Quantification of the extracellular matrix of the Listeria monocytogenes biofilms of different phylogenic lineages with optimization of culture conditions

Authors

  • T. Combrouse,

    1. Université Lille1, UGSF (Unité de Glycobiologie Structurale et Fonctionnelle), Université des Sciences et Technologies de Lille, Villeneuve d'Ascq, France
    2. CNRS UMR 8576, Université des Sciences et Technologies de Lille, Villeneuve d'Ascq, France
    3. ANSES, Laboratoire des Produits de la Pêche, Unité microbiologie, Bassin Napoléon, Boulogne-sur-Mer, France
    Search for more papers by this author
  • I. Sadovskaya,

    1. Université du Littoral-Côte d'Opale, Bassin Napoléon, Boulogne-sur-Mer, France
    Search for more papers by this author
  • C. Faille,

    1. INRA, UR638 PIHM (Interface Processes and Hygiene of Materials), Villeneuve d'Ascq, France
    Search for more papers by this author
  • O. Kol,

    1. Université Lille1, UGSF (Unité de Glycobiologie Structurale et Fonctionnelle), Université des Sciences et Technologies de Lille, Villeneuve d'Ascq, France
    Search for more papers by this author
  • Y. Guérardel,

    1. Université Lille1, UGSF (Unité de Glycobiologie Structurale et Fonctionnelle), Université des Sciences et Technologies de Lille, Villeneuve d'Ascq, France
    2. CNRS UMR 8576, Université des Sciences et Technologies de Lille, Villeneuve d'Ascq, France
    Search for more papers by this author
  • G. Midelet-Bourdin

    Corresponding author
    • ANSES, Laboratoire des Produits de la Pêche, Unité microbiologie, Bassin Napoléon, Boulogne-sur-Mer, France
    Search for more papers by this author

Correspondence

Graziella Midelet-Bourdin, ANSES, Laboratoire des Produits de la Pêche, Bassin Napoléon, F-62200 Boulogne-sur-Mer, France. E-mail: graziella.bourdin@anses.fr

Abstract

Aims

The purpose of this study was to quantify the extracellular matrix of Listeria monocytogenes biofilm. A preliminary study was carried out to establish a relationship between phylogenetic lineage of 27 strains and their ability to form biofilm in various conditions.

Methods and Results

Biofilm formation on microtitre plates of 27 strains of L. monocytogenes belonging to lineages I or II was evaluated in different conditions [two temperatures (37 and 22°C) and two media (tryptone soy broth yeast extract medium (TSBYE) and MCDB 202 defined medium)] using crystal violet assay. Lineage II strains produced significantly more biofilm than lineage I strains. In microtitre plates assay, biofilm quantities were greater in MCDB 202 vs TSBYE medium [confirmed by scanning electron microscopy (SEM) analysis] and at 37 vs 22°C. Cultivable bacteria from biofilm population on Petri dishes were enumerated in greater quantities in TSBYE than in MCDB 202 medium. The SEM investigation established that L. monocytogenes biofilms produce extracellular matrix in both media at 37°C. The amount of exopolymers in the extracellular matrix and the pH values were significantly higher in TSBYE than in MCDB 202 medium. The exception was the ScottA strain that presented similar pH values and exopolymer contents in both media. Proteins were the most abundant exopolymer components, followed by DNA and polysaccharides.

Conclusions

The interpretation of results of biofilm quantification was depending on the growth conditions, the viability of the bacteria and the analysis method. The quantities of proteins, DNA and polysaccharides were different according to the strains and the medium.

Significance and Impact of the Study

This study screened the potential of a wide panel of L. monocytogenes strains to synthesize exopolymers in biofilm growing condition. The characterization of L. monocytogenes biofilm composition may help to develop new strategies to prevent the formation of biofilms and to remove the biofilms.

Introduction

Listeria monocytogenes is an ubiquitous (Gelbíčová and Karpíšková 2012) and opportunist pathogen that causes listeriosis, a rare but deadly foodborne disease (Mead et al. 1999). Listeriosis occurs as sporadic cases or as epidemic cases with fatality rates of 25–30% (Farber and Peterkin 1991). In recent years, an increase in the number of listeriosis cases has been reported in several European countries (Goulet et al. 2008; Allerberger and Wagner 2010; Kvistholm Jensen et al. 2010; Muñoz et al. 2012).

Multiplex PCR permits accurate identification of serogroup (Doumith et al. 2004) and lineage (Nadon et al. 2001; Ward et al. 2004; Tamburro et al. 2009) of L. monocytogenes. Listeria monocytogenes may be separated into four major evolutionary lineages. Lineage I is responsible for more than 95% of human infections (Rocourt et al. 2000; Nelson et al. 2004; Swaminathan and Gerner-Smidt 2007). However, recently, some authors have observed an increase in epidemics caused by strains belonging to lineage II (Lukinmaa et al. 2003; Parihar et al. 2008; Gianfranceschi et al. 2009), which is associated with food and environment isolates. Finally, strains belonging to lineage III are rarely isolated (1%) from human listeriosis cases (Roberts et al. 2006; Tamburro et al. 2009). Lineage IV is poorly understood because this lineage is rare and was only recently described.

Listeria monocytogenes attaches to a variety of surfaces including stainless steel, polystyrene and glass (Mafu et al. 1991; Blackman and Frank 1996; Di Bonaventura et al. 2008). The existence of a direct relationship between lineage and the ability to form biofilm is still controversial. Some authors have reported that lineage II strains produced more biofilm than strains of lineage I (Norwood and Gilmour 2001; Borucki et al. 2003). In contrast, Djordjevic et al. (2002) observed an opposite trend. Such discrepancies, often found in the literature (Borucki et al. 2003; Harvey et al. 2007; Takahashi et al. 2009), may be related to differences in experimental designs and in the methods used to study biofilm. Indeed, both strains and culture conditions (medium, temperature, incubation time and surface) differed between the studies reported above. However, all agreed that L. monocytogenes biofilms are significantly influenced by temperature (Fletcher 1977; Moltz and Martin 2005; Mai and Conner 2007; Di Bonaventura et al. 2008), strain (Chae and Schraft 2000; Borucki et al. 2003; Tresse et al. 2007), incubation time (Fletcher 1977; Harvey et al. 2007), medium (Moltz and Martin 2005; Mai and Conner 2007) and nature of adhesion surface (Midelet and Carpentier 2004; Midelet et al. 2006).

The extracellular matrix surrounding bacteria in biofilm is composed of water and extracellular polymeric substances (EPS), including exopolysaccharides, proteins and nucleic acids. Exopolysaccharides are often considered to be the most abundant and important components in biofilm development (Sutherland 2001). Chae et al. (2006), using a colorimetric method, determined that the total carbohydrate (free glucose, oligosaccharides, polysaccharides) content of 21 L. monocytogenes strains ranged from 6·26 to 12·12 μg log10 colony-forming unit (CFU)−1 after 3 h of attachment on glass surface, in PBS medium. Other studies used ruthenium red staining to quantify the matrix polysaccharides (Borucki et al. 2003; Zameer et al. 2010). However, ruthenium red also binds to the bacterial cell wall carbohydrates, which may lead to an overestimation of the carbohydrate content (Borucki et al. 2003; Renier et al. 2011). Other studies have demonstrated that surface and extracellular proteins were important factors controlling L. monocytogenes biofilm formation (Smoot and Pierson 1998; Jordan et al. 2008; Longhi et al. 2008).

Nucleic acids, and more particularly extracellular DNA (eDNA), also seem to be of importance. Indeed, numerous studies have reported that eDNA is an essential component in biofilm formation and architecture in several bacteria such as Staphylococcus (Qin et al. 2007; Rice et al. 2007; Izano et al. 2008; Das et al. 2010) or Pseudomonas (Whitchurch et al. 2002; Allesen-Holm et al. 2006). For L. monocytogenes, Harmsen et al.(2010) have recently shown the important role of eDNA, both in initial adhesion and in biofilm formation steps.

To summarize, few data on the matrix composition of L. monocytogenes biofilms are available. However, it is well known that the extracellular matrix plays an important role in biofilm development, resistance to chemicals and to detachment (Sutherland 2001; Kirby et al. 2012). The purpose of the present study was to characterize the ability of L. monocytogenes strains to form biofilm, taking into account strain lineage, growth medium (poor or rich) and incubation temperature (22 or 37°C), using complementary methods. For the first time, a quantitative analysis of polysaccharide, protein and eDNA content in the biofilm matrix was carried out.

Materials and methods

Bacterial strains

Twenty-seven strains of L. monocytogenes (from environmental, animal, food, clinical and unknown sources) were used in this study (Table 1). All strains were characterized using serogroup PCR as previously described by Doumith et al. (2004), with some modifications (Midelet-Bourdin et al. 2008), and were classified by lineage (Ward et al. 2004; Tamburro et al. 2009). Cells were stored in brain–heart infusion (BHI) medium (AES, Combourg, France) supplemented with glycerol (18% v/v) at −20°C. Cultures were maintained for up to 1 month on tryptone soy agar yeast extract slopes (TSAYE) (Oxoid, Dardilly, France) at 4°C. The culture media used in this study were a rich undefined medium tryptone soy broth yeast extract (TSBYE) (Oxoid) and a defined medium the MCDB 202 (CryoBioSystem, L'Aigle, France) (Hébraud and Guzzo 2000). Refrigerated cultures were transferred into 5 ml of TSBYE at 22 or 37°C for 24 h. This suspension was transferred to 35 ml of TSBYE and then incubated at 22 or 37°C for 24 h. Concentrations of the inoculum were adjusted to 107 CFU ml−1 (OD630 nm in 1-cm diameter tubes: 0·15 = 108 CFU ml−1) in TSBYE or MCDB 202 media.

Table 1. Source, origin and lineage of the 27 strains Listeria monocytogenes studied
StrainLineageaOriginbSource
  1. a

    Lineage I correspond to serovars 1/2b, 3b, 4b, 4d and 4e and lineage II correspond to serovars 1/2c, 3c, 1/2a and 3a.

  2. b

    (–) Signify that the origin is unknown.

442IFoodMidelet-Bourdin et al. (2010)
446IFoodRoche et al. (2001), Midelet-Bourdin et al. (2010)
CIP 78·38IMidelet-Bourdin et al. (2010)
CIP 103322IOutbreakMidelet-Bourdin et al. (2006)
CS 455 S1IRaw herringMidelet-Bourdin et al. (2007)
GPL 2192ISporadicInstitut Pasteur de Lille, France
GPL 5653IMusselsInstitut Pasteur de Lille, France
HIP 506ICheeseInstitut Pasteur de Lille, France
Lm14IAugustin et al. (2011)
LmAIDairy productCarpentier and Chassaing (2004)
ScottAIOutbreakTresse et al. (2007)
TQA 200/04ISmoked salmonAugustin et al. (2011)
ADQP 105IIFishAugustin et al. (2005)
C 3978 PIISalmonMidelet-Bourdin et al. (2007)
CL 283 BS1IISmoked salmonMidelet-Bourdin et al. (2007)
CL 297IISmoked salmonMidelet-Bourdin et al. (2007), Tresse et al. (2007)
CL 86 AP1IISmoked salmonMidelet-Bourdin et al. (2007), Tresse et al. (2007)
CNL 895807IIDairy productRoche et al. (2001), Carpentier and Chassaing (2004)
CS 464 CS1IISmoked tuneMidelet-Bourdin et al. (2007)
D 3175 PT2IIHerringMidelet-Bourdin et al. (2007)
D 3188IISmoked herringMidelet-Bourdin et al. (2007)
DPF 234 HG2IIRaw herringMidelet-Bourdin et al. (2007), Tresse et al. 2007
DSS 1130 BFA2IISmoked herringMidelet-Bourdin et al. (2007), Tresse et al. 2007
DSS 863IISmoked salmonMidelet-Bourdin et al. (2006, 2007)
E 1012 PT1IISmoked salmonMidelet-Bourdin et al. (2006, 2007), Tresse et al. (2007)
EE 64IIMarinated salmonMidelet-Bourdin et al. (2006, 2007)
EGDIIAnimal outbreakMidelet-Bourdin et al. (2006)

The microtitre plate bacterial growth assay

Four wells of a 96-well polystyrene tissue-culture-treated microtitre plate (Becton Dickinson Labware, Dutscher, Brumath, France) were filled with 100 μl of inoculum per strain. Four wells of sterile medium were included as control. The cell turbidity was monitored using a microtitre plate reader (SpectrofluotimeterSafasflx-XeniusXn; SAFAS, Principality of Monaco) during 48 h at 22 or 37°C at an optical density at 590 nm and was recorded every 15 min after 5 s of agitation. The microtitre plate growth assay was performed twice for each growth temperature and medium. The average OD of the control wells was subtracted from the OD of the test wells.

Microtitre plate biofilm production assay

The microtitre plate biofilm screening assay described by O'Toole and Kolter (1998)and improved for L. monocytogenes by Djordjevic et al. (2002) and Borucki et al. (2003) was slightly modified and used to quantify biofilm production by L. monocytogenes strains. Four wells per strain were filled with 300 μl of inoculum. Four wells of sterile medium were included as control wells. Plates were covered and incubated at 22 or 37°C for 48 h. Growth media were removed, and the wells were washed three times with 300 μl of sterile distilled water to remove loosely associated bacteria. Each well was stained with 300 μl of 0·1% (v/v) crystal violet solution in water for 45 min in the dark. After staining, plates were washed with sterile distilled water three times to remove unbound crystal violet. The quantitative analysis of biofilm production was performed by adding 300 μl of 95% ethanol to detach crystal violet from the bacterial biofilm. The level (OD) of the destaining solution was measured at 590 nm, by a microtitre plate reader (Biotek Power wave XS MQX 200R; Biotek, Colmar, France). The average OD from the control wells was subtracted from the OD of the test wells. The microtitre plate biofilm assays were performed three times.

Biofilm formation assay for scanning electron microscopy

Six strains of L. monocytogenes were used for this assay: CNL 895807, DPF 234 HG2, DSS 1130 BFA2, HIP 506, Lm A and ScottA. The surface used for attachment was polystyrene slides (75 × 25 × 1·1 mm, Evergreen; Dutscher), which were placed in Petri dishes. Each polystyrene slide was covered in 20 ml of inoculum at 107 CFU ml−1 in TSBYE or MCDB 202 media. Following a 48-h incubation at 37°C, the growth medium was removed and the slides were washed three times with 5 ml of sterile distilled water. The slides were placed in Falcon tubes containing 50 ml of fixation solution (glutaraldehyde 2·5% (v/v), ruthenium red 0·05% (v/v) and cacodylate buffer 0·1 mol l−1, pH 7·0). Ruthenium red was used to improve the image resolution (Zottola 1991; Hood and Zottola 1995; Priester et al. 2007). The tubes were left at 4°C overnight. The biofilms were finally subjected to critical point drying and coated with gold–palladium for 1·5 min. Scanning electron microscopy (SEM) images were taken at the INRA-PIHM laboratory (Villeneuve d'Ascq, France), using a S-3000 Hitachi scanning electron microscope operating at 15 kV.

Quantitative analysis of extracellular matrix [extracellular polysaccharides, extracellular proteins and extracellular free deoxyribonucleic acid (DNA)]

Inocula were prepared as described above, but biofilms were grown statically in 150-mm-diameter, tissue-culture-treated, polystyrene Petri dishes (Grosseron, Saint Herblain, France) containing 60 ml of the medium. This characterization was carried out for six strains (CNL 895807, DPF 234 HG2, DSS 1130 BFA2, HIP 506, Lm A and ScottA), grown in biofilms in MCDB 202 or TSBYE at 37°C for 48 h. After a 48-h incubation period, medium was removed from Petri dishes and its pH was evaluated using pH papers with a precision of 0·2. Biofilms were washed three times with 5 ml of sterile distilled water to remove loosely associated bacteria. Biofilms were scraped off by hand, wearing a sterile glove. Biofilms were harvested with 4 ml of sterile distilled water and placed in Falcon tube. Ten Petri dishes were prepared for each strain to obtain enough material. The extracellular matrix was detached from the bacterial cells by sonication on ice (IKA Laboteckniksonicator, Staufen, Germany) for 3 × 30 s, 50% cycle, at a 0·5 intensity, as recommended for staphylococcal or pseudomonal biofilms (Sadovskaya et al. 2004, 2010). All samples were kept on ice during and between sonication steps. The cells integrity was checked by transmission electron microscopy (TEM) observation after sonication step to insure that intracellular material form lysed cells did not contribute to the analyses of extracellular matrix (data not shown) (Lequette et al. 2011). Bacteria cells were removed by centrifugation (8700 g, 10 min, 4°C). The supernatant, which contained extracellular matrix, was dialysed (dialysis bags with MW cut-off 3500 Da; Dutscher) for 48 h and lyophilized. Extracellular matrix was resuspended with 1 ml of ultrapure water. Polysaccharide concentrations were determined by phenol–sulfuric acid assay (Dubois et al. 1956) and expressed as equivalents of glucose. The protein content was measured using the Bio-Rad Protein Assay, with bovine serum albumin (Sigma, Lyon, France) as a standard. The eDNA concentration was determined by OD260 nm spectrophotometry (Eppendorf Biophotometer, Le Pecq, France) after extraction with a DNeasy blood and tissue kit (Qiagen, Courtaboeuf, France). These experiments were carried out twice, and polysaccharide and protein assays were carried out in triplicate for each experiment. In parallel to each experiment, CFUs in biofilm were counted by spiral plating (Spiral System®; Interscience, Saint-Nom-La-Bretèche, France) using an TSAYE plate. Plates were incubated at 37°C for 24 h.

Statistical analysis

All the results of destained biofilm were replicated three times and were performed with four technical replicates. A variance analysis was used to determine the significant differences between strains and between culture conditions. This was followed by a multiple comparison procedure using Tukey's grouping (Alpha level = 0·05). All calculations were performed with Statgraphics, version 5 (Sigma Plus, Paris, France), and were analysed by general linear model procedures using SAS V8.0 software (SAS Institute, Gary, NC, USA), and statistical significance was evaluated at < 0·05. Analyses were carried out for biofilm assays with media, temperature and lineage factors and for polysaccharide and protein results with media and lineage factors.

Results

The microtitre plate growth and biofilm assay

The planktonic growth was carried out on the 27 strains. Based on the suspension turbidimetry, results of E1012 PT1 strain are shown in Fig. 1. In TSBYE medium, the exponential increase in OD occurred between 5 and 10 h of incubation at 22°C and between 4 and 7 h of incubation at 37°C. In TSBYE medium, the final OD at 22°C was around 0·1 unit higher than the final OD at 37°C. On the contrary, a very small growth was observed in MCDB 202 medium at both temperatures. All strains exhibited similar growth kinetics (data not shown).

Figure 1.

Growth of E1012 PT1 strain Listeria monocytogenes in the MCDB 202 medium at 22°C (▲) at 37°C (Δ) and in the Tryptone Soy Broth Yeast Extract medium at 22°C (■) and at 37°C (□). The bars indicate the error (n = 8).

The capacities of 27 L. monocytogenes strains to form biofilm were assayed at 22 and 37°C in MCDB 202 and TSBYE media. As shown in Fig. 2, the quantity of biofilm obtained after 48 h was highly dependent on strain, medium and temperature. The OD values were generally higher at 37°C (average OD590 nm = 0·681) than at 22°C (average OD590 nm = 0·311) (P < 0·0001). Regardless of incubation temperature, the amount of biofilm was systematically higher in MCDB 202 medium than in TSBYE medium (P < 0·0001). Lastly, the amount of biofilms generated by lineage I strains (average OD590 nm = 0·402) was often lower than those of lineage II (average OD590 nm = 0·572) (P < 0·0001). However, strains CS455 S1, Lm14 and TQA 200/04 belonging to lineage I produced large amounts of biofilms when grown in MCDB 202 medium at both temperatures. ScottA strain only presented this characteristic at 37°C. On the other hand, the ADQP 105 strain produced low amount of biofilm at both growth temperatures, while D 3188 PT strain and DSS 1130 BFA2 strain produced low amount of biofilm only at 22°C and CS 464 CS1 strain at 37°C. Statistical analyses confirmed these observations. The variance analysis indicated that 70% of the variability in the OD was explained by the three parameters tested: strain (< 0·0001), temperature (< 0·0001) and medium (< 0·0001).

Figure 2.

Destained biofilm (measured at OD590 nm) of Listeria monocytogenes strains incubated at 22°C (a) or 37°C (b) for 48 h in MCDB 202 medium represented in black and in Tryptone Soy Broth Yeast Extract medium in grey. The bars indicate standard deviations.

To evaluate the possible influence of lineage on the ability to form biofilms, a second variance analysis was performed, taking into account lineage, temperature and medium. The three parameters still account for 65% of the variability, and the P values of the three parameters were lower than 0·0001.

Scanning electronic microscope

All strains exhibited similar biofilm architectures after a 48-h incubation period in both growth media (results not shown). Representative micrographs by SEM of biofilms produced by L. monocytogenes LmA on polystyrene surfaces are shown in Fig. 3. In both media, the bacterial spreads were relatively homogenous. The biofilms were composed of a dense bacterial carpet mainly composed of isolated bacteria and some cell aggregates with a three-dimensional structure. As indicated by the white arrows, an extracellular matrix was also observed, mainly on the cell surface to hold bacteria together. Biofilm population produced in MCDB 202 medium seemed important than those produced in TSBYE medium.

Figure 3.

Scanning electron microscopy images of LmA strain Listeria monocytogenes on polystyrene at 37°C in MCDB 202 medium (a) and Tryptone Soy Broth Yeast Extract medium (b). White arrows indicate the extracellular matrix.

Quantitative analyses of polysaccharides, proteins and free DNA (eDNA) in extracellular matrix and CFU count of biofilm performed in Petri dishes

Three strains of each lineage exhibiting different abilities to form biofilm were selected for further analyses. A study of the extracellular matrix of L. monocytogenes biofilms was carried out on six strains in both media at 37°C temperature, allowing the formation of a high biofilm. These strains were HIP 506, LmA and ScottA for lineage I and CNL 895807, DPF 234 HG2 and DSS 1130 BFA2 for lineage II. At 37°C, ScottA strain and DPF 234 HG2 strain formed significant biofilms, CNL 895807 strain and LmA strain were average for their lineage and DSS 1130 BFA2 strain and HIP 506 strain formed less than average biofilm.

For the six selected strains of L. monocytogenes, bacteria within the biofilm were enumerated using the plate count method (Table 2). Medium and strain effects were observed, with more adherent bacteria in biofilms produced in TSBYE than in MCDB 202 medium (average surface contaminations of 3·89 × 106 and 4·90 × 104 CFU cm², respectively), except for ScottA strain, which was not or very little affected by the growth medium.

Table 2. Polysaccharides, proteins and extracellular DNA content of the matrix of Listeria monocytogenes biofilm, pH of medium after 48-h biofilm and enumeration of bacteria in biofilm. Values are average
StrainsMediumCells number(log10 CFU cm²)aProtein content (μg cm²)bPolysaccharide content (μg cm²)bDNA content (μg cm²)apHa
  1. TSBYE, tryptone soy broth yeast extract.

  2. a

    n = 2.

  3. b

    n = 6.

CNL 895807MCDB 2025·300·1450·1450·0934·5
TSBYE7·050·6040·0860·1565·6
DPF 234 HG2MCDB 2023·540·1420·0470·0874·5
TSBYE5·990·5340·1030·0575·4
DSS 1130 BFA2MCDB 2024·230·2750·0580·1864·6
TSBYE5·870·7880·1180·1175·4
HIP 506MCDB 2023·910·5730·0910·1494·9
TSBYE7·140·8440·1260·2415·8
LmAMCDB 2024·680·4980·1100·3534·8
TSBYE6·850·9230·1630·2525·5
ScottAMCDB 2026·470·4380·0880·1225·5
TSBYE6·620·4980·0950·1025·7

The quantity of each measured extracellular component (proteins, polysaccharides and DNA) is shown in Table 2. The amount of EPS (sum of measured extracellular components) varied from strain to strain, with higher diversity in TSBYE medium (between 0·28 and 0·96 μg cm² in MCDB 202 medium, and between 0·69 and 1·34 μg cm² in TSBYE medium). Indeed, except for L. monocytogenes ScottA, the highest quantities of EPS were obtained when L. monocytogenes was grown in TSBYE medium. L. monocytogenes LmA and HIP 506 produced the highest amounts of EPS on both media. Extracellular proteins were the most abundant exopolymers within L. monocytogenes biofilms' matrix, whatever the strain (between 38 and 77%), while relatively similar amounts of carbohydrate and DNA were found (between 10 and 38% and between 8 and 37%, respectively). Moreover, extracellular proteins and polysaccharide contents were significantly higher in TSBYE medium than in MCDB 202 medium (P < 0·0001) and were significantly affected by the lineage (< 0·0001) and strain (< 0·0001) parameters. For example, biofilms from lineage I contained 0·10 μg cm² of polysaccharides in MCDB 202 medium and 0·13 μg cm² in TSBYE medium, whereas biofilms from lineage II contained 0·05 μg cm² in MCDB 202 medium compared with 0·10 μg cm² in TSBYE medium. Conversely, no significant differences in quantities of eDNA were observed between strains and between growth media. Elsewhere, no correlation could be established between the amount of each extracellular polymer and the numbers of bacteria in the biofilms or OD values.

As low pH may hinder the biofilm formation of L. monocytogenes, we then checked the pH drifts of the growth media following a 48-h biofilm formation. The initial pH of both media was 7·4, but dropped to pH 4·8 and pH 5·6 after 48-h biofilm culture of all strains in MCDB 202 medium and TSBYE medium, respectively. Again, the ScottA strain was an exception, presenting similar pH in both media. Moreover, pH values seemed to depend on lineage, with lineage II being more acidic than lineage I. A difference of one unit of pH was observed between media. In the MCDB 202 medium, which became more acidic, the EPS content was less than in TSBYE medium, except for ScottA that had similar pH values and EPS content in both media. Thus, a relationship can be seen between the final pH and the amount of EPS.

Discussion

A first set of experiments were carried out on 27 L. monocytogenes strains isolated from various environments in 96-well microtitre plates (Djordjevic et al. 2002). Indeed, it is widely accepted that biofilm production varies widely from strain to strain (Chae and Schraft 2000; Lunden et al. 2000; Norwood and Gilmour 2001; Djordjevic et al. 2002; Borucki et al. 2003). Our results confirmed these observations and showed that, under the experimental conditions used, most strains were able to produce a relatively high amount of biofilm (17 of 27 strains with OD over 1·0). In accordance with previous works (Lunden et al. 2000; Norwood and Gilmour 2001; Borucki et al. 2003), we also demonstrated that strains of lineage I often produced less biofilm than strains of lineage II. However, some other authors also showed that lineage I was a better biofilm former than was lineage II (Djordjevic et al. 2002; Takahashi et al. 2009). One could suggest that the protocol used to produce and/or to quantify biofilms in microplates is responsible for the reported discrepancies. However, from the results obtained by Borucki et al. (2003), the possible precipitation of crystal violet during the staining step would not significantly affect the final OD. Conversely, the culture conditions (for instance temperature or medium) used in the different studies might be the underlying cause for the divergence of reported data. In this work, we demonstrated that biofilm formation was strongly affected by environmental conditions. As suggested by some investigators (Chavant et al. 2002; Djordjevic et al. 2002; Borucki et al. 2003; Moltz and Martin 2005), who demonstrated that a minimal defined medium improved biofilm formation, we found that high amounts of L. monocytogenes biofilms were produced in the chemically defined MCDB 202 medium than in the rich TSBYE medium. Growth temperature also played a major role in the growth of L. monocytogenes (optimal temperature was 37°C) and in the biofilm productions. In fact, the growth of L. monocytogenes and the biofilm formation were significantly higher at 37°C than at 22°C as previously described (Djordjevic et al. 2002; Moltz and Martin 2005). From this preliminary study, a panel of six L. monocytogenes strains were selected for further analyses. The panel is composed of three strains per lineage that formed low, medium and high biofilm production at 37°C, which is the optimal temperature for biofilm formation.

Scanning electron microscopy analysis first showed that the six strains were able to adhere and to form relatively homogeneous biofilm in both media, with multilayered aggregates of bacteria and extracellular matrix on polystyrene. Similar electron microscopic observations have been described by others (Chavant et al. 2002; Borucki et al. 2003; Zameer et al. 2010), but essentially on stainless steel or PTFE. SEM biofilm population in MCDB 202 medium seemed important than biofilms produced in TSBYE medium, which was in agreement with the results of the measured OD in microtitre plates assay. Conversely, the CFU method of biofilm performed in Petri dishes revealed that the number of bacteria in the biofilm was higher in TSBYE medium than in MCDB 202 medium. Some authors have already observed some contradictions between results obtained using differing methods (Trémoulet et al. 2002; Nilsson et al. 2011). We propose that these discrepancies are the result of the presence of unculturable cells in the biofilms. Indeed, the staining step in the microplates assay allows the detection of all bacteria, dead or alive, culturable or not. Conversely, bacterial enumeration of biofilm performed in petridishes does not allow the detection of injured or dead cells. This phenomenon has been previously reported on L. monocytogenes by Nilsson et al. (2011), who showed that the number of CFU was lower than the value expected from the colorimetric assay at acidic pH. It is worth noting that, after 48-h biofilm formation, the pH value in the growth medium was significantly lower in MCDB 202 than in TSBYE medium, with similar values of pH presented by Nilsson et al. (2011). The hypothesis was that uncultivable and/or dead cells were a component of biofilm biomass, at least in MCDB 202 medium and that the presence of many unculturable cells might be attributable to the low pH of the growth medium. Moreover, differences were also observed between strains, ScottA producing the densest biofilm in both media, while DPF 234 HG2 and DSS 1130 BFA2 strains produced relatively small amounts of biofilms in both media. It should be noted that HIP 506 strain was a low biofilm producer in microplates. As was the case for the differences observed between growth media, these results suggest that the level of dead or injured cells inside biofilms depended strongly on the L. monocytogenes strain. No correlation was observed between the amount of eDNA and the number of dead or injured cells.

Besides bacterial cells, biofilms are also largely composed of an extracellular matrix, which is a complex combination of exopolysaccharides, nucleic acids and proteins and plays a structure-stabilizing and protective role for the biofilm (Sutherland 2001). Herein, we established that the EPS content was largely strain dependent. As in the case with CFU, the amount of EPS was significantly higher on TSBYE than in MCDB 202 medium, except for ScottA strain. We failed to demonstrate any clear relationship between EPS content and CFU count. Previous works have suggested a positive correlation between the ability to form biofilm and polysaccharide production (Borucki et al. 2003; Zameer et al. 2010), but in these studies, polysaccharides were quantified following staining with ruthenium red. Indeed, this stain is known to bind both EPS and cell surface carbohydrates, for example, teichoic acids (Renier et al. 2011), that would explain the correlation between the ruthenium red staining and the number of CFU reported in these works.

Within these extracellular polymers, extracellular proteins were the most abundant component for every strain in both media, follow by eDNA, while polysaccharides were the least represented component in most conditions. Although polysaccharides have often been reported to be the most abundant constituent of EPS (Kumar and Anand 1998; Branda et al. 2005), some investigators have also demonstrated that proteins constitute the major fraction of extracellular matrix(Nielsen et al. 1997; Jahn et al. 1999). It is a fact that only few data are available in the literature on L. monocytogenes EPS. Furthermore, Harmsen et al. (2010) have established that eDNAis essential for the initial attachment of L. monocytogenes and biofilm formation. A 99·9% reduction in the adhered cell population was observed after the addition of 0·01% trypsin to the attachment medium (Smoot and Pierson 1998). They suggested that proteins play a role in the initial attachment process of L. monocvtogenes. In a previous work on L. monocytogenes biofilms (Nilsson et al. 2011), the authors failed to detect any polysaccharides, although they used the same methodology as we did. It is noticeable that, due to the very low amounts of polysaccharides, we had to concentrate the exopolysaccharide samples before quantifying them. As Chae et al. (2006), we gently rinsed the biofilms to remove planktonic cells and culture medium that contain significant amounts of soluble sugars before collecting and sonicating them. Only then the cells were centrifugated once to remove planktonic bacterial cells. These washing/centrifugation steps could remove a quantity of the extracellular matrix, which was less adherent on the surface of cell bacteria. Before the concentration step, we developed a protocol with a dialysis step that ensured the total removal of soluble carbohydrates other than the high molecular weight polysaccharides, including free glucose, but also high concentrations of soluble oligosaccharides and sugar metabolites. The molecular weight cut-off of 3500 Da permits to select those molecules. Our results could not be compared with those of Chae et al. (2006) because the culture conditions were different and they worked with planktonic cell and with 3-h adherent bacteria. In these conditions, the cells were viable and were enumerated on the gelose. They presented their results in μg log10 CFU−1. In contrast, we worked on a mature 48-h biofilm, which is composed of viable cells, nonculturable cells and dead cells. We presented our results in μg cm−2 because it was not realistic to present our results in μg log10 CFU−1 (it must take into account of the viability of all cells present in the biofilm). Furthermore, we concentrated our analyses on polysaccharides, and Chae et al. (2006) quantified total carbohydrates. For all these reasons, comparison between results of these two studies is somehow difficult.

In conclusion, biofilm production was clearly affected by environmental conditions, such as medium composition and temperature. Herein, we demonstrated that the EPS extracted from the biofilm of L. monocytogenes were in majority composed of proteins and that the polysaccharide content was generally lower than the DNA content. Finally, the ability of lineage II strains to form biofilm indicates that risks from Listeria biofilm must be taken seriously in sensitive food environments. The next step is the characterization of the exopolysaccharides that are associated to virulence.

Acknowledgements

This work forms part of T. Combrouse's PhD thesis financially supported by Région Nord-Pas de Calais (France) and the ANSES (Agence Nationale de Sécurité Sanitaire). This project is part of a French research program, called ARCIR, which was supported by grants from the Région Nord-Pas de Calais and the FEDER program (Fonds Européen de Développement Régional). We are grateful to G. Leleu and C. Couvreur (ANSES Boulogne-sur-Mer) and G. Ronse (INRA in Villeneuve d'Ascq) for their technical assistance.

Ancillary