A bacterium newly named as a species in the Bacillus cereus group has shown a striking split personality in the laboratory: it slowed the growth of every one of eight fungal crop pathogens placed in direct contact with it, yet it failed to kill any of four major insect pests fed its spores and crystals. The organism, Bacillus silvicola strain IBL-03679, was isolated from a mature hardwood forest soil sample in Maryland, USA, and was recently classified as a novel species. Because the strain forms a bipyramidal parasporal crystal in pure culture, a feature long associated with insecticidal activity in Bacillus thuringiensis, researchers suspected it might harbor untapped biocontrol potential. A new study in Frontiers in Horticulture reports the first systematic look at what this bacterium can and cannot do against agricultural pests and pathogens.
The research team, led by Holly P. Bartholomew of the USDA-ARS Invasive Insect Biocontrol and Behavior Laboratory in Beltsville, Maryland, combined genome mining with a battery of laboratory bioassays. They compared B. silvicola IBL-03679 against a well-studied commercial workhorse, Bacillus thuringiensis subsp. kurstaki strain IBL-01447, also known as HD-1. Using the antiSMASH program, the team identified biosynthetic gene clusters, or BGCs, in both genomes. B. silvicola carried 15 such clusters in total, eleven of which had no similarity match to known compounds in the database, including categories such as ribosomally synthesized and post-translationally modified peptides, deazapurine, terpene, a ranthipeptide, and two class II lanthipeptides.
Three clusters stood out as shared between the two bacteria. Both species contained high-similarity matches to the siderophores bacillibactin and petrobactin, iron-scavenging molecules with documented antifungal and stress-related roles, and a medium-similarity match to the lasso peptide paeninodin. The Bt strain, however, held two cards that B. silvicola lacked: a high-similarity cluster for the antibiotic zwittermicin A, an antifungal compound purified from B. cereus with activity against numerous fungal phytopathogens, and a medium-similarity cluster for the bacteriocin thurincin H, an antibacterial agent. The authors caution that linking the observed antifungal activity to any specific cluster remains hypothetical, and that identifying the actual products made during bacterial-fungal encounters is a priority for future mechanistic work.
On the insecticidal side, a search of the B. silvicola genome against the Bacterial Pesticidal Protein Resource Center database, using an E-value threshold of 1E-15, turned up 23 putative matches to genes encoding anti-insect products. Fifteen of these shared homology with Cry proteins, specifically from the Cry32, Cry41, and Cry73 families. None of the database matches were identical, with percent identity scores ranging from 29.1 to 81.6 percent, suggesting some could be novel cry genes or homologs with unknown functions. The genome also yielded one annotated chitin-associated gene, chbG, which encodes a chitooligosaccharide deacetylase known from E. coli work but without demonstrated direct chitinase-related insecticidal or antifungal activity, alongside 30 protease genes.
To test whether any of this genomic potential translated into real toxicity, the team fed spore and crystal suspensions of B. silvicola to second-instar larvae from three insect orders: the European spongy moth Lymantria dispar dispar and the corn earworm Helicoverpa zea, both Lepidoptera; the Colorado potato beetle Leptinotarsa decemlineata, a Coleopteran; and the yellow fever mosquito Aedes aegypti, a Dipteran. Each assay was repeated twice, with 32 larvae per trial for the spongy moth and potato beetle, 30 for the mosquito, and 16 for the corn earworm, incubated at 28 degrees Celsius and scored for mortality after 72 hours, or 96 hours for mosquitoes. The result was unambiguous. B. silvicola and the negative control Bt subsp. finitimus produced limited mortality, if any, with no observed sublethal effects compared to the water control. The positive control Bt strains, by contrast, achieved greater than 93 percent mortality in their respective target insects.
The antifungal story unfolded differently. In direct plating assays on trypticase soy agar, both B. silvicola and Bt subsp. kurstaki significantly reduced the growth of all eight fungal phytopathogens tested compared to a water control: four apple pathogens, Alternaria alternata, Botrytis cinerea, Colletotrichum fioriniae, and Penicillium expansum, and four coffee pathogens, Colletotrichum pseudoboninensis, Paramyrothecium roridum, Cercospora coffeicola, and Fusarium oxysporum. Hygromycin treatment outperformed both bacteria. The Bt strain showed slightly increased efficacy against F. oxysporum and C. pseudoboninensis relative to B. silvicola, while all other fungi were equally affected by the two strains. B. cinerea showed the greatest mean reduction in colony size, 68.6 percent with B. silvicola and 67.1 percent with Bt, whereas F. oxysporum colony size fell by only 23.0 percent with B. silvicola compared to 40.5 percent with Bt.
When the researchers separated the organisms to test whether volatile organic compounds alone could do the work, the effect largely vanished. In bi-plate assays, B. silvicola caused only mild reduction in growth of P. expansum and A. alternata, while Bt showed no significant impacts except that C. pseudoboninensis colonies were slightly larger in its presence. Filtered cell-free supernatants from seven-day bacterial cultures likewise produced no significant growth difference compared to uninoculated broth controls. The authors suggest several explanations: the antifungal effect may require direct physical contact, or fungal signals that induce production of inhibitory compounds to effective concentrations, or nutrient competition and local chemical changes in the shared space between colonies. Altered growth conditions, such as liquid broth versus plate growth, could also reduce expression of antifungal compounds and peptides.
The team then moved to living plant tissue, wounding Opal apples and inoculating them first with Bacillus suspensions and, one hour later, with the four apple pathogens. After seven days of storage at room temperature, none of the fungi showed significantly reduced lesion diameters compared to water controls. Detached coffee leaves pretreated with the bacteria and then challenged with C. coffeicola or P. roridum likewise showed no significant difference in lesion diameters after 14 days. The authors attribute this failure to timing: with only one hour between bacterial and fungal inoculation into the same wound, the bacteria had insufficient time to colonize, form a biofilm, or induce plant resistance before the fungus arrived. Suboptimal conditions in the apple fruit, such as low pH, may also have limited bacterial growth and antifungal compound production.
The study also flags safety considerations. B. silvicola IBL-03679 belongs to the Bacillus cereus senso lato group and carries annotated nonhemolytic and hemolytic enterotoxin gene homologues associated with B. cereus related illness. The authors note, however, that Bt strains are phylogenetically even closer to B. cereus and that many serovars also contain these genes; while many Bt strains have been shown safe for human consumption, B. silvicola must still be tested for off-target effects before any utilization. The strain is also psychrotolerant, meaning it tolerates cold, which the authors suggest could allow it to persist in the low temperatures of postharvest storage, and its ability to form spores points toward shelf-stable formulations similar to known Bt products.
Why did the crystals fail to kill insects? The authors point to the specific Cry families found in the genome. Cry32 proteins have shown weak, species-specific activity or no activity against lepidopterans, and a range from no activity to high efficacy against dipterans. Cry41 has no demonstrated activity, while Cry73 family proteins have shown hemipteran and nematode toxicity. The remaining homologs belonged to families associated with other target groups, including coleopteran, hemipteran, dipteran, blattodean, and neuropteran activity. The tested insects may simply fall outside the toxin’s spectrum, or activity might emerge against untested orders, nematodes, or narrow species-specific targets. Characterizing the purified crystal by SDS-PAGE, and testing purified or pre-solubilized toxin, especially against insects with more acidic midguts such as the Colorado potato beetle, could sharpen the picture. As with any biocontrol candidate, the authors stress that non-target effects on beneficial insects must be investigated before commercial use.
Subject of Research: Antagonistic interactions of the novel bacterium Bacillus silvicola against insect pests and fungal phytopathogens
Article Title: Investigating interkingdom interactions of Bacillus silvicola against insect pests and agricultural phytopathogens
Article References: Bartholomew, H. P., Lovelace, A. H., Jurick, W. M., II, Sparks, M. E., Gundersen-Rindal, D. E., & Blackburn, M. B. (2026). Investigating interkingdom interactions of Bacillus silvicola against insect pests and agricultural phytopathogens. Frontiers in Horticulture, 5, Article 1966024. https://doi.org/10.3389/fhort.2026.1966024
Image Credits: AI Generated
DOI: 10.3389/fhort.2026.1966024
Keywords: Bacillus silvicola, Bacillus thuringiensis, biocontrol, phytopathogens, Cry toxins, parasporal crystals, biosynthetic gene clusters, spongy moth, apple pathogens, coffee leaf pathogens, antifungal activity, insect bioassays
News Source: Gavin Prescott. (October 28, 2026). Research report: Investigating interkingdom interactions of Bacillus silvicola against insect pests and agricultural phytopathogens. Scienmag.
