Hydrodynamic shear stress to remove Listeria monocytogenes biofilms from stainless steel and polytetrafluoroethylene surfaces

Authors


Correspondence

Maria S. Gião, Environmental Healthcare Unit, Centre for Biological Sciences, University of Southampton, Life Sciences Building, Highfield Campus, Southampton SO17 1BJ, UK. E-mail: M.S.Giao@soton.ac.uk

Abstract

Aims

To calculate the shear stress needed to remove sessile Listeria monocytogenes cells from stainless steel (SS) and polytetrafluoroethylene (PTFE) surfaces.

Methods and Results

Listeria monocytogenes biofilms were formed on SS and PTFE surfaces. Shear stress was calculated using a radial flow chamber device and cells quantified by staining with 4′,6-diamidino-2-phenylindole. Results showed that shear stress between 24 and 144 N m−2 removed up to 98% of cells from SS surfaces. PTFE presents a very hydrophobic surface, and a significant lower removal (P < 0·05) of only 63% was achieved; moreover, on PTFE discs, detachment of L. monocytogenes biofilms was more efficient at a lower shear stress (between 8·6 and 34 N m−2).

Conclusions

Water flow is more effective in removing L. monocytogenes biofilms from SS surfaces than from PTFE materials.

Significance and Impact of the Study

This work clearly demonstrates that water flow does not have the same efficiency in removing cells from different material surfaces and shows the need to optimize cleaning and sampling procedures by considering the conditions in which cells attach to surfaces and the physicochemistry of the surfaces.

Introduction

It is generally accepted that micro-organisms have a natural ability to adhere to surfaces, forming a biofilm. By definition, a biofilm consists of one or several species of micro-organisms (including bacteria, viruses and/or fungi) embedded in a complex matrix composed of exopolymeric substances (EPS) and water (Donlan 2002). The formation of biofilms essentially occurs in four steps (transport of the micro-organisms to the surface, initial adhesion, maturation and detachment), but as they are dynamic structures, all the stages occur simultaneously during the biofilm lifetime. The adhesion of micro-organisms is determined by the equilibrium of the attraction/repulsion forces that are established by the micro-organisms and surfaces (Ellwood et al. 1982; Rutter et al. 1984; Gantzer et al. 1989; Oliveira 1992). The detachment of biofilms can result from the application of a shear stress on the biofilm (Donlan 2002).

The presence of biofilms in different systems, such as drinking water pipelines, heat exchangers, contact lenses, catheters or food industry surfaces, is of great concern, as they are responsible for economic losses and may pose a public health threat when they incorporate pathogens (McFeters et al. 1984; Characklis and Wilderer 1989; Flemming 1996; Keevil 2002; Parsek and Singh 2003). The most common pathogens found in food industry biofilms include Salmonella spp., Campylobacter spp., Bacillus spp. and Listeria monocytogenes; the latter being responsible for several outbreaks worldwide with a high rate of mortality. Listeria monocytogenes is a pathogen associated with meat and derivates, milk and dairy products and vegetables (Harvey and Gilmour 1994; Dalton et al. 1997; Vitas et al. 2004). It causes listeriosis, a disease that includes meningitis, meningoencephalitis and sepsis particularly in newborns, elderly and immunocompromised people and in pregnant women might lead to miscarriage (Schuchat et al. 1991). The contamination of ready-to-eat food is normally due to contact with contaminated surfaces, as this pathogen has been isolated from several surfaces in the food industry, including floors, working surfaces, cutting boards, walls, gaskets and conveyer belts (Beresford et al. 2001; Vitas et al. 2004). Disinfection and removal of food industry biofilms are therefore mandatory for the efficient control of food safety, especially because it has been shown that sessile cells (cells in biofilms) are much more resistant to disinfectants and biocides than in the planktonic phase (Keevil et al. 1990; de Beer et al. 1994; Costerton et al. 1995). The calculation of the shear forces needed to remove biofilms from different surfaces can be of great help in the selection of the appropriate methods to effectively remove cells from food industry surfaces, whether it is for cleaning or sampling purposes. Shear force is the tangential component of the force that a fluid exerts on a surface. If the fluid is stagnant, that is, there is no flow, the shear force is zero. Shear stress is the shear force per unit of area (Çengel and Cimbala 2006).

The aim of this work was to measure the shear stress necessary to detach L. monocytogenes biofilms of different ages (0·5, 4 and 24 h) from stainless steel (SS) and polytetrafluoroethylene (PTFE) surfaces following exposure in tap water or growth in brain heart infusion (BHI) broth. For that, a radial flow cell (RFC) was used (Fowler and McKay 1980) and detachment was promoted by the passage of laminar and turbulent flow for 1, 5 and 10 min.

Material and methods

Culture maintenance

In this work, L. monocytogenes NCTC 13372 serotype 1/2c (clinical isolate) was used to study the detachment of biofilm from SS and PTFE surfaces. The culture was maintained in protective vials at −80°C and recovered onto BHI agar (Oxoid, Basingstoke, UK) prior to each experiment and incubated at 37°C for 48 h. Cultures were subcultured once and incubated at 37°C for 24 h before beginning the experiment.

RFC disc preparation

Five-cm-diameter SS AISI 316 (Brighton Systems, Newhaven, UK) and PTFE (Brighton Systems) discs were used as substrata to grow biofilms. The SS coupons were polished with 0·05-μm alumina (Buehler, Coventry, UK) and a leather cloth (Buehler) to remove residues of biofilm, organic compounds and/or oxides. The discs were rinsed with distilled water, scrubbed with a leather cloth soaked in distilled water to remove any alumina particles remaining and rinsed again with distilled water. The PTFE discs were cleaned by soaking them for 5 min in warm water with detergent (Guard Professional, UK), washed with a bottle brush, rinsed three times in distilled water and left to air-dry. Both SS and PTFE discs were then wrapped in aluminium foil and autoclaved at 121°C and 1 atm for 15 min.

Biofilm formation on SS and PTFE discs in BHI broth

Listeria monocytogenes from a 24-h culture was suspended in BHI broth to give a final concentration of c. 107 cells ml−1. Ten millilitres of the bacterial suspension was added to a 55-mm-diameter Petri dish (Fisher, Loughborough, UK) containing one SS or PTFE disc and incubated at 22°C. After 0·5 and 4 h, three discs, and after 24 h, four discs, were removed from the bacterial suspension and gently washed with distilled water to remove loosely attached planktonic cells. One disc was placed in a Petri dish, stained with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma, Gillingham, UK) and observed in situ under a Nikon Eclipse E800 episcopic differential interference contrast/epifluorescence (EDIC/EF) microscope (Best Scientific, Wroughton, UK), using a DAPI filter (Keevil 2003). The centre of the disc was localized by the use of the EDIC illumination as it was well distinguished. From this point, precise radial distance points (Table 1) were selected and eight fields of 0·01 mm2 each and distant of 45° from each other were observed to quantify cells. To avoid biofilm dehydration, the other discs were immediately transferred to the RFC to promote biofilm detachment as described below. The experiments were repeated at least three times.

Table 1. Radial distance to the centre of the disc where the number of cells was quantified and correspondent shear stress for laminar and turbulent flow calculated by the application of the equation described in the 'Material and methods' section
Radius (mm)0·51·02·03·05·07·08·09·01015
Shear stress (N m−2)
Laminar flow34178·65·73·42·42·11·91·71·1
Turbu-lent flow14472362414109·08·07·24·8

To verify whether the forces required to detach hydrated and dehydrated biofilms from SS surfaces were different, the experiment described previously was repeated but biofilms were left to air-dry for 30 min before being placed in the RFC. The experiments were repeated five times. All the experiments for hydrated and dehydrated biofilms formed on SS discs and biofilms formed on PTFE discs were performed at different times.

Biofilm formation on SS discs in tap water

Listeria monocytogenes from a 24-h-old culture was suspended in phosphate saline buffer pH 7 (PBS) (Oxoid) and washed three times by centrifugation (Heraeus, Southend-on-Sea, UK) at 855 g for 1 min, to remove residues of BHI that could be attached to the cells. The cells were then suspended in filter-sterilized dechlorinated tap water to a final concentration of c. 107 cells ml−1. Biofilms were formed as described above.

Biofilm detachment

To calculate the shear stress needed to detach L. monocytogenes biofilms, a RFC device (Brighton Systems) was used (Fowler and McKay 1980). In Fig. 1, a diagram of the RFC is shown with the respective dimensions, as well as two photographs of the top, where it is possible to observe the interior of the RFC and bottom, where it is possible to see that the inlet flow exits from the centre of the disc; an inserted SS is shown. The inlet flow runs across the disc with a decrease in the shear stress on the disc, which is related to the radial position (Eqn (2)). The discs covered with biofilm were placed in the RFC and a water flow applied at 1·62 × 10−6 m3 s−1 or 6·80 × 10−6 m3 s−1 corresponding to a Reynolds number (Re) of 1028 (laminar flow) and 4333 (turbulent flow), respectively. Laminar flow is defined as a highly ordered flow that runs in parallel layers, while turbulent flow is highly disorganized and characterized as chaotic or zigzag motion. In pipes, a flow is considered laminar when Re is lower than 2300 and turbulent when higher than 4000 (Çengel and Cimbala 2006). The Re is calculated using Eqn (1).

display math(1)

where Q is the volumetric flow (m3 s−1), ρ is the volumetric weight of the fluid (103 kg m−3, for water at 20°C), μ is the kinetic viscosity of the fluid (10−3 Pa s, for water at 20°C) and D is the hydraulic diameter, in this case, the diameter of the pipe where the flow exits in the RFC to the disc (2 × 10−3 m).

Figure 1.

Diagram of the radial flow cell (RFC) (a). Photograph of the RFC from the top with the cell open and with no disc (b) and from the bottom with a stainless steel disc inside (c).

The time of exposure of the discs to the water flow was 1 and 5 min for 0·5- and 4-h-old biofilms and 1, 5 and 10 min for biofilms grown for 24 h. Dehydrated biofilms were only exposed to the turbulent flow.

After the application of the flow, the discs were transferred to a Petri dish, stained with DAPI and observed under the EDIC/EF microscope, as described previously. For each radial distance, the shear stress, τ (N m−2), was calculated using the following equation:

display math(2)

where Q is the volumetric flow (m3 s−1), μ is the kinetic viscosity of the fluid (10−3 Pa s, for water at 20°C), r is the radial distance to the centre of the disc (m) and h is the gap between the fluid exit and the disc (3 × 10−4 m, for this RFC).

Roughness measurement

The average roughness, Ra, of the SS and the PTFE surfaces was measured using a Talysurf 120L profilometer (Taylor-Hobson Lda., Leicester, UK).

Contact angle measurement

To evaluate the hydrophobicity of the SS and PTFE surfaces, the free energy, ΔG, of the surfaces was calculated. A surface is hydrophilic if ΔG > 0 and hydrophobic if ΔG < 0 (van Oss and Giese 1995). For that, the contact angles θ formed between distilled water, formamide (Fisher) and α-bromonaphthalene (Aldrich, Gillingham, UK) and the surfaces were measured, using a Kruss DSA30 Contact Angle Goniometer (Krüss GmbH, Hamburg, Germany). The contact angle was calculated using Drop Shape Analysis software with the tangent fit method. The value of the superficial tension components (γ+, γ and γLW) of the liquids was obtained from the literature (van Oss 2006), and the superficial tension components of the surfaces were calculated using the equations described by van Oss et al. (1988; van Oss and Giese 1995).

Statistical analysis

The homogeneity of the variance of total cell number with the radial distance and flow time was checked by the Levene test for equality of variances using a statistical package (SPSS Inc., Chicago, IL, USA). The results obtained for cell numbers were transformed in logarithmic numbers, and for the percentage of removal, the decimal values were used (e.g. 0·98 instead of 98%). Results were subsequently compared by a one-way anova using the Bonferroni post hoc test. The Bonferroni test establishes that differences are statistically significant if P < 0·05.

Results

This work investigated the shear stress necessary to remove L. monocytogenes biofilms of different ages on surfaces with different physicochemistries, SS and PTFE, using a RFC device. Discs were visualized directly by microscopy after the biofilms were stained by DAPI and cells quantified. It was observed that in the discs where no flow had been applied (control), the concentration of cells was practically uniform across the surface (P = 1·000), while on discs to which a water flow had been applied, there was, in general, a lower concentration of cells closer to the disc centre. In Fig. 2, the different radial positions are shown of a SS disc covered with a 24-h-old biofilm formed in BHI and after the application of a turbulent water flow for 10 min. In this figure, it can be clearly observed that the increase in cell concentration corresponds with the increase in the distance to the centre of the disc. On the SS control discs, the concentration of cells were, on average, 1·57 (±0·88) × 103 cells mm−2, 5·55 (±4·46) × 103 cells mm−2 and 2·11 (±1·98) × 104 cells mm−2, for biofilms grown during 0·5, 4 and 24 h, respectively. For biofilms grown on PTFE control discs, those numbers were 1·41 (±0·22) × 103 cells mm−2, 1·77 (±0·48) × 103 cells mm−2 and 5·32 (±0·76) × 103 cells mm−2, respectively.

Figure 2.

Microphotographs of a stainless steel disc covered with Listeria monocytogenes biofilm formed in brain heart infusion medium at a radial distance of 0·5 (a), 3 (b), 5 (c) and 15 mm (d) from the disc centre, after being exposed to a turbulent flow for 10 min. The cells were stained with 4′,6-diamidino-2-phenylindole (DAPI) and visualized under EDIC/EF microscopy using a DAPI filter and a 100 times magnification lens.

Results showed that for experiments performed on SS surfaces in BHI medium for 0·5 and 4 h, the removal of biofilm was not statistically significant (P > 0·05) after the application of laminar or turbulent flow (results not shown). Figure 3a shows the variation in cell numbers with the radial distance of discs covered by a 24-h biofilm without and after the application of a laminar water flow. It was observed that, after the application of a laminar flow for 5 and 10 min, there was a significant decrease (P < 0·05) in the number of cells, in particular for radial distances <3 mm from the disc centre. Figure 3b describes the results obtained after the passage of a turbulent flow over SS discs covered with a 24-h-old biofilm. It was observed that the application of a turbulent flow for 1, 5 and 10 min promoted a significant removal of biofilm (P < 0·05), again for radial distances of 3 mm or less to the disc centre. Comparing Fig. 3a,b, that is, the results obtained for laminar and turbulent flow, it is possible to observe a higher decrease in the number of cells when the turbulent flow was applied.

Figure 3.

Number of sessile cells remaining at a radial distance r of a stainless steel disc where a Listeria monocytogenes biofilm has grown for 24 h in brain heart infusion medium after the passage of a laminar flow (a) and a turbulent flow rate (b) for 0 (image_n/jam12032-gra-0001.png), 1 (image_n/jam12032-gra-0002.png), 5 (image_n/jam12032-gra-0003.png) and 10 min (image_n/jam12032-gra-0004.png). Standard deviation was always lower than 10% of the average number (n = 3).

In general, for biofilms grown on SS surfaces in tap water and biofilms dried prior to the application of the flow rate (results not shown), the detachment of L. monocytogenes cells was very low (P > 0·05).

For all the cases tested, it was observed that the detachment of cells from SS discs occurred mainly for radial distances smaller than 3 mm, that is, the shear stress needed to remove a considerable amount of biofilm varied between 34 and 5·7 N m−2 and between 144 and 24 N m−2, corresponding to a shear stress originated from the application of a laminar and turbulent flow rate, respectively.

The results obtained for PTFE surfaces were surprising because a higher detachment occurred at lower shear forces (data not shown). In fact, the application of a turbulent water flow did not promote a significant removal of biofilm (P > 0·05), regardless of the age of the biofilm. In contrast, the application of a laminar water flow to 24-h-old biofilms resulted in a significant decrease in cells attached to the surface (P < 0·05), in particular for radial distances <2 mm. Contrary to SS surfaces, where the detachment of L. monocytogenes decreased with radial distance, for PTFE surfaces, the removal of cells was highly variable across the disc. The Ra values obtained for the SS and PTFE discs were 0·73 and 2·25 μm, respectively, and therefore, because PTFE was rougher than the SS surface, a similar or higher detachment was expected. As this was not observed and to attempt to explain this, the hydrophobicity of the surfaces was determined. The results presented in Table 2 showed that while SS discs were hydrophilic (ΔG = 33·6 mJ m−2), PTFE surfaces were highly hydrophobic (ΔG = −96·5 mJ m−2), which can affect the detachment of biofilms due to slippage of the water flow.

Table 2. Values for the contact angles (average numbers and respective standard deviation) formed by each surface and distilled water (θw), formamide (θf) and a α-bromonaphthalene (θb) and respective values of superficial tension components (γ) and free energy (ΔG)
DiscContact angles (degrees) average/standard deviationSuperficial tension (mJ m−2)ΔG (mJ m−2)
θwθfθbγ+γγLW
  1. PTFE, polytetrafluoroethylene; SS, stainless steel.

SS19·4/7·422·2/6·412·4/2·70·3453·8643·1833·6
PTFE113·8/8·288·7/2·772·3/0·770·020·0218·87−96·5

It was also observed that, in several cases, after the application of a flow rate for a longer time, there were more cells on the disc surface than when no flow has been applied or the flow has been applied for less time. This could have been due to variations in the initial biofilm but makes direct comparison difficult between different test conditions. Therefore, the percentage of removal of biofilm was determined using the total concentration of sessile cells calculated at 15 mm from the disc centre, as data have shown that the removal of biofilm at this distance was insignificant. The results for the maximum percentage achieved for each condition are shown in Fig. 4.

Figure 4.

Maximum percentage removal of Listeria monocytogenes biofilms grown on stainless steel discs in brain heart infusion (BHI) medium (a), in tap water (b) and in BHI but dehydrating the biofilm prior to flow passage (c) and on polytetrafluoroethylene discs is BHI medium (d). Bright colour bars correspond to biofilm removal by the application of a laminar flow, and dark colour bars correspond to biofilm removal by the application of a turbulent flow.

Results showed that, for hydrated biofilms formed on SS surfaces (Fig. 4a,b), the percentage of removal was higher when a turbulent flow was applied compared to a laminar flow, although the difference was not statistically significant (P > 0·05). It was also observed that, in some cases, there were some fields with no cells. Conversely, the removal of biofilm never reached 100%, the highest removal obtained being 98%, for 24-h-old biofilms grown in BHI and submitted to 10 min of turbulent water flow. The application of laminar and turbulent flow also revealed different tendencies. As such, the application of a laminar water flow to hydrated biofilms on SS discs resulted in a slightly increased removal with flow time (P > 0·05), while the application of a turbulent flow resulted in a similar percentage of removal for the different times studied (P = 1·00). The exception happened for 24-h-old biofilms grown on tap water where the passage of both laminar and turbulent flow for longer periods resulted in a decreased removal, although not statistically significant (P > 0·05). Comparing the percentages of removal obtained for biofilms grown in BHI and in tap water, it was verified that the detachment of biofilms is slightly superior when biofilms grew in rich media (P > 0·05).

Considering the results obtained for hydrated biofilms, where higher removal was always obtained for turbulent flow, it was decided that a water flow only in a turbulent regime would be applied for the dehydrated biofilm studies. Figure 4c shows the results obtained, where it is possible to observe that for 0·5- and 4-h-old biofilms, the application of a water flow for 5 min led to major detachment compared to 1-min flow, while for 24-h-old biofilms, the time of flow does not influence the percentage of cell detachment. This differences are not, however, statistically significant (P > 0·05). It was also observed that the percentage of removal was slightly lower for biofilms that were previously dehydrated than for hydrated biofilms (P > 0·05).

For biofilms grown on PFTE surfaces (Fig. 4d), the percentage of removal was in general higher after the application of a laminar water flow than for turbulent flow. It was also observed that when a laminar flow was applied, the increase in flow time led to a small decrease in the detachment (P > 0·05). Comparing the results obtained for biofilms formed on SS surfaces in BHI medium and results from PTFE experiments, it was verified that the passage of a water flow is more effective in detaching cells from SS surfaces than from PTFE, where removal was never higher than 63%, significantly lower than the 98% achieved for SS surfaces (P < 0·05).

Discussion

Listeria monocytogenes is a foodborne pathogen that can cross-contaminate ready-to-eat food stuffs. The effective cleaning of food industry contact surfaces is therefore fundamental to the safety of food products. On the other hand, the efficiency of sampling methods is also essential to ensure the safety of the working surfaces, because in many countries there is zero tolerance for the presence of this pathogen on food industry surfaces, in particular industries that prepare ready-to-eat products. Both have in common the need to detach L. monocytogenes cells that have adhered to industrial surfaces, in many cases forming biofilms, which are more resilient to detachment and application of sanitizers. The use of the RFC device to study the detachment forces needed to remove adhered cells was developed by Fowler and McKay (1980) and has been used by other authors to study the attachment and/or detachment of different types of cells, from human to bacterial cells (Yung et al. 1996; Tegoulia and Cooper 2002; Perni et al. 2007). The device allows the application of a mathematical model to calculate the shear force that acts on a surface area (shear stress) and correlate this with the number of cells attached or detached. This study aimed to calculate the shear stress required to detach sessile L. monocytogenes cells attached to surfaces of different physicochemistries, SS and PTFE. Results showed that, in the majority of the cases, the water flow, either applied in a laminar or turbulent regime, can detach biofilms from both surfaces. However, for biofilms formed on SS surfaces, it was observed that the maximum detachment was achieved at different radial positions, between 0·5 and 3 mm, that is, at different shear stresses. This can be due to small variations in the initial biofilm concentration, which can affect the number of cells removed, the topography of the disc surface or even small variations in flux due to the presence of cells. It was also observed that there was a tendency for a higher removal of biofilm up until >3 mm away from the disc centre, which means that when water was applied under a laminar regime, the detachment occurred at a shear stress between 34 and 5·7 N m−2, while when a turbulent flow was applied, the shear stress needed to detach cells varied between 144 and 24 N m−2. The maximum percentage of removal was higher when a turbulent flow was applied as expected, because the hydrodynamic forces generated are also higher. However, it was observed that, for the same shear stress, the removal was higher when the flow regime was laminar. As the value of shear stress is dependent on the flow velocity, the same shear stress occurred at different radial positions according to the flow applied, that is, closer to the disc centre (flow exit) for laminar flow and further for turbulent. As the flow spreads through the disc, the flow can be subject to other variables, such as variations in roughness, presence of cells or other particles that can interfere with the flow. On the other hand, while laminar flow runs straight across the surface, turbulent flow runs in random zigzags (Çengel and Cimbala 2006), which can cause cell detachment from one location but with reattachment at other locations. The direction of the laminar flow as a straight line and a turbulent flow as vortexes in the RFC was experimentally confirmed by observing the water prestained with a red dye.

For biofilms grown on SS surface in BHI medium, it was also observed that, for the laminar flow, increasing the time of flow increased the removal of cells, while for turbulent regimes, time did not affect the concentration of cells that detached. This shows that, for a higher shear stress, the application of a hydrodynamic force for longer is not able to break the forces established between the cells and the surface. Instead to achieve a higher removal, the shear stress needed to be increased. In contrast, when biofilms grew in tap water for 0·5 and 4 h, the increased time of flow increased the detachment of biofilm, for both laminar and turbulent regimes. When bacterial cells come in contact with surfaces, different types of forces are established between the substratum and the cell walls. These forces include attractive and repulsive forces, and the sum of them will determine whether the cell adheres or is repelled by the material (Ellwood et al. 1982; Rutter et al. 1984; Gantzer et al. 1989; Oliveira 1992). The medium in which the biofilm is formed can influence the adherence by changing the superficial characteristics of the substratum and the cell wall with implications for the forces that are established between them. This can explain why when biofilms grew in tap water, the percentage of removal was lower, and therefore, the time of turbulent flow influenced the detachment of cells. Biofilms formed in different conditions can also be structurally different (Moltz and Martin 2005), which can also affect cell detachment. For 24-h-old biofilms, increasing the time of flow decreased the percentage of removal, possibly due to the thickness of the biofilm, because more cells are detached, the flow is not able to carry all of them outside of the RFC, and therefore, they can reattach, possibly travelling back to radial positions closer to the disc centre. When biofilms are dried before being placed on the RFC device, it was also observed that the removal was lower compared to hydrated biofilms grown in BHI. Moreover, for 0·5- and 4-h-old biofilms, the increase in flow time promotes detachment, while for 24-h-old biofilms, time has no effect. Biofilms are composed of 90% water so it is possible that, when biofilms dry out, they become more compact, making the penetration of water inside of the biofilm and between the cells and surface difficult (Melo 2005). As a result, there is less disruption of the biofilm and lower detachment of biofilm from surfaces. As the water flow passes over, the biofilm is rehydrated and the flux of water can then remove cells. These results also suggest that the dehydration of biofilms strengthens the forces established between the cells and the surface, because the loss of water stresses cells and changes the biofilm structure, but more work regarding the adhesion forces of hydrated and dehydrated biofilms is required. It would also be interesting to use stains that can identify the biofilm matrix and distinguish viable from dead cells in both dehydrated and hydrated biofilms and study the correlation between shear stress values and biofilm physiology. In general, the results obtained for the SS discs reveal that the shear forces needed to detach L. monocytogenes biofilms can be affected by different parameters and that total removal is not possible. The complete removal of biofilms from surfaces is very difficult and has been observed by other authors (Gibson et al. 1999; Perni et al. 2007). These results can be very helpful when designing cleaning procedures or sampling methods as it shows that pressure and proximity of the jet flow are more important than time. The results presented in this work are in accordance with the results published by Perni et al. (2007), which showed that L. monocytogenes cells adhered to SS can be removed by the application of a shear stress between 52 and 101 Pa (1 Pa = 1 N m−2). However, in their work, the authors aimed to compare strains isolated from several sources, while in the present work, different conditions of biofilm formation and detachment were compared, including substrata, application of laminar and turbulent flow, as well as the time of flow application. Detachment of L. monocytogenes biofilms on PTFE as shown here required much lower shear stress.

The results obtained for PTFE substratum were completely unexpected. Firstly, the percentage removal was in general higher under laminar flow than under turbulent flow. Secondly, it was observed that detachment was very variable across the discs, that is, the percentage did not decrease with distance to the disc centre but fluctuated within the radial distance, suggesting that lower shear stress can remove more biofilm than high shear. The detachment of cells from surfaces is normally a function of the shear stress applied but can also be related to the friction factor of the flow on the surface. It is known that surface roughness can influence the friction factor, being higher for rougher surfaces. This factor decreases with flow velocity; however, it is in general higher for turbulent flows near the transition than for laminar flows (Çengel and Cimbala 2006). A higher friction coefficient would result in greater removal; however in this work, when PTFE surfaces were studied, opposite results were obtained, indicating that other factors can be influencing the removal of biofilm. It has been shown by Elkhyat et al. (2004) that hydrophobicity can change the friction forces between the flow and surface, and it has also been demonstrated by Tretheway and Meinhart (2002) that, when water flows through hydrophobic surfaces, a slippage of the fluid is promoted creating a gap on the surface where the velocity of the fluid is considerably lower than in the bulk. For these reasons, the hydrophobicity of the SS and PTFE discs used in this work was calculated. Results demonstrated that SS discs were hydrophilic, while PTFE discs were highly hydrophobic, corroborating the hypothesis that, when biofilms are formed on hydrophobic surfaces, there is an apparent decrease in the impact that water has on the bacterial cell removal. The results obtained showed that laminar flows are more effective in removing cells from hydrophobic surfaces, possibly because at lower velocities the gap between flow and surface is lower, because the friction force should theoretically be higher for the turbulent flow rate applied here (Çengel and Cimbala 2006). The hydrophobicity and roughness of a surface have an important role on microbiological adhesion (Tegoulia and Cooper 2002; Tang et al. 2009), and it has been shown by other authors that there is no correlation between material hydrophobicity or roughness and attachment of some L. monocytogenes strains (Sinde and Carballo 2000; Silva et al. 2008); however, this work clearly shows that hydrophobicity has a great influence on biofilm detachment. These results suggest that the determination of the shear stress needed to remove biofilms from hydrophobic materials is not predictable and shows that it is important to study the detachment of biofilm from each material separately. It is also important to recognize that pure biofilms are not common in nature or in industrial environments, and therefore, it would be very interesting to repeat this work using mixed biofilms, in particular with micro-organisms commonly found on food industries' surfaces.

The cross-contamination of food products can be the origin of L. monocytogenes outbreaks, and although listeriosis is not a common infection, it has a high mortality rate, c. 30% of the diagnosed cases (Todd and Notermans 2011). The efficient disinfection of working surfaces where food is handled and the efficient sampling procedure to determine cleanliness are fundamental to ensure the safety of ready-to-eat products. Therefore, the complete detachment of cells from food industry surfaces is the key for the success of safe ready-to-eat food. This work clearly demonstrates that the comprehensive removal of L. monocytogenes biofilms is not easy to achieve and the determination of the conditions to be applied to remove cells should considere parameters such as surface physicochemistry (in particular, hydrophobicity), humidity and availability of nutrients during cell attachment. This work raises new concerns about the application of the same method to clean or sample from different material surfaces. The failure of a complete detachment and resistance of surviving biofilm to sanitizers can lead to cross-contamination of food by L. monocytogenes resulting in infections or can lead to false-negative results that can give a false sense of safety, conversely leading to cross-contamination of products and consequent listeriosis.

Acknowledgements

The authors would like to acknowledge Dr Sandra Wilks, from the Centre for Biological Sciences – University of Southampton, for the writing support and Dr Barbara Cortese and Prof. Hywel Morgan, from the School of Electronics and Computer Science, for lending the equipment to measure the contact angles of the surfaces as well as the support during those experiments. Acknowledgements also go to Dr Terry Harvey from School of Engineering Sciences – University of Southampton, for the equipment and support on the experiments to determine the roughness of the surfaces. This work was supported by the European Commission within the Seventh Framework Programme, ‘BioliSME – Speedy system for sampling and detecting Listeria monocytogenes in agri-food and related European industries', no. FP7-SME-232037. Disclaimer stating that the author is solely responsible for the work, it does not represent the opinion of the Community and the Community is not responsible for any use that might be made of data appearing there in.

Ancillary