ACR3 as a Key Regulator Antimicrobial Lipopeptide Synthesis and Efflux in Bacillus velezensis HN‐1

ABSTRACT

Bacillus velezensis is known for producing and secreting antimicrobial lipopeptides, yet the regulatory mechanisms behind these processes remain unclear. In this study, we constructed arsenite transporter protein gene (acr3)‐deficient strains (HN‐1ΔQ acr3) and complementary strains (HN‐1ΔC acr3) to elucidate the roles of ACR3 in antimicrobial lipopeptide synthesis and efflux. Our findings demonstrate that ACR3 dysfunction compromised swarming and biofilm formation, completely abolishing the antifungal capabilities of the HN‐1ΔQ acr3 strains. Transcriptomic analysis and reverse transcription‐quantitative PCR revealed significant downregulation of lipopeptide biosynthesis genes, including sfp, fenD and bmyC, in HN‐1ΔQ acr3. Notably, the addition of purified ACR3 protein restored the expression of lipopeptide biosynthesis genes and partially rescued antifungal activities in the HN‐1ΔQ acr3 strains. Furthermore, yeast two‐hybrid assays indicated an interaction between ACR3 and proteins involved in cell survival, substrate uptake and antifungal activities, such as ribose ABC transporter substrate‐binding protein (RbsB), fengycin synthetases and phosphopantetheine transferases. In parallel, our findings indicate that ACR3 interacts with the ABC transporter pathway to tune antifungal lipopeptide transport in Bacillus velezensis . Collectively, these results suggest that ACR3 functions as a crucial regulator, mediating both the synthesis and efflux of lipopeptides in B. velezensis . These research establish an experimental foundation for elucidating the mechanisms of antimicrobial compound synthesis and secretion in B. velezensis and provide a theoretical basis for the optimisation and engineering of high‐yield biocontrol strains.

Lipopeptides are natural antibiotics produced by Bacillus velezensis . Mutation of acr3 led to the complete loss of antimicrobial activity. Following the biological assays for validation, we found ACR3 protein may serve as a master switch of lipopeptides synthesis regulation, and control transmembrane transport of lipopeptides by the ABC pathway.

1. Introduction

Chemical pesticides are the most effective and widely used agents for controlling plant pests and pathogens. Currently, they are integrated into plant protection strategies to enhance the preservation of crops, forestry and animal husbandry, forming an essential backbone for modern agricultural productivity (Afridi et al. 2022). However, the long‐term, large‐scale application of chemical pesticides poses severe threats to soil microfauna, floral biodiversity and the survival of beneficial insect pollinators (Hou and Wang 2022). Furthermore, the intensive use of the same type of chemical pesticides contributes to the development of resistance among phytopathogenic microbes and insect pests (Gui et al. 2022). This phenomenon creates a vicious cycle, increasing the demand for pesticide application and negatively impacting the environment while disrupting the ecological balance (Pathak et al. 2022). Residues of these chemicals often remain in crops and can leach into water sources, posing direct and indirect risks to humans and animals. This situation underscores the urgent need for research into eco‐friendly biopesticides derived from plants or microorganisms to support sustainable agriculture, food security and the preservation of biodiversity (Trivedi et al. 2020; Zvinavashe et al. 2019).

Bacillus is a widely distributed genus of gram‐positive bacteria, characterised by a rich diversity of physiological traits that are integral to the microbial communities in soil and plant roots. Antimicrobial substances produced by Bacillus species have demonstrated inhibitory effects against plant‐pathogenic microbes. For instance, non‐ribosomal pathway‐associated lipopeptide antibiotics, including surfactin, iturin and fengycin (Abriouel et al. 2011), have shown high efficacy against various phytopathogenic viruses, bacteria and fungi (Saggese et al. 2018; Zeriouh et al. 2014). Researchers have indicated that the application of iturin induces cell death in fungi by disrupting the integrity of the cytoplasmic membrane, leading to vesicles formation and the release of lysozymes and macromolecules and the degradation of phospholipids (Xiao et al. 2021). A prevailing theory suggests that these secondary metabolites induce cytoplasmic exocytosis by interacting with the membranes of pathogenic bacteria, creating ion leakage channels (Pérez‐García et al. 2011). Despite their potential, the low yield and high production cost of antimicrobial lipopeptides from B. velezensis hinder their competitiveness for large‐scale production (Du and Shen 2001; Rabbee et al. 2019). Thus, there is a pressing need for efficient, reliable strains capable of producing antimicrobial lipopeptides. Gaining insights into the regulatory mechanisms governing the synthesis and secretion of antimicrobial lipopeptides in Bacillus could significantly improve strain development, increase production and enhance the efficient utilisation of these compounds for sustainable pest and pathogen management.

During bacterial growth and development, interactions with the external environment occur, and small molecules such as cyclic dipeptide lactones can serve as signalling molecules that cross the cell membrane to promote physiological processes, including nutrient acquisition, environmental adaptation and virulence (Balachandra et al. 2021; Silpa and Rupachandra 2020). The cell membrane acts as a barrier to the intracellular space, facilitating the exchange of substances between the intracellular and extracellular environments. Transport across the cell membrane can occur via passive transport through ion channels, facilitated diffusion, or active transport that requires energy (Dergunov and Baserova 2022; van den Berg et al. 2015). In gram‐negative bacteria, selective secretion systems, including types I–VI and type IX, mediate protein secretion via exocytosis (Costa et al. 2015). The efflux of substances typically requires passage through both the cytoplasmic membrane and the peptidoglycan layer. The most common transmembrane efflux channels in bacteria are the type IV and VII secretion systems, along with Sec, Tat, ABC and Com transporter pathways. The ABC transporter pathway is notable for transporting substrates unidirectionally, including organic substances such as amino acids, proteins, sugars, peptides and antibiotics, by using the energy generated from ATP hydrolysis (Orelle and Jault 2016; Wilkens 2015). The ABC transporter superfamily plays a critical role in nutrient uptake, secondary metabolite efflux and biofilm formation across all bacterial growth stages (Hollenstein et al. 2007; Moussatova et al. 2008). However, the overlapping functions of ABC transporter proteins complicate the identification of specific roles. Moreover, the ABC transporter regulatory system often complements the two‐component regulatory system, contributing to resistance against antimicrobial lipopeptides. Previous studies on prokaryotes, including Bacillus, indicate that surfactin activity is dependent on the membrane dipole potential and that increases in cardiolipin within the cell membrane can protect Bacillus subtilis from surfactin‐induced permeabilization (Ostroumova et al. 2010; Pinkas et al. 2020). Nevertheless, the mechanisms by which lipopeptides achieve transmembrane transport and the corresponding secretion pathways in Bacillus remain unclear.

In this study, we constructed arsenite transporter protein gene (acr3)‐deficient strains (HN‐1ΔQ acr3) and complementary strains (HN‐1ΔC acr3) to investigate the direct and indirect influences of ACR3 on antimicrobial properties and the expression of lipopeptide biosynthesis genes in Bacillus velezensis . We demonstrated that the mutation of acr3 abolished the haemolytic and antagonistic capabilities of the HN‐1ΔQ acr3 strains against fungi, while enhancing the accumulation of intracellular bacillomycin D. Furthermore, we observed a significant downregulation of genes associated with lipopeptide biosynthesis in the HN‐1ΔQ acr3 strains. Notably, treatment with purified ACR3 partially restored antibacterial activity and the expression of lipopeptide‐related genes in the HN‐1ΔQ acr3 strains. We also identified RbsB from the ABC transporter superfamily as a reciprocal protein interacting with ACR3. Collectively, these results suggest that ACR3 functions as a crucial regulator, mediating both the synthesis and efflux of lipopeptides in B. velezensis HN‐1. These findings provide a foundational basis for future studies on the efflux pathways of lipopeptides from Bacillus species, highlighting their potential as effective biomolecules for developing new biopesticide formulations against pathogenic bacteria while mitigating the risk of stimulating antibiotic resistance.

2. Results

2.1. Identifications of the acr3 Gene in B. velezensis HN‐1

A mutant library of B. velezensis strain HN‐1 was constructed using TnYLB‐1 transposon, and mutants exhibiting a loss of haemolytic activity were screened. The mutant sequence of these strains was subsequently determined through reverse PCR (Figure S1A). BLAST comparison revealed that the amino acid sequence of the identified gene showed 99% homology with the arsenic‐resistant protein ACR3 of B. velezensis FZB42 (Figure S1B). Further analysis using an amino acid phylogenetic tree indicated that ACR3 possesses a conserved site but also exhibits some mutated structural domains when compared to other functionally similar membrane transport‐associated proteins (Figure S2). Additionally, predictions made using DeepTMHMM (https://dtu.biolib.com/DeepTMHMM/) indicated that ACR3 contains 10 transmembrane structural domains (Figure S2C).

To further investigate the function of ACR3, we performed a targeted insertion resulting in the strain HN‐1ΔQ acr3 and constructed the complementary strain HN‐1ΔC acr3. To evaluate the influence of ACR3 on arsenic ion stress tolerance in B. velezensis, we tested the strains in media supplemented with varying concentrations of arsenic (III). The results showed that the HN‐1ΔQ acr3 was slightly sensitive to 40 μM of arsenic (III) exposure (Figure 1A), indicating that ACR3 partly contributes to arsenic (III) stress tolerance in B. velezensis .

FIGURE 1.

Characteristics of ACR3 on Bacillus velezensis HN‐1 phenotypes. (A) ACR3 function analysis on arsenic ion (III) resistance test. (B, C) Biofilm formation determination. (D) Swarming motility test. (E–G) Determination of oil drainage capacity of intracellular and extracellular crude extracts. Experimental data were analysed through one‐way ANOVA. Asterisks indicate a significant difference (***p ≤ 0.001, *p < 0.05), and ns indicates no significant difference by one‐way ANOVA.

2.2. Characteristics of ACR3 on B. velezensis HN‐1 Phenotypes

A targeted insertion mutant of B. velezensis demonstrated that ACR3 plays a weak role in regulating arsenic ion stress tolerance. However, this mutation impacted various phenotypic characteristics beyond stress tolerance. In biofilm formation assays, a notable difference was observed between the HN‐1 wild‐type strain (WT) and the HN‐1ΔQ acr3 mutant. At 16 h, all strains initiated biofilm formation, creating a thin, non‐folding smooth layer. By 24 h, the WT strain HN‐1 and the complementary strain HN‐1ΔC acr3 had developed folded, smooth biofilms, whereas the HN‐1ΔQ acr3 strain failed to form similar folds. This trend continued over time, with the HN‐1 strain and HN‐1ΔC acr3 strain producing thick, folded biofilms by 48 h, while the mutant strain HN‐1ΔQ acr3 formed noticeably thinner biofilms (Figure 1B,C). Additionally, the absorbance recorded at 570 nm, indicative of biofilm mass, was significantly lower for the HN‐1ΔQ acr3 mutant compared to the WT strain. These results suggest that ACR3 plays a critical role in promoting biofilm formation in B. velezensis . The reduced biofilm‐forming capacity of the HN‐1ΔQ acr3 mutant highlights its importance in regulating biofilm‐associated phenotypes.

In the swarming motility experiments, notable differences were observed between the WT HN‐1 strain and the HN‐1ΔQ acr3 mutant in their ability to exhibit swarming movement on 0.7% agar Luria Bertani (LB) medium (Figure 1D). The WT strain HN‐1 and the complementary strain HN‐1ΔQ acr3 showed rapid expansion across the surface of the medium, spreading over the entire bottom of the dish within a short time. In contrast, the HN‐1ΔQ acr3 strain displayed severely impaired motility, growing very slowly and remaining in a punctiform distribution, with no visible signs of expansion. We also analysed the active compounds to determine the quality of extracellular and intracellular extracts using n‐butanol extraction. The results showed that the weight of the extracellular extract of the HN‐1ΔQ acr3 strain was only 29.8% but the weight of the intracellular extract was 12.9‐fold compared to WT strain HN‐1 (Figure S3A). In summary, these findings suggest that ACR3 plays a crucial role in regulating swarming motility in B. velezensis . Moreover, the observed shift in the balance between extracellular and intracellular components in HN‐1ΔQ acr3 indicates that ACR3 is essential for proper cellular efflux, which could affect the secretion of key metabolites and overall cellular regulation.

To further explore the role of ACR3 in B. velezensis phenotypic traits, we investigated the haemolytic activity and oil drainage capacity in HN‐1ΔQ acr3. The results showed that the fermentation broth extract of the HN‐1 exhibited significant haemolytic activity on Columbia blood plate, with a haemolytic zone of 12.26 ± 0.16 mm (Figure S3B). The complementary strain HN‐1ΔC acr3 also showed haemolytic activity, though with a slightly small zone of 10.60 ± 0.26 mm, while the intracellular extract displayed no haemolytic activity. The HN‐1ΔQ acr3 strain exhibited negligible haemolytic activity, both intracellularly and extracellularly. Furthermore, oil discharge assays, which offer a more sensitive measure of haemolytic activity, confirmed these findings. The diameter of the oil discharge circle for the WT fermentation broth extract was 3.5 cm, whereas that of the HN‐1ΔQ acr3 extract was significantly smaller at 0.7 cm. Notably, the oil discharge circles of the intracellular extracts of both strains measured 0.7 cm (Figure 1E–G). Taken together, these results highlight the multifaceted role of ACR3 in biofilm formation, swarming motility, active compounds efflux, haemolytic activity and oil drainage capacity in B. velezensis .

2.3. Deletion of acr3 Hinders Active Component Efflux in B. velezensis HN‐1 Against Colletotrichum fructicola

The deletion of acr3 significantly impacted the efflux of active components in B. velezensis HN‐1, as demonstrated by its reduced antagonistic activity against C. fructicola. The fermentation broth and intracellular components of strains HN‐1, HN‐1ΔQ acr3 and HN‐1ΔC acr3 were extracted using the n‐butanol extraction method and assessed for their ability to inhibit C. fructicola HD‐1 growth (Figure 2A,C). The wild‐type strain HN‐1 exhibited a robust inhibition zone of 30.3 ± 0.6 mm, whereas the intracellular extract had a smaller inhibition zone of 17.0 ± 0.7 mm. In contrast, the HN‐1ΔQ acr3 showed a reduced inhibition zone, yielding a diameter of 21.7 ± 0.4 mm and the intracellular extract showing 17.2 ± 2.2 mm. The complementary strain HN‐1ΔC acr3 restored inhibition to a level similar to the wild‐type, with inhibition zones of 23.1 ± 1.0 mm (fermentation broth) and 14.2 ± 0.8 mm (intracellular).

FIGURE 2.

ACR3 hinders active component efflux in Bacillus velezensis HN‐1. (A) Measurement of inhibition zones of Colletotrichum fructicola HD‐1 antagonism by intra‐ and extracellular active substances from B. velezensis HN‐1. (B) Measurement of HD‐1 EC₅₀ by intra‐ and extracellular active substances; EE, extracellular extracts; IE, intracellular extract. (C) Inhibitory effect of B. velezensis intracellular extracts on HD‐1. (D) The extracellular concentrations of bacillomycin D produced by the wild‐type (HN‐1), mutant (HN‐1ΔQ acr3), and complemented strains (HN‐1ΔC acr3) were quantified using high‐performance liquid chromatography. Significant difference, ***p < 0.001. (E) Effects of crude extracts of B. velezensis on the morphology of HD‐1 mycelium.

The EC₅₀ values of HN‐1 against HD‐1 were 53 μg/mL for extracellular extracts and 2293 μg/mL for intracellular extracts. In comparison, the HN‐1ΔQ acr3 had an EC₅₀ of 1709 μg/mL for extracellular extracts and 888 μg/mL for intracellular extracts, representing a 2.6‐fold increase. The EC₅₀ values for the complemented strain HN‐1ΔC acr3 were 54 μg/mL (extracellular) and 2863 μg/mL (intracellular), similar to the wild‐type (Figure 2B). The crude fermentation extracts from the wild‐type strain HN‐1, the mutant strain (HN‐1ΔQ acr3), and the complemented strain (HN‐1ΔC acr3) were analysed using semi‐preparative HPLC. Bacillomycin D, previously isolated and preserved in our laboratory, was used as a reference standard. The retention time of bacillomycin D was observed at approximately 42.5–43 min. The results demonstrated that, based on a 20 μL injection volume, the amount of bacillomycin D in the wild‐type strain was 77.0 ± 6.5 μg. In contrast, the mutant strain HN‐1ΔQacr3 yielded only 1.9 ± 0.7 μg, representing a drastic reduction of approximately 97.5% compared to the wild‐type. The complemented strain HN‐1ΔC acr3 showed a trend of recovery, with an amount of 7.3 ± 1.1 μg. These findings confirm that the deletion of the acr3 gene significantly impairs the production of antifungal active components, predominantly bacillomycin D, thereby demonstrating that acr3 is essentially involved in regulating the biosynthesis of lipopeptides (Figure 2D). After treatment with the HN‐1 strain, the hyphae of C. fructicola exhibited extensive deformation and swelling, whereas no such morphological abnormalities were observed following treatment with the mutant strain (Figure 2E). These findings demonstrate that acr3 is essential for the efficient export of antifungal compounds in B. velezensis . The deletion of acr3 disrupts the balance of active component efflux, leading to a significant reduction in the extracellular inhibitory activity against C. fructicola . Collectively, the marked increase in bacillomycin D accumulation within the mutant strain further highlights the role of acr3 in regulating intracellular lipopeptide levels. This suggests that acr3 not only facilitates the secretion of antifungal compounds but also prevents their excessive intracellular accumulation, maintaining a critical balance in the cellular export processes of B. velezensis .

2.4. ACR3 Involvement in Lipopeptide Synthesis Pathway in B. velezensis HN‐1

To further explore this phenomenon, reverse transcription‐quantitative real‐time PCR (RT‐qPCR) analysis was conducted on genes related to lipopeptide synthesis, swarming motility and biofilm formation. The transcriptional profiles revealed a significant downregulation of key regulatory genes in HN‐1ΔQ acr3 compared to the wild‐type strain HN‐1. The sfp, srfAA and srfAC genes, which regulate surfactin synthesis, were highly downregulated, as were fenD, fenA, bmyC, bmyD and bmyA—genes responsible for fengycin and iturin production. Additionally, genes involved in fatty acid chain synthesis (comP, ycsE, comA, comX, comQ, comK, comS) and those regulating swarming motility and biofilm formation (flgB, degU, degS, abrB, rapA, sigH, kinA) also showed significant downregulation in the mutant strain (Figure 3). Importantly, comP, a membrane protein responsible for the phosphorylation of comA and the subsequent regulation of srfA for surfactin synthesis, was also notably downregulated. This collective gene expression analysis supports the hypothesis that ACR3 plays a vital regulatory role in lipopeptide synthesis, motility and biofilm formation in B. velezensis by facilitating the efflux of secondary metabolites and ensuring proper gene regulation.

FIGURE 3.

ACR3 regulates the expression of genes related to antimicrobial activity substances. Experimental data were analysed through one‐way ANOVA. For multiple comparison of means and evaluation of significance by Duncan's test, asterisks indicate a significant difference (***p ≤ 0.001).

2.5. ACR3 Restores Antimicrobial Activity of HN‐1ΔQ acr3 In Vitro

To further investigate the role of ACR3 in regulating antimicrobial activity, we examined whether the exogenous expression and purification of ACR3 could restore the lost antimicrobial activity in the HN‐1ΔQ acr3. Purified ACR3 protein at various concentrations was co‐cultured with the mutant strains. The results showed that at a concentration of 10 μg/mL, the antimicrobial activity of the mutant strain HN‐1ΔQ acr3 was restored, exhibiting inhibition zones of approximately 1.18 cm compared to the untreated HN‐1ΔQ acr3 strain. Interestingly, higher concentrations of ACR3 did not restore the antimicrobial activity as expected, probably due to the potential inhibition of active substance synthesis within HN‐1 cells (Figure 4A,B).

FIGURE 4.

ACR3 restores antimicrobial activity of Bacillus velezensis HN‐1ΔQ acr3 in vitro. (A, B) Varying concentrations of recombinant fusion proteins and HN‐1ΔQ acr3 co‐cultures were tested for Colletotrichum fructicola HD‐1 antagonistic activity. CK: The purified ACR3 protein at 100 μg/mL. (C) Reverse transcription‐quantitative PCR was used to determine the expression of genes involved in the regulation of the synthesis of antimicrobial activity substances. Experimental data were analysed through one‐way ANOVA. Asterisks indicate a significant difference (***p ≤ 0.001) by Duncan's test.

Additionally, we analysed the expression levels of genes involved in lipopeptide biosynthesis, such as srfAC, comP, fenD and bmyC. The results showed significant upregulation of these genes in HN‐1ΔQ acr3 when co‐cultured with 10 μg/mL ACR3. Conversely, co‐culturing with 1 mg/mL of ACR3 resulted in downregulation of these lipopeptide biosynthesis genes, suggesting that high concentrations of ACR3 may inhibit their biosynthesis (Figure 4C). Together, these findings indicated that ACR3 serves as a crucial regulator of antimicrobial activity in B. velezensis , modulating lipopeptide synthesis pathways.

2.6. Transcriptomic, Proteomic and Metabolomic Analysis

To elucidate the molecular mechanisms underlying the function of the acr3 gene in B. velezensis , integrated transcriptomic, proteomic and metabolomic analysis was performed. Differentially expressed genes (DEGs), differentially expressed proteins (DEPs) and differentialy expressed metabolites (DEMs) were identified to delineate the global regulatory network influenced by acr3 deletion. For transcriptomic data, a false discovery rate (FDR) < 0.05 and a fold‐change ≥ 2 were applied, while proteomic and metabolomic analyses employed a t‐test (p < 0.05) and a fold‐change ≥ 2.

In Figure 5A,D, transcriptomic analysis identified 685 DEGs, comprising 275 significantly upregulated and 383 significantly downregulated transcripts. KEGG enrichment analysis revealed the functional implications of these DEGs. The most significantly enriched pathways include nonribosomal peptide structures, the phosphotransferase system (PTS), starch and sucrose metabolism and nucleotide sugar biosynthesis. This indicates profound perturbations in secondary metabolism, carbohydrate transport and central carbon metabolism. Notably, nonribosomal peptide biosynthesis, a key pathway for antimicrobial metabolite production, is highly enriched, suggesting altered secondary metabolism. Carbohydrate‐related pathways, such as the PTS and starch/sucrose metabolism, are also significantly affected, implying dysregulated sugar utilisation. ABC transporters and quorum sensing, which contain the largest number of DEGs, are additionally altered, reflecting broad changes in nutrient transport and cell–cell communication. Collectively, these results demonstrate that the acr3 mutation triggers widespread transcriptional reprogramming, primarily impacting secondary metabolism, carbohydrate utilisation and membrane transport in B. velezensis HN‐1.

FIGURE 5.

Integrated transcriptomic, proteomic and metabolomic analyses of the acr3‐deficient strain HN‐1ΔQ acr3. (A) Volcano plot of differentially expressed genes. The x‐axis represents the log₂ fold‐change (FC) of expression levels between groups, and the y‐axis indicates the −log₁₀(p _adjust) values. Red and green dots denote significantly upregulated and downregulated genes, respectively. (B) Volcano plot of differentially expressed proteins. Red and green dots represent significantly upregulated and downregulated proteins, respectively. (C) Volcano plot of differentially expressed metabolites. Each dot represents an individual metabolite; the x‐axis shows the log₂ fold‐change, and the y‐axis shows the −log₁₀(p value). The size of each dot reflects the VIP value in the OPLS‐DA model‐larger dots indicate higher VIP values and greater reliability of differential expressions. Red and blue dots indicate significantly upregulated and downregulated metabolites, respectively, while grey dots represent non‐significant metabolites. The top five metabolites, ranked by p‐value, are labelled. (D) KEGG enrichment analysis of differentially expressed genes in the acr3‐deficient strain. (E) KEGG enrichment analysis of differentially expressed proteins. (F) Classification of differentially expressed metabolite pathways. Distinct colours represent KEGG level 2 pathway categories, and bar lengths indicate the number of differential metabolites associated with each pathway.

The proteomic quantification further delineated the phenotypic differences at the protein level (Figure 5B,E). Proteins exceeding the significance thresholds are marked in red (upregulated) and green (downregulated). The analysis revealed 221 DEPs, with 119 being upregulated and 102 downregulated in the acr3 mutant. Subsequent KEGG pathway enrichment analysis for these DEPs identified the biological pathways most significantly impacted by the proteomic alterations. The most significantly enriched pathway is nonribosomal peptide structures, demonstrating profound dysregulation in secondary metabolite biosynthesis. Purine metabolism, mismatch repair and quorum sensing are also highly enriched, suggesting disruptions in nucleotide metabolism, DNA repair and intercellular communication. Moreover, the biosynthesis of antibiotics pathway contains the largest number of DEPs, further confirming a widespread perturbation of secondary metabolism. Additionally, ABC transporters and central carbon metabolism (glycolysis/gluconeogenesis) are also significantly altered, reflecting changes in nutrient uptake and membrane transport processes. Consistent with the transcriptomic findings, the ABC transporters pathway is reported to be significantly enriched among the DEPs as well. This concordance indicates that the transcriptional changes in transport‐related genes are translated to the protein level, validating the crucial role of ABC transporter‐related proteins in the regulatory network of acr3 in B. velezensis HN‐1.

The metabolomic profiling provided a downstream phenotypic snapshot (Figure 5C,F), the volcano plot visualises these DEMs, with point size representing VIP values, and the most significant metabolites driving group separation are highlighted. Significant downregulated and upregulated DEMs are coloured blue and red, respectively. A total of 821 DEMs were identified, including 389 upregulated and 432 downregulated metabolites. This bar plot presents the KEGG pathway enrichment analysis of significantly altered metabolites, with the bar length indicating the metabolite count. The most prominent finding was the predominant enrichment of biosynthetic pathways for complex secondary metabolites. This included a significant effect on the production of macrolide antibiotics, as evidenced by the strong enrichment for 12‐, 14‐ and 16‐membered macrolide biosynthesis. A concurrent, global activation of diverse antimicrobial synthesis was further underscored by enriched pathways for aminoglycosides (neomycin, kanamycin and gentamicin) and indole alkaloids, complemented by notable alterations in terpenoid and polyketide metabolism. In contrast, perturbations in core metabolic processes, such as amino acid metabolism (e.g., arginine and proline) and carbohydrate metabolism (e.g., amino sugar and nucleotide sugar metabolism) were present but less prominent. Crucially, a significant number of metabolites were annotated to the ABC transporters pathway, with key members like glycerol, uridine and oleandomycin being markedly downregulated. This indicates substantial disruptions in precursor availability for membrane synthesis (glycerol), nucleotide metabolism and biofilm processes (uridine) and antibiotic efflux (oleandomycin). Collectively, these results suggest that acr3 mutation may indirectly affect cell membrane formation and antibiotic efflux by regulating upstream genes or signalling pathways of ABC transporters.

Notably, integrative network analysis revealed that ACR3 probably functions as a pivotal node between lipopeptide biosynthesis and efflux, linking transporter activity with secondary metabolite regulation, providing mechanistic insight into how B. velezensis HN‐1 regulates antimicrobial lipopeptide synthesis and efflux. Detailed pathway annotations are provided in Figure S4.

2.7. ACR3 Interacts With Lipopeptide Biosynthesis and Efflux Protein to Tune Lipopeptide in B. velezensis HN‐1

To elucidate the role of ACR3 in regulating lipopeptide synthesis and efflux in B. velezensis HN‐1, a cDNA library of HN‐1 and two decoy vectors were constructed for a membrane yeast two‐hybrid (Y2H) assay. The vectors were screened through self‐activation and functional assays, leading to the selection of 47 positive clones after several rounds of screening on drop‐out medium. Sequence analysis of these clones identified several interacting proteins, including ribose ABC transporter permeases, cytochrome P450, type VII secretion system proteins, including the ribose ABC transporter permease RbsB, which is associated with lipopeptide efflux, and the non‐ribosomal peptide synthetases SrfAA, BmyC, FenC and Sfp, which are involved in lipopeptide synthesis. These interactions were further validated using pairwise yeast two‐hybrid assays, confirming strong interactions between ACR3 and the ribose ABC transporter substrate‐binding protein (RbsB), surfactin synthetase SrfAA, bacillomycin D synthetase BmyC, fengycin synthetase FenC and phosphopantetheinyl transferase Sfp (Figure S5). Furthermore, GST‐pulldown assays corroborated these findings by demonstrating robust interactions, thereby indicating that the ACR3 protein interacts with and regulates lipopeptide synthesis‐associated proteins (Figure 6A–C). Additionally, the expression of genes related to the ABC transporter superfamily was analysed. The results revealed a highly significant downregulation of these genes in the mutant strain HN‐1ΔQ acr3 compared to the wild‐type strain (Figure 6D).

FIGURE 6.

ACR3 interacts with RbsB. (A) Yeast two‐hybrid assay tested ACR3 interacted with RbsB. NubG‐ACR3 (pPR3‐N‐ACR3) and RbsB–Cub (pBT3‐STE‐RbsB) were grown on SD/−Leu−Trp (SD/−LW) and SD/−Leu−Trp−His−Ade (SD/−LWHA) plates containing 30 mM 3‐AT and X‐α‐gal. The co‐transformation of NubI and Cub was used as a positive control. (B) GST pull‐down validated protein interaction between ACR3 and RbsB in vitro. (C) Bimolecular fluorescence complementation (BiFC) assays verified the interaction of ACR3 and RbsB in Nicotiana benthamiana mesophyll cells. Agrobacterium harbouring ACR3 and RbsB expression vectors were co‐injected into N. benthamiana leaves and photographed by laser scanning confocal microscopy at 48 h post‐infiltration. Bar = 50 μm. (D) Expression of ABC transporter‐related genes in mutant strains was determined by reverse transcription‐quantitative PCR (RT‐qPCR). (E) ABC transporter‐related gene expression was assessed by RT‐qPCR after treatment with various concentrations of recombinant fusion proteins. Experimental data were analysed through one‐way ANOVA. For multiple comparison of means and evaluation of significant differences, asterisks indicate a significant difference (***p ≤ 0.001) by Duncan's test.

To further assess the impact of exogenous ACR3 on ABC transporter‐related gene expression, co‐cultures of the mutant strain HN‐1ΔQ acr3 with the recombinant ACR3 fusion protein were analysed at the transcriptional level. The expression of the ABC transporter genes opuAA, opuBC and opuE in the mutant strain treated with a low concentration (10 μg/mL) of ACR3 protein showed a highly significant upregulation, reaching nearly twice the expression level of the untreated mutant strain. Conversely, treatment with a high concentration (1 mg/mL) of ACR3 protein resulted in significant downregulation, with gene expression reduced to just 20% of that in the untreated mutant strain (Figure 6E). Combined with Figures S3 and S5, these results indicated that the ACR3 protein acts as a regulator of lipopeptide synthesis and efflux in B. velezensis HN‐1.

2.8. Lipopeptide Activity Was Suppressed by Broad‐Spectrum ABC Transporter Inhibitors in B. velezensis HN‐1

Tariquidar, a broad‐spectrum inhibitor, was used to assess its impact on the wild‐type HN‐1 strain. As the concentration of Tariquidar increased, the inhibitory activity of the treated strain HN‐1 against C. fructicola HD‐1 gradually decreased. At a final concentration of 50 μM, the inhibitory activity of the HN‐1 strain was nearly lost, with the inhibition zone measuring only 1.1 cm, approximately 50% of the size of the inhibition zone observed in the untreated HN‐1 strain (Figure 7A,B). Conversely, the intracellular oil drainage capacity increased significantly, with the diameter of the oil drainage circle reaching 2.0 cm, markedly larger than the 0.7 cm observed in the untreated wild‐type intracellular extract, representing a 2.9‐fold increase (Figure 7C,D).

FIGURE 7.

Lipopeptide activity was suppressed by broad‐spectrum ABC transporter inhibitors in Bacillus velezensis HN‐1. (A, B) Colletotrichum fructicola HD‐1 inhibition activity by HN‐1 strains treated with different concentrations of inhibitor. (C, D) Effect of intracellular extracts of HN‐1 strains treated with different concentrations of inhibitors on oil drainage capacity. (E) Reverse transcription‐quantitative PCR to detect the expression of genes related to the regulation of antibacterial active substance synthesis after treatment of HN‐1 strains with different concentrations of inhibitors. Experimental data were analysed through one‐way ANOVA. Asterisks indicate a significant difference (***p ≤ 0.001) by Duncan's test.

We selected representative synthetic regulatory genes to assess expression differences relative to the wild‐type HN‐1 strain. The results showed that the expression of these genes in the Tariquidar‐treated HN‐1 strain was significantly downregulated compared to the untreated wild‐type strain (Figure 7E). In subsequent experiments, we knocked out the rbsB gene by CRISPR (Clustered, Regularly interspaced, Short Palindromic Repeats) technology. The results showed that after knocking out the rbsB gene, the content of intracellular and extracellular antifungal lipopeptides of B. velezensis HN‐1 was significantly reduced, the antifungal activity was almost lost, and the genes related to lipopeptide synthesis were significantly downregulated (Figure S6). Taken together, these results suggested that ABC transporters pathway may act as crucial transporters pathway in lipopeptide efflux in B. velezensis HN‐1.

3. Discussion

Bacillus velezensis is an important biocontrol agent because it synthesises a diverse suite of antimicrobial lipopeptides (surfactin, fengycin, iturin/bacillomycin) that suppress phytopathogens and shape rhizosphere interactions (Devi et al. 2019; Hashem et al. 2019; Kenfaoui et al. 2024; Koumoutsi et al. 2004). While biosynthetic non‐ribosomal peptide synthase (NRPS) clusters responsible for these compounds are well described, the mechanisms that coordinate secretion with biosynthesis have remained unclear (Fazle Rabbee and Baek 2020). Here, using genetic, biochemical and interaction assays, the results suggest that ACR3 functions as a crucial regulator, mediating both the synthesis and efflux of lipopeptides in B. velezensis HN‐1.

Genetic disruption of acr3 dramatically altered lipopeptide dynamics and phenotypes. The HN‐1ΔQ acr3 mutant showed loss of extracellular lipopeptide activity (reduced haemolysis, oil‐spreading and antifungal activity; extracellular EC₅₀ increased from 53 μg/mL in wild type to 1709 μg/mL in the mutant), whereas intracellular extracts accumulated active compounds. HPLC/TOF data showed that the amount of bacillomycin D in the wild‐type strain was 76.95 ± 6.48 μg. In contrast, the mutant strain HN‐1ΔQ acr3 yielded only 1.89 ± 0.66 μg, representing a drastic reduction of approximately 97.5% compared to the wild type, consistent with impaired export and intracellular accumulation. Complementation restored lipopeptide secretion and swarming/biofilm phenotypes, indicating that ACR3 is necessary for effective lipopeptide export and for maintenance of extracellular antimicrobial activity.

Therefore, it is crucial to experimentally unravel the specific mechanisms of ACR3‐mediated lipopeptide regulation and evaluate the significance of lipopeptide‐protein interactions on the antimicrobial properties of the HN‐1 strain. Bacillus species are known for secreting diverse and bioactive antibacterial substances, including lipopeptide antibiotics such as surfactin, fengycin and iturin, which are synthesised via non‐ribosomal pathways. However, the regulatory mechanisms underlying the synthesis of each peptide differ. The surfactin synthase genes srfAA, srfAB, srfAC and srfAD form the srfA operon, which activates the downstream phosphopantetheine transferase gene sfp to synthesise surfactin (Beld et al. 2014; Junqing Qiao et al. 2024; Zhi et al. 2017). B. subtilis OKB105 is a backfill mutant of strain 168 lacking the sfp gene. Strain FZB42, which produces surfactin, exhibits better swarming motility compared to strain 168 or its sfp‐deficient mutants (Bremer et al. 2023; Kinsinger et al. 2003; Xie et al. 2015). Research has shown that B. subtilis OKB105 and OKB105ΔyczE spread more slowly than strains FZB42 and FZB42ΔyczE, suggesting a strong correlation between surfactin synthesis and swarming ability (Qiao et al. 2024). In swarming ability tests, the wild‐type HN‐1 strain spread uniformly across the culture dish, while the mutant HN‐1ΔQ acr3 strain exhibited a dotted distribution. The complemented strain HN‐1ΔC acr3 demonstrated swarming levels equivalent to the wild‐type, indicating that deletion of the acr3 gene impairs surfactin synthesis in Bacillus. Furthermore, B. subtilis OKB105 struggled to form biofilms, while Bacillus amyloliquefaciens FZB42 began forming a thin flat membrane layer after 12 h of incubation, which thickened over time (Qiao et al. 2015). Consistent with this, the wild‐type HN‐1 strain formed a biofilm within 16 h, developing a more mature three‐dimensional structure with folds by 24 h. The mutant strain HN‐1ΔQ acr3 also formed a biofilm by 16 h, but it did not exhibit a folded structure until 48 h. These results support the conclusion that surfactin significantly influences swarming motility and biofilm formation (Aleti et al. 2016; Gao et al. 2012). The mutation of the acr3 gene led to a loss of surfactin production in the mutant strain; however, the presence of exopolysaccharides (EPS) allowed for a thin biofilm to form during the later growth stages, which resulted in a marked reduction in both swarming and biofilm formation abilities. This underscores the direct correlation between acr3 gene expression and the regulation of lipopeptide (surfactin) synthesis.

Antimicrobial active substances are secondary metabolites produced by microorganisms, higher plants and animals that influence various cellular developmental functions. Landy et al. first identified the lipopeptide analogue iturin, synthesised via non‐ribosomal peptide synthase (NPRS) systems from Bacillus sp. in 1948 (Finking and Marahiel 2004; Stein 2005). Iturins, a class of active lipopeptide compounds with potent antifungal properties, are primarily regulated by the synthesis of iturin A, iturin C, bacillomycin D, F, L, LC and mycosubtilin (Maget‐Dana and Peypoux 1994; Stein 2005). While bacillomycin D is synthesised by the bmy gene cluster, mycosubtilin is produced by the myc gene cluster (Beld et al. 2014; Qiao et al. 2024; Zhi et al. 2017). The iturin produced by Bacillus exhibits significant antifungal activity; for instance, two congeners of the iturin family and bacillomycin D were isolated and purified from the fermentation broth of B. subtilis AU195, demonstrating 99.9% inhibition of Aspergillus flavus by Day 6 of fermentation (Moyne et al. 2001). The fen operon synthesises fengycin, which also possesses broad‐spectrum antifungal activity (Steller et al. 1999). In our previous studies, time‐of‐flight mass spectrometry results revealed that the fermentation broth of the wild‐type HN‐1 strain contained bacillomycin D compound, demonstrating significant inhibitory effects against various fungal species. The HN‐1 strain notably inhibited the vegetative growth of C. fructicola HD‐1, with an EC₅₀ of 53.1 μg/mL. Current insights into lipopeptide biosynthesis in Bacillus indicate that a cluster of npr genes, including biosynthetic and regulatory genes, is involved. Selected strains, such as B. subtilis 168, possess a complete lipopeptide synthase gene cluster but are unable to synthesise antimicrobial substances due to mutations in the sfp gene, which is essential for the non‐ribosomal pathway of lipopeptide synthesis (Bremer et al. 2023; Kinsinger et al. 2003; Xie et al. 2015). In this study, the extracellular extract of the mutant HN‐1ΔQ acr3 strain exhibited diminished activities, including reduced haemolytic and fungal inhibitory effects, with an EC₅₀ of 1709 μg/mL—only 3% of that of the wild‐type strain. Conversely, the intracellular extract was 2.6 times more effective in inhibiting the wild‐type strain, aligning with previously reported findings. Analysis of the liquid‐phase peak time of bacillomycin D from the intracellular extract of the wild HN‐1 strain indicated a peak time of 42.5–43.0 min, and the amount of bacillomycin D in the wild‐type strain was 76.95 ± 6.48 μg. In contrast, the mutant strain HN‐1ΔQ acr3 yielded only 1.89 ± 0.66 μg, representing a drastic reduction of approximately 97.5% compared to the wild type, consistent with impaired export and intracellular accumulation. Although a significant amount of active antimicrobial substance accumulates due to the acr3 gene mutation, it may remain unmodified and untranslated, leading to a failure in intracellular expression. These results suggest that the acr3 gene plays a direct role in regulating the biosynthesis of active antimicrobial substances.

Based on our experiments regarding antifungal activity and the results from the HPLC assay, we hypothesise that the acr3 gene may be involved in the regulation of lipopeptide synthesis. Drawing from extensive literature and transcriptomic, proteomic and metabolomic data, we tested 22 genes, including srfAA. Deletions within srfAA and extending to srfAB abolished the activity and production of all three enzymes in B. subtilis (Vollenbroich et al. 1994). Additionally, bmyB, which encodes C14‐iturin A, plays a crucial role in inhibiting fungal growth in B. amyloliquefaciens NCPSJ7 (Wang et al. 2022). The degU mutation has been shown to prevent both flagellar and biofilm formation in B. subtilis (Kobayashi 2007). Ashokkumar et al. (2018) isolated B. amyloliquefaciens VCRCB 483, which contains the gene encoding bmyC, and its crude extract exhibited significant inhibitory effects against Fusarium and Carvularia species. The sfp gene encodes a phosphotransmercaptoethyl transferase necessary for surfactin synthesis (Mootz et al. 2001). However, we demonstrated that the genetic deactivation of acr3 significantly suppressed the expression of genes coding for phosphotransmercaptoethyl transferase in the HN‐1ΔQ acr3 strains, indicating that the loss of function of acr3 impacts the lipopeptide synthesis pathway. The comA gene is essential for the expression of the surfactin synthase gene srfAC, as it stimulates binding to the srfA promoter via the transmembrane protein kinase ComP, thereby inducing srfA transcription and surfactin production (Roggiani and Dubnau 1993; Solomon et al. 1995). In our study, we observed significant downregulation of comA, comP and other com genes, suggesting that acr3 operates upstream and regulates lipopeptide synthesis. Additionally, the sfp gene was also significantly downregulated, located downstream of the regulatory cascade. To further elucidate these interactions, we employed ACR3 as a bait protein in GST pull‐down assays to investigate protein interactions with BmyC, FenC and Sfp. We found that ACR3 exhibited a strong interaction with both FenC and Sfp proteins. Combining all experimental data, we conclude that the acr3 gene regulates lipopeptide synthesis at the most upstream point in the regulatory network.

There are limited reports on the mechanisms underlying the transmembrane efflux of lipopeptides, with the yerP gene being the first functionally characterised RND transporter associated with surfactin efflux (Ostroumova et al. 2010; Tsuge et al. 2001). Previous studies indicated that B. subtilis enhances cardiolipin production to protect against surfactin‐induced membrane permeabilization (Pinkas et al. 2020). In our study, we observed that while the independent ACR3 protein exhibited no inhibitory activity, the HN‐1ΔQ acr3 mutant regained some inhibitory capacity upon the exogenous addition of ACR3 fusion protein at low concentrations. This was accompanied by significant upregulation of comP, sfp, fenD and bmyC compared to the untreated group. Our findings suggest that ACR3 functions as a key gene that senses intracellular lipopeptide concentrations and subsequently regulates their synthesis. When lipopeptide levels decrease within the cell, phosphorylation of comP triggers increased lipopeptide synthesis. Conversely, the absence of ACR3 results in lipopeptide accumulation, nullifying the antifungal properties of B. velezensis HN‐1. This led us to hypothesise that ACR3 interacts with proteins involved in lipopeptide efflux. To investigate this further, we constructed a cDNA library of B. velezensis HN‐1 using ACR3 as a decoy protein to screen for interacting downstream proteins, resulting in the identification of multiple proteins from the ABC transporter superfamily that interacted with ACR3. Interestingly, integrative analyses of the transcriptomic, proteomic and metabolomic datasets consistently revealed that the ACR3 protein is closely associated with the ABC transporter family, particularly influencing the expression of the rbsB gene. This observation further corroborates the findings obtained from the yeast library screening. Active drug efflux mediated by multidrug transporter proteins is a primary mechanism behind multidrug resistance in bacterial pathogens (Li and Nikaido 2004; Munita and Arias 2016). In recent years, ABC transporter proteins have emerged as one of the largest protein superfamilies, using ATP binding and hydrolysis to transport a variety of substrates across membranes (Dawson et al. 2007; Gerber et al. 2008; Patzlaff et al. 2003). Li et al. (2023) identified two upregulated genes (rbsB and degP) that are closely linked to processes such as cell membrane synthesis, energy metabolism and transmembrane transport. Régis Stentz's study on Lactobacillus sakei ΔrbsR demonstrated that transcription of the Rbs manipulator increased in glucose conditions, implicating the Rbs family in ribose regulation as novel transporter proteins (Stentz and Zagorec 1999). Consequently, we selected RbsB, a ribose ABC transporter substrate‐binding protein, and RbsC, a ribose ABC transporter, for reciprocal screening using the GST pull‐down technique with ACR3 as bait. The results revealed a strong interaction between RbsB and ACR3. Notably, the HN‐1ΔQ acr3 mutant exhibited significantly downregulated expression of ABC transporter‐related genes compared to the wild‐type strain. However, treatment with a low concentration of recombinant fusion protein resulted in marked upregulation compared to the untreated mutant. Thus, the interaction between ACR3 and ABC transporter family proteins suggests a critical role for these transmembrane channels in lipopeptide efflux.

The ABC transporter superfamily encompasses proteins such as P‐glycoprotein (p‐GP) and multidrug resistance‐associated proteins (MRPs) (Juliano and Ling 1976). Tariquidar, a broad‐spectrum P‐gp inhibitor, was used to assess its impact on the wild‐type HN‐1 strain (Senior and Bhagat 1998). Treatment with Tariquidar resulted in a dose‐dependent loss of inhibitory activity, with 50 μM leading to a significant increase in oil excretion capacity and a marked reduction in the expression of genes involved in lipopeptide synthesis. Additionally, RT‐qPCR analysis of the ACR3 mutant strain revealed that the deletion of acr3 impaired lipopeptide synthesis and altered antibacterial activities and lipopeptide content compared to the HN‐1 strain. In summary, ACR3 serves as a sensor that modulates intracellular lipopeptide levels. We propose that when lipopeptide concentrations exceed a certain threshold, ACR3 initiates secretion via the ABC transporter system while simultaneously downregulating biosynthesis. Conversely, a drop in intracellular lipopeptide levels below an unidentified threshold activates ACR3‐dependent signalling to enhance the expression of both biosynthetic and efflux genes. Thus, ACR3 plays a critical sensory regulatory role in lipopeptide biosynthesis and secretion in Bacillus, functioning as a pivotal hub gene in the regulation of antibacterial compounds and transmembrane transport in B. velezensis (Figure 8).

FIGURE 8.

A mechanistic network model of ACR3‐mediated regulation in lipopeptide biosynthesis and transmembrane transport.

In conclusion, this study involved the extraction of intracellular and extracellular fractions, along with testing their biological properties through the construction of fixed‐point mutant HN‐1ΔQ acr3 and complemented HN‐1ΔC acr3 strains. The development of prokaryotic expression vectors facilitated protein interaction screening using a yeast two‐hybrid system to identify molecules that interact with ACR3. GST pull‐down assays confirmed strong interactions between ACR3 and proteins RbsB. Importantly, this study is the first to demonstrate that ACR3 serves a crucial regulatory role in the synthesis and efflux of lipopeptides in Bacillus, specifically implicating the ABC transporter pathway in this process. These findings represent a significant initial step toward understanding the roles of ACR3 and the ABC transporter superfamily in maintaining lipopeptide levels and antimicrobial activity in Bacillus. Future research will delve deeper into these components to further elucidate their functions.

4. Experimental Procedures

4.1. Bioinformatics Analysis and Construction of the acr3 Mutant in Strain HN‐1

In our laboratory's previous research, a mutant library was established in strain HN‐1 through the application of TnYLB‐1 transposon‐mediated chemical transformation on pMarA. A strain that exhibited a loss of haemolytic capability was identified by assessing the presence of a haemolysis zone, and the acr3 gene in this strain was disrupted by transposon TnYLB‐1, as determined through RT‐PCR analysis and subsequent BLAST comparison. The amino acid sequence of acr3 was obtained from NCBI‐Blast, and ClustalW‐based comparison was performed using MEGA 7 software (Tamura et al. 2011) to construct a phylogenetic tree by the neighbour‐joining method. The structure of the transmembrane structural domain of ACR3 was predicted on the DeepTMHMM web server (https://dtu.biolib.com/DeepTMHMM).

The genomic DNA of HN‐1 was extracted using a TIANamp Bacteria DNA kit (TIANGEN Biotech (Beijing) Co. Ltd.). The upstream and downstream fragments of the acr3 were amplified by PCR with Q5 High‐Fidelity DNA polymerase (New England Biolabs Inc.) using HN‐1 genomic DNA as a template and primers (Qacr3‐F1/F2, Qacr3‐F5/F6, Table S1). Qacr3‐F3/F4 were used as templates to amplify the TnYLB‐1 transposon fragment using pMarA. The three fragments were fused by overlapping PCR to form the UP‐TnYLB‐1‐DOWN fusion fragment and recovered by DNA purification. The recovered product was subsequently ligated to pMD18‐T to form the recombinant plasmid pMD18‐T‐acr3 and transformed into Escherichia coli DH5α for bulk amplification. The backfill fragment was amplified using a similar method with primer pMK3‐acr3‐SalIF/SmaIR, and the backfill vector pMK‐acr3 was constructed using T4 ligase after double digestion of pMK3 using SalI and SmaI. After recovery, the homologous recombination method of construction was applied to construct mutants.

Strain HN‐1 was inoculated into GMI medium, cultured overnight at 28°C, 180 rpm, and then shaken the next day under the same passaging conditions to OD₆₀₀ = 1.0, followed by taking an appropriate amount of bacterial solution and cultured in GMII for 2 h (37°C, 180 rpm). 1 mL of the bacterial solution was centrifuged at 5000 rpm for 5 min and the bacteria resuspended with 100 μL of the supernatant to obtain the HN‐1 sensory state. pMD18‐T‐acr3 was added to the sensitive state and mixed well, placed in a 37°C water bath for 45 min, then Luria Bertani (LB) medium containing the appropriate antibiotics added and incubated at 37°C, 180 rpm for 3 h. This was centrifuged at 5000 rpm for 1 min, 100 μL of the resuspended bacterial body spread on an LB plate containing kanamycin (10 μg/mL) and incubated at 37°C for 12 h. Positive clone transformants were picked and placed on an LB plate containing kanamycin (10 μg/mL). Positive clone transformants were picked and placed in LB medium containing kanamycin (10 μg/mL) and incubated overnight at 37°C. DNA was then extracted from the positive clone transformants. The DNA was verified by PCR using validation primers. The correct transformants were sent to Shanghai Sangong Biological Co. The HN‐1ΔQ acr3 mutant strain was obtained. A similar approach was used to construct the complemented strain HN‐1ΔC acr3.

4.2. Bacterial and Fungal Strains and Growth Conditions

Bacillus velezensis HN‐1 was isolated from rice tissues sampled from rice fields in the Hainan Province, China, inoculated on LB medium, and cultured at 28°C. Colletotrichum fructicola HD‐1 was isolated, identified and preserved in our laboratory and grown on potato dextrose agar (PDA) and carboxymethylcellulose sodium (CMC) medium at 28°C. Saccharomyces cerevisiae NMY51, provided by Shanghai Biogene Biotechnology company, was inoculated on YPDA medium and cultured at 28°C. pMarA, a shuttle vector carrying TnYLB‐1 transposon and mariner‐Himar1 transpose, pUC replicon and thermosensitive replicon for gram‐positive hosts were generous gifts from Prof. Gao at Nanjing Agricultural University (Table S2).

4.3. Arsenic Tolerance Analysis

The target strains were transferred to LB medium in 250 mL Erlenmeyer flasks and incubated overnight at 28°C with shaking at 180 rpm until they reached an OD₆₀₀ of 1.0. The seed solution was transferred to 100 mL LB containing different concentrations of arsenic ions in a 1:100 ratio of seed solution to culture medium and incubated for 24 h at 28°C and the OD₆₀₀ was measured. Consistent results from three independent biological experiments with triplicates were used to analyses arsenic tolerance in HN‐1ΔQ acr3, HN‐1ΔQ acr3 and HN‐1 strains.

4.4. Biofilm Formation and Swarming Motility Assay

The target strains were selected in a sequential manner on LB plates, then inoculated in LB liquid medium and incubated at 28°C with 180 rpm until the OD₆₀₀ reached 1.0. Firstly, 5 μL of the bacterial suspension was pipetted into the centre of the swarm plates, which had been prepared a day prior and contained 25 mL of LB medium fortified with 0.7% agar (10 g tryptone, 5 g yeast extract, 5 g NaCl per litre) (Patrick and Kearns 2009; Salvetti et al. 2011). The plates were placed in an incubator at 28°C and observed every 2 h, and the radius of colony expansion was recorded until the entire plate was expanded. Subsequently, 20 μL of the bacterial suspension was combined with 2 mL of minimal salts glycerol glutamate (Msgg) medium and introduced to the 24‐well cell culture plate. The plate was incubated at 28°C, and the formation of biofilms by the six strains of Bacillus was observed. Swarm assays were conducted using three independent biological experiments with three technical replicates.

4.5. Extraction of Active Substances by n‐Butanol Extraction

The target strains were transferred into 250 mL Erlenmeyer flasks containing LB medium and incubated at 28°C in a shaking incubator at 180 rpm for 48 h. Subsequently, centrifugation was performed at 10,000 rpm for 10 min to remove the bacteria, resulting in the isolation of the fermentation supernatant. The fermentation supernatant was extracted with n‐butanol at a ratio of 1:1 to obtain the upper n‐butanol extracts, which were then freeze‐dried after rotary evaporation to obtain the n‐butanol crude extracts (Stein 2005).

4.6. Haemolytic Activity and Oil Drainage Capacity Assay

Under sterile conditions, the n‐butanol‐based crude extracts were dissolved in methanol to a final concentration of 50 mg/mL, and 10 μL was added dropwise on a filter paper with a diameter of 6 mm in a Petri dish. After drying, the filter paper was placed symmetrically on the blood agar plates, blown to dryness and placed upside down in an incubator at 28°C. The diameter (mm) of the transparent circle (zone) was measured at 12 h using a Vernier calliper (Arnaouteli et al. 2021).

The oil expelling assay was used to detect the surface activities of the crude extracts using the methods described by Liang et al. (2020). To better show the result of oil expelling, a 10 mL light fraction of paraffin was dyed red with 10 mg Sudan III, and 1 mL red oil was dropped into 90 mm plates containing 20 mL double‐distilled water. After spreading the red oil across the whole surface, 1 μL crude extract of surfactin was dropped at the centre of the plates, and 1 μL sterile water was used as control (Liang et al. 2020).

4.7. Effect of n‐Butanol Crude Extracts on C. fructicola

Colletotrichum fructicola HD‐1 was transferred in PDA in 250 mL Erlenmeyer flasks and subjected to 3–5 days of shaking incubation at 180 rpm and 28°C, and the medium was filtered with sterile gauze to remove mycelia, resulting in a spore suspension that was mixed 1:1 with PDA. The crude n‐butanol dissolved in sterile water to a final concentration of 20 mg/mL was pipetted into a 10 μL drop onto a filter paper with a diameter of 6 mm. Sterile water was used as a negative control. After drying, the filter paper was placed symmetrically on the plate and incubated at 28°C for 48 h. The diameter of the inhibition ring (mm) was measured with Vernier callipers (Jin et al. 2020).

4.8. High‐Performance Liquid Chromatography (HPLC)

HPLC analysis of the crude lipopeptide extract was performed using an Agilent 1260 HPLC system (Agilent Technologies) equipped with a Diamonsil C₁₈ column (5 μm, 250 × 10.0 mm). The column temperature was maintained at 35°C. The mobile phase comprised ultrapure water containing 0.1% phosphoric acid (solvent A) and acetonitrile (solvent B) at a flow rate of 1 mL/min. A gradient elution was applied, with solvent A decreasing linearly from 100% to 0% over 54 min. The injection volume was 20 μL, and the detection wavelength was set to 210 nm. Prior to analysis, the crude extract of B. velezensis was reconstituted in double‐distilled water to a concentration of 20 mg/mL. Bacillomycin D previously isolated in our laboratory was used as a reference standard (Gong et al. 2014).

4.9. Multi‐Omics Analysis

Single colonies were picked and inoculated into 50 mL LB medium, followed by incubation overnight at 37°C with shaking at 180 rpm. Cells were harvested at the logarithmic growth phase (OD₆₀₀ = 0.8). An appropriate amount of the bacterial culture was transferred into suitable centrifuge tubes, and cells were collected by high‐speed centrifugation (4°C, 12,000 g for 3 min). The supernatant medium was discarded, and the cells were quickly frozen in liquid nitrogen before being sealed with Parafilm. Transcriptome, proteome and metabolome sequencing were conducted at Beijing Biomarker Technologies Co. Ltd. and generated data were analysed at the online Majorbio Cloud Platform (https://www.biocloud.net/).

4.10. Lipopeptide Synthesis Regulatory Gene RT‐qPCR Assay

To determine the changes in the expression level of lipopeptide synthesis regulating genes in mutant strains, HN‐1ΔQ acr3 strains were grown in LB medium under shaking at 180 rpm until the cultures reached the log growth phase and then the Bacillus cells were collected by centrifugation at 12,000 rpm at 4°C. Total RNA was extracted from the Bacillus cells using Bacterial RNA Kit DP441 (Tiangen). RNAs were reverse transcribed into cDNA with Thermo Scientific RevertAid First Strand cDNA Synthesis following protocols described by the manufacturer. qPCR‐based transcriptomic analysis of the expression of genes associated with the regulation of lipopeptides biosynthesis in the HN‐1, HN‐1ΔQ acr3 and HN‐1ΔC acr3 strains was carried out using gene‐specific primers. Transcribed cDNAs were used as the tested templates, DNA (positive control) and RNA (negative control). Real‐time qPCR was carried out using SYBR Premix Ex Taq (Tiangen) in a 7500 Fast RealTime PCR Detection System (Applied Biosystems). The expression of the 16S rDNA housekeeping gene from B. velezensis was used for internal normalisation, and analysis was carried out with the ΔΔC _T method using the 7500 system SDS software (Applied Biosystems). The expression data represent the average expression records from three independent experiments conducted in triplicate.

4.11. Validation of Exogenous Expression of ACR3

For ACR3 protein expression, the acr3 gene was amplified from the HN‐1 chromosomal DNA with pGST‐acr3‐BF and pGST‐acr3‐ER primer pairs. The PCR products were purified, cloned into the BamHI and EcoRI sites of the expression plasmid pGoodp, and transformed into E. coli BL21(DE3). The positive colonies harbouring the expression plasmids were cultured in liquid LB medium supplemented with 100 μg/mL ampicillin and incubated in a steady shaking incubator at 180 rpm and 37°C with shaking at 180 rpm. At OD₆₀₀ of 0.5–0.7, isopropyl thio‐β‐D‐galactoside (IPTG) at a final concentration of 1 mM was added to the cultured strains in the flasks. After 20 h of incubation at 16°C with shaking at 120 rpm, the cells were harvested by centrifugation at 5000 rpm for 20 min (Liang et al. 2020; Wu et al. 2011). The bacterial cell lysates were loaded onto the columns. The recombinant proteins were washed and eluted according to the instructions in the His‐tag protein purification kit. The proteins were detected and identified by SDS‐PAGE and western blot.

The purified recombinant protein pGoodp‐acr3 was used to extract lipopeptides after co‐culture in mutant strain HN‐1ΔQ acr3 at concentrations of 10 μg/mL, 50 μg/mL, 100 μg/mL, 1 mg/mL and 10 mg/mL, and the C. fructicola HD‐1 inhibitory activity was measured according to the method in 2.7. RNA was extracted using purified recombinant protein pGoodp‐acr3 at 10 μg/mL and 1 mg/mL concentrations co‐cultured with mutant strain HN‐1ΔQ acr3 for gene expression assays of genes regulating lipopeptide synthesis.

4.12. Y2H Assay

The Shanghai Biogene Biotechnology company constructed cDNA from the genomic library of HN‐1. The acr3 gene was amplified from HN‐1 chromosomal DNA using primer pBT3‐acr3F and pBT3‐acr3R, cloned into the SfiI site of the expression plasmid pBT3‐STE and pBT3‐SUC, and transformed into E. coli DH5α. The DH5α cells harbouring the expression plasmids were cultivated at 37°C with shaking at 180 rpm in LB medium containing 100 μg/mL ampicillin, and recombinant plasmid pBT3‐STE‐acr3 and pBT3‐SUC‐acr3 were obtained. The two recombinant plasmids were separately co‐transferred into the recipient yeast strain NMY51 with the following various plasmids for autoactivation assay and functional assay in Table S2. NMY51 yeast transformants containing the correct pBT3‐SUC‐acr3 decoy plasmid were used as the recipient bacteria to prepare the sensory state, into which the library plasmid pPR3N‐acr3 was transferred and plated on selective solid medium (SD−Trp−Leu−His) supplemented with 3‐amino‐triazole (3AT).

To further validate the interactions identified from the library screening as well as the interactions with lipopeptide synthetases, pairwise Y2H assays were performed. The coding sequences of the candidate interacting proteins (RbsB, SrfAA, BmyC and FenC) were amplified and cloned into the corresponding prey or bait vectors. Specifically, using RbsB as an example, the designated vector pair (NubG‐ACR3 and pBT3‐STE‐RbsB) was co‐transformed into the yeast strain NMY51. Co‐transformants were initially selected on double‐dropout medium (SD/−Leu−Trp). Subsequently, the positive clones were spotted onto quadruple‐dropout medium (SD/−Leu−Trp−His−Ade) supplemented with 30 mM 3‐AT to evaluate protein–protein interactions. Where applicable, X‐α‐gal was added for blue‐white colony screening. Co‐transformants harbouring NubI/Cub served as the positive control, while the empty vector pair (NubG/Cub) and the respective autoactivation test pairs (NubG‐ACR3/Cub and NubG/Cub‐target proteins) served as negative controls. All plates were incubated at 30°C for 3–5 days prior to photography.

4.13. Bimolecular Fluorescence Complementation (BiFC) Assays

To confirm the interaction of ACR3 and RbsB in planta, BiFC assays were performed in Nicotiana benthamiana using Agrobacterium‐mediated transformation. The specific procedures were conducted according to the methods described by Xie et al. (2023).

4.14. GST Pull‐Down

To express the ACR3 and RbsB, RbsC, the acr3, rbsB, rbsC genes were amplified from HN‐1 chromosomal DNA using primer pGST‐acr3‐BF/pGST‐acr3‐ER, pET32a‐rbsCF/pET32a‐rbsCR, pET32a‐rbsBF/pET32a‐rbsBR, cloned into the BamHI and EcoRI sites of the expression plasmids pGST and pET3‐32a, and transformed into E. coli BL21(DE3). The cells harbouring the expression plasmids were cultured at 37°C with shaking at 180 rpm in an LB medium containing 100 μg/mL ampicillin until an OD₆₀₀ nm of 0.5–0.7 was reached. Flasks containing the cultures were supplemented with IPTG at a final concentration of 1 mM. After 16 h of incubation at 28°C with shaking at 120 rpm, the cells were harvested by centrifugation at 5000 rpm for 20 min. A high‐pressure homogeniser was used to lyse the cell pellets and centrifuged at 5000 rpm for 20 min to remove cell debris. The cell supernatant was collected and fully conjugated with BeyoGold GST‐tag Purification Resin and BeyoGold His‐tag purification resin for 1 h. The purification of recombinant proteins was then purified using 5 mL GSTrap HP and HisTrap HP columns (Liang et al. 2020; Wu et al. 2011). The column was loaded with the bacterial cell lysate, and the recombinant proteins were washed and eluted according to the instructions in the GST‐tag protein purification kit and His‐tag protein purification kit. The proteins were detected and identified by SDS‐PAGE and western blot. The purified pGST‐acr3 protein was fully conjugated with BeyoMag anti‐GST Magnetic Beads, and then the pET3‐32a‐rbsC protein or pET3‐32a‐rbsB protein were added. The conjugated intercalated proteins were eluted with eluent, and SDS‐PAGE and western blot detected the samples to be tested.

4.15. Effects of Broad‐Spectrum Inhibitors of the ABC Transport Superfamily on the Activity of Strains

The ABC transporter protein superfamily broad‐spectrum P‐glycoprotein inhibitor Tariquidar was co‐cultured with strain HN‐1 at 50 nM, 500 nM, 1 μM, 10 μM and 50 μM to extract lipopeptides and to test the antagonistic activity of the extracellular extract against HD‐1 and the oil‐draining ability of the intracellular extract, RT‐qPCR‐based analysis of expression of lipopeptide biosynthesis genes in the wild‐type strain HN‐1 treated with different concentrations of Tariquidar inhibitors.

4.16. Data Analysis

Experimental data were analysed through one‐way ANOVA using SPSS v. 2.5 software and presented as mean values of the treatment replicates ± standard deviation. For multiple comparisons of means and evaluation of significant differences among the treatments, Duncan's testwas used at p = 0.05. The values were mean ± standard deviation. The results were obtained from three biological experiments, each with three technical replicates. Asterisks (*) indicate a significant difference (p ≤ 0.05), asterisks (**) indicate a highly significant difference (p ≤ 0.01) and (***) indicate a significant difference (p ≤ 0.001) between the two groups by Duncan's test.

Author Contributions

Ren Zeng: validation, formal analysis, investigation. Zheng Lin: investigation, validation, formal analysis. Yu Wang: investigation, validation. Yukai Fang: investigation, validation, formal analysis, writing – original draft. Yuying Shen: investigation, validation, formal analysis, writing – original draft. Jiatong Wang: validation, formal analysis, investigation. Weiguo Miao: writing – review and editing, project administration, supervision. Qingbiao Xie: project administration, resources, writing – review and editing. Haonan Wang: data curation, project administration, formal analysis. Pengfei Jin: writing – review and editing, writing – original draft, funding acquisition, project administration, methodology, conceptualization. Wenbo Liu: project administration, resources, formal analysis.

Funding

This work was supported by grants from the National Natural Science Foundation of China (Grants 32560681, 32260698, 31960552), Hainan Provincial Natural Science Foundation‐the Scientific Research Foundation for Advanced Talents (Grants 324RC455, 322RC591). Hainan Province Science and Technology Talent Innovation Project (No. KJRC2023B14).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Mutant library screening and identification of acr3‐deficient mutants.

Figure S2: Bioinformatics analysis of ACR3.

Figure S3: (A) Yield of crude extracts from wild‐type strain HN‐1 and its mutant strains. (B) Haemolytic activity assays of the wild‐type strain HN‐1 and its mutant strains.

Figure S4: Detailed pathway annotations of transcriptomic, proteomic and metabolomic analyses in the acr3‐deficient strain HN‐1ΔQ acr3.

Figure S5: ACR3 interacts with Sfp, SrfAA, FenC, BmyC.

Figure S6: Construction of HN‐1ΔrbsB, colony morphology and activity assays.

Table S1: Primers for specific PCR.

Table S2: Strains and plasmids used in this study.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.