Membrane disruption of Fusarium oxysporum f. sp. niveum induced by myriocin from Bacillus amyloliquefaciens LZN01

Myriocin damaged Fon cell membrane. The damage to the cell membrane of Fon was caused by the altered or disrupted expression of some membrane‐related genes and proteins at the mRNA and protein levels; these genes and proteins mainly including those related to sphingolipid metabolism, glycerophospholipid metabolism, steroid biosynthesis, ABC transporters, and protein processing in the endoplasmic reticulum.

Summary

Myriocin, which is produced by Bacillus amyloliquefaciens LZN01, can inhibit the growth of Fusarium oxysporum f. sp. niveum (Fon). In the present study, the antifungal mechanism of myriocin against Fon was investigated with a focus on the effects of myriocin on the cell membrane. Myriocin decreased the membrane fluidity and destroyed the membrane integrity of Fon. Significant microscopic morphological changes, including conidial shrinkage, the appearance of larger vacuoles and inhomogeneity of electron density, were observed in myriocin‐treated cells. A membrane‐targeted mechanism of action was also supported by transcriptomic and proteomic analyses; a total of 560 common differentially expressed genes (DEGs) and 285 common differentially expressed proteins (DEPs) were identified. The DEGs were further verified by using RT‐qPCR. The combined analysis between the transcriptome and proteome revealed that the expression of some membrane‐related genes and proteins, mainly those related to sphingolipid metabolism, glycerophospholipid metabolism, steroid biosynthesis, ABC transporters and protein processing in the endoplasmic reticulum, was disordered. Myriocin affected the serine palmitoyl transferase (SPT) activity as evidenced through molecular docking. Our results indicate that myriocin has significant antifungal activity owing to its ability to induce membrane damage in Fon.

Introduction

Watermelon (Citrullus lanatus) is one of the most widely distributed fruit crops in the world (Dou et al., 2018). However, the production of watermelon is seriously threatened by Fusarium wilt. Fusarium oxysporum f. sp. niveum (Fon), the causal agent of Fusarium wilt in watermelon, can survive for long periods and accumulate in soil. The symptoms of Fusarium wilt involve necrotic lesions, foliar wilting during pathogen invasion of the vascular system of watermelon plants, immature fruit and plant death (Costa et al., 2018). The control of Fusarium wilt has been a challenge to watermelon producers as it is a devastating soilborne disease. Therefore, producers have opted to use chemical fungicides (Everts et al., 2014); however, this is not an environment‐friendly strategy.

Biological control is one of the most potential methods for soilborne disease owing to its advantages of environmental friendliness, safety and sustainability (Jiang et al., 2015). Some bacterial strains have been evaluated as efficient biocontrol agents (BCAs) for suppressing soilborne disease; for example, Bacillus amyloliquefaciens SN16‐1 has a high potential of controlling F. oxysporum f. sp. lycopersici (Wan et al., 2018). B. amyloliquefaciens L3 can control Fon by producing diffusible and volatile organic compounds (Wu et al., 2019). B. velezensis F21 can induce systemic resistance to Fon in watermelon through related genes and phytohormone signalling factors (Jiang et al., 2019).

Studies on the biocontrol mechanisms of BCAs have progressed in recent years. The biocontrol mechanisms operating against pathogens involve changes in soil microbial diversity, inducing systemic resistance and the production of diffusible and volatile antimicrobial compounds. The biocontrol activity of BCAs is related to the production of antibiotic compounds including lipopeptides (LPs) (Koumoutsi et al., 2004; Molinatto et al., 2016). These LPs are of the iturin (such as bacillomycin D/F/L/Lc, iturin A/C and mycosubtilin), fengycin (fengycin A/B and plipastatin A/B) and surfactin (halobacillin, pumilacidin and surfactin) classes; these are major contributors to Bacillus biocontrol activity. For instance, B. amyloliquefaciens FZB42 produces fengycin and bacillomycin D, which show synergistic antagonistic activity against Fusarium oxysporum (Koumoutsi et al., 2004). Fengycin and surfactin of the B. subtilis strain SG6 have been considered to contribute to the inhibition of F. graminearum growth (Zhao et al., 2014). Surfactin, iturin and fengycin produced by B. velezensis Y6 and Y7 were reported to be responsible for the antimicrobial activity against Ralstonia solanacearum and Fusarium oxysporum (Cao et al., 2018). Bacillomycin D, which is secreted by B. velezensis SQR9, can bind to the iron transport regulator Btr and modulate biofilm formation through the KinB‐Spo0a‐Sinl‐SinR pathway to suppress the growth of Fusarium oxysporum (Xu et al., 2013). Phenazine‐1‐carboxamide, which is secreted by Pseudomonas piscium, can affect the activity of the fungal protein FgGcn5 that would thereby lead to the deregulation of histone acetylation at H2BK11, H3K14, H3K18 and H3K27 in Fusarium graminearum and eventually cause fungal death (Chen et al., 2018). The plasma membrane is the primary target of some LPs and antimicrobial compounds (Li et al., 2015). In a study, iturin has been reported to negatively affect the cytoplasmic membranes of yeast cells wherein the leakage of K⁺ ions and the growth of cell were affected (Ongena and Jacques, 2008). Fengycin also has been reported to cause perturbation, bending and pore formation on artificial membranes (Patel et al., 2011; Falardeau et al., 2013). Antimicrobial substances directly act on the cell membrane of pathogens, enhance membrane permeability and cause cell death (Li et al., 2015). The electrostatic interactions between chensinin‐1b and lipopolysaccharide can facilitate the disruption of the cytoplasmic membranes of Escherichia coli and Staphylococcus aureus (Sun et al., 2015).

Myriocin, a sphingosine analogue with antifungal antibiotic properties (De Melo et al., 2013), reduces the synthesis of sphingolipid compounds by inhibiting the activity of serine palmitoyl transferase (SPT) (Wadsworth et al., 2013). Sphingolipids are plasma membrane lipid constituents in eukaryotic cells (Yamaji‐Hasegawa et al., 2005). In a previous study, we found that sphingofungin compounds are secreted by B. amyloliquefaciens LZN01; myriocin was one of the secreted sphingofungin compounds that inhibited the growth of Fon, and its minimum inhibitory concentration (MIC) was 1.25 µg ml⁻¹ (Xu et al., 2019). However, there is limited information about the effect of myriocin on the cell membrane of Fon.

Omics technologies provide powerful tools for characterizing the modes of action of antimicrobial compounds in pathogens (Massart et al., 2015). However, the study of the mechanisms of action of myriocin on the cell membrane of Fon through proteomic and transcriptomic analyses has not yet been reported. In this study, we aimed (i) to evaluate the effect of myriocin on the microscopic morphology of conidia and membrane integrity and fluidity of Fon, (ii) to examine membrane perturbation by measuring calcein leakage from calcein‐loaded liposomes, (iii) to reveal the effect of myriocin on the gene expression at the mRNA and protein levels that would subsequently provide insights into the mechanism of myriocin action on the cell membrane of Fon and (iv) to analyse the enzyme activity of SPT and characterize the mode of myriocin action on SPT by simulating molecular docking. This study aimed to understand the damage mechanisms of myriocin on the cell membrane of Fon and to lay a good foundation for the application of B. amyloliquefaciens LZN01 in the future.

Results

Effect of myriocin on the microscopic morphology of Fon conidia

The morphological changes observed in Fon conidia that were exposed to myriocin for 12 h are shown in Figure 1. SEM imaging showed that Fon conidia were replete and had smooth surfaces in the negative control (CK1). In contrast, the Fon conidia treated with myriocin (MIC‐8MIC, 1.25 μg ml⁻¹–10 μg ml⁻¹) and amphotericin B (CK2) exhibited morphological abnormalities including surface indentation and shrinkage deformation; the indentation on the conidial surface deepened as the concentration of myriocin was increased (Fig. 1A). TEM micrographs of Fon conidia in the negative control (CK1) showed that the cell nucleus and a small number of vacuoles were visible and that the electron density in the cytoplasm was homogeneous. In contrast, the Fon conidia treated with amphotericin B (CK2) and myriocin exhibited abnormalities, including the appearance of larger vacuoles (blue arrows), separation of the cell wall and cytoplasm (red arrows), and inhomogeneity of electron density in the cytoplasm (Fig. 1B). This result indicated that myriocin damaged the morphology and affected the internal structure of the conidia; these effects contributed to the growth inhibition of Fon.

Fig. 1.

SEM (A) and TEM (B) images of Fon conidia exposed to amphotericin B and various myriocin concentrations. CK1, 0 μg ml⁻¹ myriocin; MIC, 1.25 μg ml⁻¹ myriocin; 2MIC, 2.5 μg ml⁻¹ myriocin; 4MIC, 5 μg ml⁻¹ myriocin; 8MIC, 10 μg ml⁻¹ myriocin; CK2, 1.25 μg ml⁻¹ amphotericin B.

Effect of myriocin on Fon cell membrane integrity

SYTO 9 and propidium iodide (PI) dye, which can distinguish the intact from membrane‐damaged cells, were used to analyse the membrane integrity of Fon. SYTO 9 generally stains live and dead cells in a population, whereas PI penetrates cells with damaged membranes and causes a reduction in SYTO 9 fluorescence when both dyes are present. Thus, the cells with intact membranes emit green fluorescence, whereas the cells with damaged membrane emit red fluorescence. The Fon conidia in the negative control (CK1) emitted green fluorescence under confocal laser scanning microscopy (CLSM); this indicated that the cell membranes were intact. After 12 h amphotericin B (CK2), a small amount of green fluorescence was observed; this may be related to the reduction in membrane integrity. After culturing with different concentrations of myriocin for 12 h, the intensity of red fluorescence increased as the concentration of myriocin was increased (Fig. 2). In myriocin‐treated cells (MIC‐8MIC, 1.25 μg ml⁻¹–10 μg ml⁻¹), the dead cells (red fluorescence) increased from 54.9% to 77.3%, and the hypoactive cells (yellow fluorescence) decreased from 26.5% to 9.1% (Fig. S1). This suggests that myriocin damaged the cell membranes of Fon, which thereby resulted in a large number of cell deaths.

Fig. 2.

Effect of myriocin on Fon cell membrane integrity. CK1, 0 μg ml⁻¹ myriocin; MIC, 1.25 μg ml⁻¹ myriocin; 2MIC, 2.5 μg ml⁻¹ myriocin; 4MIC, 5 μg ml⁻¹ myriocin; 8MIC, 10 μg ml⁻¹ myriocin; CK2, 1.25 μg ml⁻¹ amphotericin B.

Membrane fluidity and calcein leakage

The 1,6‐diphenyl‐1,3,5‐hexatriene (DPH) fluorescence polarization is closely related to the fluidity of the cell membrane; in other words, the increase in membrane fluidity implies that the fluorescence polarization (P) and anisotropy (r) will decrease (Cavanagh et al., 2019). Compared with CK1, the P and r of the Fon conidia exposed to different concentrations of myriocin significantly increased (Fig. 3A). The increase in P and r indicated that exposure to myriocin decreased the membrane fluidity of Fon. However, no significant difference in the P and r of the Fon conidia among MIC, 2MIC and 4MIC treatments was found. Maybe, the conidia survived in similar way after 12‐h exposure to those myriocin treatments.

Fig. 3.

Influences of myriocin on the membrane fluidity of Fon (A) and time courses for the leakage of calcein (B) after the addition of myriocin. CK1, 0 μg ml⁻¹ myriocin; MIC, 1.25 μg ml⁻¹ myriocin; 2MIC, 2.5 μg ml⁻¹ myriocin; 4MIC, 5 μg ml⁻¹ myriocin; 8MIC, 10 μg ml⁻¹ myriocin. The arrow indicates the addition of 10% Triton X‐100 at 12 min. Significant differences between treatment and control (CK1) are indicated by ‘*’ (*P ≤ 0.05 and **P ≤ 0.01, independent samples t‐test).

To determine whether the Fon cell membrane can be preferentially targeted by myriocin, membrane perturbation was examined by measuring the calcein leakage from phosphatidylcholine (PC)/phosphatidylethanolamine (PE)/phosphatidylinositol (PI)/ergosterol (5:4:1:2, w/w/w/w) liposomes. As shown in Figure 3B, encapsulated calcein was rapidly leaked from the liposomes upon exposure to different concentrations of myriocin. MIC (1.25 μg ml⁻¹), 2MIC (2.50 μg ml⁻¹), 4MIC (5.0 μg ml⁻¹) and 8MIC (10 μg ml⁻¹) of myriocin induced 41.7%, 54.5%, 63.4% and 68.3% leakage from the liposomes in 12 min, respectively. However, the calcein leakage from the liposomes in the control (CK1) was only 3.7%. These results demonstrated that myriocin caused damage to the membrane and that calcein escaped the liposomes through myriocin‐induced pores. Upon the addition of 10% Triton X‐100 at 12 min, an additional 20–76% increase in calcein leakage from liposomes was observed, and complete leakage of calcein (100%) was detected at 18 min.

GO functional annotation, KEGG pathway enrichment and expression analysis of DEGs and DEPs

In the present study, membrane‐related DEGs and DEPs were investigated by using combined transcriptome and proteome analyses. At the protein level, a total of 651 DEPs were identified by using a statistical criterion (fold changes ≥ 1.2 or ≤ 0.83, P‐value < 0.05). Among them, 584 and 352 DEPs belonged to CK1_VS_MIC and CK1_VS_8MIC, respectively; 285 common DEPs in both CK1_VS_MIC and CK1_VS_8MIC (Fig. 4A) were found. Of the 285 common DEPs, 135 and 150 proteins were significantly downregulated and upregulated under myriocin exposure respectively (Fig. 4B).

Fig. 4.

Venn diagram, number and GO functional annotations of the DEGs and DEPs (P < 0.05). CK1, 0 μg ml⁻¹ myriocin; MIC, 1.25 μg ml⁻¹ myriocin; 8MIC, 10 μg ml⁻¹ myriocin.

A. Venn diagram of the DEGs and DEPs in CK1_VS_MIC and CK1_VS_8MIC. This figure was performed by using Funrich 3.0.

B. The number of common DEGs or DEPs in both CK1_VS_MIC and CK1_VS_8MIC.

C: GO functional annotation of DEPs and DEGs. GO categories significantly enriched (P < 0.05) are shown in the figure. The ordinate represents the GO term, the nether abscissa represents the number of DEPs or DEGs (column) and the upper abscissa represents the per cent (dot). The biological process, cellular component and molecular function are represented by blue, yellow and green, respectively.

At the mRNA level, 1161 DEGs were identified (fold changes ≥ 2 or ≤ 0.5, P‐value ≤ 0.05). There were 560 common DEGs in both CK1_VS_MIC and CK1_VS_8MIC (Fig. 4A), of which 226 genes were significantly upregulated and 334 genes were significantly downregulated under myriocin exposure (Fig. 4B).

The common DEPs and DEGs were classified by using GO functional annotation. The DEPs and DEGs were mainly enriched in cellular components, such as membrane part, membrane, cell and cell part. Among them, the DEPs were mainly involved in the cell (66.6%) and cell part (66%), and the DEGs were involved in the membrane (93.6%) and membrane part (88.9%) (Fig. 4C). As shown in Figure 4C, a large percentage of DEPs and DEGs were involved in the membrane and membrane parts.

KEGG pathway enrichment analysis was performed. As shown in Figure 5A, all the common DEPs and DEGs were combined and significantly enriched in the top five pathways: ABC transporters, steroid biosynthesis, sphingolipid metabolism, glycerophospholipid metabolism and protein processing in the endoplasmic reticulum. A total of 27 DEPs (Table S2) and 26 DEGs (Table S3) were enriched in these five pathways. Protein processing in the endoplasmic reticulum was the most enriched pathway in the proteome analysis, and steroid biosynthesis was the most enriched pathway in the transcriptome analysis. In addition, three pathways (steroid biosynthesis, sphingolipid metabolism and glycerophospholipid metabolism) belonged to the lipid metabolism category; protein processing in the endoplasmic reticulum belonged to the folding, sorting and degradation category, and ABC transporters belonged to the membrane transport category (Table S4). The results indicated that some membrane‐related genes and proteins were affected by myriocin and were involved in some metabolic pathways.

Fig. 5.

KEGG pathway enrichment and clustering heatmaps of the DEGs and DEPs.

A. Significantly enriched KEGG pathways are shown in the figure. The ordinate represents the KEGG pathway names, the nether abscissa (column) represents the number of DEGs or DEPs, and the upper abscissa (dot) represents the significance level of enrichment.

B. Clustering heatmap of DEPs and DEGs. The colours indicate the expression level of the gene/protein [log₁₀(TPM + 1)]. A tree diagram of gene/protein clustering is shown on the left. The gene names/protein NCBI accession numbers are shown on the right. The two genes/proteins belong to same branch, which indicates their expression levels are similar.

As shown in the clustering heatmap (Fig. 5B and Table 1), 14 DEGs and 13 DEPS were downregulated, and 12 DEGs and 14 DEPs were upregulated. The upregulated DEGs were mainly enriched in sphingolipid metabolism and steroid biosynthesis. The downregulated DEGs were mainly enriched in protein processing in the endoplasmic reticulum and glycerophospholipid metabolism. The upregulated DEPs were mainly enriched in glycerophospholipid metabolism, sphingolipid metabolism and protein processing in the endoplasmic reticulum. The downregulated DEPs were enriched in glycerophospholipid metabolism and protein processing in the endoplasmic reticulum. In addition, the expression of genes and proteins was significantly changed in the ABC transporter pathway.

Table 1.

The number of upregulated and downregulated DEGs (26) and DEPs (27) in five pathways.

Pathway ID	Pathway name	Number of DEGs		Number of DEPs
		Up	Down	Up	Down
map00564	Glycerophospholipid metabolism	0	3	3	3
map00600	Sphingolipid metabolism	3	0	3	0
map00100	Steroid biosynthesis	7	2	2	2
map04141	Protein processing in endoplasmic reticulum	0	7	5	7
map02010	ABC transporters	2	2	1	1

Five pathways are glycerophospholipid metabolism pathway, sphingolipid metabolism pathway, steroid biosynthesis pathway, protein processing in endoplasmic reticulum pathway and ABC transporters pathway, respectively.

Correlation of expression levels and protein interaction network analysis

To further determine the target of membrane damage, a correlation analysis of DEG expression levels was performed by using the Spearman algorithm (correlation coefficients ≥ 0.5, q‐value = 0.05), and protein–protein interaction (PPI) network analysis was performed by using the STRING database (combined score ≥ 0.4).

The correlation of the expression levels of 26 DEGs was divided into three parts. The first part consisted of 7 DEGs in steroid biosynthesis, 3 DEGs in sphingolipid metabolism and 2 DEGs in ABC transporters. The second part consisted of 2 DEGs in glycerophospholipid metabolism and 1 DEG in protein processing in the endoplasmic reticulum. The third part consisted of 2 DEGs in glycerophospholipid metabolism, 1 DEG in ABC transporters, 2 DEGs in steroid biosynthesis and 6 DEGs in protein processing in the endoplasmic reticulum. FOXG_08532 in the third part was correlated with the expression levels of the 7 DEGs, and the correlation of FOXG_08532 with other genes was the greatest among all DEGs (Fig. 6A). The results indicated that 26 DEGs from different pathways were correlated, and we speculate that myriocin affected the expression of FOXG_08532 and triggered a series of reactions related to FOXG _08532.

Fig. 6.

Correlation analysis of expression levels among 26 DEPs, interaction networks of DEPs and distribution of network centre coefficient. (A) Each node represents a DEG, and a connection represents the correlation of DEGs. The size of the node represents correlations with the other DEGs. The colours of node represent five pathways. (B) Diagram of interaction network among DEPs. The nodes represent DEPs which are shown by NCBI accession ID. The size of a node is proportional to its connectivity (degree). The larger the node is, the more important it is in the network. (D) Prediction of DEGs‐encoded proteins interaction networks. The nodes represent DEGs‐encoded proteins, and the names of the nodes are displayed by gene name. (C and E) Distribution of the network centre coefficient. The abscissa represents the NCBI accession ID/gene name. The ordinate represents the centrality, including betweenness centrality, degree centrality and closeness centrality. The larger the three values are, the more important the node is in maintaining connection of entire network, the higher the centrality degree of the node is, and the closer the node is to the network centre.

In the PPI network, a total of 24 DEPs interacted. XP_018237142.1 (methylene‐fatty‐acyl‐phospholipid synthase) is involved in the glycerophospholipid metabolism pathway. The degree centrality, closeness centrality and betweenness centrality values of XP_018237142.1 were 0.3, 0.44 and 0.64 (Fig. 6C), respectively, and these values were higher for XP_018237142.1 than for the other proteins. The results showed that methylene‐fatty‐acyl‐phospholipid synthase was the core protein, which thereby indicated that it plays an important role in interacting with the protein network (Fig. 6B). This methylene‐fatty‐acyl‐phospholipid synthase interacted not only with 5 DEPs in the glycerophospholipid metabolism pathway but also with XP_018249071.1 (DnaJ‐like subfamily B member 12) and XP_018240962.1 (hypothetical protein FOXG_05558) in protein processing in the endoplasmic reticulum pathway. In addition, the top six DEPs in Figure 6C are involved in the glycerophospholipid metabolism pathway, sphingolipid metabolism pathway and steroid biosynthesis pathway. The results showed that the expression levels of some proteins involved in the above metabolism pathway were affected by myriocin.

The relationship between the DEGs and DEPs was analysed by combining the transcriptome data, proteome data and PPI network. The interactions of proteins encoded by 25 DEGs were predicted, and a total of 6 candidates overlapped in the transcriptome and proteome data sets (Table S5). Six candidate DEPs belonged to the PPI network (Fig. 6B); FOXG_01590, FOXG_03186, FOXG_11535, encoding C‐8 sterol isomerase (XP_018234407.1), methylene‐fatty‐acyl‐phospholipid synthase (XP_018237142.1) and serine palmitoyl transferase (XP_018249811. 1) played an important role in the interaction of genes (Fig. 6D and E). The methylene‐fatty‐acyl‐phospholipid synthase (XP_018237142.1) and serine palmitoyl transferase (XP_018249811. 1) were 2 of the top six DEPs (Fig. 6C). The FOXG_01590 and FOXG_11535 correlated with the expression levels of other DEGs (Fig. 6A). The combined analysis of the transcriptomic and proteomic data indicated that myriocin might target FOXG_01590 or FOXG_11535, change the expression of the proteins and finally affect the function of the cell membrane.

RT‐qPCR analysis

To confirm the RNA‐Seq data, the expression of 9 typical genes was analysed by using RT‐qPCR. The expression levels of FOXG_04166 and FOXG_11535 in myriocin‐treated cells were significantly increased with reference to those in the control; and the expression levels of FOXG_07972, FOXG_03186, FOXG_01897, FOXG_07760, FOXG_04703, FOXG_02300, FOXG_04252, FOXG_06096 and FOXG_01429 were significantly decreased in myriocin‐treated cells (Fig. 7). The trend of expression of 9 genes was consistent with the transcriptome analysis results.

Fig. 7.

Validation of RNA‐seq data using RT‐qPCR. The left panel represents the RNA‐seq data, and the right panel represents the mean values of log₂ (fold change) observed for the 8MIC myriocin treatment vs. CK1 (*P ≤ 0.05, **P ≤ 0.01). The different colours of the lined boxes represent different pathways.

Assay of SPT activity and molecular docking with myriocin

To confirm the differential expression of SPT according to proteomics analysis, SPT activity was measured by using ELISA. As shown in Figure 8A, the SPT activities were higher in the MIC and 2MIC treatments than in the control (CK1). Interestingly, there was a significant decrease in SPT activity in 8MIC treatment compared with that in CK1. These suggest that the response of SPT activity to different concentrations of myriocin were different.

Fig. 8.

SPT activity of Fon cells in myriocin‐treated, 3D diagram, and surface drawing of molecular docking of myriocin to SPT. CK1, 0 μg ml⁻¹ myriocin; MIC, 1.25 μg ml⁻¹ myriocin; 2MIC, 2.5 μg ml⁻¹ myriocin; 4MIC, 5 μg ml⁻¹ myriocin; 8MIC, 10 μg ml⁻¹ myriocin.

A. SPT activity of Fon. Significant differences between treatment and control (CK1) are indicated by ‘*’ (**P ≤ 0.01, independent samples t‐test).

B. 3D structures of SPT and the best 3D conformation of myriocin. The protein is shown as ribbons with one subunit in green and the other subunit in blue. Myriocin is shown as a stick: C in purple, N in dark blue, O in red, H in white.

C. A detailed 3D diagram of molecular docking of myriocin to SPT. Surface drawing of myriocin‐SPT binding is shown: one subunit is in purple, the other subunit is in yellow, and myriocin is in green (left). 3D diagram of myriocin bind to the internal active site of SPT (right). Myriocin and amino acid residues are shown as sticks. Carbon, nitrogen, oxygen and hydrogen atoms in amino acid residues are shown in green, blue, red, and white, respectively. Atoms in myriocin are coloured (see B legend for details). Hydroxyl groups of myriocin are indicated by different letters (a, b, c, and d). The distance between the two electronegative atoms in the hydrogen bond is 3 Å. Hydrogen bonds are shown as yellow dashed lines, and the residues in the binding site are labelled.

Molecular docking is a key tool in predicting the binding modes of ligand–protein interactions and is increasingly popular for studying ligand–protein interactions (Thomsen and Christensen, 2006). Based on proteomics analysis, the SPT (XP_018249811.1) encoded by FOXG_11535 exhibited differential expression in myriocin‐treated cells. To verify the interaction between myriocin and SPT, a molecular docking of myriocin to the 3D protein model of SPT (XP_018249811.1) was performed. Figure 8B shows the 3D conformation of myriocin with the best docking model and with a low grid score of −60.5 kcal mol⁻¹. Shown in Figure 8C is a myriocin bound to the internal active site of SPT. The fatty acid hydrocarbon chain of myriocin inserted into a hydrophobic pocket of SPT and formed a hydrophobic interaction. The polar portion of myriocin interacted with several amino acid residues located in the active site of SPT. Two hydroxyl groups (a and b) of myriocin and Asn325 formed hydrogen bonds. The hydroxyl group c of myriocin and Leu324 formed a hydrogen bond, and the hydroxyl group d of myriocin and Ala563 formed another hydrogen bond. The amidogens of myriocin and Ala563 and Pro562 also formed hydrogen bonds. Moreover, the carboxyl group of myriocin and Arg330 formed a hydrogen bond. Myriocin provided hydrogen, and the amino acid residues of SPT accepted hydrogen, leading to an interaction between myriocin and SPT. The results showed that myriocin and SPT formed hydrophobic interactions and seven hydrogen bonds; this might be what could have led to the stable interaction between myriocin and SPT, wherein the biological activity and function of SPT were inhibited.

Discussion

In the present study, we focused on understanding the mechanism of action of myriocin in Fon. SEM and TEM observations showed that the growth inhibition of Fon in response to myriocin was accompanied by conspicuous morphological and cytological changes, including shrivelled cell surface, condensation of the cytoplasm, appearance of larger vacuoles, and the separation of cell wall and cytoplasm. These effects were similar to those of some antibiotics (Yan et al., 2010; Zhang et al., 2013, 2016). In a study, the formation of larger vacuoles has been reported to be closely related to antifungal activity (Thimon et al., 1995).

Some studies have indicated that some antifungals target the plasma membrane (Spampinato and Leonardi, 2013; Lee and Lee, 2015). The propidium iodide influx assay can reveal the loss of membrane integrity and pore formation in the cell membrane (Spyr et al., 1995). Hence, cells without a cell membrane or damaged cell membrane are considered dead cells (Gumbo et al., 2014; Cronjé and Gaynor, 2019). In this study, we determined by using CLSM that the number of dead cells increased as the concentration of myriocin was increased (Fig. 2); this indicates that myriocin damaged membrane integrity and caused cell death (Li et al., 2012). A similar result was observed in the mimic membranes of liposomes. Myriocin can induce calcein leakage from PC/PE/PI/ergosterol liposomes (Fig. 3B), which thereby suggests that calcein escape from liposomes might be related to myriocin‐induced pore formation. Damage to membrane integrity and shrinkage of cell occurs along with the loss of ions or intracellular components (Cho et al., 2013). In this study, the shrinkage of myriocin‐treated Fon conidia was observed by using SEM (Fig. 1A); this result might be related to myriocin inducing injury to the membrane.

Membrane fluidity affects membrane functions such as biochemical reactions, protein secretion, ion transport and permeability (Arora et al., 2000). In this study, myriocin decreased membrane fluidity (Fig. 3A), which is consistent with the decrease in membrane fluidity observed in silymarin‐treated cells in another study (Yun and Lee, 2017).

Phospholipids are the main components of eukaryotic cell membranes and consist of glycerophospholipids and sphingolipids. In glycerophospholipid metabolism, the main metabolites are glycerophospholipids, including PC (Morii et al., 2014), PE (Kimblehill et al., 2018), PG (Patiño‐Márquez et al., 2018) and PI (Collin and Cochet, 2013). The hypothetical protein FOXG_06093 (XP_018241738.1), which is involved in the synthesis of PE, FOXG_03186 and its encoded protein (methylene‐fatty‐acyl‐phospholipid synthase), which are involved in the synthesis of PC, phosphatidylserine (XP_018244080.1), which is involved in the synthesis of CDP‐diacylglycerol synthase, and FOXG_10104, FOXG_04329 and phosphatidylserine synthase (XP_018246094.1), which are involved in the synthesis of PG, were all significantly downregulated at the mRNA and protein levels in myriocin‐treated cells. Moreover, phospholipase D (XP_018240842.1), which is involved in the synthesis of 1,2‐dimyritoyl‐sn‐glycero‐3‐phosphatidylcholine (DMPC), and CDP‐diacylglycerol‐inositol 3‐phosphatidyltransferase (XP_018243765.1), which is involved in the synthesis of PI, were significantly upregulated in myriocin‐treated cells. Myriocin significantly reduced the expression levels of genes and proteins and affected the content of phospholipids, which thereby led to the damage to cell membrane and ultimately inhibited the growth of Fon. Methylene‐fatty‐acyl‐phospholipid synthase was a key enzyme in the synthesis of PC and played an important role in the protein interaction network (Fig. 6). The decrease in the expression of methylene‐fatty‐acyl‐phospholipid synthase might have led to the decrease in the expression of multiple proteins and change the content of multiple phospholipids. Moreover, the contents of PC and DMPC on the surface of the lipid membrane were changed; this may cause damage to the cell membrane (Nabika et al., 2016). In a study, changes in the content of PI have been reported to result in the decreased fluidity of the membrane (Andoh et al., 2016). In this study, the decrease of membrane fluidity in myriocin‐treated cells might be due to the influence of myriocin on PI content.

Sphingolipid metabolism has been known to play a role in cell death, survival and therapy resistance (Beckham et al., 2013). De novo sphingolipid metabolism is initiated by serine palmitoyl transferase (SPT) (Perry, 2002). In this study, myriocin induced the overexpression of FOXG_11535 and SPT, or it directly interacted with SPT (Fig. 8) and then caused the overexpression of FOXG_11535. The overexpression of FOXG_11535 (Fig. 6A) led to the upregulation of other genes in the pathway, and then, SPT interacted with other proteins (Fig. 6), which may cause the overexpression of proteins in the entire pathway. Finally, sphingolipid metabolism was upregulated in myriocin‐treated cells. Organisms are exposed to life‐threatening risks that may disrupt sphingolipid metabolism. The upregulation of sphingolipid metabolism may be related to the resistance of Fon to myriocin treatment (Tani and Funato, 2018). L‐Serine and palmitoyl‐CoA are catalysed by SPT, which reacts with other enzymes to synthesize excess dihydrosphingosine and dihydroceramide. These reactions may cause a decrease in membrane fluidity (Andoh et al., 2016) and result in the damage to the Fon cell membrane and eventually cell death (Suh et al., 2018). In this study, SPT exhibited higher activities at MIC (1.25 μg ml⁻¹) and 2MIC (2.50 μg ml⁻¹) than at 4MIC (5.0 μg ml⁻¹) and 8MIC (10 μg ml⁻¹) (Fig. 8B). This might be due to the resistance of Fon to myriocin treatment (MIC, 1.25 μg ml⁻¹ and 2MIC, 2.50 μg ml⁻¹) by increasing SPT activity. The inhibition of SPT activities at 4MIC (5.0 μg ml⁻¹) and 8MIC (10 μg ml⁻¹) might reduce the total level of sphingolipids in cells and cause a strong growth defect (Tani and Funato, 2018). Furthermore, myriocin affected SPT activity as evidenced through molecular docking (Fig. 8).

Nascent membrane proteins are located adjacent to ER membrane proteins throughout their integration and translation (Thrift et al., 1991). In this study, the expression levels of genes and proteins in protein processing in the endoplasmic reticulum pathway were decreased; this might lead to defects in ER membrane and membrane proteins. In a study, the protein transporter sec‐13 (XP_018241549.1) was reported to be involved in the formation of double‐membrane transport vesicles from the ER to the Golgi (Bouain et al., 2014). Moreover, the expression levels of the protein transporter sec‐13 were significantly downregulated when Fon was exposed to myriocin, which could block the synthesis of transport vesicles and ultimately prevent membrane proteins from reaching the cell membrane (Kaneko et al., 2019). In addition, the E3 ubiquitin‐protein ligase synoviolin (XP_018238877.1), which reduce the ubiquitination of target proteins and retard the cells growth (You et al., 2016), was significantly downregulated in myriocin‐treated cells.

ATP‐binding cassette (ABC) transporters constitute a superfamily of membrane proteins that are responsible for the translocation of many molecules across membranes with ATP (Rees et al., 2009). In this study, the expression levels of some genes and proteins related to ABC transporters were altered or disrupted by myriocin. For example, the FOXG_20796 gene and protein (XP_018251339.1), which are related to the translocation of iron‐sulfur clusters, were significantly downregulated in myriocin‐treated cells; this effect may cause abnormal metabolism and membrane damage (Do et al., 2018). Moreover, FOXG_15633 was upregulated by myriocin; this upregulation might be related to the resistance of Fon to myriocin (Dolberg and Reichl, 2017) and indicate the adjustment of Fon growth to myriocin.

Steroid biosynthesis is a classic target of antifungal compounds (Song et al., 2019b). In the steroid biosynthesis pathway, the main metabolites are sterols, including cholesterol, lanosterol, sitosterol, sterol and ergosterol (Fujita et al., 2018). Ergosterol plays an important role in membrane‐bound enzyme activity, membrane fluidity, membrane integrity and substrate transport (Ferrante et al., 2016). In this study, the expression levels of FOXG_01590 and C‐8 sterol isomerase (XP_018234407.1), which are related to the synthesis of ergosterol, were upregulated in myriocin‐treated cells (Fig. 6). Furthermore, methylsterol monooxygenase (XP_018244831.1) and hypothetical protein FOXG_09168 (XP_018246232.1) were significantly downregulated in myriocin‐treated cells. Also, the expression of FOXG_04703 and methylsterol monooxygenase (XP_018244831.1) was significantly downregulated, which would thereby reduce steroid biosynthesis, destroy the cell membrane structure, change membrane‐bound enzyme activity, decrease membrane fluidity and cause damage to the Fon cell membrane (Scheinpflug et al., 2017). These results suggest that some membrane‐related genes and proteins are targets for myriocin action.

Conclusion

The mechanisms of myriocin on the Fon cell membrane are shown in Figure 9. Myriocin decreased cell membrane fluidity, destroyed membrane integrity, induced leakage of calcein, caused shrinkage of cell morphology and inhibited Fon growth or induced cell death. The damage to the cell membrane of Fon was caused by the altered or disrupted expression of some membrane‐related genes and proteins at the mRNA and protein levels; these genes and proteins mainly include those related to sphingolipid metabolism, glycerophospholipid metabolism, steroid biosynthesis, ABC transporters and protein processing in the endoplasmic reticulum. In conclusion, myriocin exhibited potent antifungal effects on Fon via membrane disruption.

Fig. 9.

Schematic diagram of the mechanism of myriocin on the Fon cell membrane. Myriocin damaged Fon cell membrane. The damage to the cell membrane of Fon was caused by the altered or disrupted expression of some membrane‐related genes and proteins at the mRNA and protein levels; these genes and proteins mainly including those related to sphingolipid metabolism, glycerophospholipid metabolism, steroid biosynthesis, ABC transporters and protein processing in the endoplasmic reticulum. The red and green letters represent upregulated and downregulated of genes and proteins, respectively. The DEGs and DEPs are listed in Table S6.

Experimental procedures

Experimental preparation

The Fon used in this study was provided by the Laboratory of Microbial Ecology, Qiqihar University in China (Xu et al., 2019). Fon mycelium was incubated in potato dextrose agar (PDA) Petri dishes in the dark at 28°C for 5 days to induce sporulation. The dishes were soaked in sterile distilled water, and the conidia were separated from the culture surface with a brush. Afterwards, sterile four‐layer cheesecloth was used for the filtration of the suspension and removal of mycelial fragments. A hemocytometer was used to establish a conidia concentration of 1.0 × 10⁶ CFU (colony‐forming units) ml⁻¹. Myriocin and amphotericin B were dissolved in DMSO and diluted into final concentrations (MIC, 1.25 μg ml⁻¹; 2MIC, 2.5 μg ml⁻¹; 4MIC, 5 μg ml⁻¹ and 8MIC, 10 μg ml⁻¹; amphotericin B, 1.25 μg ml⁻¹) with DMSO at concentrations less than or equal to 1% (Wadsworth et al., 2013).

Observation of conidial morphologies by scanning and transmission electron microscopy

To observe the morphological changes of conidia caused by myriocin, electron microscopy was used. The Fon conidial suspensions were incubated with myriocin for 12 h at final concentrations of MIC (1.25 μg ml⁻¹), 2MIC (2.50 μg ml⁻¹), 4MIC (5.0 μg ml⁻¹) and 8MIC (10 μg ml⁻¹). For scanning electron microscopy (SEM), the myriocin‐treated conidia were centrifuged at 8000 r.p.m. for 5 min and prefixed with 2.5% glutaraldehyde. The fixed cells were washed three times for 10 min with 100 mM phosphate buffer, post‐fixed for 3 h in 1% osmium tetraoxide and dehydrated through a series of ethanol gradient solutions. The samples were then coated with gold and observed on a scanning electron microscope (S‐3400, Hitachi, Japan). For transmission electron microscopy (TEM) observation, the samples were embedded in Epon 812, sectioned using an ultramicrotome and imaged under a transmission electron microscope (H‐7650, Hitachi, Japan). The negative control (CK1) consisted of Fon conidial suspensions mixed without myriocin, and the positive control (CK2) consisted of conidial suspensions mixed with amphotericin B at a final concentration of 1.25 μg ml⁻¹. The experiment was repeated three times, and each treatment consisted of three replicates.

Analysis of Fon membrane integrity

According to the methods described by Sun et al. (2015), the effect of myriocin on Fon membrane integrity was analysed by using LIVE/DEAD BacLight™ Viability Kit (MA, USA) under confocal laser scanning microscopy (CLSM). SYTO 9 stains both live and dead cells, whereas propidium iodide (PI) permeates cells only when the membrane is broken, reducing the SYTO 9 fluorescence when both dyes are present. Thus, cells with intact membranes are stained by SYTO 9 and emit green fluorescence, whereas the cells with damaged membranes are stained by PI and emit red fluorescence (Liu et al., 2017). The Fon conidial suspensions were mixed with myriocin at final concentrations of MIC (1.25 μg ml⁻¹), 2MIC (2.50 μg ml⁻¹), 4MIC (5.0 μg ml⁻¹) and 8MIC (10 μg ml⁻¹), and the mixtures were incubated for 12 h. The mixtures incubated for 12 h without myriocin was considered the negative control (CK1), and the mixtures with amphotericin B (final concentrations: 1.25 μg ml⁻¹) was considered the positive control (CK2). The incubated conidial suspensions were centrifuged at 8000 r.p.m. for 5 min at 4°C. Then, the conidia were resuspended in phosphate‐buffered saline (PBS). Subsequently, 1.5 µl of SYTO 9 and 1.5 µl of PI were added, and the mixture was incubated for 30 min at 25 ± 2°C in the dark. Afterwards, the mixtures were centrifuged. The localization of fluorescent dye in the conidia was visualized by using a confocal laser scanning microscope (TCS SP8, Leica, Germany) at 63× and excitation wavelengths of 488 nm (green conidia) and 535 nm (red conidia).

Determination of cell membrane fluidity

The fluorescence polarization (P) and anisotropy (r) were determined by using 1,6‐diphenyl‐1,3,5‐hexatriene (DPH) (Wang et al., 2019). The Fon conidial suspensions were incubated for 12 h with myriocin at final concentrations of MIC (1.25 μg ml⁻¹), 2MIC (2.50 μg ml⁻¹), 4MIC (5.0 μg ml⁻¹) and 8MIC (10 μg ml⁻¹). The mixtures incubated for 12 h without myriocin was considered the control (CK1). Afterwards, the mixtures were centrifuged at 8000 r.p.m. for 5 min, washed three times and resuspended in PBS. Then, 5 μl of 100 μmol l⁻¹ DPH solution was mixed with 10 ml of the conidial suspension. After incubation at 30°C for 30 min, fluorescence polarization and anisotropy were measured at excitation/emission wavelengths λ_ex/λ_em = 360 nm/427 nm by using a fluorescence spectrophotometer (F‐7000, Hitachi, Japan) equipped with a polarizer.

Preparation of calcein‐loaded liposomes and leakage assay

Phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol (PI) and ergosterol were purchased from Sigma‐Aldrich, Shanghai, China. One type of liposome was prepared as follows: PC/PE/PI/ergosterol with a mass ratio of 5:4:1:2 to mimic the fungal cell membrane was mixed (Sun et al., 2015). Calcein‐loaded liposomes were prepared according to methods described by Lee & Lee, (2015). Phospholipids were dissolved in chloroform and methanol at a ratio of 5:4:1:2. After vacuum evaporation and drying overnight, 1.5 ml of dye buffer (70 mM calcein, 10 mM Tris, 150 mM NaCl, and 0.1 mM EDTA, pH 7.4) was added to the dried sample. Then, the mixture was freeze‐thawed in liquid nitrogen until homogeneous and passed through 0.22‐μm polycarbonate filters. The uncoated calcein was removed by gel filtration through a Sephadex G‐50 column and elution with phosphate buffer. The calcein‐loaded liposomes were diluted to 90 mM for further testing. For the calcein leakage assay, a suspension of liposomes containing calcein was treated with myriocin at final concentrations of MIC (1.25 μg ml⁻¹), 2MIC (2.50 μg ml⁻¹), 4MIC (5.0 μg ml⁻¹) and 8MIC (10 μg ml⁻¹). The sample without myriocin was considered the control (CK1). The leakage of calcein from liposomes was measured by using a fluorescence spectrophotometer (F‐7000, Hitachi, Japan) at an excitation wavelength of 485 nm and an emission wavelength of 515 nm. Then, 20 µl of 10% Triton X‐100 (pH 7.4) was added into the mixture at 12 min to determine the maximum fluorescence intensity of 100% calcein leakage. The percentage of calcein leakage induced by myriocin was calculated according to the equation calcein leakage (%) = 100 × (F–F ₀)/(F_t–F ₀), where F represents the fluorescence intensity acquired after the addition of myriocin, and F₀ and F_t represent the fluorescence intensities without myriocin and with Triton X‐100 respectively (Bakonyi et al., 2018). The statistical data are presented as mean ± standard deviation of three replicates.

Illumina library and RNA‐Seq

According to the returned results, treatment including myriocin at MIC and 8MIC were selected for further study, and 0 μg ml⁻¹ myriocin was considered the control (CK1). Three biological replicates were prepared for the treatment groups and the control. Fon conidia were exposed to various myriocin concentrations for 12 h; the mixtures were centrifuged at 8000 r.p.m. for 10 min and then frozen in liquid nitrogen immediately for RNA extraction and RNA sequencing (RNA‐Seq).

The total RNA was extracted from the samples by using TRIzol^® Reagent according the manufacturer’s instructions (Invitrogen, USA) and was treated with DNase I (TaKara, Kofu, Japan) to remove genomic the DNA. Then, the quality and concentration of the RNA were determined by using a 2100 Bioanalyzer (Agilent, Palo Alto, CA, USA) and a NanoDrop spectrophotometer (ND‐2000, Thermo, USA). Afterwards, high‐quality total RNA samples (OD_260/230 ≥ 2.0, OD_260/280 = 1.8–2.2, RIN ≥ 6.5, 28S: 18S ≥ 1.0, > 2 μg) were used to construct the sequencing libraries.

The RNA‐Seq library was prepared following the TruSeq^TM RNA sample preparation kit from Illumina (San Diego, CA). First‐strand cDNA was synthesized by using random hexamer primers, and double‐stranded cDNA was subsequently synthesized by using a SuperScript Double‐Strand cDNA Synthesis kit (Invitrogen, Carlsbad, CA, USA). PCR amplification was performed by using Phusion DNA polymerase (NEB) for 15 PCR cycles. The PCR amplification products were quantified by using TBS380 (PicoGreen), and the pair‐end RNA sequencing library was sequenced by using the Illumina HiSeq 4000 platform with 2 × 150 bp reads.

The raw paired‐end reads were trimmed and controlled for quality by using SeqPrep (https://github.com/jstjohn/SeqPrep) and Sickle (https://github.com/najoshi/sickle). Then, the clean reads were separately aligned to the reference genome with orientation mode by using TopHat software, and the mapped results were then queried by using BLAST against the database (ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/2.9.0/). The FPKM method was used to compare the differences in gene expression among samples. The software package EdgeR was used to determine the fold change of transcripts and to screen differentially expressed genes (DEGs) (Song et al., 2019a). The criteria of significant differential expression were fold change ≥ 2 or ≤ 0.5 and adjusted P‐value ≤ 0.05 (Jian et al., 2017). Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) were used for annotation (Wang et al., 2016).

Protein extraction and LC‐MS/MS analysis

Treatment with myriocin at MIC (1.25 μg ml⁻¹) or 8MIC (10 μg ml⁻¹) was used for the proteomics study, and 0 μg ml⁻¹ myriocin was considered the control (CK1). Fon conidia were exposed to various myriocin concentrations for 12 h; the mixtures were centrifuged at 8000 r.p.m. for 10 min and then frozen in liquid nitrogen immediately for label‐free quantitative proteomic analysis. Each treatment was performed in triplicate for proteomic analysis. The proteins from samples were extracted by using an extraction buffer (1% SDS, 200 mM dithiothreitol, 50 mM Tris‐HCl and 1% protease inhibitor, pH 8.8). Then, the samples were centrifuged at 12 000 r.p.m. at 4°C for 20 min, and the supernatants were precipitated with precooled acetone at −20°C overnight. Afterwards, the precipitate was collected and washed twice with 90% precooled acetone. Then, the protein precipitate was redissolved in lysis buffer (8 M urea, 1% SDS and 1% protease inhibitor). The lysates were centrifuged, and the protein concentration of the supernatant was quantified by using the bicinchoninic acid (BCA) method. Protein digestion was performed according to standard procedure previously reported (Morrin et al., 2019). Briefly, the appropriate amount of trichloroethyl phosphate was added to each sample tube containing 150 μg of protein and incubated at 37°C for 60 min. The protein sample was resuspended in 150 μl of buffer (100 mM TEAB). Trypsin was added to approximately 150 μg of protein for each sample at an enzyme:protein ratio of 1:50 and digested at 37°C overnight. The peptides were desalted with OASIS^® HLB μElution plates and vacuum dried. Peptide concentrations were determined by using a peptide quantification kit (Thermo, Cat. 23275). LC‐MS/MS analysis was performed on a Q Exactive™ mass spectrometer (MS) (1200‐6430A, Agilent, USA) (Zhong et al., 2018). The peptides were loaded onto a C18‐reverse phase column (75 μm × 25 cm, Thermo, USA) with buffer A (2% acetonitrile and 0.1% formic acid) and separated with a linear gradient of buffer B (80% acetonitrile and 0.1% formic acid) at a flow rate of 300 μL/min. Afterwards, the peptides were subjected to nanoelectrospray ionization followed by tandem MS in Q Exactive (Thermo, USA). Intact peptides were detected in the orbitrap at a resolution of 70 000. Peptides were selected for MS/MS using a high‐energy collisional dissociation (HCD) setting of 20. Ion fragments were identified in the orbitrap at a resolution of 17 500.

Proteomics data processing

MS/MS spectra were searched using ProteomeDiscover™ Software 2.2 against the Fusarium oxysporum database (27347s, 20190328). Trypsin digestion was selected with cleavage specificity allowing up to two missing cleavages. Carbamidomethyl on cysteine was assigned as a fixed modification, and oxidation on methionine and protein N‐terminal acetylation were specified as variable modifications. Differentially expressed proteins (DEPs) between the two samples were screened with fold change > 1.2 or < 0.83 and P‐value < 0.05. The analysis of membrane‐related DEPs was performed, including protein functional annotation, GO functional annotation and KEGG pathway enrichment analysis.

The correlation of DEG‐DEG or DEP‐DEP was revealed by using cluster heat maps and network interactions, and the relationship between DEGs and their corresponding DEPs was explored.

RT‐qPCR

To verify the RNA‐Seq results, 9 DEGs from the RNA‐Seq analysis were selected, and RT‐qPCR was performed to confirm the expression changes of these 9 DEGs. RT‐qPCR was performed on the RNA samples used for RNA‐seq. Total cellular RNA was quantified at OD₂₆₀ and OD₂₈₀ by using a SMA400 microspectrophotometer (Merinton, Beijing, China). The single‐stranded cDNA of total RNA was generated by using RevertAid Premium Reverse Transcriptase (EP0733, Thermo Scientific, MA, USA), and the cDNA samples were stored at −20°C for later use (Hu et al., 2019a). Quantitative real‐time PCR was performed on a StepOne Plus Real‐Time PCR instrument (ABI, Waltham, MA, USA). The primers used for RT‐qPCR are listed in Table S1, and the 18S rRNA gene was used as the internal control to which normalization of the mRNA abundance between samples was performed. Each qRT‐PCR reaction system (20 μl) included 10 μl of 2 × SYBR Green qPCR Master Mix, 0.4 μl of forward primer (10 μM), 0.4 μl of reverse primer (10 μM), 2 μl of template (cDNA) and 7.2 μl of ddH₂O. PCR was performed as follows: 3 min at 95°C, 45 cycles of 5 s at 95°C and 30 s at 60°C. A melting curve was generated to detect the specificity of the amplification. A relative quantitative analysis was calculated by using the 2^−ΔΔCt method. Each measurement was performed in triplicate.

Assay of SPT activity and molecular docking

The myriocin and conidia solution were mixed at different concentrations (MIC, 1.25 μg ml⁻¹; 2MIC, 2.50 μg ml⁻¹; 4MIC, 5.0 μg ml⁻¹; and 8MIC, 10 μg ml⁻¹; CK1, 0 μg ml⁻¹). After 12‐h incubation, the mixtures were centrifuged at 8000 r.p.m. for 10 min, washed three times and resuspended in PBS. The conidia were cryo‐crushed by using an ultrasonic cell disrupter (JY‐92IIDN, Scientz, China) according to the following procedure: 45 cycles of 3‐s ultrasonic time and 3‐s interval time. Then, the supernatants were centrifuged at 8000 r.p.m. for 10 min and transferred onto 96‐microwell plates. The SPT in the supernatant was extracted by using an enzyme‐linked immunosorbent assay kit (ELISA, Blobase, Zhangqiu, China) according to the manufacturer’s protocol, and the activity of SPT was measured by using a microplate reader (PT‐3502, Potenov, Tongzhou, China) (Hu et al., 2019b). Each treatment was performed in triplicates.

Homology modelling is a computational procedure that aims to build a three‐dimensional (3D) protein model (for unknown crystallographic structures) by using known structures of homologous protein as templates. In this study, we used SWISS‐MODEL to derive the 3D structure of SPT (XP_018249811.1) from SPT amino acid sequences (query sequence). The amino acid sequence of the SPT in this study was acquired through NCBI and automatically compared. The template of the homology modelling was from the SPT crystal structure (PDB ID: 4bmk.1. A) in the Protein Data Bank (PDB). The homology model was constructed through the SWISS‐MODEL (http://swissmodel.expasy.org) server (Waterhouse et al., 2018).

For molecular docking, the 3D structure of myriocin was downloaded from the PubChem (https://pubchem.ncbi.nlm.nih.gov/) website. The molecular docking of myriocin (ligand) to the 3D protein model of SPT (XP_018249811.1) was performed by using UCSF DOCK 6 (Xie and Minh, 2019). Both the protein (SPT) and ligand (myriocin) molecules were prepared prior to docking analysis by using the Prepare Protein and Prepare Ligand steps, respectively. The generation of the 3D conformation of myriocin was attained by using the Prepare Ligands tool, and the 3D conformation of myriocin with the lowest grid score was considered the 'best docking model'. Afterwards, protein–ligand (myriocin) interactions were analysed. A diagram of the docking model was made by using the FREE MAESTRO module in Schrodinger 2018 Software and MS PowerPoint (PPT) 2013. The docking process was performed by using the default parameters.

Statistical analysis

spss 26.0 software was used for the statistical analysis of the data. The data are expressed as mean ± standard error from three biological replicates. The t‐test was used for statistical analysis; P ≤ 0.05 was considered significant.

Conflict of interest

None declared.

Supporting information

Fig. S1. The proportions of differentially coloured of conidia treated with myriocin. MIC, 1.25 μg ml⁻¹ myriocin; 2MIC, 2.5 μg ml⁻¹ myriocin; 4MIC, 5 μg ml⁻¹ myriocin; 8MIC, 10 μg ml⁻¹ myriocin; CK1, 0 μg ml⁻¹ myriocin as negative control; CK2, 1.25 μg ml⁻¹ amphotericin B as positive control.

Table S1. Genes and primers used in RT‐qPCR validation.

Table S2. The information and pathway names of 27 DEPs.

Table S3. The information and pathway names of 26 DEGs.

Table S4. The KEGG pathway category and ID and name of pathway.

Table S5. DEGs and its corresponding DEPs.

Table S6. Names and spots of 26 DEGs and NCBI accessions and spots of 27 DEPs.

Microbial Biotechnology (2021) 14(2), 517–534