Histone acetyltransferase TGF-1 regulates Trichoderma atroviride secondary metabolism and mycoparasitism

Some filamentous fungi of the Trichoderma genus are used as biocontrol agents against airborne and soilborne phytopathogens. The proposed mechanism by which Trichoderma spp. antagonizes phytopathogens is through the release of lytic enzymes, antimicrobial compounds, mycoparasitism, and the induction of systemic disease-resistance in plants. Here we analyzed the role of TGF-1 (Trichoderma Gcn Five-1), a histone acetyltransferase of Trichoderma atroviride, in mycoparasitism and antibiosis against the phytopathogen Rhizoctonia solani. Trichostatin A (TSA), a histone deacetylase inhibitor that promotes histone acetylation, slightly affected T. atroviride and R. solani growth, but not the growth of the mycoparasite over R. solani. Application of TSA to the liquid medium induced synthesis of antimicrobial compounds. Expression analysis of the mycoparasitism-related genes ech-42 and prb-1, which encode an endochitinase and a proteinase, as well as the secondary metabolism-related genes pbs-1 and tps-1, which encode a peptaibol synthetase and a terpene synthase, respectively, showed that they were regulated by TSA. A T. atroviride strain harboring a deletion of tgf-1 gene showed slow growth, thinner and less branched hyphae than the wild-type strain, whereas its ability to coil around the R. solani hyphae was not affected. Δtgf-1 presented a diminished capacity to grow over R. solani, but the ability of its mycelium -free culture filtrates (MFCF) to inhibit the phytopathogen growth was enhanced. Intriguingly, addition of TSA to the culture medium reverted the enhanced inhibition growth of Δtgf-1 MFCF on R. solani at levels compared to the wild-type MFCF grown in medium amended with TSA. The presence of R. solani mycelium in the culture medium induced similar proteinase activity in a Δtgf-1 compared to the wild-type, whereas the chitinolytic activity was higher in a Δtgf-1 mutant in the absence of R. solani, compared to the parental strain. Expression of mycoparasitism- and secondary metabolism-related genes in Δtgf-1 was differentially regulated in the presence or absence of R. solani. These results indicate that histone acetylation may play important roles in the biocontrol mechanisms of T. atroviride.

Introduction of the glycosyl hydrolases, or prb-1, which codes for a basic protease of the subtilisin type, provide more protection to plants against root and foliar pathogens [28][29][30][31].
Intriguingly, the cooperative effect between lytic enzymes and antibiotics to disrupt cell walls has been proposed [32]. In this sense, Trichoderma spp. use non-ribosomal peptide synthetases (NRPS) to produce linear peptides called peptaibols, which exert antibacterial, antifungal, and occasionally antiviral activities [33][34][35]. The peptaibol synthetase PBS-1 of T. atroviride contains 19 typical peptide synthetase modules with the required additional modifying domains at its N-and C-termini [36]. In fungi, the 4-phosphopantetheinyl transferase (PPTase) activates enzymes involved in primary and secondary metabolism [37]. In T. virens, a Δppt1 mutant was unable to synthesize peptaibols, and was also incapable of inhibiting the growth of phytopathogenic fungi and oomycetes [38]. Moreover, in T. virens, a secondary metabolite deficient mutant showed a reduced expression of genes whose products are related to secondary metabolism. In agreement with these results, the fungus was incapable of synthetizing the antibiotics viridiol and viridin [39].
The search to identify regulators of mycoparasitism and secondary metabolism-related genes in the presence of external stimuli, such as the presence of a host, needs to consider the way that leads to chromatin modifications in Trichoderma.
In this work, we analyzed the effect of Trichostatine A (TSA), an inhibitor of HDACs from class I and II [40], in mycoparasitism, in the production of antibiotics, and in the expression of mycoparasitism-and secondary metabolism-related genes in T. atroviride during its interaction with the phytopathogenic fungus Rhizoctania solani. The role of TGF-1, the orthologous to Gcn5 of S. cerevisiae in T. atroviride, on mycoparasitism and on the transcriptional regulation of mycoparasitism and secondary metabolism-related genes was also assessed. Our results provide new insights into the biological control mechanisms of phytopathogens by Trichoderma.

Microorganisms and growth conditions
T. virens Gv29-8 [41], T. atroviride IMI206040 and Δtgf-1 (Uresti-Rivera et al., in preparation) were used throughout this study. Rhizoctonia solani, Trichoderma citrinoviride and Trichoderma harzianum strains were isolated from a tomato field in San Luis Potosi, Mexico (22˚38' 39.84" N, 100˚50' 56.4" W), and identified by PCR amplification of the internal transcribed spacer (ITS) of ribosomal deoxyribonucleic acid (rDNA) sequences, using the oligonucleotides ITS1 and ITS4 [42]. Fungal strains were routinely grown in potato dextrose agar (PDA) or potato dextrose broth (PDB) (both from Difco™, BD Becton, Dickinson and Company, New Jersey, USA), as indicated for each experiment. When indicated, 300 nM of TSA (Sigma-Aldrich, Taufkirchen, Germany) was added to the media. All the strains were incubated at 28 C at the indicated times for each experiment.
rpm. The mycelium was centrifuged 5 min at 2000 rpm and used to inoculate PDA plates at 3 cm from the middle, and allowed to grow for 72 h before inoculating with the other fungal strains, as described above. Two independent experiments were performed in triplicate. Photographs were taken at the indicated times. In addition, dual cultures of T. atroviride wt or Δtgf-1 against R. solani were carried out on PDA plates with or without TSA (300 nM). The different fungal strains were grown alone on PDA plates with or without TSA as controls of the interactions.
Antimicrobial activity of T. atroviride wt and Δtgf-1 mycelium-free culture filtrates versus R. solani T. atroviride wt and Δtgf-1 strains were grown in PDB or Vogel minimal medium supplemented with 1.5% sucrose [43] in the presence or absence of 300 nM TSA for 7 days. Then, cultures were filtered using Corning 1 500 ml bottle top vacuum filter, 0.2 μm-pore 33.2 cm 2 nylon membrane (Corning Life Sciences, Massachusetts, USA). Mycelium-free culture filtrates (MFCF) were used to prepare 1 × PDA plates containing 60% of MFCF and 40% of sterile distilled water. A mycelial plug of an actively growing colony of R. solani was inoculated on the center of a PDA-MFCF plate, incubated at 28˚C, and radial growth was measured after 12, 24, 36, 48, and 60 h post-inoculation. PDA plates without MFCF inoculated with R. solani were included as controls. Two independent experiments were performed in triplicate.

Induction of Trichoderma cultures with R. solani mycelium
R. solani mycelium was grown in PDB medium for 7 days and filtered as described for MFCF. After filtration, R. solani mycelium was lyophilized, frozen in liquid nitrogen, and ground. For the induction of T. atroviride wt or Δtgf-1, 0.7 g of R. solani sterilized powder was added to 500 ml of PDB previously autoclaved. The induction medium was inoculated with 10 plugs (0.5 cm 2 ) of T. atroviride wt or Δtgf-1 strains and incubated at 28˚C with agitation at 250 rpm for 7 days in the darkness. The wt and Δtgf-1 induced MFCF were obtained by filtration as described above for R. solani. The mycelium was discarded and the MFCF were concentrated by lyophilization. The wt and Δtgf-1 lyophilized samples were resuspended in 500 μl of phosphate buffer, and protein concentration was determined as described by [44].

Relative expression of mycoparasitism (ech-42 and prb-1) and secondary metabolism-related genes (tps-1 and pbs-1) in dual culture experiments
Dual cultures of T. atroviride wt and Δtgf-1 against R. solani were performed as described above, but placing a sterile cellophane sheet on the PDA plates before inoculation of the fungi. Dual cultures were incubated for 36, 48, and 60 h at 28˚C. Mycelia of the different Trichoderma strains were collected at the indicated times, frozen in liquid nitrogen, and ground for total RNA extraction by the Trizol 1 method (Invitrogen, New York, USA). Total RNA (5 μg) was treated with Turbo DNase as described by the manufacturer (Ambion, Life Technologies, New York, USA). cDNA was synthesized using the Superscript 1 kit (Invitrogen), following the manufacturer's protocol. Subsequently, cDNA was quantified using a Nanodrop spectrophotometer (ND-1000, Thermo Scientific, Delaware, USA). Relative gene expression was assessed by quantitative reverse transcription PCR (RT-qPCR), using specific oligonucleotides (Table 1) in an Applied Biosystems 7500/7500 Fast Real-Time PCR System, using the Fast SYBR 1 Green Master Mix (Applied Biosystems, Foster City, California), in a final volume of 20 μl per reaction. Reactions were set up by using the standard settings of the system, except for the oligonucleotides annealing temperature (Table 1), which was 62˚C. Two independent experiments were performed in triplicate. Relative genes expression was calculated using the 2 (ΔΔ-Ct) method [45].

Δtgf-1 versus R. solani
For SEM analysis, dual cultures of T. atroviride wt and Δtgf-1 strains against themselves or against R. solani were performed as described in the previous section. For SEM analysis, 0.5-cm 2 agar pieces of the different confrontations at the indicated times were fixed in 3% glutaraldehyde in PBS (phosphate buffered saline) for 2 h. Samples were rinsed three times with cold PBS for 15 min each. Afterwards, samples were post-fixed with 1% OsO 4 in PBS solution for 1 h, followed by three rinses in PBS, 15 min each, and then dehydrated in graded ethanol. Critical point drying was done in a Tousimis Samdri-PVT-3D, mounted, and gold-coated sputtering in the Cressington Model 108 auto, and examined in a FEI model Quanta 200 SEM. The SEM was adjusted at 18 kV, spot 5, WD 10 mm, and the photomicrographs were taken with Everhart Thornley Detector (ETD).

Proteinase activity
Proteinase activity was determined in a 500-μl reaction mix containing 100 μl of Trichoderma atroviride wt and Δtgf-1 strains induced MFCF, 0.5 mM Suc-Ala-Ala-Pro-Phe-pNA (Sigma, Cat number S7388) in 50 mM MOPS, pH 7.0. The reaction mix was incubated at 37˚C for 20 min, and 500 μl of ice-cold water was added to stop the reaction, and measured spectrophotometrically at 405 nm. The activity was expressed as millimoles of p-nitroanilide released per minute [46]. Specific activity was referred to 1 mg of protein.

Chitinolytic activity
Chitinolytic activity was assayed by estimating the reducing ends of sugars as described elsewhere [47]. The reactions were performed in a 600μl mix containing 300 μl of 1% colloidal chitin, pH 6.6, and 300 μl of a 1/100 dilution of the wt or Δtgf-1 concentrated induced MFCF. The reaction mix was incubated 30 min at 37˚C, and the hydrolysis reactions were terminated and analyzed by adding 600 μl of dinitrosalicylic acid reagent (DNS). Then, the reaction mix was boiled for 15 min, and the insoluble chitin was removed by centrifugation, absorbance was measured spectrophotometrically at 540 nm. One unit of chitinase activity is defined as the amount of enzyme required to release 1 μmol of detectable reducing sugars in 1 min at 37˚C.

Data analysis
Experiments were statistically analyzed in the SPSS 10 program (SPSS, Chicago). Multivariate analyzes with a Tukey's post hoc test were used for testing differences in antimicrobial activity of MFCF, RT-qPCR and biochemical analysis (proteinase and chitinolytic activity). Linear regression analysis was done using the SPSS 10 program (SPSS, Chicago). Different letters are used to indicate means that differ significantly (P 0.05).

Results
Trichostatin A (TSA) did not affect the growth of T. atroviride over R. solani To (Fig 1B) or absence of TSA ( Fig 1A).

Addition of TSA to T. atroviride-growing medium enhanced its inhibitory effect on R. solani growth
As byproducts of their metabolism, most Trichoderma species are capable of synthetizing a wide range of antimicrobial compounds, used to combat phytopathogenic fungi. To determine whether TSA has an effect on the synthesis of antimicrobial compounds in T. atroviride, the fungus was grown for 7 days in PDB amended or not with 300 nM TSA. Thereafter, Trichoderma cultures were filtered to remove the mycelium. Plugs of R. solani were inoculated at the center of PDA plates amended with mycelium-free culture filtrates (MFCF) (final concentration of 60%) and the radial growth of R. solani was measured at 12, 24, 36, 48, and 60 h. As shown in Fig 2, addition of MFCF to PDA plates (crosshatched bars) inhibited R. solani radial growth compared to R. solani growth on PDA control plates with (arrow filled bars) or without TSA (black bars) (Fig 2). Furthermore, addition of MFCF plus 300 nM TSA to PDA plates (white dotted bars) enhanced the inhibition of R. solani radial growth compared to PDA-MFCF medium without TSA (crosshatched bars) (Fig 2).

TSA and R. solani presence negatively affected T. atroviride ech-42, prb-1, pbs-1, and tps-1 gene expression
To investigate whether the addition of TSA to the culture medium affected the expression of T. atroviride mycoparasitism-related genes ech-42 and prb-1, and those related to secondary metabolism, pbs-1 and tps-1, dual cultures of this fungus against R. solani were performed on PDA plates amended or not with 300 nM TSA. In the absence of TSA but in the presence of R.

Deletion of tgf-1 did not affect T. atroviride mycoparasitic effect against R. solani
The mutant growing phenotype of Δtgf-1, whose wild-type gene encodes for a histone acetyltransferase, was analyzed by inoculating T. atroviride wt and Δtgf-1 on PDA plates during 72 h under 12:12 light:dark regime (Fig 4A and 4B). The Δtgf-1 strain showed a slow growth phenotype and lack of the typical green colored conidia compared to the wt strain (Fig 4A and  4B). SEM photomicrographs of the Δtgf-1 strain showed a thin, flat and less branched mycelium compared to the wt strain (Fig 4C and 4D).
To determine whether the product of tgf-1 is involved in the antagonistic capacity of T. atroviride against phytopathogenic fungi, dual cultures of wt or Δtgf-1 strains against R. solani were conducted. The wt strain was able to growth over R. solani, whereas the Δtgf-1 strain was scarcely able to come in contact ( Fig 5). Moreover, in confrontation assays of Δtgf-1 against the phytopathogen, the growth of the latter was slightly delayed as compared when R. solani was co-cultured with the wt strain ( Fig 5). Linear regression analysis was used to find the relationship between the strains used in dual cultures and the growth inhibition of R. solani (S1 Table). Positive relationships were found between growth inhibition of R. solani by the wt, and growth inhibition of R. solani by the Δtgf-1 mutant. A strong significant positive (R 2 = 0.802, Sig. = 0.000) relationship was observed in growth inhibition of R. solani by the wt, followed by a low, but significant positive relationship in growth inhibition of R. solani by Δtgf-1 mutant (R 2 = 0.382, Sig. = 0.014). Since the Δtgf-1 mutant was barely able to enter in contact with the phytopathogen at the different times tested, the 192 h interaction zone of both dual cultures was taken and analyzed by SEM. Fig 6 shows that both, the Δtgf-1 and the wt, strains were able to coil around the phytopathogen hyphae (Fig 6C and 6D).
Dual cultures of Δtgf-1 against itself (Fig 7B), its parental strain (Fig 7C), or other Trichoderma species (Fig 7D-7I) showed that Δtgf-1 could be overgrown by other Trichoderma species compared to its parental strain. Also, the Δtgf-1 mutant was able to stop the growth of other species, showing, in some cases, an enhanced brown color in the contact area, which is indicative of lysed hyphae (Fig 7G and 7I). Confrontation of tgf-1 against itself did not show   3. TSA and R. solani presence negatively affected T. atroviride ech-42, prb-1, pbs-1, and tps-1 gene expression. T. atroviride and R. solani were co-cultured on PDA (A, B, C, and D) or PDA amended with 300 nM TSA (E, F, G, and H) at 28˚C. Total RNA was extracted from T. atroviride mycelium collected at 36, 48, and 60 h. Relative expression was calibrated using act-1 as housekeeping gene and normalized against the wt strain in the absence of the phytopathogen. ech-42 = 42-kDa endochitinase gene, prb-1 = basic proteinase gene, pbs-1 = peptaibol synthase gene, and tps-1 = terpene synthase gene. Black bars, T. atroviride wt strain growing alone (Ta); grey bars, T. atroviride vs R. solani (Ta+Rs) co-cultures. The bars show the mean ± SD of three independent biological replicates. Different letters are used to indicate means that differ significantly (P < 0.05). This assay was repeated twice per triplicate with similar results.    significant changes in any of the two colonies ( Fig 7B); however, when the Δtgf-1 strain was confronted against its parental strain, a lysing zone was observed (Fig 7C).

MFCF obtained from a Δtgf-1 strain shows enhanced inhibition of R. solani growth
Since the Δtgf-1 strain inhibited the growth of R. solani slightly more than the wt strain before they entered in contact (Fig 5 and S1 Table), we decided to assess the capacity of MFCF from Δtgf-1 to inhibit R. solani growth. PDA or PDA plus MFCF plates were inoculated in the center with mycelial plugs of R. solani and the radial growth of the phytopathogen was measured at 12, 24, 36, 48, and 60 h post-inoculation. MFCF from Δtgf-1 inhibited the growth of R. solani more effectively than that obtained from the wt strain (dotted black and black bars, respectively) (Fig 8). Intriguingly, addition of TSA to the Δtgf-1 growing medium diminished the negative effect of MFCF on the phytopathogen radial growth (arrow filled bars), compared to its respective control without TSA (dotted black bars), but MFCF from Δtgf-1 reached the inhibition effect observed with the wt MFCF plus TSA (cross-hatched bars) (Fig 8). Furthermore, the MFCF from the wt and Δtgf-1 grown in liquid Vogel's minimal medium showed similar results compared when they were grown in PDB medium (S1 Fig and Fig 8, respectively). Addition of TSA to the wt, grown in Vogel's minimal medium showed also similar results to those observed in PDB (Fig 8). Contrastingly, addition of TSA to the Vogel's minimal medium The Δtgf-1 and wt strains were grown for 7 days in PDB medium amended or not with 300 nM TSA, at 28˚C. MFCFs obtained from each of these cultures were added to PDA 1× medium at a final concentration of 60%. R. solani was inoculated into the different media and its radial growth was determined at 12, 24, 36, 48, and 60 h. Radial growth of R. solani on PDA containing T. atroviride wt strain MFCF without TSA (black bars) or amended with TSA (crosshatched bars) was determined at the indicated times. Radial growth of R. solani on PDA containing T. atroviride Δtgf-1 strain MFCF without TSA (black dotted bars) or with TSA (arrow filled bars) was determined at the indicated times. The bars show the mean ± SD of three independent biological replicates. Different letters are used to indicate means that differ significantly (P < 0.05). Three replicate plates were established for each treatment, and the experiment was repeated twice.

Δtgf-1 strain shows constitutive chitinolytic but inductive proteolytic activity
Lysing enzymes such as chitinases and proteases are used by some Trichoderma strains as biocontrol mechanisms against phytopathogenic fungi and oomycetes [23]. Aimed at elucidating whether chitinolytic and proteolytic activities were affected in the Δtgf-1 strain, and if they could be contributing to the enhanced inhibition of growth observed against R. solani, these activities were determined in cultures of the wt and Δtgf-1 strains induced by the presence of R. solani mycelium (Fig 9). Wild-type and Δtgf-1 MFCFs showed similar proteolytic activity under control conditions, whereas in the presence of R. solani, this activity increased at similar levels in both strains (Fig 9A). In contrast, the wt strain presented an increased chitinolytic activity in the presence of R. solani, compared to the mocked control, whereas the Δtgf-1 MFCF showed enhanced chitinolytic activity in control conditions, which was significantly high than that observed for the wt MFCF under the induced condition ( Fig 9B). Addition of R. solani to the growing medium did not increase the chitinolytic activity of Δtgf-1 as compared to its own control ( Fig 9B).

TGF-1 and TSA differentially regulate ech-42, prb-1, pbs-1, and tps-1 expression in the presence or absence of R. solani
To investigate whether TGF-1 plays a role in the regulation of mycoparasitism-(ech-42 and prb-1) and secondary metabolism-related genes (pbs-1 and tps-1), their expression levels were analyzed in a Δtgf-1 background and compared to its parental strain during their co-culture with R. solani, in media amended or not with TSA (300 nM). In the absence of TSA and R. solani, ech-42 showed significantly changes in its expression after 48 and 60 h (black bars) ( Fig  10A), whereas pbs-1, and tps-1 were considerably upregulated after 36 and 48 h, but downregulated at 60 h (black bars). prb-1 was upregulated after 48 h but downregulated at 60 h (Fig 10B,  10C and 10D). The presence of R. solani, increased~2-fold the expression of ech-42 after 36 h of co-culture, but it was downregulated after 48 and 60 h (gray bars) (Fig 10A). prb-1 expression suffered no significant changes in the presence of R. solani (Fig 10B). However, pbs-1, and tps-1 were downregulated in the presence of the phytopathogen after 36 and 48 h compared to their controls in absence of R. solani (gray bars) (Fig 10B, 10C and 10D). In the presence of TSA, ech-42 did not increase its expression after 36 h; however, it was marginally upregulated after 48 and 60 h (black bars) (Fig 10E). pbs-1 was downregulated in the presence of TSA after 60 h, whereas prb-1 and tps-1 were repressed 48 and 60 h after the addition of TSA, respectively (black bars) (Fig 10E, 10F and 10G). The presence of both, the phytopathogen and TSA, upregulated the expression of ech-42 and pbs-1 after 36 and 48 h (gray bars) (Fig 10E and 10G). prb-1 and tps-1 were marginally upregulated in the presence of TSA and R. solani after 48 h, and downregulated after 60 h (gray bars) (Fig 10G and 10H).

Discussion
Some species of the Trichoderma genus are used as biocontrol agents of phytopathogenic fungi and oomycetes that affect negatively important crops [22]. The molecular mechanisms governing mycoparasitism and secondary metabolism in Trichoderma have been intensely investigated. However, the role of histone acetylation in the biocontrol repertoire shown by these fungi is poorly understood. Chitinolytic activity was determined using colloidal chitin by estimating the reducing sugars with dinitro-salicylic acid (DNS) method, measuring absorbance spectrophotometrically at 540 nm. One unit of chitinase activity was defined as the amount of enzyme required to increase absorbance at 540 nm by 1 OD unit ml -1 h -1 . Proteinase activity was determined using mM Suc-Ala-Ala-Pro-Phe-pNA (0.5 mM) as substrate, the liberation of pNA was measured spectrophotometrically at 405 nm. The activity was expressed as p-nitroanilide released per min. Specific activity was referred to 1 mg of protein. The effect of induced MFCF of the wt and Δtgf-1 strains was tested over both activities. The bars show the mean ± SD of three independent biological replicates. Different letters are used to indicate means that differ significantly (P < 0.05). Chitinolytic and proteolytic activities were tested in three replicates and repeated two times.
https://doi.org/10.1371/journal.pone.0193872.g009 Fig 10. TGF-1 and TSA differentially regulate T. atroviride ech-42, prb-1, pbs-1 and tps-1 in the presence or absence of R. solani. Dual culture assays of T. atroviride wt and Δtgf-1 against R. solani, or grown alone as control were performed on PDA in absence (A, B, C, D) or in presence of 300 nM TSA (E, F, G, H) at 28˚C. Total RNA was extracted from T. atroviride mycelium collected at 36, 48, and 60 h. Relative expression was calibrated using act-1 as housekeeping gene and normalized against the wt strain in absence of the phytopathogen. ech-42 = 42-kDa endochitinase gen, prb-1 = basic proteinase gene, pbs-1 = peptaibol synthase gene, tps-1 = terpene synthase gene. Black bars represent Δtgf-1 strain, whereas gray bars represent Δtgf-1 vs R. solani interaction. Addition of TSA is indicated at the top of the panels. The bars show the mean ± SD of three independent biological replicates. Different letters are used to indicate means that differ significantly (P < 0.05). This assay was repeated twice per triplicate with similar results. https://doi.org/10.1371/journal.pone.0193872.g010

Trichostatin A (TSA) did not affect the overgrowth of T. atroviride on R. solani
In this work, we reported that TSA, a potent inhibitor of histone deacetylases class I and II, slightly affected the growth of T. atroviride and R. solani; however, the capability of T. atroviride to grow over the phytopathogen was not affected. Since TSA is per se an antifungal compound, it may have a negative effect on the growth of both fungi [48]. We also speculate that the inhibition of growth in both fungi could be due to the enhanced production of antimicrobial compounds promoted by the addition of TSA. In this regard, addition of TSA to the growing medium increases antimicrobial activity in a number of fungi against facultative bacterial and yeast pathogens [49]. Furthermore, T. virens strains that overproduce the antimicrobial compound gliovirin grow slower than the wt strain [49,50]. Contrastingly, mutants lacking genes whose products are involved in the synthesis of antimicrobial compounds, such as the potent antimicrobial and cytotoxic compound gliotoxin grow faster than the wt strain [38,51,52].

Addition of TSA to T. atroviride-growing medium enhanced its inhibitory effect on R. solani growth
Fungi are excellent producers of biological active compounds, as a consequence of their secondary metabolism, including antibiotics, plant growth-regulating molecules, cell toxic compounds, mutagenic, immunosuppressants, enzyme inhibitors, and molecules with other biological effects [25,53,54]. Here, we show that the addition of TSA enhanced the ability of T. atroviride MFCF to inhibit R. solani growth. These data suggest that inhibition of HDACs class I and II, such as Rpd3p and Hda1p, by TSA may be affecting the production of antimicrobial compounds in T. atroviride, probably by an increase in the acetylation of histone tails. This increased acetylation could de-repress silenced chromatin regions that contain genes involved in secondary metabolism and/or in the synthesis of antimicrobial compounds in this fungus. A genomic analysis of three Trichoderma species, including T. atroviride, showed that all of them contain the orthologous genes to RPD3 and HDA1 of S. cerevisiae [12]. It was recently demonstrated that the addition of HDAC inhibitors, including TSA, induces the antimicrobial activity of extracts on several fungi [49]. Moreover, A. nidulans mutants lacking components involved in chromatin compaction, such as HepA (heterochromatin protein 1) show derepression of genes involved in the biosynthesis of secondary metabolites. In agreement with these results, the silent sterigmatocystin gene cluster shows low levels of repressive histone marks (trimethylated H3K9), whereas histone H3 acetylation is increased [55].

TSA and R. solani presence negatively affected T. atroviride ech-42, prb-1, pbs-1, and tps-1 gene expression
Expression analysis of mycoparasitism and secondary metabolism-related genes showed that R. solani induced the expression of ech-42 and prb-1, as previously reported [28,56]. In addition, the presence of R. solani also induced the transcription of tps-1 and the repression of pbs-1. A plausible explanation for the repression of pbs-1 in the presence of R. solani is that the phytopathogen is trying to block the production of the peptaibols produced by T. atroviride through the synthesis of effector molecules. However, this is not the case for tps-1, which is involved in the synthesis of volatile terpenes. This fact, points to different regulation pathways for pbs-1 and ech-42, prb-1 and tps-1 in the presence of the host. In this regard, it is well-known the strong battle that takes place between parasites and their hosts to counteract each other by means of lysing enzymes as well as effector molecules [57]. For instance, fungal effectors that promote pathogen virulence by suppressing the chitin-triggered immunity in plants, including LysM (effectors that carry no recognizable protein domains other than lysin motifs) have been identified also in saprophytic and beneficial fungi [58]. In this case, probably the phytopathogen is attempting to evade the mycophagous behavior of T. atroviride by producing effectors that, in the presence of TSA, enhance their production in the phytopathogen to silence genes such as tps-1.
Intriguingly, pbs-1 and ech-42, prb-1 and tps-1 were positively affected by the presence of TSA. These results indicate that TSA may be affecting the transcription of these genes, by inhibiting Rpd3-and Hda1-encoding genes through histone acetylation in silenced chromatin regions as reported for other organisms [59,60]. In this sense, it is well known that TSA induces histone acetylation, which is often associated with an increased expression of a number of genes [61][62][63]. Frequently, these changes impact the biology of the organism, such as virulence and pathogenicity [64][65][66]. Moreover, the presence of both TSA and the phytopathogen downregulated the expression of all four genes, suggesting a positive effect of TSA in the synthesis of effector molecules in R. solani to counteract the mycoparasitic response of T. atroviride. Another plausible explanation is that both, the presence of TSA and the phytopathogen, but none of them alone enhances the expression of a negative regulator of mycoparasitism and secondary metabolism-related genes in this fungus.

Deletion of tgf-1 did not affect T. atroviride mycoparasitism effect against R. solani
Dual cultures of Δtgf-1 and R. solani showed that the mutant coiled around the phytopathogen hyphae; however, Δtgf-1 was deficient to grow over the pathogen. This result suggests that TGF-1 is not involved in coiling, but the reduced growth rate observed in the Δtgf-1 strain could be affecting its capacity to grow over R. solani. Probably, the Trichoderma growth rate is an important trait for competence. In this regard, it has been determined that growth inhibition of fungi by Trichoderma is directly proportional to the growth rate of the antagonists under a given condition [67].

MFCF obtained from a Δtgf-1 strain shows enhanced inhibition of R. solani growth
Antimicrobial analysis of MFCF obtained from a Δtgf-1 strain, as well as dual cultures against R. solani, showed an enhanced growth inhibition of the phytopathogen compared to MFCF obtained from the wt strain. These results point out to a negative role of TGF-1 in secondary metabolism in T. atroviride. In this regard, TGF-1, orthologous in other organisms, also regulates negatively gene expression. For instance, in N. crassa and in T. atroviride, the orthologous to Gcn5 positively regulates gene photoinduction, but negatively the expression of the al-3 gene in darkness [5] (Uresti-Rivera et al., in preparation). In S. cerevisiae, stress response adaptation requires Gcn5 activity for the activation or repression of genes that are physically associated with this protein. This fact provides support for a role of Gcn5 and its orthologous proteins as co-activators as well as co-repressors [6].
Addition of TSA to the Δtgf-1 growing medium (PDB) reverted the effect of the lack of TGF-1 in the antibiosis assays, which suggests that histone acetylation induced by TSA is enough to totally revert the effect of tgf-1 deletion, and that probably its product regulates negative elements. In this regard, it has been demonstrated that addition of TSA rescues the phenotypes of HAT mutants in yeast and HeLa cells [68][69][70]. Intriguingly, addition of TSA to the Vogel's minimal medium, did not revert the effect of Δtgf-1 MFCF on R. solani to wild-type phenotype, on the contrary, it was enhanced. This result indicates that the components of the PDB medium could be affecting the acetylation pattern on the promoters of secondary metabolism-related genes in a Δtgf-1 background, but not in the wt.

The Δtgf-1 strain shows constitutive chitinolytic but inductive proteolytic activity
Our results indicate that TGF-1 is a negative regulator of chitin degrading proteins, but not of those implicated in protein degradation, since chitinase activity showed an increased basal activity in the Δtgf-1 as compared to the wt. These results together with those of growth inhibition of R. solani may explain in part the augmented growth inhibition exerted by Δtgf-1 MFCF on the phytopathogen compared to wt MFCF, since some Trichoderma chitinases are secreted to the medium [71]. Trichoderma strains overexpressing either ech-42 or prb-1 provide more protection to plants against root and foliar pathogens [29][30][31]. In addition, mutants in the mitogen-activated protein kinase encoding gene, tvk1, from T. virens, showed increased transcription levels of mycoparasitism-related genes, high production of lytic enzymes and were considerably more effective in disease control than the wt strain [72].

TGF-1 and TSA differentially regulate ech-42, prb-1, pbs1, and tps-1 expression in the presence or absence of R. solani
Expression analysis of mycoparasitism-and secondary metabolism-related genes in the Δtgf-1 strain showed that probably the product of tgf-1 negatively regulates prb-1, tps-1, ech-42 and pbs-1 in absence of the phytopathogen, since all tested genes were expressed in almost all tested times. Based on this, we can propose that TGF-1 could be a direct repressor through its putative histone acetyltransferase activity. Or it could be also an indirect repressor of these genes through acetylation of the promoter of a negative regulator-encoding gene, whose transcript is absent by the loss of TGF-1. In this regard, a negative role of Gcn5 has been observed in yeast and filamentous fungi in a number of processes [5,6,72] (Uresti-Rivera et al., in preparation).
Dual cultures of Δtgf-1 against R. solani showed that the expression of all tested genes in the mutant strain was downregulated at almost all tested times. A plausible explanation for these results is that TGF-1 is necessary to maintain the homeostasis with the effector molecules produced by R. solani, which in the mutant background favors the suppression of Trichoderma genes by the phytopathogen. Together, these results indicate that TGF-1 is necessary for the induction of all four genes in the presence of the phytopathogen or to counteract the negative effect of R. solani on T. atroviride. The exceptions in gene repression by R. solani at all tested times were ech-42 and pbs-1. Probably, the induction of such genes under this condition may be due to the activity of other HATs such as Sas3 (ID Ta_223094; [12]), which has been shown to present overlapped roles with Gcn5 in S. cerevisiae [73]. Addition of TSA to the culture medium induced the expression of all tested genes in the Δtgf-1 strain, but at different times compared with its control. Probably this is a result of the promotion of histone acetylation, but it could be also a compensatory effect of other HAT in absence of TGF-1. The addition of TSA in the presence of R. solani induced the transcription of ech-42 and pbs-1, which were repressed at 60 h of co-culture. Nevertheless, the expression of prb-1 and tps-1 genes was downregulated. A possible explanation for these results is that the sole presence of R. solani is not enough to fully repress the four genes at early times as it happened during confrontation with the wt strain in the presence of TSA and that such suppression was delayed by the absence of TGF-1.
The fact that other Trichoderma strains can grow over the Δtgf-1 mutant, and that some of them, including its parental strain, showed a marked lysing zone, could be explained as due to the slow growth of the Δtgf-1 strain. It is feasible that Δtgf-1 cannot compete with the other Trichoderma strains, although it showed an enhanced production of antibiotics that led to the lysis of other fungi hyphae [74], including its parental strain.
A detailed study focused on the interaction of HATs/HDACs with chromatin and acetylation balance, by ChIP on chip and ChIP-Seq assays, will be required to fully understand the role of chromatin modifications that regulate the antagonistic activity of T. atroviride and to determine if TGF-1, in these processes, acts as a co-activator as well as a co-repressor.
Supporting information S1 Fig. Antibiosis assay using MFCF obtained from T. atroviride wt or Δtgf-1 strains grown in Vogel's minimal medium in the presence or absence of TSA against R. solani. The Δtgf-1 and wt strains were grown for 7 days in Vogel's minimal medium amended or not with 300 nM TSA, at 28˚C. MFCFs obtained from each of these cultures were added to PDA 1× medium at a final concentration of 60%. R. solani was inoculated into the different media and its radial growth was determined at 12, 24, 36, and 48 h. Radial growth of R. solani on PDA containing T. atroviride wt strain MFCF without TSA (black bars) or amended with TSA (crosshatched bars) was determined at the indicated times. Radial growth of R. solani on PDA containing T. atroviride Δtgf-1 strain MFCF without TSA (black dotted bars) or with TSA (arrow filled bars) was determined at the indicated times. The bars show the mean ± SD of three independent biological replicates. Different letters are used to indicate means that differ significantly (P < 0.05). Eight replicate plates were established for each treatment, and the experiment was repeated twice. (TIF) S1 Table. Relation between growth inhibition of R. solani and the T. atroviride strain used in dual cultures assays. (DOCX)