Phylogenetic Affinity in the Potential Antagonism of Trichoderma spp. against Moniliophthora roreri

Abstract

Frosty pod rot, caused by the fungus Moniliophthora roreri, is one of the most important diseases of cacao in Peru. Sustainable alternatives to control it include timely cultural practices such as pruning, and the application of biological control agents such as Trichoderma spp. We isolated 234 Trichoderma strains native to the department of Amazonas in Northern Peru from soil samples in cacao farms. These strains belong to at least eighteen species within four phylogenetic clades in the genus (Harzianum, Longibrachiatum, Hamatum, and Brevicompactum clades). We aimed to assess the in vitro biocontrol potential of these strains against M. roreri. We evaluated their mycoparasitism, antibiosis, and potential antagonism to select candidate strains for efficient biocontrol of M. roreri. We found evidence (Kruskal–Wallis test, p < 0.005) that strains belonging to the Harzianum clade tend to have higher mycoparasitism, antibiosis, and potential antagonism levels than strains in the Longibrachiatum and Hamatum clades. Additionally, this study constitutes the first report for antagonistic behavior against M. roreri for T. parareesei, T. lentiforme, T. orientale, T. asperelloides, T. inhamatum, T. zelobreve, T. afarasin, T. ghanense, T. rifaii, and T. breve. These results will be foundational for further M. roreri biocontrol studies.

1. Introduction

Moniliophthora roreri is the causal agent of frosty pod rot (FPR), one of the most destructive diseases of cacao (Theobroma cacao L.) in Latin America [1]. It can cause 100% yield losses under poor management conditions, which has led to the abandonment of cacao plantations [2]. In Peru, it was first reported in Quebrada Seca district, Utcubamba province, Amazonas department, in 1989, becoming the main cacao pathogen of the country since then [1]. Frosty pod rot is a disease twice as destructive as the black pod rot caused by Phytophthora spp., and more difficult to control than witches’ broom, caused by M. perniciosa [3]. Therefore, its control is a major challenge in cacao cultivation [4]. To achieve sustainable cacao production, environmentally friendly control strategies must be implemented to overcome the problems caused by traditional chemical control methods [5]. Multiple strategies have been evaluated and recommended for the control of FPR. Cultural management practices such as purges, pruning, and timely harvests are very effective [6,7,8,9,10,11,12]. Additionally, the use of natural enemies of M. roreri such as Trichoderma spp. has been shown to be a sustainable alternative to mitigate the impacts of FPR [2,13,14].

Trichoderma spp. are filamentous fungi that live naturally in the soil in a close relationship with plant roots [15]. They are considered beneficial for agriculture because of their ability to protect the crops by acting as biological control agents to fungal pathogens [5,16,17,18,19]. Various plant diseases such as root rot, wilt, pod rot, and others can be mitigated by Trichoderma spp. [20,21,22,23,24,25].

In cacao, Trichoderma spp. has an important biocontrol potential against M. roreri and other major pathogens [2,9,13,14,19,26,27,28]. Trichoderma biocontrol of fungal pathogens can be achieved through mycoparasitism or antibiosis. Mycoparasitism is the ability of the fungus to grow and feed on other fungi [29]. Antibiosis is related to the production of metabolites that inhibit the growth of other organisms [30]. Most studies evaluating the Trichoderma spp. activity against M. roreri have identified strains and species with stronger biocontrol potential. However, they have only considered either mycoparasitism or antibiosis separately in their analysis. The integration of both biocontrol mechanisms, referred to as potential antagonism, has been proposed to more accurately evaluate the biocontrol capacity of Trichoderma spp. [13,14,31].

The objective of this study was to evaluate the in vitro potential antagonism against M. roreri of 234 strains in four phylogenetic clades of Trichoderma native to the Amazonas department in Northern Peru, based on their levels of mycoparasitism and antibiosis.

2. Materials and Methods

2.1. Acquisition of Native Strains of Trichoderma spp.

We evaluated 234 strains of Trichoderma recovered from the rhizosphere soil of T. cacao from farms growing the native fine aroma cacao (CNFA, Spanish acronym for “cacao nativo fino de aroma”) from the provinces of Condorcanqui, Bagua, Utcubamba, Chachapoyas, Luya, and Rodríguez de Mendoza in the department of Amazonas, Peru (Table S1). All strains were isolated using the dilution technique [32]. Monosporic cultures were then obtained from each identified colony [33]. Two hundred and sixteen strains belonged to eighteen species [34], while the identification of eighteen strains remained ambiguous based on the identification protocol proposed by Cai and Druzhinina [35]. However, all strains belonged to four major Trichoderma phylogenetic clades previously described [35,36] and interactively presented at © 2022 Trichokey (https://trichokey.com/; last accessed on 1 July 2022) (Table S1). One hundred and sixteen strains belonged to the phylogenetic clade 1 (also known as Harzianum clade), one hundred and four to clade 3 (Longibrachiatum clade), thirteen to clade 5 (Hamatum clade), and one strain to clade 6 (Brevicompactum clade) (Table S1). The strains are conserved in the fungal collection of the plant health laboratory of the Universidad Nacional Toribio Rodríguez de Mendoza de Amazonas (LABISANV).

2.2. Acquisition of Moniliophthora roreri

A fully characterized strain of M. roreri (MBG_01) was used for the assays in this research. This strain was isolated from cacao pods with visible M. roreri mycelia, collected from Quebrada Seca, Utcubamba, Amazonas (UTM: 17M 779480 9368166). This strain was identified, characterized, and conserved in the LABISANV fungal collection.

2.3. Mycoparasitism Test

Mycoparasitism was evaluated by using the method of pre-colonized Petri plates [37]. A five-mm-diameter disc from 12-day-old M. roreri colonies were placed on the periphery of a 90-mm Petri plate containing potato dextrose agar (PDA) medium (Figure 1 and Figure S1). Cultures were incubated in a bioclimatic chamber at 28 °C with a 12 h light/day regime (all incubations were conducted at these conditions unless specified). When M. roreri colonized 100% of the plate (three to five weeks later), a 2.5 × 0.5 cm² Trichoderma inoculum strip, taken from the edge of a three-day-old freshly sporulating colony, was placed on the opposite side from where M. roreri was originally seeded (Figure 1). The cultures were then further incubated for two weeks. Then, ten 5-mm discs were evenly extracted from each plate, including the original M. roreri disc that was originally inoculated. The discs were placed in two Petri plates with 20% PDA culture medium (five discs per plate) and further incubated. We monitored the growth of either Trichoderma or M. roreri for two weeks (Figure 1). Five repetitions of the test per Trichoderma strain were performed. The percentage of mycoparasitism was estimated by using the following formula from Evans et al. [37]:

$$\mathrm{PP}=\left(\mathrm{TG}\times 100\right)/\mathrm{N};$$

where:

PP = Parasitism (%),

TG = Number of discs with Trichoderma growth,

N = Total number of discs taken for each repetition (plate).

Figure 1. Mycoparasitism tests against Moniliophthora roreri. (A) Example of a test with a high mycoparasitism level (Trichoderma longibrachiatum strain F14M3) (100%). A_Left: dual culture at 21 days after inoculation. A_Center and A_Right: discs taken from A_left plate showing Trichoderma growth. (B) Example of a test with a medium mycoparasitism level (T. parareesei strain CRSF3_C2) (50%). B_Left: dual culture at 21 days after inoculation. B_Center: discs with M. roreri growth taken from B_left plate. B_Right: discs taken with Trichoderma growth from B_left plate. (C) Example of a test with low mycoparasitism level (T. longibrachiatum strain F21M5) (0%). C_Left: dual culture at 21 days after inoculation. C_Center and C_Right: discs with M. roreri growth taken from C_left plate.

Figure 1. Mycoparasitism tests against Moniliophthora roreri. (A) Example of a test with a high mycoparasitism level (Trichoderma longibrachiatum strain F14M3) (100%). A_Left: dual culture at 21 days after inoculation. A_Center and A_Right: discs taken from A_left plate showing Trichoderma growth. (B) Example of a test with a medium mycoparasitism level (T. parareesei strain CRSF3_C2) (50%). B_Left: dual culture at 21 days after inoculation. B_Center: discs with M. roreri growth taken from B_left plate. B_Right: discs taken with Trichoderma growth from B_left plate. (C) Example of a test with low mycoparasitism level (T. longibrachiatum strain F21M5) (0%). C_Left: dual culture at 21 days after inoculation. C_Center and C_Right: discs with M. roreri growth taken from C_left plate.

2.4. Antibiosis Test

The antibiosis of Trichoderma strains was evaluated using the paired culture method [38]. Here, 5-mm discs of M. roreri (previously grown on PDA for 7 days) were placed on the periphery of a 90-mm PDA Petri plate and incubated for 10 ± 1 days or until a 1.3–3.3 cm-wide colony was established. From this colony, 5-mm Trichoderma discs from 3–4-day old PDA cultures were collected and transferred on the opposite side of the M. roreri colony previously established (Figure 2). Then, the confrontation plates were incubated. Five replicates were established for each Trichoderma strain, with five controls (non-confronted M. roreri plates). Daily measurements of the M. roreri radial growth were taken, until mycelia from both fungi came into contact in one of the five replicates. Finally, the inhibition of radial growth (i.e., antibiosis) of the phytopathogen was evaluated using the following formula from Holmes et al. [38]:

$$\mathrm{PA}=\left[\left(\mathrm{RG}-\mathrm{RGT}\right)/\mathrm{RG}\right]\times 100;$$

where:

PA = Antibiosis (%),

RG = Radial growth of M. roreri in control plates (mm),

RGT = Radial growth of M. roreri confronted with Trichoderma spp. (mm).

Figure 2. Antibiosis test against Moniliophthora roreri. (A) Example of a test with high antibiosis level (Trichoderma afroharzianum strain F13M4 in a dual culture with M. roreri at 62, 72, and 96 h of confrontation). (B) Example of a test with medium antibiosis level (T. reesei strain UCF7_M2 in a dual culture with M. roreri at 62, 72, and 96 h of confrontation). (C) Example of a test with low antibiosis level (T. afroharzianum strain F12M3 in a dual culture with M. roreri at 62, 72, and 96 h of confrontation).

Figure 2. Antibiosis test against Moniliophthora roreri. (A) Example of a test with high antibiosis level (Trichoderma afroharzianum strain F13M4 in a dual culture with M. roreri at 62, 72, and 96 h of confrontation). (B) Example of a test with medium antibiosis level (T. reesei strain UCF7_M2 in a dual culture with M. roreri at 62, 72, and 96 h of confrontation). (C) Example of a test with low antibiosis level (T. afroharzianum strain F12M3 in a dual culture with M. roreri at 62, 72, and 96 h of confrontation).

2.5. Potential Antagonism

The potential antagonism was calculated by averaging both the mycoparasitism and antibiosis results, as proposed by Reyes-Figueroa et al. [14]:

$$\mathrm{AP}=\left(\mathrm{PP}+\mathrm{PA}\right)/2;$$

where:

AP = Potential antagonism,

PP = Mycoparasitism of Trichoderma against M. roreri (%),

PA= Antibiosis of Trichoderma against M. roreri (%).

2.6. Data Analysis

The mycoparasitism, antibiosis, and potential antagonism levels of strains were analyzed under a completely randomized experimental design using ANOVA and the Scott Knott test for separation of means, when ANOVA detected significant differences, in Infostat V.2020 software [39]. The Scott Knott test allowed us to classify the strains into groups with different levels of mycoparasitism, antibiosis, and potential antagonism (starting from “a”, which grouped strains at the highest levels of these antagonistic features). Additionally, a Kruskal–Wallis test to compare the mycoparasitism, antibiosis, and potential antagonism levels of the strains, according to the phylogenetic clade they belong to, was performed using the R package agricolae [40]. Finally, these data were integrated (Table S3) and used to perform a minimum spanning network analysis with R package poppr [41] to investigate the association of strains according to their potential antagonism.

3. Results and Discussion

3.1. Mycoparasitism

We found highly significant differences ($\mathrm{Scott} \mathrm{Knott}\mathrm{test},p<0.0001$) in the mycoparasitic behavior of Trichoderma strains against M. roreri. The 234 strains were distributed in six groups significantly different to each other (Table S2).

All species had strains with some parasitic behavior against M. roreri. One hundred and forty strains belonging to seventeen species reached parasitism percentages above 50%, and more remarkably, fifty-nine strains reached 100% parasitism against M. roreri (Table S2). These strains belonged to: T. afarasin (1 strain), T. afroharzianum (22 strains), T. asperelloides (1 strain), T. brevicompactum (1 strain), T. ghanense (1 strain), T. inhamatum (4 strains), T. lentiforme (4 strains), T. longibrachiatum (3 strains), T. orientale (4 strains), T. parareesei (5 strains), T. reesei (4 strains), T. zelobreve (1 strain), Trichoderma spp. (ambiguous) (5 strains), and Trichoderma sp. 4 (3 strains) (Table S2). On the other hand, only 25 strains did not have parasitic behavior on M. roreri. These belonged to: T. afroharzianum (6 strains), T. asperelloides (1 strain), T. lentiforme (1 strain), T. longibrachiatum (3 strains), T. parareesei (5 strains), T. reesei (2 strains), Trichoderma sp. 1 (3 strains), and Trichoderma spp. (ambiguous) (4 strains) (Table S2).

In general, the interspecific mycoparasitism levels found in this study were variable, which was also reported in previous studies. Hoyos-Carbajal et al. [42] and Reyes-Figueroa et al. [14] found T. harzianum strains with high and low mycoparasitism levels against Sclerotinia spp. and Rhizoctonia spp. Reyes-Figueroa et al. [14] reported mycoparasitism levels against M. roreri ranging between 10% and 100% for strains of T. harzianum, T. virens, T. spirale, T. brevicompactum, T. koningiopsis, and T. asperellum. Evans et al. [37] reported a 29% colonization level on other cacao pathogen, M. perniciosa, by a T. longibrachiatum strain, and 100% by T. virens and T. spirale strains. Galarza et al. [27] reported mycoparasitism of 100% on various phytopathogenic fungi by T. asperellum strains. Additionally, strains of T. asperelloides, T. lentiforme, and T. reesei have been reported as parasitic on other pathogens, such as Phytophthora palmivora in palm crops [43], and T. koningiopsis, on Macrophomina phaseolina in peanuts [44]. In this study, we report for the first time the parasitic activity against M. roreri of T. parareesei, T. longibrachiatum, T. lentiforme, T. orientale, T. asperelloides, T. inhamatum, T. zelobreve, T. afarasin, T. ghanense, T. rifaii, and T. breve.

When we looked at these results in terms of the phylogenetic clades the strains belong to, we found that the Harzianum clade had strains with higher mycoparasitism levels than the Longibrachiatum and Hamatum clades (Figure 3). Galarza et al. [27] found that T. asperellum (Hamatum clade) and T. harzianum (Harzianum clade) had similar levels of mycoparasitism against M. roreri. They also found that the only strain of the Longibrachiatum clade they worked with (T. virens) had no mycoparasitism and a very low antibiosis capacity against the pathogen. The different mycoparasitic levels of strains within the different Trichoderma clades depend on their enzymatic matrix production capability. This enzymatic matrix includes cellulases, glucanases, lipases, proteases, and chitinases that help them to parasitize M. roreri with greater or lesser aggressiveness [2,14,45]. The production of this enzymatic repertoire may have phylogenetic patterns that future studies must investigate.

3.2. Antibiosis

We also found highly significant differences ($p<0.0001$) in the antibiosis activity of Trichoderma strains against M. roreri (Table S2). All species had at least one strain with antibiosis behavior. Only two strains of T. parareesei and T. afroharzianum did not show antibiosis levels, UJF2-C2 (−1.98% ± 5.23%) and UCF21_M2 (−3.45% ± 2.81%), respectively.

The antibiosis levels of the 234 strains ranged from −3.45% to 31.09%. Sixty strains belonging to: T. afarasin (2 strains), T. afroharzianum (15 strains), T. asperelloides (4 strains), T. brevicompactum (1 strain), T. inhamatum (5 strains), T. lentiforme (10 strains), T. longibrachiatum (8 strains), T. orientale (1 strain), T. parareesei (9 strains), T. reesei (1 strain), T. zelobreve (2 strains), Trichoderma sp. 4 (1 strain), and Trichoderma spp. (ambiguous) (1 strain), had antibiosis levels greater than 20%, while strains F13M4 (T. afroharzianum), UCF5A-C1 (T. longibrachiatum), and F9M3 (T. parareesei) outperformed the others, reaching antibiosis levels greater than 30% (Table S2).

The analysis of antibiosis by phylogenetic clade revealed that the Harzianum clade had significantly higher antibiosis levels than strains in the Longibrachiatum and Hamatum clades (Figure 3). Galarza et al. [27] found that strains of T. asperellum (Hamatum clade) had, for the most part, significantly higher antibiosis levels against M. roreri than strains of T. harzianum (Harzianum clade). Additionally, T. asperellum was found to have higher antibiosis levels than T. harzianum and T. longibrachiatum (Longibrachiatum clade) against other pathogens, such as Fusarium solani [46]. However, these studies have no more than 15 isolates analyzed in total. Our study presents antibiosis information of more than 100 strains for Harzianum and Hamatum clades, providing statistical evidence of their biocontrol capacity against M. roreri (Figure 3). Other cacao pathogens, such as M. perniciosa, were also confronted against potential Trichoderma strains, finding no significant differences between T. asperellum and T. harzianum strains [27]. This may suggest that the antagonistic mode of action of Trichoderma spp. against M. perniciosa may be different than M. roreri.

We identified three strains, BIF7_C1 (T. asperelloides) in the Hamatum clade, and BMF19_C4 (T. parareesei) and CNF3_C1 (T. parareesei) in the Longibrachiatum clade, showing high levels of antibiosis (all in the Scott Knott group “a”, Table S2) without mycoparasitism levels against M. roreri. This situation is not rare as similar findings were reported by Bailey et al. [2] and Reyes-Figueroa [14], in which strains within the Hamatum (T. koningiopsis and T. ovalisporum) and Longibrachiatum (T. longibrachiatum and T. reesei) clades had high levels of antibiosis but low levels of mycoparasitism against M. roreri. Moreover, this situation can extend to strains in the Harzianum clade, as Reyes-Figueroa [14] found that strains of T. pleuroticola had antibiosis but no mycoparasitism activity on M. roreri. The opposite situation is also possible. For example, we found strains F21M2 (T. harzianum) and BIF27_C2 (Trichoderma sp. ambiguous), both in the Harzianum clade, with 96% of mycoparasitism, but 3.94% and 1.71% of antibiosis, respectively. A similar scenario was reported by Bailey et al. [2], in which they found that strains within the Harzianum clade (mostly T. harzianum) had both the highest mycoparasitism levels and the lowest antibiosis levels [2].

3.3. Potential Antagonism

The potential antagonism in terms of phylogenetic clade of strains revealed a similar scenario of significance, in which strains in the Harzianum clade tended to have higher potential antagonism levels than strains in the Longibrachiatum and Hamatum clades (Figure 3). Due to the known intraspecific variability of Trichoderma species, we might expect that the phylogenetic affinities towards the potential antagonism against M. roreri of strains in the Harzianum clade does not apply to all species evaluated. However, when we dissected the data by species, we surprisingly found that T. afarasin, T. breve, T. inhamatum, Trichoderma sp. 2, Trichoderma sp. 3, T. zelobreve, Trichoderma sp. 4, T. lentiforme, and T. afroharzianum, which all belong to the Harzianum clade, had strains with high potential antagonism, mostly above 50% (Figure 4 and Figure S2). Conversely, the species T. ghanense, T. orientale, Trichoderma sp. 1, T. longibrachiatum, T. parareesei, and T. reesei (all in the Longibrachiatum clade), as well as T. asperelloides and T. koningiopsis (both in the Hamatum clade), had strains with potential antagonism mostly below 50% (Figure 4 and Figure S2). This provides another perspective to visualize the data and to corroborate the phylogenetic affinity of the evaluated Trichoderma species in the Harzianum clade to potentially have a higher antagonistic behavior against M. roreri than at least the evaluated species in the Longibrachiatum clade. We cannot draw the same conclusion for species in the Hamatum clade because we only worked with 13 strains in 2 species, which is far less than the number of strains within the other two clades (over 100 each).

Even though our only strain of T. brevicompactum (BBF1_C3) outperformed all other strains, we cannot conclude much about the species and the phylogenetic clade it belongs to (Brevicompactum) (Figure 4). Reyes-Figueroa et al. [14] evaluated four T. brevicompactum strains, finding variable levels of potential antagonism, ranging between 36.6% and 57.0%. Therefore, the phylogenetic affinity of T. brevicompactum in terms of M. roreri biocontrol needs further revision. However, the potential use of the strain BBF1_C3 must be considered in future biocontrol studies.

3.4. Association Analysis

The association analysis considering mycoparasitism, antibiosis, and potential antagonism of the 234 strains also showed that proportionally, strains with higher potential antagonism (Scott Knott group “a”) were composed mostly from strains in the Harzianum clade and were more intimately interconnected in terms of their antagonistic features against M. roreri (Figure 5). Conversely, strains with lower potential antagonism (Scott Knott groups “d–g”) were composed mostly from strains in the Longibrachiatum clade (Figure 5). These results confirmed our findings in terms of the phylogenetic affinities of Trichoderma clades to control M. roreri.

Phylogenetic affinities in Trichoderma spp. regarding their parasitic behavior have been previously reported in human and animal clinical samples [47]. Strains belonging to the Longibrachiatum clade were more commonly isolated from human respiratory tract samples, as opposed to strains in the Harzianum clade, which mainly came from superficial animal tissues [47]. To the best of our knowledge, this is the first study to report a phylogenetic pattern in the Trichoderma genus regarding the antagonistic behavior against the cacao pathogen M. roreri.

The high variability observed on the mycoparasitism and antibiosis between and within species confirms that it is necessary to calculate the potential antagonism (i.e., the average of mycoparasitism and antibiosis) when selecting Trichoderma strains as biocontrol agents [30]. In this study, we combined mycoparasitism and antibiosis, following the procedures of previous studies [13,14,31], to evaluate mycoparasitism and antibiosis simultaneously, which have been used separately in many other studies [2,27]. This allowed us to establish a better approach to select strains with stronger biocontrol potential. Then, we identified several strains, mainly in the Harzianum clade, as the most promising biocontrol agents against M. roreri. Finally, the phylogenetic pattern we found in this study has implications for future studies to test Trichoderma strains in the field, either individually or as a consortium.

4. Conclusions

The results observed for mycoparasitism and antibiosis against M. roreri support the existence of a high intra- and inter-specific variability in Trichoderma. However, despite this variability, we found evidence of a phylogenetic pattern in terms of the antagonistic behavior of Trichoderma strains against M. roreri. The majority of strains in all evaluated species in the Harzianum clade had levels of potential antagonism greater than 50% as opposed to the species in the Longibrachiatum and Hamatum clades, confirming the phylogenetic affinity of the Harzianum clade towards the antagonism against M. roreri. Finally, we demonstrated that the in vitro evaluation of the biocontrol potential of Trichoderma strains, even within the same species, is variable and should be assessed integrating the mycoparasitism and antibiosis levels, in the form of potential antagonism. This process is a necessary step for the selection of biocontrol agents of plant pathogens.