Visualization and quantification of the cellular and extracellular components of Salmonella Agona biofilms at different stages of development

Salmonella is a major food-borne pathogen able to persist in food processing environments because of its ability to form biofilms. A Salmonella enterica serotype Agona isolate from poultry (S24) was grown at 37°C in biofilms for up to 144 hours (H144) in attachment to polystyrene surfaces. Biofilm structures were examined at different stages in their development (H3, H24, H48, H72, H96 and H144) using confocal laser scanning microscopy (CLSM) in conjunction with fluorescent dyes for live cells (SYTO 9), dead cells (propidium iodide), proteins (fluorescein isothiocyanate isomer I), lipids (DiD’oil), α-polysaccharides (concanavalin A, tetramethylrhodamine conjugate), and β-polysaccharides (calcofluor white M2R). Strain S24 developed a robust biofilm at H72 (biovolume of 166,852.5 ± 13,681.8 μm3 in the observation field of 16,078.2 μm2). The largest biovolume of live cells was also detected at H72 (128,110.3 ± 4,969.1 μm3), decreasing thereafter, which was probably owing to the detachment of cells prior to a new phase of colonization. The percentage of dead cells with regard to total cells in the biofilms increased throughout the incubation, ranging from 2.3 ± 1.1% (H24) to 44.2 ± 11.0% (H144). Proteins showed the greatest biovolume among the extracellular components within the biofilms, with values ranging from 1,295.1 ± 1,294.9 μm3 (H3) to 19,186.2 ± 8,536.0 μm3 (H96). Maximum biovolume values of 15,171.9 ± 660.7 μm3 (H48), 7,055.3 ± 4,415.2 μm3 (H144), and 2,548.6 ± 1,597.5 μm3 (H72) were observed for β-polysaccharides, α-polysaccharides and lipids, respectively. A strong (P < 0.01) positive correlation was found between the total biovolume of biofilm and the biovolume of live cells, proteins and β-polysaccharides, which may serve as useful markers of biofilm formation. The present work provides new insights into the formation of S. Agona biofilms. Our findings may contribute to the designing of reliable strategies for preventing and removing these bacterial communities.

Introduction Salmonella is a globally widespread food-borne pathogen. In the European Union, this bacterium has become the most frequently reported cause of outbreaks of food-borne illness, resulting in a total of 94,530 confirmed cases of human salmonellosis in 2016, with an incidence rate of 20.4 cases per 100,000 population. Salmonella enterica serotype Agona was among the fifteen most common serotypes in human cases arising in the European Union during 2016 [1].
Salmonella has the ability to form biofilms, which have been identified as an important factor for the persistence of food-borne pathogens in food-processing environments [2,3,4]. Salmonella Agona strains have been classified as strong biofilm producers [5]. Biofilms have been shown to be the main source of the contamination of foodstuffs and have been associated with many outbreaks of food-borne illness, becoming a significant problem in the food industry [6]. Serious engineering problems are also associated with the presence of biofilms on the equipment and in the installations of food-processing facilities [5]. Additionally, there is a correlation between biofilm formation and the ability of strains to colonize and replicate within the intestines of multiple host species, including humans [7].
Biofilms are structured microbial communities enclosed in a self-produced matrix of hydrated extracellular polymeric substances (EPS) and are adhered to an inert or living surface [8]. The EPS, which is composed of a wide variety of organic materials, including alfa-polysaccharides (e.g. α-mannopyranosyl and α-glucopyranosyl residues), beta-polysaccharides (e.g. cellulose), proteins (e.g. flagella and curli fimbriae), nucleic acids and lipids [9,10], is thought to maintain the biofilm architecture and functions, holding biofilm cells together and protecting them from environmental stresses. On these lines, it has been reported that cells in biofilms are far less sensitive to sanitizing agents than their non-attached individual planktonic counterparts [11].
In view of the crucial role that the EPS matrix plays in biofilm development, resistance to chemicals and detachment [12,13], a thorough understanding of the extracellular matrix ultrastructure is critical for the rational design of strategies aimed at disrupting biofilms. This would help in preventing the numerous critical problems in terms of public health and financial losses that are associated with these structures in both industrial and clinical settings.
Confocal laser scanning microscopy (CLSM) is among the most versatile and effective approaches for understanding the architecture of biofilms and their development over time. When CLSM is combined with a range of fluorescent probes, a three-dimensional image of various cellular and extracellular biofilm constituents can be obtained together with detailed quantitative information about them [14]. Thus, the quantitative three-dimensional results recorded by CLSM are an important basis for understanding, controlling and modelling of biofilms formed by pathogenic bacteria. In previous work, the structural parameters and cellular viability of 48-hour-old Salmonella Typhimurium biofilms were studied using CLSM [2]. However, it would appear that the individual cellular and EPS matrix components of biofilms formed by Salmonella at different stages of development have not hitherto been visualized or quantitatively characterized.
Taking into account the significant impact of Salmonella biofilms in the food industry, studies require to be conducted to elucidate the dynamics of the various different components of these structures during the biofilm formation process. This would facilitate progress toward the goal of developing more effective specific methods to control these bacterial communities. Therefore, the aim of this study was to visualize and characterize quantitatively, through CLSM together with image analysis techniques, the main cellular and extracellular components of biofilms formed at 37˚C on polystyrene surfaces by a Salmonella Agona strain of food origin over the course of 144 hours. This would appear to be the first time that a quantitative analysis of live cells, dead cells, proteins, lipids, α-polysaccharides and β-polysaccharides content in the S. Agona biofilm at different stages of development has been performed.

Bacterial strain and culture conditions
A Salmonella enterica serotype Agona strain (S24) available in our laboratory, previously isolated from a chicken carcass in a local slaughterhouse, was used. The strain was preserved in tryptone soya broth (TSB, Oxoid Ltd., Hampshire, England) supplemented with 20% (v/v) glycerol at -80˚C. Prior to the experiments, the frozen cells were sub-cultured twice in TSB at 37˚C. Working cultures were kept at 4˚C on plates of tryptone soya agar (TSA, Oxoid).

Biofilm formation
S24 cultures were grown in TSB for 24 hours at 37˚C, and appropriate (two-fold) dilutions in the same culture broth were made to obtain a concentration of approximately 10 7 cfu/ml. A volume of 250 μl of this culture was added to the wells of Nunc™ MicroWell™ 96-Well Optical-Bottom Plates with Polymer Base (Thermo Fisher Scientific, New Hampshire; reference number 165305). These are of high optical quality, and have a low fluorescent background and overall flatness, which allowing high resolution imaging.
After one hour of adhesion at 37˚C, the wells were rinsed with 150 mM of NaCl in order to eliminate any non-adherent bacteria before being refilled with TSB. The plates were then incubated for 3, 24, 48, 72, 96 or 144 hours at 37˚C. After incubation, the wells were rinsed with 150 mM of NaCl.

Staining procedure
Six fluorescent dyes were used (Table 1). SYTO 9 and propidium iodide (PI) from the LIVE/ DEAD 1 BacLight TM Bacterial Viability Kit, DiIC 18 (5) oil, 1,1'-dioctadecyl-3,3,3',3'-tetramethylindodicarbocyanine perchlorate (DiD'oil) and concanavalin A, tetramethylrhodamine Five working solutions of stains were prepared in NaCl 150 mM: SYTO 9 (stock 3.34 mM in DMSO) plus PI (stock 20 mM in DMSO) at 1.0 μl/ml each (A), FITC (stock 2 mg in 100 μl absolute ethanol) at 46.6 μg/ml (B), DiD'oil (stock of 25 mg in 2.5 ml of absolute ethanol) at 79.4 μg/ml (C), ConA-TMR (10 mg in 2 ml distilled water with 16,8 mg NaHCO 3 ) at 944.8 μg/ ml (D) and CFW at 189 μl/ml (E). To avoid overlapping spectra, a volume of 250 μl of each of the five solutions was added to five different wells (A to E, respectively). Additionally, in order to calculate the structural parameters of the complete biofilm (all components simultaneously), a sixth well (F) was refilled with 250 μl of NaCl 150 mM including all stains at the above-mentioned concentrations. The microtiter plate was then incubated in the dark at 37˚C. After 25 minutes wells B, C, D, E and F were rinsed with NaCl at 150 mM and refilled with 250 μl of this saline solution.

CLSM and image analysis
Confocal laser scanning microscopy (CLSM) image acquisition was performed using a Zeiss LSM 800 Airyscan confocal laser scanning microscope with ZEN 2.3 software (Carl Zeiss, Jena, Germany). Channel mode visualization was done using the 63× (0.8 NA) objective with oil immersion.
Three stacks of horizontal plane images (512 × 512 pixels corresponding to 126.8 × 126.8 μm) with a z-step of 0.25 μm, were acquired for each well from three different randomly chosen areas. Two independent experiments were performed for each development stage. Thus, a total of 216 CLSM images were obtained (6 development stages by 2 replicates by 6 wells by 3 surfaces in each well). For image analysis, original Zeiss files (CZI format) were imported into the IMARIS 9.1 software package (Bitplane, Zurich, Switzerland) for modelling in three dimensions.
The quantitative structural parameters of the biofilms were calculated using the BioRCA 1.7 software, previously developed by some members of our Research Group using the Lazarus Integrated Development Environment (IDE). From a threshold value, each pixel is labeled as empty or marked. With this information, a three-dimensional model of the biofilm under study is generated and the structural parameters are calculated. This computer program allowed quantification of the total biofilm biomass (comprising all cellular and extracellular components of biofilms) as well as of the individual components of biofilm, represented by fluorescence emitted by SYTO 9 (from cells with intact membranes), PI (from bacteria with damaged membranes), FITC (proteins), DiD'oil (lipids), ConA-TMR (α-polysaccharides) and CFW (β-polysaccharides). Biovolume represented the amount of biofilm (μm 3 ) in the observation field of 16,078.2 μm 2 . Surface coverage (%) reflected the efficiency of substratum colonization by the populations of bacteria. Roughness provided a measure of how much the thickness of the biofilm varied and was thus an indicator of biofilm heterogeneity [15]. A roughness with a value of zero indicates a biofilm of uniform thickness, and the greater the roughness coefficient, the rougher the surface. The maximum thickness (μm) of biofilms was determined directly from the confocal stack images.

Statistical analysis
The quantitative structural parameters of the biofilms were compared for statistical significance using analysis of variance techniques (ANOVA) and Duncan's multiple range test. A correlation analysis was performed to determine the relationship between the different structural parameters and components of the biofilms. Data processing was performed using the Statistica1 8.0 software package (StatSoft Ltd., Tulsa, Oklahoma).

Structural parameters of the Salmonella Agona biofilms
The structural parameters of the S. Agona biofilms at different stages of development were evaluated by CLSM observations and digital image analysis. As expected, the lowest total biomass of biofilms was observed at H3 (6,600.7 ± 1,245.6 μm 3 in the observation field of 16,078.2 μm 2 ), as shown in Table 2. Biovolume increased strongly up to H72 (166,852.5 ± 13,681.8 μm 3 ) and decreased thereafter. Biofilm expansion during the period from H3 to H72 was multidirectional, in both horizontal (increase in surface coverage) and vertical (increase in thickness) directions. After 48 hours, biofilms covered most of the surface available. On the other hand, a thin biofilm (from 0.44 ± 0.04 μm to 2.68 ± 0.15 μm, as average) was observed up to H48, with growth in depth becoming significant after this point. The greatest thickness was reached at H72, 10.17 ± 1.89 μm being the average thickness and 32.10 ± 2.14 μm the maximum. Roughness values decreased from H3 to H72 and increased thereafter. A positive change in the roughness coefficient indicated an increase in biofilm heterogeneity.
A positive correlation (from 0.630 to 0.941; P < 0.01) was found between biovolume, surface coverage, maximum thickness and average thickness of biofilms. All of these were negatively correlated with biofilm roughness (from -0.915 to -0.609; P<0.01), as shown in Table 3. Initially (H3) only single S. Agona live cells (SYTO 9-stained) and small clusters consisting of a few live bacteria were seem by CLSM. After 24 hours of incubation, the quantity of sessile cells had increased, and irregularly shaped micro-colonies were observed. As the incubation time increased, the images of live cells changed from micro-colonies to compact structures growing up until H72. Moreover, the contact surfaces were completely covered by dense and homogeneous biofilm of live cells when cultured for 72 hours. At this point the strain exhibited the greatest biofilm thickness under these conditions (Fig 1). Thus, the largest biovolume of live cells (128,110.3 ± 4,969.1 μm 3 ) and the highest percentage of biovolume of live cells within the total biovolume of the biofilm (74.9 ± 2.5%) was seen at H72. A decline in the biovolume and percentage of live cells was observed after this point, with the biofilm becoming thinner and more wrinkled. Staining the biofilm with propidium iodide (PI) showed the presence of red aggregates of materials probably formed by a mixture of dead or damaged cells and extracellular DNA (eDNA). These increased significantly over the course of storage, from 123.2 ± 70.2 μm 3 at H3 to 28,176.5 ± 9,509.6 μm 3 at H144 (Fig 1). The percentage of dead cells (PI-stained) relative to total cells ranged from 2.3 ± 1.1% at H24 to 44.2 ± 11.0% at H144 (Fig 2).

Cellular and extracellular components within the Salmonella Agona biofilm
The individual components within the EPS matrix of S. Agona biofilms at various stages of development was also monitored. Proteins and β-polysaccharides (FITC-and CFW-stained compounds, respectively) were major components of the biofilm matrix. The biovolume of proteins in the observation field ranged from 1,295.1 ± 1,294.9 μm 3 at H3 to 19,186.2 ± 8,536.0 μm 3 at H96, which may be seen from Fig 1. A low biovolume of β-polysaccharides was observed after 3 hours of incubation (761.7 ± 256.7 μm 3 ). Quantification of this biofilm component showed a progressive accumulation of biomass during the first 48 hours, the point at which the highest biovolume was observed (15,171.9 ± 660.7 μm 3 ). On the other hand, BioRCA analysis of CLSM images revealed low biovolumes for fluorescent labelled lipids (stained with DiD'oil, these lying between 799.6 ± 26.4 μm 3 at H3 and 2,548.6 ± 1,597.5 μm 3 at H72) and α-polysaccharides (ConA-TMR; ranging between 306.8 ± 255.7 μm 3 at H3 and 7,055.3 ± 4,415.2 μm 3 at H144) (Fig 1). The biovolume of proteins and β-polysaccharides increased in a similar way to the total biovolume of the biofilms. Significant positive correlations (from 0.685 to 0.778; P < 0.01) were found between these three structural parameters. On the other hand, no significant correlations were observed between the biovolume of lipids, the biovolume of α-polysaccharides and the total biovolume of biofilm (Table 3).
The formation of micro-colonies and the EPS matrix went through three distinct developmental phases during the formation of biofilm. The early phase (H3 to H24) was characterized by the attachment of a number of bacterial cells to the polystyrene surface, forming small clusters and micro-colonies, and by the start of the formation of the EPS matrix. Both cellular and EPS biomass showed a substantial increase from H3 to H24. During this phase, similar values were observed for the biovolume of cells and of EPS (the total of the extracellular components measured), as may be observed from Fig 3A. In the second phase (H24 to H72) a compact biofilm was formed, resulting in a marked increase in its cellular biomass. During this period, the biovolume of EPS and their individual components did not show any significant changes (Fig 3A and 3C). After 72 hours of biofilm formation, a third phase saw an increase in the biovolume of dead cells, as well as a marked decrease in the biovolume of live cells (probably as a consequence of biofilm dispersal) (Fig   significantly different (P < 0.05). The images correspond to three-dimensional reconstructions obtained from confocal stack images by the IMARIS 9.1 software, virtual projections being included on the right.

Structural parameters of the Salmonella Agona biofilms
In the food industry, biofilms formed by food-borne pathogens have become a major public health issue and a great financial concern. This is because of the potential contamination of foodstuffs and the greater resistance to disinfectants of biofilm relative to planktonic forms. Indeed, biofilm-related contamination of equipment has been estimated to cause nearly 60% of outbreaks of food-borne disease [16]. A better understanding of the formation and ultrastructure of biofilms may open up avenues for the control of these structures. The work reported here would appear to be the first presentation of a numerical characterization over time of the biofilms formed on polystyrene by a S. Agona strain of food origin. In this study, Salmonella biofilms were studied at 37˚C in order to test the behaviour of this pathogen at extreme room temperatures, and because this is the optimal growth temperature for Salmonella and the human body temperature.
Salmonella biofilms quickly and easily established themselves on polystyrene surfaces by building a three-dimensional structure with multiple layers of cells. This finding is a matter for concern because plastic materials are frequently used in the installations and equipment of food-processing facilities [5]. It should be noted that certain of the results of the research being reported here show substantial standard deviations. This may be explained by taking into account the heterogeneity of the biofilms, possibly due to the presence of clusters and deeper emptier areas.
The results, as shown in Table 2, are in agreement with the findings of Bridier et al. [17], who demonstrated that S. enterica strains of different serotypes (including a S. Agona isolate) produced biofilms on hydrophobic (polystyrene) surfaces that had biovolume values of approximately 100,000 μm 3 in the observation field of 14,209 μm 2 , with a maximum thickness of 35 μm, and a roughness of 0.4. Structural parameter values in the present work are also similar to those previously observed for biofilms formed on polystyrene by a S. Typhimurium isolate from food (S175), as recorded by Capita et al. [2]. It should be noted that in both the above-mentioned papers only fluorescent dyes for staining nucleic acids were used, and the EPS matrix was not studied.

Cellular and extracellular components within the Salmonella Agona biofilm
Biovolume and percentage of live cells within the biofilm decreased after 72 hours of incubation. This decrease, which has previously been observed for several groups of food-borne pathogenic bacteria, is likely to be due to the detachment of cells prior to a new phase of colonization [18,19]. The increase in the biovolume of dead or damaged cells could also contribute to the decrease of percentage of live cells from H72.
The presence of eDNA in Salmonella biofilms is in accordance with previously published reports [20]. The origin of the eDNA composing the biofilms remains unclear but might involve cellular lysis as a consequence of the biofilm life cycle or the release of small vesicles, as has formerly been put forward [21,22]. A key role of eDNA from lysed cells in the intercellular adhesion and stability of biofilm structure has also been suggested [20].
The major part played by the EPS matrix in biofilms includes the adhesion and aggregation of cells, their cohesion, the retention of water, the sorption of organic and inorganic material, enzymatic activity, acting as a nutrient source and exchange of genetic information, the export of cell components and cell-to-cell communication [23]. Moreover, the EPS matrix increases resistance to antimicrobial agents, in comparison with that of more vulnerable planktonic cells. Thus, the control of biofilms does not necessarily require the direct killing of the bacteria in the biofilm but might be directed toward the degradation or dispersal of the EPS matrix. This would reverse the sessile mode of growth to a planktonic state, which is significantly easier to remove or delete, for example by the use of disinfectants. In such a scenario, a better understanding of the organization and development of the EPS matrix within biofilms formed by pathogenic bacteria is essential for devising effective strategies for the control and eradication of biofilms. However, there is a lack of studies focusing on individual components of the EPS matrix of Salmonella biofilms at different stages of development. In this study, a quantitative characterization of fluorescently labelled structural components of the EPS in biofilms was performed. CLSM combined with image analysis techniques allowed the determination of the predominant matrix components and their spatial distribution at different time points during the formation of biofilm.
The composition of the EPS matrix within Salmonella biofilms has been demonstrated to be highly variable, depending upon the environment and the substratum upon which they are formed [10]. In the present study, proteins and β-polysaccharides, in this order, were the predominant components inside the EPS matrix of S. Agona biofilms throughout the course of incubation, occurring at a significantly greater volume than lipids and α-polysaccharides. Other authors have also shown proteinaceous components, such as flagella, and polysaccharides to be the key components of the biofilm matrix [6,10,20,23,24,25,26]. As indicated by the positive correlations found, the greater the biovolume of biofilms, the more abundant were components stained with FITC, the dye for proteins, and with CFW, that for βpolysaccharides.
It has been suggested that proteinaceous components provide structural integrity, while polysaccharide components function in a chemically or immune-protective manner or both. It has also been put forward that structural integrity requires complex interactions between both polysaccharides and proteins [10]. According to Arciola et al. [27], polysaccharides are associated with the retention of water, sorption of organic and inorganic compounds, protection against biocides and cohesion. On these lines, Houari et al. [28] demonstrated that an increase in the proportion of polysaccharides (i.e. rendering the matrix dense) ensures better cohesion. Some authors have even suggested that the consistent predominance of polysaccharides reported in the literature for biofilms formed under different growing conditions indicates that carbohydrates may play a greater role in the structural stability of biofilms than do proteins [29].
In agreement with the study being presented here, similar biomass values for cells and for the EPS matrix have also been observed by other authors [9] in Salmonella biofilms after a few days of incubation. On the other hand, these results do not agree with most studies, which identified EPS as the main component in mature bacterial biofilms, comprising up to 90% of the biofilm mass in different microbial groups [20,23]. According to several authors [24,30], the dominance of EPS over cells might be due to the need for greater protection of the cells from a challenging environment (such as the presence of disinfectants), a protection which was unnecessary under the favourable conditions tested in our study. The variations in biofilm composition observed between different reports can also be due to the use of diverse microbial groups and environmental conditions, such as the temperature at which biofilms were grown. It has been found that the genes involved in curli fimbriae and cellulose production (two important EPS matrix components) by Salmonella enterica strains are highly induced when biofilm is formed at temperatures below 30˚C [31].

Conclusions
A better understanding of the biofilms formed by pathogenic bacteria would lead to the design of more efficient control strategies. The present work appears to be the first to present the visualization and quantitative characterization of individual components of the biofilms formed on polystyrene by a S. Agona strain of food origin during the course of 144 hours of incubation. This study demonstrates that the EPS matrix of S. Agona biofilms contains mainly proteins and β-polysaccharides, which contribute to the biofilm architecture and may be used to identify biofilm formation. The findings of this study may help in determining the physiological state of the biofilms and in identifying and attacking targets so as to prevent and remove these bacterial communities in food industry establishments.