Antifungal activity of volatile compounds generated by endophytic fungi Sarocladium brachiariae HND5 against Fusarium oxysporum f. sp. cubense

The soil-born filamentous fungal pathogen Fusarium oxysporum f. sp. cubense (FOC), which causes vascular wilt disease in banana plants, is one of the most economically important Fusarium species. Biocontrol using endophytic microorganisms is among the most effective methods for controlling banana Fusarium wilt. In this study, volatile organic compounds (VOCs) showed strong antifungal activity against FOC. Seventeen compounds were identified from the VOCs produced by endophytic fungi Sarocladium brachiariae HND5, and three (2-methoxy-4-vinylphenol, 3,4-dimethoxystyrol and caryophyllene) showed antifungal activity against FOC with 50% effective concentrations of 36, 60 and 2900 μL/L headspace, respectively. Transmission electron microscopy (TEM) and double fluorescence staining revealed that 2-methoxy-4-vinylphenol and 3,4-dimethoxystyrol damaged the plasma membranes, resulting in cell death. 3,4-dimethoxystyrol also could induce expression of chitin synthases genes and altered the cell walls of FOC hyphae. Dichloro-dihydro-fluorescein diacetate staining indicated the caryophyllene induced accumulation of reactive oxygen species (ROS) in FOC hyphae. FOC secondary metabolism also responded to active VOC challenge by producing less fusaric acid and expressions of genes related to fusaric acid production were interrupted at sublethal concentrations. These findings indicate the potential of S. brachiariae HND5 as a biocontrol agent against FOC and the antifungal VOCs as fumigants.

This statement is required for submission and will appear in the published article if the submission is accepted. Please make sure it is accurate.

Unfunded studies
Enter: The author(s) received no specific funding for this work. The authors have declared that no competing interests exist. NO  widely used to manage this disease [5]. However, the abuse of synthetic antifungal agents may pose risks to the environment and human health [6]. Therefore, it is essential to develop alternative, environmentally friendly methods to control FOC. For this purpose, biological control agents have attracted increasing attention because of their low mammalian toxicity, low target specificity, and environmental friendliness [7,8].
Endophytic fungi, by definition, reside in the tissues beneath the epidermal cell layers and cause no apparent harm to the host [9]. As they can confer abiotic and biotic stress tolerance and increase the biomass of the host plant, many endophytic fungi have been studied as biological control agents, including species of Sarocladium (re-allocated from Acremonium genus), Aspergillus, Fusarium and Penicillium [10,11,31]. Among them, Sarocladium species are well studied and suitable for developing biological control agents because, as fungal endophytes, they can easily colonize host plants [12].
Sarocladium species are known to produce a wide range of bioactive compounds; thus, they could inhibit pathogen growth directly or indirectly via the stimulation of induced systemic resistance [13][14][15]. Additionally, Sarocladium species might help host plants accumulate nutrients and increase organic nitrogenous compounds [16].
In addition to the well-documented suites of soluble anti-microbial compounds found in endophytic fungi, these species may also emit a wide range of volatile organic compounds (VOCs) with strong inhibitory activity against microbial competitors [17,18]. VOCs are carbon-based solids and liquids that readily enter the gas phase by vaporizing at 0.01 kPa at a temperature of approximately 20°C; they include acids, alcohols, alkyl pyrones, ammonias, esters, hydrogen cyanides, ketones, and lipids [19].
Different kinds and amounts of VOCs are produced by microorganisms during both primary and secondary metabolism [20]. VOCs emitted by microorganisms have a variety of applications; for example, they are used to indicate biocontamination in the food industry and in indoor environments, and to identify and separate microorganisms [21][22][23]. In recent years, VOCs produced by microorganisms have been shown to be effective and eco-friendly biocontrol agents [24,25]. Most of these VOCs have antibacterial or anti-fungal activity, and some can induce defence responses and promote plant growth [26,27].
In a survey on the diversity of the endophytic fungi of Brachiaria brizantha, we isolated the strain HND5, which can effectively inhibit the growth of FOC mycelia. Based on the LSU (the large subunit rDNA), ITS (rDNA transcribed spacer region) sequences along with the culture morphology and whole genome sequence data, we characterized HND5 as a new species, Sarocladium brachiariae [28,29]. Although HND5 has antiphytopathogen activity, the active compounds and the mechanism of action remain unknown. In this study, we found that the VOCs emitted by HND5 significantly affect FOC growth. Three antifungal VOCs were identified from HND5, and the 50% effective concentration (EC50) values of these VOCs were analysed. Biochemical, microscopic and molecular biological analyses revealed that the antifungal VOCs resulted in the leakage of cytoplasm, cell death, inhibition of spore germination, differential gene expression, and significant alterations to secondary metabolism in FOC. In summary, the results of this study indicate that the VOCs emitted by HND5 show potential as biological control agents against FOC in agricultural production systems.

Microorganisms and culturing conditions
The antagonistic strain HND5, which was isolated from healthy leaf of Brachiaria  [30,31]. Fusarium oxysporum f. sp. cubense, which exhibits high virulence against banana, was used in this study [32,33]. The fungal strains used in this study were grown on potato-dextrose agar (PDA) plates at 28°C for 3-7 d. To produce microconidia, 5 mm-discs from FOC culture plates were placed into potato dextrose broth (PDB) at 28°C under shaking at 170 rpm/min for 3 d.

Antagonistic assay of VOCs emitted from HND5 against FOC mycelium growth and conidia germination
The VOCs produced by HND5 were tested according to the method of Raza with some modifications [34]. h. Active carbon was used as an absorbent to remove VOCs in the HND5 and FOC assay according to Gong's method with minor modification [35].

Collection and analysis of VOCs
For the production and collection of VOCs, 5-mm-diameter discs of HND5 from PDA culture plates were inoculated in vials with 50

Transmission electron microscopy (TEM) observation of FOC
PDA plates with FOC plugs were first inoculated at 28°C for 5 d, and the EC50 concentrations of different VOCs were added separately onto the covers. After 12 h, the VOC-treated and untreated mycelia were extracted and fixed with 2.5% glutaraldehyde overnight. The fixed cells were rinsed three times for 10 minutes with 100 mM phosphate buffer, post-fixed for 3 h in 1% osmium tetroxide, and dehydrated using an ethanol gradient. The samples were then embedded in Epon 812, sectioned using an ultra-microtome, and examined under a Hitachi HT-7700 transmission electron microscope [37]. The samples were checked with a Nikon NI/E microscope.

Extraction and detection of fusaric acid
Autoclaved corn kernels with 45% added water were used to test fusaric acid production using the method of Bacon with minor modification [39]. Corn kernels (10 g HC-C18 column (4.6 × 250 mm) was employed to analyse fusaric acid. Quantification was performed as previously described [40].  Table S4.

VOCs of HND5 inhibit mycelium growth and conidium germination in FOC
In this study, a dual-plate assay system was used to test the antifungal activities of VOCs emitted by HND5 against FOC mycelium growth and conidium germination because there was no physical contact between FOC and HND5 colony. As shown in Fig. 1, the diameter of untreated FOC mycelium reached 3 cm after three days of inoculation, and untreated conidia germinated after 48 h on the PDA plate. In contrast, the HND5-treated mycelium grew to a diameter of only 1 cm, and no conidia germinated. To verify that the VOCs produced by HND5 were directly responsible for the antifungal activity, charcoal, which can absorb VOCs, was added into the system. The added charcoal reduced the antifungal activity of HND5, indicating that the activity against FOC was indeed caused by the VOCs produced by HND5 (Fig. 1).

Figure 1. VOCs of HND5 inhibit mycelia growth and conidia germination of FOC.
CK: PDA with FOC plug or FOC conidia covered with petris containing PDA medium; HND5 treatment: PDA with FOC petri or FOC conidia covered with Petris containing PDA medium in-oculated with HND5 plug; HND5+Charcoal treatment: HND5 treatment with 5g charcoal. All plates were inoculated at 28℃, 3 d for mycelia growth and 48 h for conidia germination.

Gas chromatography/mass spectrometry (GC/MS) analysis of VOCs produced by HND5
Because SPME fibres with different coating materials have different absorption characteristics [36], three different SPME fibre coatings (PDMS, 100 μm; DVB, 65 μm; DVB/CAR/PDMS, 50/30 μm) were used in this study to obtain a complete picture of the VOCs produced by HND5. The total ion current chromatograms and identified compounds of the HND5 cultures and pure PDA medium after HS extraction were compared (Table S1: PDMS, 100 μm results;

Antifungal activities of individual VOC against FOC and EC50 analysis
Among the 17 identified VOC compounds, only three (2-methoxy-4-vinylphenol, benzene, 2M4V; 3,4-dimethoxystyrol, 34D; and caryophyllene, β-C) were available from regent companies. These three VOCs were selected for further testing of antifungal activity against FOC. As shown in Fig. 2, all three VOCs inhibited the growth of FOC mycelia at a concentration of 10 μL/plate. 2M4V and 34D showed a stronger antifungal activity than β-C at the same concentration. We also tested a range of concentrations to determine the EC50 values of the selected VOCs against FOC. The results showed that the inhibitory activities of the VOCs increased with VOC concentration (Fig. 3). Colony diameter was measured, and the EC50 was calculated via statistical analysis. As the volume of free space of the plates was 90 mL, we then transformed unite μL/plate into μL/L. The EC50 values of 2M4V, 34D and β-C against FOC were found to be 36, 60 and 2900 μL/L, respectively.

Micro-and ultrastructural changes to FOC hyphae induced by VOCs
To determine the mechanism of VOC activity, TEM was used to evaluate the ultrastructural damage to hyphae caused by the selected VOCs. As shown in Fig. 4, the normal FOC hyphae showed distinct cell walls, intact plasma membranes, uniformly distributed electro-dense cytoplasm and clearly visible cell organelles. In contrast, the FOC hyphae treated with 2M4V and 34D showed completely different and irregular structures without intact plasma membranes, smeared cytoplasm and leaked cell content.
In addition, the cell walls of the 34D-treated hyphae were two to three times thicker than those in the other groups. Treatment with the EC50 concentration of β-C did not drastically change the hypha structure; the β-C-treated hyphae had distinct cell walls, intact plasma membranes and visible cell organelles (Fig. 4).

The selected VOCs affect chitin synthesis in FOC
Chitin is one of the major components of cell walls in FOC and play an important role in pathogenesis [41]. TEM analysis showed that cell walls of 34D-treated FOC hyphae were thicker than normal hyphae (Fig. 4). Thus, we hypothesized that 34D could affect chitin synthesis in FOC. We analysed the expression levels of three different types of chitin synthase gene related with pathogenicity in FOC with or without VOCs treatment (at sublethal concentration) with quantitative real-time PCR: Class V (FOIG_06738) [42], ChsVb (FOIG_06735) [43] and Class 4 (FOIG_00580) [44]. Result showed expression level of ChsVb (FOIG_06735) and Class V (FOIG_06738) chitin synthase genes increased significantly after 34D treatment (Fig. 6).

VOC-induced accumulation of reactive oxygen species (ROS) in FOC
High concentrations of ROS are harmful to cells and can result in cell death [45]. To determine whether FOC cells accumulate ROS as a result of treatment with VOCs at sublethal concentration, a DCFH-DA-based ROS assay kit was used. As shown in Fig.   7, the untreated hyphae did not show any green fluorescence. Only a few cells treated with 2M4V or 34D showed green fluorescence. Unlike the other treatments, nearly all the β-C-treated mycelia showed strong green fluorescence. These findings indicate that treatment withβ-C can lead to the accumulation of ROS in FOC.

Selected VOCs reduce fusaric acid production in FOC
To determine the involvement of the three selected VOCs in the biosynthesis of fusaric acid in FOC, we inoculated FOC to a mix of sterilized wheat kernels/oats/corn (1:1:1) kernels with or without VOCs at the sublethal concentrations (2M4V and 34D: 1 μL/plate, β-C: 40 μL/plate) [39,46]. After 23 days of incubation, the VOCs had significantly reduced the accumulation of fusaric acid in the kernels (Fig. 8). In Fusarium, 1 gene cluster have been identified as the biosynthetic gene cluster of fusaric acid [47]. We analysed the expression levels of two genes of this cluster, FUB2 and FUB5. As shown in Fig. 9, 2M4V, 34D, and β-C all deduced the expressions of FUB2 and FUB5. These results suggest that at sublethal concentrations (2M4V and 34D: 1 μL/plate, β-C: 40 μL/plate), the three selected VOCs can negatively affect fusaric acid biosynthesis.  worldwide [49]. Low-toxicity and environmentally friendly biological control agents are required to control FOC. Because of their potential as biocontrol agents for fungal diseases, endophytic fungi have attracted considerable attention [49,50]. Sarocladium spp. have been identified as a promising agents for the biocontrol of plant diseases [15,51]. Sarocladium brachiariae HND5, an endophytic fungus isolated from healthy Brachiaria brizantha leaf, is a new species of Sarocladium [28]. In this study, the VOCs emitted by HND5 effectively inhibited FOC growth. Using SPME-GC-MS, we shown to have anti-inflammatory effects [53,54]. Volatile sesquiterpenes have been identified in many fungi-produced VOCs, including Streptomyces albulus, Fusarium oxysporum, and Gliocladium sp. [52,55,56]. Two sesquiterpenes, caryophyllene and α-cadinol, were identified from HND5-generated VOCs. Based on anti-FOC assay, caryophyllene was found to have weak antifungal activity with EC50 > 2900 μL/L headspace, in agreement with a previous report [57]. We also identified three derivatives of naphthalene in the HND5-derived VOCs in this study. Naphthalene derivatives have been reported as an antimicrobial VOC in the essential oils of wood or volatile constituents of propolis [58].
TEM was used to study the ultra-structure of FOC after treatment with VOCs. 2M4V and 34D were found to induce cytoplasm leakage by disrupting the plasma membranes of FOC hyphae (Fig. 4), consistent with the FDA/PI double fluorescence staining results (Fig. 5). The plasma membrane is the target of many other antifungal VOCs, such as 4-methoxystyrene produced by S. albulus and oxygenated aromatic essential oil compounds [52,59]. The TEM analyses also indicated the incrassation of FOC cell walls after treatment with 34D (Fig. 4). Many other VOCs also affect cell walls in fungi, including the VOCs produced by S. albulus and farnesol. However, these compounds disrupt the integrity of the cell wall; they do not cause incrassation [60,61]. Chitin is a major constituent of fungal cell walls synthesized by different types of chitin synthase genes. Besides of maintaining of cell wall integrity and structure, several chitin synthases play vital role in infection process of Fusarium, such as Class V, ChsVb and Class 4 types [42][43][44]. Class 4 chitin synthase co-regulate virulence, DON production and septum formation with chitin synthases in F. graminearum [45]. Class V and ChsVb chitin synthase is critical for pathogenicity and cell wall assembly in F. oxysporum [42,43]. Marta et al. found Class V chitin synthase is hypersensitive to plant antimicrobial defence compounds such as the tomato phytoanticipin a-tomatine or H2O2 and speculated that F. oxysporum requires a specific Class V chitin synthase for pathogenesis, most probably to protect itself against plant defence mechanisms [42].
Gene expression analysis indicated that expression levels of Class V (FOIG_06738) , ChsVb (FOIG_06735) chitin synthase genes increased significantly after treatment with 34D at sublethal concentration (Fig. 6), consistent with the TEM results (Fig. 4). Based on these evidences, we conjecture that FOC identify 34D as plant antimicrobial defence compounds and activate defence system including cell wall enhancement. Low ROS concentrations act as intracellular messengers for many molecular events, whereas large amounts of ROS are associated with cell death [25,62]. N-butanol, a volatile compound identified from Muscodor albus, induces ROC accumulation in bacteria [63]. Among the three VOCs evaluated in this study, only β-C caused the accumulation of ROS in FOC (Fig. 7).
Fusaric acid is a well-known nonspecific toxin produced by all Fusarium species.
Fusaric acid can kill banana cells and protoplasts and causes symptoms including the rotting of roots and pseudostems and the wilting of seedling leaves. Siwen Liu et al.
found both banana leaves and pseudostems exhibited increased sensitivity to Foc4 invasion when pretreated with fusaric acid and suggested that fusaric acid functions as a positive virulence factor and acts at the early stage of the disease development before the appearance of the fungal hyphae in the infected tissues [46]. Although fusaric acid is not considered to be a mycotoxin with significant health consequences to humans, it still causes pathological disorders in experimental animals and human cell lines [64].
In this study, we found that treatment with 2M4V, 34D and β-C decreased the production of fusaric acid in FOC (Fig. 8). Previous studies revealed that the production of fusaric acid is encoded by the fusaric acid biosynthetic gene cluster containing 12 genes (FUB1-FUB12) [47,65]. These results indicated that target VOCs could influence the biosynthesis of fusaric acid and this indicated that S. brachiariae HND5 could delay the invasion of banana by FOC by decreasing the production of fusaric acid with VOCs.
As FOC is a soil-born pathogen, soil sterilizers and fumigants are frequently used to control this pathogen [66]. Methyl bromide is an effective fumigant against soil-borne pathogens and was broadly used worldwide on many crops until 2015, when it was phased out because it depletes the ozone layer [67,68]. Many chemicals have been studied as alternatives to methyl bromide, including metham sodium, 1,3dichloropropene, chloropicrin, sulfuryl fluoride and methyl iodide [69,70]. In this study, three antifungal VOCs were identified from HND5. Of these, two possess high anti-FOC activity and show potential as methyl bromide alternatives. The mechanisms of antifungal activity of these VOCs against FOC were also clarified. The findings suggest that HND5 and the VOCs it generates show promise for use as biological control agents or fumigants against FOC in agricultural production systems.

Conclusion
This study identified seventeen compounds from the volatile organic compounds