https://orcid.org/0000-0002-2358-9642
Figures
Abstract
Bacillus anthracis and the emerging pathogen Bacillus cereus biovar anthracis (Bcbva) are causative agents of the lethal disease anthrax. Their ability to form highly resistant endospores enables them to persist in the environment, posing significant threats to wild herbivores, livestock, and also humans globally, especially considering their potential use as bioweapons. Despite the importance of these pathogens, the exact mechanisms underlying host infection remain poorly understood. Notably, both require relatively high infectious doses to cause disease. Proposed transmission routes include soil-based transmission and dissemination by carrion flies, which contaminate the area surrounding infected carcasses. However, considering the substantial dilution of bacteria and spores in these processes alongside the high infectious dose required, sustained transmission under natural conditions appears improbable. An alternative hypothesis is that B. anthracis and Bcbva can survive and proliferate in the environment by forming biofilms, structured bacterial communities attached to surfaces, which may serve as reservoirs for infection. While biofilm formation has been demonstrated for B. anthracis, data on Bcbva are lacking. To address this, we examined the biofilm-forming abilities of multiple B. anthracis and Bcbva strains using a three-dimensional, soil-inspired porous glass bead model (PGB) system. We applied confocal laser scanning microscopy, scanning electron microscopy, and quantitative cell counts to characterize biofilm development in detail. Our results confirm that both B. anthracis and Bcbva form biofilms under somewhat soil-like conditions. Furthermore, differences observed in biofilm structure and density between the above-mentioned species suggest that Bcbva may prefer additional environmental niches beyond soil, such as plant surfaces or small water bodies, potentially expanding our understanding of its environmental persistence and infection routes.
Author summary
Anthrax is a deadly disease caused by bacteria that can form highly resistant spores, allowing them to survive in the environment for long periods. Two closely related bacteria, Bacillus anthracis and Bacillus cereus biovar anthracis (Bcbva), cause anthrax but differ in their natural habitats and likely in their modes of infecting animals and humans. Unfortunately, B. anthracis can also be misused as a biological weapon, giving it a grim reputation. While B. anthracis typically infects herbivores in grassland areas, Bcbva is mainly found in tropical rainforest environments, where its transmission and infection routes remain unclear. Our research investigates the possibility that both bacterial species can survive outside their hosts by forming biofilms, communities of bacteria that adhere to surfaces and protect themselves, thereby increasing the chance of infection. Using a soil-inspired laboratory model, we studied biofilm formation by these bacteria. Our findings confirm that both B. anthracis and Bcbva are capable of biofilm formation and suggest that Bcbva may also inhabit environments beyond soil, such as plant surfaces or small water bodies. Understanding these survival strategies is important for improving predictions of anthrax spread in nature and could aid in developing better methods for disease control and prevention, especially considering the potential misuse as a bioweapon.
Citation: Borst L, Howaldt S, Berdaguer R, Schaudinn C, Dupke S, Stämmler M, et al. (2026) Anthrax-causing bacteria form biofilms on a soil-inspired porous glass bead model. PLoS Negl Trop Dis 20(3): e0014043. https://doi.org/10.1371/journal.pntd.0014043
Editor: Joseph M. Vinetz, Yale University School of Medicine, UNITED STATES OF AMERICA
Received: September 26, 2025; Accepted: February 16, 2026; Published: March 3, 2026
Copyright: © 2026 Borst et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Anthrax is a lethal zoonotic disease caused by Bacillus anthracis, a spore-forming, Gram-positive bacterium with global distribution. Although human infections are rare (around 2000 cases per year according to WHO [1]), anthrax poses a persistent threat to livestock and wild herbivores, particularly in arid and semi-arid grassland ecosystems. The pathogen has also attracted considerable attention due to its potential use as a biological weapon, as demonstrated by the anthrax letter attacks in the United States in 2001 [2], and its relevance in biodefense and biosecurity contexts [3]. However, unintentional releases of the pathogen, as observed in cases of contaminated heroin [4], animal hides [5] or the accident in the Sverdlovsk military microbiology facility [6] can also pose severe public health threats.
A closely related pathogen, Bacillus cereus biovar anthracis (Bcbva), also belonging to the B. cereus group, is endemic in tropical rainforests of sub-Saharan Africa [7,8], and poses a significant threat to susceptible wildlife populations such as duikers, monkeys and even elephants [9,10] by causing an anthrax-like disease. The pathogenicity of both B. anthracis and Bcbva is harbored on two large virulence plasmids, pXO1/pBCXO1 and pXO2/pBCXO2, which are nearly identical in the two species. However, their chromosomal backgrounds diverge substantially. Bcbva exhibits greater genomic similarity to B. cereus, which is reflected phenotypically by traits such as motility, a characteristic absent in B. anthracis [11]. These genomic and phenotypic differences likely contribute to the occupation of distinct ecological niches and may influence both infection pathways and environmental persistence.
Current evidences suggest that both B. anthracis and Bcbva proliferate almost exclusively within a host and persist in the environment solely as spores until further infection. B. anthracis is typically associated with soil-based transmission in arid environments, where spores persist in the soil and infect hosts through feeding or inhalation at contaminated sites [12]. In contrast, an alternative transmission route, especially relevant for Bcbva commonly inhabiting densely vegetated rainforest ecosystems and also infecting strictly arboreal monkey species [10,13], involves necrophagous flies. These flies can disseminate bacteria from infected carcasses into the surrounding environment, potentially contaminating areas above the ground [10,14,15]. However, these modes of transmission both entail substantial dilution effects, thereby limiting the probability of subsequent infections [15,16]. Consequently, environmental proliferation and persistence mechanisms outside the host may be critical for maintaining transmission cycles. In this context, biofilm formation on abiotic surfaces such as soil particles, sand, and within small water bodies, or on biotic surfaces like roots, leaves and fruits, may represent an important proliferation and survival strategy. By supporting bacterial proliferation in the environment, biofilms could substantially increase the infectious dose and therefore the likelihood of subsequent infection.
Biofilms are generally defined as structured communities of bacteria that often adhere to surfaces and are frequently embedded within a self-produced matrix of extracellular polymeric substances [17]. For bacteria, biofilms can provide protection against a variety of environmental stresses, including desiccation and antimicrobial agents as well as enhance intercellular communication and cooperation [18]. While biofilm formation has been extensively studied for B. cereus group members like B. thuringiensis and B. cereus sensu stricto [19], other members of this group, especially the highly pathogenic B. anthracis and Bcbva have been largely neglected. Despite their biosecurity importance and their beforementioned potential ecological relevance, data on biofilm formation for B. anthracis remain limited [20,21], particularly with regard to spatial distribution and structural organization in three-dimensional environments and are still lacking for Bcbva.
To address these gaps, the present study investigates biofilm formation by both species using porous glass beads (PGB) as a model system [22]. The aim was to elucidate their biofilm-forming capabilities and to identify potential species-specific patterns in biofilm architecture and attachment behavior. To this end, ten distinct strains of B. anthracis and Bcbva were cultivated on PGBs over a period of seven days. Eight of these strains required handling in special biosafety level 3 laboratories (BSL-3). Biofilm development was assessed at multiple time points using quantitative cell counts and microscopic techniques.
Materials and methods
Bacterial strains and media
In total, ten different strains were used in this study (Table 1), which comprised two attenuated risk group 2 strains (B. anthracis Sterne and Bcbva CAR-H) for protocol establishment and eight virulent risk group 3 strains (B.a. Vollum, B.a. 14RA5915 (hereafter called Dobichau), B.a. 19/39, B.a. VAR28A88, Bcbva CI, Bcbva CA, Bcbva DRC, Bcbva RCA). B. anthracis Sterne and Bcbva CAR-H are both lacking the capsule plasmid pXO2/pBCXO2, and Bcbva CAR-H also lost the ability to produce the hyaluronic acid capsule due to a deletion of the hasA gene. Risk group 3 strains were handled in a BSL-3 laboratory.
All strains were cultivated in Brain-Heart Infusion (BHI) broth (BBL, #211059, Becton Dickinson, Fisher Scientific, Schwerte, Germany) at 37 °C overnight with constant agitation. Afterwards, 25% (v/v) glycerol was added, aliquots of 500 µl were prepared as stocks, stored at -70 °C and used for all subsequent experiments.
Porous glass bead model
A modified bead assay originally established with Pseudomonas aeruginosa [22] was used for the PGB model. Biofilm cultivation was performed using PGBs (ROBU Glasfilter-Geraete GmbH, Hattert, Germany, Ø 4 mm, pore size: 60 µm) in 24-well plates. A single sterile PGB was placed upright in each well. Bacterial suspensions were prepared from the previously stored stocks with 6.7 × 106 colony-forming units (CFU)/ml in BHI medium (BBL, #211059, Becton Dickinson, Fisher Scientific, Schwerte, Germany), and each PGB was inoculated by dropping 15 µl (105 cells) of the respective strain on top. Following inoculation, the 24-well plate with the PGBs was carefully transferred into an airtight box containing a moistened towel to maintain high humidity (~100%). The humidity chamber was incubated at 30 °C for up to 7 days without shaking. At four time points (0 h, 24 h, 72 h, and 168 h), CFU and genome equivalents (GE) were quantified, and microscopy was performed.
Colony forming units and genome equivalents
To quantify bacterial load on PGBs, both CFU and GE were determined from biofilm samples. For each time point, including the starting point (0 h) immediately after inoculation, PGBs were washed by dipping them once into 1 ml of phosphate-buffered saline (PBS) to remove any unattached bacteria. PGBs were then transferred into cryo-tubes containing 250 µl PBS and each cryo-tube was additionally placed into a 15 ml tube (serving as a secondary container in compliance with BSL-3 safety regulations) filled with 3 ml of tap water. To ensure submersion of the cryo-tube inside during treatment, an additional empty cryo-tube was placed on top as a weight.
Samples were subjected to ultrasonication for 10 minutes in a SONOREX Super RK 514 bath (Bandelin, Berlin, Germany; nominal power: 215 W, 35 kHz) to detach biofilm-associated bacteria. Successful and complete detachment was verified beforehand by confocal laser scanning microscopy (CLSM) of treated PGBs. The cryo-tube containing the PGB was then vortexed for eight seconds, and 100 µl of the supernatant was transferred into two sterile 1.5 ml screw-cap tubes. One tube was heat-treated at 65 °C for 30 minutes to eliminate vegetative cells. The second tube was mixed with 100 µl of BHI medium and incubated at 37 °C for 45 minutes to induce germination of spores, improving the efficiency of subsequent DNA extraction. After incubation, 100 µl were again transferred into two fresh screw-cap tubes. One of these was stored at –70 °C until DNA extraction.
For CFU enumeration, both heat-treated (65 °C) and germinated (37 °C) samples were serially diluted in PBS using a 96-well plate, with each dilution performed in duplicate. For each dilution, 5 µl were spot-plated [22] onto tryptone soya agar (TSA, Thermo Scientific Oxoid, Wesel, Germany) and incubated overnight at 37 °C. Colonies were counted manually to calculate CFU.
For GE determination, the frozen samples were thawed and DNA was extracted using the Qiagen DNeasy Blood & Tissue Kit (Qiagen GmbH, Hilden, Germany), following the manufacturer’s protocol for Gram-positive bacteria. DNA was eluted in 100 µl of EB buffer (Qiagen GmbH, Hilden, Germany). Quantitative PCR (qPCR) targeting the rpoB gene of the B. cereus group was performed using 5 µl of the extracted DNA in a total reaction of 25 µl, with 6,25 µl TaqMan Environmental Master Mix (Applied Biosystems, Darmstadt, Germany), 300 nM each of forward and reverse primers (5′ AAC CGC CTG ACG TTG AAA 3′ and 5′ CGG CAG CGA CAG CTT GTA 3′) and 100 nM of a FAM-TAMRA-labeled probe (5′ FAM-CTA ACC GCG CAC TTA TGG GAG CGA AC-TAMRA 3′, Primers and Probe were purchased from Metabion, Planegg, Germany). Each replicate was measured in duplicates with cycling conditions as described earlier [28]. Standard curves were generated using genomic DNA from Bcbva CI and B. anthracis Vollum. DNA concentrations were measured using a Qubit 4 fluorometer (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. Genome copy numbers were calculated based on DNA concentration and genome size.
Confocal laser scanning microscopy
Bacteria were cultivated on PGBs as described above. At each time point (24 h, 72 h, and 168 h) PGBs were carefully picked using sterile forceps and washed by dipping them in PBS to remove any residual media and unattached bacteria. The samples were then fixed by placing the PGBs in 1 ml of 4% paraformaldehyde (PFA) for at least 24 hours.
For staining, a staining solution was prepared by mixing water and dimethyl sulfoxide in a 1:1 (v:v) ratio, and adding 4′,6-Diamidin-2-phenylindo (DAPI, 40 µg/ml, Fisher Scientific, Schwerte, Germany). Following fixation, the PGBs were removed from the PFA and stained by placing 15 µl of the staining solution on top of the sample. Afterwards, the samples were incubated for 30 minutes at room temperature in the dark.
After staining, the samples were transferred to a glass-bottom chamber (Ibidi GmbH, Graefelfing, Germany) onto a 3 µl droplet of water, positioning the PGBs with the flat side down.
Microscopy was performed using a confocal laser scanning microscope (LSM780, Carl Zeiss Microscopy GmbH, Jena, Germany). Imaging parameters (resolution, contrast and brightness) were optimized for all samples. Image acquisition was carried out using the Zeiss ZEN 2.3 SP1 FP3 black edition software.
Scanning electron microscopy
For scanning electron microscopy (SEM), the same samples prepared for CLSM were further processed. Samples were additionally fixed in a mixture of 2.5% glutaraldehyde and 1% paraformaldehyde in 50 mM HEPES for 24 hours at room temperature. After fixation, dehydration was carried out using a graded ethanol series with six different concentrations (30%, 50%, 75%, 90%, 95%, 100% and again 100%), with each concentration applied for a minimum of 20 minutes at room temperature. Subsequently, samples were incubated in hexamethyldisilazane (HMDS) until the HMDS had fully evaporated. Dried samples were then mounted on aluminum stubs and sputter-coated with a 10 nm gold-palladium alloy. SEM imaging was performed using a 1530 Gemini microscope (Carl Zeiss GmbH, Germany) at 3 kV.
Fourier transform Raman spectroscopy
Fourier transform (FT)-Raman measurements were performed using a Bruker MultiRAM NIR FT-Raman spectrometer (Bruker Optics, Ettlingen, Germany). Raman measurements were conducted using Bruker’s OPUS 7.0 data acquisition software with the integrated OpusLab extension for automated measurements. The OPUS software was also used for pre-processing the spectra, such as averaging and baseline correction.
NIR FT-Raman spectra were recorded in the spectral range between 100 and 4000 cm-1. A total of 250 individual scans were averaged per single spectrum. The laser power was set to 350 mW, and the nominal spectral resolution was set to 4 cm-1 with a zero-filling factor of 4, giving a final point spacing of approximately 1 cm-1. To prepare the samples, circa 50 PGBs were filled in a 50 ml tube and vortexed for 10 minutes. The detached glass sinter particles were used for analysis, while the sand particles stemmed from the Sahara Desert (collected before 2000 in the Kingdom of Morocco).
Statistical analysis and software
Statistical analysis was done with R software (Edition 4.3.0 and package “car”) [29,30]. All quantitative results are presented as means ± standard deviations, unless otherwise specified.
Due to limited sample sizes, formal tests for normality and homogeneity of variances (e.g., Shapiro-Wilk normality test, Levene’s test and visual inspection of Q-Q plots) may have low statistical power. Consequently, non-parametric Wilcoxon rank-sum tests were used to compare groups. To control for the increased risk of Type I errors from multiple pairwise comparisons, p-values were adjusted using the Bonferroni correction. Data visualization was performed with Graph Pad Prism (version 9.1.0, GraphPad Software, San Diego, California USA, www.graphpad.com).
Language assistance and literature guidance were supported using the large language model ChatGPT-4 (OpenAI). No content was generated autonomously without author review. All AI-generated suggestions were critically assessed and edited by the authors, who take full responsibility for the final manuscript.
Results
Biofilm model
The biofilm model for this study needed to fulfill several criteria. It should be highly reproducible, capable of producing large amounts of biofilm, it should be easy to handle and ideally resemble a soil-like structure as this is the typical reservoir for B. anthracis. For all these reasons, a model using porous glass beads (PGB, Fig 1a-c) was chosen since it met all of our criteria, particularly, as the PGBs are standardized according to ISO 4793-80, Por. 2 promising high reproducibility.
a | Macro image of a PGB showing side (left) and top views (right). b | SEM image of the flat bottom area of a PGB. c | SEM image of a PGB. d | Macro image of the Sahara sand used for FT-Raman. e | SEM image of Sahara sand (arrow) alongside a PGB particle. f | NIR FT-Raman spectra of PGB (1) and Sahara sand (3) compared to a reference fused silica (amorphous quartz glass, 2) and rose quartz (crystalline quartz, 4). The Raman intensities of all spectra shown were min-max normalized and vertically shifted for better visualization.
Furthermore, PGBs resemble a sandy, soil-like structure comparable to Sahara sand (Fig 1d-e), a similarity confirmed by FT-Raman spectroscopy (Fig 1f). Both PGBs and Sahara sand exhibited Raman SiO2 features characteristic for either amorphous quartz glass (fused silica), or crystalline quartz structures (Fig 1f). While amorphous quartz glass and crystalline quartz are chemically nearly identical (SiO2), the primary distinction between the samples lies in their structure: Raman spectra of PGBs closely resemble amorphous quartz glass spectra whereas the Sahara sand spectrum is almost identical to the Raman spectrum of crystalline rose quartz (Fig 1f, rose quartz contains aside from SiO2 as the major component trace amounts of titanium, iron and manganese).
To better depict a natural arid soil environment and to prevent an exaggerated growth of planktonic bacteria, the existing PGB model was modified by using only a small inoculum volume.
To include a broad range of B. anthracis and Bcbva for this study, five different strains of each species, one attenuated strain for protocol establishment and four virulent wild-type strains, were used. Biofilm formation was confirmed using CLSM and SEM as well as by counting cells of the biofilm via CFU and qPCR.
Biofilm formation observed by confocal laser scanning microscopy
To assess the biofilm-forming capability of the tested strains, CLSM was performed at different timepoints post-inoculation (24 h, 72 h, 168 h) with DAPI stained cells. Due to the inherent limitations of CLSM in imaging deeper layers, analysis was restricted to the flat bottom surface of the PGBs, potentially missing bacterial cells located within the inner part or at the lateral surfaces of the beads.
CLSM analysis demonstrated that all tested strains were capable of forming biofilms on the PGBs after 24 hours of incubation. Prominent bacterial aggregates were primarily observed along the outer rim of the PGBs, whereas only few or no bacterial cells were detected at the center region or inside the pores (Fig 2).
Images show the entire flat-bottom area of the PGB measuring 4 mm in diameter. Bacteria are stained with DAPI (cyan). All strains exhibit strong biofilm formation with bacterial aggregates predominantly localized at the rim of the PGBs.
After 72 and 168 hours, only aggregates of B. anthracis strains remained primarily at the outer rim of the PGBs (Fig 3a-e and Fig 4a-e).
Images show the entire flat-bottom area of the PGB measuring 4 mm in diameter. Bacteria are stained with DAPI (cyan). At this time point, differences in biofilm formation between B. anthracis (a–e) and Bcbva strains (f–j) were evident. Bcbva strains displayed an overall reduced signal with a patchy distribution.
Images show the entire flat-bottom area of the PGB measuring 4 mm in diameter. Bacteria are stained with DAPI (cyan). For B. anthracis strains (a-e), bacteria remained concentrated at the rim of the PGBs, though signal intensity decreased compared to earlier time points. This further decrease in signal was also observed for the Bcbva strains, which maintained their patchy distribution between the PGB particles.
In contrast, Bcbva strains were predominantly located inside the pores of the PGBs at both time points (Fig 3f-j and Fig 4f-j). One exception was the attenuated strain Bcbva CARH, which appeared to remain on the outer surface of the PGB even after 72 and 168 hours (Figs 3j and 4j).
Since Bcbva cells were mostly located inside the pores of the PGBs, the overall biofilm mass on the PGB appeared reduced in the optical section of the CLSM images compared to B. anthracis after 72 and 168 hours.
Quantitative analysis of biofilm formation
In order to quantify the biofilm formation observed via microscopy on the PGBs, CFU counts and qPCR were performed. Additionally, spore counts were determined by CFU enumeration of heat-treatment samples.
For each time point, three independent biological replicates (n = 3) were analyzed, each comprising two additional replicates. Furthermore, each sample was measured in duplicates in both CFU and qPCR. Since all samples were washed in PBS prior to CFU enumeration and qPCR, the cell count values represent only the cells attached to PGBs and should thus correspond to the microscopy results.
CFU enumeration revealed that all strains exhibited a significant increase in cell numbers during the first 24 hours of incubation (Wilcoxon rank-sum test, W = 67, p = 1.5 × 10-8). Following this initial growth phase, CFU counts remained relatively stable over the 168-hour observation period (Fig 5a-b and S1 Fig).
Cell and spore counts (CFU and qPCR) of all Bcbva (a) and B. anthracis strains except VAR28A88 (b), which was excluded due to negligible CFU growth, making comparison unreliable. c | Spore ratio comparison between B. anthracis and Bcbva. B. anthracis VAR28A88, Bcbva CI, and Bcbva CAR-H were excluded due to their differing sporulation behavior. Significance was tested by Wilcoxon rank-sum test with Bonferroni correction. p-values are shown as: * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001.
An exception was B. anthracis VAR28A88, which exhibited a markedly lower cell count of 6.37 × 104 CFU/PGB ± 9 × 104 after 24 hours (S1 Fig c), with even fewer cells detected at subsequent time points.
Similar trends were observed using qPCR, as GE/PGB values increased significantly for all strains during the first 24 hours (Wilcoxon rank-sum test, W = 0, p = 3 × 10-11). In addition, all GE/PGB values were consistently higher than their respective CFU/PGB values, up to tenfold for B. anthracis strains, though to a lesser extent for Bcbva strains (Fig 5a-b and S1 Fig).
Over the course of 168 hours, GE/PGB values slightly decrease for Bcbva, whereas B. anthracis strains showed a moderate increase, particularly up to 72 hours, a trend more clearly detected by qPCR than by CFU enumeration (Fig 5a-b and S1 Fig). Due to the atypically low CFU counts of B. anthracis VAR28A88 compared to the other B. anthracis strains, it was excluded from subsequent CFU-based comparisons.
After 24 hours, both species exhibited comparable growth (Table 2). By 72 hours, however, B. anthracis cell counts had slightly increased, while Bcbva counts had decreased, resulting in a statistically significant difference (Table 2). At 168 hours, a notable difference in cell counts between B. anthracis and Bcbva persisted, although it was no longer statistically significant for CFU (Table 2), whereas the difference remained highly significant in GE values (Table 2).
These results supported the microscopy results, which also revealed reduced biofilm formation of Bcbva compared to B. anthracis strains after 72 and 168 hours (Figs 3-4).
Interestingly, despite the lower CFU counts, GE/PGB values for B. anthracis VAR28A88 increased comparably to those of other strains (S1 Fig c). This pattern was consistent with microscopy observations, which showed robust biofilm formation for VAR28A88 throughout the entire 168-hour period.
Spore counts showed differences between B. anthracis and Bcbva strains. To facilitate comparison, sporulation ratios [%] were calculated relative to the total CFU counts.
Excluding strain VAR28A88, B. anthracis exhibited a significantly lower sporulation rate after 24 hours, reaching only 13.0% ± 12.0%, and showed overall less sporulation over the 168-hour period compared to Bcbva. In contrast, Bcbva strains CA, RCA, and DRC achieved much higher sporulation rates of 58.0% ± 13.0% already at 24 hours, with only minor increases observed thereafter (Fig 5c). These differences between B. anthracis (excluding VAR28A88) and the aforementioned Bcbva strains were highly significant (Wilcoxon rank-sum test, 24 h: W = 1, p = 4.1 × 10 ⁻ ⁵; 72 h: W = 6, p = 0.0004, 168 h: W = 20, p = 0.04; adjusted by Bonferroni correction; B. anthracis VAR28A88, Bcbva CI, and Bcbva CAR-H excluded).
Exceptions to this pattern included the attenuated Bcbva strain CAR-H, which showed sporulation rates comparable to B. anthracis (3.4% ± 1.3%), and Bcbva strain CI, which initially exhibited a very low sporulation rate at 24 hours (<0.1% ± < 0.1%) but then showed a pronounced increase up to 55.1% ± 16.1% at 72 hours post-inoculation.
Comparison of biofilm formation by scanning electron microscopy
To further investigate biofilm architecture and sporulation dynamics, SEM was performed on all strains at the specified time points (24 h, 72 h, and 168 h).
At 24 hours post-inoculation, no notable morphological differences were observed between B. anthracis and Bcbva strains. All samples exhibited densely packed bacterial chains characteristic of B. anthracis, with minimal extracellular matrix production and no visible spores (Fig 6, S2-S4 Figs).
SEM images of B. anthracis Dobichau (a-c) and VAR28A88 (d-f) as well as Bcbva CI (g-i) and CA (j-l) with stepwise increasing magnification from left to right at the marked areas.
By 72 hours, strain-specific differences became evident. While B. anthracis 19/39 and Vollum began to show early signs of sporulation, B. anthracis Dobichau and VAR28A88 still lacked visible spores. In contrast, Bcbva strains displayed more pronounced sporulation at this time point (Fig 7, S2-S3 Figs and S5 Fig).
SEM images of B. anthracis Dobichau (a-c) and VAR28A88 (d-f) as well as Bcbva CI (g-i) and CA (j-l) with stepwise increasing magnification from left to right at the marked areas. Spores are exemplary marked with white arrows.
SEM images also revealed a patchy distribution of biofilm on the PGBs, consistent with observations made by CLSM. Bcbva strains formed denser biofilm structures within the interior regions of the PGBs (Fig 7g-l, S3 Fig and S5 Fig), whereas overall biofilm density appeared lower compared to both B. anthracis and the earlier time point, again consistent with CLSM observations. Among the B. anthracis strains, only Vollum and 19/39 showed a somewhat similar patchy distribution pattern (S5 Fig a-f), while still forming large bacterial aggregates at the rim of the PGBs. Enhanced matrix production was also more prominent in Bcbva strains at this stage (Fig 7g-l, S3 Fig and S5 Fig).
After 168 hours, both B. anthracis and Bcbva strains exhibited extensive sporulation and abundant extracellular matrix production, with no substantial differences between the species (Fig 8, S6 Fig).
SEM images of B. anthracis Dobichau (a-c) and VAR28A88 (d-f) as well as Bcbva CI (g-i) and CA (j-l) with stepwise increasing magnification from left to right at the marked areas. Spores are exemplary marked with white arrows.
A notable exception was strain VAR28A88, which did not show any sporulation or matrix production (Fig 8d-f), retaining a biofilm morphology similar to that observed at 24 and 72 hours, lacking the typical progression seen in other strains.
Discussion
Although biofilm formation in the Bacillus cereus group has been extensively studied [19], the virulent species B. anthracis and, in particular, Bcbva have been largely overlooked. To date (2025), only a few studies have addressed biofilm formation in B. anthracis [20,21,31,32], and no research has been published regarding Bcbva. Our study addressed this gap by examining biofilm formation across diverse strains, including attenuated and laboratory strains as well as wild-type isolates.
Using several microscopic techniques and cell enumeration methods, we demonstrated that different B. anthracis and Bcbva strains are capable of forming biofilms on a modified version of the PGB model described by Konrat et. al [22]. This model was chosen, because it demonstrated high biofilm production and excellent reproducibility, making it well-suited for our study. Furthermore, Raman spectroscopy confirmed that these PGBs closely resemble the chemistry of Sahara sand, which is a component frequently found in soils of sub-Saharan regions [33], where B. anthracis is commonly found [13,34]. Both surfaces have virtually identical chemical compositions, being predominantly composed of SiO2. The main difference between PGB and Sahara sand appeared in their differing atomic arrangement, i.e., their crystallinity. While Sahara sand exhibits a well-defined crystalline structure (quartz), the PGBs display an amorphous structure with glass-like characteristics. As biofilm attachment properties are primarily influenced by chemical composition, surface energy, roughness, and topography of the substrate [35], this difference has been shown to have minimal influence on biofilm formation, as supported by Almaguer-Flores et al. [36]. Additionally, the SEM images (Fig 1e) corroborate that PGBs and Sahara sand particles share a nanotopography.
To characterize biofilm formation on the PGBs, we employed CLSM and SEM, which are powerful and widely used tools for biofilm analysis [37]. However, the three-dimensional structure of the PGB model imposes limitations for microscopy, especially when imaging the inner regions of the PGB. Consequently, classical microtiter plate assays, superior for microscopy, are more commonly used [38–40]. Nonetheless, the structural similarity to natural sand and the often overlooked three-dimensionality of the model outweigh these microscopy limitations in our study.
To compensate for the microscopy constraints, we also performed CFU enumeration. Cell count data complement the imaging by quantifying all bacteria on a PGB, including spores. Because CFU enumeration can be hindered by bacterial aggregation or chaining, particularly relevant for B. anthracis, we supported it with qPCR, which mitigates this issue [41].
Microscopic analysis revealed clear biofilm formation by all tested B. anthracis and Bcbva strains. This finding aligns with previous studies for B. anthracis [20] and is consistent with expectations for Bcbva, which also belongs to the B. cereus group, known for robust biofilm formation [19].
Interestingly, distinct differences were observed between B. anthracis and Bcbva in biofilm formation on PGBs. Over seven days, Bcbva generally exhibited weaker biofilm formation, with smaller and fewer aggregates, especially at days three and seven. Partly, CFU and qPCR data corroborated this observation, as all B. anthracis strains except B. anthracis VAR28A88 showed more biofilm formation after 72 and 168 hours compared to Bcbva. Another notable difference was in sporulation, which was higher in most Bcbva strains at earlier time points. Previous studies which compared sporulation of B. anthracis and Bcbva in liquid cultures confirmed the higher sporulation efficiency of Bcbva [28,42]. Since spores are smaller and more difficult to stain than vegetative cells [43], this may further explain the lower biofilm signal in CLSM images. SEM also confirmed a higher abundance of spores in Bcbva biofilms during later stages.
Another notable observation is that Bcbva cell counts tended to decline slightly after three and seven days, whereas B. anthracis counts showed a slight increase. This, along with the faster and more extensive sporulation in Bcbva, may indicate a reduced proliferation and persistence capacity on the PGB substrate. Given that B. anthracis’s typical reservoir is soil [44], its strong performance in our model is unsurprising. For Bcbva, the PGB substrate may be suboptimal, possibly explaining its limited geographical distribution within rainforest ecosystems [13]. In these environments, soil persistence may be only one transmission route. As various studies have shown [1,10,45–47], carrion flies can transmit B. anthracis and Bcbva, potentially playing a more important role in rainforests than in arid ecosystems. These flies transport bacteria to substrates such as leaves, fruits, and stones [15], suggesting that alternative strategies for proliferation and persistence are critical. This could explain the poorer performance of Bcbva on PGBs, which may not reflect its adaptation to other substrates such as leaves. Indeed, B. cereus [48] and the closely related B. thuringiensis [49] were already found on leave surfaces.
Since both species demonstrated overall robust biofilm formation, this supports the hypothesis of a lifestyle beyond host infection, challenging the common assumption that their proliferation is strictly host-dependent and that they survive solely as dormant spores in the environment [50,51]. Several studies already suggest that B. anthracis has a mode of replication outside a host [15,52,53]. This is particularly important given the puzzling observation of anthrax infections occurring years after the last known cases, which cannot be readily explained when considering the expected environmental dilution and decline of spores over time [16].
Interestingly, the B. anthracis strain VAR28A88, a rare B-branch (B-clade) strain [25], behaved differently from the other B. anthracis and Bcbva strains. It showed virtually no sporulation over seven days but exhibited exceptional biofilm formation. Despite this, we could not cultivate the bacteria, resulting in low CFU counts. However, the bacteria appeared healthy after seven days, suggesting they may have entered a viable but non-culturable (VBNC) state [54]. Supporting this, staining with 5-cyano-2,3-di-(p-tolyl)-tetrazolium chloride (CTC), which detects active respiratory chains on bacterial surfaces [55], produced a signal (S7 Fig), indicating the bacteria were still metabolically active.
In summary, our study provides novel insights into the biofilm formation capabilities of B. anthracis and Bcbva, demonstrating that both species can robustly form biofilms on a soil-analogous substrate. This challenges the traditional view that their replication is strictly confined to animal hosts and suggests that environmental biofilms may play a significant role in their lifecycle and persistence. The observed differences in sporulation dynamics and biofilm architecture between the species likely reflect distinct ecological adaptations, with important implications for their environmental reservoirs and transmission routes. In particular, for Bcbva, further investigations are necessary to explore biofilm formation on alternative substrates, such as fruits or leaves. Overall, our findings emphasize the necessity of considering biofilm-mediated survival and transmission strategies to deepen our understanding of anthrax ecology.
Supporting information
S1 Fig. Quantitative analysis (cell and spore count) of all strains.
Cell and spore count of B. anthracis (a-e) and Bcbva (f-j) biofilms on PGBs over a period of 168 hours. Shown here are the means of three independent biological replicates (n = 3). Error bars show the standard deviation. In green with diamonds the GE, in orange with triangles whole CFU and in grey with circles Spores per PGB are shown. Each PGB corresponds to 15 µl of media used for inoculation.
https://doi.org/10.1371/journal.pntd.0014043.s001
(TIFF)
S2 Fig. SEM images of B. anthracis Sterne 24 – 168 hours post-inoculation.
https://doi.org/10.1371/journal.pntd.0014043.s002
(TIFF)
S3 Fig. SEM images of Bcbva CAR-H 24 – 168 hours post-inoculation.
https://doi.org/10.1371/journal.pntd.0014043.s003
(TIFF)
S4 Fig. SEM images of remaining B. anthracis and Bcbva strains 24 hours post-inoculation.
https://doi.org/10.1371/journal.pntd.0014043.s004
(TIFF)
S5 Fig. SEM images of remaining B. anthracis and Bcbva strains 72 hours post-inoculation.
https://doi.org/10.1371/journal.pntd.0014043.s005
(TIFF)
S6 Fig. SEM images of remaining B. anthracis and Bcbva strains 168 hours post-inoculation.
https://doi.org/10.1371/journal.pntd.0014043.s006
(TIFF)
S7 Fig. Detection of respiratory activity of B. anthracis VAR28A88 with CTC staining.
CLSM images of B. anthracis VAR28A88 (A-C) and B. anthracis Vollum (D) on PGBs stained with CTC (yellow) and DAPI (blue). Respiratory activity for B. anthracis VAR28A88 can be observed at all tested timepoints (24 h (A), 72 h (B) and 168 h (C)) and is comparable to the positive control with B. anthracis Vollum (24 h, D).
https://doi.org/10.1371/journal.pntd.0014043.s007
(TIFF)
Acknowledgments
We especially thank Klaus Heuner for fruitful discussions and valuable assistance during the finalization of the manuscript. We also gratefully acknowledge Michael Laue for his support with the light and electron microscopy aspects of this study. Furthermore, we want to thank W. Beyer, K. Antonation, M. Mori and M. Elschner for providing valuable strains for this study.
References
- 1.
WHO. Anthrax in humans and animals. 4th edition. World Health Organization, International Office of Epizootics; 2008.
- 2. Jernigan JA, Stephens DS, Ashford DA, Omenaca C, Topiel MS, Galbraith M, et al. Bioterrorism-related inhalational anthrax: the first 10 cases reported in the United States. Emerg Infect Dis. 2001;7(6):933–44. pmid:11747719
- 3. Goel AK. Anthrax: A disease of biowarfare and public health importance. World J Clin Cases. 2015;3(1):20–33. pmid:25610847
- 4. Keim P, Grunow R, Vipond R, Grass G, Hoffmaster A, Birdsell DN, et al. Whole Genome Analysis of Injectional Anthrax Identifies Two Disease Clusters Spanning More Than 13 Years. EBioMedicine. 2015;2(11):1613–8.
- 5. Bennett E, Hall IM, Pottage T, Silman NJ, Bennett AM. Drumming-associated anthrax incidents: exposures to low levels of indoor environmental contamination. Epidemiol Infect. 2018;146(12):1519–25. pmid:29970201
- 6. Meselson M, Guillemin J, Hugh-Jones M, Langmuir A, Popova I, Shelokov A, et al. The Sverdlovsk anthrax outbreak of 1979. Science. 1994;266(5188):1202–8.
- 7. Leendertz FH, Ellerbrok H, Boesch C, Couacy-Hymann E, Mätz-Rensing K, Hakenbeck R, et al. Anthrax kills wild chimpanzees in a tropical rainforest. Nature. 2004;430(6998):451–2. pmid:15269768
- 8. Leendertz FH, Lankester F, Guislain P, Néel C, Drori O, Dupain J, et al. Anthrax in Western and Central African great apes. Am J Primatol. 2006;68(9):928–33. pmid:16900500
- 9. Antonation KS, Grützmacher K, Dupke S, Mabon P, Zimmermann F, Lankester F, et al. Bacillus cereus Biovar Anthracis Causing Anthrax in Sub-Saharan Africa - Chromosomal Monophyly and Broad Geographic Distribution. PLoS Negl Trop Dis. 2016;10(9):e0004923. pmid:27607836
- 10. Hoffmann C, Zimmermann F, Biek R, Kuehl H, Nowak K, Mundry R, et al. Persistent anthrax as a major driver of wildlife mortality in a tropical rainforest. Nature. 2017;548(7665):82–6. pmid:28770842
- 11. Klee SR, Brzuszkiewicz EB, Nattermann H, Brüggemann H, Dupke S, Wollherr A, et al. The genome of a Bacillus isolate causing anthrax in chimpanzees combines chromosomal properties of B. cereus with B. anthracis virulence plasmids. PLoS One. 2010;5(7):e10986. pmid:20634886
- 12. Dragon DC, Rennie RP. The ecology of anthrax spores: tough but not invincible. Can Vet J. 1995;36(5):295–301. pmid:7773917
- 13. Romero-Alvarez D, Peterson AT, Salzer JS, Pittiglio C, Shadomy S, Traxler R, et al. Potential distributions of Bacillus anthracis and Bacillus cereus biovar anthracis causing anthrax in Africa. PLoS Negl Trop Dis. 2020;14(3):e0008131. pmid:32150557
- 14. Hugh-Jones ME, de Vos V. Anthrax and wildlife. Rev Sci Tech. 2002;21(2):359–83. pmid:11974621
- 15. Jiranantasak T, Benn JS, Metrailer MC, Sawyer SJ, Burns MQ, Bluhm AP, et al. Characterization of Bacillus anthracis replication and persistence on environmental substrates associated with wildlife anthrax outbreaks. PLoS One. 2022;17(9):e0274645. pmid:36129912
- 16. Lindeque PM, Turnbull PC. Ecology and epidemiology of anthrax in the Etosha National Park, Namibia. Onderstepoort J Vet Res. 1994;61(1):71–83. pmid:7898901
- 17. Costerton JW, Stewart PS, Greenberg EP. Bacterial biofilms: a common cause of persistent infections. Science. 1999;284(5418):1318–22. pmid:10334980
- 18. Flemming H-C, Wingender J, Szewzyk U, Steinberg P, Rice SA, Kjelleberg S. Biofilms: an emergent form of bacterial life. Nat Rev Microbiol. 2016;14(9):563–75. pmid:27510863
- 19. Majed R, Faille C, Kallassy M, Gohar M. Bacillus cereus Biofilms - Same, Only Different. Front Microbiol. 2016;7:1054. pmid:27458448
- 20. Lee K, Costerton JW, Ravel J, Auerbach RK, Wagner DM, Keim P, et al. Phenotypic and functional characterization of Bacillus anthracis biofilms. Microbiology (Reading). 2007;153(Pt 6):1693–701. pmid:17526827
- 21. Arora G, Sajid A, Virmani R, Singhal A, Kumar CMS, Dhasmana N, et al. Ser/Thr protein kinase PrkC-mediated regulation of GroEL is critical for biofilm formation in Bacillus anthracis. NPJ Biofilms Microbiomes. 2017;3:7. pmid:28649408
- 22. Konrat K, Schwebke I, Laue M, Dittmann C, Levin K, Andrich R, et al. The Bead Assay for Biofilms: A Quick, Easy and Robust Method for Testing Disinfectants. PLoS One. 2016;11(6):e0157663. pmid:27315246
- 23. Brézillon C, Haustant M, Dupke S, Corre J-P, Lander A, Franz T, et al. Capsules, toxins and AtxA as virulence factors of emerging Bacillus cereus biovar anthracis. PLoS Negl Trop Dis. 2015;9(4):e0003455. pmid:25830379
- 24. Klee SR, Ozel M, Appel B, Boesch C, Ellerbrok H, Jacob D, et al. Characterization of Bacillus anthracis-like bacteria isolated from wild great apes from Cote d’Ivoire and Cameroon. J Bacteriol. 2006;188(15):5333–44. pmid:16855222
- 25. Wattiau P, Klee SR, Fretin D, Van Hessche M, Ménart M, Franz T, et al. Occurrence and genetic diversity of Bacillus anthracis strains isolated in an active wool-cleaning factory. Appl Environ Microbiol. 2008;74(13):4005–11. pmid:18487406
- 26. Klee SR, Nattermann H, Becker S, Urban-Schriefer M, Franz T, Jacob D, et al. Evaluation of different methods to discriminate Bacillus anthracis from other bacteria of the Bacillus cereus group. J Appl Microbiol. 2006;100(4):673–81. pmid:16553722
- 27. Elschner MC, Busch A, Schliephake A, Gaede W, Zuchantke E, Tomaso H. High-Quality Genome Sequence of Bacillus anthracis Strain 14RA5914 Isolated during an Outbreak in Germany in 2014. Genome Announc. 2017;5(40).
- 28. Gummelt C, Dupke S, Howaldt S, Zimmermann F, Scholz HC, Laue M, et al. Analysis of Sporulation in Bacillus cereus Biovar anthracis Which Contains an Insertion in the Gene for the Sporulation Factor σK. Pathogens. 2023;12(12):1442. pmid:38133325
- 29.
R Core Team. R: A Language and Environment for Statistical Computing. 4.3.0 edition. Vienna, Austria: R Foundation for Statistical Computing; 2023.
- 30.
Fox J, Weisberg S. An R Companion to Applied Regression. 3 edition. Thousand Oaks (CA): Sage; 2019.
- 31. Schuch R, Fischetti VA. The secret life of the anthrax agent Bacillus anthracis: bacteriophage-mediated ecological adaptations. PLoS One. 2009;4(8):e6532. pmid:19672290
- 32. Hermanas TM, Subramanian S, Dann CE 3rd, Stewart GC. Spore-Associated Proteins Involved in c-di-GMP Synthesis and Degradation of Bacillus anthracis. J Bacteriol. 2021;203(17):e0013521. pmid:34096779
- 33. Sunnu A, Afeti G, Resch F. A long-term experimental study of the Saharan dust presence in West Africa. Atmospheric Research. 2008;87(1):13–26.
- 34. Chikerema SM, Pfukenyi DM, Hang’ombe BM, L’Abee-Lund TM, Matope G. Isolation of Bacillus anthracis from soil in selected high-risk areas of Zimbabwe. J Appl Microbiol. 2012;113(6):1389–95. pmid:22984812
- 35. Song F, Koo H, Ren D. Effects of Material Properties on Bacterial Adhesion and Biofilm Formation. J Dent Res. 2015;94(8):1027–34. pmid:26001706
- 36. Almaguer-Flores A, Silva-Bermudez P, Galicia R, Rodil SE. Bacterial adhesion on amorphous and crystalline metal oxide coatings. Mater Sci Eng C Mater Biol Appl. 2015;57:88–99. pmid:26354243
- 37. Lawrence JR, Neu TR. Confocal laser scanning microscopy for analysis of microbial biofilms. Methods Enzymol. 1999;310:131–44. pmid:10547787
- 38. Christensen GD, Simpson WA, Younger JJ, Baddour LM, Barrett FF, Melton DM, et al. Adherence of coagulase-negative staphylococci to plastic tissue culture plates: a quantitative model for the adherence of staphylococci to medical devices. J Clin Microbiol. 1985;22(6):996–1006. pmid:3905855
- 39. Merritt JH, Kadouri DE, O’Toole GA. Growing and analyzing static biofilms. Curr Protoc Microbiol. 2005;Chapter 1:Unit 1B.1. pmid:18770545
- 40. Auger S, Krin E, Aymerich S, Gohar M. Autoinducer 2 affects biofilm formation by Bacillus cereus. Appl Environ Microbiol. 2006;72(1):937–41. pmid:16391139
- 41. Childers NK, Osgood RC, Hsu K-L, Manmontri C, Momeni SS, Mahtani HK, et al. Real-time quantitative polymerase chain reaction for enumeration of Streptococcus mutans from oral samples. Eur J Oral Sci. 2011;119(6):447–54. pmid:22112030
- 42. Jiranantasak T, Bluhm AP, Chabot DJ, Friedlander A, Bowen R, McMillan IA, et al. Toxin and capsule production by Bacillus cereus biovar anthracis influence pathogenicity in macrophages and animal models. PLoS Negl Trop Dis. 2024;18(12):e0012779. pmid:39715264
- 43. Lechtman MD, Bartholomew JW, Phillips A, Russo M. Rapid methods of staining bacterial spores at room temperature. J Bacteriol. 1965;89(3):848–54. pmid:14273671
- 44. Carlson CJ, Kracalik IT, Ross N, Alexander KA, Hugh-Jones ME, Fegan M, et al. The global distribution of Bacillus anthracis and associated anthrax risk to humans, livestock and wildlife. Nat Microbiol. 2019;4(8):1337–43. pmid:31086311
- 45. Fasanella A, Scasciamacchia S, Garofolo G, Giangaspero A, Tarsitano E, Adone R. Evaluation of the house fly Musca domestica as a mechanical vector for an anthrax. PLoS One. 2010;5(8):e12219. pmid:20808920
- 46. Blackburn JK, Van Ert M, Mullins JC, Hadfield TL, Hugh-Jones ME. The necrophagous fly anthrax transmission pathway: empirical and genetic evidence from wildlife epizootics. Vector Borne Zoonotic Dis. 2014;14(8):576–83. pmid:25072988
- 47. Braack LE, De Vos V. Feeding habits and flight range of blow-flies (Chrysomyia spp.) in relation to anthrax transmission in the Kruger National Park, South Africa. Onderstepoort J Vet Res. 1990;57(2):141–2. pmid:2216348
- 48. Antequera-Gomez ML, Diaz-Martinez L, Guadix JA, Sanchez-Tevar AM, Sopena-Torres S, Hierrezuelo J, et al. Sporulation is dispensable for the vegetable-associated life cycle of the human pathogen Bacillus cereus. Microb Biotechnol. 2021;14(4):1550–65.
- 49. Bizzarri MF, Bishop AH. The ecology of Bacillus thuringiensis on the Phylloplane: colonization from soil, plasmid transfer, and interaction with larvae of Pieris brassicae. Microb Ecol. 2008;56(1):133–9. pmid:17973155
- 50. Turnbull PCB. Introduction: anthrax history, disease and ecology. Curr Top Microbiol Immunol. 2002;271:1–19. pmid:12224519
- 51. Titball RW, Turnbull PC, Hutson RA. The monitoring and detection of Bacillus anthracis in the environment. Soc Appl Bacteriol Symp Ser. 1991;20:9S–18S.
- 52. Saile E, Koehler TM. Bacillus anthracis multiplication, persistence, and genetic exchange in the rhizosphere of grass plants. Appl Environ Microbiol. 2006;72(5):3168–74. pmid:16672454
- 53. Dey R, Hoffman PS, Glomski IJ. Germination and amplification of anthrax spores by soil-dwelling amoebas. Appl Environ Microbiol. 2012;78(22):8075–81. pmid:22983962
- 54. Golpayegani A, Douraghi M, Rezaei F, Alimohammadi M, Nodehi RN. Propidium monoazide-quantitative polymerase chain reaction (PMA-qPCR) assay for rapid detection of viable and viable but non-culturable (VBNC) Pseudomonas aeruginosa in swimming pools. J Environ Health Sci Eng. 2019;17(1):407–16. pmid:31297217
- 55. Stockert JC, Horobin RW, Colombo LL, Blázquez-Castro A. Tetrazolium salts and formazan products in Cell Biology: Viability assessment, fluorescence imaging, and labeling perspectives. Acta Histochem. 2018;120(3):159–67. pmid:29496266