A soil bacterium that protects plants against fungal disease does not respond the same way to a living pathogen as it does to the mere wreckage of one. A new study of Bacillus siamensis strain SSVP1 shows that the bacterium’s extracellular chemical output shifts markedly depending on whether it meets live Fusarium oxysporum or Macrophomina phaseolina, two major soil-borne plant pathogens, or simply encounters their heat-killed, processed biomass. The work, published in Discover Molecules, used untargeted metabolomics to map these condition-dependent differences and flags a set of candidate metabolites for future validation as biological control agents.

The research team, Vikram Poria and Surender Singh, grew SSVP1 alone and under four interaction conditions: co-culture with live Fusarium, co-culture with live Macrophomina, exposure to processed heat-killed Fusarium biomass, and exposure to processed heat-killed Macrophomina biomass. Fungal mycelia destined for the non-viable treatments were dried, ground under liquid nitrogen and autoclaved, a procedure intended to eliminate fungal metabolic activity while retaining the processed biomass as a chemical stimulus. Culture supernatants were collected after 72 hours of incubation and analysed by ultra-high-performance liquid chromatography coupled to tandem mass spectrometry, with all metabolite annotations treated as putative because the workflow was untargeted.

The raw datasets revealed a striking asymmetry. An average of 1,242 metabolite features were detected in bacterial monoculture, rising to 1,693 with live Fusarium and 1,719 with live Macrophomina, but jumping to 2,593 and 2,625 with the dead Fusarium and dead Macrophomina conditions respectively. After background filtering, 3,327 features were retained for downstream analysis. The authors suggest that non-viable fungal biomass may provide accessible chemical cues, substrates or biomass fragments associated with a broader extracellular profile, whereas live fungal interactions may constrain the bacterial metabolite output, potentially through competition or exposure to inhibitory compounds.

Differential abundance analysis, comparing each interaction condition against monoculture with thresholds of a log2 fold change of at least 1 and a Benjamini-Hochberg adjusted p-value of at most 0.05, reinforced the pattern. Under the live Fusarium condition, 251 features showed significantly lower relative abundance and 1,313 showed significantly higher abundance than in monoculture. Under dead Fusarium, those figures rose to 288 and 1,509. The dead Fusarium condition carried 14.74 percent more features with significantly negative fold changes and 14.92 percent more with significantly positive fold changes than the live Fusarium condition. A similar, though milder, trend appeared for Macrophomina: dead biomass yielded 1,507 significantly positive and 116 significantly negative features, against 1,376 and 121 for live biomass.

Fungal species identity mattered as much as viability. Direct comparison showed that the live Fusarium condition contained 107.43 percent more metabolite features with significantly negative log2 fold changes than the live Macrophomina condition, and the dead Fusarium condition contained 148.27 percent more such features than its dead Macrophomina counterpart, while the counts of significantly positive features remained broadly comparable. In other words, encountering Fusarium was associated with a far larger pool of metabolites whose abundance dropped relative to monoculture, a signature the authors link to the known capacity of fungal pathogens to influence bacterial metabolism through toxins, bioactive metabolites and nutrient competition.

Venn analysis separated conserved responses from condition-specific ones. A core set of 787 metabolite features, 48.33 percent of the upregulated pool, showed significantly positive fold changes across all four interaction conditions, suggesting a general bacterial response to fungal presence regardless of species or viability. Yet each condition also carried its own chemical fingerprint: dead Macrophomina produced the largest set of uniquely upregulated features at 99, followed by live Macrophomina with 89, dead Fusarium with 69 and live Fusarium with 25. Among downregulated features, dead Fusarium again led with 57 unique entries, and no exclusive overlap appeared between the two Macrophomina conditions, underscoring that both the fungal species and the interaction state shape the metabolite profile.

After stringent filtering of the 2,329 differentially abundant features, removing unnamed compounds, features without MS/MS spectra, redundant ion adducts, high mass-error entries and compounds of synthetic, human or plant origin, a curated set of 233 metabolites remained. Chemically, the pool was dominated by peptide and non-ribosomal peptide synthetase-derived compounds at 56.95 percent and polyketides at 25.17 percent, the two classes most often behind the bioactive arsenal of Bacillus species. Because source annotations came from database and literature matching rather than direct demonstration, the authors treated them strictly as references for structurally similar compounds, not as evidence of which organism actually produced each metabolite.

Within the 151 putatively bacterial metabolites, functional classes behaved differently across conditions. Antifungal-associated metabolites, including members of the iturin, surfactin and fengycin lipopeptide families and the peptide antibiotic bacilysin, showed their highest mean log2 fold change under live Macrophomina at 1.72, followed by dead Macrophomina at 1.17 and live Fusarium at 0.75, with dead Fusarium lowest at 0.20. Antibacterial-associated metabolites, by contrast, stayed comparatively stable across all four conditions. Siderophore-associated metabolites, notably bacillibactin and its derivatives, maintained strongly positive fold changes everywhere, and signalling-associated metabolites, including N-acyl homoserine lactones and solonamides, remained positive throughout. At the level of individual compounds, the picture fragmented further: bacilysin showed fold changes of 9.77 to 10.33 across conditions, while fengycin C and several heinamide derivatives were consistently negative, and iturin A7 and A8 turned positive only under live Macrophomina. Individual members of the same biosynthetic family thus responded in opposite directions, which is why the authors argue that class-level summaries alone can miss the real story.

Metabolite-to-pathway mapping added a physiological layer. Bacillibactin biosynthesis, a siderophore pathway tied to iron acquisition, carried the highest positive fold changes of any mapped pathway, ranging from 7.13 to 9.32 across conditions. Pathways for L-cysteine biosynthesis, the S-adenosyl-L-methionine cycle, demethylmenaquinol-6 biosynthesis and heme biosynthesis all showed positive fold changes, pointing to heightened sulfur metabolism, methylation, electron transport and energy metabolism. Cell envelope pathways diverged by condition: peptidoglycan biosynthesis metabolites were negative under dead Fusarium, live Fusarium and live Macrophomina but positive under dead Macrophomina, while adenosylcobalamin biosynthesis was positive only under the two dead-biomass conditions, cleanly separating live from non-viable interactions.

Crucially, despite all this compositional variation, supernatants from every interaction condition still inhibited both fungi in agar well diffusion assays, with no inhibition in the negative control. The authors are careful about what their data do and do not show: this is a discovery-based, untargeted study describing relative, condition-associated abundance patterns rather than absolute metabolite production or pathway activity, and the extraction method selectively enriched acid-precipitable compounds such as lipopeptides and polyketides rather than the global metabolome. They call for bacterial biomass normalisation, isotope tracing, transcriptomics, targeted metabolomics and biochemical validation to pin down mechanisms. The practical message stands, however: evaluating a biological control agent only in monoculture may misrepresent its behaviour in the crowded, chemically negotiated environment of the rhizosphere, where a pathogen’s corpse can elicit a very different response than its living self.

Subject of Research: Condition-dependent variation in extracellular bioactive metabolites of Bacillus siamensis during interaction with live and heat-killed fungal pathogens

Article Title: Condition dependent variation in bioactive metabolites of Bacillus siamensis during interaction with fungal pathogens

Article References: Poria, V., & Singh, S. (2026). Condition dependent variation in bioactive metabolites of Bacillus siamensis during interaction with fungal pathogens. Discover Molecules, 3(1), Article 30. https://doi.org/10.1007/s44345-026-00073-9

Image Credits: AI Generated

DOI: 10.1007/s44345-026-00073-9

Keywords: Bacillus siamensis, metabolomics, UHPLC-MS/MS, Fusarium oxysporum, Macrophomina phaseolina, lipopeptides, siderophores, biological control, soil-borne pathogens, non-ribosomal peptides, polyketides, rhizosphere