Pathogen

The Fusarium graminearum F224a used in this study belongs to the Microbial Culture Collection of the Área Microbiología, Facultad de Química, at Universidad de la República (UDELAR), Montevideo, Uruguay. It was selected based on the elevated levels of deoxynivalenol (DON) produced in culture, its capacity of generating perithecia in wheat straw in controlled (humidity and temperature) conditions, and the high aggressiveness level when inoculated in wheat flower anthers in greenhouse assays (^{Umpiérrez et al., 2013}).

Antagonist isolation

Trichoderma spp. strains utilized in this study were isolated from 20 wheat straw samples from the main production zone in our country located eastward from Rio Uruguay in the departments of Río Negro, Soriano, and Colonia. For this purpose, 10 g of wheat straw were placed in 100 mL of sterile physiological serum and homogenized in Stomacher 400^® (Stomacher, UK) for two min at maximum speed. Dilutions of this suspension were sown on surface Potato Dextrose Agar (PDA, Oxoid) Petri dishes containing neomycin (50 mg mL^-1). Typical colonies of Trichoderma spp. were cultured in PDA Petri dishes; subsequently, monosporic cultures were performed, and the isolates obtained were preserved in PDA test tubes at 5 °C until use.

Antagonist identification

Trichoderma strains were identified by studying the sequences of two genomic regions, ITS1-ITS2 and one part of the gene codifying for the translational elongation factor 1 alpha (tef1α) comprising the totality of intron 4 and part of the exons 4 and 5 of the gene (^{Jaklitsch and Voglmayr, 2015}). DNA extraction was carried out according to ^{Garmendia and Vero (2016)}. The amplification reactions were performed in a total volume of 25 μL utilizing 0.2 μL of each primer 25 μM, 0.15 U of Taq polymerase (Invitrogen, Thermo Fisher Scientific, Waltham MA, USA), 0.75 μL of MgCl₂ 50 mM, 1 µL of each dNTP 5 mM solution (Invitrogen, Thermo Fisher Scientific, Waltham MA, USA), 2.5 µL of amplification buffer (Invitrogen, Thermo Fisher Scientific, Waltham MA, USA), and approximately 20 ng of DNA, completing the reaction volume with sterile distilled water. In the case of the ITS1-ITS2 region amplification, the primers used were ITS1 (5´-TCCCGGTTCGCTCGCCGTTACTA-3´) and ITS4 (5´-TCCTCCCGCTTATTGATATGC-3´) (^{White et al., 1990}). The reaction mixture was subjected to 5 min denaturation at 95 °C, followed by 35 cycles of one-min denaturation at 95 °C; 30 sec alignment at 55 °C; one min extension at 72 °C, ending with a 10 min extension at 72 °C.

In the case of the tef1α gene, the primers used were EF1-728F (5´-CATCGAGAAGTTCGAGAAGG-3´) (^{Carbone and Kohn, 1999}) and Tef1 rev (5´-GCCATCCTTGGAGATACCAGC-3´) (^{Samuels et al., 2006}). The PCR cycle was similar to the previous one except for the annealing temperature, which was 51 °C. The primers used in both cases were synthetized by Byo Synthesis Company (Lewisville, TX, USA). The PCR products were analyzed in agarose gel at 0.8% containing ethidium bromide (5 μg mL^-1). A band of around 600 base pairs was obtained. The amplification products of both regions were purified and sequenced by Macrogen Inc. (Seoul, Korea). The sequences of both regions were compared with those deposited in Genbank (https://www.ncbi.nlm.nih.gov/genbank/) database using the BLAST tool.

Phylogenetic analyses

The phylogenetic analyses of the concatenated sequences (ITS1-ITS2 and tef1α gene) were aligned with the concatenated sequences of type strains obtained by GenBankusing MEGA program version 6 (^{Tamura et al., 2013}). The phylogenetic trees were constructed by the Neighbor-Joining (^{Saitou and Nei, 1987}) method, and the evolutionary distances were computed by the ^{Jukes-Cantor (Jukes and Cantor, 1969}) model. All the positions containing gaps or missing data were eliminated with the option of paired-sequence comparison (Pairwise deletion option). Group stability was determined using a bootstrap of 1000 replicates (^{Felsenstein, 1985}).

Characterization of Trichoderma strains

Sensitivity to tebuconazole. The minimum inhibitory concentration (MIC) of the fungicide - defined as the lowest concentration that prevented visible growth of each Trichoderma strain - was determined. For this purpose, the fungicide Folicur 450⁽ (Bayer Crop Science, Monheim am Rhein, DE) was diluted in sterile water and incorporated into PDA to achieve concentrations of 0, 2, 4, 8, 16, 32, and 64 mg L^-1 of tebuconazole (^{Umpiérrez et al., 2013}). In timely manner, 5 μL of conidial suspension adjusted to a concentration of 1 × 10⁵ conidium mL^-1 of each Trichoderma strain were inoculated on the surface of fungicide amended PDA plates Eight strains were inoculated per dish, and two replicates per treatment were performed. The Petri dishes were incubated at 25 °C for five days. Fungal growth was evaluated visually.

Determination of xylanase production capacity. For each Trichoderma spp. strain, the capacity of producing xylanase was determined in minimal medium containing xylan as carbon source. Cultures were performed in liquid Yeast Nitrogen Base (YNB) (BioChemica, Sigma Chemical, St. Louis MO, USA) medium amended with Xylan (Sigma Chemical, St. Louis, MO, USA) at 0.5% as the only carbon source. Each 40 mL Erlenmeyer flasks with culture medium were inoculated with 100 μL of conidial suspension 2 ×10⁴ conidia mL^-1) of one of the strains and incubated at 25 °C.

Enzymatic activity was determined by mixing 100 μL of citrate-phosphate buffer (pH 5.6), 50 μL of culture supernatant (centrifuged and filtered through a 0.45 μm pore filter), and 50 μL of xylan at 1% in distilled water. The mixture was incubated in water bath at 40 °C for one hour; subsequently, 300 μL of dinitrosalicylic acid (DNS) were added and heated in water bath to boil for five min. Finally, 500 μL of distilled water were added, and absorbance was measured at 540 nm against each blank, which was prepared similarly to the reaction mixture but adding the DNS and heating immediately after the mixture was prepared. Each reaction was performed in triplicate. One unit of enzyme was defined as the amount released by 1 μmol of reducing sugar (expressed in glucose equivalent) per minute in the reaction conditions.

Production of chitinases associated to fungal cell wall degradation. The capacity of producing chitinases by Trichoderma spp. strains was evaluated in YNB amended with fungal cell walls of F. graminearum in a concentration of 0.5% (m/v) as the only carbon source (^{Vero et al., 2013}). The cultures were performed individually by inoculating 100 μL of a suspension of 2 × 10⁴ conidia mL^-1 of each Trichoderma spp. strain in Erlenmeyer 200 mL flasks containing 40 mL of the medium. Incubation was performed at 25 °C without agitation for 10 days. Each culture was performed in triplicate. Once incubation ended, each culture was centrifuged, and the supernatant was filtered through a 0.45-μm pore cellulose acetate filter.

Chitinolytic activity from the culture filtrates was measured according to the method of ^{Mahadevan and Crawford (1997)} with some modifications. Briefly, 90μL of culture filtrate and 10 μL of a solution 0.18 mM of p-nitrophenyl-N-acetyl-β-D-glucosaminide (Sigma, St. Louis, MO, USA) dissolved in sodium phosphate buffer 50 mM, pH 6.1 were placed in wells of a flat-bottom microtitre plate. The plate was incubated at 25 °C for 24 h without agitation, and the reaction ended by the addition of 10 μL of NaOH 1 M. Absorbance was read at 405 nm in an automatic Microplate Autoreader (Bio-tek Instruments, Winooski, VT, USA) against a blank of each sample, in which NaOH solution was added previous to incubation. The para-nitrophenol (pNP) concentration obtained in each case was determined using a standard curve, one unit of enzyme was defined as the amount needed to release 1 μmole of pNP in one min. The assay was performed in triplicate per each strain.

Production of Volatile Antifungal Compounds. The production of volatile antifungal compounds was analyzed according to ^{Arrarte et al. (2017)} with modifications. Petri dishes containing PDA were inoculated centrally with a 5 mm mycelial agar plug from a non-sporulating, two-day-old PDA culture of each Trichoderma strain. At the same time, other plates containing PDA were inoculated in the same manner with a mycelial agar plug of the pathogen.

The lids of the dishes were removed, and the bases containing the pathogen were placed over those with the antagonist, in such a way that the cultures would be confronted. Each set of confronted dishes was sealed with Parafilm and incubated at 25 °C. As control, pairs with F. graminearum in both Petri dishes were used. The F. graminearum colony diameter was measured at day 5, at the moment in which the pathogen covered all the control dish. Inhibition was expressed as the reduction percentage of the colony diameter compared with the control. The assay was performed in triplicate.

Production of soluble antifungal compounds in liquid medium. This assay followed the methodology of ^{Marques et al. (2018)} with some modifications. The different Trichoderma strains were inoculated in flasks containing 50 mL of YES medium (yeast extract, 2%; sacarose, 15%, pH = 6) adding 1 mL of 1 × 10⁶ conidia mL^-1 suspension. The cultures were cultivated at 25 °C, without agitation for seven days. At the end of this interval, the culture supernatant was filtered with 0.45-μm pore filters to obtain a cell-free solution, which was incorporated to an equal volume of PDA at double concentration and poured into Petri dishes previously adjusted to pH = 6. As control, PDA plates were prepared in the same manner but adding non-inoculated YES medium. On these dishes, a 5 mm mycelial agar plug of the pathogen was placed and incubated at 25 °C. At the end of day 4 of incubation, F. graminearum colony diameters were measured and compared with those in control plates. Inhibition was expressed as colony growth reduction percentage in the Petri dishes containing different filtrates compared with the average growth in control plates. The assay was performed in triplicate.

Dual pathogen-antagonist cultures

The capacity of the different Trichoderma spp. strains of inhibiting Fusarium graminearum F224a growth was determined by means of dual cultures in PDA Petri dishes at 25 °C. Disks of 5-mm from a three-day-old culture of antagonists and pathogen were cut from the border of each colony with a hole.puncher. Both pathogen and antagonist disks in Petri dishes with PDA were confronted, each one at 2.5 cm from the dish center. As control, a pathogen disk was sown in a Petri dish with PDA. The Petri dishes were incubated at 25 °C, and the diameter of each colony was measured after five days, moment in which the pathogen colony had covered the control dish. The antagonistic capacity of each species was established according to the following scale: (1) the F. graminearum colony radius was greater than 2.5 cm; (2) the F. graminearum radius was less than 2.5 cm; (3) the Trichoderma spp. strain grew over all the Petri dish covering the F. graminearum colony. The assay was performed in triplicate for each case.

Production inhibition of the pathogen perithecia in the pathogen-antagonist culture on wheat straw

Five Trichoderma spp. strains of the Viride clade were selected to perform this assay. The selection was based on the results obtained in the dual culture study, in such a way to have representatives of all the defined categories in such assay. This study assessed the inhibition in perithecium production by the pathogen on wheat straw caused by Trichoderma spp. strains (^{Bujold et al., 2001}). Ten pieces of wheat straw of approximately 7 cm in length - each one with a knot - were previously inoculated by 1-min immersion in 50 mL of a suspension of conidia from the pathogen and one antagonist (1 × 10⁴ conidia mL^-1). The conidial suspensions were prepared from the cultures in the leaned PDA tubes using physiological serum with Tween 80 at 0.1% to extract conidia. The inoculated wheat straw was placed on Petri dishes containing humid sand previously sterilized. Each pathogen-antagonist treatment was performed in triplicate. Both wheat straw and sand used in this assay were sterilized by two autoclave sessions at 121 °C for 30 min in two consecutive days. As control, wheat straw inoculated with a suspension containing only pathogen conidia was used. The Petri dishes containing treated wheat straw were incubated at a temperature of 25 °C with light-darkness cycles of 12 h to promote perithecia development. Sand humidity of each Petri dish was controlled each five days, and sterile distilled water was added when necessary to maintain initial weight of each Petri dish. At 40 days, all perithecia in the upper face of the wheat straw of each Petri dish were counted under a magnifying glass. The results were expressed as inhibition percentage with respect to the average of the perithecium amount found in the control Petri dishes.

Correlation between antagonism characteristics

The correlation between chitinase and antifungal compounds (volatile and soluble) production, pathogen growth in dual cultures in Petri dishes and perithecia development on wheat straw, was determined by means of Spearman correlation coefficient, using the statistical program Infostat⁽ (^{Di Rienzo, 2015}).

Statistical analyses

The results were analyzed with one-way Analysis of Variance (ANOVA) using Infostat⁽ (^{Di Rienzo, 2015}). Previous to the analysis, the percentage values were divided by 100 and transformed to arcsine square root. The comparison between averages was performed with Fisher’s least significant difference (LSD) (P < 0.05).

Antagonist isolation and identification

A total of 16 isolates of Trichoderma spp. from 20 samples of wheat straw from the different cultivation areas were identified at species level by the corresponding sequence analysis of the ITS1-ITS2 region and part of the tef1α gene. By comparing with the corresponding sequences of the type strains deposited in the GenBank and by phylogenetic analysis, the strains were determined to belong to five different species. The identification of each one of the strains is shown in Table 1. Nine strains were identified as T. koningiopsis (57%), three as T. atroviride (19%), two as T. longibrachiatum (12%), one as T. afroharzianum and another one as T. rifaii.

The phylogenetic tree obtained in such analysis included type strain sequences that showed similarity (98-100%) with both sequences of the strains in study (Figure 1). The presence of three great groups with bootstrap values of 100% could be established. One of them corresponded to Viride clade where 12 of the isolated strains were included, of which nine were consistently grouped with T. koningiopsis GJS 93-20 (100% bootstrap) type strain and three of them with T. atroviride CBS 142.95 (100% bootstrap) which was also type strain.

Sequences from type strains retrieved from GenBank were included. The tree was constructed by Neighbour-Joining and evolutive distances were calculated by Jukes-Cantor method. Bootstrap values were calculated from 1000 replications. Phylogenetic analysis was carried out with MEGA6.

Figure 1: Phylogenetic tree based on concatenated ITS1-ITS2 y tef1α gene sequences of uruguayan Trichoderma spp. strains used in this work.  

Another group, in which the sequences corresponding to two strains (T35 and T4) were located, corresponded to the Harzianum clade. One of the sequences was grouped with T. afroharzianum GJS 04-186 (type strain) and the other one with T. rifaii DIS 355B (type strain), clearly separated from the type strains of the other species in the same clade. The group formed by the T3 and T2 strains corresponded to the Longibrachiatum clade. The Uruguayan strains grouped with T. longibrachiatum ATCC 18648 type strain and were separated from T. bissettii UTHSC 08‑2443 type strain, which was located in the same clade.

Most of the strains were identified as T. koningiopsis within the Viride clade. The presence of this species has already been described in warm regions of South America (Brazil and Peru) and in colder regions of North America in Canada and eastern United States (^{Samuels, 2006}). Additionally, within the same clade, other three strains were identified as T. atroviride - species that has also been reported as cosmopolitan (^{du Plessis et al., 2018}). Two isolates were identified as T. longibrachiatum, a species of wide geographic distribution which has been reported as a causal agent of localized and systemic infections in immunodepressed or dialyzed patients (^{Druzhinina et al., 2008}). These characteristics preclude the use of this species as a biocontrol agent, especially in extensive crops, and consequently, the strains of these species were not considered in further studies. The remaining isolates were identified as T. afroharzianum and T. rifaii, which belong to the species complex T. harzianum within the Harzianum clade (^{Chaverri et al., 2015}). This complex gathers at least 14 species, including T. harzianum sensu stricto which gives its name to the complex. A recent work about species reclassification within such complex has determined that T. harzianum could only be found in the northern hemisphere (^{Chaverri et al., 2015}). On the other hand, other species of the clade have different geographic distribution patterns. For example, T. rifaii, has only been found in South America, while T. afroharzianum has a worldwide distribution. Thus, finding these two species in Uruguay was not strange. To our knowledge, this is the first report of T. afroharzianum and T. rifaii in Uruguay.

Characterization of Trichoderma spp. strains

Sensitivity to Tebuconazol. The sensitivity of Trichoderma spp. native strains to this fungicide was studied to determine the possibility of a joint application in an integrated disease management. The MIC values obtained for each strain were concordant between replicates. The strains of the Harzianum and Longibrachiatum clades, which represented 25% of the total, showed a MIC of 16 mg L^-1 while those of the Viride clade were inhibited with 32 mg L^-1 of the fungicide. The sensitivity to tebuconazole of Trichoderma spp. strains was very similar to that found for F. graminearum strains in a previous study (^{Umpiérrez et al., 2013}), In turn, the Trichoderma strains isolated in this study resulted more sensitive to the fungicide than those of European origin described by ^{Hatvani et al. (2006)}, which showed MIC values of 100 mg L^-1 for tebuconazole. According to the results, a joint application of the fungicide with the biocontrol agent would not be effective since it could compromise its own viability.

Xylanase production. The capacity of Trichoderma spp. strains to produce xylanases was determined. Xylanases are enzymes associated to plant tissue degradation, which would facilitate the degradation of wheat straw that acts as support for the pathogen. All the strains were capable of producing xylanases in the assayed conditions (Figure 2). Significant differences (P ≤ 0.05) were found between strains of the Viride clade species, while the strains identified as T. longibrachiatum showed a very similar and relatively low activity. At the same time, the strain with less activity was TR4 identified as T. rifaii, which belongs to the Harzianum clade.

Error bars represntatandard deviation of 3 repetitions. Distinct letters in the same column indicate significant differences (LSD P 0.05).

Figure 2: Xylanases (UE mL^-1) produced by Trichoderma spp. strains in presence of xylan a carbon source. 

Chitinase production. The capacity of all Trichoderma strains to produce chitinases and glucanases, enzymes associated to fungal cell wall polymers degradation was evaluated in study. The extracellular production of such enzymes has been related to the capacity of mycoparasitism, mechanism associated to Trichoderma spp. with demonstrated biocontrol activity against fungal plant diseases (^{Matarese et al., 2012}). The chitinolytic capacity of the different Trichoderma spp. strains are shown in Figure 3 and represents the average of three independent determinations performed at 25 °C. In such conditions, all the strains produced extracellular chitinases. However, when the activity assay was performed at 40 °C for 30 min (protocol developed for xylanases), the activity values were very low in average and not detectable in some cases.

Error bars represent standard deviation of 3 repetitions. Distinct letters in the same column indicate significant differences (LSD P ≤ 0.05).

Figure 3: Chitinolytic activity (UE mL^-1) produced of different Trichoderma spp. strains.  

The strains belonging to the Viride clade showed different levels of activity with significant differences (P ≤ 0.05) in activity for strains of the same species, such as in the case of those of T. koningiopsis species. The strains of the Longibrachitum clade developed very similar activities with no significant differences (P ≤ 0.05) between them or with those that belonged to T. atroviride species of the Viride clade. The chitinolytic activity within the Harzianum clade was variable. The activity associated to the strain identified as T. afroharzianum was significantly (P ≤ 0.05) greater than that of the T. rifaii strain.

The results evidenced that all the strains in study can produce enzymes capable of degrading chitin, a fundamental activity to act as mycoparasite.

Pathogen growth inhibition due to volatile and soluble compounds in the culture medium

Most of the Trichoderma spp. strains in the assay were capable of producing volatile and soluble metabolites with antifungal activity against the pathogen of study. The greatest pathogen inhibition percentages (around 40%) caused by volatile compounds were achieved with T1 and T11 strains identified as T. koningiopsis, which resulted significantly different (P ≤ 0.05) to the rest (Figure 4). Nevertheless, the inhibition achieved by other strains of the same species (TR8, TR7, and T251) was lower, which shows that the production of volatile antifungal metabolites is a strain dependent characteristic. Several volatile compounds have been characterized as responsible for the antifungal activity of the Trichoderma strains. Among them, simple aromatic compounds, pyrones, volatile terpenes, and isocyanates have been reported (^{Stoppacher et al., 2010}). The 6-pentyl-2H-pyran-2-ona (one pyrone) - a volatile compound associated to coconut aroma, characteristic of T. atroviride (^{Garnica-Vergara et al., 2016}) - has antifungal activity. Its capacity of inhibiting in vitro growth of different fungal pathogens, such as Fusarium oxysporum f. sp. lycopersici and Rhizoctonia solani, has been demonstrated (^{Reino et al., 2008}). According to these authors, the volatile compounds act at a distance empowering the localized action of the enzymes and the non-volatile antifungal compounds. The strains that inhibited the pathogen at a greater degree by producing soluble antifungal compounds were those that belonged to T. atroviride (TmE, TmB, and To) and T35 (T. afroharzianum). They produced more than 60% of inhibition, showing significant differences (P ≤ 0.05) with the rest of the strains. On the other hand, T1 (T. koningiopsis), T2, T3 (T. longibrachiatum) practically did not inhibit pathogen growth in this assay, showing values lower than 15%. The strains of the genus Trichoderma produced a variety of secondary metabolites with antibacterial, antifungal, and antiviral activities. Among those that stand out are some non-ribosomal peptide synthesis formed from five to 20 amino acids, some of which do not form part of proteins, such as alpha aminoisobutyric acid and isovaline (^{Marik et al., 2018}). These compounds receive the name of peptaibols and act at membrane level. Some of them are characteristic of determined species (^{Oh et al., 2005}). For example, koninginins are produced by T. koninigiopsis (^{McMullin et al., 2017}) while T. atroviride produces atroviridis and neoatroviridins (^{Komon-Zelazowska et al., 2007}). Since these compounds are secondary metabolites, the production levels depend on the medium where the producing fungus develops, so the real role of these compounds in a specific pathosystem should be verified in situ. This study considered the production of soluble antimicrobial substances in YES medium, which was recommended by ^{Paterson and Bridge (1994)} to promote secondary metabolite production. Therefore, the results obtained in this study should be considered only as indicative of the antifungal activity of each strain. It would be important to corroborate its relationship with the antagonistic activity on wheat straw.

Error bars represent standard deviation of 3 repetitions. Distinct letters in the same column indicate significant differences (LSD P ≤ 0.05).

Figure 4: Inhibition of F. graminearum growth by volatile and soluble antifungic compounds produced by Trichoderma spp. strains.  

Antagonist-pathogen dual cultures

Pathogen growth inhibition was evaluated confronting the different Trichoderma spp. strains in dual culture in PDA Petri dishes at 25 °C. The pathogen colony radius in the confrontation zone with the antagonist was determined on day 5 of culture, moment at which the F. graminearum mycelium covered all the surface of the medium in the control Petri dishes. TmE, TmB, and To (T. atroviride) and TR8 (T. koningiopsis) produced the total inhibition of the pathogen, growing and sporulating over its mycelium. T2, T3, T4, TR4, T12, TR9, and T251 inhibited slightly the pathogen growth in the assayed conditions, and the rest of the strains showed an intermediate behavior. Figure 5 shows the radius of the pathogen colony in the confrontation zone of dual cultures.

Error bars represent standard deviation of 3 repetitions. Distinct letters in the same column indicate significant differences (LSD P ≤ 0.05).

Figure 5: Radius of F. graminearum colony facing Trichoderma spp. strains in dual cultures.  

Figure 6 shows examples of dual cultures. In (a) the antagonist (T. atroviride TmE) grew over the pathogen, and in (b) the pathogen grew beyond the confrontation zone with T. koningiopsis 251a, which is a weak antagonist in the assayed conditions.

Figure 6: Inhibition of F. graminearum in dual cultures against Trichoderma spp.; a) TmE - F. graminearum; b) T251a - F. graminearum y c) control. 

Pathogen inhibition in dual cultures is one of the most utilized assays to study the biocontrol capacity of a microorganism. The extended use of this method is due to the simplicity of the assay. Nevertheless, it should be considered that it only provides indicative - non-definitive - results. In the first place, growth velocity of the two confronted microorganisms in the culture medium depended on the culture conditions, which are not always similar to those where the interaction would be expected to occur. In second place, the capacity of the biocontrol agent to inhibit the pathogen mostly depends on the inducible compound production or secondary metabolism products, so inhibition in the culture medium would not necessarily correspond to inhibition in the field (^{Mondino and Vero, 2006}).

Inhibition of perithecia production on wheat straw

The production inhibition of perithecia by the pathogen on wheat straw was the definitive assay of this study since it reproduced - in the most faithful manner - what was expected to occur in field. The previous assays were only indicative of the potential biocontrol of the antagonists. All the Trichoderma spp. strains could significantly inhibit the production of perithecia on wheat straw (Figure 7).

Error bars represent standard deviation of 3 repetitions. Distinct letters in the same column indicate significant differences (LSD P ≤ 0.05).

Figure 7: Perithecia produced by the pathogen on wheat straw y absence (a) and presence (b) of Trichoderma spp. from Viride clade. 

An effective colonization of wheat straw by the different Trichoderma strains could also be observed (Figure 8). Of the strains selected, TmE and T11 produced higher levels of volatile antifungal compounds, which could contribute to control at distance. TmE, in turn, showed a significantly greater production of antifungal soluble compounds and also an elevated production of chitinases. The synergic action between the antimicrobial compounds and the hydrolytic enzymes have already been reported for Trichoderma spp. strains (^{Saravanakumar et al., 2017}) and linked with the capacity of acting by mycoparasitism. In turn, TmE was the strain that produced greater xylanase levels, which could degrade the xylans in wheat straw, favoring its degradation. According to the results of this study, TmE identified as T. atroviride could be a good antagonist capable of colonizing wheat straw and inhibiting the development of F. graminearum mycelium and perithecia, thus contributing to Fusarium Head Blight control by decreasing the primary inoculum.

Figure 8: Growth of F. graminearum on wheat straw in absence (a) and presence (b) of T. atroviride TmE. Black points in a) are F. graminearum perithecia observed under optical microscope with a magnification of 50X. 

Correlation between antagonism characteristics

A Spearman correlation analysis was performed with the results obtained for the strains from the Viride clade in the assays for chitinolytic activity, pathogen inhibition by volatile or soluble antifungal compounds, pathogen growth in dual cultures in plates, and the production of perithecia by the pathogen in dual cultures on wheat straw. Table 2 shows the results of the analysis. The main diagonal shows the correlation of variables with themselves (Value 1). The values above the main diagonal correspond to the probabilities associated to the null correlation hypothesis between the variables. The values located beneath the main diagonal show the correlation coefficients between the variables. Through these analyses, a significantly (P ≤ 0.05) strong positive correlation (0.85) was determined between the pathogen growth in dual culture against strains from the Viride clade and the production of perithecia by the pathogen on wheat straw in the presence of those antagonists. Moreover, this study confirmed that the more an antagonist affected pathogen growth in the dual culture in Petri dish, the greater results in inhibition of perithecia production by the pathogen in wheat straw were observed in the presence of such antagonist. Likewise, a good negative correlation (-0.78) with a lower level of significance (P = 0.12) was obtained between chitinase production by the antagonists and the production of perithecia by the pathogen in wheat straw in presence of such antagonists. The chitinase production also showed a negative correlation (-0.80) with pathogen growth in the dual culture in the presence of the different Trichoderma spp. strains. In this case a greater chitinase production was correlated with lower pathogen growth in the dual culture (P = 0.10). The inhibition of the pathogen by volatile and soluble antifungal compounds produced by the Trichoderma spp. strains did not show a significant correlation with the inhibition of perithecium production by the pathogen or the antagonist activity in the dual culture. According to these results, only would the antagonism study in dual culture in Petri dish and the chitinase production in presence of the pathogen walls be necessary to determine the potential of the Trichoderma strains as biocontrol agents of F. graminearum production of perithecia on wheat straw, making it unnecessary to perform the rest of the assays to estimate such potential.