Introduction

Bacterial endospores are formed by strains of Bacillus during nutrient limitation or when conditions are unfavorable to support vegetative growth. Spores are a major concern to the food industry because of spoiling in addition to the pathogenic potential of Bacillus cereus producing toxins inducing diarrhea and dehydration1. Spores are extremely resistant and durable; they can remain metabolically inactive where their survival is protected by multiple factors including low core water content, low permeability of the inner membrane, in addition to genome protection by specific DNA binding proteins, as well as a proteinaceous multilayering that additionally enables them to resist harsh physicochemical and environmental conditions that would otherwise kill vegetative cells2. In particular, B. cereus is present in a wide range of environments, such as soils, sediments, dust, or plants3. Bacillus cereus spores have been detected in rain water at concentrations ranging from 101 to 2 × 102 CFU/mL, and at 105 CFU/mL in soil4. Bacillus thuringiensis spores have been extensively used as biopesticides by the agriculture and forestry industries5,6, and direct application of these spores to rivers and lakes is used to control mosquitoes and blackflies7.

Collecting microorganisms from contaminated water constitutes the important first step towards detection and enumeration of target microorganisms. Conventional methods are based on vacuum filtration of manually collected water samples8. Examples of the filtration methods and type of filters investigated for this purpose are provided in Table 1. The reported concentration factors (CF) of microorganisms achieved with filtration are in the 5–586 range9,10,11,12. It is important to indicate that CF depend on the filtered material and the backwash volumes. As an example, CF = 666 could be obtained for the filtered volume of 100 L and the backwash of 150 mL if all microbes were recovered10. But, it is unlikely to recover 100% of the microbes, not to mention that collecting 100 L samples is highly unpractical for conducting aquasi-continuous analysis of environmental water samples.

Table 1 Examples of filtration methods and concentration factors in microbial sampling.
Full size table

The presence of spores in water has typically been determined by enumeration of colony growth on nutrient agar following contact of filtration membranes13. Methodology for filtration and pre-concentration of spores intended for further processing and analysis has largely been missing in the literature. Our preliminary results to pre-concentrate B. cereus spores with a polyvinylidene difluoride (PVDF) filter-based system were unsatisfactory. This was due to the excessive adhesion of spores to the filter material and surfaces of the components employed for constructing the filtration system, originally designed for filtration of L. pneumophila12. That observation is consistent with the reports of strong adhesion of bacterial spores to hydrophobic surfaces14,15.

Furthermore, it has been reported that the electrical polarity of surfaces, as well as spore morphology can affect the adhesion of spores to different solid surfaces16. For instance, it has been observed that B. cereus spores show high affinity to surfaces of polyethylene of high density (PEHD), Teflon and polyvinyl chloride (PVC), while they exhibit lower adherence to stainless steel, polyamide-6, and minimal adherence to glass17. The hydrophobic characteristic of B. cereus and B. thuringiensis spores is mainly attributed to their outermost layer called exosporium with collagen-like region and the C-terminal domain of a predominantly abundant surface proteoglycan called Bacillus collagen like protein A (BclA)18,19,20,21.

To investigate the pre-concentration of spores intended for biosensing, we have examined PVDF, fiberglass (FG), and polyether sulfone (PES) filters installed in a water sampling module dedicated for collecting spores (WSM-S). The filter with superior performance was employed for processing Bacillus spores and evaluating further its repetitive use.

Materials and methods

B. cereus group strains and reagents

Bacillus thuringiensis subspecies (subsp.) kurstaki strain HD1 (B. thuringiensis kurstaki) derived from Foray 48B, and Bacillus cereus ATCC 11,778 obtained from the American Type Culture Collection were used in a spore state for the concentration performance evaluation of WSM-S. The B. cereus and B. thuringiensis spores were grown for 8 days to sporulate22. The sporers were purified by using an aqueous two-phase system comprising polyethylene glycol (PEG) and potassium phosphate with ~ 90% purity for B. cereus spores and ~ 93% for B. thuringiensis spores, as confirmed by Schaeffer–Fulton staining method and microscopy23. Phosphate buffered saline (PBS) solution (10×: 20 mM phosphate-1.4 M NaCl, pH 7.4) used after dilution to 1 × PBS to backwash spores, sodium chloride employed for the preparation of the spores suspensions to be filtered, Luria Bertani broth (LB) and agar were all obtained from Sigma-Aldrich (Ontario, Canada). Hydrophilic polyvinylidene fluoride (PVDF) membranes (0.45 µm, Sigma-Aldrich, ON, Canada), fiber glass (FG) membranes (0.3 µm, Sterilitech, WA, USA) and polyethersulfone (PES) membranes (0.45 µm, Cole Parmer, QC, Canada) were all 47 mm in diameter.

Water sampling module

A schematic diagram of the WSM-S unit is shown in Fig. 1. The unit employs a peristaltic pump and a set of pinch valves that allow processing of bacterial suspensions while eliminating contact of metallic or plastic parts of the valves employed for building a conventional WSM12. That minimized chemical degradation of the system and eliminated adventitious contamination of processed samples. The C-FLEX type tubing with high biochemical compatibility did not interact with the investigated spores as verified with a series of dedicated experiments (Fig. S1, Supplementary information). The automated process of filtration, backwashing, disinfecting, and cleaning was ensured by an Arduino microcontroller connected to a relay of 12 V sources employed for switching on/off individual valves.

Figure 1
figure 1

Schematic diagram of the water sampling module (WSM-S) used to filter and concentrate spores into a 5 mL backwash solution from a 250 mL water reservoir. The filtration (blue arrows) and backwash (red arrows) runs are indicated by respective flow directions controlled by the peristaltic pump and a set of electromagnetic pinch valves.

Full size image

The operation of the system is based on the application of a user-friendly Python coded interface. The first step involves filtration of a pre-determined volume of bacterial suspension in water (250 mL in the current experiments) through either a PVDF, FG or PES filter (blue arrows in Fig. 1). The filtered water is collected in a waste bottle and the spores are retained by the filter membrane. During the second step, the spores retained by the membrane are collected into a Falcon tube by reversing the flow direction and backwashing with 1 × PBS (red arrows in Fig. 1). Only 5 mL of the backwash are collected (monitored with a sensitive photodiode) to avoid excessive dilution of spores. A time sequence of the processing steps executed by the WSM-S unit (cleaning of the system, filter installation, suction of spore suspension, and backwashing) is schematically shown in Supplementary information (Fig. S2).

Filtration of B. thuringiensis kurstaki and B. cereus spores

Initial suspensions were prepared by spiking saline water (9 g/L NaCl) with investigated spores. The suspensions were circulated through the WSM-S module equipped with either a PVDF, FG or PES filter placed in the holder. To determine the initial concentration, 100 μL of samples were taken from B. thuringiensis kurstaki or B. cereus suspensions and seeded on LB-agar. Following a 24-h growth period at 35 °C, the concentration was determined by the plate count method. From a tenfold diluted backwash (BW), several 100 μL samples were seeded in LB agar and the numbers of grown colonies were counted after a 24-h growth period at 35 °C.

Before each filtration run, the module was decontaminated by incubating with a 1% bleach solution for 60 min. The system was then rinsed by 3 successive washes using hot water (~ 80 °C), and 3 times with room-temperature deionized (DI) water. The filtration performance was evaluated by calculating the concentration factor (CF) acording the following formula:

$$\mathrm{CF }= \frac{{\text{Spore concentration in the backwash suspension}}}{{\text{Spore concentration in the filtered suspension}}}$$

Under ideal conditions, where 100% of spores were collected from a 250 mL sample of suspension, the maximum concentration factor (CFmax), calculated as the ratio of the initial to the final volume, is 50×.

To evaluate the adherence of B. cereus spores to FG and PES filters, the after-backwashing filters were ground in 50 mL of 1 × PBS using a bleach-sterilized high-speed hand blender (Cuisinart Smart Stick Two-Speed Hand Blender—CSB-75BKC). 200 µL of the resulting filters micro pulp suspensions were used to inoculate three LB agar plates and incubated for 24 h at 35 °C. The number of colonies present helped determine the adherence of B. cereus spores to the filter material. In the second experiment, after backwashing FG, PES or PVDF filters employed for filtration of the B. cereus suspension at initial concentration of 1 × 102 spores/mL, the filters were placed in a Petri dish with LB agar, making sure that all the filter surface remained in contact with the medium. The filters were removed after 1 min, and the growth of B. cereus was evaluated after 24-h incubation at 35 °C.

Reusability of FG filters

Repetitive runs with FG filters were carried out for up to 6 consecutive filtrations of B. thuringiensis kurstaki spore suspensions. Each filtration run consisted of the following steps: (1) Flow of 250 mL of a saline water suspension of B. thuringiensis kurstaki spores, (2) Production of 5 mL of backwashed suspension in 1 × PBS, (3) 1 h exposure of the WSM-S system to 1% Bleach, (4) 3-times washing in hot (80 °C) DI water, and (5) 3-times washing with DI water at room temperature. The steps were repeated with the same filter, and 100 μL of filtrate collected after each filtration run were used to seed LB-agar to enumerate spores not captured by the filter.

Confocal microscopy measurements

The morphology of FG filters was investigated with a confocal microscope (KEYENCE model VK-X1100). The instrument works with a 404 nm laser operating in the reflection mode. The images collected with a 150× lens had a maximum examination depth of 150 µm, and 9375 images collected with a 0.016 µm step were used for the 3D reconstruction. The surface area of 92.1 µm × 72.8 µm was scanned with a 0.13 µm resolution, 1.6 nm vertical resolution and the overall magnification of 3600×. The average roughness (Sa), defined as a mean difference in height of each point compared to the arithmetical mean of the surface, was used to characterize the texture of the filter.

Concentrating B. thuringiensis kurstaki spores from cooling tower water samples

Samples of cooling tower water (CTW) were acquired from one of the cooling towers of the Université de Sherbrooke. The OD600nm of CTW samples was at 0.01 compared to "zero" of a blank saline water sample. The CTW samples were inoculated with B. thuringiensis kurstaki spores to produce a suspension at 2.75 × 102 spores/mL.

Four filtration runs, each of a 250 mL suspension, were completed with WSM-S using a FG filter and 5-mL backwashes were produced each time. Between each filtration run, the WSM-S system was washed only with DI water. The samples were heated for 10 min at 80 °C to inactivate the CTW flora. The samples prepared in this way were employed to inoculate LB-agar and the results of the cultivation at 35 °C were analyzed after the standard incubation period of 24 h.

Results

Concentration factors for B. thuringiensis kurstaki and B. cereus spores using different types of filters

The CF values achieved for B. thuringiensis kurstaki and B. cereus spores with FG, PES, and PVDF filters are presented in Fig. 2. CFs of ~ 20 were obtained with the FG filter for both types of spores. The PES filter provided CF of ~ 21 and ~ 4 for B. thuringiensis kurstaki and B. cereus spores, while CF values of ~ 11 and ~ 7 were obtained with the PVDF filter for B. thuringiensis kurstaki and B. cereus spores. Thus, FG and PES filters concentrate B. thuringiensis kurstaki spores by the same order of magnitude, but only the FG filter concentrates B. cereus spores better than the other two filters.

Figure 2
figure 2

Concentration factors of B. thuringiensis kurstaki and B. cereus spores obtained with FG, PES, and PVDF filters for suspensions at initial concentration of 3.14.2 × 102 spores/mL. The data bars represent standard errors of the mean from three separate experiments.

Full size image

Following the filtration of the suspension containing B. cereus spores, the WSM-S containing each filter type was backwashed and LB-agar was stamped with the filters. The resulting growth indicated that lower numbers of B. cereus spores were retained on the FG filter while most of the spores were retained by the PES filter (Fig. 3).

Figure 3
figure 3

Images of stamp-grown B. cereus spores from backwashed FG (a), PVDF (b) and PES (c) filters used to retain spores from 250 mL suspensions at 1 × 102 spores/mL.

Full size image

The number of B. cereus spores enumerated in suspensions produced by ground FG and PES filters are presented in Table 2.

Table 2 Number of B. cereus spores remaining on FG and PES filters following the backwashing procedure.
Full size table

This experiment demonstrates that ~ 26% and 68% of the initial concentrations of B. cereus spores remain on the FG and PES filters, respectively, which is consistent with the greater efficiency of pre-concentrating spores with the FG filters.

Concentration factors for FG filters collecting B. thuringiensis kurstaki and B. cereus spores from low concentration suspensions (N ≤ 3.2 × 102 spores/mL)

Further investigation of the WSM-S performance was carried out using low concentration (10–3.2 × 102 spores/mL) suspensions of B. thuringiensis kurstaki or B. cereus. Table 3 shows the CFs values of 22.5 ± 1.5 and 19.5 ± 1 for B. cereus at 10 and 3.1 × 102 spores/mL, respectively. For B. thuringiensis kurstaki spores at initial concentrations of 5 × 101, 2 × 102, and 3.2 × 102 spores/mL, the CF values were 23.8 ± 3.4, 22 ± 1.5 and 20.3 ± 2, respectively. Thus, the CFs values appear independent of the initial concentrations for the investigated concentration range.

Table 3 Concentration factors of B. thuringiensis kurstaki and B. cereus spores for weakly contaminated saline water suspensions obtained with the FG filter.
Full size table

Reusability of FG filters

To investigate the reusability of FG filters, 8 consecutive filtrations of B. thuringiensis kurstaki spores suspensions (4.3 × 102 spores/mL) were carried out using the same filter. In this experiment no bleach treatment was used between filtration runs. Spore concentrations in the backwash and in the filtrate were determined by plate counting and CFs were calculated for each filtration run (Table 4). CF values between 21.7 ± 0.5 and 18 ± 0.5 were achieved for the first 6 filtration runs but a significant drop in CF was observed for the 7th and 8th filtrations (14 ± 2.2 and 8.9 ± 1.8, Table 4). The loss of the filter’s ability to retain spores for more than 6 runs was confirmed by the presence of spores in the 7th and 8th filtrates.

Table 4 Concentration factors of B. thuringiensis kurstaki spores achieved for backwashing of the same FG filter and concentration of spores in filtrates obtained with bleach-free filtration runs (The limit of detection at 10 spores/mL for 100 µL samples).
Full size table

The effect of exposing FG filters to 60 min bleaching was investigated with repeated filtration runs using suspension at 2.5 × 102 spores/mL. The results illustrate that bleaching had a negligible effect on the CF values for up to 3 filtration runs, which suggests that the structural integrity of an FG filter is maintained under these conditions (Table 5). It is possible that a shorter than 1 h exposure of the filters to bleach may also play a role in the longer-term maintenance of their structural integrity, but this effect was not investigated in the current experiment.

Table 5 Concentration factors of B. thuringiensis kurstaki spores achieved for backwashing and bleaching FG filters reused for up 6 times.
Full size table

Representative confocal microscopy micrographs of a fresh filter and of a filter used for 3 consecutive filtration runs with bleaching applied after each backwashing step are shown in Fig. 4. The roughness of the filters is illustrated by the decrease of the Sa value from 5.87 µm (fresh filter) to 2.81 µm (reused filter). A significant drop in the Sa value to 4.12 µm was observed for the filter backwashed and bleached only once (Table 6). The reuse of filters further reduced their Sa values to 2.4–2.9 µm, without a significant difference observed for 3-times or 6-times re-used filters (Table 6). Ultimately, filtration of a B. thuringiensis kurstaki spore suspension at 1 × 106 spores/mL without backwashing produced a filter with Sa = 0.94 µm (see Supplementary information for examples of confocal microscopy images of the related filters, Fig. S3).

Figure 4
figure 4

Confocal microscopy images of (a) a fresh FG filter, and (b) a FG filter after 3 filtration runs with backwashing and bleach decontamination. The scale bar corresponds to 10 μm.

Full size image
Table 6 Sa values for FG filters used for the indicated number of successive filtrations including backwashing and bleaching.
Full size table

Culture analysis showed that repeated filtrations (up to 4 times) of 3 × 103 B. thuringiensis kurstaki spores/mL generated filtrates free of spores. However, 8 × 101 and 7.5 × 102 spores/mL were detected in the filtrates produced with the 5th and 6th filtrations, respectively, which constitutes 2.6 and 25% of the original spore suspension (see Fig. S4). These results are consistent with the significantly reduced CF values observed for the 5th and 6th filtration runs (Table 5). The repetitive collection of B. thuringiensis kurstaki spores suspended in CTW using the same FG filter re-used up to 4 times (washed with DI water between each filtration step) revealed that the CF decreased from 18.9 to 16.7 (Supplementary information, Table S1).

Discussion

The experiments involving PVDF, FG, and PES filters for collecting B. thuringiensis kurstaki and B. cereus spores revealed the superior performance of the FG filter delivering CF ≅ 20 for both types of spores. The performance of a filter is of particular importance to the backwashing step, and dependant on the electrostatic interaction between hydrophobic Bacillus strains and membranes of the investigated filters. For instance, zeta (ζ)‐potential of the glass surface is equal to − 50 mV at pH = 724, thus the repulsive interactions of the hydrophilic FG filter with hydrophobic spores is expected to facilitate the release of spores during the backwashing step17,25. This compares to the reduced performance of PES filters whose material is characterized by ζ‐potential of − 20 mV in a comparable pH environment26. Furthermore, the poorest filtration performance of PVDF filters observed in our experiments (Fig. 2) must be related to the augmented hydrophobicity of the filter material characterized by ζ‐potential of + 35 mV. However, other authors have reported that the ζ‐potential of PVDF filters grafted with hydrophilic molecules, such as lysine or sulfone, could become as high as − 50 mV27. The analysis of used filters revealed that some spores remained on the filter material after backwashing (Fig. 3), which led to reduced CF values for all the investigated filters. The repetitive use of a filter could lead to further reduced CFs, as observed with the 4th filtration run involving the bleaching step included in the process (Table 4).

The adhesion of spores to different surfaces has been the subject of extensive investigations. For instance, the extended Derjaguih, Landau, Verwey and Overbeek (XDVLO) theory and environmental scanning electron microscopy (ESEM) were used to predict and experimentally verify the attachment of four Penicillium strains to cedar wood28. The model was corroborated by data concerning a specific Penicillium strain, but did not fit the other three strains investigated by those authors28. That illustrates the complexity of the phenomena related to the microbial spore interactions with different surfaces.

The results discussed in this report suggest that filtration of B. cereus spores that leads to CF = 20 with FG filters reused up to 5 times is attractive for supporting the operation of a semi-automated workstation designed for monitoring environmental or agricultural water environments. We note that a CF = 50 would correspond to the perfect filtration of a 250 mL suspension and collecting all the spores in a 5 mL volume of the backwash material. Such a result seems unrealistic in view of the studies of cake backwashing from membrane filters that illustrated a dependence of the process on parameters such as turbulence, backwash duration, as well as flow speed and pressure29,30,31. Given the reasonable reproducibility of the backwashing conditions achieved with our WSM-S unit, the difference in CF values obtained in our experiments with different filters is mainly related to the weaker adherence of the B. thuringiensis kurstaki spores to the FG filter material 17. The evaluation of a trapping efficiency of B. cereus spores after backwashing showed that 68% of the filtered spores were still attached to the PES filter compared to 26% of these spores attached to the FG (Table 2). That result confirms that B. cereus spores interact strongly with the PES filter, which is consistent with the lower CF values observed for this filter.

Conclusions

A filtration module (WSM-S) dedicated for collecting water samples containing spores from members of the Bacillus cereus group was designed and evaluated for the potential application as a quasi-autonomously operating tool. The application of pinch valves eliminated the risk of cross-contamination due to the potential reaction between materials commonly used in construction of electromagnetic valves and sampled liquids. The backwashing of PES and PVDF filters revealed the difficulty in releasing spores trapped by those filters, while FG filters were found efficient for the repetitive collection of spores. The structural integrity of FG filters investigated with confocal microscopy revealed a reduction of the filter porosity indicated by the Sa diminution from 5.87 µm (fresh filter) to 2.81 µm (3 times used filter). The 20× concentration factor of B. thuringiensis kurstaki spores obtained with the FG filter and reproduced with 4 consecutive runs (backwashed and bleached between each run) suggests attractive application of such a filter for the operation of a semi-automated water sampling module. We argue that a greater number of repetitive filtrations runs using similar filters should be possible with the further advanced knowledge of materials and the advancement of backwashing protocols. This should lead towards practical advancement of the concept of a quasi-autonomous biosensing workstation for monitoring the presence of bacteria and spores in the environmental and anthropogenic water.