Potential of Trichoderma spp. for Biocontrol of Aflatoxin-Producing Aspergillus flavus

The inhibitory action of 20 antagonistic Trichoderma isolates against the aflatoxigenic isolate A. flavus ITEM 9 (Af-9) and their efficacy in reducing aflatoxin formation in vitro were examined. Production of metabolites with inhibitory effect by the Trichoderma isolates was also investigated. Antagonistic effect against Af-9 was assessed by inhibition of radial growth of the colonies and by fungal interactions in dual confrontation tests. A total of 8 out of 20 isolates resulted in a significant growth inhibition of 3-day-old cultures of Af-9, ranging from 13% to 65%. A total of 14 isolates reduced significantly the aflatoxin B1 (AfB1) content of 15-day-old Af-9 cultures; 4 were ineffective, and 2 increased AfB1. Reduction of AfB1 content was up to 84.9% and 71.1% in 7- and 15-day-old cultures, respectively. Since the inhibition of Af-9 growth by metabolites of Trichoderma was not necessarily associated with inhibition of AfB1 production and vice versa, we investigated the mechanism of reduction of AfB1 content at the molecular level by examining two strains: one (T60) that reduced both growth and mycotoxin content; and the other (T44) that reduced mycotoxin content but not Af-9 growth. The expression analyses for the two regulatory genes aflR and aflS, and the structural genes aflA, aflD, aflO and aflQ of the aflatoxin biosynthesis cluster indicated that neither strain was able to downregulate the aflatoxin synthesis, leading to the conclusion that the AfB1 content reduction by these Trichoderma strains was based on other mechanisms, such as enzyme degradation or complexation. Although further studies are envisaged to identify the metabolites involved in the biocontrol of A. flavus and prevention of aflatoxin accumulation, as well as for assessment of the efficacy under controlled and field conditions, Trichoderma spp. qualify as promising agents and possible alternative options to other biocontrol agents already in use.


Introduction
Aflatoxins are a group of potent mycotoxins produced by moulds of the genus Aspergillus in the course of spoilage of agricultural products, stored commodities, feeds and foods. Among the different forms of aflatoxins, aflatoxins B 1 , B 2 , G 1 and G 2 are especially important [1], since they exhibit carcinogenic, mutagenic and hepatotoxic effects [2]. The most potent form is aflatoxin B 1 (AfB 1 ), which is classified as carcinogenic to humans (group I) by the International Agency for Research on Cancer [3,4]. The main aflatoxigenic species is Aspergillus flavus [5], a saprophytic soil fungus with a large degree of genetic diversity, hundreds of different vegetative compatibility groups [6] and morphologically variable types, distinguished into two morphotypes based on sclerotia size, i.e., Group I (S strains) with sclerotia <400 µm in diameter and Group II (L strains) with sclerotia >400 µm in diameter [7]. L strains produce aflatoxins B 1 and B 2 , are quite variable in the levels of aflatoxin produced and also include non-producing strains (non-aflatoxigenic strains). Conversely, S strains show less variation in aflatoxin production, are generally higher aflatoxin producers than L strains and can produce aflatoxins G 1 and G 2 in addition to aflatoxins B 1 and B 2 [8]. Another important aflatoxigenic species is A. parasiticus, which is able to produce all four of the above aflatoxins. Aspergillus flavus and A. parasiticus differ in host range and habitat. Aspergillus flavus occurs more widely on cereals, oilseeds and dried fruits, including economically important crops, such as peanuts and corn [9][10][11], while A. parasiticus is more strictly associated with soil environment and infections of belowground plant organs [12].
The occurrence of aflatoxins in foods and feeds is a major economic and food safety issue worldwide. Due to the danger to human and animal health, aflatoxins are highly regulated in most countries in the world, although regulations are not harmonized. To protect humans and animals from exposure to aflatoxins, the US Food and Drug Administration has set a limit of 20 µg/kg (ppb) of total aflatoxins in food and feed [13]; in China, according to the National Standards on Mycotoxins Limit in Food, the limits for AfB 1 are 20 µg/kg in peanut, 10 µg/kg in rice and 5 µg/kg in wheat and related cereals [14]. In the European Union, the guidelines are more strict, setting a limit of 8 µg/kg AfB 1 in food and 0.1 µg/kg AfB 1 in cereal-based foods for babies and young children [15]. According to the Rapid Alert System for Food and Feed reports [16], in the last ten years AfB 1 has been a frequent cause of rejection of imported products, leading to severe economic losses, especially for developing countries [17].
Aflatoxin-producing Aspergillus spp. generally infect plants in the field, leading to contamination of crops with aflatoxins at harvest and further aflatoxin accumulation under storage if temperature and humidity are not controlled properly [18]. Infections and aflatoxin contamination are more severe under plant-stressing climatic conditions, such as high heat and drought [19,20], thus the climate change and global warming are expected to increase the risk of aflatoxin occurrence in temperate areas in the years to come. In this scenario, management of aflatoxigenic fungi at pre-harvest becomes a fundamental part of a multiple action strategy to reduce the risk of aflatoxins at consumption [18]. Good cultural practices, such as balanced watering and fertilization, which reduce plant stress and injuries caused by pests, have mitigating effects on A. flavus infections and aflatoxin occurrence. However, these practices are not always possible or sufficient to lower the infections of A. flavus and aflatoxin formation to acceptable levels. The development of cultivars that are less subjected to A. flavus infections or aflatoxin biosynthesis still encounters major hurdles, including a lack of resistant genotypes in some crop species (e.g., cotton), the long time required for breeding programs and questionable stability of the resistance conferred by the currently available genes [21], while transgenic approaches are not universally accepted.
Studies on the genetic and aflatoxigenic diversity in A. flavus and on the dynamics of A. flavus populations of high and low aflatoxin-producing genotypes sharing the same niche have led to the development of a biological control strategy based on competitive exclusion [18,22]. The biocontrol relies on the inundative introduction of atoxigenic strains that are incapable of producing aflatoxins because they either lack necessary aflatoxin biosynthesis genes or have deleterious mutations in critical biosynthesis genes [23]. The atoxigenic A. flavus strains, once spread on the soil surface, with time displace the native aflatoxigenic genotypes by a mechanism of competition [22,24]. Currently, this strategy is being intensively studied and applied in an attempt to replace existing methods of chemical control, which may leave toxic residues or lead to development of resistant biotypes of the pathogen [25]. Nevertheless, the use of non-aflatoxigenic strains is not without drawbacks [26,27]. Aspergillus flavus is a heterothallic fungus, capable of sexual reproduction between individuals with different mating type loci (Mat1-1 and Mat1-2). Recombination can occur during sexual crosses and has also been detected between aflatoxigenic and non-aflatoxigenic strains. Therefore, the possibility exists that the capability to produce aflatoxins is transferred to the progeny of atoxigenic strains. In addition, other toxic metabolites on top of aflatoxins may be produced by A. flavus, including cyclopiazonic acid, aflatrem, aflavinines, paxillines, paspalicines and pseurotin [28,29]. These mycotoxins are not regulated and are not tracked by surveillance programs, but epidemiological data suggest that they might exert toxic effects on their own or act synergistically with aflatoxins [30,31]. Since the biosynthesis of those toxins is genetically regulated independently of aflatoxins, the non-aflatoxigenic biocontrol strains might still be capable of producing one or more of these other mycotoxins. Finally, it should be noted that, while the biocontrol A. flavus strains are non-aflatoxigenic, they are not necessarily non-pathogenic. In other words, they are able to reduce the level of aflatoxin contamination but might still cause severe crop disease and yield losses.
Apart from non-aflatoxigenic A. flavus strains, antagonistic microbes, including various bacteria [32][33][34][35][36][37][38][39][40] and yeasts [41][42][43][44], have shown remarkable inhibitory effects on the growth and/or on aflatoxin production of aflatoxigenic strains and have been considered as potential candidates for biocontrol of aflatoxin contamination at pre-harvest. Among the antagonists, fungi of the genus Trichoderma are possibly the most diffused biocontrol agents worldwide [45]. Trichoderma are free-living, mostly soil-resident, filamentous fungi that comprise rhizosphere-competent strains associated with root ecosystems [46]. Beneficial effects of Trichoderma include the ability to suppress both soil-borne and foliar plant pathogens, improvement of plant resistance to abiotic stresses and diseases, enhancement of mineral plant nutrition, promotion of plant growth and increase in crop productivity [47,48]. The outstanding success of Trichoderma spp. as biocontrol agents arises from their capability to exhibit multiple mechanisms of antagonistic action against plant pathogens [49]. Being fast-growing and metabolically flexible fungi able to use a variety of organic materials as food source, they are excellent competitors in soil environment. In addition, they produce a number of secondary metabolites inhibitory to plant pathogens [50], which function as chemical weapons within the so-called interference competition [51]. Although Trichoderma spp. are historically among the most studied agents for the biocontrol of soil pathogens, so far only few Trichoderma species and strains have been tested for biological control of A. flavus and aflatoxin production [52][53][54]. In this study we examined the inhibitory action of 20 antagonistic Trichoderma isolates against an aflatoxigenic strain of A. flavus and their effect on aflatoxin production, with an emphasis on the effect of Trichoderma metabolites and their involvement in the mode of action. In this regard, the effect of metabolites produced by two Trichoderma strains (namely T. harzianum T44 and T. polysporum T60) on the regulation of aflatoxin biosynthesis genes was investigated. In particular, we focused on the expression of the two regulatory genes aflR and aflS and four structural genes of the aflatoxin biosynthesis cluster, aflA, aflD, aflO and aflQ, by using the reverse transcriptase quantitative approach. Our results suggest that, unlike other microbial agents [55], Trichoderma are not able to downregulate the toxin biosynthesis, and aflatoxin reduction is based on other mechanisms, such as enzyme degradation or complexation.

Antagonism of Trichoderma Isolates against A. flavus
The average daily radial growth of A. flavus ITEM 9 (Af-9) was 4.6 mm/day; this is considerably slower than that of most of the Trichoderma isolates. The daily radial growth of Trichoderma colonies varied from 4.9 to 22.9 mm/day; that is up to three to four times faster than Af-9 ( Figure 1). The fast-growing Trichoderma isolates rapidly spread and surrounded the colony of A. flavus, thus stopping its further enlargement. In several confrontations, the early contact between Trichoderma isolates and A. flavus resulted in the arrest of A. flavus growth before the differences between R A1 and R A2 became apparent (Figures 2 and 3). As a result, in these confrontations, the value of percent inhibition of Af-9 colony growth was not statistically different from the control values (Table 1), in spite of the capability of Trichoderma to stop the growth of Af-9. More insight into the antagonistic potential of the Trichoderma isolates against A. flavus was obtained with the study of colony interactions, observed after 21 days of co-culture (Table 1, Figure 3). Overall, only type 1, 3 and 4 colony interactions were observed (Table 1); in none of the interactions have we observed overgrowth of A. flavus on Trichoderma spp. The Trichoderma isolates T32, T50, ITEM 908, T61, T60 and T62 arrested the growth of A. flavus after contact and aggressively overgrew the A. flavus colony (interaction type 1). The Trichoderma isolates T58, T38, T48, T8, T36 and T46 showed a different type of aggressive interaction with A. flavus, which resulted in a clear, ≥4 mm-wide inhibition zone (type 4). The rest of the strains (ITEM 4484, T54, ITEM 908-5, T11, T37, T41, T44, T51) showed a less aggressive interaction (type 3), with a mutual ≤2 mm-wide inhibition zone.  Representation of dual culture tests with Trichoderma spp. and A. flavus Af-9. T and A are the inoculation points of Trichoderma and Af-9, respectively. Undisturbed growth of either Af-9 or Trichoderma was assumed to be R A1 and R T , respectively. The percent inhibition of Af-9 radial growth was calculated as %I DC = (R A1 − R A2 )/R A1 × 100%.

Inhibitory Effect of Trichoderma Metabolites on Growth of A. flavus
Production of metabolites inhibitory to Af-9 by the 20 Trichoderma spp. isolates was investigated in both the standard Czapek Dox-Agar medium (CDA) and in Czapek Dox-Agar supplemented with 2% peanut flour (CDP). For all the isolates of Trichoderma tested, the inhibitory effect on growth of A. flavus was lower in CDP compared to CDA ( Figure 4A,B). On CDA, 16 out of 20 isolates of Trichoderma significantly inhibited Af-9; in percentages, that ranged from 9 ± 1% to 100 ± 0% ( Figure 4A). On CDP, only 8 out of 20 isolates resulted in a significant growth inhibition of Af-9; in percentages, that ranged from 13 ± 1% to 65 ± 3% ( Figure 4B). All of them were also inhibitory on CDA. The most effective strains belonged to the species T. atroviride (T32 and T50), T. citrinoviride (ITEM 4484), T. harzianum species complex (ITEM 908-5, T11, T41, T61) and T. polysporum (T60). The strain T60 was the most effective in both media, resulting in 65 ± 3% and 100 ± 0% growth inhibition on CDP and CDA, respectively. Czapek Dox-Agar supplemented with 2% peanut flour (B) (see text for more details) for 5 or 3 days, respectively. After removal of the cellophane sheet, the dishes were inoculated with Af-9 and incubated at 25 • C with 12/12 photoperiod for 6 days. Values are the means ± SD (n = 3) of the percent reduction of colony diameter with respect to control. Asterisks indicate statistically significant values at p < 0.05 (*), p < 0.01 (**) or = p < 0.001 (***) by one-way ANOVA.

Inhibitory Effect of Trichoderma Metabolites on AfB 1 Production by A. flavus
The inhibitory effect of Trichoderma metabolites on AfB 1 production by Af-9 was initially investigated using both the media CDA and CDP. Analyses of AfB 1 content in the media carried out after 7 and 15 days of growth showed that Af-9 was not able to produce AfB 1 on CDA, regardless of the presence or absence of Trichoderma metabolites (data not shown). On the contrary, 7.3 ± 1.0 and 11.2 ± 2.6 µg/g of AfB 1 were detected in control plates of CDP, respectively after 7-and 15-day growth of Af-9. Therefore, only CDP was used for the assessment of the inhibitory activity of Trichoderma metabolites on AfB 1 production.
The inhibitory effect of Trichoderma isolates on AfB 1 production is shown in Figure 5. Most Trichoderma isolates (14 out of 20) released metabolites in the medium that reduced significantly (p < 0.001) the production of AfB 1 , both at 7 and 15 days ( Figure 5). In two cases (isolates T54 and T51), the inhibitory effect was temporary, since it was observed at 7 days but not at 15 days ( Figure 5). Two isolates (T38 and T37) did not inhibit production of AfB 1 either at 7 or 15 days ( Figure 5). The reduction of AfB 1 production by Af-9 ranged from 5.7 to 84.9% and from 2.5 to 71.1%, respectively, after 7 and 15 days. Interestingly, the isolates ITEM 4484 and T8 increased the production of AfB 1 by Af-9, in spite of the former isolate being significantly inhibitory and the latter not being significantly inhibitory to Af-9 growth ( Figure 4). The increase in AfB 1 production was initially as high as 40% and then decreased to 30% at 15 days of growth ( Figure 5). Figure 5. Percent reduction of AfB 1 production by Af-9 grown for 7 and 15 days on CDP containing Trichoderma metabolites, with respect to control. In control plates, 7.3 ± 1.0 and 11.2 ± 2.6 µg/g of AfB 1 were produced after 7 and 15 days, respectively. Values are the means ± SD of three replicates; statistically significant differences with control by one-way ANOVA are indicated by asterisks above the bar (*** = p < 0.001).

Analysis of Aflatoxin Biosynthesis Gene Expression in Relation to the Control of Trichoderma
In order to investigate the molecular mechanism of the inhibitory effect exerted by Trichoderma metabolites on AfB 1 production, the expression levels of aflatoxin biosynthesis genes were studied in Af-9 grown on CDP plates that were pre-inoculated with either Trichoderma isolate T60 or isolate T44. These two isolates were presumed to have different modes of inhibition of AfB 1 production. Indeed, while T60 metabolites inhibited both mycelial growth and AfB 1 production, T44 was very effective in inhibiting AfB 1 production but did not significantly inhibit the growth of Af-9 (Figures 4 and 5).
In this regard, the AfB 1 production was determined per milligram of mycelium fresh weight so that, through the normalization of AfB 1 content per unit of Af-9 biomass, the effect of reduction of AfB 1 due to reduced growth was distinguishable from reduction due to the downregulation of biosynthesis. AfB 1 was measured in the mycelium of Af-9 after 2 and 5 days of growth. As shown in Figure 6A, in the presence of T44 metabolites, a significant reduction (90.3%) of AfB 1 production was observed after 5 days of growth. Conversely, when A. flavus was grown on plates pre-inoculated with T60, no significant difference was observed in the normalized production of AfB 1 either at 2 or 5 days after inoculation (d.a.i.). The expression levels of the two regulatory genes (aflR and aflS) and four structural genes (aflA, aflD, aflO and aflQ) of the aflatoxin biosynthesis cluster were analyzed in Af-9 after 2 days of growth in control conditions or in plates pre-inoculated with the strain T44 or the strain T60. In Af-9 grown on CDP that was pre-inoculated with the isolate T44, the two regulatory genes aflR and aflS had a similar transcriptional trend, with a higher expression level than in control; on the contrary, no difference from control was observed when strain T60 was used ( Figure 6B). A different transcriptional profile was observed among the structural genes ( Figure 6B). In particular, for both the Trichoderma strains the expression levels of aflA and aflD showed no significant variation between control and treatment. On the other hand, the expression level of aflO decreased when Af-9 grew on medium pre-inoculated with either Trichoderma strains, more markedly with strain T44. By contrast, the expression level of aflQ increased significantly, as did that of the regulatory genes aflR and aflS when strain T44 was used in the pretreatment of the growth medium. flavus Af-9 mycelium after 2 and 5 days of growth on medium untreated (control, c) and pre-inoculated with Trichoderma strains T44 or T60; (B) expression analyses of aflatoxin biosynthesis genes in Af-9 after 2 days of growth on medium untreated (control, c) and pre-inoculated with Trichoderma strains T44 and T60. The β-tubulin gene was used as reference gene. Significantly differential gene expression is indicated (* p-value ≤ 0.05; ** p-value ≤ 0.01).

Discussion
Contamination of commodities with aflatoxins is a major issue for food safety and trade worldwide. Control of contamination at multiple points of the food chain, cultivation, transit, storage and processing, is necessary, but the control of infections by aflatoxigenic fungi at pre-harvest appears especially critical [11]. Spreading non-toxigenic A. flavus strains that competitively exclude aflatoxigenic strains in the field was first introduced by Cotty and Bayman [56], and this approach has proved to be successful in many cases, leading up to 80% reduction of aflatoxin contamination. It is currently utilized in cottonand maize-growing areas of the USA and in Kenya and has led to the development of commercial biopesticides AF36 (non-aflatoxigenic strain NRRL 18543) and Afla-Guard ® (non-aflatoxigenic strain NRRL 21882) that are marketed in the USA for biocontrol of aflatoxins in groundnut and maize, respectively. However, some drawbacks of the method have been pointed out [26,27], which prompt the search for additional low-environmentalimpact methods of control of aflatoxigenic fungi and aflatoxin in the field. Trichoderma spp. are well known and widely used biocontrol agents of root and foliar plant pathogens. Apart from the direct action against the pathogens, these beneficial fungi are also able to enhance plant mineral nutrition [47] and activate plant defenses and resistance against abiotic and biotic stresses [57,58], including drought [59] and herbivore insects [60,61]. Lately, a few research works reported on the capability of Trichoderma spp. to control A. flavus in the field, both in maize and peanut [62][63][64]. In the present research work, we have investigated the potential of 20 Trichoderma isolates belonging to different species for biological control of A. flavus and aflatoxins, with a focus on the effect of metabolites on A. flavus growth, AfB 1 production and expression of genes of aflatoxin biosynthesis.
The confrontation test, also known as dual culture, is a method widely utilized for the in vitro selection of effective biocontrol strains of Trichoderma against various pathogens and has also been utilized for picking isolates antagonistic to A. flavus [65][66][67]. Usually, in this bioassay, the antagonistic capability of the biocontrol agent against a given pathogen is assessed by determination of the percent inhibition of colony radial growth of the challenged pathogen (%I). In our trials, we have observed that this parameter is not per se enough for a correct evaluation of the potential antagonistic capability of Trichoderma isolates against Af-9. Indeed, some aggressive and fast-growing Trichoderma isolates did not score high %I values, since the early contact between Trichoderma and Af-9 resulted in the arrest of pathogen growth before significant differences between the challenged colony growth and the unchallenged colony growth of Af-9 appeared. Therefore, it is very important that the %I value is carefully considered along with the type of colony interaction that occurs after the contact between the colonies, particularly the capability of the Trichoderma isolate to aggressively overgrow A. flavus (type 1 interaction).
Our experiments showed that in the antagonism of Trichodema spp. against Af-9, an important role is played by the production of metabolites. Metabolites produced by Trichoderma isolates and released in the growth medium were able to inhibit both the colony growth of Af-9 and its AfB 1 production. However, it should be stressed that the effect of metabolites in inhibiting Af-9 growth was strongly affected by the medium used for the assessment. For all the isolates of Trichoderma tested, the inhibitory effect on A. flavus growth was higher in CDA compared to the same medium supplemented with 2% peanut flour (CDP). It is conceivable that some components of peanut flour either reduce the production of inhibitory metabolites by Trichoderma or enhance the A. flavus resistance. From a practical standpoint, this highlights the necessity to use media that contain natural components of the host plant for testing antagonistic efficacy of Trichoderma biocontrol isolates. Likewise, the presence of peanut flour in the medium enhanced the mycotoxigenicity of Af-9 and AfB 1 biosynthesis, consistently with what had been observed in previous studies with other plant components [68,69]. It should also be noted that, for this experiment, "true" control plates were difficult to design. The pre-growth of Trichoderma on the medium that was subsequently exploited by Af-9 resulted in partial depletion of nutrients, which may have affected both Af-9 growth and AfB 1 production. Conversely, depletion did not occur in control plates. Nevertheless, the experiment allowed us to assess the biocontrol capability of the different Trichoderma isolates comparatively. The inhibition of A. flavus colony growth by metabolites of Trichoderma was not necessarily associated with inhibition of AfB 1 production and vice versa. Metabolites of the isolates T44 and T60 that were released in CDP proved to have a strong inhibitory effect on AfB 1 production (approx. by 65-85% for T44 and 48-63% for T60, Figure 5). However, these two strains had a different effect on A. flavus growth (Figure 4). While T44 did not significantly inhibit mycelial growth, T60 had the highest inhibitory effect among all the strains tested (approx. by 65% on CDP). A similar differential effect on fungal growth and aflatoxin biosynthesis has been previously reported for some plant metabolites as well [70]. On this premise, it seemed interesting to investigate which inhibitory mechanism led to the reduction of AfB 1 content, given the fact that only in the case of T60, but not in that of T44, could the lower amount of the mycotoxin be due to the reduction of Af-9 growth. Hence, the effect of metabolites of T44 and T60 on production of AfB 1 was studied in correlation with the expression of some genes of aflatoxin biosynthesis cluster, in order to investigate the possible interference of Trichoderma metabolites at some points of the aflatoxin biosynthesis pathway. When the AfB 1 content was analyzed in the 5-day-old mycelium of Af-9, the results confirmed that the reduction caused by metabolites of T60 was due to the inhibition of growth, while for metabolites T44, a different mechanism was conceivable. Recent studies have reported that the use of biocontrol agents or natural products can inhibit the production of AfB 1 by downregulation of the genes involved in the biosynthesis of AfB 1 , even though the molecular mechanism behind this inhibition activity has not yet been completely determined [71]. For our study, we selected the two aflatoxin biosynthesis regulatory genes aflR and aflS, and the four structural genes, aflA, aflD, aflO and aflQ, for analysis of gene expression in 2-day-old mycelium of Af-9, since the regulation and expression of biosynthetic genes usually precede the production and accumulation of aflatoxin in the mycelium [72]. The aflR and aflS genes encode cluster-specific transcriptional factors that were suggested to interact to modulate and coordinate the expression of aflatoxin structural genes [73,74]. Among the selected structural genes, aflA and aflD act at the early phase of the biosynthesis pathway. The aflA gene encodes one of the two fatty acid synthases responsible for the first polyketide structure of the aflatoxin molecule, and the aflD gene encodes the enzyme responsible for the formation of the intermediate averantin. The other two genes aflO and aflQ are involved in the production of sterigmatocystin and hydroxyl-methylsterigmatocystin, the precursor of AfB 1 , respectively, acting in the later stages of the biosynthesis pathway [71]. From the analysis of expression levels in A. flavus grown in the presence of T60 metabolites, no difference was found compared to control. These findings were consistent with the results of AfB 1 content, confirming that the T60 metabolites did not affect the biosynthetic mechanism of the mycotoxin. In the case of T44 metabolites, which led to a reduction in AfB 1 production, the expression levels of most biosynthetic genes did not appear to be directly related to inhibitory activity. In fact, both the regulatory genes aflR and aflS and the structural gene aflQ showed an upregulation compared to the control. Conversely, the transcriptional levels of aflA and aflD genes remained unchanged, while only the aflO gene seemed downregulated compared to the control and apparently in line with the reduction in AfB 1 production. Usually, the inhibitory activity exerted by biological control agents on the AfB 1 biosynthesis by A. flavus is correlated with the downregulation of all or most of the approximately 27 genes composing the aflatoxin biosynthesis cluster [55,75]. In our study, we observed an increase in the expression of aflR and aflS in the presence of T44 metabolites, and only the aflO was downregulated. Further analyses are necessary to examine whether the expression of other genes of the biosynthesis cluster may be affected and establish whether the negative regulation of AfB 1 production occurs at the biosynthesis cluster level or during translational and/or post-translational stages, causing low levels of proteins or lack of functionality. It cannot be excluded that some of the metabolites produced by strain T44 could activate a signaling pathway, leading to the upregulation of regulatory genes aflR and aflS of the aflatoxin cluster. On the other hand, the observed reduction in aflatoxin accumulation in the culture media could be due to mechanisms triggered by other Trichoderma metabolites. Several works have reported the ability of microorganisms, including Trichoderma species, to produce enzymes capable of degrading or modifying mycotoxin molecules [76]. Finally, it should be considered that in confrontation tests, 2 out of the 20 Trichoderma strains tested, namely ITEM 4484 and T8, resulted in an augmented capability of Af-9 to produce AfB 1 (Figure 5), even if the antagonists significantly reduced Af-9 growth. While upregulation of AfB 1 biosynthesis genes cannot be ruled out, other hypotheses, including digestion of medium components by Trichoderma enzymes, which generate compounds boosting AfB 1 production, also deserve further investigation.
The option of using Trichoderma spp. for biocontrol of aflatoxigenic fungi may offer a few advantages in respect to the use of non-aflatoxigenic A. flavus strains. First, the non-aflatoxigenic A. flavus strains are not necessarily non-pathogenic, and they might still cause disease and yield loss. As broad-spectrum biocontrol agents, Trichoderma may also protect the plants from the attack of other plant pathogens in addition to A. flavus [58,77,78]. Finally, the indirect action of Trichoderma on the enhancement of plant resilience to drought stress and on the prevention of insect pest damages, both of which are factors facilitating aflatoxin occurrence [79], is one more point for consideration of Trichoderma in the context of A. flavus control and prevention of aflatoxin occurrence.

Conclusions
In conclusion, the results reported herein have shown the potential of isolates belonging to different species of Trichoderma to be biocontrol agents for A. flavus and prevent aflatoxin accumulation. Both of these properties may be heavily influenced by the presence of host-plant-derived components in the growth medium; therefore, the in vitro selection of effective strains for further assessment should be compiled, taking into account the crop on which they will be used. The biocontrol Trichoderma strains may arrest fungal growth, reduce aflatoxin production or both. Metabolites of Trichoderma spp. play a role in their mechanism of action, although with a diversity of modes. Some are inhibitory to A. flavus growth; others reduce the accumulation of aflatoxin, presumably via degradation. Thus, an effective biocontrol strategy may be based on the combined use of multiple isolates with different mechanisms of action. Although further studies are envisaged to identify the metabolites involved in the biocontrol of A. flavus and prevention of aflatoxin accumulation, as well as for assessment of the efficacy under controlled and field conditions, Trichoderma spp., including a few strains tested within the present work, qualify as promising agents and possible alternative options to other biocontrol agents already in use.

Fungal Strains
The isolates of Trichoderma spp. used in this study were collected from soil or plant debris, or obtained from the culture collection of the Institute of Science of Food Production (ITEM Collection, http://www.ispa.cnr.it/Collection/ (accessed on 2 December 2021), Bari, Italy) ( Table 2). All the isolates were identified morphologically, according to Gams and Bisset [80]. The commercial biocontrol strain T. atrobrunneum ITEM 908 (T. harzianum species complex, formerly T. harzianum ITEM 908) was previously molecularly characterized [81] by analysis of the sequences of the internal transcribed spacer regions ITS-1 and ITS-2 of the nuclear rDNA and of a fragment of the translation elongation factor gene TEF-1α, according to Chaverri et.al. [82]. The mutant strain ITEM 908-5 was generated by UV irradiation of the parental strain ITEM 908 [83]. The AfB 1 -producing strain Aspergillus flavus ITEM 9 (Af-9 = NRRL 3251, ATCC 36061) was used for all the tests throughout the work. Af-9 was originally isolated from walnuts in the USA and is a high AfB 1 -producing strain [84]. Fungal cultures were maintained in purity on potato dextrose agar (PDA, Oxoid, Italy) slants at + 5 • C, which were used for preparation of fresh cultures and inocula.

Antagonism of Trichoderma spp. against A. flavus
The antagonism and colony interaction between each Trichoderma spp. isolate and Af-9 were studied in vitro by dual cultures. Petri dishes of 9 cm in diameter containing 20 mL of PDA were inoculated 1 cm apart from the edge of the plate with a 6 mm-diameter mycelial plug from a fresh culture of Trichoderma and on the opposite side, at a 7 cm distance from the Trichoderma inoculation point, with a 10 µL drop of spore suspension of Af-9 containing 1 × 10 7 conidia/mL in sterile distilled water. All the pathogen-antagonist co-cultures were incubated at 25 • C ± 1 with 12/12 photoperiod. The radial growth of both the colonies was measured daily until contact or until the arrest of colony growth if contact did not occur because of mutual inhibition, and the average daily radial growth (mm/day) of each fungus was calculated. The percent inhibition of Af-9 radial growth in dual cultures (%I DC ) was calculated as where R A1 was the longest radius of the Af-9 colony, and R A2 was the radius of the Af-9 colony along the line that connected the Af-9 and the Trichoderma inoculation points (Figure 2), measured on the day of contact or the last day of incremental growth, if contact did not occur. At 21 days after inoculation (d.a.i.), the colony interactions were assessed visually and classified in accordance with Whipps [52] as: 1 = Trichoderma overgrowing Af-9 and Af-9 stopped; 1/2 = Trichoderma overgrowing Af-9 but Af-9 still growing; 2/1 = Af-9 overgrowing Trichoderma but Trichoderma still growing; 2 = Af-9 overgrowing Trichoderma and Trichoderma stopped; 3 = slight mutual inhibition (inhibition zone ≤ 2 mm-wide); 4 = strong mutual inhibition (inhibition zone ≤ 4 mm-wide). The experiment was carried out in triplicate.

Effect of Non-Volatile Metabolites of Trichoderma spp. on A. flavus Growth
The inhibitory effect of Trichoderma metabolites on the growth of A. flavus was studied in Czapek Dox-Agar (CDA, Biolife Italiana, Monza, Italy) and Czapek Dox-Agar supplemented with 2% (w/v) peanut flour (CDP). For preparation of CDP, peanuts were finely ground in a laboratory mill (Mulino Cyclone, International PBI, Milano, Italy) to particles of <0.2 mm; 0.4 g of ground peanuts were transferred into 2.5 cm-diameter and 15 cm long test tubes that were filled with 20 mL of melted CDA and autoclaved at 121 • C for 15 min. After cooling at 55 • C, the medium was thoroughly mixed by a vortex mixer and quickly poured into 9 cm-diameter Petri dishes. After solidification of the medium, a pre-autoclaved (121 • C, 15 min) cellophane disc was laid on the medium to cover the entire surface. Mycelial plugs of 6 mm in diameter were removed from the edge of 5-day-old cultures of Trichoderma and transferred to the center of the Petri dishes, onto the cellophane disc. The cellophane membrane was used to keep the mycelium from invading the medium, at the same allowing time the metabolites to diffuse into the substrate through the cellophane pores. The colonies of Trichoderma were grown for 5 days on CDA and for 3 days on CDP, at 25 ± 1 • C with 12/12 photoperiod; then, the cellophane sheets with the overlying Trichoderma colonies were removed. The plates were center-inoculated with a 10 µL drop of a 1 × 10 7 conidia/mL spore suspension of Af-9 and incubated at 25 ± 1 • C with 12/12 photoperiod for 6 days. Control plates were prepared by inoculating Af-9 on CDA and CDP without previous cultivation of Trichoderma. The experiment was carried out in triplicate. The growth of A. flavus was assessed by the colony diameter, measured with a ruler under a dissecting microscope every 24 h. The percent inhibition of the diametral growth of Af-9 colonies (%I D ) caused by Trichoderma metabolites was calculated with the following formula %I D = (D C − D T )/D C × 100 (2) where D T was the average diameter of Af-9 colonies grown on media pre-inoculated with Trichoderma, and D C was the average diameter of Af-9 colonies grown in control plates.

Determination of AfB 1 by UPLC
AfB 1 was determined by Ultra Performance Liquid Chromatography (UPLC) using the Acquity UPLC system (Waters, Milford, MA, USA). Data acquisition and instrument control were performed by Empower 2 software (Waters). The column used was a 100 mm × 2.1 mm i.d., 1.7 µm, Acquity UPLC1 BEH RP-18, with an Acquity UPLC column in-line filter (0.2 µm), detected by fluorometric detector without post-column derivatization. The fluorometric detector was set at wavelengths of 365 nm (excitation) and 435 nm (emission). The mobile phase was a mixture of water-acetonitrile-methanol (64:18:18, v/v/v) at a flow rate of 0.4 mL/min. The temperature of the column was maintained at 40 • C. In these experimental conditions, the retention time of the AfB 1 standard was 3.7 min (see Supplementary Figure S1). AfB 1 was quantified by measuring the peak areas of the samples at the retention time of the aflatoxin standard and comparing these areas with the calibration curve of AfB 1 in the range of 0.2 to 10.0 ng/mL. The limit of quantification (LOQ) of the method was 0.2 ng/mL for AfB 1 , based on a signal to noise ratio of 10:1.

Effect of Trichoderma Metabolites on Aflatoxin Production by A. flavus
Media, inoculations and controls were made in the same way as described for the experiment on the effect of Trichoderma non-volatile metabolites on the growth of Af-9. Each Trichoderma isolate was tested in three replicates. After 7-and 15-day growth of A. flavus, the cultures were sampled for the determination of AfB 1 . An 8 mm-diameter cork-borer was used to excise mycelial discs at regular distances along the radius of the colony, starting from the inoculation point to the edge of the colony. Excised discs were then transferred to test tubes. The samples (approximately 1 g) were precisely weighted and extracted with 5 mL of a methanol-water (80: 20, v/v) solution in a KS 4000i orbital shaker (IKA Werke GmbH & Co. KG., Staufen, Germany) at 250 rpm for 60 min at room temperature. Samples were then centrifuged, filtered and diluted to obtain extracts, which were stored at −20 • C until AfB 1 analysis, according to previously described methods [85,86]. The inhibition rate of AfB 1 production (%I AfB1 ) was calculated with the following formula where AfB 1T was the amount of AfB 1 produced by A. flavus on medium pre-inoculated with Trichoderma, and AfB 1C was the amount of AfB 1 in the control plates. For this study, the Trichoderma strains T44 and T60 and Af-9 were grown on CDP. Preparation of CDP and cultivation of the fungi on cellophane sheets were conducted as described in Section 5.3. After 3 days of growth, the membrane and the colony of Trichoderma were removed from the Petri dish. Then, the medium was inoculated with 100 µL of a spore suspension of Af-9 containing 1 × 10 7 spores/mL, which were spread evenly with an L-shaped spreader. As controls, Af-9 was inoculated on cellophane sheets laid on CDP not pre-inoculated with Trichoderma. Cultures of A. flavus were incubated at 25 ± 1 • C with 12/12 photoperiod, and the mycelium was collected by scraping it off the cellophane at two time points, i.e., 2 and 5 d.a.i. The mycelium samples were stored at −20 • C until extraction and AfB 1 analysis, according to the procedure described in Section 5.4. For the 2 d.a.i. samples, the cellophane sheets were cut in half, and the mycelium grown in each half was scraped off, weighted, immediately frozen in liquid nitrogen and stored at −80 • C. The mycelium from one half plate was used for RNA extraction, while the mycelium from the other half plate was used for AfB 1 analysis. At each time point, analyses were performed using the pooled mycelium from 10 plates.

RNA Extraction and cDNA Synthesis
Total RNA was extracted using the RNeasy kit (Qiagen, Valencia, CA, USA), and an on-column DNase I treatment was performed, according to the manufacturer's protocols. The A 260 /A 280 ratio was determined using a NanoDrop 1000 (Thermo Fisher Scientific, Waltham, MA, USA), and gel electrophoresis was performed for qualitative analysis of total RNA. The cDNA was synthesized using 2.0 µg total RNA, oligo (dT) 18 primer and random hexamers, and SuperScript III Reverse Transcriptase (Invitrogen, San Diego, CA, USA), according to instructions of the manufacturer.

Expression Analysis of Aflatoxin Biosynthesis Gene by Quantitative Reverse Transcriptase PCR (RT-qPCR)
The StepOne™ Real-Time PCR System (Applied Biosystem, Waltham, MA, USA) was used to carry out RT-qPCR assays. The transcription profiles of 6 genes of the aflatoxin biosynthesis cluster (aflR, aflS, aflA, aflD, aflO, aflQ) were analyzed, and the housekeeping β-tubulin gene was used as reference gene. Sequences of primers listed in Table 3 were retrieved from previous publications [55,72]. The sequences of all genes were obtained from the A. flavus NRRL3557 genome database at NCBI (BioProject PRJNA575750). Each reaction was performed in a total volume of 20 µL containing cDNA (3.8 ng/µL), 12.5 µL of SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) and different concentrations of primers (Table 3). Nuclease-free water was added to make the total volume 20 uL. The PCR reaction consisted of initial denaturation at 95 • C for 5 min, 40 cycles at 94 • C for 30 s, 57 • C for 30 s and 72 • C for 15 s. The relative quantification of gene expression was established by using the 2 −∆∆Ct method [87]. Statistical analysis was performed by using the t-test.

Statistical Analysis
Data were analyzed by one-way analysis of variance (ANOVA) and Tukey-Kramer multiple comparison test. The statistical analyses were performed using the GraphPad Instat 3.0 software (GraphPad Software, San Diego, CA, USA). Figures were drawn using OriginPro9.0 software (OriginLab Corporation, Northampton, MA, USA).

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare that they have no competing interest.