To develop a staining method for specific detection of metabolically active (viable) cells in biofilms of the foodborne pathogen Campylobacter jejuni.
To develop a staining method for specific detection of metabolically active (viable) cells in biofilms of the foodborne pathogen Campylobacter jejuni.
Conversion of 2,3,5 triphenyltetrazolium chloride (TTC) to insoluble, red 1,3,5-triphenylformazan (TPF) was dependent on metabolic activity of Camp. jejuni. When used with chicken juice, TTC staining allowed quantification of Camp. jejuni biofilm levels, whereas the commonly used dye, crystal violet, gave high levels of nonspecific staining of food matrix components (chicken juice). The assay was optimized to allow for monitoring of biofilm levels and adapted to monitor levels of Camp. jejuni in broth media.
Staining with TTC allows for the quantification of metabolically active Camp. jejuni and thus allows for quantification of viable cells in biofilms and food matrices. The TTC staining method can be adapted to quantify bacterial cell concentration in a food matrix model, where the accepted method of A600 measurement is not suitable due to interference by components of the food matrix.
2,3,5 Triphenyltetrazolium chloride (TTC) staining is a low-cost technique suitable for use in biofilm analysis, allowing rapid and simple imaging of metabolically active cells and increasing the methods available for biofilm assessment and quantification.
Biofilms are defined as either single or multiple species of bacteria that are attached to a surface and embedded in an extracellular polymeric matrix, which can include, but is not limited to a mixture of polysaccharides, proteins, nucleic acids and phospholipids (Siringan et al. 2011). The study of bacterial biofilms is highly relevant as they impact on many aspects of human activity, ranging from biofouling in marine and liquid-handling settings (De Beer et al. 1994), survival of various bacteria in the food chain (Hood and Zottola 1995) to involvement in debilitating infections in patients with existing medical conditions or medical implants (Hunter 2008). Biofilms have also been utilized in various artificial systems such as waste water treatment (Sutherland 2001) and fuel cells (Erable et al. 2010).
Diarrhoeal illness impacts heavily on the health of the global population, accounting for 15% of fatalities in children under 5 years old (Mathers et al. 2009). Many bacterial infections are caused by ingestion of food contaminated with bacteria, such as Escherichia coli, Salmonella, Listeria and Campylobacter (Sofos and Geornaras 2010). All of these pathogens have been shown to form biofilms (Harmsen et al. 2010; Iibuchi et al. 2010; Teh et al. 2010; Beauchamp et al. 2012), and such biofilms are thought to aid survival of these pathogenic bacteria in the adverse conditions found within the food production and storage chain, ultimately ending with consumption of the contaminated food and subsequent infection (Hall-Stoodley et al. 2004).
Commonly used experimental methods to study biofilms include flow cell models and static cultures (Hall-Stoodley et al. 2004), but often rely on standard bacteriological media for consistency and reproducibility, making the data obtained difficult to translate to the situations found in food production chains. For Campylobacter jejuni, such a translation for the food chain matrix was previously addressed in an experimental model based on chicken juice (an exudate recovered from frozen chicken carcasses upon thawing) (Birk et al. 2004). Chicken juice, obtained from commercially sourced chickens, is readily available and has been shown to enhance the survival of Camp. jejuni in storage conditions relevant to the food industry (Birk et al. 2006), but to our knowledge has not been used to study biofilm formation in Camp. jejuni. Biofilm models that model industrial conditions are increasingly important to further understanding of bacterial survival in the food chain (Birk et al. 2004; Balamurugan et al. 2011).
Levels of biofilm formation are usually assessed by dye-based staining techniques, which broadly fall into two categories: nonspecific dyes and dyes targeting specific molecules within the biofilm, such as fluorescent dyes that can be imaged using confocal laser scanning microscopy (Lawrence and Neu 1999). Such specific dyes are useful for detailed and dynamic imaging of biofilms (Baird et al. 2012), but are costly and have labour-intensive and time-consuming methodologies, making them an inappropriate tool when rapid quantification of biofilms is required. Conversely nonspecific dyes may overestimate the quantity of viable biofilm cells present due to nonspecific binding of the matrix or surface components.
The most commonly used method to detect and quantify biofilms is by staining with crystal violet (Chavant et al. 2007). This dye is nonspecific, as it binds to surface molecules with a negative charge, found both on the bacteria and in the extracellular matrix of the biofilm (Extremina et al. 2011). Crystal violet will bind both planktonic cells and biofilm material (Pan et al. 2010). This has led some authors to criticize crystal violet for its relatively high interassay variability, especially when compared to other imaging methods (Li et al. 2003).
Tetrazolium salts have been used previously to study cell growth and biofilm formation in various bacterial models (Tengerdy et al. 1967; Schaule et al. 1993), but not with Camp. jejuni biofilms. In this study, we demonstrate that crystal violet gives high levels of nonspecific (false-positive) staining in the presence of food matrices and offer an alternative staining method based on an optimized methodology utilizing the reduction of the metabolic stain 2,3,5 triphenyltetrazolium chloride (TTC) to insoluble, red crystals of 1,3,5-triphenylformazan (TFP) (Bačkor and Fahselt 2005; Berends et al. 2010), which allows for the specific staining of adhered, metabolically active Camp. jejuni. We have also adapted the TTC biofilm staining protocol to provide an alternative method for measuring cell concentration of Camp. jejuni in a food matrix model, where A600 measurements are not possible due to the high level of background particulate formation.
Campylobacter jejuni strain NCTC 11168 was routinely cultured in a MACS-MG-1000 controlled atmosphere cabinet (Don Whitley Scientific, Shipley, UK) under microaerobic conditions (85% N2, 5% O2, 10% CO2) at 37°C, essentially as described previously (Reuter et al. 2010). For growth on plates, strains were either grown on Brucella agar or blood agar base (BAB) with Skirrow supplement (10 μg ml−1 vancomycin, 5 μg ml−1 trimethoprim and 2·5 IU polymyxin-B) (BAB/Skirrow, Oxoid, Basingstoke, UK). Broth culture was carried out in Brucella broth (Becton Dickinson, Oxford, UK). An Innova 4230 (New Brunswick Scientific, Enfield, CT, USA) incubator was used for aerobic culture at 37°C.
Chicken juice was prepared as described previously (Birk et al. 2004). Thawing of frozen chickens was performed overnight at room temperature, and the juice was collected before filtering using a 0·2-μm sterile polyethersulfone (PES) syringe filter (Millipore, Watford, UK). Chicken juice was aliquoted and stored at −20°C until use. Immediately before use, the juice was refiltered as described. A 5% (w/v) skimmed milk powder solution was prepared by mixing commercially available skimmed milk powder (Co-operative instant dried skimmed milk powder, UK) into Brucella broth and allowed to dissolve thoroughly before filtering using a 0·2-μm sterile PES syringe filter.
Campylobacter jejuni strains were started from a single-use glycerol stock, thawed and inoculated onto BAB/Skirrow plates and grown overnight at 37°C in microaerobic conditions (5% O2, 10% CO2 and 85% N2). Brucella broth cultures were inoculated with cells from the BAB/Skirrow plate, grown overnight with shaking (37°C, microaerobic conditions) and used for biofilm experiments as previously described (Reuter et al. 2010). Briefly, cell cultures were adjusted to an A600 of 0·05, and 1 ml of this solution was added to a sterile borosilicate glass test tube (Corning). Tubes were incubated at 37°C in either microaerobic or atmospheric air conditions. Following incubation for 48 h, biofilms were stained with either crystal violet or TTC as described in the relevant methodology sections.
Cell suspensions were removed from the test tubes, and the tubes were subsequently washed once with demineralized water and dried at 60°C for 30 min. One ml of 1% w/v crystal violet solution in demineralized water was added, and tubes were further incubated on a rocker at room temperature for 30 min. After incubation, the nonbound dye was removed from the tubes by thorough washing with demineralized water, followed by drying at 37°C. Bound crystal violet was dissolved by adding 20% acetone/80% ethanol and incubating on a rocking platform for 15 min at room temperature. The absorbance levels of dissolved dye were measured at a wavelength of 590 nm using a Biomate 5 spectrophotometer (Thermo, Loughborough, UK).
2,3,5 Triphenyltetrazolium chloride staining was essentially performed as described below with minor adaptations during the method optimization process. Following the 48 h incubation, the cell suspension was removed and tubes were washed twice by adding 0·6 ml of sterile phosphate-buffered saline (PBS) pH 7·4 and swirled gently to rinse the attached biofilm, and remove unbound cells. An excess (1·2 ml) of Brucella broth, supplemented with TTC (0·01% w/v), was added to each tube before incubating at 37°C in microaerobic conditions for 24 h. The remaining Brucella broth/TTC solution was then removed and the tubes air-dried. Bound TTC dye was dissolved using 20% acetone/80% ethanol and the A500 of the solution measured. During assay optimization experiments, the following parameters were altered as follows: rigour of the biofilm washing (no wash, 1 × 0·5, 2 × 0·5, 1 × 1 and 2 × 1 ml), TTC concentration (0·01, 0·02, 0·05, 0·1 and 0·5%), Camp. jejuni cell concentration (initial A600 values of 0·01, 0·02, 0·05 and 0·1) and TTC-incubation time (6, 24, 48 and 72 h). Based on the optimization experiments described in the results section, the following method is recommended: following the 48 h incubation, the cell suspension was removed and tubes were washed twice by adding 1 ml of sterile PBS pH 7·4 and swirled gently to rinse the attached biofilm and remove unbound cells. An excess (1·2 ml) of Brucella broth, supplemented with TTC (0·05% w/v), was added to each tube before incubating at 37°C in microaerobic conditions for 48 or 72 h. A schematic representation of this method, compared to crystal violet staining, is shown in Fig. 1.
To determine the number of viable cells, a sample of the planktonic cell suspension was serially tenfold diluted in phosphate-buffered saline (PBS, pH 7·5) to 1 × 10−8, and 5 μl of each dilution plated on a Brucella agar plate and incubated at 37°C for 48 h. After 2 days of growth, CFU ml−1 was calculated. Cell viability was assessed upon initial addition of cultures into static culture, before washing and addition of TTC containing media, and following incubation with TTC containing media.
Overnight broth cultures of Camp. jejuni were grown as described above. A set of culture dilutions, ranging from A600 = 0·001–1·5 were prepared, by diluting cells in Brucella broth supplemented with 0·05% TTC. Supplemented cultures were incubated at 37°C in microaerobic conditions for 30 min. Formazan crystals were then dissolved by adding an equal volume of 20% acetone/80% ethanol and incubating at room temperature for 30 min. Cells were removed by centrifugation (16 000 g, 10 min at room temperature), and the A500 of the supernatant was measured. For comparison with A600 values, a duplicate set of dilutions was prepared, incubated for 30 min (37°C, microaerobic) and A600 measured. Up to five, independent replicates were used to calculate mean absorbance values and standard error of mean.
Statistical analysis was carried out using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA, USA). Significance was assessed by calculating Student's t-test and Bonferroni post-test probabilities, and a P-value below 0·05 was considered significant. Linear regression was used to compare the cell density measurements presented in the section entitled ‘TTC reduction in broth cultures correlates with A600’.
Crystal violet staining is a frequently used method for detection and quantification of Camp. jejuni biofilm formation (Joshua et al. 2006; Reeser et al. 2007; Reuter et al. 2010). However, we observed that chicken juice supplemented Camp. jejuni biofilms showed nonspecific, false-positive staining (Fig. 2a): tubes incubated with chicken juice but without Camp. jejuni gave high levels of staining, visualized as a heavily stained ring at the air-media interface (Fig. 2a), independent of the presence or absence of Camp. jejuni. Similar results were obtained when replacing the chicken juice with skimmed milk solution, another food-relevant matrix that has previously been used (Chmielewski and Frank 2003). Nonspecific staining was also observed with the alternative biofilm stain Congo Red (Reuter et al. 2010). Hence, we conclude that crystal violet and similar dyes cannot be used to discriminate between the bacterial biofilm and adhered components of the food matrix.
To study biofilm formation using the chicken juice model, we therefore developed an alternative biofilm staining method using the redox dye TTC to stain a preformed biofilm. Biofilms were cultivated using static culture (primary incubation), and then growth media were replaced by fresh TTC-supplemented media (secondary incubation). During this secondary incubation, cells were able to reduce the TTC, leading to the formation of a visible red formazan ring, which could be quantified by dissolving the bound formazan precipitate in a solvent and measuring the absorbance at a wavelength of 500 nm (Fig. 1). Our initial results with TTC showed that, in chicken juice containing medium, while the crystal violet method gave a clear ring of staining in both the presence and absence of Camp. jejuni, tubes stained with TTC only showed a ring of adhered material in the presence of Camp. jejuni (Fig. 2b).
Initial experiments using TTC in conjunction with chicken juice were conducted in microaerobic conditions. We previously demonstrated that Camp. jejuni forms biofilms during incubation at 37°C in aerobic conditions (Reuter et al. 2010) and therefore investigated the reduction in TTC by aerobically incubated cells, and compared this with crystal violet staining (Fig. 3) in Brucella media. Levels of TTC staining were significantly lower when compared with crystal violet following secondary incubation in air, suggesting that TTC conversion by aerobically incubated Camp. jejuni does not correlate with the total level of biofilm, as determined by crystal violet. This difference was not observed in cultures from primary incubation in air followed by secondary incubation in microaerobic conditions (Fig. 3). There was no difference in the number of viable cells in the culture supernatants following the secondary incubation in air compared with the secondary incubation in microaerobic conditions. Hence, all secondary incubations were performed in microaerobic conditions. The final level of biofilm formation was not affected by the transfer of tubes to microaerobic conditions after an aerobic incubation, as assessed by both crystal violet staining and TTC staining (Fig. 3), suggesting that the secondary microaerobic incubation, while allowing TTC conversion and biofilm quantification, does not increase the total amount of adhered cells in the biofilm.
In order to optimize the assay for use with Camp. jejuni and reduce variability, we explored the effect of several methodological variables on assay reproducibility: rigour of the biofilm washing, TTC concentration, Camp. jejuni cell concentration and finally TTC-incubation time.
To investigate the effect of the washing procedure on quantification of biofilm formation and cell loss in the TTC assay, we compared the following washing procedures prior to the TTC-incubation step: no washing, a single wash with 0·5 ml PBS, two washes with 0·5 ml PBS, a single wash with 1 ml of PBS or two washes with 1 ml PBS. None of these conditions resulted in a statistically significant difference in the A500 values obtained (data not shown), suggesting that the washing procedure does not impact the final levels of biofilm staining. We subsequently calculated and analysed the coefficient of variance (%CV) to show the levels of reproducibility between replicates. Results were more reproducible (i.e. lower% CV) using two washes with 1 ml of PBS. Therefore, on all subsequent assays, tubes were washed twice with 1 ml sterile PBS.
2,3,5 Triphenyltetrazolium chloride has been reported to be toxic to bacteria at high concentrations and has been used in selective media for Gram-negative bacteria (Weinberg 1953). Thus, we compared different concentrations of TTC to assess optimal staining while not affecting viability of Camp. jejuni during the incubation. While there was a trend towards less TTC reduction at higher concentrations, there was no statistically significant difference between any of the TTC concentrations tested (from 0·01–0·5%), demonstrating that TTC reduction occurs at all tested concentrations (Fig. 4a). However, cell viability assays of the planktonic cells showed that TTC concentrations of ≥0·1% lead to a dramatic loss of Camp. jejuni viability (Fig. 4b). To avoid artefacts resulting from such cytotoxicity, all future assays used TTC at a final concentration of 0·05%.
To test if the total number of cells used to inoculate the static cultures influenced the final levels of TTC staining, overnight Camp. jejuni cultures were diluted to initial A600 of 0·01, 0·02, 0·05 and 0·1 prior to innoculation. When assessing biofilm formation using the TTC method, no significant difference in TTC staining was seen over this range of inocula (Fig. 4c).
Following the addition of fresh TTC-supplemented media, various secondary incubation times, ranging from 6 to 72 h, were compared in order to establish the optimal secondary incubation period. Staining was significantly increased 48 h after the addition of TTC, although there was little increase from 48 to 72 h (Fig. 4d).
During the course of our assays, we observed a high level of precipitation and aggregation in cultures supplemented with chicken juice, which hampered accurate tracking of growth by measuring A600 values and also adversely affected making accurate serial dilutions of culture samples of viable cell counting. As TTC conversion is dependent on metabolic activity, we hypothesized that it can be used as an alternative to A600 measurement in media supplemented with meat exudates, such as chicken juice. In order to test this, stationary-phase Camp. jejuni cultures were diluted to a range of concentrations in Brucella media, supplemented with 0·05% TTC. Following a static 30-min incubation, both the A600 and A500 were determined (Fig. 5). At an A600 ≤ 0·8, TTC conversion (as measured by A500) showed a linear correlation with the A600 (Fig. 5). The gradients of A500 and measured A600 values did not significantly differ (P-value = 0·59). Above A600 = 0·8, A600 and A500 measurements did not correlate, with A500 values beginning to plateau (Fig. 5a). This could be due to maximal conversion of the dye, or poor recovery of the dissolved dye due to the high cell concentration. Thus, for tracking bacterial growth at A600 = 0–0·8, TTC supplementation can be used where A600 measurement is not possible due to interfering components present in food matrices.
In this study, we have shown that TTC staining of Camp. jejuni biofilms is a suitable alternative to crystal violet dye, giving similar levels of sensitivity and assay reproducibility. TTC also has the advantage of allowing metabolic activity detection and visualization of biofilms in matrices relevant to the food chain, in this case chicken juice, and by extension other meat exudates and skimmed milk. We have previously shown that viable cells could be recovered from an aerobically grown biofilm (Reuter et al. 2010). Using our optimized TTC staining methodology, we can also show that viable cells contribute to the adhered population of aerobically cultured biofilms.
Crystal violet is commonly used in biofilm studies; however, various authors have suggested that information gained from crystal violet staining alone may be misleading due to nonspecific staining (Pan et al. 2010; Skogman et al. 2012). Other techniques, such as cell enumeration, must be carried out alongside crystal violet staining, leading to increased study cost and time. Crystal violet is also not suitable for use in combination with high protein content matrices, or where nonspecific binding to components in the growth matrix is expected. Congo red also showed a high level of background binding in chicken juice models (data not shown), suggesting the presence of amyloid tissue or β1-3/β1-4 linked oligosaccharides in chicken juice. Calcofluor-white has also been reported to stain Camp. jejuni biofilms (McLennan et al. 2008); however, as this is another carbohydrate component stain, we anticipate similar issues in food matrix models. High levels of background are also expected with the protein stain Coomassie blue (Austin et al. 1998; Rogan et al. 2004).
The method presented here allows rapid, nontoxic and low-cost quantification of metabolically active attached cells. TTC has been used to detect cell viability in a wide range of tissues (Steponkus and Lanphear 1967; Adegboyega et al. 1997). A number of studies have reported the use of TTC for speciation of Campylobacter species (Skirrow and Benjamin 1980; Luechtefeld and Wang 1982), due to different sensitivity of Camp. jejuni and C. foetus to TTC. TTC has been used to assess the levels of biofilm formation and growth of Staphylococcus aureus and E. coli (Tengerdy et al. 1967; Skogman et al. 2012), but a review of the literature shows that TTC has not been used to assess Camp. jejuni biofilm formation, either in monoculture or in a food matrix model.
As TTC is a metabolic dye, growth conditions should be optimal when using the TTC stain. In our hands, suboptimal growth or stressful conditions (such as atmospheric levels of oxygen for a microaerobic organism or nutrient limitation) led to inefficient reduction in the TTC during the staining step. Rapid loss of signal, due to starvation conditions, has also been observed for 5-cyano-2,3-ditolyl tetrazolium chloride detection of Pseudomonas putida in a drinking water system (Schaule et al. 1993) and during assessment of antibiotic activity against Streptomyces venezuelae (Brooks et al. 2012). For detection of adhered cells in suboptimal conditions, conventional staining and microscopy still have a role to play in measuring biofilm formation. It needs to be stressed that although this method is suitable for quantification of adhered bacterial populations, and by extension biofilm formation, it does not allow for quantification of extracellular matrix. If additional quantification of extracellular matrix is required, then it is recommended that TTC staining is used alongside other straining techniques.
The washing procedure, performed before the secondary incubation, did not significantly affect the levels of TTC conversion. We hypothesize that a proportion of the Camp. jejuni biofilm forms tight attachments to the abiotic surface, which cannot be overcome by the mild shearing forces generated during the washing stages. Although this phenomenon has not been reported previously in Camp. jejuni, it has been observed in Pseudomonas fluorescens biofilms, where washing with a citrate buffer detergent removed <1% of the cell population from a mature biofilm (Simões et al. 2008). Similarly, strong attachment of E. coli to abiotic and biotic surfaces could not be reversed by repeated washing (Silagyi et al. 2009).
2,3,5 Triphenyltetrazolium chloride has also been suggested as a tool for allowing rapid enumeration of Campylobacter colonies in selective media (Line 2001). In this study, concentrations of 0·1, 0·5, 1, 2 and 5 mg ml−1 were tested, and our results were consistent with that of the previous authors (Veron and Chatelain 1973; Butzler and Skirrow 1979; Skirrow and Benjamin 1980; Line 2001), showing normal growth of Camp. jejuni at TTC concentrations of 0·2 and 0·4 mg ml−1, but very weak growth at 1 mg ml−1 and no growth observed above this concentration.
We have observed that the speed of TTC conversion is dependent on the culture conditions. For example, it is possible to measure the levels of Camp. jejuni in shaking cultures following a 30min incubation with TTC; however, to achieve the same signal levels with biofilm cultures, a 48 h incubation time is required. Previous studies have shown that bacteria in biofilms have reduced metabolic activity, due to gradients of nutrients and oxygen within the biofilm itself (Fux et al. 2005; Høiby et al. 2010), leading to differential conversion of TTC.
Finally, it is advisable to further optimize the TTC methodology presented here for specific applications. This phenomenon was reported by Klancnik et al. (2010), who found that Camp. jejuni and E. coli did not respond to the TTC staining protocol used for Bacillus cereus, Staphylococcus aureus and Salmonella infantis.
In conclusion, the method presented here offers an additional, complementary low-cost technique suitable for use in biofilm analysis. TTC staining is especially valuable when traditional dyes such as crystal violet cannot be used due to concerns around nonspecific staining of medium components. Moreover, we demonstrate the use of TTC conversion to monitor growth in matrices, such as meat juices and skimmed milk, which do not allow measurement of A600 values. The method allows rapid and simple imaging of metabolically active cells within a biofilm.
The authors wish to thanks members of the IFR Campylobacter research group and Gary Barker for helpful discussions and Maddy Houchen for microbiology media support. We gratefully acknowledge the support of the Biotechnology and Biological Sciences Research Council (BBSRC) via the BBSRC Institute Strategic Program (IFR/08/3 and BB/J004529/1) and a BBSRC CASE studentship (BB/I15321/1) with CASE funding from Campden BRI.
No conflict of interest declared.