Edeine B1 produced by Brevibacillus brevis reduces the virulence of a plant pathogenic fungus by inhibiting mitochondrial respiration

ABSTRACT Plant pathogenic fungi cause serious diseases, which result in the loss of crop yields and reduce the quality of crops worldwide. To counteract the escalating risks of chemical fungicides, interest in biological control agents to manage plant diseases has significantly increased. In this study, we comprehensively screened microbial culture filtrates using a yeast screening system to find microbes exhibiting respiratory inhibition activity. Consequently, we found a soil-borne microbe Brevibacillus brevis HK544 strain exhibiting a respiration inhibitory activity and identified edeine B1 (EB1) from the culture filtrate of HK544 as the active compound of the respiration inhibition activity. Furthermore, against a plant pathogenic fungus Fusarium graminearum, our results showed that EB1 has effects on multiple aspects of respiration with the downregulation of most of the mitochondrial-related genes based on transcriptome analysis, differential EB1-sensitivity from targeted mutagenesis, and the synergistic effects of EB1 with electron transport chain complex inhibitors. With the promising plant disease control efficacy of B. brevis HK544 producing EB1, our results suggest that B. brevis HK544 has potential as a biocontrol agent for Fusarium head blight. IMPORTANCE As a necrotrophic fungus, Fusarium graminearum is a highly destructive pathogen causing severe diseases in cereal crops and mycotoxin contamination in grains. Although chemical control is considered the primary approach to control plant disease caused by F. graminearum, fungicide-resistant strains have been detected in the field after long-term continuous application of fungicides. Moreover, applying chemical fungicides that trigger mycotoxin biosynthesis is a great concern for many researchers. Biocontrol of Fusarium head blight (FHB) by biological control agents (BCAs) represents an alternative approach and could be used as part of the integrated management of FHB and mycotoxin production. The most extensive studies on bacterial BCAs-fungal communications in agroecosystems have focused on antibiosis. Although many BCAs in agricultural ecology have already been used for fungal disease control, the molecular mechanisms of antibiotics produced by BCAs remain to be elucidated. Here, we found a potential BCA (Brevibacillus brevis HK544) with a strong antifungal activity based on the respiration inhibition activity with its active compound edeine B1 (EB1). Furthermore, our results showed that EB1 secreted by HK544 suppresses the expression of the mitochondria-related genes of F. graminearum, subsequently suppressing fungal development and the virulence of F. graminearum. In addition, EB1 exhibited a synergism with complex I inhibitors such as rotenone and fenazaquin. Our work extends our understanding of how B. brevis HK544 exhibits antifungal activity and suggests that the B. brevis HK544 strain could be a valuable source for developing new crop protectants to control F. graminearum.

of synthetic fungicides; however, many researchers are greatly concerned about these chemical fungicides, such as carbendazim, tebuconazole, and azoxystrobin that trigger mycotoxin deoxynivalenol (DON) biosynthesis (16,17).Biocontrol with BCAs can be an alternative approach to chemical control and could be used as part of the integra ted management of FHB and mycotoxin production (18).This study aimed to find a useful microbe exhibiting respiratory inhibitory activity, considering that SDHI and QoI fungicides targeting mitochondrial respiration are biologically active against all the stages of fungal growth (5,19).Furthermore, an initial discovery that the culture filtrate (CF) of the Brevibacillus brevis HK544 strain inhibited mitochondrial respiration led to the discovery of an active compound edeine B 1 (EB 1 ) responsible for the respira tion inhibition activity.The biological function of EB 1 through transcriptome analysis, drug-induced haploinsufficiency (DIH) analysis, and targeted mutagenesis revealed that EB 1 has effects on multiple aspects of respiration and that most of the mitochondrialrelated genes in F. graminearum are downregulated by EB 1 .With an emphasis on understanding the molecular basis of EB 1 in antifungal activity, our findings show that the B. brevis HK544 strain could be a useful resource for developing new natural agents to control FHB.

B. brevis HK544 exhibited a respiration inhibitory activity with a robust antifungal activity
In the search for new natural resources exhibiting respiratory inhibition activity, we found that the HK544 CF (20%, vol/vol) inhibited the growth of Saccharomyces cerevisiae in the NFYG medium but not in the YG medium, which was similar to the strobilurin fungicide kresoxim-methyl that has a role as a mitochondrial cytochrome bc1 complex inhibitor (see Fig. S1).At various concentrations, HK544 CF inhibited the growth of S. cerevisiae in the range of 55%-99% when they were grown in NFYG medium, whereas the growth inhibition of S. cerevisiae grown in YG medium ranged from 3% to 11% (Fig. 1A).These results suggest that the HK544 strain might produce active compounds exhibiting a respiratory inhibition activity against S. cerevisiae.
To explore the HK544 strain in the control of plant pathogenic fungi, we investigated the antifungal activity of HK544 against F. graminearum as a model pathogenic fungus in terms of mycelial and conidial growth.As a result, we observed that the HK544 strain inhibited the mycelial growth of F. graminearum based on the dual culture assay.The HK544 CF also had a minimum inhibitory concentration (MIC) of 1.25% against F. graminearum when a conidial suspension was used (Fig. 1B).To investigate whether the HK544 CF has a respiratory inhibition activity against F. graminearum, we measured the ATP production of F. graminearum with and without HK544 CF treatment.When the mycelia were treated by 0.1% and 1% of the HK544 CF, the ATP production decreased by 21.9% and 64.5% compared to the non-treatment control after a 1 h treatment incuba tion, respectively.At 4 h after the treatment, the ATP production decreased by 83.5% and 100% compared to the non-treatment control (Fig. 1C).These results support the possibility that the HK544 CF is involved in the respiratory inhibition of F. graminearum.
The HK544 strain was previously identified as a B. brevis at the genome level (20).Prior to the chemical identification of the respirator inhibitor from B. brevis HK544 CF, we analyzed and compared the secondary metabolites derived from the HK544 genome with several Brevibacillus and Bacillus species (21,22).Based on the antiSMASH analysis, we found that B. brevis HK544 contains 17 biosynthesis gene clusters (BGCs), including non-ribosomal peptide, polyketide synthases, and post-translationally modified peptide clusters, suggesting that the HK544 strain can produce various bioactive compounds such as tyrocidine and zwittermicin A (Fig. 1D).When compared with other Brevibacillus spp., the composition of HK544 BGCs was similar with other Brevibacillus spp.although the number of BGCs differed (Fig. 1D).Based on the previously reported bioactive compounds from B. brevis, limited information is available regarding the respiratory inhibitory activity of the bioactive compounds by B. brevis.

B. brevis HK544 produced EB 1 as a major active compound for respiratory inhibition
To isolate the active compounds showing respiratory inhibition activity from the B. brevis HK544 CF, we performed chromatography analyses based on the growth comparison of S. cerevisiae in NFYG and YG media.Consequently, the active compound HKC1 (117 mg) was obtained as colorless needles from the HK544 CF.The UV/Vis spectrum of HKC1 showed an end absorbance below 210 nm.The electrospray ionization mass spectrome try (ESI-MS) spectrum of HKC1 showed molecular ions at m/z 797 [M+H] + and m/z 819 [M+Na] + .The nuclear magnetic resonance (NMR) data of HKC1 (see Table S1) wholly agreed with those of EB 1 , which is comprised of β-tyrosine, β-serine, α,β-diamino-pro pionic acid, α-2,6-diamino-7-hydroxyazelaic acid, glycine, and polyamine (Fig. 2A) (23).EB 1 inhibited the growth of S. cerevisiae in the NFYG medium in a dose-dependent manner.In contrast, growth inhibition in the YG medium was less than 6% at the tested concentrations (Fig. 2B).These results suggest that EB 1 has effects on the respiration of S. cerevisiae.To explore how EB 1 inhibits the growth of F. graminearum, we investigated the inhibitory activity of EB 1 for conidial germination and mycelial growth.When F. grami nearum conidia were treated with EB 1 , conidial germination was inhibited in a dosedependent manner compared to the non-treatment control; at a concentration of 20 µg/mL, the conidial germination was inhibited by 75% (Fig. 2C).Furthermore, at the same concentration, we observed that EB 1 has higher antifungal activity for inhibiting conidial germination compared to mycelial growth (Fig. 2C and D).Considering that AMPs can cause morphological changes in the plasma membrane and cell wall and increase cell permeability, we observed morphological alterations on F. graminearum grown on potato dextrose agar (PDA) medium supplemented with EB 1 .The morphology of hyphae treated with EB 1 was similar to that of the normal control hyphae based on SEM observation (Fig. 2E).When conidia treated with EB 1 (10 µg/mL) were stained by propidium iodide, the EB 1 -treated conidia exhibited no red fluorescence similar to the non-treatment control (Fig. 2F).By contrast, ethanol-treated conidia as a positive control exhibited a much stronger red fluorescence than that of the control and EB 1 treatment (Fig. 2F), suggesting that EB 1 did not trigger fungal cell membrane defects.Given that EB 1 exhibits respiratory inhibition activity, we observed the mitochondria of F. graminea rum treated with EB 1 using transmission electron microscopy.Our results showed that there were no observable morphological changes in mitochondrial structure by EB 1 treatment (see Fig. S2).Furthermore, along with no morphological changes of mitochon dria, we observed there was no change in mitochondrial superoxide generation in F. graminearum treated with EB 1 (see Fig. S3).Taken together, our results suggest that EB 1 seems to indirectly regulate mitochondrial respiration in F. graminearum.
Considering that the peptide antibiotic edeine binds between the P-site and the E-site of the small subunit of the ribosome and consequently impairs protein synthesis (24), we investigated the effects of EB 1 on protein synthesis based on fluorescence in fungal cells.As a result, we observed strong green fluorescence signals from the germinated conidia of F. graminearum (Fig. 2G).However, when F. graminearum was treated with 5 and 10 µg/mL of EB 1 , the fluorescence intensity was reduced by 72% and 91%, respectively (Fig. 2G); as a positive control, we observed that a translation inhibitor cycloheximide treatment also exhibited a reduced fluorescence intensity (Fig. 2H).Furthermore, when F. graminearum was co-treated with EB 1 and cycloheximide, the fluorescence intensity decreased more than that of the single treatment, supporting that EB 1 and cycloheximide target different sites on the eukaryotic ribosome (24).

EB 1 downregulated genes involved in oxidative phosphorylation
Considering that edeines have been widely used as transcriptional inhibitors to study ribosomal function and protein synthesis (25), we performed transcriptomic analysis of the F. graminearum wild-type strain from which the fungal mycelia were treated with EB 1 and incubated for 1, 2, and 4 h (Fig. 3A).We considered differentially expressed genes (DEGs) showing differences in transcript accumulation with a log 2 fold change of greater than 1 or less than −1 (P < 0.05).Compared to the non-treatment control, a total of 7,876 DEGs were identified from F. graminearum mycelia, with 982 and 635 genes exhibit ing a significant increase and decrease in all time points, respectively (Fig. 3B).Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis on the 1,617 DEGs revealed that the up-and downregulated genes were mainly enriched in "ubiquinone and other terpenoid-quinone biosynthesis (fgr00130), " "pentose phosphate pathway (fgr00030), " "glycolysis/gluconeogenesis (fgr00010), " "ABC transporters (fgr02010), " and "nucleotide metabolism (fgr01232)" (Fig. 3C).In particular, all analyzed genes within the "ubiquinone and other terpenoid-quinone biosynthesis" pathway exhibited changes in expression levels in the samples treated with EB 1 for 4 h (Fig. 3D).Given the cru cial role of ubiquinone as a key component of the electron transport chain in the oxidative phosphorylation process (26,27), these results imply that EB 1 has effects on oxidative phosphorylation.Furthermore, when Gene Ontology (GO) enrichment analysis was performed, the DEGs were highly enriched in GO terms related to ribosomal function such as rRNA processing (GO:0006364), ribosome biogenesis (GO:0042254), and ribosomal large subunit assembly (GO:0000027) in the biological process category (see Fig. S4).In the category of molecular function, ATP binding (GO:0005524) was a dominant term, and translation initiation-related GO terms were identified, including translation initiation factor activity (GO:0003743), RNA helicase activity (GO:0003724), and translation initiation factor binding (GO:0031369).For the cellular component, the DEGs were enriched in nucleolus (GO:0005730), mitochondrion (GO:0005739), and mitochondrial inner membrane (GO:0005743).These results suggest that EB 1 may play an important role in the ribosome biogenesis, translation, and mitochondrial function of F. graminearum.
Considering that EB 1 exhibited a respiratory inhibition activity against S. cerevisiae, we further explored genes involved in oxidative phosphorylation of F. graminearum based on the KEGG database.From the RNA-Seq results, we selected 47 genes for the ETC of mitochondria.These genes were mostly downregulated 2 and 4 h after EB 1 treat ment, except for FGSG_01981 and FGSG_04854, which encode succinate dehydrogenase and ATP synthase, respectively (Fig. 4A).To validate the RNA-Seq results, we arbitra rily selected eight genes, seven downregulated genes, and one upregulated gene, for quantitative real-time PCR (qRT-PCR) analysis.Although the fold changes were different, the expression trends of these eight genes were consistent in both the RNA-Seq and qRT-PCR results (Fig. 4B).

Haploinsufficiency profiling of Schizosaccharomyces pombe strains treated with EB 1
To further investigate mitochondrial-related genes affected by EB 1 , we exploited a DIH assay with heterozygous deletion mutants of 342 genes related to mitochondria among a Schizosaccharomyces pombe genome-wide deletion mutant library (4,845 genes).EB 1 showed potent antifungal activity against wild-type S. pombe cells for 21 h (GI 50 = 80 µM; see Fig. S5).DIH by EB 1 in various S. pombe heterozygous deletion mutants revealed that the fitness scores of 10 gene deletion mutants were over 2.5 among the mutants, and a mpc1-deleted strain had the highest fitness value of 3.99 (Fig. 5A and  B).Among the F. graminearum homologs to the 10 S. pombe genes, we found that six homologs were differentially expressed in response to EB 1 (Fig. 5C).In particular, two genes FGSG_09250 encoding NADH dehydrogenase flavoprotein 1 in mitochondrial respiratory chain complex I and FGSG_07074 encoding mitochondrial pyruvate carrier 1 were assigned to the oxidative phosphorylation pathway (fgr00190) and mitochon drial biogenesis (fgr03029) of F. graminearum KEGG database, respectively.To further determine whether the target deletion of six homologs (FGSG_07074, 08530, 09250, 09683, 13011, and 13152) leads to chemical sensitivity to EB 1 , we generated gene deletion strains (see Fig. S6) and compared the growth of the wild-type and the deletion strains according to the treatment of EB 1 .As a result, we found that there was no significant difference in EB 1 sensitivity between the wild-type and deletion strains (see Fig. S7).

Deletion of FgSdhC1 increased the sensitivity to EB 1
Considering that EB 1 changed the expression of genes involved in oxidative phosphor ylation, we explored whether SDHI and QoI fungicides have effects on the transcript level of succinate dehydrogenase (FgSdhC1; FGSG_01981) and cytochrome c oxidase (FgCox7A; FGSG_09879) subunits.They were selected because their expressions were significantly increased and decreased by EB 1 , respectively (Fig. 6A).FgSdhC1 was highly induced and suppressed by the SDHI fungicide fluopyram and QoI fungicide kresoximmethyl, respectively, and both chemicals suppressed FgCox7A.These results were sillier than the RNA-Seq and qRT-PCR results derived from the EB 1 treatment.However, in contrast to the EB 1 treatment results, we observed that FgCox7A was induced by both chemicals 1 h after the treatment (Fig. 6A).
To further determine whether the deletion of succinate dehydrogenase genes leads to chemical sensitivity to EB 1 , we generated an FgSdhC1 deletion strain that was named Δ01981 (see Fig. S8) and compared the growth of the wild-type and Δ01981 strains according to the treatment of EB 1 .As a result, the Δ01981 strain was more sensitive to EB 1 compared to the wild-type strain below a concentration of 12.5 µg/mL; in particular, the greatest difference (2.7-fold) was shown at a concentration of 3.1 µg/mL (Fig. 6B).In addition, when we investigated the growth difference to fluopyram which binds to the ubiquinone-binding sites in the succinate dehydrogenase complex composed of four subunits (SdhA, SdhB, SdhC, and SdhD), the deletion of FgSdhC1 also led to a change in fluopyram sensitivity (Fig. 6B).Therefore, our observations that the deletion of FgSdhC1 increased the sensitivity to EB 1 support that EB 1 has effects on the mitochon drial respiration of F. graminearum.

HK544 CF and EB 1 reduced the development of FHB
As proof of concept that the in vivo antifungal activity of HK544 CF and EB 1 can be used as a BCA, we investigated their ability to control or reduce the FHB by treating flowering wheat heads with HK544 CF and EB 1 before inoculation with F. graminearum conidial suspension.The non-treatment control exhibited typical head blight symptoms, which manifested as discoloration, at 5 days post-inoculation (dpi), whereas the treatments with HK544 CF and EB 1 led to suppressed FHB development in a dose-dependent manner (Fig. 7A and B).In addition, given that chemical fungicides inhibiting mitochondrial respiration induce trichothecene production of F. graminearum (16,17), we investigated whether HK544 CF containing EB 1 induces trichothecene production.The results showed that the cultures treated with HK544 CF produced reduced trichothecenes in a dose-dependent manner without the differential expression of the trichothecene biosynthetic genes TRI5 (FGSG_03537; trichodiene synthase) and TRI6 (FGSG_03536; transcription factor) between the treated and untreated cultures (Fig. 7C and D).This result suggests that the reduced trichothecene production is from the inhibition of fungal growth by HK544 CF.Taken together, our results indicate that B. bacillus HK544 CF can control FHB with a reduction of trichothecene production.

Synergistic effects of the combinations of EB 1 and respiratory inhibitors
Considering that the combination assay can provide information on improved antifungal potency for the application and also help unravel the mechanism of action of the drugs, we investigated the synergistic effects between EB 1 and ETC inhibitors (rotenone for complex I, thenoyltrifluoroacetone (TTFA) for complex II, antimycin A for complex III, and potassium cyanide (KCN) for complex IV) against F. graminearum.First, we determined the MIC values of each ETC inhibitor against F. graminearum.The results showed that the MICs of rotenone and antimycin A were above the highest concentration tested (>200 µg/mL), and TTFA and KCN had MIC values of 25 and 12.5 µg/mL, respectively (Fig. 8).Because the fractional inhibitory concentration index estimates the drug interactions at the MIC levels, we used a zero interaction potency (ZIP) model to understand the interactions over the entire range of concentrations (28).When EB 1 (ranging from 1.6 to 25 µg/mL; MIC value of 12.5 µg/mL; see Fig. S9) was combined with ETC inhibitors, the highest average synergy score (18.337) was observed in the combination of EB 1 + rotenone among the tested combinations although the 2D and 3D plots of the modelspecific statistic showed regions of synergy in all the tested combinations (Fig. 8).An EB 1 + TTFA combination exhibited the lowest average synergy score (0.796) and also showed regions of antagonism in the matrix (Fig. 8).
Based on our observation that the combinations EB 1 + rotenone and EB 1 + TTFA exhibit the highest and lowest average synergy scores, respectively, we fur ther investigated the synergistic interactions between EB 1 and commercial fungicides targeting complexes I and II.When EB 1 was combined with the complex I inhibitor fenazaquin (MIC, >200 µg/mL) and complex II inhibitor fluopyram (MIC, 100 µg/mL), the average ZIP score was 14.107 and 5.864 (Fig. 8), respectively, suggesting that there was a synergistic effect between EB 1 and the complex I inhibitor fenazaquin.These results show that EB 1 synergistically interacts with complex I inhibitors rotenone and fenazaquin for the respiratory inhibition of F. graminearum.The summary ZIP score calculated over the full dose-response matrix and landscape is marked on top of the panels.The gray-shaded area with a white line indicates the most synergistic area in that particular dose-response matrix.

DISCUSSION
The mitochondrial machinery facilitating ATP synthesis by oxidative phosphorylation is a promising target for active compounds derived from microbial metabolites (9,29).The natural compound strobilurin A was discovered in the wood-rotting basidiomycete fungus Strobilurus tenacellus, and synthetic strobilurin fungicides (e.g., azoxystrobin and kresoxim-methyl) were developed from β-methoxyacrylate through optimization of their photostability and activity (5).Although fungicides targeting mitochondrial respiration are the dominant chemical groups used in the market (30), it has been reported that strobilurin fungicides do not inhibit the respiration of several fungal pathogens, such as Magnaporthe oryzae and Botrytis cinerea, in which an alternative oxidase is activated (31)(32)(33).Respiration is maintained through an alternative pathway when strobilurin binds to cytochrome bc 1 complex in the ETC (31)(32)(33).Due to issues of resistance and reduced effectiveness for these chemicals in several fungal pathogens, this study aimed to find natural resources exhibiting respiratory inhibitory activity.To this end, numer ous microbial CFs were explored using a yeast screening system for their respiratory inhibitory activity.As a result, we found a soil-borne microbe B. brevis HK544 strain, and identified EB 1 as an active compound for the respiration inhibition activity.
Brevibacillus is a genus of bacteria reclassified from Bacillus based on the 16S rRNA sequence analysis (14).This genus has been considered a rich resource for antibac terial and antifungal activity with the isolation and characterization of many active compounds such as gramicidin S, tyrocidines, tauramamide, bogorols, and laterosporu lin (14,(34)(35)(36).Most of the peptides produced by Brevibacillus have been known to damage the cytoplasmic membrane, but edeines exhibit a different mode of action from most Brevibacillus peptides (14).Edeines initially identified in B. brevis Vm4 are linear non-ribosomal peptides with a structurally unusual backbone that includes four non-proteinogenic amino acids and an N-terminal polyamine cap (37).This unique structure of edeines contributes to various biological functions, inhibiting the growth of numerous bacteria, fungi, mycoplasma, and tumor cells (38,39).Furthermore, it has been reported that low concentrations of edeines inhibit DNA synthesis (22,40).In contrast, high concentrations of edeine prevent translation initiation by inhibiting the binding of fMet-tRNA to the P site of the 30S ribosomes in prokaryotes (22,41).Although significant progress has been made by the docking model between edeine and prokaryotic/eukaryotic ribosomes, relatively little is known mechanistically about the biological function of edeines, particularly against plant pathogenic fungi.In this study, EB 1 exhibited a respiratory inhibition activity against the necrotrophic fungus F. graminearum with low ATP production and downregulation of ETC genes.Additionally, the deletion of FgSdhC1, one of the succinate dehydrogenase complex components in ETC, increased the sensitivity of F. graminearum to EB 1 .With previous reports that edeines inhibit protein synthesis by binding to ribosomes, our results postulated that EB 1 inhibits the growth of F. graminearum by inhibiting mitochondrial translation.This hypothesis can be explained by the fact that some ribosome-targeting antibiotics can inhibit mitoribosomes, considering that mitochondria are of prokaryotic origin, and the bacterial and mitochondrial protein translation machinery share similarities (42,43).Indeed, tetracycline binding to ribosome subunits has been reported to interfere with mitochondrial proteostasis, leading to changes in nuclear gene expression and altered mitochondrial dynamics and function in eukaryotes (44).When we calculated the predicted binding free energies with molecular dynamics simulations to compare the binding affinity of EB 1 to mitochondrial ribosome and cytosolic ribosome, our results showed that the EB 1 in the mitochondria ribosome exhibited higher binding affinity than cytosolic ribosome although the EB 1 can bind to both ribosomes which were located in cytosolic and mitochondria (see Table S2).Therefore, our results cannot exclude the possibility that inhibition of mitochondrial respiration by EB 1 is a major byproduct of an antibiotic-target interaction.
Biological control by antagonistic microbes, including yeasts, fungi, and bacteria, can be used to manage fungal diseases, and the BCAs can also solve the resistance and toxicity issues derived from chemical pesticides (10,45,46).Although numerous antagonistic strains of Brevibacillus spp.have been identified, most research has focused on in vitro antifungal activity with identifying active compounds, not disease control efficacy.In this study, we showed that the reduction of FHB by HK544 CF and EB 1 ranged from 54% to 96% at the tested concentrations.Unlike the chemical fungicides triggering mycotoxins (16,17), the HK544 CF reduced trichothecene production in the cultures, suggesting that B. brevis HK544 can decrease both fungal growth and trichothe cene production.Furthermore, our observation that EB 1 inhibited the fungal growth more effectively with the complex I inhibitor fenazaquin suggests that EB 1 is a potent chemosensitizer.These findings provide further information for the effectiveness of its in vivo antifungal activity, given that the disease control efficacy of B. brevis HK544 seems to be insufficient when used alone.Considering that antifungal chemosensitization is a novel antifungal intervention strategy in which one drug is used to enhance the efficacy of conventional agents against fungal pathogens (46), our results suggest that B. brevis HK544 could be used in combination with chemical fungicides to enhance the efficacy of antifungal agents.Further research, such as the formulation of HK544, field trials with a mixture of HK544 and chemicals, and HK544 sensitivity analyses against chemical fungicides, will be necessary to prove the feasibility of the B. brevis HK544 as a chemosensitizer for FHB control.
Regarding in vitro antifungal activity, we observed that EB 1 is more effective against conidial germination than mycelial growth.Similarly, it has been reported that chemical fungicides targeting mitochondrial respiration exhibit better fungicidal activity against spore germination than mycelial growth in many filamentous fungi (16).For example, the mycelial growth of B. cinerea was less sensitive to the SDHI fungicides fluopyram and boscalid compared to conidial germination (47,48).A similar dominant effect on the inhibition of spore germination has also been described from the QoI fungicides ( 16).This phenomenon can be explained by the differences in the target site sensitivity and physiological processes between spores and mycelia (49).Considering that SDHI and QoI fungicides target energy production in the mitochondria, and fungal spores have a lower metabolic activity with the transient softer cell walls compared to the growing mycelia, mitochondrial respiration inhibitors can lead to the different susceptibility between spores and mycelia (5).Therefore, we speculate that EB 1 has a similar action with SDHI and QoI fungicides in inhibiting fungal development.
In summary, our findings shed light on the in vitro antifungal activity of B. brevis HK544 and its major active compound EB 1 , with multifaceted evidence of respiratory inhibition.The promising results of the disease control efficacy in wheat heads infected with F. graminearum with the treatment of HK544 CF and EB 1 highlight their practical potential in crop protection against pathogenic fungi.Furthermore, the synergistic interactions between EB 1 and respiratory inhibitors, particularly complex I inhibitors, reveal novel opportunities for enhancing antifungal activity.Further research into the optimization of formulations and investigation of the activity spectrum could pave the way for the practical applications of HK544 containing EB 1 as an effective biofungicide in agriculture.

Microbial strains and culture conditions
The soil-borne bacterium B. brevis HK544 strain was isolated from the soil at Daejeon, South Korea, and deposited as a patent microorganism to the Korean Agricultural Culture Collection (KACC, Jeonju, South Korea; No. KACC 81093BP).The bacterium was main tained on tryptic soy broth (TSB; BD Difco, San Jose, CA, USA) agar plate at 30°C.For the in vitro and in vivo antifungal activity, the F. graminearum wild-type strain Z-3639 was used as a target fungal pathogen and grown in complete medium (CM) or potato dextrose broth (PDB; BD Difco) supplemented with 1.5% agar as needed (1).Conidia from F. graminearum were induced in a carboxymethyl cellulose medium (50).A minimal medium including 5 mM agmatine (MMA) was used to evaluate trichothecene produc tion (51).All microbial strains used in this study were stored in 20% glycerol at −80°C.

Mitochondrial respiratory inhibition assay
The inhibitory activity of mitochondrial respiration was investigated using the method described by Han et al. (9).Briefly, the growth inhibition of S. cerevisiae A-139 treated with CF or EB 1 was compared in two different media: YG (1% yeast extract and 2% glucose) and NFYG (1% yeast extract and 1% glycerol).After 24-h incubation, the optical density (OD 600 ) of each well was measured using a microplate reader (Bio-rad, Hercules, CA, USA), and the growth inhibition (%) of the A-139 strain was calculated as follows: [1 − (OD 600 of treatment/OD 600 of control)] × 100.The QoI fungicide kresoxim-methyl and 1% dimethyl sulfoxide (DMSO) were used as positive and negative controls, respectively.

Measurement of ATP production
To measure the ATP production of F. graminearum treated with the HK544 CF, a conidial suspension (5 × 10 7 conidia/mL) was inoculated into 50 mL of PDB.After 24 h incu bation, HK544 CF was added to the cultures at a final concentration of 0.1% and 1%, respectively.After 1-and 4-h incubation, mycelium was harvested and ground in liquid nitrogen with a pestle and mortar.ATP production in the ground mycelia was determined with the ATP Determination Kit (Invitrogen, Eugene, OR, USA) following the manufacturer's instructions.Fungal tissue (10 mg) was reacted with 1.25 µg/mL firefly luciferase, 50 µM D-luciferin, 1 mM dithiothreitol, and 1× reaction buffer.After a 15-min incubation, luminescence was measured by the BioTek Synergy LX multimode reader (Agilent Technologies).
The chemical structure of HKC1 was determined by spectroscopic analyses and comparisons with values in previous literature (23).The ESI-MS data were recorded on a single-quadruple mass spectrometer (Acquity QDa; Waters, Milford, MA, USA).All nuclear magnetic resonance (NMR) spectra were recorded in deuterium oxide (Cam bridge Isotope Laboratories, Tewksbury, MA, USA) at 25°C on the Bruker Advance 500 MHz spectrometer (Bruker BioSpin, Rheinstetten, Germany).

In vitro antifungal activity assay
For the inhibitory effect of EB 1 on the conidial germination of F. graminearum, EB 1 dissolved in DMSO was added to the microplate wells containing a conidial suspension (5 × 10 5 conidia/mL of PDB).After a 10 h incubation at 25°C, the number of germinated conidia was counted by microscopic observation in 100 conidia.To investigate the mycelial growth inhibition of EB 1 , a mycelial disk (5 mm in diameter) of F. graminearum was inoculated onto CM supplemented with different concentrations of EB 1 , and then, the radial growth of F. graminearum was measured at 5 dpi.To determine the MIC values of CF and EB 1 against F. graminearum, we used the broth microdilution method as previously described (52).Briefly, a conidial suspension (1 × 10 4 conidia/mL of PDB) of F. graminearum was added to the wells of a 96-well microtiter plate, and then, EB 1 stock solutions dissolved in DMSO were added at an initial concentration of 25 µg/mL by twofold serial dilutions.The 1% DMSO was used for a negative control.After incubation for 24 h at 25°C, the MIC was determined by visual examination and corresponded to the lowest concentration that caused complete growth inhibition.

Microscopic observation
The F. graminearum mycelia grown on CM containing the HK544 CF were observed under an FEI Quanta 250 FEG scanning electron microscope (Hillsboro, OR, USA) at 10 kV.To investigate the effect of EB 1 on F. graminearum membrane permeability, the conidia were incubated with EB 1 (12.5 µg/mL) for 4 h and then stained with 2 µM propidium iodide for 1 h.As a positive control, F. graminearum conidia was treated with 70% aqueous ethanol.Images were captured on an Olympus BS53 microscope (Münster, Germany) with a filter set of 494 nm/515 nm (excitation/emission wavelength).

Protein synthesis assay
To investigate the effects of EB 1 on protein synthesis in fungal cells, we evaluated the initial protein level using a Click-iT HPG Alexa Fluor 488 Protein Synthesis Assay kit (Thermo Fisher, Massachusetts, UK) according to the manufacturer's instructions.Briefly, a conidial suspension (1 × 10 6 conidia/mL of CM) of F. graminearum was treated with EB 1 (or cycloheximide) with 4 h incubation.After 1 h-additional incubation with HPG for 90 min, the fungal cells were fixed using a 3.7% formaldehyde solution and permeabilized with 0.5% Triton X-100 in PBS.The Click-iT reaction buffer containing Alexa Fluor dye with the azide moiety was added to the samples and then incubated for 30 min at room temperature in the dark.After washing with the Click-iT reaction rinse buffer, the samples were stained with 1× HCS NuclearMask blue stain solution, and the newly synthesized proteins were examined by an Olympus BS53 microscope (Münster, Germany) with a consistent exposure time.For the quantification, the fluorescence signal intensity was determined using the BioTek Synergy LX multimode reader (Agilent Technologies).

Transcriptome analysis
The F. graminearum Z-3639 strain was grown in 50 mL of CM at 25°C with shaking for 72 h and then incubated with EB 1 (10 µg/mL).After 1-, 2-, and 4-h incubation, the fungal mycelia were collected and washed twice with distilled water before being ground in liquid nitrogen.Total RNA was extracted with an Easy-Spin Total RNA Extraction Kit (iNtRON Biotechnology, Seongnam, South Korea) following the manufac turer's instructions, and RNA quality was assessed on the Agilent 2100 Bioanalyzer system (Agilent Technologies, Santa Clara, CA, USA).The total RNA was subjected to RNA-Seq library preparation using a TruSeq Stranded Total RNA Sample Prep Kit (Illumina, San Diego, CA, USA) following the manufacturer's instructions.Each library from three biological replicates per treatment (total 12 libraries) was sequenced for 101 bp paired-end reads with an Illumina NovaSeq6000 platform (Illumina) by Macro gen in Seoul, South Korea.The resulting sequences were analyzed using the Galaxy web-based platform.Trimming of reads was conducted using Trimmomatic (Galaxy version 0.38.1)(53), and the paired-end reads of each sample were aligned to the F. graminearum reference genome sequence using RNA STAR (Galaxy version 2.7.10b) (54).Gene expression quantification was computed using featureCounts (Galaxy version 2.0.3)(55), and the analysis of differential gene expression was performed using the DESeq2 R package (56).Enrichment analysis was performed in DAVID Bioinformatics Resources (57).

Quantitative real-time PCR analysis
Total RNA was extracted using an Easy-Spin Total RNA Extraction Kit (iNtRON Bio technology), and first-strand cDNA was synthesized from the total RNA using Super Script III First-Strand Synthesis SuperMix (Invitrogen).qRT-PCR was performed using SsoFast EvaGreen Supermix (Bio-Rad) with the CFX Real-Time PCR System (Bio-Rad).An endogenous housekeeping gene ubh (FGSG_01231, ubiquitin C-terminal hydrolase) of F. graminearum was used as an internal control for normalization (58).Relative expression levels were calculated through the 2 −ΔΔCT method (59).The experiments were repeated twice, with three replicates for each.Primers used in the qRT-PCR are listed in Table S3.

Generation of knockout strain
Gene knockout constructs were prepared by double-joint PCR as previously described (60).Briefly, the fungal genomic DNA of Z-3639 was extracted from freeze-dried mycelia powder according to the Fusarium laboratory manual (1).The 5′ and 3′ flanking regions (approximately 1.5 kb) of a target gene were amplified from F. graminearum gDNA and then were fused into a hygromycin resistance cassette (Hyg) amplified from pGEM-T_Hyg (61).The resulting constructs were transformed into the Z-3639 protoplasts using polyethylene glycol (PEG)-mediated transformation (50).Transformants grown on CM medium containing hygromycin were subjected to PCR-based screening using specific primer pairs.Primers used in the generation of knockout strains are listed in Table S3.

EB 1 -induced haploinsufficiency analysis
DIH assay was performed by using S. pombe heterozygous deletion mutants from Bioneer (Daejeon, South Korea) based on the measurement of cell growth inhibition (62).Briefly, S. pombe cells were grown at 30°C in a complete YES medium containing 0.5% yeast extract, 2% glucose, and various supplements as described (63).The GI 50 value of EB 1 was determined against the S. pombe wild-type strain (SP286; h+/h+, ade6-M210/ ade6-M216, ura4-D18/ura4-D18, leu1-32/leu1-32).Spore suspensions (1 × 10 6 cells/mL of YES medium) of a mitochondrial-related functional group subset, including 342 heterozygous deletion mutants of S. pombe, were added to the wells of a 96-well microtiter plate.Then, the EB 1 stock solution dissolved in DMSO was added at a final concentration of 80 µM.As a control, we used a YES medium containing 0.06% DMSO.The microtiter plates were incubated at 30°C for 20 h.The fitness value that represents the growth inhibitory potency of EB 1 at the GI 50 dose in each deletion mutant was calculated by dividing the cell mass (A 600 ) of the DMSO control with the one for the treated drug.Gene information on S. pombe was obtained from the PomBase database (64).

FHB disease control efficacy
To investigate the disease control efficacy of HK544 CF and EB 1 , the spray inoculation method was used with wheat (Triticum aestivum cv.Eunpa) that was grown in a glasshouse at 25°C ± 5°C for 4 weeks (65).Briefly, mid-anthesis stage wheat heads were sprayed with HK544 CF and EB 1 dissolved in sterile water using a hand-held atomizer; both samples contained 0.025% Tween 20 solution as a wetting agent.The sterile water alone served as a negative control.After the treated wheat plants were air dried for 24 h, 20 of each treated wheat head were sprayed with a conidial suspension (1 × 10 6 conidia/mL) of F. graminearum containing 0.025% Tween 20 solution.The inoculated plants were kept in the humidified chamber (25°C; 12 h photoperiod) for 3 days and then moved to a glasshouse at 25°C ± 5°C.After 2 days of additional incubation, we recorded a percentage of the wheat head surface showing FHB symptoms.The FHB disease control efficacy was calculated with the following equation: control efficacy (%) = 100 × (1 − B/A), where A is the mean of the lesion area (%) on the wheat head of the control plants and B is the mean of lesion area (%) on the wheat head of the treated plants (65).

Trichothecene analysis
A conidial suspension (1 × 10 6 conidia/mL) was incubated in MMA medium supplemen ted with or without HK544 CF at 25°C under stationary growth conditions to measure trichothecene production.Trichothecenes were extracted from 7-day-old MMA cultures with an ethyl acetate-methanol mixture (4:1, vol/vol) as described previously (49).Briefly, the extracts were dried and derivatized with Sylon BTZ (Supelco, Bellefonte, PA, USA).Sequentially, the derivatized samples were mixed with an equal volume of n-hexane and distilled water.Then, the hexane layer was analyzed for trichothecene production using a Shimadzu QP2020 gas chromatograph-mass spectrometer (GC-MS) (Shimadzu, Kyoto, Japan).Quantification was performed by comparing the peak areas of the samples with those of external standards of DON and 15-ADON (Sigma-Aldrich, St. Louis, MO, USA), and then normalized by biomass.

Synergetic effect assay
As previously described, a checkerboard assay was performed to determine synergism between EB 1 and respiratory inhibition compounds (66).Briefly, EB 1 and each inhibitor dissolved in DMSO were added using twofold serial dilutions in the x-axis and y-axis of a 96-well microtiter plate, respectively, containing an F. graminearum conidial suspension (1 × 10 4 conidia/mL).Rotenone (complex I), TTFA (complex II), antimycin A (complex III), and KCN (complex IV) were used as specific inhibitors for ETC complexes, and fenazaquin and fluopyram were used as chemical fungicides targeting complex I and II, respectively.As a negative control, we used a 1% DMSO treatment.After an 18 h incubation, the OD 600 value of each well was measured using a microplate reader (Bio-rad), and the growth inhibition (%) was calculated as follows: [1 − (OD 600 of combined treatment/ OD 600 of negative control)] × 100.Obtained data were analyzed with SynergyFinder (version 1.6.1),and the synergy score was evaluated based on the ZIP method (67,68).The results were interpreted by the ZIP scores as synergism, >10; indifference, −10 -10; and antagonism, <−10 (68).

Statistical analysis
All experiments were repeated at least two times with three replicates each unless otherwise noted.Data were expressed as the mean ± standard deviation.Statistical analysis was performed using R version 4.1.2software.A two-tailed unpaired Student's t-test was used to compare the two groups.P-values less than 0.05 were considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001).For FHB disease control efficacy, data were subjected to one-way ANOVA, and the means of the treatments were separated by Duncan's multiple range test (P < 0.05).Significant differences were indicated with different small letters in each bar.

FIG 1
FIG 1 Discovery of B. brevis HK544 exhibiting a respiratory inhibitory activity.(A) Growth comparison of S. cerevisiae in NFYG and YG media containing HK544 CF.PC, treatment of kresoxim-methyl (10 μg/mL).(B) Antifungal activity of B. brevis HK544.The upper layer shows a dual culture assay of HK544 strain against F. graminearum grown on PDA+TSA (1:1, vol/vol) medium, and the photos were taken 4 days post-inoculation (dpi) at 25°C.The bottom layer shows the antifungal activity of the HK544 CF against an F. graminearum conidial suspension.After a 24-h incubation, each well was stained with PrestoBlue reagent.(C) ATP production of F. graminearum mycelia treated with HK544 CF. (D) Genome-wide comparative analysis.Heatmap and dendrogram of average nucleotide identity based on the whole genome illustrating the phylogenetic relationships between the HK544 strain and nine other bacterial strains.Bars depict the number of biosynthetic gene clusters (BGCs) belonging to each genome, which were identified by antiSMASH analysis.Colored boxes represent different BGC types as provided by the legend.Asterisks indicate a statistically significant difference in mean values (n = 3; *P < 0.05; **P < 0.01; ***P < 0.001).NC, negative control.

FIG 2
FIG 2 Identification of EB 1 from B. brevis HK544 as an active compound for respiratory inhibition.(A) Chemical structure of EB 1 .DAPA, α,β-diamino-propionic acid; DAHAA, α-2,6-diamino-7-hydroxyazelaic acid.(B) Respiratory inhibition activity of EB 1 based on the growth comparison of S. cerevisiae in NFYG and YG media containing EB 1 .PC, treatment of kresoxim-methyl (10 μg/mL).(C) Inhibitory effects of EB 1 on the conidial germination of F. graminearum.(D) Inhibitory effects of EB 1 on the mycelial growth of F. graminearum.(E) Mycelial growth and morphological observation of F. graminearum on complete medium (CM) supplemented with EB 1 .Scale bar, 10 μm.(F) Effects of EB 1 on fungal membrane permeability.The conidial suspension was treated with EB 1 and then stained with propidium iodide.The conidia were treated with 70% aqueous ethanol as a positive control.Scale bar, 50 µm.(G) Inhibition of protein synthesis by EB 1 in germinating conidia of F. graminearum wild-type strain.Newly synthesized proteins were visualized using the Click-iT protein synthesis assay kit and quantified by relative fluorescence intensity.Scale bar, 50 µm.(H) Inhibitory effects of the combinations of EB 1 and cycloheximide on protein synthesis of F. graminearum wild-type strain.Scale bar, 50 µm.Asterisks indicate a statistically significant difference in mean values (n = 3; *P < 0.05; **P < 0.01; **P < 0.001).

FIG 3
FIG 3 Transcriptome analysis of F. graminearum treated with EB 1 .(A) Hierarchical clustering heat map of DEGs in three biological replications of different time points after the EB 1 treatment.The red and blue colors denote highly and weakly expressed genes, respectively.The color gradient represents the Z-score for the log 2 FC-based normalized values.(B) Venn diagrams representing the number of differentially up-and downregulated genes.The number of overlapping genes is found in both conditions.(C) Top 10 enriched KEGG pathways of the DEGs.(D) Visualization of the expression of genes enriched in the KEGG pathway "ubiquinone and other terpenoid-quinone biosynthesis." The color gradient represents the fold change value of genes from the EB 1 4-h treatment compared to the non-treatment control.

FIG 4
FIG 4 Differential expression of genes involved in oxidative phosphorylation of F. graminearum.(A) Heatmap visualization of transcriptional profiles of oxidative phosphorylation genes.The red and blue colors denote highly and weakly expressed genes, respectively.The color gradient represents the Z-score for the log 2 FC-based normalized values.C I-IV, respiratory chain complex I-IV; AS, ATP synthase.(B) Validation of the representative DEGs using qRT-PCR.Asterisks indicate a statistically significant difference in mean values (n = 3; *P < 0.05; **P < 0.01).

FIG 5
FIG 5 Identification of potential targets of EB 1 using the S. pombe mutants.(A) S. pombe deletion mutants of the mitochondrial-related functional group subset were treated with the GI 50 value (80 μM) of EB 1 .The fitness values of each mutant based on the cell growth inhibition were plotted as the mean of three biological replicates.Red dots indicate the top 10 target candidates of EB 1 .(B) Description of the top 10 target candidates of EB 1 .(C) Transcript levels of F. graminearum homologs to the target candidates derived from the RNA-Seq results.The red and blue colors denote highly and weakly expressed genes, respectively.The color gradient represents the Z-score for the log 2 FC-based normalized values.

FIG 6
FIG 6 Effects of chemical fungicides and EB 1 on F. graminearum strains.(A) Transcript levels of the genes encoding the putative succinate dehydrogenase (FGSG_01981) and cytochrome c oxidase (FGSG_09879) subunit by treatment of chemical fungicides.(B) Growth inhibition effects of EB 1 and fluopyram on the Δ01981 strain.FgWT, F. graminearum wild-type strain; Δ01981, an FGSG_01981 gene deletion strain.Asterisks indicate a statistically significant difference in mean values (n = 3; *P < 0.05; **P < 0.01).

FIG 7
FIG 7 FHB control efficacy and trichothecene analysis.(A) Each wheat head was inoculated with an F. graminearum conidial suspension (1 × 10 6 conidia/mL) 24 h after treatment with HK544 CF and EB 1 .NC, non-treatment control; Mock, mock inoculation was performed with the 0.025% Tween 20 solution.The photos were taken at 5 dpi.(B) The disease control efficacy (%) was calculated by the lesion area (%) of wheat heads compared to the non-treatment control.(C) Trichothecenes were extracted from 7-day-old cultures in liquid minimal medium including agmatine (MMA) supplemented with or without HK544 CF and quantified by gas chromatograph-mass spectrometer (GC-MS).(D) Transcript levels of the trichothecene biosynthetic genes Tri5 and Tri6 were measured by qRT-PCR from 4-day-old cultures in liquid MMA medium supplemented with or without HK544 CF.Different letters indicate a statistically significant difference at P < 0.05.

FIG 8
FIG 8 Synergistic effects of the combinations of EB 1 and respiratory inhibitors on F. graminearum.In each drug combination (A-F), ZIP synergy scores based on the F. graminearum growth were calculated by the SynergyFinder software, and the interaction landscapes are shown in both 2D (left) and 3D (right) with red color (synergism) and green color (antagonism).