Environ. Sci. Technol. 2004, 38, 4596-4602
Electrokinetic Transport of PAH-Degrading Bacteria in Model Aquifers and Soil L U K A S Y . W I C K , * ,†,‡ PHILIPP A. MATTLE,† P I E R R E W A T T I A U , §,# A N D H A U K E H A R M S ‡ Swiss Federal Institute of Technology Lausanne (EPFL), School of Environmental Sciences, Architecture and Civil Engineering, Laboratory of Soil Science, CH-1015 Lausanne, Switzerland, UFZ Centre for Environmental Research Leipzig-Halle, Department of Environmental Microbiology, D-04318 Leipzig, Germany, and Catholic University of Louvain (UCL), B-1348 Louvain-la-Neuve, Belgium
An investigation of the mobility, viability, and activity of polycyclic aromatic hydrocarbon (PAH) degrading bacteria in an electric field is presented. Bench-scale model aquifers were used to test electrophoresis and electroosmosis as potential mechanisms for bacterial dispersion in contaminated sites. Glass beads, alluvial sand from Lake Geneva, and historically polluted clayey soil were used as packing materials. The green-fluorescent protein labeled PAH-degrading bacteria Sphingomonas sp. L138 and Mycobacterium frederiksbergense LB501TG were used as test organisms because of the known differing physicochemical surface and adhesion properties of the corresponding wild-type strains. No adverse effects of the electric current on bacterial viability and PAH-degradation were observed in the system chosen. Up to 90% of the weakly negatively charged and moderately adhesive cells of strain L138 were transported by electroosmosis, whereas 0-20% were transported by electrophoresis. By contrast, poor electrokinetic transport of strongly charged and highly adhesive cells of M. frederiksbergense LB501TG occurred in the different model aquifers. Treatment of bacteria with the nonionic surfactant Brij35 resulted in up to 80% enhanced electrokinetic dispersion of both strains. Our findings demonstrate that electroosmosis may be a valuable mechanism to transport bacteria in the subsurface with transport efficiencies heavily depending on the retention of the bacteria by the solid phase.
Introduction One of the major obstacles to biotechnological soil remediation is the limited access of bacteria to hydrophobic organic contaminants (HOC) (1). The concurrence of sorption-retarded HOC transfer and the restricted mobility of soil bacteria limit bioavailability and, consequently, pollutant biodegradation. Soil bacteria are believed to form microcolonies with estimated distances in the range of 100 µm * Corresponding author phone: + (341) 2352523; fax: + (341) 2352247, e-mail:
[email protected]. † Swiss Federal Institute of Technology. ‡ UFZ Centre for Environmental Research. § Catholic University of Louvain. # Present address: Veterinary and Agricultural Research Institute (VAR), Brussels, Belgium. 4596
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004
between pollutant-degrading microcolonies (2) and contaminant sources such as droplets of nonaqueous phase liquids (NAPL). To achieve HOC degradation, it is thus crucial to enhance the contact probability of the bacteria and their HOC substrates either by enabling faster contaminant transport to the bacteria or by enhancing the mobility of bacteria. Several studies have shown that the application of direct current to soil (electrokinesis) moves charged chemicals (electromigration), whereas uncharged compounds can be transported by electrically induced interstitial water movement (electroosmosis). Capillary electrophoresis nowadays is an established tool for the separation and analysis of bacteria in the laboratory (for a review, see ref 3), but little is known on the influence of electrokinetic mobilization on biodegradation (electro-bioremediation) (4) and the electrokinetic dispersion of microorganisms in the subsurface (5-8). As bacteria are charged at circumneutral pH, dispersion of bacteria by electrokinesis appears to be a way to stimulate biotechnological cleanup. Electrokinetic treatment is an emerging engineering technique for the remediation of contaminated soils (9). It has been applied to both metal and organic contamination (10, 11). It can be performed in situ and is particularly effective for fine-grained soils of low hydraulic conductivity, which are difficult to treat by other in situ methods. A constant voltage direct current (dc) generating potential gradients in the range of 20-200 V m-1 is often used so that a small variable current passes through the soil (12). For nonionic substances such as gasoline hydrocarbons, phenol (13), and trichloroethylene (14), electroosmosis, that is, the creation of waterflow along charged surfaces, is the predominant transport mechanism. It appears, however, that electroosmotic longdistance transport of HOC is only feasible when they are present as droplets or colloidal particles (12, 15, 16). For in situ bioremediation, however, long-distance transport may not be needed, as bacteria are ubiquitous in soil and transport of either pollutant or bacteria in the millimeter range is likely to improve mutual contact rates. Here we present work aimed at studying the mobilization, viability, and activity of two specialized PAH (polycyclic aromatic hydrocarbon) degrading bacteria (17), M. frederiksbergense LB501TG and Sphingomonas sp. L138, under electrokinetic conditions. To our knowledge no data on the impact of direct current (dc) on the viability of bacteria and biodegradation of dissolved PAH exist in the literature. This knowledge is of particular relevance for bioremediation as sphingomonads and mycobacteria are thought to be major agents of PAH biodegradation in contaminated soils (17-19). To study the mobility of these nonmotile strains, bench-scale model aquifers were constructed to test electrophoresis and electroosmosis as potential mechanisms of bacterial dispersion in different porous matrixes, such as glass beads, alluvial sand from Lake Geneva (Switzerland), and historically PAH-polluted clayey soil from Andujar (Spain).
Material and Methods Organisms and Culture Conditions. Sphingomonas sp. L138 and M. frederiksbergense LB501TG (20) are green fluorescent protein (GFP)-labeled homologues of Sphingomonas sp. LB126 and M. frederiksbergense LB501T (17), respectively. Bacteria were cultivated on 2 g L-1 glucose in a minimal medium (21) containing 5 µg mL-1 of tetracycline (strain L138) and 15 µg mL-1 of kanamycin (strain LB501TG) to maintain the gfp marker. Cultures were grown at 25°C on a gyratory shaker at 130 rpm in 300-mL Erlenmeyer flasks 10.1021/es0354420 CCC: $27.50
2004 American Chemical Society Published on Web 07/31/2004
FIGURE 1. Schematic view of the setup used for electrokinetic experiments, consisting of the acrylic electrokinetic chamber, two titanium-iridium electrodes traversing the width of the electrode compartments, and a buffer recirculation system. The electrokinetic apparatus (40 cm × 7 cm × 3.5 cm i.d.) is composed of two electrode chambers, a lid-covered model aquifer chamber, and a bypass channel placed below the aquifer chamber connecting both electrode compartments. Titanium-iridium electrodes were inserted at 1 cm from both front ends. Sefar Tetex membranes were placed 1 cm from the electrodes to separate the electrode compartments from the aquifer materials. containing 100 mL of medium. To avoid biasing of the growth phase, cells were always harvested in the stationary phase (after 3 and 9-10 days for strains L138 and LB501TG, respectively) washed three times with 40 mL cold 10 mM phosphate-buffered saline (PBS) and resuspended in 2 mL of 10 mM PBS resulting in an optical density at 578 nm of 40 to 60. Brij35-treated cells were obtained accordingly by suspending washed bacteria in 2 mL of 10 mM PBS containing 100 times the critical micelle concentration (cmc) of Brij35 (cmcBrij 35 ) 0.05-0.1 mM)). To test growth on Tris acetate buffer (TA) (22) used in electrokinetic experiments, cultures were grown at 25°C and 130 rpm in 300-mL Erlenmeyer flasks containing 100 mL of 0.05 M TA. Glucose agar plates amended with kanamycin and tetracycline were used for plate counting. All solutions were prepared with deionized water (Nanopure Cartridge System, SKAN, Basel, Switzerland). gfp Labeling by Transposon Insertion. Transposon mutagenesis of the fluorene-degrading strain Sphingomonas sp. LB126 was carried out by biparental mating essentially as described earlier (23). Briefly, concentrated cell suspensions (OD600 ≈ 50) derived from fresh cultures of each donor and recipient strain were washed three times with 0.9% (w/ v) NaCl. One hundred µL of each suspension were gently mixed and deposited on a TSA agar plate. After overnight incubation at 30°C, the resulting spot was collected and homogenized in 500 µL of 0.9% (w/v) NaCl. Serial dilutions were plated on phosphate agar plates supplemented with 0.2% (w/v) glucose, 100 µg mL-1 streptomycin, and 4 µg mL-1 tetracycline. Colonies of Sphingomonas sp. LB126 recombinants appeared after 4-7 days with a frequency of about one recombinant for 107 recipients. Five fluorescent colonies were purified on the same medium and tested for their ability to grow on both glucose and fluorene. One labeled strain (L138) was kept for later experiments. Characterization of Bacterial Cell Surface Properties. Physicochemical cell surface properties were determined using standard methods. The electrophoretic mobility of bacterial suspensions in 10 mM PBS at pH 7.2 was determined in a Doppler electrophoretic light scattering analyzer (Zetamaster, Malvern Instruments Ltd., Malvern, Worcestershire, United Kingdom) at 100 V. The zeta potential (ζ) as an indirect measure of cell surface charge was approximated from the electrophoretic mobility according to the method of Helm-
holtz-Smoluchowski (24). The zeta potential of the grampositive strain LB501TG, however, may have been underestimated by the Smoluchowski relation due to the occurrence of high conductances at the surface of the cell walls at the low electrolyte concentration used (25). Bacterial lawns needed for contact angle (Θw) measurements were prepared by collecting cell suspensions in 10 mM PBS on 0.45-µm pore-size Micropore filters (Schleicher & Schuell, Dassel, Germany), mounting the filters on glass slides, and drying them for 2 h at room temperature. Cell surface hydrophobicities were derived from the contact angles θw of water drops on the bacterial lawns using a microscope with a goniometric eyepiece (Kru ¨ ss GmbH, Hamburg, Germany) (26, 27). Analytical Methods. Suspended cell density was determined spectrophotometrically at 578 nm after carefully shaking the sample. Cell suspensions of OD578 > 0.5 were diluted. High performance liquid chromatography (HPLC) (HP Series 1050; Hewlett and Packard) analysis was performed on an RP-18 column (Vydac, RP 18 (4.6 × 250 mm; 5 µm, 90A); mobile phase: MeOH/water (70:30 v/v); flow: 1 mL min-1) and the PAH detected by fluorimetry (anthracene: λex ) 251 nm, λem ) 400 nm; fluorene: λex ) 270 nm, λem ) 340 nm). Electrokinetic Apparatus and Setup. The electrokinetic apparatus consisted of a Plexiglas trough (inner dimensions: length 40 cm, depth 7 cm, and width 3.5 cm) that is divided into four compartments: two electrode chambers (2 × 7 × 3.5 cm) at both ends, a lid-covered model aquifer chamber (35.5 × 4 × 3.5 cm) between, and a bypass channel (35.5 × 2 cm × 3.5 cm) below the aquifer compartment having hydraulic contact with both electrode compartments to exclude bacterial transport by advective hydraulic water flow (Figure 1). The lid that contained sample ports was in contact with the aqueous phase to avoid preferential bacterial dispersion at the air-water interface. In preliminary experiments using troughs without a lid, we observed that evaporation created microeddies that efficiently spread bacteria. Titanium-iridium electrodes (10 × 4 cm) of 1.5 mm thickness (Electro Chemical Services International SARL, Chaˆtelaine, Switzerland) were inserted at 1 cm from both ends. Sefar Tetex membranes (Sefar Tetex Mono 05-1001K 079, SEFAR AG, Heiden, Switzerland) were placed at 1 cm VOL. 38, NO. 17, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4597
from the electrodes to separate the electrode compartments from the solid packings. At each front end of the apparatus an inlet and outlet valve (1.5 and 9.5 cm above the bottom, respectively) was mounted. The whole chamber was filled with 0.05 M Tris-acetate-buffer (TA) of pH 7 and the electrolyte in both electrode chambers was recirculated to a common 1-L reservoir of TA buffer at a flow rate of 120 mL/h with a peristaltic pump (Ismatec, Glattbrugg-Zu ¨ rich, Switzerland) to avoid pH changes in the electrode chambers. No pH-gradient and no significant temperature change (
