Impact de la force ionique sur le transport d’E. coli analysé par imagerie confocale

Strain, plasmid and fluorescent labeling

The bacterial strain used in this study was Escherichia coli K12 MG1655 ompR234 (Vidal et al. 1998) (PHL1314) and was kindly given by P. Lejeune (French Institute of Applied Science, Lyon, FRANCE). This strain was selected for its astonishing mobility capacities in porous media as described previously (Jacobs et al. 2007). PHL1314 transport behavior was presumably due to the hydrophilic negatively charged outer cell surface characteristics (Jacobs et al. 2007). The E. coli strain used in this study was genetically fluorescently labeled using the pDsred-express plasmid (Clonetech, USA). Cells were transformed using the chemical TSS method (Transformation Storage Solution: PEG3350 10%, MgCl2 10mM, MgSO4 10mM, DMSO 5%). The pDsredexpress plasmid inserted is carrying the dsred gene witch encodes for the red fluorescent protein from Discosoma sp. reef coral. Dsred maximum excitation wavelength is 556nm and maximum emission is 586nm. The bacterial cells were fluorescently labeled in order to perform cytometry counting and in situ confocal imaging. The pDsred-express plasmid also provides resistance to ampicillin antibiotic. Prior experiments, bacteria were cultivated overnight in 100ml of half diluted Luria Bertani media at 30°C, 100rpm. 100mg.L-1 of ampicillin was added to ensure the selection of cells carrying the pDsredexpress plasmid. 2.2. The electrostatic free energy of interaction. To be able to quantify the influence of ionic strength on bacterial interaction with the porous media the electrostatic free energy of interaction from the classical DLVO theory was calculated. The DLVO theory also includes Lifshitz-van der Waals interactions and has been calculated in our previous study (Jacobs et al. 2007). Although a raw approximation, for practical reasons we consider porous media grains as flat plates and the bacteria as spherical colloids since porous media grains are at least 100 times greater than bacterial cells. The electrostatic interaction free energy as a function of distance y between a spherical bacterium (1), and a flat plate (2), immersed in water (3)  can be calculated from (Chen et Strevett 2002). ζ potentials (V) of bacteria were determined by micro-electrophoresis using a Zetaphoremeter model (II) (CAD Instrumentation, Limours, France). The electrophoretic mobility was measured with the bacteria collected after overnight culture, suspended in either 10-1M, 10-2M, 10-3M and 10-4M NaCl solutions at pH7. Bacterial cells were placed in a measurement cell under a microscope equipped with a CCD camera. Their displacements in response to an applied electric field of 8V.cm-1 were recorded and the velocity of individual cells was calculated. The ζ potentials were calculated with the conventional Smoluchowski theory for each cell. The displacements of about one hundred cells were recorded at each measurement. The experiment was repeated three times from separate cultures for every ionic strength. The electro kinetic properties of the sand and the glass beads were measured by a streaming potential analyzer Zetacad (CAD Instrumentation, Limours, France) in a either 10-1M, 10-2M, 10-3M and 10-4M NaCl solutions at pH7. The zeta potential was calculated from the streaming potential experiments as described by Elimelech et al. (Elimelech et al. 2000). 

Column experiments

The experimental setup consisted in Plexiglas columns (Ø = 50mm, length = 150mm) filled either with clean sand or glass beads. Porosity and porous media characteristics are listed in table 1. The sand used in this study is called “sand of Fontainebleau” (south Paris, France) and is very homogeneous (over 99.6% silica). Before use, both porous media were thoroughly rinsed with milliQ water on a 40µm filter (VWR international, 11cm, type 417) then sterilized and oven dried for at least 2 hours at 120°C. Next the columns were saturated with sterilized water (10-1M or 10-2M NaCl pH7). For each experiment porous media were renewed in the columns. Bacterial cells were collected after overnight culture, centrifuged for 5min at 5000tr/min and suspended in a sterile 10-1M or 10-2M NaCl pH7 solution. Before use, concentration of the bacterial solution was adjusted approximatively to 5.107 cells/ml. Injection of the bacteria in the column lasted 10 minutes at the approximate flow rate of 3 mL.min-1 (flow rates are listed in table 1). Circulation of the bacteria through the column was obtained by mean of a peristaltic pump (Ismatec SA reglo digital MS-4,Switzerland). Each column experiment was repeated four times with bacteria cells from separate cultures. Experiments were also carried out with decreasing the ionic strength (IS) of NaCl (10-1M, 10-2M, 10-3M, 10-4M and finally milliQ water). These experiments were repeated two times with cells from independent cultures for both the sand and the glass beads columns. Times of IS change of injected solutions are listed in table 3 (indicated in pore volumes). Bacterial concentration was measured by taking 0.5ml samples at the outlet of the column for flow cytometer enumeration. The flow cytometer (Beckman Coulter, EPIC XL, Fullerton, USA) settings were as follows: excitation source 488nm, fluorescence detection on FL2 channel (emission 510-530nm), analysis time 1 minute, no compensation was used.

Fluorence time lapse confocal imaging

 Bacterial displacement through the two porous media (sand and glass beads) was observed with a LEICA SP2 AOBS spectral confocal microscope (Leica Microsystems, France) at the MIMA2 microscopic plateform (http://voxel.jouy.inra.fr/mima2). A single channel micro-flowcell (length: 50mm volume: 250mm3 , Ibidi, Germany) was filled with sand or glass beads and connected to a peristaltic pump (Watson-Marlow, serie200, UK) running at 8.5µL.min-1 to obtain about the same flow rate as in the column experiments. The fluid running through the porous media was a pH7 10-1M or 10-2M NaCl solution to visualize IS effect on bacterial adhesion/detachment phenomena. Once the bacteria were attached to the matrix with 10-1M NaCl solution, distilled water was also injected in the flow cell. Bacterial concentration of the injected solution was approximatively 106 cells.ml-1. Since both sand particles and glass beads are not fluorescent FITC was added to the bulk fluid when necessary to create a negative fluorescent staining. FITC is a green fluorescent molecule with maximum excited/emission wavelengths of 506/529nm. For time lapse series, a fluorescent picture was taken every 3 seconds from a same field to view the dynamic of bacterial adhesion/detachment process in situ and in real time. We also looked for specific sites favorable for adhesion and wedging in both porous media such as particle-particle junctions, dead end paths or cell scale furrows were bacteria might be trapped. LSClite v2.5 software (Leica Microsystems, Germany) was used to quantify the fluorescence from the time series obtained.  

Ionic strength and electrostatic repulsions 

Both the PHL1314 strain and the porous media studied showed negative potential zeta values in all the ionic solutions (see table 2). Zeta potentials of the sand grains and the glass beads were nearly the same whatever the IS. The Gibbs free energies of interaction (i.e. GEL + GLW) between the sand and the PHL1314 cells as a function of distance and ionic strength was calculated using the classical DLVO theory (see figure 1). In the highest ionic conditions (10-1M NaCl, pH7) electrostatic repulsions between the bacteria and the porous media were the lowest (404kT). Lowering the IS causes a decrease of the zeta potential of both the cells and the porous media and consequently an increase of electrostatic repulsions. At 10-4M NaCl electrostatic repulsions were the highest (2300kT). Lowering the IS also increased the range of electrostatic interactions from 4nm at 10-1M NaCl to 350nm in 10-4M NaCl solutions. The secondary minimum provided by LW interactions was extremely weak compared to the electrostatic barrier and ranged from -3.69kT to -0.04kT. Moreover increasing the electrostatic barrier when the IS was decreased caused the secondary minimum to be attenuated (see figure 1). 

Breakthrough curves 

For both porous media studied ionic strength was a main factor for transport behavior of the PHL1314 strain and results are listed in table 3. In a 10-1M NaCl solution transport of the PHL1314 strain was negligible for both the sand and the glass beads. In 10 fold lower ionic conditions the cells were significantly better transported: 89.8% (SD=2%) on the sand and 94.9% (SD=4%) on the glass beads. The bacteria breakthrough peak approximatively corresponded to the breakthrough peak of the tracer and was similar for the sand and the glass columns (about 1 pore volume, see figure 2). To further investigate the influence of electrostatic interactions on bacterial adhesion in porous media the ionic strength was progressively decreased inside the column binjecting consecutively lower NaCl concentrated solutions starting at 10-1M NaCl (pH remained constant at 7). Each time the IS was lowered variable amounts of bacteria cells were removed from the porous media (figure 3). The percentages of removed bacteria are given in table 3. Lowering the IS had a stronger effect in the glass beads column than in sand columns: 42% of the cells were released from the glass beads but only 18% from the sand matrix. Moreover it is interesting to notice that the amount of removed cells varied with the IS of the incoming solutions. Bacterial detachment was much better using the 10- 3M or 10-4M solutions than with the 10-2M solution (figure 3 and table 3). However lowering the IS does not enable to transport as many cells as with a low ionic solution on its own: for example 89% of the cells were transported through sand with a 10-2M solution whereas only 2% of the cells traveled through the sand matrix when the IS was lowered from 10-1M to 10-2M (table 3).