7, and tvc10 for crude chitin). SgC gene expression analysis of T. virens cultures under carbon starvation showed that several sgC genes were induced by starvation, but again, a different subset (tvc3, tvc4, tvc5, tvc6, tvc7, tvc9, tvc10, and tvc12) than under all previously tested growth conditions. Hyphal development and colony formation. We also tested the expression of sgC chitinase genes during growth and hyphal development in two types of cultivation (Fig. 2): (i) Starting from a spore suspension and harvesting the mycelial mat at different time points and (ii) starting from an agar plug with mycelium and harvesting the peripheral and central hyphal zones of the colony (see Fig. S2b in the supplemental material). By analogy to growth on cell walls and chitin, both cultivations were carried out in liquid static cultures, but with glucose as the carbon source. Additionally, gene expression in conidia was tested. In conidia, only tvc4 was expressed, which was also found under almost all other tested conditions. In young mycelium (15 h), starting from a spore suspension, we found that all but the four seemingly inactive sgC chitinase genes were expressed. Expression patterns decreased to only tvc4, tvc6, tvc7, and tvc9 at 25 h, and their transcript abundances decreased strongly at later time points. The observed expression pattern at 65 h, where the onset of autolysis can be assumed, differs strongly from that during starvation (0.05% peptone but no carbon source) but bears some similarity to growth on T. virens cell walls, as tvc4, tvc7, and tvc9 were also expressed under these growth conditions. When the inner and peripheral hyphal zones were investigated, again, all 11 apparently active chitinases were expressed in the inner zone of the fungal colony but only a few of them in the peripheral hyphal zone, and those were at lower levels. In the context of these results, it should be taken into account that in the confrontation assay controls, only the first few millimeters of the hyphae were harvested from agar plates (see Fig. S2a and b in the supplemental material), and no sgC chitinase genes were expressed. The results from gene expression analysis during hyphal development and colony for-
VOL. 77, 2011
CHITINASE GENE REGULATION IN TRICHODERMA
7223
mation, therefore, clearly showed that in T. virens, sgC chitinase genes are abundantly expressed during “normal” growth, i.e., in the absence of a potentially inducing carbon source, but their expression patterns vary strongly with the mode of cultivation and the hyphal zone that is harvested. Since the same cultivation conditions were used for T. atroviride, the possibility that the observed differences are due to the experimental setup can be excluded. Differences in biomass formation, hyphal growth, and colony morphology between T. virens and T. atroviride. The most conspicuous detail of the gene expression analysis in T. virens was the abundant expression of several chitinases during growth without the presence of potentially inducing carbon sources. In order to elucidate why the respective sgC gene expression patterns of T. virens are so substantially different from those of T. atroviride, we investigated a number of parameters, such as the influence of the medium composition, morphology, and biomass formation, in more detail. Different standard media are usually used for cultivation of T. atroviride and T. virens (3, 12), and attempts to grow T. virens in T. atroviride MM led to low biomass formation (see Fig. S4 in the supplemental material). Upon cultivation of T. virens in MM with glucose, we found a slightly altered expression profile: tvc4 was not expressed, but tvc6, tvc7, and tvc9 were still expressed, and in addition, tvc2, tvc3, tvc12, and tvc14 were very weakly induced (data not shown). This indicates that the slow growth on this medium slightly altered the expression patterns but resulted in expression of even more sgC chitinase genes. In contrast, when T. atroviride was grown in VMM, the expression patterns of sgC genes were identical to those during growth in MM (data not shown). Therefore, although change of the medium revealed some differences in sgC expression profiles, it did not explain the complex and abundant expression patterns observed in T. virens in comparison to the simple patterns in T. atroviride. Thus, we concluded that the medium composition itself cannot explain the observed differences between T. atroviride and T. virens. Further, it should be noted that for confrontation assays, the same medium (PDA) was used for both fungi. Next, we investigated the morphological development of the two fungi at different time points and with various media and carbon sources. Microscopic analysis revealed clear differences in hyphal development between T. atroviride and T. virens grown on glucose independent of the medium used. Young T. virens hyphae were strongly branched and showed irregular constrictions, whereas T. atroviride hyphae grew as straight, long, thin tubes, with very few branches (Fig. 3a). Strong macroscopic differences in the formation of the mycelial mat in static cultures were also observed. While T. atroviride formed a uniform mycelial mat when starting from a spore suspension, T. virens initially formed small clumps and clusters (Fig. 3b), indicating different hyphal network formation due to morphogenetic parameters. Cell wall stress is not a significant regulatory center for sgC chitinases. In order to test whether stimuli involved in cell wall formation and remodeling could influence sgC chitinase expression, we tested the influence of cell wall stress on sgC gene expression. The dyes Congo red and calcofluor white intercalate with glucan and chitin polymers and therefore interfere with fungal cell wall formation. Addition of these agents to the medium causes cell wall stress and was shown to lead to in-
creased chitin polymerization and thicker cell walls (9). The aim was to analyze whether chitinase expression patterns are altered during cell wall stress. T. virens was found to be very sensitive to Congo red and calcofluor white, and while for other fungi concentrations in the range of 50 to 200 g/ml of these substances are used (9, 14, 17), for T. virens, addition of more than 5 g/ml Congo red to the medium with the spore suspension in liquid, static cultures led to such a drastic decrease in growth that comparison with normal growth was not possible. Unfortunately, growth on calcofluor white was completely abolished even at 1 g/ml under these growth conditions. The results with Congo red (Fig. 4a) showed that chitinase patterns were not different but that the expression levels were higher under cell wall stress. In this experiment, the sgC chitinase genes tvc4, tvc6, and tvc7, which were found to be induced in glucose cultivations (see above), showed enhanced expression. Starting from an agar plug in a static liquid culture, somewhat higher concentrations of the cell wall stress reagents were tolerated, but growth was strongly delayed and colonies with an irregular shape were formed; therefore, only the central hyphal zones were harvested (Fig. 4b). Similar to the control, a large set of genes was expressed, and the response of most sgC chitinase transcript levels to calcofluor white was different from that to Congo red. Only tvc4 was induced by both cell wall stress reagents, and only tvc5 and tvc11 were clearly downregulated by both. These results show that cell wall stress has an influence on sgC chitinase expression in T. virens, supporting our hypothesis that sgC genes are regulated by colony-internal morphological parameters. However, cell wall stress does not appear to be the main factor for upregulation of these genes. Further, these data showed that Congo red and calcofluor white have different modes of action during this process. DISCUSSION In this study, we investigated the gene expression of sgC chitinases in T. virens. Mycoparasitic Trichoderma spp. have significantly higher numbers of sgC chitinases than, e.g., the saprotrophic species T. reesei, and it was thus previously hypothesized that they are involved in the mycoparasitic attack (6). The results presented in this study shed a completely new light on sgC chitinase regulation and showed only very few parallels between T. atroviride and T. virens. Even orthologous genes showed completely different expression patterns. All five orthologous sgC chitinase pairs are located in small, nonsyntenic islands within regions of high synteny, and in T. virens, they are located rather close to the ends of scaffolds. This could indicate their proximity to repetitive regions that favor genomic rearrangements. For tvc1, tvc8, and tvc13, no expression was detected, but phylogenetically closely related sgC chitinases were (strongly) expressed. These data point to gene duplications with subsequent inactivation of one of the copies, and tvc1, tvc8, and tvc13, and possibly also tvc15, could therefore be inactive pseudogenes. With respect to mycoparasitism, confrontation assays of T. virens with itself led to the induction of strictly the same chitinases that were also induced by mycoparasitism of R. solani and B. cinerea, whereas when T. virens was alone on an agar plate, no expression of sgC chitinases was observed in the
7224
GRUBER ET AL.
APPL. ENVIRON. MICROBIOL.
FIG. 3. Growth morphology of T. atroviride and T. virens in static cultivations after 25 h. (a) Microscopic images. (b) Photographic images.
hyphal front. This suggests that in T. virens, nonmycoparasitic inducers derived from living fungal cell walls triggered the expression of these genes. This is in strong contrast to T. atroviride, where during confrontation with itself and with R. solani, none of the sgC chitinase genes was induced. Thus, the expression pattern observed during the confrontation assays in T. virens is, interestingly, not specific to a mycoparasitic interaction but, rather, for sensing the presence of another fungus
in a more nonspecific way. In this context, it should be mentioned that, while T. atroviride can always (partially) overgrow itself, T. virens colonies do not come into physical contact when grown in constant light or in a 12-h light/12-h darkness cycle but are able to (partially) overgrow each other when cultivated in constant darkness (see Fig. S5 in the supplemental material). This further underlines the different lifestyles and sensing mechanisms of the two mycoparasites.
VOL. 77, 2011
CHITINASE GENE REGULATION IN TRICHODERMA
7225
FIG. 4. Gene expression analysis of sgC chitinases during cell wall stress. (a) Static cultivations starting from a spore suspension in VMM with glucose (control) and supplemented with Congo red (5 g/ml). Cultures were harvested after 48 h. Expression levels were measured by ImageJ software (optical density [OD]) and normalized to tef1. (b) Static cultivations starting from a conidiated agar plug in VMM with glucose (control) and supplemented with Congo red (5 g/ml) and calcofluor white (10 g/ml). The central hyphal zones were harvested, and gene expression levels were analyzed as described for panel a.
In T. virens, growth on fungal cell walls as carbon sources led to sgC expression profiles substantially different from those observed in confrontation assays. This suggests a different utilization potential of the cell walls that leads to different inducing stimuli, which could be due to, e.g., limited accessibility of the chitin, melanization of the cell wall, or general differences in cell wall composition. Our data showed that in T. virens, sgC chitinase genes are differentially expressed in various zones of the fungal colony (i.e., the central versus the peripheral hyphal zone in liquid, static cultures) without a potentially inducing carbon source, thus indicating that colony-internal signals derived from hyphal differentiation influence the expression of these genes. It can be expected that these internal stimuli overlap with external nutritional stimuli and together regulate sgC chitinase gene expression. We were therefore not able to find typical “inducing” growth conditions as is usually the case for glycoside hydrolases. Why was this strong influence of morphogenetic parameters apparently not observed in T. atroviride? There are two possible explanations, which are not mutually exclusive: either the regulation of these genes is substantially different in the two species or the observed differences in growth morphology are pivotal for the regulation of sgC chitinase genes. Some of the results are probably due to differences in the regulatory regions of the respective genes. It should be considered, how-
ever, that tvc2 and tvc7 are, even in their regulatory regions, almost identical at the DNA sequence level to their T. atroviride orthologues, with the exception of a small gap in the promoter regions. Nonetheless, we found strongly divergent transcript patterns. In view of the fact that the complete genes and large parts of the regulatory regions show ⬎90% identity, we consider it unlikely that the strong differences in the obtained expression profiles could be explained solely by the short nonhomologous gaps in the promoter regions. Speciesspecific transcription factors could therefore be responsible for the differential regulation. Genome analysis showed large sets of transcription factors that were unique to either of the two species. Differences in signal transduction and transcriptional activation could therefore be an important factor in the observed differences in sgC chitinase gene regulation between T. atroviride and T. virens. In order to test whether the regulation of sgC chitinases is species specific, heterologous expression of these genes in the other species could be carried out. This approach could also show whether the differential transcription of genes of orthologue pairs has an effect on the growth morphology of each species. In addition to the regulatory aspects, we hypothesize, based on our observations, that the observed morphogenetic differences between T. atroviride and T. virens are connected to sgC chitinase regulation via colonyinternal inducing stimuli. Further supporting this hypothesis,
7226
GRUBER ET AL.
we found that cell wall stress influences the expression of these genes. Cell wall stress has an effect on the cell wall composition, i.e., increased chitin content, and is also accompanied by altered hyphal morphology and colony formation (as also observed in this study). Calcofluor white and Congo red are nonspecific fluorochromes that bind to cellulose and chitin polymers in the cell walls of fungi (9, 16). The strong toxicity of calcofluor white and Congo red for T. virens was somewhat surprising in view of the fact that the organism is a mycoparasite and thus potentially has well-protected cell walls. However, for the biopolymer chitosan, which has antibacterial and antifungal properties due to its free amino groups, an increased sensitivity was also already reported for mycoparasitic Trichoderma spp. in comparison to, e.g., entomopathogenic or plant-pathogenic fungi (8). Calcofluor white and Congo red cause the formation of thicker cell walls and lead to the upregulation of chitin synthases (9). The fact that we found altered, but not strictly increased, sgC expression patterns in the presence of these cell wall stress agents suggests that sgC chitinase expression is not coupled to chitin synthesis rates. The growth of T. virens on glucose led to abundant sgC chitinase expression, which indicates that sgC chitinases are actually multipurpose chitinases and are involved in the degradation and recycling of (exogenous) chitin from dead and living hyphal material. This is in agreement with the fact that fungi with a less aggressive life style, e.g., saprotrophic fungi, also have several sgC chitinases. Although increased numbers of sgC chitinases were found in the genomes of mycoparasitic Trichoderma spp., they are not exclusively induced during mycoparasitism, but rather, their expression data indicate a general involvement in various degradation processes of chitinous carbon sources, including their own and other fungal cell walls. Our data link for the first time the two main functional aspects of chitinases—morphogenetic development and exogenous chitin degradation—and also underline a potential interplay of these two aspects in the mycoparasitic attack. ACKNOWLEDGMENTS This work was supported by the FWF Austrian Science Fund (T390 and P20559 to V.S.-S.). The work conducted by the U.S. Department
APPL. ENVIRON. MICROBIOL. of Energy Joint Genome Institute is supported by the Office of Science of the U.S. Department of Energy (contract no. DE-AC0205CH11231). REFERENCES 1. Benítez, T., A. M. Rincon, M. C. Limon, and A. C. Codon. 2004. Biocontrol mechanisms of Trichoderma strains. Int. Microbiol. 7:249–260. 2. Boraston, A. B., D. N. Bolam, H. J. Gilbert, and G. J. Davies. 2004. Carbohydrate-binding modules: fine-tuning polysaccharide recognition. Biochem. J. 382:769–781. 3. Djonovic´, S., M. J. Pozo, and C. M. Kenerley. 2006. Tvbgn3, a beta-1,6glucanase from the biocontrol fungus Trichoderma virens, is involved in mycoparasitism and control of Pythium ultimum. Appl. Environ. Microbiol. 72:7661–7670. 4. Foreman, P. K., et al. 2003. Transcriptional regulation of biomass-degrading enzymes in the filamentous fungus Trichoderma reesei. J. Biol. Chem. 278: 31988–31997. 5. Gruber, S., et al. 2011. Analysis of subgroup C of fungal chitinases containing chitin-binding and LysM modules in the mycoparasite Trichoderma atroviride. Glycobiology 21:122–133. 6. Karlsson, M., and J. Stenlid. 2008. Comparative evolutionary histories of the fungal chitinase gene family reveal non-random size expansions and contractions due to adaptive natural selection. Evol. Bioinform. Online 4:47–60. 7. Kubicek, C. P., et al. 2011. Comparative genome sequence analysis underscores mycoparasitism as the ancestral life style of Trichoderma. Genome Biol. 12:R40. 8. Palma-Guerrero, J., H. B. Jansson, J. Salinas, and L. V. Lopez-Llorca. 2008. Effect of chitosan on hyphal growth and spore germination of plant pathogenic and biocontrol fungi. J. Appl. Microbiol. 104:541–553. 9. Roncero, C., and A. Duran. 1985. Effect of calcofluor white and Congo red on fungal cell wall morphogenesis: in vivo activation of chitin polymerization. J. Bacteriol. 163:1180–1185. 10. Seidl, V. 2008. Chitinases of filamentous fungi: a large group of diverse proteins with multiple physiological functions. Fungal Biol. Rev. 22:36–42. 11. Seidl, V., B. Huemer, B. Seiboth, and C. P. Kubicek. 2005. A complete survey of Trichoderma chitinases reveals three distinct subgroups of family 18 chitinases. FEBS J. 272:5923–5939. 12. Seidl, V., B. Seiboth, L. Karaffa, and C. P. Kubicek. 2004. The fungal STRE-element-binding protein Seb1 is involved but not essential for glycerol dehydrogenase (gld1) gene expression and glycerol accumulation in Trichoderma atroviride during osmotic stress. Fungal Genet. Biol. 41:1132– 1140. 13. Seo, J., and B. Shneiderman. 2002. Interactively exploring hierarchical clustering results. Computer 35:80–86. 14. Song, W., et al. 2010. R-SNARE homolog MoSec22 is required for conidiogenesis, cell wall integrity, and pathogenesis of Magnaporthe oryzae. PLoS One 5:e13193. 15. Vogel, H. J. 1956. A convenient growth medium for Neurospora (Medium N). Microbiol. Genet. Bull. 13:42–43. 16. Wood, P. J. 1980. Specificity in the interaction of direct dyes with polysaccharides. Carbohydr. Res. 81:271–287. 17. Zhang, S., Y. Xia, and N. O. Keyhani. 2011. Contribution of the gas1 gene of the entomopathogenic fungus Beauveria bassiana, encoding a putative glycosylphosphatidylinositol-anchored beta-1,3-glucanosyltransferase, to conidial thermotolerance and virulence. Appl. Environ. Microbiol. 77:2676–2684.
