Antifungal Agents of Bacillus velezensis 5RB: Complex Action of Lipopeptides, Polyketides, and Subtilosin A

Abstract

Bacillus velezensis is a promising candidate for the biocontrol of plant pathogens due to its production of broad-spectrum antimicrobial metabolites. The present study evaluates the antifungal potential and metabolic profile of the non-pathogenic rhizobacterial strain Bacillus velezensis 5RB. Modifying the mineral composition across three different growth media (A, B, and C) resulted in distinct secondary-metabolite profiles of B. velezensis 5RB. Notably, the free-cell supernatant of strain 5RB cultivated in medium A and C demonstrated the most potent antifungal activity against the gray mold pathogen Botrytis cinerea. For the first time, the antimicrobial metabolites produced by B. velezensis 5RB across these three media were comprehensively characterized using ultra-performance liquid chromatography-quantitative time-of-flight mass spectrometry (UHPLC-QTOF-MS). Mass spectrometric analyses (LC-ESI-MS) identified predominantly lipopeptides from the surfactin and fengycin families, alongside three classes of polyketides: macrolactins, difficidins, and bacillaenes. Multiple lipopeptide homologues, including a novel surfactin isoform (B*), were confirmed by tandem mass spectrometry (LC-ESI-MS/MS). Additionally, this study provides the first experimental evidence for subtilosin A production by B. velezensis 5RB. The multicomponent fermentation matrix of 5RB conferred robust biocontrol protection to tomato leaves against B. cinerea and Phytophthora infestans in an in vitro disease model, underscoring its strong potential for sustainable agricultural applications.

1. Introduction

Plant pathogenic fungi and oomycetes continue to pose a serious challenge to global agriculture. Their rapid dissemination and broad host range threaten the production of numerous economically important crops, leading to reduced yield, deterioration of product quality, and substantial financial losses. Consequently, the development of environmentally friendly strategies for disease management has become a priority in sustainable crop protection [1,2,3]. Although synthetic fungicides remain the principal approach for controlling plant diseases, their long-term use has raised concerns about environmental contamination, the persistence of chemical residues, and the rapid emergence of resistant pathogen populations [4]. These limitations have stimulated the search for biological alternatives based on beneficial microorganisms. Among them, Bacillus velezensis has become one of the most extensively investigated rhizobacteria due to its ability to suppress diverse plant pathogens while promoting plant health [5,6,7]. Rather than relying on a single mode of action, B. velezensis employs a complex arsenal of complementary mechanisms. Different strains have been shown to activate host immune responses, stimulate the expression of defense-related enzymes, secrete hydrolytic enzymes capable of degrading fungal cell walls, release volatile organic compounds, and synthesize a wide range of antimicrobial secondary metabolites [8,9,10,11,12]. The relative contribution of these mechanisms varies among strains and strongly depends on environmental conditions, making B. velezensis an attractive model for investigating the relationship among cultivation conditions, metabolite production, and biocontrol efficacy.

In general, the most promising biofungicide strains of B. velezensis combat fungi by producing lipopeptides or polyketides, thereby directly inhibiting pathogen growth [13,14]. Microbial lipopeptides have attracted considerable attention for their structural and functional diversity, antifungal activity, low resistance levels, and low toxicity [15,16]. Lipopeptides (LPs) are a class of non-ribosomally synthesized (NRPS) peptides in which a hydrophilic oligopeptide ring is linked to a hydrophobic fatty acid chain [17]. The amino acid sequence of the oligopeptide classifies lipopeptides into three families: iturin, fengycin, and surfactin [7]. The composition of lipopeptides produced by Bacillus species is highly strain-dependent and strongly influenced by cultivation conditions [18]. Among them, surfactins are the most widely distributed cyclic lipopeptides, being reported in several Bacillus species, including B. velezensis, B. amyloliquefaciens, B. licheniformis, B. methylotrophicus, and B. thuringiensis [19]. They comprise a family of homologous heptapeptides linked to β-hydroxy fatty acids of different chain lengths, giving rise to numerous molecular variants that differ by 14 Da increments [20]. Although surfactins exhibit only moderate direct antifungal activity, they play an important role in biocontrol by destabilizing biological membranes, facilitating biofilm formation and root colonization, and contributing, in combination with other lipopeptides, to overall antifungal efficacy enhancement [14]. Fengycins represent the principal antifungal lipopeptides synthesized by B. velezensis [13]. Their structural diversity arises from variations in both the peptide sequence and the β-hydroxy fatty acid chain, resulting in multiple homologs and isoforms, including the fengycin A and B families and several recently described variants [21,22,23]. This molecular diversity is associated with a broad spectrum of activity against filamentous fungi, particularly Botrytis, Fusarium, and Rhizoctonia species. Among the known isoforms, fengycin B generally displays stronger antifungal activity than fengycin A [13].

In addition to lipopeptides, B. velezensis synthesizes several polyketides that further contribute to pathogen suppression. Genome analyses have consistently identified biosynthetic gene clusters for bacillaene, difficidin, and macrolactins [24,25]. Bacillaene and difficidin are primarily known for their antibacterial properties, whereas macrolactins also exhibit antifungal activity and occur as several naturally modified derivatives [25,26,27,28]. The functional diversity of these metabolites has spurred extensive research to isolate novel B. velezensis strains and sequence their genomes. To date, more than 1600 genomes of this species have been deposited in the NCBI database, revealing a highly conserved genomic architecture of 3.9–4.2 Mb that encodes approximately 3700–4200 proteins. We previously reported the whole-genome sequence of the Bulgarian isolate B. velezensis 5RB [29], which harbors seven complete biosynthetic gene clusters for polyketides and lipopeptides. Although the genomic potential of strain 5RB has been established, the metabolite profile and the relative contribution of individual metabolite groups to its antifungal activity have not yet been investigated.

The present study aimed to characterize the qualitative metabolite profiles of B. velezensis 5RB cultivated at different media and to determine which antifungal compounds are relevant to the biological control of B. cinerea and Phytophthora infestans. Using UHPLC-QTOF-MS, we compared secondary metabolite production across three growth media and mapped the identified compounds to their corresponding genome-predicted biosynthetic gene clusters. Particular focus was placed on identifying metabolites responsible for the strain’s pronounced antifungal activity and evaluating the individual contributions of distinct metabolic fractions to its overall bioactivity.

2. Materials and Methods

2.1. Microbial Strains and Cultivation Conditions

B. velezensis 5RB, previously isolated from rhizosphere-associated lake sediments (NBIMCC no. 9094), was used throughout this study. The strain was maintained at −70 °C in Luria–Bertani (LB) medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 20% (v/v) glycerol. For all experiments, the inoculum was prepared by transferring a frozen stock into 100 mL of LB medium (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, and 10 g/L glucose) in 500 mL Erlenmeyer flasks. Cultures were incubated at 37 °C on a rotary shaker at 200 rpm until an optical density of ~2.0 at OD₆₀₀. All fermentation experiments were performed in three independent biological replicates (n = 3).

Production of extracellular antimicrobial metabolites was carried out in three production media (media A, B, and C) that differed in mineral composition but contained identical carbon and nitrogen sources (

Table 1. Composition of the production medium for B. velezensis 5RB.

Component (g/L)	Medium A	Medium B	Medium C
Sucrose	10	10	10
Soybean meal	20	20	20
MgSO4	1	5	1
CaCl2	1	3	3
MnSO4	–	0.2	0.2
K2HPO4	–	1	–
FeSO4	–	–	0.2

). Each flask contained 200 mL of production medium and was inoculated with a 5% (v/v) seed culture. Fermentations were performed at 37 °C and 200 rpm for 24–72 h.

2.2. Phytopathogenic Fungi Cultivation and Antifungal Activity of Cell-Free Supernatant

B. cinerea (NBIMCC 120; ATCC^® 28985™) and Ph. infestans (CBS 120920, Westerdijk Fungal Biodiversity Institute, the Netherlands) were used as target phytopathogens for evaluation of the antifungal activity of B. velezensis 5RB.

B. cinerea was routinely cultivated on potato dextrose agar (PDA) plates at 25 °C for 5–7 days until abundant sporulation was observed. Ph. infestans was maintained on rye agar medium and cultivated at 18 °C in the dark for 7 days.

The antifungal activity of the cell-free supernatant of B. velezensis 5RB or of the HPLC lipopeptide fractions was assessed against B. cinerea using the resazurin microtiter test, as previously described by Armenova et al. [30]. The test compounds from the dried HPLC-separated peaks were prepared as stock solutions in 0.5% dimethyl sulfoxide (Merck KGaA, Darmstadt, Germany) and then diluted in sterile distilled water to obtain the desired working concentrations.

The antifungal activity assays were performed in three independent biological replicates (n = 3). Data are presented as mean ± SD. Statistical significance was evaluated using one-way ANOVA followed by Tukey’s post hoc test, with a significance threshold set at p < 0.05.

2.3. Antifungal Activity on Detached Leaves

Healthy, detached tomato leaves at a similar developmental stage (the fourth leaf) were surface-disinfected with 70% ethanol, rinsed twice with sterile distilled water, and air-dried under sterile conditions. The leaves were then placed on moist, sterile filter paper in sealed plastic containers to maintain high relative humidity throughout the experiment.

For the biocontrol assays, the entire B. velezensis 5RB fermentation broth collected after 24 h of cultivation was used without removing the bacterial cells. The fermentation broth was diluted 10-fold with sterile distilled water and applied to the surface of detached leaves using a spray bottle. Leaves were sprayed thoroughly until visibly wet, then allowed to dry for 1 h before pathogen inoculation. Approximately 2.5 mL of the original, undiluted fermentation broth was used to treat 24 leaves.

Subsequently, the center of each leaf was inoculated with a 5 × 5 mm agar plug excised from an actively growing 7-day-old colony of B. cinerea or Ph. infestans. The inoculated leaves were placed in sterile Petri dishes lined with moist filter paper to maintain high humidity and incubated at 22 ± 0.5 °C under a 16 h light/8 h dark photoperiod. Disease development was monitored daily, and lesion diameter was measured after incubation. Each experimental group consisted of 24 independent tomato leaves. Representative leaves from each treatment group were selected for photographic documentation.

2.4. Extraction of Secondary Metabolites from Bacillus velezensis 5RB

After cultivation, bacterial biomass was removed by centrifugation (12,000× g, for 30 min, at 4 °C), and the supernatant was filtered through a 0.45 μm membrane filter to obtain a cell-free culture filtrate. Extracellular metabolites were recovered by acid precipitation after adjusting the filtrate to pH 2.0 with 6 M HCl. After overnight incubation at 4 °C, the precipitate was collected by centrifugation (10,000× g for 15 min) and extracted twice with methanol (≥99.9%, suitable for HPLC, Sigma Aldrich, Milwaukee, Germany) after brief vortexing. The insoluble impurities were removed by centrifugation at 8000× g for 5 min. The solvent was removed by vacuum evaporation at 37 °C. The resulting metabolite extracts from cultures of B. velezensis 5RB grown in media A, B, and C were stored at −80 °C until the next analysis.

2.5. Identification of Metabolites Produced by B. velezensis 5RB Using Mass Spectrometric Studies

Extracted metabolites from B. velezensis 5RB cultivated in media A, B, and C were dissolved in methanol and filtered through a 0.22 μm membrane syringe filter. Metabolite profiling was carried out using a reversed-phase ultra-performance liquid chromatography coupled with a time-of-flight mass spectrometer (UHPLC/QqTOF Bruker Compact) (Bruker Daltonics, Bremen, Germany). Separation was achieved by RP-UHPLC, using an Acquity® BEH C18 VanGuard pre-column (130 Å, 1.7 μm, 2.1 × 5 mm) (Waters Corporation, Milford, MA, USA) coupled with a Solo C18 RP analytical column (2.1 × 100 mm, 2.0 μm) (Bruker Corporation, Billerica, MA, USA). The chromatographic system employed a binary solvent gradient consisting of 0.1% formic acid (reagent grade, ≥95%, Sigma Aldrich, Milwaukee, Germany) in water (A) and 0.1% formic acid in acetonitrile (B) (≥99.9%, gradient grade, Sigma Aldrich, Milwaukee, Germany). An aliquot of 15 μL of each sample was injected and eluted at a flow rate of 0.10 mL min⁻¹. The gradient elution program was as follows: 0.0–1.5 min: column equilibration (50% B); 1.5–2.5 min: isocratic at 50% B; 2.5–24.0 min: linear gradient from 50% to 100% B; 24.0–30.0 min at 100% B; 30.0–31.0 min: return to initial conditions (100% to 50% B); 31.0–33.0: final re-equilibration at 50% B. Eluted lipopeptides were monitored by UV detection at 226 nm. Components eluted in each peak were determined by ESI-MS and ESI-MS/MS in positive ionization mode in the range m/z 250–2500. The main source parameters were set as follows: capillary voltage 4.0 kV; drying gas flow 8.0 L min⁻¹; nebulizer pressure 1.9 bar; collision energy: 10.0 eV; collision RF: 2500.0 Vpp; transfer time: 90.0 µs; and ion source temperature 200 °C. The intense precursor ions were fragmented in positive ionization mode via Collision-Induced Dissociation (CID) using auto nMS scan, in the m/z range of 150–2500, at capillary voltage: 4.5 kV and ion source temperature 250 °C, the remaining parameters were the same as for the ESI-MS method. Raw data acquisition and spectral processing were performed with Bruker DataAnalysis software (v. 4.0, Bruker Daltonics). Mass accuracy for precursor ions (MS1) was evaluated with a strict tolerance threshold of ≤10 ppm. Spectra were averaged over three scans. MS/MS fragmentation spectra were interpreted by sequence ions (b-ions, y-ions) and diagnostic fragments matched within a fragment mass tolerance of 0.05 Da. Molecular ion peaks of lipopeptides and polyketides produced by B. velezensis 5RB were identified based on their mass-to-charge (m/z) ratios in LC-ESI-MS spectra, which were compared with previously reported data in studies [21,31,32,33,34,35,36,37,38] and commercial standards (Fengicin with purity ≥90%, Sigma-Aldrich Co., Saint-Quentin-Fallavier, France; Surfactin with purity ≥98%, Sigma-Aldrich Co., St. Louis, MO, USA). The polyketides were also confirmed by comparing the isotopic distribution in the ESI-MS spectra with the calculated one.

The relative abundance of metabolite groups was estimated by calculating the relative proportion of the summing peak areas of all detected homologs within each family, expressed as a percentage of the total metabolite signal, as previously described by Barale et al. [31].

2.6. Lipopeptide Fractions Isolated by RP-HPLC from Cell-Free Supernatant of B. velezensis 5RB

The crude lipopeptide extract derived from the cell-free supernatant of B. velezensis 5RB was fractionated by reverse-phase high-performance liquid chromatography (RP-HPLC) using a Nucleosil C18 column (250 × 4.6 mm, 3 µm, 100 Å). Chromatographic separation was performed at a flow rate of 2.0 mL/min with a binary mobile phase system consisting of 0.1% (v/v) trifluoroacetic acid (TFA) in ultra-pure water (Phase A) and 0.1% (v/v) TFA in acetonitrile (Phase B). A 1.0 mL sample volume (containing 5.0 mg/mL of methanol extract) was injected on the column. The following gradient elution profile was used: 0.0–5.0 min 20% B; 5–19.0 min linear gradient from 20% to 100% B; 19.0–22.0 min at 100% B; 22.0–24.0 min: return to initial conditions (100% to 20% B); 24.0–28.0: final re-equilibration at 20% B. The eluate was monitored by UV absorbance at 216 nm. The highest intensity fractions eluted in individual peaks were collected, dried using a vacuum concentrator (SpeedVac) and used for further mass spectrometric characterization.

3. Results

3.1. Antifungal Activity of Bacillus velezensis 5RB in Selected Production Media

The antifungal activity of cell-free supernatants from B. velezensis 5RB against Botrytis cinerea varied with both the production medium and the assay duration (Figure 1). All three media supported the production of extracellular antifungal compounds; however, the magnitude and persistence of inhibition differed markedly among them. The highest initial activity was observed in cultures grown in media A and C. After 24 h of fungal exposure, the corresponding cell-free supernatants inhibited B. cinerea growth by 98.3% and 98.1%, respectively. Thus, despite their different mineral compositions, both media supported nearly complete inhibition of fungal growth at this time point. After 48 h, inhibition remained high, reaching 86.6% for medium A and 93.75% for medium C. At 72 h, inhibition decreased to 53.8% and 62.6%, respectively.

In comparison, medium B showed lower antifungal activity throughout the experimental period. The differences among the media became particularly evident at later time points, with medium C maintaining the highest inhibitory activity at 48 and 72 h. This indicates that the cultivation conditions affected not only the initial antifungal potential of the cell-free supernatant but also the persistence of its activity during fungal growth.

Because the three production media differed in mineral composition while the carbon and nitrogen sources were kept constant, these results indicate that the mineral environment of the cultivation medium influences the antifungal phenotype of B. velezensis 5RB. However, the experimental design does not allow the specific contribution of an individual salt or mineral ion to be distinguished. Several mineral components were changed simultaneously across the three media, and soybean meal, used as a complex nutrient source, also contributes to the basal mineral composition of the cultivation medium. Therefore, the observed differences cannot be attributed to the effect of individual supplemented salts or mineral ions in isolation. Rather, the results reflect the influence of the overall mineral balance of the production medium on antifungal activity and secondary metabolite production in B. velezensis 5RB.

Given their strong and sustained antifungal activity, media A, B, and C were subsequently selected for comparative characterization of the antimicrobial metabolites produced by B. velezensis 5RB. Medium B generally showed lower activity throughout the test period, whereas medium C, supplemented with MnSO₄ and FeSO₄, showed the highest activity up to 72 h of fungal cultivation. Nevertheless, the higher activity observed with medium C should not be interpreted as evidence of a specific stimulatory effect of MnSO₄ or FeSO₄ individually, because the three media differed in several mineral components, and contributions from minerals in SBM cannot be excluded.

3.2. Characterization of Metabolite Components Produced by B. velezensis 5RB by Mass Spectrometry

3.2.1. LC-MS Profiling of Secondary Metabolites from B. velezensis 5RB

The extracellular secondary metabolites produced by B. velezensis 5RB, cultivated in three distinct media (A, B, and C), were analyzed using reversed-phase ultra-performance liquid chromatography coupled with electrospray ionization quadrupole time-of-flight mass spectrometry (UHPLC-ESI-Q-TOF-MS). The resulting analytical profiles revealed complex metabolic patterns across all tested media (Figure 2). The secondary metabolites were detected mainly as lipopeptides from the fengycin and surfactin families, with surfactins with retention times (RT) between 19 and 27 min being dominant. Initial identification of the lipopeptides was achieved based on mass-to-charge ratios (m/z), retention times (RT) and comparison with literature values [21,31,32,33,34,35,36,37]. The majority of these identities were further confirmed by characteristic fragment ions obtained by tandem mass spectrometry (MS/MS) of the corresponding precursor ions.

3.2.2. Structural Diversity and Profiling of the Fengycin Class

The LC-ESI-MS analysis of the first cluster of peaks eluting predominantly between 8 and 17 min in all three extracts (cultivated in fermentation media A, B and C) revealed the presence of lipopeptides belonging to the fengycin family, detected predominantly as doubly charged ions [M+2H]²⁺, [M+2Na]²⁺ or [M+H+Na]²⁺ in the m/z range of 724–774. The identification of the lipopeptides was achieved based on the ratio m/z, retention time (RT) and comparison with previously reported molecular ion data of lipopeptides. The results showed the presence of isoforms of fengycin A and fengycin B and their homologues containing saturated and monounsaturated β-hydroxy (β-OH) fatty acid chains ranging from C14 to C19. Mass peaks of some fengicin forms of B. velezensis 5RB were confirmed by LC-MS of the commercial standard Fengicin (with purity ≥ 90%, Sigma-Aldrich Co.) (Figure S1).

Significant differences in the distribution of these homologues were observed depending on the cultivation conditions (Figure 3a–c). As shown in Figure 3a, the dominant ions in the profile of the extract cultured in medium A were doubly charged ions [M+2H]²⁺ at m/z 753.32 (RT 12.9–13.8 min) and [M+2H]²⁺ at m/z 738.31 (RT 14.1–15.7 min), respectively, identified as C17 fengycin B (28.5%) and C17 fengycin A with monounsaturated fatty acid (19.4%) (

Table 2. Identified main mass peaks of fengycin in extract of B. velezensis 5RB cultivated in medium A based on LC-MS and LC-MS/MS analyses.

Peak	Retention Times (min)	Observed Ion (m/z)	Identified Metabolite	Peak Area	Relative Abundance (%)
6	8.6–9.2	748.4102 [M+2H]2+ 1495.8204 [M+H]+	Linear C17 Fengycin A	6013	1.5
		755.3837 [M+2H]2+ 1509.7674 [M+H]+	Linear C18 Fengycin A	2618	0.6
7	9.5–10.3	762.4221 [M+2H]2+ 1523.8441 [M+H]+	Linear C19 Fengycin A	19,953	4.9
8	10.3–10.7	732.3081 [M+2H]2+ 1463.6161 [M+H]+	C16 Fengycin A/C14 Fengycin B	9248	2.3
9	10.7–11.1	732.2972 [M+2H]2+ 1463.5865 [M+H]+	C16 Fengycin A	9446	2.3
		747.3161 [M+2H]2+ 1493.6222 [M+H]+	C16 Fengycin B	1069	0.3
10	11.1–11.5	769.3884 [M+2H]2+ 1537.7768 [M+H]+	Linear C18 Fengycin B	6066	1.5
11	11.5–11.9	746.3103 [M+2H]2+ 1491.6128 [M+H]+	C16 Fengycin B/C17 Fengycin B2	19,047	5.0
		757.2995 [M+2Na]2+ 1513.5917 [M+Na]+		1615
12	11.9–12.5	739.3040 [M+2H]2+ 1477.6237 [M+H]+	C17 Fengycin A	26,438	7.1
		750.2928 [M+2Na]2+ 1499.5757 [M+Na]+		2147
13	12.5–12.9	754.3262 [M+2H]2+ 1507.6685 [M+H]+	Linear C17 Fengycin B2 with monounsaturated fatty acid	21,533	5.2
14	12.9–13.8	753.3170 [M+2H]2+ 1505.6490 [M+H]+	C17 Fengycin B	107,731	28.5
		764.3066 [M+2Na]2+ 1527.6048 [M+Na]+		9610
15	13.8–14.1	724.3001 [M+2H]2+ 1469.5711 [M+Na]+	C15 Fengycin A	12,002	3.2
		735.2901 [M+2Na]2+ 1469.5711 [M+Na]+		1019
16	14.1–14.5	740.3580 [M+2H]2+ 1479.7068 [M+H]+	C16 Fengycin B/C18 Fengycin A	14,682	3.6
17	14.5–15.7	738.3138 [M+2H]2+ 1475.6417 [M+H]+	C17 Fengycin A with monounsaturated fatty acid	79,862	19.4
18	15.7–17.0	745.3201 [M+2H]2+ 1489.6555 [M+H]+	C18 Fengycin A with monounsaturated fatty acid	49,445	12.0
		760.3236 [M+2H]2+ 1519.6395 [M+H]+	C18 Fengycin B/C19 Fengycin B2	10,767	2.6
			Total	410,311	100%

, Figure 3a). The next high-intensity ions are [M+2H]²⁺ at m/z 745.32 (RT 15.7–17.0 min) and [M+2H]²⁺ at m/z 739.3040 (RT 11.9–12.5 min), corresponding to C18 fengycin A with monounsaturated fatty acid and C17 fengycin A.

The major peaks in the profile of the methanol extract obtained after cultivation of B. velezensis 5RB in medium B were [M+2H]²⁺ ions at m/z 745.32 (RT 16.0–16.5 min), m/z 738.33 (RT 14.3–14.8 min), and m/z 760.33 (RT 14.8–15.3 min), which correspond to monounsaturated C18 fengycin A, monounsaturated C17 fengycin A, and C18 fengycin B/C19 fengycin B2, respectively.

The dominant homologues in the methanol extract from medium C are C16 fengycin B (with relative abundance 31.2%), detected as [M+2H]²⁺ ion at m/z 747.3894 (Figure 3c), followed by the linear form of C16 fengycin B/linear C18 fengycin A (18.9%) and linear C19 fengycin A (16.7%), identified as [M+2H]²⁺ ions at m/z 754.3977 and m/z 762.3825, respectively (

Table 3. Identified main mass peaks of fengycin in B. velezensis 5RB extract cultivated in medium C based on LC-MS and LC-MS/MS analyses.

Peak	Retention Times (min)	Observed Ion (at m/z)	Identified Metabolite	Peak Area	Relative Abundance (%)
11	9.2–10	755.3836 [M+2H]2+ 1509.7672 [M+H]+	Linear C18 Fengycin A/Linear C16 Fengycin B	1279	1.8
		748.4105 [M+2H]2+ 1495.821 [M+H]+	Linear C17 Fengycin A	1879	2.6
12	10.9–11.3	773.3824 [M+H+Na]2+	Linear C19 Fengycin A	6755	16.7
		762.4221 [M+2H]2+ 1523.8442 [M+H]+		5413
		733.3732 [M+2H]2+ 1465.7336 [M+H]+	C14 Fengycin A2	3683	10.6
		744.3641 [M+H+Na]2+		4028
13	11.6–12.0	747.3894 [M+2H]2+ 1493.7688 [M+H]+	C16 Fengycin B	11,471	31.2
		758.3797 [M+H+Na]2+		11,179
14	13.8–14.1	754.3977 [M+2H] 2+ 1507.7954 [M+H]+	Linear C17 Fengycin B2/Linear C16 Fengycin B with monounsaturated	7806	18.9
		765.3881 [M+H+Na]2+	fatty acid	5879
		753.3170 [M+2H]2+ 1505.6490 [M+H]+	C17 Fengycin B/C19 Fengycin A	2069	2.9
15	16.2–16.5	769.4364 [M+2Na]2+ 1515. 8628 [M+Na]+	C18 Fengycin A/C16 Fengycin B	2245	3.1
		782.3721 [M+2Na]2+ 1541.7348 [M+Na]+	C18 Fengycin B/C19 Fengycin B2	2225	7.7
		760.3536 [M+2H]2+ 1519.7348 [M+H]+		3356
		745.3201 [M+2H]2+ 1489.6402 [M+H]+	C17 Fengycin B2 with monounsaturated fatty acid	3298	4.5
			Total	72,545	100%

, Figure 3c).

The relative abundance values presented in Table 2 and Table 3 were calculated by summing the peak areas of all detected homologues within the fengycin family and expressing them as percentages of the total peak area.

The main difference between the fengycin profiles of the extracts cultured in the three different media was the significantly higher total content (by approximately two orders of magnitude) of fengycin compounds in the extract produced by medium A compared to those from media B and C. Overall, cultivation of B. velezensis 5RB in medium A resulted in the greatest diversity of fengycin isoforms and homologues, followed by medium C, in contrast to medium B.

The structural characterization of the different fengycin forms determined by LC-ESI-MS was confirmed by LC-ESI-MS/MS analyses. Interpretation of the ESI–MS/MS spectrum of the precursor ion at m/z 753.47 (Figure 4a), which dominates the LS-MS profile of the relative quantitative distribution of fengycins, revealed diagnostic ions at m/z 994.43 (y8) and m/z 1108.51 (y9), along with [M+2H]²⁺ at m/z 554.735 (y9), confirming the structure of C17 fengycin B. Identification of fengycin A homologs was achieved by determining other characteristic fragment ions as [M+H]⁺ at m/z 966.46 and m/z 1080.54 and the doubly charged ion [M+2H]²⁺ at m/z 540.77, observed in the MS/MS spectrum of the precursor ion [M+2H]²⁺ at m/z 739.36, confirming C17 fengycin A (Figure 4b). The identified fragment ions of fengycin B differ from those of fengycin A by 28 Da due to the substitution of Ala at position 6 in the amino acid sequence of fengycin A with Val 6 in the amino acid chain of fengycin B. The above-mentioned fragment ions can be used as fingerprints for the rapid detection of homologues of fengycin A and B [34]. The presented isoforms of fengycin differ in molecular mass, elution time, and amino acid sequence (AAS), but both isoforms contain a C17 β-OH fatty acid.

In this way, other fengycin homologs were also confirmed. The observed high microheterogeneity in fengycins across all three B. velezensis 5RB extracts, confirmed by LC-MS/MS analyses, is due to (i) variations in chain length and the presence of unsaturation and (ii) variations in AAS at positions 6 (Ala, Val, or Ile/Leu) and 10 (Val or Ile).

3.2.3. Structural Diversity and Profiling of Lipopeptides Belonging to the Surfactin Family

Lipopeptides belonging to the surfactin family demonstrated the highest relative abundance among all detected metabolites across all three extracts of B. velezensis 5RB when cultivated in media A-C. The various surfactin forms, predominantly detected as protonated molecular ions [M+H]⁺ and sodium adducts [M+Na]⁺ in the m/z range of 994–1090 (Figure 5), were interpreted in light of previously published data for surfactin molecular ions [21,31,32,33,34,35,36,37]. Additionally, mass peaks of some surfactin forms found in B. velezensis 5RB extracts were confirmed by LC-MS of the commercial standard surfactin Fengicin (with purity ≥ 90%, Sigma-Aldrich Co.) (Figure S2).

In the methanol extract of secondary metabolites of B. velezensis 5RB cultivated in medium A, a significant diversity of cyclic and linear forms of surfactin A and B bearing chains C13-C16 β-hydroxy fatty acid was eluted between 17 and 27 min. The highest abundance was observed for [M+H]⁺ at m/z 1036.5, followed by 1022.5, corresponding to C15 surfactin A and C14 surfactin A (Figure 5a). In contrast, cultivation in medium C shifted the metabolic profile toward [M+H]⁺ at m/z 1050.7 and [M+Na]⁺ at m/z 1072.7, corresponding to C16 surfactin A, which emerged as the dominant homolog under these conditions (Figure 5c). In medium B, the surfactin pool was once again dominated by C15 surfactin A, as evidenced by intense peaks for [M+H]⁺ at m/z 1036.5 and [M+Na]⁺ at m/z 1058.5, followed by C16 surfactin A represented as [M+H]⁺ at m/z 1050.6 and [M+Na]⁺ at m/z 1072.6 (Figure 5b). Long-chain β-hydroxy fatty acid forms of surfactin were detected in all three extracts as [M+H]⁺ at m/z 1064.7 and [M+Na]⁺ at m/z 1086.7, corresponding to C17 surfactin A/C18 surfactin B.

The presented results (Figure 2 and Figure 5) clearly demonstrate that among all detected lipopeptide components, surfactin homologues constitute the predominant fraction in B. velezensis 5RB extracts, irrespective of the culture medium utilized.

Following their initial detection via LC-ESI-MS, the surfactin isoforms were confirmed and structurally characterized using tandem mass spectrometry (ESI-MS/MS). ESI-MS/MS conditions induced cleavage of the cyclic surfactin lactone ring, leading to the identification of characteristic fragmentation profiles based on y- and b-product ion series and enabling the determination of the AAS.

ESI-MS/MS analysis of the [M+H]⁺ precursor ions at m/z 1036.6502 (RT 23.4–24.2 min) yielded a series of product ions shown in the spectrum on Figure 6a. The ion at m/z 1018.6372 corresponded to the loss of water (−18 Da) from the precursor. The remaining product ions were identified as two product ion series deriving from the initial opening of the lactone ring. The first series (b-fragmented ions) includes the fatty acid chain alongside the N-terminal amino acids, yielding b₇ at m/z 923.54, b₆ at m/z 810.46, b₅ at m/z 695.44 and b₄ at m/z 596.38, corresponding respectively to the loss of Leu (113.08 Da), Leu-Leu (226.16 Da), Leu-Leu-H₂O (244.16 Da), and Leu-Leu-Asp-Val (440.26 Da) from the precursor ion at m/z 1036.6502. The second series (comprising C-terminal product ions) contains the peptidic fragments at m/z 685 (y6+H₂O), 554.30 (y5), and 441.22 (y4). The most intense ion at m/z 685.39 (y6+H₂O) corresponded to the loss of the C15 β-hydroxy fatty acid chain linked with Glu (~351 Da) from the precursor ion via a double hydrogen transfer mechanism. This fragment ion is a diagnostic ion for the surfactin A isoform. Interpretation of the MSMS spectrum in Figure 5a confirmed the structure of C15 surfactin A. In this way, other forms of surfactin were also identified (

Table 4. Major mass peaks of lipopeptides from the surfactin family in an extract of B. velezensis 5RB cultured in medium A detected by LC-ESI-MS.

Peak	Retention Times (min)	Observed Ion (m/z)	Identified Metabolite	Peak Area	Relative Abundance (%)
19	17.1–18.6	1040.6442 [M+H]+	Linear C14 Surfactin A	30,370	3.2
		1062.6235 [M+Na]+		18,832
20	18.7–19.7	1054.6782 [M+H]+	Linear C15 Surfactin A	44,886	4.7
		1076.6586 [M+Na]+		28,305
21	19.7–20.2	1022.6255 [M+H]+ 1044.6048 [M+Na]+	C14 Surfactin B*	27,103 19,061	3.0
22	20.2–21.0	1008.6120 [M+H]+	C13 Surfactin A/C14 Surfactin B	63,921	7.1
		1030.5915 [M+Na]+ 515.7361 [M+2Na]2+		45,607
23	21.1–21.9	1036.6399 [M+H]+	C14 Surfactin B	46,982	4.9
		1058.5193 [M+Na]+		29,095
24	22.0–23.2	1022.6256 [M+H]+ 1044. 6052 [M+Na]+	C14 Surfactin A	343,298	22.3
25	23.4–24.2	1036.6401 [M+H]+ 518.8641 [M+2H]2+	C15 Surfactin A	527,433	37.7
		1058.6188 [M+Na]+		53,822
26	24.2–24.7	1022.6272 [M+H]+ 1044.6070 [M+Na]+	C15 Surfactin B	42,711 12,586	3.6
27	24.9–25.7	1036.6422 [M+H]+	C16 Surfactin B	71,820	6.4
		1058.6212 [M+Na]+		26,227
		1050.6556 [M+H]+	C16 Surfactin A	52,799	5.1
		1072.6346 [M+Na]+		25,929
28	25.7–26.3	1036.6411 [M+H]+	C16 Surfactin B	17,936	1.2
29	26.3–26.9	1050.6545 [M+H]+	C16 Surfactin A/C17 Surfactin B	4028	0.5
		1072.6344 [M+Na]+		2696
		1064.6688 [M+H]+	C17 Surfactin A	3990	0.5
		1086.6487 [M+Na]+		3451
			Total	1,542,888	100.0

and

Table 5. Main mass peaks of surfactins detected by LC-ESI-MS in an extract of B. velezensis 5RB cultured in medium C.

Peak	Retention Times (min)	Observed Ion (m/z)	Identified Metabolite	Peak Area	Relative Abundance (%)
20	22–23.2	1022.6017 [M+H]+	C14 Surfactin A	203,250	19.9
		1044.6058 [M+Na]+		32,012
		1030.6385 [M+Na]+	C14 Surfactin B	43,379	5.7
		1008.6486 [M+H]+		23,357
21	23.8–24.2	1036.6820 [M+H]+	C15 Surfactin A	50,769	13.1
22	24.2–24.6	1058.664 [M+Na]+		104,478
23	25.0–25.9	1050.7004 [M+H]+	C16 C Surfactin A	161,025	25.2
		1072.6853 [M+Na]+		136,872
		1036.6882 [M+H]+	C16 C Surfactin B	4830	0.4
24	25.9–26. 4	1050.7131 [M+H]+	C17 Surfactin B	201,282	30.2
		1072.6945 [M+Na]+		156,407
25	26.4–26.7	1064.7166 [M+H]+	C17 Surfactin A	19,658	1.9
		1086.6881 [M+Na]+		2489
26	26.7–27.1	1064.7141 [M+H]+	C17 Surfactin A/C18 Surfactin B	36,756	3.6
		1086.6983 [M+Na]+		7467
			Total	1,184,031	100.0

).

Interpretation of the precursor ion at m/z 1022.627 (RT 23.4–24.2 min) revealed a fragmentation profile distinct from that shown in Figure 5a, characteristic of the surfactin B isoform (Figure 5b). Its MS/MS spectrum is dominated by the C-terminal fragment ion [y6+H₂O]⁺ at m/z 671.5, which is a diagnostic ion for the surfactin B isoform. Additional fragments were assigned as y₃ at m/z 328.14 (DLV), y4 at m/z 441.1 (VDLV), and y₅ at m/z 540.29 (LVDLV), fully supporting the sequence structure of C15 surfactin B. The observed series of b-ions from the N-terminus confirms the proposed structure of C15 surfactin B with the AAS Glu¹-Leu/Ile²-Leu³-Val⁴-Asp⁵-Leu⁶-Val⁷ (Figure 5b).

Although typical fragmentation of most surfactin A and B precursors results in the characteristic diagnostic ions mentioned above, some exceptions have been observed. For instance, MS/MS fragmentation analysis of the precursor ion at m/z 1022.6056 (RT 19.7–20.2) (Figure 6c) generated a prominent series of b-type ions at m/z 356.2 (b₂), 451.26 (b₃), 469 (b₃ + H₂O), 582.36 (b₅), 810.45 (b₆), and 909.53 (b₇) from the N-terminus and y-type ions at m/z 441.02 (y₄), 554.30 (y₅), and 685.39 (y₆+H₂O). Despite the presence of an intense diagnostic ion at m/z 685.39, the interpretation elucidated a specific surfactin B* isoform with the amino acid sequence Glu¹–Leu/Ile²–Leu³–Leu/Ile⁴–Asp⁵–Leu⁶–Val⁷. This sequence corresponds to a novel C14 Surfactin B* isoform that has not been previously reported in B. velezensis strains. To our knowledge, a structurally comparable isoform has only been documented in a methanol extract from marine Bacillus megaterium [35]. The proposed structure is unequivocally supported by the diagnostic y- and b-type fragment ions observed in the MS/MS spectrum. For comparison, the fragmentation pattern of C15 Surfactin B, which shares the same molecular mass but differs in both amino acid sequence and β-hydroxy fatty acid chain length, is also presented (Figure 6b).

3.2.4. Identification of Linear Forms of Fengycin and Surfactin Produced by B. velezensis 5RB

The presence of linear homologues of fengycin and surfactin produced by B. velezensis 5RB, initially predicted via LC-MS, was unequivocally confirmed through LC-MS/MS analyses. The linear form of C17 fengycin A (Ala 6) was validated by ESI-MS/MS analysis of the doubly charged precursor ion [M+2H]²⁺ at m/z 748.40, which showed a series of characteristic y- and b-ions at m/z 520.26 (y4), 804.39 (y7), 675.44 (b3), 905.51 (b5) and 1201.64 (b8) (Figure 7). Using this approach, four linear homologues of fengycin A and two linear homologues of surfactin A were determined (

Table 6. Linear forms of fengycin and surfactin identified in a strain of B. velezensis 5RB confirmed by ESI-MS/MS.

Lipopeptide	Observed Ion (m/z)	Molecular Mass (Da)	Amino Acid Sequence
Linear C17 Fengycin A	748.4102 [M+2H]2+ 1495.8204 [M+H]+	1494.84	Glu-Orn-Thr-Glu-Ala-Pro-Gln-Tyr-IleCOOH
Linear C18 Fengycin A	755.3806 [M+2H]2+ 1509.7612 [M+H]+	1508.84	Glu-Orn-Thr-Glu-Ala-Pro-Gln-Tyr-IleCOOH
Linear C19 Fengycin A	762.4221 [M+2H]2+ 1523.8442 [M+H]+	1522.85	Glu-Orn-Thr-Glu-Ala-Pro-Gln-Tyr-IleCOOH
Linear C20 Fengycin A/ Linear C18 Fengycin B	769.3884 [M+2H]2+ 1537.7768 [M+H]+	1536.87	Glu-Orn-Thr-Glu-Ala-Pro-Gln-Tyr-IleCOOH/ Glu-Orn-Thr-Glu-Val-Pro-Gln-Tyr-IleCOOH
C14 Surfactin A	1040.6442 [M+H]+	1039.68	Glu-Leu/Ile-Leu-Val-Asp-Leu-LeuCOOH
C15 Surfactin A	1054.6782 [M+H]+	1053.69	Glu-Leu/Ile-Leu-Val-Asp-Leu-LeuCOOH

). Crucially, these linear structures are reported here for the first time in a methanol extract of a B. velezensis strain. Similar linear lipopeptide forms have been previously found in culture filtrates of B. subtilis C-1 isolated from petroleum sludge [36] and in methanol extracts of marine B. megaterium [35].

3.2.5. Polyketides Identified in Free-Cell Extracts of c 5RB

In addition to lipopeptides, analysis of the cell-free extracts from the fermentation broth of B. velezensis 5RB (cultivated in media A–C) revealed the presence of three distinct polyketide classes: macrolactins, dificidin, and bacillaenes. These secondary metabolites were detected via LC-ESI-MS as protonated molecular ions [M+H]⁺ and their corresponding sodium [M+Na]⁺ or potassium [M+K]⁺ adducts (

Table 7. Polyketides identified in a strain of B. velezensis 5RB, cultivated in medium C.

Peak	Retention Time (min)	Observed Ion (m/z)	Theoretical Mass (Da)	Relative Error Δppm (ppm)	Molecular Formula	Identified Metabolite	Peak Area	Relative Abundance (%)
3	2.8–3.1	581.3538 [M+H]+	581.3590	−8.95	C34H48N2O6	Bacillaene A	2653
		603.3414 [M+Na]+	603.3410	+0.66			921	2.0
		619.3195 [M+K]+	619.3149	+7.43			5094
4	3.1–4.4	905.2484 [M+Na]+	905.2476	+2.98	C39H42N6O18	Bacillibactin	28,078	6.7
		883.2664 [M+H]+	883.2637	+3.06			1384
		567.2885 [M+Na]+	567.2852	+5.82	C31H45O6P	Dificidin	21,751	5.0
5	4.6–5.1	583.3724 [M+H]+	583.3742	−3.08	C34H50N2O6	Dihydrobacillaene	9496	2.8
		605.3548 [M+Na]+	605.3567	−3.14			2855
		527.2022 [M+ K]+	488.2410	−3.79	C27H36O8	7-O-Malonyl macrolactin A	23,904	5.5
8	6.5–6.8	703.2745 [M+K]+	703.2745	+1.85	C34H48O13	Macrolactin D	1184	0.4
		687.3005 [M+Na]+	664.3095	+1.75			557
27	25.7–26.3	511.2298 [M+Na]+	511.2308	−1.96	C27H36O8	7-O-Malonyl macrolactin A	73,174	26.6
		527.2018 [M+K]+	527.2042	−4.55			67,808
28	27.8–28.3	525.2416 [M+Na]+	525.2465	−9.33	C28H38O8	7-O-Succinyl macrolactin A	76,491	35.4
		541.2205 [M+K]+	541.2199	+1.11			78,746
29	28.3–28.8	425.2314 [M+Na]+	425.2298	+2.35	C24H34O5	Macrolactin A	25,987	15.6
		441.2065 [M+K]+	441.2038	+4.99			42,219
						Total	438,398	100

), which are in very good agreement with both the theoretically calculated monoisotopic masses and the previously reported macrolactins, bacillibactin, bacillaenes and dificidin produced by B. velezensis 32a [38]. Additionally, the identity of the detected polyketides was confirmed by comparing the observed isotopic distributions of the most intense ions in the LC-MS spectra with their theoretical models, showing good agreement (Figure S3), based on the natural abundance of stable isotopes of the elements constituting the given molecule. The results show a high agreement of the characteristic peak profile of the experimentally determined ions with the corresponding theoretical profile (Table 7, Figure S3).

The relative abundance of each polyketide was expressed as a percentage by integrating its main peak area relative to the total peak area of all three polyketide classes (Table 7). It was found that 7-O-succinylmacrolactin A was the most abundant derivative (35%), followed by 7-O-malonylmacrolactin A and macrolactin A, whereas macrolactin D and bacillaene A were the least abundant components.

3.2.6. Identification of Ribosomally Synthesized Sactipeptide Subtilosin A

For the first time, LC-ESI-MS and ESI-MS/MS analyses established that the metabolic profile of the purified, cell-free extract of B. velezensis 5RB, cultivated in media A and B, includes the antimicrobial sactipeptides subtilosin A and its modification A1. Based on LC-MS analysis, subtilosin A was identified in the medium A extract by [M+4H]⁴⁺ at m/z 850.7762, [M+3H]³⁺ at m/z 1134.0354 and [M+2H]²⁺ at m/z 1700.5803 (Figure 8a). In the medium B extract, subtilosin A was detected via [M+3H]³⁺ at m/z 1134.0537 and [M+3Na]³⁺ at m/z 1141.3812 (Figure 8b). Subtilosin A1 was detected exclusively in the medium B extract via [M+3H]³⁺ at m/z 1138.4 and confirmed by [M+3Na]³⁺ at m/z 1146.41 (Figure 8b). A high agreement was found between the calculated isotopic molecular mass of the peptide (3399.0951 Da) and the molecular mass of subtylosin A (3399.7 Da) previously determined by MALDI-TOF-MS as reported in [39]. Subtylosin A1 is a variant of subtylosin A, distinguished by the substitution of a Thr residue at position 6 in the cyclic chain by an Ile residue. The experimentally determined monoisotopic mass from the LC-MS analysis is consistent with the reference monoisotopic mass for subtilosin A1 ([M+H]⁺ ion at m/z 3412.5 Da or approximately 3411.5 Da); this represents a +12 Da mass shift compared to subtilosin A, corresponding to the substitution of Thr with Ile [40].

The presence of sactipeptides was also confirmed by ESI-MS/MS analysis of the [M+3H]³⁺ precursor ion at m/z 1134.0537. The interpretation of the fragment ions is shown in Figure 9.

The macrocycle of subtilosin A is stabilized by three internal thioether bridges (Cys4–Phe31, Cys⁷–Thr²⁸, and Cys¹³–Phe²²), which block standard linear fragmentation and result in an ESI–MS/MS spectrum dominated by internal fragments originating from the flexible loop regions between these bridges.

Monitoring the series of fragment ions at m/z 388.14, 459.18, 516.20, and 587.2325 in the MS/MS spectrum identified the peptide fragment FEIAGA (Figure 9). The observed mass of 587.23 Da for FEIAGA reflects the loss of one hydrogen atom (1 Da) from Phe22, confirming its involvement in a saccipeptide modification with Cys13. The series of [M+H]⁺ ions at m/z 630.24, 694.29, 747.28, 838.35, and 976.4 indicates internal fragmentation of the peptide chain at positions 6–15. The hydrophobic pentapeptide fragment GAACL, containing cysteine, is commonly observed during ring opening and internal fragmentation in Subtilosin A. The spectrum shows a series of peptide [M+H]⁺ ions at m/z 630.24, 694.29, 747.28, 838.35, and 976.4, confirming sequential internal fragmentation from the same peptide region. The assigned ion [M+H]⁺ at m/z 976.4, corresponding to TCSIGAACLV, was obtained after opening of the macrocycle and cleavage near the thioether-linked Cys7, resulting in the generation of a series of internal ions shown in the spectrum. The fragment ions confirming the presence of subtilosin A identified based on the MS/MS spectrum are presented in

Table 8. MS/MS spectrum fragment ions confirming subtilosin A in media A and B.

Observed Ion (m/z)	Residue Range	Description	Fragment Sequence
388.1427	22–24	Tetrapeptide fragment from the hydrophobic loop	FEI
459.1769	22–25	Sub-fragment of the FEIAGA series	FEIA
516.1988	22–26	Sub-fragment of the FEIAGA series	FEIAG
587.2325	22–27	Sub-fragment of the FEIAGA series	FEIAGA
480.1984	16–20	Sub-fragment	DGPIP
595.2196	16–21	Sub-fragment of a proline-rich loop located between Cys13 and Phe22	DGPIPD *
630.24	10–14	Sub-fragment	GAACL
694.29	9–14	Sub-fragment	IGAACL
747.28	8–14	Sub-fragment	SIGAACL
838.35	6–14	Sub-fragment	TCSIGAACL
856.37	6–14	Sub-fragment	TCSIGAACL+H2O
976.4	6–15	Large internal fragment containing the previously identified sulfur-rich core, after ring-opening	TCSIGAACLV

* This sub-fragment is prone to cleavage as it is not directly linked by a thioether bond.

.

The experimentally identified sactipeptide subtilosin A in an extract of B. velezensis 5RB (cultivated in fermentation media A and B) was also confirmed by the complete genome and in silico analysis of the B. velezensis 5RB strain. A putative subtilosin biosynthetic gene cluster was identified on contig 2 of the draft genome assembly (between 299 and 320 kb), supporting the subsequent LC-MS identification of subtilosin A.

3.3. Genome Mining of Biosynthetic Gene Clusters in B. velezensis 5RB

To validate the experimentally determined metabolite profile, genome mining of B. velezensis 5RB was performed using the draft genome sequence (DDBJ/ENA/GenBank accession QXJL00000000) [29]. The draft genome comprises 26 contigs and harbors multiple biosynthetic gene clusters (BGCs) encoding non-ribosomal peptide synthetases (NRPS), polyketide synthases (PKS), and ribosomally synthesized and post-translationally modified peptides (RiPPs). The identified BGCs correspond to several of the major antimicrobial metabolites detected by LC–MS/MS. The major biosynthetic gene clusters identified in the draft genome are distributed across several genomic contigs. The surfactin (srf) NRPS cluster is located on contig 7, whereas the fengycin (fen) cluster is located on contig 4, together with the bacillaene (bae) PKS/NRPS hybrid cluster. Additional PKS clusters associated with macrolactin and difficidin biosynthesis were detected on contigs 3 and 1, respectively. The bacillibactin (dhb) siderophore biosynthetic cluster was identified on contig 9. In addition, genome analysis supported the presence of a putative subtilosin A/A1 biosynthetic gene cluster on contig 2 of the draft genome assembly, spanning approximately 299–320 kb.

3.4. Comparative Distribution of Antimicrobial Metabolites in the Selected Production Media

To compare metabolic profiles across the selected production media, the relative abundance of major antimicrobial metabolite groups was estimated from the LC-MS peak areas of the extracts (

Table 9. Relative distribution of surfactins, fengycins, subtilosins, and polyketides in extracts obtained from the selected production media.

Metabolite Group	Medium A (%)	Medium B (%)	Medium C (%)
Surfactins	44.3 ± 1.0	34.0 ± 1.0	27.5 ± 1.0
Fengycins	15.5 ± 1.0	3.0 ± 1.0	2.5 ± 1.0
Subtilosins	0.26 ± 0.02	0.50 ± 0.04	–
Polyketides	1.4 ± 0.02	3.49± 0.02	8.85 ± 0.50
Relative ratio	Surfactins:Fengycins:Subtilosins:Polyketides = 2.8:1:0.02:0.9	Surfactins:Fengycins:Subtilosins:Polyketides = 11.3:1:0.17:1.2	Surfactins:Fengycins:Polyketides = 11:1:3.5

).

Surfactins were the predominant metabolite family in all three media, accounting for 44.3%, 34.0%, and 27.5% of detected compounds in media A, B, and C, respectively. Medium A also promoted the highest accumulation of fengycins (15.5%), whereas their abundance decreased markedly in media B (3.0%) and C (2.5%). In contrast, subtilosins were detected only in media A and B, with the highest relative abundance in medium B (0.50%). Polyketides were produced almost exclusively in media C (8.85%), compared to those produced in media B (3.49%) and A (1.4%). The production media directed the metabolism of B. velezensis 5RB toward distinct antimicrobial profiles, ranging from lipopeptide-rich extracts (medium A), through subtilosin-enriched extracts (medium B), to polyketide-enriched extracts (medium C). The high antifungal activity observed in medium C coincided with a pronounced enrichment in polyketides, suggesting that these metabolites may contribute to the enhanced activity. However, the substantially greater activity of the unfractionated metabolite mixture compared with the isolated lipopeptide fractions indicates that the antifungal phenotype is likely associated with the combined action of multiple metabolite classes rather than with a single metabolite group.

3.5. Antifungal Activity Against B. cinerea of Lipopeptide Fractions Isolated by RP-HPLC from Cell-Free Supernatant of B. velezensis 5RB

The metabolite extract from B. velezensis 5RB cultivated in medium C was separated by RP-HPLC, yielding two high-abundance fractions. After LC-MS analysis, it was found that the fraction with a retention time (RT) 12.3 min, eluted at 71% Buffer B (Fraction I), contained primarily lipopeptides from the fengycin family, with molecular masses ranging from 1465 Da to 1520 Da. In contrast, the second fraction (Fraction II, RT 17.2 min, eluted at 83% Buffer B) contained mainly homologues of surfactins A and B, with molecular masses between 1022.0 and 1064.6 Da.

Evaluation of antifungal activity against B. cinerea revealed that Fraction I inhibited fungal growth by 43.64% to 51.19% after 24 h. This inhibitory effect peaked at 48 h (45.39–53.55%) before declining to 41.73% (at 44.6 µg/mL) after 72 h, indicating that lipopeptides containing different forms of fengycin provide strong initial suppression but limited persistence over time. In contrast, Fraction II (dominated by surfactins) exhibited a markedly lower inhibition rate of 36.65 ± 0.5% after 48 h (at 44.6 µg/mL), which further decreased to 27.78–30.57% by 72 h. These results confirm that lipopeptides of surfactins possess significantly lower antifungal activity compared to fengycins (

Table 10. Antifungal activity of HPLC-purified fractions from cell-free extract from B. velezensis 5RB grown for 24 h in medium C against B. cinerea.

Purified Fraction (min)	Main Metabolites	Concentration (µg/mL)	Growth Inhibition (%) at 24 h	Growth Inhibition (%) 48 h	Growth Inhibition (%) at 72 h
Fraction I	Mostly fengycin A and B isoforms and their homologues	11.2	43.64 ± 0.5	47.09 ± 0.5	36.45 ± 0.5
(Rt 12.3 min,		22.3	47.67 ± 0.5	45.39 ± 0.4	39.91 ± 0.5
Buffer B 71%)		44.6	51.19 ± 0.5	53.55 ± 0.6	41.73 ± 0.4
Fraction II	Predominantly Surfactin A and B and their homologues	11.2	35.91 ± 0.4	31.51 ± 0.4	27.78 ± 0.4
(Rt 17.2,		22.3	37.28 ± 0.5	32.20 ± 0.5	28.05 ± 0.4
Buffer B 83%)		44.6	39.09 ± 0.5	36.65 ± 0.5	30.57 ± 0.4
Whole cell-free supernatant	All metabolites	Complex mixture (not quantified)	98.3 ± 3.0	86.6 ± 2.8	53.8 ± 2.8

). It is characteristic of both studied fractions that increasing the concentration does not lead to a significant increase in the degree of inhibition, which indicates the presence of a saturation effect, typical of compounds with a membrane-directed mechanism of action.

In contrast to the activity of the individual fractions, the unfractionated cell-free supernatant of B. velezensis 5RB showed almost complete inhibition of B. cinerea growth (98.3 ± 3.0% at 24 h) (Table 10). Therefore, these results suggest a clear combination effect between the different classes of bioactive compounds produced by B. velezensis 5RB. The interaction between lipopeptides (fengycins and surfactins), among other secondary metabolites (such as polyketides and subtilosin A/A1), leads to enhanced antifungal activity that exceeds that of the isolated fractions. The observed antifungal effect is likely due to a combination of actions on different cellular targets, including cell membrane disruption and inhibition of cell wall synthesis.

3.6. Ex Vivo Biocontrol Activity of the Fermentation Broth of Bacillus velezensis 5RB Against Botrytis cinerea and Phytophthora infestans

The protective activity of the fermentation broth of B. velezensis 5RB was evaluated using detached tomato leaves infected with B. cinerea and Ph. infestans (Figure 10).

In untreated controls, both pathogens rapidly developed the infection and produced expanding necrotic lesions over a five- to seven-day incubation period. Lesion development was particularly severe on water-treated control leaves inoculated with B. cinerea. Necrotic lesions expanded rapidly from the inoculation site, reaching approximately 8 mm on the fifth day and 12.0–14.5 ± 1.0 mm on the seventh day, resulting in extensive tissue maceration. In contrast, leaves pretreated with B. velezensis 5RB fermentation broth remained largely protected throughout the experiment. Only a small, localized necrotic spot, typically 1–3 mm in diameter, developed at the inoculation site, with no visible lesion expansion or secondary tissue colonization. A similar high protective effect was observed against Ph. infestans, with disease progression markedly suppressed compared with the water-treated control leaves, which had lesions measuring 19.5–21.5 ± 1.0 mm on the fifth day and were completely yellow, wilted, and destroyed on the seventh day. It should be noted that the detached-leaf assay was performed using the fermentation broth rather than a cell-free supernatant or a purified metabolite fraction. Consequently, the protective effect observed in this assay cannot be attributed solely to the metabolites present in the fermentation broth. Viable B. velezensis 5RB cells may also contribute to the observed antifungal effect, either through direct antagonistic activity or by continuing to produce bioactive metabolites during the assay. Thus, the detached-leaf results reflect the combined biological activity of the fermentation broth and should not be interpreted as evidence of the activity of individual metabolites alone. The biocontrol effect observed in vivo on foliage is a cumulative result of both the pre-formulated metabolites and the live bacterial cells, which may contribute via tissue colonization, nutrient competition, and continued in situ synthesis of antifungal compounds.

4. Discussion

Modern plant protection strategies, which rely heavily on the breeding of resistant crop cultivars and the intensive application of synthetic chemical fungicides, incur substantial economic costs and raise serious concerns regarding environmental safety and human health [41,42]. In the search for sustainable and eco-friendly alternatives, members of the genus Bacillus have been extensively investigated as antagonistic microorganisms due to their widespread distribution, rapid growth rates, high stress tolerance, and plant growth-promoting capabilities [43,44,45,46,47,48].

Various Bacillus strains synthesize distinct profiles of cyclic lipopeptides—either as discrete families or as complex mixtures—as secondary metabolites possessing diverse hydrophilic and hydrophobic domains [23]. Among them, B. velezensis strains have emerged as particularly promising in modern agricultural biotechnology due to their efficacy against phytopathogens across a broad taxonomic spectrum, ranging from oomycetes (Phytophthora) to necrotrophic fungi (Botrytis) and mycotoxigenic strains (Aspergillus spp.), thereby highlighting their potential as viable alternatives to chemical treatments.

In the present study, we focused on the qualitative modulation of the secondary metabolome of B. velezensis 5RB under varying cultivation regimes. The extracellular metabolites present in the culture supernatants of various B. velezensis strains predominantly comprise lipopeptides and polyketides. Lipopeptides (particularly members of the iturin and fengycin families) are renowned for their antifungal activity against plant pathogens. In contrast, surfactins primarily exhibit antibacterial efficacy, although they can also display antifungal activity against specific fungal pathogens. In addition to lipopeptides, polyketides (such as macrolactins, bacillaenes, and difficidins) act predominantly as antibacterial agents [47,48]. Despite macrolactins being recognized primarily for their antibacterial action, a growing body of scientific literature demonstrates their inhibitory activity against specific fungal pathogens, including key phytopathogens such as B. cinerea and Fusarium species [49,50].

For the first time, a precise qualitative characterization of the extracellular antimicrobial metabolites produced by B. velezensis 5RB is presented, based on LC-ESI-MS and tandem mass spectrometry (LC-MS/MS) analyses. The discovered compounds, derived from non-ribosomal and ribosomal biosynthetic pathways, demonstrate the antimicrobial capacity of B. velezensis 5RB, which produces a remarkable diversity of lipopeptides—in particular surfactants and fengycins—along with four macrolactylan analogs (macrolactin A, 7-O-malonyl macrolactin A, 7-O-succinyl macrolactin A, and macrolactin D), bacillaen A, dihydrobacillaen, dificidin, and the sacipeptides subtylosin A/A1. This rich profile of secondary metabolites suggests a broad-spectrum antimicrobial activity driven by complex, potentially complementary mechanisms.

In recent years, numerous studies have validated biocontrol strategies for B. velezensis, which encompass direct antagonism—through lipopeptide mixtures, highly active cell wall-degrading enzymes, and volatile organic compounds—as well as induction of systemic resistance in host plants [51,52,53,54,55,56,57,58,59,60,61,62,63,64,65] (Table S1).

Among plant pathogens, B. cinerea and Ph. infestans represent two of the most economically critical targets for biological control [7]. B. cinerea, the causative agent of gray mold, infects over 200 plant species and causes severe pre- and post-harvest losses in vegetables, fruits, and ornamental crops [66]. In contrast, Ph. infestans, the causal agent of late blight, is an oomycete and remains one of the most destructive pathogens of tomatoes and potatoes worldwide [2,4]. Their distant taxonomic positions, distinct infection strategies, and differing cell wall compositions make this pair of phytopathogens an ideal model for evaluating the antifungal activity of novel biocontrol strains such as B. velezensis 5RB.

In the present study, mass spectrometric analyses of three extracts from B. velezensis 5RB grown in culture media of varying composition revealed a complex array of metabolites synthesized via both non-ribosomal (lipopeptides and polyketides) and ribosomal (subtilosin A/A1) pathways. The production of antimicrobial metabolites by B. velezensis 5RB, as with other Bacillus spp., is strongly influenced by cultivation conditions, particularly medium composition, carbon and nitrogen sources, and the availability of minerals [13]. For the first time, based on LC–ESI–MS and LC–ESI–MS/MS analyses, the metabolic profiles of B. velezensis 5RB extracts cultivated in three distinct media are presented and comparatively evaluated.

It was established that minor alterations in the mineral composition of the cultivation medium redirect secondary metabolism toward distinct classes of bioactive compounds. Medium A, characterized by lower mineral supplementation and the absence of MnSO₄, FeSO₄, and K₂HPO₄, favored the production of a rich diversity of lipopeptides. We identified 16 homologs predominantly comprising fengycins A and B (Table 2), 13 homologs of surfactins A and B (Table 4), subtilosin A, and a specific relative quantitative ratio among the compositional components (Table 9). The emergence of molecules with identical mass characteristics (m/z) but distinct retention times (Table 2 and Table 4) clearly indicates structural microheterogeneity within the two primary lipopeptide groups—fengycin and surfactin. Notably, we elucidated the structure of a specific C14 surfactin B* isoform (Leu/Ile⁴, Val⁷), characterized by the amino acid sequence Glu–Leu/Ile–Leu–Leu/Ile–Asp–Leu–Val linked to a C14 β-hydroxy fatty acid chain. To date, a similar structure has only been identified in a study by Ma et al. (2016) in a methanol extract of marine Bacillus megaterium [35] and has not been reported in other B. velezensis strains. Furthermore, several linear surfactin and fengycin homologs were identified that have not been previously reported in the metabolic profiles of other B. velezensis strains (Table 6). Therefore, the observed high antifungal activity against B. cinerea exhibited by the extract of B. velezensis 5RB cultivated in medium A (Figure 1) is precisely due to the specificity of the metabolic profile, which includes a rich diversity of isoforms and homologues of fengycin and surfactin and subtilosin A, although the relative abundance of polyketides is the lowest compared to that determined in the extracts obtained after cultivation of 5RB in media B and C (Table 9). Medium B, formulated with higher salt content, promoted the simultaneous production of subtilosins A/A1 while reducing lipopeptide diversity. Additional trace elements in Medium C increased the production of polyketides (Table 7), including bacillaene, dificidin, macrolactins and the siderophore bacillibactin. These results indicate that mineral composition can modulate the balance between NRPS and PKS-synthesized metabolites in B. velezensis 5RB (as shown in Table 9) and can also affect the structural diversity of lipopeptide homologues and isomers. Medium C promoted polyketide production to the greatest extent compared to the other two nutrient media (Table 7 and Table 9). The increased levels of 7-O-malonyl macrolactin A, 7-O-succinyl macrolactin A and macrolactin A, in combination with long-chain homologues of surfactin A and B (C15–C18) and homologues mainly of fengycin B (known for its higher antifungal potential compared to fengycin A), explain the highest inhibitory activity observed against B. cinerea spores (Figure 1). This hypothesis is supported by the study [38], which demonstrated the strong synergistic effect of dificidin and macrolactins against crown gall disease caused by Agrobacterium tumefaciens, against which lipopeptides showed weak antagonistic activity. The concentrations of the supplemented salts were selected empirically based on commonly used cultivation conditions and were not intended to represent optimized concentrations for antimicrobial metabolite production.

Therefore, the present results show a relationship between the total mineral composition of the culture medium and antifungal activity and do not establish the individual effects or optimal concentrations of specific mineral components. In order to determine their individual and combined effects on the production of secondary metabolites and antifungal activity, further systematic optimization of the mineral composition and salt concentrations will be carried out, using factorial or response-surface experimental designs.

An important finding of this study is the identification of the sactipeptide subtilosin A and its analog, subtilosin A1, among the extracellular antimicrobial compounds of B. velezensis 5RB. To date, subtilosin A has been reported in B. subtilis 168 [67], B. amyloliquefaciens [68], and B. atrophaeus [69], whereas its production by B. velezensis has remained unexplored. Subtilosin A belongs to the recently recognized class of ribosomally synthesized and post-translationally modified bacteriocins known as sactibiotics [70]. Its characteristic structure includes three sulfur-to-α-carbon thioether bridges (Cys⁴–Phe³¹, Cys⁷–Thr²⁸, and Cys¹³–Phe²²) and a head-to-tail cyclization linking the N-terminal asparagine to the C-terminal glycine. These extensive post-translational modifications confer exceptional stability against proteolytic degradation and enable efficient disruption of target cell membranes. The detection of both subtilosin A and subtilosin A1 therefore substantially broadens the known antimicrobial repertoire of the species B. velezensis.

The genome mining results provide genetic support for the metabolite profile obtained by UHPLC-QTOF-MS/MS. The identification of biosynthetic gene clusters for surfactin, fengycin, bacillaene, macrolactin, difficidin, bacillibactin, and subtilosin A/A1 demonstrates that the strain encodes the complete machinery required for this metabolic diversity. In particular, the presence of the srf and fen NRPS clusters provides genomic support for the production of the major lipopeptides detected in the culture supernatants, while the identified PKS and PKS/NRPS hybrid clusters support strain 5RB’s capacity to produce additional antimicrobial polyketides [29]. The concordance between the genome-mining and metabolomic data therefore indicates that the experimentally observed antimicrobial metabolite profile aligns with the biosynthetic potential encoded in strain 5RB’s genome.

Presented results clearly show that the antifungal efficacy of strain 5RB cannot be attributed to a single metabolite class. Bioactivity-guided fractionation showed that the separated fractions exhibited only moderate fungistatic activity, with maximal inhibition of B. cinerea ranging from 28.05 ± 0.4% to 53.55 ± 0.6%, depending on concentration and incubation. Specifically, isolated Fraction I (containing fengycin) and Fraction II (containing surfactin) showed substantially lower individual activities. In contrast, the crude cell-free supernatant produced almost complete inhibition of fungal growth, reaching 98.3 ± 3.0% inhibition of B. cinerea at 24 h, and maintained high activity after 48 h. These results indicate that fractionation disrupts the natural metabolic balance established during fermentation and suggest that maximal antifungal efficacy results from the cooperative action of multiple metabolite classes rather than from the activity of individual compounds. Such combined effects are expected and physiologically justified because fengycins directly disrupt fungal membranes, surfactins facilitate membrane permeabilization and enhance the activity of other lipopeptides, and subtilosins and polyketides contribute additional antimicrobial mechanisms, together generating a broader and more durable inhibitory effect [71,72]. The practical antifungal efficacy was further evaluated in a detached-leaf bioassay. The whole fermentation broth—comprising both secreted bioactive metabolites and live bacterial cells—provided strong bioprotection on tomato foliage against phytopathogenic fungi and oomycetes. While this ex vivo setup demonstrates the protective potential of strain 5RB, it represents an initial evaluation. The presence of live cells in the fermentation broth likely contributes to the observed efficacy through tissue colonization and continued in situ metabolite production, aspects that warrant further validation on intact plants under greenhouse and field conditions.

The promising results in this study outline the main focus of our subsequent research to establish a fully predictive framework for biocontrol based on the integration of multivariate metabolomics (PCA/PLS-DA) and quantitative synergistic modeling. Combining genomics, metabolomics and co-cultivation strategies will be the next step in our studies.

5. Conclusions

This study demonstrates that B. velezensis 5RB produces a notably diverse repertoire of antifungal metabolites, the composition of which can be selectively modulated by the cultivation medium. Comprehensive UHPLC-QqTOF-MS/MS analysis identified numerous surfactin and fengycin homologs, several polyketides, and, for the first time in B. velezensis, the sactipeptides subtilosin A/A1. Furthermore, a previously undescribed C14 surfactin B* isoform and several linear fengycin homologs were characterized, expanding the known chemical diversity of antimicrobial metabolites produced by this species. Overall, our findings demonstrate that the secondary metabolism of B. velezensis 5RB can be strategically directed by modifying the culture medium to enrich specific bioactive compounds. Advanced genome mining effectively mapped the non-ribosomal and ribosomal biosynthetic pathways (NRPS, PKS, and RiPP) responsible for this diverse chemical profile.

The remarkable consistency among our metabolic profiling, genetic predictions, and preliminary ex vivo bioprotection assays confirms that the 5RB isolate represents a promising candidate for the development of multi-target biofungicides. However, further studies on intact plants under greenhouse and field environments are required to fully evaluate its biocontrol efficacy, persistence, and practical application in agricultural settings.

6. Patents

Patent Application No. BG/P/2026/114370/21 May 2026, Patent Office of the Republic of Bulgaria; Title: “Biologically active composition with antifungal activity, obtained from a culture of the bacterial strain Bacillus velezensis 5RB, containing lipopeptides, sactipeptides and/or polyketides”.