Environ. Sci. Technol. 2007, 41, 198-205
Influence of Extracellular Polymeric Substances on Pseudomonas aeruginosa Transport and Deposition Profiles in Porous Media YANG LIU,† CHING-HONG YANG,‡ AND J I N L I * ,† Department of Civil Engineering and Mechanics, and Department of Biological Sciences, University of Wisconsins Milwaukee, Milwaukee, Wisconsin 53201
The impact of cell surface extracellular polymeric substances (EPS) on bacterial transport and retention profiles was investigated in saturated columns packed with glass beads. Three genetically well-defined isogenic Pseudomonas aeruginosa strains with different EPS secretion capability and EPS composition were used to systematically examine their deposition behavior over a range of solution chemistry. The presence of EPS on nonmucoid strain PAO1 and mucoid strain PDO300 significantly increased bacterial adhesion over the EPS deficient PAO1 psl pel mutant strain despite their similar surface charge as indicated by the zeta potential measurements. Retained bacterial profiles show the deposition rate coefficients with various shapes and degrees of deviation from those expected from the classic filtration theory. Non-monotonic deviations from the log-linear deposition pattern with the majority of the bacteria retained downgradient of the column inlet were observed when bacterial cells were encapsuled by EPS under both high and low ionic strength conditions. In contrast, the EPS-deficient strain exhibited monotonic deviation from theory only under low ionic strength conditions. The results demonstrate that the non-monotonic deviation from filtration theory observed in this study was driven by steric interactions between extracellular polymers and glass beads. Analysis of the retained polysaccharides (carbohydrates and uronic acids) and protein profiles suggests that bacterial re-entrainment and re-entrapment may have contributed to the downgradient movement of the maximum retained bacteria. The detachment of bacteria may leave behind various constituents of EPS as their “footprints,” which can be a valuable tool for tracking the trajectory of bacterial transport.
Introduction The transport and deposition of microbial particles in porous media are important phenomena in a variety of environmental contexts, such as water and wastewater filtration, bioremediation of contaminated soil and aquifers, and pathogen transport in subsurface (1, 2). The overall retention of colloids, including biocolloids (e.g., bacteria, viruses, and protozoa), in porous media is controlled by the mass transfer * Corresponding author phone: (414) 229-6891; fax: (414) 2296958; e-mail:
[email protected]. † Department of Civil Engineering and Mechanics. ‡ Department of Biological Sciences. 198
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007
of suspended particles from the bulk flow to the surface of collector grains and the attachment of particles to solid surfaces as a result of colloid-surface interaction. A number of physical and chemical processes, such as interception, sedimentation, and diffusion, determine the mass transport of colloid particles in porous media, as described in the classic colloid filtration theory first developed by Yao et al. (3). According to Yao’s conceptual model, particle concentration in the fluid phase is represented by first-order kinetics with a spatially and temporally constant colloid deposition rate coefficient. The suspended and retained particle concentrations in the porous media are therefore predicted to decrease exponentially with transport distance. Based on this assumption, the fraction of colloids recovered from the effluent of packed bed columns or aquifers is typically used to estimate the deposition rate coefficient (kd), or alternatively, the sticking efficiency (R) defined as the ratio of the number of particles that attach to collector grains to the number of particles that collide with collector grains (4). A growing number of studies, however, have demonstrated that discrepancies exist between experimental data and the log-linear decrease pattern anticipated from the “clean bed” theory under certain experimental conditions by investigating the concentration profiles of retained colloids. In most cases, the deposition rate coefficients of particles were found to decline monotonically with travel distance. This type of divergence of experimental data from theory prediction has been attributed to a variety of factors, including straining (11-13), heterogeneity in collector grain and microbial population (5, 9, 15), cell motility (6-7), surface roughness (15), time-dependent attachment (16), and colloid detachment (10, 17). Until very recently, non-monotonic deviations from theory have rarely been reported in the literature. Redman et al. (18) observed that the maximum retained profile of recombinant Norwalk viruses was located ∼10 cm from the filter inlet when wastewater was used as the pore fluid. The presence of organic matter in the wastewater was argued for blocking the deposition of particles by forming steric and/or electrostatic barriers near the filter entrance (19). More recently, non-monotonic deviations were reported by Li et al. and Tong et al. (20, 21) by investigating the retained profiles of an adhesion deficient bacterial strain Comamonas DA001 and latex microspheres. The form and magnitude of the deviation were found to be sensitive to solution chemistry (ionic strength) and column media (22). Even though the ubiquity of non-monotonic spatial variations in particle deposition rate coefficients has been recognized, the fundamental mechanism underlying this phenomenon is as yet unclear. Previous studies have found that bacterial cell types (e.g., gram positive and gram negative), outer membrane proteins, lipopolysaccharide (LPS), fimbrae, flagella, and EPS may significantly contribute to the microbial surface heterogeneity and interfere with extended Derjaguin-Landau-VerweyOverbeek (XDLVO) forces (23-38). Interactions between bacterial surface polymers and surfaces can be attractive or repulsive depending on the rigidity of the polymer and whether or not a higher affinity exists between the bacterial surface polymers and the aqueous phase than the solid surface (39). Jucker et al. (39, 40) reported that the uneven distribution of short- and long-chain polysaccharides on bacterial outer surfaces and formation of hydrogen bonds could significantly affect bacterial adhesion to surfaces. One distinctive characteristic of biofilm-forming bacteria is the presence of EPS that encase the constituent cells (41). The EPS matrix typically consists of a mixture of macromolecules, 10.1021/es061731n CCC: $37.00
2007 American Chemical Society Published on Web 12/05/2006
including proteins, polysaccharides, nucleic acids, lipids, and other polymeric compounds. EPS is often implicated in the initial attachment of bacteria to surfaces and subsequent biofilm formation since it is present at the outermost layer of the cell. However, it is not clear whether, and to what extent, EPS functions as the initial adhesive, because the composition of EPS varies from organism to organism and its function on bacterial adhesion is not well understood (29, 42-44). Here we attempt to determine the influence of EPS on bacterial adhesion by systematically investigating cell deposition and EPS profiles in packed columns using three genetically well-defined isogenic P. aeruginosa stains with various EPS secretion characteristics. P. aeruginosa is a gram negative bacterium that is present in a diverse range of ecological niches, ranging from water and soil to plant and animal tissues (45). It is considered the paradigm organism for microbial biofilm study. Our results suggest that not only the existence but also the composition of EPS play significant roles in controlling the overall deposition rate and retained profile of P. aeruginosa in porous media. Non-monotonic deviations from the log-linear decrease deposition pattern were observed when bacterial cells were encapsuled by EPS, under both high and low ionic strength conditions. The significant difference observed between the retained EPS profiles and cell profiles implicates that bacterial reentrainment and re-entrapment may have contributed to the occurrence of non-monotonic deposition in packed columns.
Materials and Methods Bacterial Strains and Cell Preparation. The specific strains include the following: nonmucoid wild-type biofilm-grown strain PAO1; the PAO1 isogenic psl and pel mutant strain with a deficiency in exopolysaccharide production and biofilm formation; and PDO300, an isogenic, mucoid alginateoverproducing derivative of the prototypic PAO1 with a mucA22 mutation. Mucoidy is defined as the formation of shiny colonies on an agar plate (46-48). The bacterial strains were provided by Dr. M. Parsek at the University of Washington. For each experiment, the stored strains were streaked onto Luria-Bertani (LB) agar plate and incubated at 37 °C overnight. A single colony was then transferred into 10 mL of LB broth and grown in a shaker incubator (New Brunswick Scientific Co., NJ) at 200 rpm and 37 °C for 16-20 h. Stationary-phase bacterial cells were harvested by centrifugation (Avanti J-20I, Beckman Coulter, CA) at 3000g and 4 °C for 10 min. After the supernatant was decanted, the pellets were resuspended in appropriate electrolytes (0.01×, 0.1×, and 1× phosphate buffered saline (PBS) with corresponding ionic strength of 1.65, 16.5, and 165 mM). The 1× PBS was prepared by dissolving 1.093 g/L Na2HPO4, 0.3175 g/L NaH2PO4‚H2O, and 8.475 g/L NaCl in ultrapure water (Milli-Q water, Millipore Corp., MA). PBS is commonly used to suspend cells for bacterial biofilm formation and cell attachment assays. Although the presence of phosphate groups might shield chemical heterogeneity on exposed surfaces and minimize cell-glass interactions, the transport and deposition of all three bacterial strains were compared under the same solution chemistry. The centrifugation and re-suspension procedure was repeated twice to remove traces of growth media. A final cell density of approximately 106 to 107 colony-forming units (CFU)/mL was obtained by measuring the optical density (OD) using a UV/Visible spectrophotometer (Varian, Inc., CA) at 600 nm wavelength. Cell suspensions were kept on ice before the filtration experiment to minimize potential bacterial growth. The pH of the electrolytes ranged from 7.0 to 7.2.
Electrokinetic Characterization of Bacterial Cells. The rinsed stock cell suspensions were diluted in 1×, 0.1×, and 0.01× PBS to a final OD of 0.2-0.25. A ZetaPALS Analyzer (Brookhaven Instruments Corp., NY) was used to measure the electrophoretic mobility and zeta potential. The measurements were performed 10 times for each assay. Granular Porous Medium. The filtration system consisted of polycarbonate plastic columns with an internal diameter of 2.54 cm and a length of 26 cm. The packing material was spherical glass beads (MO-SCI Specialty Products, LLC., MO) with an average diameter of 0.55 mm and a specific gravity of 2.5 g/cm3. Columns were wet-packed to a height of 25 cm with vibration to minimize any layering or air entrapment. The column porosity was estimated to be 0.4 using the standard gravimetric method. Glass beads were thoroughly cleaned prior to column assemblage following the method reported previously (49). Packed-Bed Column Experiments. A peristaltic pump (Cole Parmer, IL) was used to pump the solutions in a down flow mode. Prior to each experiment, the filter was equilibrated by pumping at least 20 pore volumes of bacteria free background PBS through the column at a constant approach velocity of 0.011 cm/s. The ionic strengths of the PBS were 1.65, 16.5, and 165 mM. Approximately 6 pore volumes of bacteria suspension (106-107 CFU/mL) were injected after switching the influent from the PBS to the cell suspension. Following the bacteria injection, columns were eluted with 8 pore volumes of background PBS. The bacterial transport and deposition were also examined in a longer travel distance, using two identical columns in series. Two columns in series were assembled by connecting the bottom of the first column to the top of the second column with a 40 cm long tubing. The column effluent was collected in 50 mL polystyrene tubes and immediately placed on ice, prior to microbiological analysis. Following each transport experiment, the column media were extruded and dissected into 5 cm long segments. Approximately 48 g of glass beads were obtained from each segment and placed in a beaker containing 50 mL of PBS. Bacteria and EPS bound to the surface of glass beads were dissociated by vertexing (Vertex-genie 2, Fisher Scientific, IL) at the maximum speed for 30 s followed by 10 min of ultrasonication (FS20H, Fisher Scientific, IL). Ultrasonication is a common method for dislodging bacterial cells from surfaces (50, 51). Column effluent and the supernatant samples from the centrifugation containing bacteria recovered from glass beads were enumerated by plating appropriate dilutions. All experiments were performed in triplicate at room temperature (20-25 °C). To quantitatively compare the overall deposition of the three P. aeruginosa strains at different solution ionic strengths, the deposition rate coefficient kd was estimated using the steady state breakthrough concentrations of the cell according to the following equation (26):
kd ) -
()
U C ln L C0
(1)
where is the bed porosity, U is the approach (superficial) velocity, L is the length of the column, and C/C0 is the normalized breakthrough concentration relevant to “clean bed” conditions, which was obtained from each bacterial breakthrough curve by averaging the values measured between 8 and 10 pore volumes. According to the classic colloid filtration theory, particle deposition pattern S(X), i.e., the number of deposited bacteria per mass of the granular collector, can be calculated using the following equation (14):
S(X) )
(
)
t0kdC0 kdX exp Fb U
VOL. 41, NO. 1, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
(2) 9
199
FIGURE 1. Surface zeta potential of three Pseudomonas aeruginosa strains, PAO1 (nonmucoid wild-type biofilm-grown strain), PAO1 psl pel (mutant strain with a deficiency in exopolysaccharide production), and PDO300 (mucoid alginate-overproducing strain), as a function of solution chemistry. Ionic strengths: 1.65 mM (0.01× PBS), 16.5 mM (0.1× PBS), and 165 mM (1× PBS). Error bars represent standard deviations of ten replicate measurements.
Here, X is column depth, Fb is the porous medium bulk density, and t0 is the duration of continuous particle injection. Bacteria Enumeration Procedures. Viable bacterial cell counts were obtained using the drop plate method (50). A series of 10-fold dilutions was performed, and 10 µL of each dilution was plated on LB agar plates in triplicate. Plates were incubated at 37 °C for 20 h before counting. The lower limit of this detection method is 103 CFU/mL. Carbohydrate and Uronic Acid Analysis. EPS components were isolated from the bacteria surface by placing 5 mL of cell suspension into 15 mL sterilized plastic centrifuge tubes, and sonicating for 20 s at 3.5 Hz with a cell disruptor (Branson Sonic Power Co., CT). Most of the cell capsular material can be stripped from the bacteria surface using this method without causing cell lysis (less than 10%) (38). The bacteria were removed by ultracentrifugation (Beckman Coulter, CA) twice at 33 000g and 4 °C for 10 min each time. The supernatant was then filtered through a 0.22 µm cellulose acetate filter to ensure that samples were free of bacterial cells. The total carbohydrates were quantified by the phenolsulfuric acid method with glucose as the standard (51). Uronic acids were measured using the m-hydroxydiphenyl sulfuric acid method, using D-glucuronic acid as the standard (52).
Results Electrokinetic Potential of Bacterial Cells. The variation in the zeta potentials of bacterial cells as a function of ionic strength is shown in Figure 1. Despite the difference in magnitude, all three strains exhibited negative zeta potentials indicating negative surface charge under the carrying solution concentrations. The absolute magnitude of zeta potential was reduced with increasing electrolyte concentration due to double-layer compression at high ionic strength. The EPS composition had only a small impact on the bacterial surface electrokinetic potential based on an estimate of the 95% confidence interval for PAO1 psl pel, which appeared to include the zeta potentials for the other two strains. Bacterial Cell Transport in Packed Beds. Bacterial breakthrough curves are shown in Figure 2 by plotting normalized cell concentrations (C/C0) in the column outflow against the number of pore volumes passed through the packed bed. The injected bacteria were detected in the 200
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007
FIGURE 2. Breakthrough curves of Pseudomonas aeruginosa strains, PAO1 (nonmucoid wild-type biofilm-grown strain), PAO1 psl pel (mutant strain with a deficiency in exopolysaccharide production), and PDO300 (mucoid alginate-overproducing strain), under three solution ionic strengths: 1.65 mM (0.01× PBS) (A), 16.5 mM (0.1× PBS) (B), and 165 mM (1× PBS) (C). Bacteria injection started at pore volume 4 and ended at pore volume 10. Experimental conditions were as follows: approach velocity ) 0.011 cm/s, porosity ) 0.4, and pH ) 7.0-7.2. column effluent within one pore volume upon switching the cell free electrolytes to cell suspensions. The breakthrough concentrations appear to be constant for PAO1 and the PAO1 psl pel EPS mutant, whereas PDO300 differs significantly from its counterparts with much lower normalized cell concentrations and gradually increasing breakthrough concentrations. The slow increase of effluent concentration over time indicates a temporal decrease in the deposition rate and has been attributed to “blocking” (16) caused by previously deposited bacterial cells or simultaneous bacterial deposition and release. Upon flushing the columns with cell-free solutions of identical chemical composition, a certain degree
FIGURE 3. Comparison of bacterial deposition rate coefficient (kd) determined from the breakthrough curves using eq 1 under various solution chemistry. Experimental conditions were as follows: approach velocity ) 0.011 cm/s, porosity ) 0.4, and pH ) 7.0-7.2. Error bars represent standard deviations of triplicate measurements. of tailing was observed during the washout, indicating cell detachment. As shown in Figure 3, the deposition rate coefficients increased with higher ionic strength for all three strains. This is in qualitative agreement with the prediction provided by classic DLVO theory, which states that increasing the ionic strength will cause reduction of the electrostatic double layer repulsion. Individual bacterial strains had distinct retention behavior over the range of ionic strengths tested, with the overall deposition behavior of PAO1 being much more sensitive to solution chemistry than its counterparts. It is also noted that under the same solution chemistry, the bacterial deposition rate increased with the increasing presence of EPS, although the measured zeta potentials for all three strains were virtually identical as shown in Figure 1. Retained Bacterial Profiles. Figure 4A, B, and C compare the retained bacterial profiles obtained by plotting the number of bacteria recovered from glass beads as a function of traveled distance and those to be recovered expected from the colloid filtration theory. Log-linear retained profiles were observed for PAO1 psl pel EPS mutant in 16.5 mM and 165 mM PBS, indicating spatially constant deposition rate coefficients as implicated in the “clean-bed” theory. Under lower ionic strength (1.65 mM PBS), the retained PAO1 psl pel cell concentrations were found to decrease hyperexponentially (faster than log-linear), indicating that the deposition rate coefficients declined with transport distance through the column. A decreasing deposition rate coefficient was the most common type of deviation from the theoretical log-linear profile reported in the literature (22). For PAO1, the retained cell concentration first increased and then decreased under all ionic strength conditions, indicating that the deposition rate coefficient changed non-monotonically with the transport distance. The same deposition pattern was observed for the extracellular polysaccharide alginate over-producer PDO300 over the solution chemistry tested, although the maximum number of bacteria was located closer to the column inlet relative to PAO1. The maximum retained concentration appeared to be located at 10 cm and 20 cm for PDO300 and PAO1 respectively, under different ionic strengths. Retained Carbohydrate and Uronic Acid Profiles To explain the distinct bacterial deposition patterns observed for different P. aeruginosa strains with various EPS secretion capability, the total carbohydrate and uronic acid profiles were analyzed for experiments carried out in 165 mM PBS. The normalized carbohydrate and uronic acid
concentrations (µg/cell number) in each sectioned layer were plotted against column depth, as shown in Figure 4D. For PAO1 psl pel mutant, the normalized carbohydrate concentration remained constant throughout the column. Since this EPS mutant is deficient in extracellular polysaccharide production, the detected total carbohydrate represents cellular polysaccharide inside the cell membrane. In contrast, the ratio between the total carbohydrate detected and PAO1 cell number was significantly higher than PAO1 psl pel near the column inlet and decreased toward the downgradient direction until the maximum cell number was achieved at ∼20 cm. Similar trends were found for PDO300, which had uronic acids (alginate) as the EPS main component. No uronic acid was detected for PAO1 and PAO1 psl pel strains as expected from their EPS composition. Effect of Column Length L on Breakthrough Curves and Retained Profiles. As shown in Figure 4E, the increase-thendecrease deposition pattern was less pronounced in the second column when two identical columns were run in series for both PAO1 and PDO300 strains, although peak deposition points occurred at similar places as occurred in the single column experiments.
Discussion PAO1 psl pel and PDO300 are isogenic variants of the wild type isolate strain PAO1, and differ only in their surface EPS secretion capability. The main extracellular polysaccharide secreted by PDO300 is aliginate, a linear copolymer of mannuronic and guluronic acids (uronic acids C6H10O7) joined by β 1-4 linkage. Uronic acids are negatively charged and may contribute to the overall negative surface charge of PDO300 cells (44). The primary carbohydrate constituents of PAO1 EPS are glucose (41.0%), rhamnose (14.3%), and mannose (13.9%), which are neutral sugars and which may shield the negatively charged surface functional groups (e.g., LPS, proteins) located on the cell membrane (45). The significant variation in the overall cell retention and retained cell profiles among the three bacterial strains demonstrates that interactions between bacteria surface polymers and collector grain surfaces played important roles in controlling cell transport and deposition in porous media. Recently, non-monotonic retained profiles in porous media have gained considerable attention in the literature. The fact that downgradient maximum retained profiles were observed in our experiments under relatively simple carrying solutions corroborates Li et al. and Tong et al.’s (20, 21) conclusion that unusual water chemistry was not necessary to yield such behavior. Other potential artifacts that may interfere with experimental data include pH changes across the column length, collector grain heterogeneity due to packing and grain migration during injection, and nonideal bulk flow causing shortcut within the column. A dye tracer study was performed and the results (not shown) confirmed that the flow was evenly spread upon injection. The pH of the carrying solution also remained constant throughout the experiment. Straining is unlikely to be a major mechanism for the observed deposition pattern since the ratio between bacteria and median grain diameter was very low (