Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Secreted metabolite-mediated interactions between rhizosphere bacteria and Trichoderma biocontrol agents

  • Ningxiao Li,

    Roles Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing

    Affiliation Department of Plant Pathology and Environmental Microbiology, The Pennsylvania State University, University Park, Pennsylvania, United States of America

  • Md Tariqul Islam,

    Roles Formal analysis, Investigation, Methodology, Writing – original draft

    Affiliation Department of Plant Pathology and Environmental Microbiology, The Pennsylvania State University, University Park, Pennsylvania, United States of America

  • Seogchan Kang

    Roles Conceptualization, Funding acquisition, Supervision, Writing – review & editing

    sxk55@psu.edu

    Affiliation Department of Plant Pathology and Environmental Microbiology, The Pennsylvania State University, University Park, Pennsylvania, United States of America

Secreted metabolite-mediated interactions between rhizosphere bacteria and Trichoderma biocontrol agents

  • Ningxiao Li, 
  • Md Tariqul Islam, 
  • Seogchan Kang
PLOS
x

Abstract

Trichoderma has been used as an alternative to synthetic pesticides to control a variety of phytopathogenic fungi, oomycetes, and nematodes. Although its mechanism of pathogen suppression has been extensively studied, how Trichoderma interacts with non-target microbes is not well understood. Here, we investigated how two Trichoderma biological control agents (BCAs) interact with rhizosphere bacteria isolated from a tomato plant via secreted proteins, metabolites, and volatile compounds (VCs). Culture filtrates (CFs) of T. virens and T. harzianum, containing secreted proteins and metabolites, strongly inhibited (>75% reduction in growth) 39 and 19, respectively, out of 47 bacterial strains tested. Their CFs inhibited the remaining strains at lower degrees. Both metabolites and proteins are involved in inhibiting bacteria, but they seem to antagonize each other in inhibiting some strains. Trichoderma and bacteria suppressed the growth of each other using VCs. The secretion of antibacterial and antifungal molecules by T. virens and T. harzianum was significantly affected by VCs from some bacteria, suggesting that both Trichoderma BCAs and rhizosphere bacteria use VCs to influence each other in multiple ways. In light of these results, we discuss how metabolite-mediated interactions can potentially affect the effectiveness of biocontrol.

Introduction

Synthetic pesticides have helped reduce crop loss caused by various diseases and pests. However, heavy reliance on pesticides has caused the rapid emergence of pesticide resistance and has also degraded the environment and vital ecosystem services, underscoring the need for alternative strategies that can sustainably protect crop health [13]. Biocontrol has long been touted as one such strategy. Some members of the bacterial genera Bacillus, Pseudomonas, Stenotrophomonas, and Streptomyces and the fungal genera Ampelomyces, Coniothyrium, and Trichoderma have been tried as biocontrol agents (BCAs) for a variety of plants [47]. Among those, Trichoderma-based BCAs are most commercially successful as more than 60% of the registered biopesticides worldwide contain at least one Trichoderma strain [8,9]. Trichoderma spp. are ubiquitous probably due to their adaptation to a plethora of ecosystems as saprophytes [10]. Some Trichoderma spp. are opportunistic, avirulent plant symbionts and parasites of other fungi [1113]. Trichoderma BCAs control phytopathogenic fungi and oomycetes via parasitism [14], antibiosis [15], competition for space and nutrients [16], and induction of plant systemic resistance [1719]. There exist many reports of successful plant disease management using Trichoderma BCAs in experimental trials and small-scale production systems. For instance, the soil application of T. harzianum T-22 in tomato field trials reduced disease caused by Alternaria solani up to 80% [20].

However, inconsistent efficacy has limited the employment of biocontrol under field conditions [21] due to the following deficiencies in knowledge. First, why biocontrol fails is poorly understood. Without an adequate understanding of the underlying causes for previous failures, biocontrol would continue to be like shooting a target that is obstructed by a series of unpredictably moving obstacles. Second, although commercial biocontrol products often include diverse pathogens on their label, only sparse field efficacy data exist. Information regarding environmental parameters that potentially affect biocontrol efficacy, which is crucial to guide biocontrol application and to predict the likelihood of success under specific conditions, is even sparser. Considering that many plant-associated microbes likely affect plant health [2224], improved understanding of if and how introduced BCAs interact with non-target microbes is crucial to understand the basis of biocontrol success and failure. Trichoderma spp. are well known producers of proteins and secondary metabolites that exhibit antagonistic activities against phytopathogens [25,26]. Such molecules secreted by Trichoderma may negatively affect plants by disrupting the abundance and activity of non-target microbes.

In this study, we used two Trichoderma strains that have been commercially used as BCAs to test if their secreted metabolites, including volatile compounds (VCs), and proteins inhibit the growth of diverse rhizosphere bacteria isolated from a tomato plant. We also investigated if Trichoderma BCAs respond to VCs produced by selected bacteria as part of our on-going efforts to understand the nature and mechanism of chemical ecology underpinning biocontrol [27].

Materials and methods

Isolation of bacteria from tomato rhizosphere

Roots of a single tomato plant, collected at the Penn State Rock Springs Research Farm, were brushed gently and washed under running tap water to remove attached soil particles. The roots placed in a 500 mL flask containing 100 mL sterile MilliQ water were shaken at 150 rpm for 1 hour. After spreading serially diluted suspensions on both 1% Luria-Broth (LB) agar and 1% Tryptone Yeast Extract (TYE) agar, the plates were incubated at 25°C for 2–3 days. One gram of soil attached to the root surface was sampled from five random places and then suspended in sterile MilliQ water. The resulting suspension was diluted and plated as described above. Colony shape, smoothness, and color were used to pick diverse strains, and they were streaked on new plates for purification. Because all of them grew well on LB agar, we used LB for culturing all strains. For long-term storage, all strains in LB with 20% glycerol were kept at -80°C. We also included Escherichia coli BW25113 (obtained from the E. coli Stock Center at Yale University), a strain used for producing the Keio collection of genome-wide mutants [28]. This E. coli strain was included to evaluate if its mutant collection can be employed to study the nature and mechanism of metabolite-mediated interactions between bacteria and Trichoderma.

Sequence-based bacterial identification

Bacterial genomic DNA was extracted and purified as previously described [29]. We checked the quality and quantity of DNA samples using NanoDrop 2000 (Thermo Scientific) and gel electrophoresis. Primers 530F (5′-TGACTGACTGAGTGCCAGCMGCCGCGG-3′) and 1490R (5’-TGACTGACTGAGGYTACCTTGTTACGACTT-3’) were used to amplify the 16S ribosomal RNA (rRNA)-encoding gene. Thermal cycling conditions are initial denaturation for 5 mins at 95°C, 35 cycles of 30 secs at 95°C, 40 secs at 60°C and 40 secs at 72°C, and final extension for 10 mins at 72°C. Amplified DNA was purified using the QIAquick PCR purification kit (QIAGEN, Germantown, MD) and sent to the Penn State Genomics Core for sequencing. Sequences of both strands were manually checked for errors and then assembled using CAP3 [30]. Assembled sequences were used as queries to search the non-redundant (NR) small subunit rRNA reference dataset in the SILVA database [31]. The database provides a manually curated taxonomy for all three domains of life and helps identify bacteria [32]. A phylogenetic analysis using 16S rRNA sequences was performed using the maximum likelihood method (bootstrap value at 1000) in Mega7 [33]. We deposited their 16S rRNA sequences to GenBank (accession numbers from MK591839 to MK591885; see Table 1).

Measurement of bacterial growth in Trichoderma culture filtrates

Trichoderma strains used in this study, including T. virens G-41 and T. harzianum T-22, were isolated from commercial biocontrol products [27]. After inoculating a plug from each Trichoderma culture on potato dextrose agar (PDA) in 500 mL 1:1 mixture of potato dextrose broth (PDB) and LB, culture flasks were shaken (180 rpm) for 7 days at 25°C. Individual Trichoderma cultures were filtered sequentially through cheesecloth, Whatman filter paper, and 0.2 μm nitrocellulose filter to prepare culture filtrates (CFs). We inoculated a single bacterial colony in 2 mL LB and cultured by shaking (200 rpm) at 25°C until OD600 reached ~1 to prepare the inoculum used to assess how resulting CFs affect bacterial growth. We inoculated 10 μL bacterial culture into 2 mL Trichoderma CF and incubated them by shaking (200 rpm) at 25°C. For control, 10 μL bacterial culture was inoculated into fresh PDB+LB (1:1). After 1–2 days of culturing, OD600 was measured. We also measured bacterial growth in diluted CFs (1 mL CF+1 mL PDB+LB and 0.5 mL CF+1.5 mL PDB+LB) to determine whether bacterial growth inhibition was due to the depletion of nutrients. Growth inhibition was calculated using the following formula: , where Xcontrol and Xtreatment indicate OD600 or colony-forming units (when bacterial growth was measured by plating cultures on an agar medium) in control and treatment, respectively. Each treatment included three biological replicates and was repeated three times.

Determination of how secreted proteins from Trichoderma affect bacterial growth

Secreted metabolites in Trichoderma CFs were removed via dialysis (Spectra/Por® 3 dialysis tube with the molecular weight cut off value of 3.5 kD, Spectrum Laboratories, Rancho Dominguez, CA). After rinsing the dialysis tube with distilled water, it was autoclaved in MilliQ water at 121°C for 15 minutes. Each dialysis tube containing 70 mL CF was placed in 1 L beaker containing 930 mL PDB+LB (1:1) medium. After stirring at 4°C for 5–6 hours, dialysis was repeated twice to obtain secreted metabolite-free CF (labeled as–Met). The growth of 20 bacterial strains, including 19 noted in Table 1 and E. coli, was evaluated by inoculating 10 μL bacterial culture into 2 mL fresh PDB+LB (control) and–Met. After 1–2 days of culturing, the degree of growth inhibition was measured as described above. We used Proteinase K to confirm that some proteins in–Met of T. harzianum exhibit antibacterial activity. Proteinase K solutions in three concentrations (100, 300, and 500 μg/mL) were prepared by mixing 200, 600, and 1000 μg of Proteinase K (Promega, Madison, WI) with 2 mL CF,–Met, and control (fresh PDB+LB). After incubating at 37°C for 2 hours, LR1 and LR3 were cultured in Proteinase K-treated media by shaking (200 rpm) at 25°C for one day. OD600 was measured to calculate the degree of growth inhibition. Each treatment consisted of three biological replicates and was repeated three times.

Determination of how secreted molecules from Trichoderma affect bacterial growth using an agar medium

This experiment was conducted using 20 bacterial strains (19 noted in Table 1 and E. coli). One plug of Trichoderma culture was inoculated on a sterilized cellophane membrane (Paper Mart, Orange, CA) overlaid on PDA+LB (1:1) agar and incubated at 25°C until it covered three-quarters of the membrane. After removing the membrane along with Trichoderma culture, we plated ~500 bacterial cells. Fresh PDA+LB (1:1) agar was used as the control for this experiment. After 2–3 days of incubation at 25°C, the number of colonies was counted. Each treatment was performed in three replicates and was repeated three times.

We evaluated whether the cellophane membrane used is permeable to secreted proteins by applying drops of Precision Plus Protein Standards (Bio-Rad, Hercules, CA) on a cellophane membrane overlaid on 1.5% water agar. This mixture contains pre-stained proteins (covalently linked with a blue dye) with M.W. ranging from 10 kD to 250 kD. First, we removed glycerol, SDS, and other compounds in the mixture using Microsep® Advance Centrifugal Devices (the molecular weight cut-off of 1 kD, Pall Corporation, NY) followed by concentration using DNA120 Speed Vac® (ThermoSavant). After applying two 5 μL drops of the concentrated protein solution on 1.5% water agar and a cellophane membrane overlaid on the agar, the plate was left overnight at room temperature. We photographed the surface after removing the cellophane membrane, as well as the membrane used.

Volatile compound (VC)-mediated inhibition between Trichoderma and bacteria

We used the sandwiched plate assay (S1 Fig) previously developed [27] to determine whether VCs produced by Trichoderma affect the growth of bacterial strains. One plug from Trichoderma culture was inoculated on 0.75X PDA at 25°C for two days. After spreading ~500 bacterial cells on LB agar plate, each plate was placed on top of a Trichoderma plate and sealed with three layers of parafilm. Sandwiching a bacterial plate with an uninoculated 0.75X PDA plate served as the control for this experiment. After incubation at 25°C for 2–3 days, the number of colonies was counted to calculate the degree of growth inhibition. We also evaluated the inhibitory activity of bacterial VCs on Trichoderma using LR1, TS6, TS9 and E. coli. Approximately 1x103 cells of these strains were plated on LB agar and incubated for two days before sandwiching them with plates inoculated with a plug from each Trichoderma culture. The control used consisted of each Trichoderma plate sandwiched with an uninoculated LB plate. The colony diameter of Trichoderma was measured 1.5 days after co-cultivation. Each treatment included three replicates and was repeated three times.

Effect of bacterial VCs on the secretion of antibacterial and antifungal metabolites from Trichoderma

We determined whether bacterial VCs affect Trichoderma’s secretion of antibacterial or antifungal metabolites using the sandwiched plate assay described above with the following modifications. Approximately 1x103 cells of LR1, TS6, TS9 and E. coli were plated on LB agar and incubated for two days. After inoculating a plug from each Trichoderma culture on a sterilized cellophane membrane overlaid on PDA+LB (1:1; for testing antibacterial activity) or 0.75X PDA (for testing antifungal activity), each bacterial plate was sandwiched with a Trichoderma culture plate and sealed with three layers of parafilm. The control treatment consisted of each Trichoderma plate sandwiched with an un-inoculated LB plate. After 33 hours (for T. harzianum) or 40 hours (for T. virens) of co-cultivation when their colonies were ~8 mm away from the edge of the overlaid membrane, the membrane along with Trichoderma culture was removed. The co-cultivation time was different because T. harzianum grew faster than T. virens. We added 5 mL LB to each plate and shook the plate (70 rpm) at room temperature for three hours to allow secreted metabolites to diffuse into LB. We added 10 μL LS1 culture into 1 mL LB containing secreted metabolites and incubated overnight before measuring OD600. LS1 cells grown in fresh LB were used as the control in calculating the degree of growth inhibition. To measure the antifungal activity of secreted metabolites, a plug of Fusarium oxysporum f.sp. lycopersici NRRL54003 culture (obtained from the ARS Culture Collection at the National Center for Agricultural Utilization Research) was inoculated on each plate. Its colony diameter was measured after three days of incubation on the plates used for culturing T. harzianum and five days for those used for culturing T. virens. The difference in incubation time was due to stronger inhibition of F. oxysporum by the latter, which required more than three days of incubation to detect a measurable growth. Each treatment was performed in three replicates and was repeated three times.

Statistical analysis

We randomly placed culture tubes and Petri plates in incubators. One-way analysis of variation (ANOVA) was performed using Minitab 18 (Minitab Inc., State College, PA). The significance of each treatment was determined using the F value. When a significant F test was observed, the separation of the means was carried out using Tukey’s test. Statistical significance was determined at P≤0.05.

Results

Proteins and metabolites secreted by two Trichoderma BCAs strongly inhibit the growth of diverse bacteria isolated from the tomato rhizosphere

Forty-seven rhizosphere bacterial strains, including 22 isolated from the root surface of a tomato plant and 25 isolated from surrounding soils, were identified using 16S rDNA sequences (Table 1), and their phylogenetic positions are shown in Fig 1.

thumbnail
Fig 1. Bacterial growth responses to metabolites and proteins secreted by Trichoderma.

Tv and Th correspond to T. virens and T. harzianum, respectively. Growth of diverse bacteria (Table 1) and E. coli under the following treatments was measured: (A) CF (Trichoderma culture filtrate); (B)–Met (CF after removing secreted metabolites via dialysis); (C) AM (agar medium after removing Trichoderma cultured on cellophane membrane); and (D) VC (exposure to Trichoderma volatile compounds). OD600 (Treatments A and B) or colony number (Treatments C and D) was measured in triplicates to calculate the degree of growth inhibition/enhancement after each treatment (see S1 Table for the exact degree of growth inhibition or enhancement by each treatment). Results were color-coded to provide an overview of their growth under the four different culture conditions. The phylogenetic tree was generated using their 16S rRNA sequences.

https://doi.org/10.1371/journal.pone.0227228.g001

When these strains and E. coli were grown in culture filtrates (CFs) of T. virens and T. harzianum, their growth was inhibited at varying degrees (Fig 1 and S1 Table). We also counted cell numbers by plating diluted bacterial cultures in CFs on LB agar to determine whether the measurement of OD600 accurately indicates the degree of growth inhibition. Results from both methods were comparable (S2 Fig). The degree of inhibition by T. virens CF was higher than that by T. harzianum CF for all strains except TS10. For example, T. virens CF caused 96% and 95% inhibition of LR1 and LR3, respectively, while T. harzianum CF inhibited them by 25% and 27%, respectively (S1 Table). The CF of T. virens inhibited 40 strains at the level of 75% or higher, while T. harzianum CF inhibited only 19 strains at the same level. The CFs of T. virens and T. harzianum suppressed the growth of the remaining strains by 57–74% and 12–74%, respectively (S1 Table).

We cultured 19 strains (noted in Table 1) and E. coli in diluted CFs to test if the growth inhibition was simply due to nutrient depletion caused by prior culturing of Trichoderma. The degrees of growth inhibition of TS4, TS15, LR2, LR1, LR3, LR20, LS7 and LR21 in diluted CFs of T. virens were comparable to those observed in undiluted CF, while the growth of other strains was improved in diluted CFs (S1 Table). Most strains grew better in diluted CFs of T. harzianum with two exceptions. Strain TS4 was similarly inhibited in all three CFs (S1 Table). Surprisingly, diluted T. harzianum CFs inhibited the growth of LR1, LR3 and E. coli more strongly than undiluted CFs (Fig 2 and S1 Table).

thumbnail
Fig 2. Growth of three bacteria in diluted CFs of T. harzianum.

LR1, LR3 and E. coli were cultured in undiluted and diluted (1:1 and 1:3) CFs. Growth inhibition (%) shown corresponds to the mean ± SE of data from three replicates. Different letters indicate a significant difference between treatments according to Tukey’s test at P≤0.05.

https://doi.org/10.1371/journal.pone.0227228.g002

Both metabolites and proteins contribute at varying degrees to inhibiting bacteria

We investigated which molecules (metabolites, proteins, or both) inhibit individual bacteria. We removed secreted metabolites in CFs via dialysis (labeled as–Met) to assess the role of secreted proteins (Fig 1 and S1 Table). Although–Met derived from both CFs still inhibited most strains by 10% or higher, the degree of inhibition was much lower than that caused by the CFs without dialysis. Interestingly, E. coli grew better in–Met of both CFs compared to control (fresh medium). Similarly, TS9 also grew better in–Met of T. harzianum than control. The degree of inhibition of TS4, LR2 and LR1 by–Met of T. virens was similar to that caused by its CF, and–Met and CF of T. harzianum similarly inhibited TS4, LS11 and LR8 (Fig 1 and S1 Table). T. harzianum–Met more strongly inhibited LR1 and LR3 than its CF. The antibacterial activity of secreted proteins was further confirmed by treating–Met with Proteinase K. The inhibition of LR1 and LR3 by–Met of T. harzianum decreased as the concentration of Proteinase K increased (Fig 3). However, after T. harzianum CF was treated with Proteinase K, it exhibited even higher antibacterial activity (Fig 3).

thumbnail
Fig 3. Proteinase K treatment to evaluate the contribution of secreted proteins in inhibiting bacterial growth.

Degrees of growth inhibition of (A) LR1 and (B) LR3 by CF and–Met of T. harzianum after Proteinase K treatment are shown. Values shown correspond to the mean ± SE of data from three replicates.

https://doi.org/10.1371/journal.pone.0227228.g003

We also measured the antibacterial activity of secreted molecules from Trichoderma on an agar medium (Fig 1 and S1 Table). To determine whether the cellophane membrane used for this measurement blocks proteins from reaching medium, we evaluated its permeability to a mixture of proteins ranging from 10 kD to 250 kD (S3 Fig). The stained proteins did not appear going through the cellophane membrane (S3 Fig). All 20 bacterial strains tested could not grow on T. virens-treated plates (Fig 1 and S1 Table). However, on the plates used for culturing T. harzianum, nine strains failed to growth, and the other strains were inhibited to lesser degrees (6–74% reduction, S1 Table).

VCs produced by Trichoderma and bacteria inhibit each other

We determined the effect of VCs released by Trichoderma on bacterial growth (S1 Fig). VCs produced by T. virens and T. harzianum prevented the growth of 10 and 11 strains, respectively (Fig 1 and S1 Table). However, LS11, LR8, TS9, TS1 and E. coli were minimally affected (10% or less inhibition). LR13 and TS14 were strongly inhibited (100% and 98%, respectively) by T. harzianum VCs, whereas T. virens VCs did not significantly affect their growth (S1 Table). The remaining strains were inhibited at varying degrees (25% to 73%). We also tested whether bacterial VCs inhibit Trichoderma using LR1, TS6, TS9 and E. coli. VCs produced by these strains inhibited both Trichoderma spp. (Fig 4), with the degree of inhibition being 18–27% for T. virens and 17–25% for T. harzianum. VCs from TS9 exhibited the highest inhibitory effect on both Trichoderma spp.

thumbnail
Fig 4. Inhibition of Trichoderma by bacterial VCs.

Colony diameters of (A) T. virens and (B) T. harzianum after co-cultivation with LR1, TS6, TS9 and E. coli as well as un-inoculated LB plates (Control) are shown. Values shown correspond to the mean ± SE of data from three replicates. Different letters indicate a significant difference between treatments according to Tukey’s test at P≤0.05.

https://doi.org/10.1371/journal.pone.0227228.g004

VCs of some bacteria induced the secretion of antibacterial molecules by T. virens but suppressed the secretion of antifungal molecules by both Trichoderma spp.

Our previous study showed that T. virens increased the secretion of antifungal molecules in response to VCs produced by diverse F. oxysporum strains, while T. harzianum responded to VCs from only a few F. oxysporum strains [27]. Here, we determined whether bacterial VCs similarly affect the secretion of antibacterial/antifungal molecules by T. virens and T. harzianum. The extracts from T. virens-cultured medium inhibited LS1 more strongly than control extract (Fig 5A), indicating that VCs produced by LR1 and TS6 induced the secretion of antibacterial molecules. However, VCs from TS9 and E. coli did not cause any noticeable changes. VCs produced by all four strains did not affect the secretion of antibacterial molecules by T. harzianum (Fig 5B).

thumbnail
Fig 5. Effect of bacterial VCs on the secretion of antibacterial molecules from Trichoderma.

Degrees of growth inhibition (%) of LS1 in the extracts derived from the plates used for co-culturing (A) T. virens and (B) T. harzianum with LR1, TS6, TS9 and E. coli as well as un-inoculated LB plates (Control) are shown. Values shown correspond to the mean ± SE of data from three replicates. Different letters indicate significant differences between treatments based on Tukey’s test at P≤0.05.

https://doi.org/10.1371/journal.pone.0227228.g005

Bacterial VCs also affected the secretion of antifungal molecules by both T. virens and T. harzianum (Fig 6). Colony diameter of F. oxysporum on the medium used for treating T. virens with E. coli VCs was twice as big as that on control plates, indicating decreased secretion of antifungal molecules in response to E. coli VCs. However, VCs from the other three strains did not significantly affect the secretion of antifungal molecules (Fig 6A). VCs from all four strains decreased the secretion of antifungal molecules from T. harzianum (Fig 6B). Bacterial VCs also affected the amount of a yellow metabolite secreted by T. harzianum (S4 Fig). The effect varied depending on the fungal culture medium used. When PDA was used, VCs from all strains suppressed the secretion of this metabolite (S4A Fig). However, when T. harzianum was cultured on PDA+LB medium, VCs from TS6 increased its secretion, but the effect of VCs from the other strains did not look noticeably different from control (S4B Fig).

thumbnail
Fig 6. Effect of bacterial VCs on the secretion of antifungal molecules from Trichoderma.

Colony diameters of F. oxysporum NRRL54003 on plates used for co-culturing (A) T. virens and (B) T. harzianum with LR1, TS6, TS9 and E. coli as well as un-inoculated LB plates (Control) are shown. Values shown correspond to the mean ± SE of data from three replicates. Different letters indicate significant differences between treatments based on Tukey’s test at P≤0.05.

https://doi.org/10.1371/journal.pone.0227228.g006

Discussion

Growth of all 48 bacterial strains, including 47 rhizosphere bacteria and E. coli, was inhibited by CFs of T. virens and T. harzianum (Fig 1 and S1 Table). The degree of inhibition varied among them and did not follow their phylogenetic relationship. However, the CF of T. virens inhibited all strains, except TS10, more strongly than that of T. harzianum (Fig 1 and S1 Table), indicating that T. virens secretes more or stronger antibacterial molecules than T. harzianum. Similarly, metabolites produced by another T. virens strain more strongly inhibited Phytophthora erythroseptica than those from T. harzianum [34]. Bacterial growth in diluted CFs (Figs 1 and 2) suggested that growth inhibition was not merely due to the depletion of some essential nutrients by Trichoderma. Metabolites secreted by T. koningii also inhibited the growth of soil bacteria [35], suggesting that Trichoderma BCAs likely inhibit many neighboring bacteria while controlling pathogens. An earlier study [36] reported that peptaibols produced by T. harzianum, such as trichorzianines A1 and B1, act synergistically with cell wall degrading enzymes in inhibiting fungal pathogens. However, our results suggested antagonistic interactions between secreted proteins and metabolites in inhibiting some bacteria. Diluted CFs of T. harzianum inhibited LR1, LR3 and E. coli more strongly than undiluted CF (Fig 2). Dialysis and Protease K treatment of Trichoderma CFs (Figs 1 and 3) indicated that some metabolite(s) might antagonize the antibacterial activity of secreted proteins. Proteinase K-treated CF exhibited higher antibacterial activity than untreated CF (Fig 3), suggesting that some protein(s) may dampen the effect of antibacterial metabolites. Identification of the Trichoderma metabolites and proteins involved in inhibiting the growth of rhizosphere bacteria is needed to understand the mechanism underlying bacterial growth inhibition and antagonistic interaction between secreted proteins and metabolites.

The ability of VCs to move through the air and porous soils enables them to participate in both short- and long-distance organismal interactions within and across kingdoms [3739]. Although water may not be readily available in many environments, research on the nature and mechanism of organismal interactions has mostly focused on secreted molecules that require water as a medium for function. Recent studies suggested multiple roles of VCs in Trichoderma’s interaction with plants and other fungi [27,40,41]. Some Trichoderma species produce antifungal VCs [27,42,43], suggesting their involvement in suppressing fungal pathogens. However, how VCs affect interactions between Trichoderma and bacteria is poorly understood, which is why we chose to investigate whether Trichoderma VCs affect the growth of diverse bacteria and whether bacterial VCs affect Trichoderma BCAs. The VCs produced by T. virens and T. harzianum strongly inhibited most of the tested bacteria (Fig 1 and S1 Table) and diverse F. oxysporum isolates [27]. These results suggest that some VCs released by Trichoderma BCAs function as a fumigant, suppressing pathogens and other microbes, and likely modify rhizosphere microbial communities. Both Trichoderma strains used in our study produce various volatile alcohols, acids, esters, ketones, and sesquiterpenes [27], some of which are known antimicrobial compounds. Bacteria are also known to produce diverse VCs, some of which exhibit antibacterial and antifungal activities [4448]. For example, many strains from the genus of Pseudomonas have been shown to produce VCs inhibitory to fungi and bacteria, such as hydrogen sulfide, 2-phenylethanol, and nonanal [49]. A recent study showed that VOCs produced by Bacillus amyloliquefaciens altered the soil microbial community [50]. In this study, we found that VCs produced by all four strains tested inhibited the growth of Trichoderma (Fig 4), suggesting that more strains in the rhizosphere likely produce antifungal VCs and may affect the biocontrol activity of Trichoderma in the soil.

Our previous study on VC-mediated interactions between four Trichoderma BCAs and F. oxysporum [27] showed that besides the role of VCs as a chemical weapon, some Trichoderma BCAs recognized the presence of F. oxysporum by sensing specific F. oxysporum VCs as cues and increased the secretion of antifungal metabolites. F. oxysporum also recognized and similarly responded to VCs released by Trichoderma BCAs, suggesting that VC-mediated recognition of other microbes may be a commonly used mechanism among fungi. This study led us to check whether the secretion of antibacterial and antifungal molecules produced by T. virens and T. harzianum could be affected by bacterial VCs (Figs 5 and 6). VCs produced by LR1 and TS6 significantly increased the secretion of antibacterial molecules by T. virens, whereas VCs produced by TS9 and E. coli did not. The secretion of antibacterial molecules by T. harzianum was not affected by VCs from all strains. In contrast, VCs from all four bacterial strains significantly suppressed the secretion of antifungal molecules by T. harzianum, whereas only E. coli VCs suppressed their secretion by T. virens. As shown in S4 Fig, VCs from all strains also affected the secretion of a yellow metabolite by T. harzianum, and the effect varied depending on media used for culturing T. harzianum. The secretion of this yellow metabolite was also suppressed or induced by F. oxysporum VCs depending on strains used [27]. Although we did not identify this metabolite, we think that it is one of the chromogenic secondary metabolites called anthraquinones, which are produced by several Trichoderma spp. including T. harzianum [51]. Some anthraquinones exhibit antimicrobial activities [52]. These findings suggest that specific VCs produced by some bacteria may manipulate T. virens and T. harzianum. Similar VC-mediated interactions between bacteria and fungi were reported in several recent studies [5355]. For instance, VCs produced by Paenibacillus polymyxa, a bacterial BCA, and Verticillium longisporum, a soilborne fungal pathogen, affected the production of antimicrobial VCs and other metabolites in the other side [53]. VCs produced by Aspergillus flavus and Ralstonia solanacearum, soilborne fungal and bacterial pathogens, respectively, also affected each other [55]. Future study is needed to identify which bacterial VCs have an antibiotic or signal function and if such bacterial VCs affect the outcome of biocontrol via the use of Trichoderma.

Conclusion

Metabolites, including VCs, and proteins secreted by two Trichoderma BCAs strongly inhibited the growth of diverse bacteria isolated from the tomato rhizosphere. This observation raised the questions of whether these Trichoderma BCAs significantly modify rhizosphere bacterial communities during biocontrol and if resulting changes affect plant health and the outcome of biocontrol. Strong inhibition of many bacteria by Trichoderma VCs suggests the potential role of VCs as soil fumigants. Additionally, VCs produced by some bacterial strains affected the growth of Trichoderma BCAs and their secretion of antifungal/antibacterial metabolites. Multiple functions of VCs in microbial interactions suggest that chemical ecology may play crucial roles in biocontrol, underscoring the need for systematically exploring the nature and mechanism of VC-mediated inter-kingdom microbial interactions.

Supporting information

S1 Fig. Sandwiched plate assay used for measuring the antibacterial activity of VCs produced by Trichoderma.

Each plate of Trichoderma culture (bottom) was sandwiched with LB agar plate inoculated with bacterial cells (top) and incubated at 25°C for two days. Bacterial growth after control treatment (bacterial plate sandwiched with PDA plate without Trichoderma) is shown.

https://doi.org/10.1371/journal.pone.0227228.s001

(DOCX)

S2 Fig. Comparison of two methods used to measure the degree of growth inhibition by T. virens CF.

After culturing LR1 and LR3 in control (PDB+LB) and CF of T. virens for one day, the degree of growth inhibition was determined by (A) measuring OD600 and (B) spreading diluted bacterial cultures on LB agar. (C) Results from A and B are shown. Values shown correspond to the mean ± SE of data from three replicates. No statistically significant difference was observed between the methods according to Tukey’s test at P≤0.05.

https://doi.org/10.1371/journal.pone.0227228.s002

(DOCX)

S3 Fig. Evaluation of the protein permeability of the cellophane membrane used.

Two 5 μL drops of a pre-stained protein solution were applied on water agar (left side) and cellophane membrane (right side) overlaid on water agar. After overnight incubation at room temperature, the membrane was removed, and the plate was photographed (A). The cellophane membrane used is shown (B).

https://doi.org/10.1371/journal.pone.0227228.s003

(DOCX)

S4 Fig. Effect of bacterial VCs on the secretion of a yellow metabolite by T. harzianum.

Plates of T. harzianum inoculated on cellophane membrane overlaid on (A) PDA and (B) PDA+LB (1:1) were sandwiched with plates of LR1, TS6, TS9, and E. coli as well as un-inoculated LB agar plate (Control) for 33 h. After removing the cellophane membrane along with T. harzianum culture, they were photographed.

https://doi.org/10.1371/journal.pone.0227228.s004

(DOCX)

S1 Table. Growth inhibition by different types of the molecules secreted by T. virens and T. harzianum.

https://doi.org/10.1371/journal.pone.0227228.s005

(DOCX)

Acknowledgments

We thank Rhea Lumactud for critically reviewing the manuscript.

References

  1. 1. National Research Council. Toward Sustainable Agricultural Systems in the 21st Century. Washington, D.C.: National Academies Press; 2010 Jun. https://doi.org/10.17226/12832
  2. 2. Reid A, Greene SE. How microbes can help feed the world. American Society for Micriobiology; 2012. https://doi.org/10.1093/biosci/biu163
  3. 3. Velivelli SLS, De Vos P, Kromann P, Declerck S, Prestwich BD. Biological control agents: from field to market, problems, and challenges. Trends Biotechnol. 2014;32: 493–496. pmid:25246168
  4. 4. Berg G. Plant-microbe interactions promoting plant growth and health: perspectives for controlled use of microorganisms in agriculture. Appl Microbiol Biotechnol. 2009;84: 11–8. pmid:19568745
  5. 5. Calistru C, McLean M, Berjak P. In vitro studies on the potential for biological control of Aspergillus flavus and Fusarium moniliforme by Trichoderma species. A study of the production of extracellular metabolites by Trichoderma species. Mycopathologia. 1997;137: 115–124. pmid:16284721
  6. 6. Ejechi B. Biological control of wood decay in an open tropical environment with Penicillium sp. and Trichoderma viride. Int Biodeterior Biodegradation. 1997;39: 295–299.
  7. 7. Yedidia I, Benhamou N, Chet I. Induction of defense responses in cucumber plants (Cucumis sativus L.) by the biocontrol agent Trichoderma harzianum. Appl Environ Microbiol. 1999;65: 1061–1070. pmid:10049864
  8. 8. Mukherjee PK, Horwitz BA, Herrera-Estrella A, Schmoll M, Kenerley CM. Trichoderma research in the genome era. Annu Rev Phytopathol. 2013;51: 105–29. pmid:23915132
  9. 9. Singh A, Shukla N, Kabadwal BC, Tewari AK, Kumar J. Review on plant-Trichoderma-pathogen interaction. Int J Curr Microbiol Appl Sci. 2018;7: 2382–2397.
  10. 10. Guzmán-Guzmán P, Porras-Troncoso MD, Olmedo-Monfil V, Herrera-Estrella A. Trichoderma species: versatile plant symbionts. Phytopathology. 2019;109: 6–16. pmid:30412012
  11. 11. Monte E. Understanding Trichoderma: between biotechnology and microbial ecology. Int Microbiol. 2001;4: 1–4. pmid:11770814
  12. 12. Samuels GJ. Trichoderma: systematics, the sexual state, and ecology. Phytopathology. 2006;96: 195–206. pmid:18943925
  13. 13. Chaverri P, Samuels GJ. Evolution of habitat preference and nutrition mode in a cosmopolitan fungal genus with evidence of interkingdom host jumps and major shifts in ecology. Evolution. 2013;67: 2823–2837. pmid:24094336
  14. 14. Adams PB. The potential of mycoparasites for biological control of plant diseases. Annu Rev Phytopathol. 1990;28: 59–72. pmid:20540607
  15. 15. Ghisalberti EL, Sivasithamparam K. Antifungal antibiotics produced by Trichoderma spp. Soil Biol Biochem. 1991;23: 1011–1020.
  16. 16. Harman GE. Overview of mechanisms and uses of Trichoderma spp. Phytopathology. 2006;96: 190–194. pmid:18943924
  17. 17. Harman GE, Howell CR, Viterbo A, Chet I, Lorito M. Trichoderma species—Opportunistic, avirulent plant symbionts. Nat Rev Microbiol. 2004;2: 43–56. pmid:15035008
  18. 18. Koike N, Hyakumachi M, Kageyama K, Tsuyumu S, Doke N. Induction of systemic resistance in cucumber against several diseases by plant growth-promoting fungi: lignification and superoxide generation. Eur J Plant Pathol. 2001;107: 523–533.
  19. 19. Yedidia I, Shoresh M, Kerem Z, Benhamou N, Kapulnik Y, Chet I. Concomitant induction of systemic resistance to Pseudomonas syringae pv. lachrymans in cucumber by Trichoderma asperellum (T-203) and accumulation of phytoalexins. Appl Environ Microbiol. 2003;69: 7343–7353. pmid:14660384
  20. 20. Seaman A. Efficacy of OMRI-Approved Products for Tomato Foliar Disease Control. New York State Integr Pest Manag Progr Publ. 2003.
  21. 21. Mazzola M, Freilich S. Prospects for biological soilborne disease control: application of indigenous versus synthetic microbiomes. Phytopathology. 2017;107: 256–263. pmid:27898265
  22. 22. Mendes R, Garbeva P, Raaijmakers JM. The rhizosphere microbiome: Significance of plant beneficial, plant pathogenic, and human pathogenic microorganisms. FEMS Microbiol Rev. 2013;37: 634–663. pmid:23790204
  23. 23. Raaijmakers JM, Paulitz TC, Steinberg C, Alabouvette C, Moënne-Loccoz Y. The rhizosphere: a playground and battlefield for soilborne pathogens and beneficial microorganisms. Plant Soil. 2009;321: 341–361.
  24. 24. Saito M. Symbiotic exchange of nutrients in arbuscular mycorrhizas: transport and transfer of phosphorus. Arbuscular Mycorrhizas: Physiology and Function. Dordrecht: Springer Netherlands; 2000. pp. 85–106. https://doi.org/10.1007/978-94-017-0776-3_5
  25. 25. Mendoza-Mendoza A, Zaid R, Lawry R, Hermosa R, Monte E, Horwitz BA, et al. Molecular dialogues between Trichoderma and roots: Role of the fungal secretome. Fungal Biol Rev. 2018;32: 62–85.
  26. 26. Zeilinger S, Gruber S, Bansal R, Mukherjee PK. Secondary metabolism in Trichoderma–chemistry meets genomics. Fungal Biol Rev. 2016;30: 74–90.
  27. 27. Li N, Alfiky A, Wang W, Islam M, Nourollahi K, Liu X, et al. Volatile compound-mediated recognition and inhibition between Trichoderma biocontrol agents and Fusarium oxysporum. Front Microbiol. 2018;9: 1–16.
  28. 28. Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, et al. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol. 2006;2: 2006.0008. pmid:16738554
  29. 29. Wilson K. Preparation of genomic DNA from bacteria. Curr Protoc Mol Biol. 2001;56: 2.4.1–2.4.5. pmid:18265184
  30. 30. Huang XQ, Madan A. SymBioSys\rCAP3: A DNA sequence assembly program. Genome Res. 1999;9: 868–877. pmid:10508846
  31. 31. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2012;41: D590–D596. pmid:23193283
  32. 32. Yilmaz P, Parfrey LW, Yarza P, Gerken J, Pruesse E, Quast C, et al. The SILVA and “All-species Living Tree Project (LTP)” taxonomic frameworks. Nucleic Acids Res. 2014;42: D643–D648. pmid:24293649
  33. 33. Kumar S, Stecher G, Tamura K. MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol Biol Evol. 2016;33: 1870–1874. pmid:27004904
  34. 34. Etebarian HR, Scott ES, Wicks TJ. Trichoderma harzianum T39 and T. virens DAR 74290 as potential biological control agents for Phytophthora erythroseptica. Eur J Plant Pathol. 2000;106: 329–337.
  35. 35. Simon A, Sivasithamparam K. Interactions among Gaeumannomyces graminis var. tritici, Trichoderma koningii, and soil bacteria. Can J Microbiol. 1988;34: 871–876.
  36. 36. Schirmbock M, Lorito M, Wang YL, Hayes CK, Arisan-Atac I, Scala F, et al. Parallel formation and synergism of hydrolytic enzymes and peptaibol antibiotics, molecular mechanisms involved in the antagonistic action of Trichoderma harzianum against phytopathogenic fungi. Appl Environ Microbiol. 1994;60: 4364–4370. pmid:7811076
  37. 37. Bitas V, Kim H-S, Bennett JW, Kang S. Sniffing on microbes: diverse roles of microbial volatile organic compounds in plant health. Mol Plant Microbe Interact. 2013;26: 835–843. pmid:23581824
  38. 38. Li N, Alfiky A, Vaughan MM, Kang S. Stop and smell the fungi: fungal volatile metabolites are overlooked signals involved in fungal interaction with plants. Fungal Biol Rev. 2016;30: 134–144.
  39. 39. Schmidt R, Cordovez V, de Boer W, Raaijmakers J, Garbeva P. Volatile affairs in microbial interactions. ISME J. 2015;9: 1–7.
  40. 40. Lee S, Yap M, Behringer G, Hung R, Bennett JW. Volatile organic compounds emitted by Trichoderma species mediate plant growth. Fungal Biol Biotechnol. 2016;3: 7. pmid:28955466
  41. 41. Martínez-Medina A, Van Wees SCM, Pieterse CMJ. Airborne signals from Trichoderma fungi stimulate iron uptake responses in roots resulting in priming of jasmonic acid-dependent defences in shoots of Arabidopsis thaliana and Solanum lycopersicum. Plant Cell Environ. 2017;40: 2691–2705. pmid:28667819
  42. 42. Wheatley R, Hackett C, Bruce A, Kundzewicz A. Effect of substrate composition on production of volatile organic compounds from Trichoderma spp. inhibitory to wood decay fungi. Int Biodeterior Biodegrad. 1997;39: 199–205.
  43. 43. Meena M, Swapnil P, Zehra A, Dubey MK, Upadhyay RS. Antagonistic assessment of Trichoderma spp. by producing volatile and non-volatile compounds against different fungal pathogens. Arch Phytopathol Plant Prot. 2017;50: 629–648.
  44. 44. Weiwei L, Wie M, Bingyu Z, Feng Liu. Antifungal activities and components of VOCs produced by Bacillus subtilis G8. Curr Res Bacteriol. 2008;1: 28–34.
  45. 45. Zou C-S, Mo M-H, Gu Y-Q, Zhou J-P, Zhang K-Q. Possible contributions of volatile-producing bacteria to soil fungistasis. Soil Biol Biochem. 2007;39: 2371–2379.
  46. 46. Kai M, Effmert U, Berg G, Piechulla B. Volatiles of bacterial antagonists inhibit mycelial growth of the plant pathogen Rhizoctonia solani. Arch Microbiol. 2007;187: 351–360. pmid:17180381
  47. 47. Beneduzi A, Ambrosini A, Passaglia LMP. Plant growth-promoting rhizobacteria (PGPR): Their potential as antagonists and biocontrol agents. Genet Mol Biol. 2012;35: 1044–1051. pmid:23411488
  48. 48. Audrain B, Farag M a., Ryu C-M, Ghigo J-M. Role of bacterial volatile compounds in bacterial biology. FEMS Microbiol Rev. 2015;39: 222–233. pmid:25725014
  49. 49. Piechulla B, Lemfack MC, Kai M. Effects of discrete bioactive microbial volatiles on plants and fungi. Plant Cell Environ. 2017;40: 2042–2067. pmid:28643880
  50. 50. Yuan J, Zhao M, Li R, Huang Q, Raza W, Rensing C, et al. Microbial volatile compounds alter the soil microbial community. Environ Sci Pollut Res. 2017;24: 22485–22493. pmid:28803260
  51. 51. Reino JL, Guerrero RF, Hernández-Galán R, Collado IG. Secondary metabolites from species of the biocontrol agent Trichoderma. Phytochem Rev. 2007;7: 89–123.
  52. 52. Liu SY, Lo CT, Shibu MA, Leu YL, Jen BY, Peng KC. Study on the anthraquinones separated from the cultivation of Trichoderma harzianum strain Th-R16 and their biological activity. J Agric Food Chem. 2009;57: 7288–7292. pmid:19650641
  53. 53. Rybakova D, Rack-Wetzlinger U, Cernava T, Schaefer A, Schmuck M, Berg G. Aerial warfare: a volatile dialogue between the plant pathogen Verticillium longisporum and its antagonist Paenibacillus polymyxa. Front Plant Sci. 2017;8: 1294. pmid:28798756
  54. 54. Spraker JE, Wiemann P, Baccile JA, Venkatesh N, Schumacher J, Schroeder FC, et al. Conserved responses in a war of small molecules between a plant-pathogenic bacterium and fungi. Casadevall A, editor. MBio. 2018;9: 1–14. pmid:29789359
  55. 55. Spraker JE, Jewell K, Roze L V., Scherf J, Ndagano D, Beaudry R, et al. A volatile relationship: profiling an inter-kingdom dialogue between two plant pathogens, Ralstonia solanacearum and Aspergillus flavus. J Chem Ecol. 2014;40: 502–513. pmid:24801606