Environ. Sci. Technol. 2008, 42, 9310–9316
Antecedent Growth Conditions Alter Retention of Environmental Escherichia coli Isolates in Transiently Wetted Porous Media HSIAO-HUI YANG,† JAYNE B. MORROW,† DOMENICO GRASSO,‡ ROBERT T. VINOPAL,† ARNAUD DECHESNE,§ AND B A R T H F . S M E T S * ,§ Environmental Engineering and Microbiology Program, University of Connecticut, Storrs, Connecticut 06269 College of Engineering and Mathematical Sciences, University of Vermont, Burlington, Vermont 05405-0156, Department of Environmental Engineering, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark
Received June 20, 2008. Revised manuscript received September 30, 2008. Accepted October 1, 2008.
The physical transport of Escherichia coli in terrestrial environments may require control to prevent its dissemination from potential high-density sources, such as confined animal feedlot operations. Biobarriers, wherein convective flows carrying pathogens pass through a porous matrix with high retentive capacity, may present one such approach. Eight environmental E. coli isolates were selected to conduct operational retention tests (ORT) with potential biobarrier materials Pyrax or dolomite, or silica glass as control. The conditions in the ORT were chosen to simulate conditioning by manure solutes, a pulse application of a bacterial load followed by rainfall infiltration, and natural drainage. Removal was limited, and likely caused by the relatively high velocities during drainage, and the conditioning of otherwise favorable adhesion sites. Flagella-mediated motility showed the strongest correlation to biobarrier retention. Significant variability was observed across the E. coli isolates, but consistently higher retention was observed for cells with external versus intestinal pregrowth histories. E. coli O157: H7 was retained the least with all examined matrices, while E. coli K-12 displayed moderate retention and may not serve asrepresentativemodelstrain.Pyraxisagoodcandidatebiobarrier material given its superior removal ability across the tested E. coli strains.
Introduction Animal waste in the form of raw or composted manure is routinely applied to land as a crop fertilizer or soil amendment. A similarly concentrated input of animal fecal matter to soils occurs at concentrated animal feed operations (1). Many pathogens can be excreted in the feces of both infected and healthy carrier animals (2), and this large-scale disposal * Corresponding author phone: (+45)4525-2230; fax: ( +45)45932850; e-mail:
[email protected]. † University of Connecticut. ‡ University of Vermont. § Technical University of Denmark. 9310
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 24, 2008
of fecal waste might result in the spread of enteric pathogens (3). Therefore, the fate and transport of pathogens must be understood to evaluate and possibly mitigatesvia suitable biobarrier technologiesscontamination of water supplies from such concentrated inputs of fecal material in the environment. Bacterial retention in soils, which are characterized by periodic variations in hydraulic saturation, can be caused by attachment on surfaces and interfaces and straining or trapping in small pores spaces or in liquid films (4, 5). Attachment is commonly thought to be the main retarding factor, but straining or wedging may dominate when conditions for adhesion are unfavorable (5, 6). For straining to become an important removal mechanism, the ratio of the mean colloid diameter to the mean pore or grain diameter needs to exceed 0.5-1% (7, 8). Factors that impact bacterial adhesion are ionic strength, the pH of the aqueous phase, the surface properties of the cells and geological matrix, and the hydrodynamic conditions (9-12). In addition, bacterial motility may be an important factor in initial attachment (13, 14). The degree of bacterial adhesion can, furthermore, change dramatically with the physiological state of the bacterium and concomitant changes in cell surface properties (10, 15-18). For E. coli, various specific cell surface structures, such as fimbriae, conjugal pili, and autoaggregative adhesins may be required for adhesion and subsequent biofilm formation (19). Usually, bacterial transport studies consider one or a few strains under carefully controlled chemical and hydrodynamic conditions, and rarely consider transport variability across strains within a single species. We have previously documented significant subspecies variability among environmental E. coli isolates (20), which displayed variable but growth-condition dependent cell surface properties (21). Hence, we were especially interested in evaluating the uniformity of transport and retention behavior across this set of environmental E. coli isolates under hydrodynamic conditions that provided a rudimentary approximation to rainfall-driven infiltration. Our long-term goal is to develop reactive barriers for pathogen attenuation (biobarriers), which augment the natural filtering effect of soils. Test-minerals were, therefore, chosen based on predicted adhesion capacity (Pyrax and dolomite) predominantly due to their hydrophobic properties (22). This study summarizes the impacts of intestinal and external antecedent growth conditions on bacterial retention in three selected mineral matrices across eight E. coli isolates; significant factors governing bacterial retention efficiency were derived.
Materials and Methods Bacterial Strains and Growth Conditions. E. coli O157:H7 (ATCC 700728), E. coli K-12 (ATCC 25404), and E. coli isolates UCFL-71, UCFL-94, UCFL-131, UCFL-167, UCFL-263, and UCFL-348 were included in this study (Supporting Information Table S1). Detailed intestinal and external antecedent growth conditions media and growth conditions were described previously (21). Briefly, cells were grown without shaking at 39 °C for 6 h in screw-capped tubes previously filled with rumen medium to mimic intestinal growth condition (23). To mimic external growth condition, cells were grown in manure extract at 150 rpm, 48 h, 23 °C until stationary growth phase. Bacterial Surface Properties. Expression of FimH and Ag43 was estimated by yeast agglutination assay and immunofluorescence microscopy, respectively. Motility and electrophoretic mobility were measured by soft agar plates 10.1021/es801708h CCC: $40.75
2008 American Chemical Society
Published on Web 11/18/2008
FIGURE 1. Overview of the operational retention tests. and zeta potential, respectively. Hydrophobicity was inferred from water contact angles. Detailed methodology and results were described previously (21). Column Packing Material. Glass beads (∼500 µm in diameter) (Cole-Parmer, Vernon Hills, IL) were cleaned with 15% nitric acid, agitated on a shaker table for 12 h, rinsed with deionized water, and dried overnight at 105 °C. Pyrax (35-45% quartz and 0, hydrophilic surface, ∆Giwi applied cells). In those cases, zero removal was computed. Overall mass balances indicated cell retentions between 0-35%, 0-62%, 0-61% (glass, Pyrax, dolomite) and 2-61%, 1-65%, 0-62% (glass, Pyrax, dolomite) of the applied dosage for intestinal and externally grown cells, respectively. Comparison Across Column Matrices and Strains (Based on CRF) (Figure 2, Table S2). It is evident that the pregrowth conditions had a differential effect on strain retention. Some strains displayed a consistent effect of pregrowth condition, irrespective of the column matrix: strains 0157:H7, UCFL -167, and UCFL-348, and K-12 all showed a remarkably higher retention when pregrown under external vs intestinal conditions, while strains UCFL-131 and UCFL-263 showed the opposite trend. For strain UCFL-71, the response was different: external pregrowth resulted in higher retentions on Pyrax, but lower retention on glass and dolomite. For
FIGURE 3. Fractional cell numbers retained in the biobarriers as calculated by direct extraction of the sacrificed columns (RRF) after percolation with manure extract and diH2O. For each strain bars (1, 2), (3, 4), and (5, 6) refer to retention in glass, Pyrax, and dolomite columns with intestinal (1, 3, 5) and external (2, 4, 6) pregrowth history, respectively. Numbers are expressed as a % of the applied cells (I). strain UCFL-91, external pregrowth resulted in higher this is less the case under external conditions (F: 0.02). Second, retention on glass and Pyrax, but lower retention on dolomite. motility and Ag43 expression appear the core “correlants” Across the treatments, Pyrax consistently displayed the with cell retention in either condition. Third, and surprisingly, highest retention across strains and treatments. The only hydrophobicity does not explain cell retention well. exception was for intestinally pregrown UCFL-71, which displayed an anomalously high retention on glass beads, and Discussion for externally pregrown UCFL-131 where Pyrax retained When animal-origin fecal matter is applied to land, vertical essentially no cells. and horizontal movements of a wide variety of enteric Recovered (RRF) versus Calculated (CRF) Retained pathogens are possible. To mitigate such migration, physical Fractions. (Supporting Information Figure S1) The fraction biobarriers, which retain pathogens from the convecting of retained cells was also directly examined by sacrificing fluids, can be considered. While several low-cost materials the columns and vigorously mixing the column matrix in a can be identified with high retentive/adhesive properties for dilute surfactant solution: the recovered fraction (RRF) ranged bacterial cells, it is essential to test these materials under a from 0.5 to above 100% of the calculated fraction (CRF). range of biologically and environmentally relevant conditions: Recoveries less than 100% could indicate either overestimates across strains and different growth histories, under realistic in the CRF or truly irreversible adhesion to the grain matrix: aqueous chemistry and hydrodynamic settings. In a previous the latter was especially noted for several strains on dolomite study, we observed significant differences in surficial behavior and glass beads. Loss of cell culturability in E4 was an unlikely in a subset of environmental E. coli isolates with antecedent reason for low RRF, because the experiments were shortintestinal and external growth histories (21). Eight isolates term and cells in E4 were treated and enumerated identically were selected to conduct operational retention tests against to those in the other fractions. three candidate biobarrier materials under condition simuNevertheless, RRF can also be taken as an indicator of the lating rainfall infiltration/natural drainage conditions. retained fraction, and its values have a narrower experimental All strains displayed negative zeta potentials (0 to -14 error. An even clearer picture emerged (Figure 3, Supporting mV with PBS (I ) 0.125 M), which would translate in more Information Table S2): across the strains examined, externally negative zeta potentials under the applied conditions (mapregrown cells were retained to a larger extent than intesnure extract, 7 mOsm ∼10 mM I, assuming that all solutes tinally pregrown cells. This was only reversed for two isolates: are ionizable with an average ionic charge number of 2, and UCFL-131 and UCFL-263. Also, Pyrax was a consistently better ideal solution conditions). Hence, dominant repulsive elecbiobarrier material than glass or dolomite. Only strain UCFLtrostatic interactions are expected for all strains, as they are 263 behaved differently: dolomite was the superior retaining negatively charged under circumneutral conditions (22). matrix, while intestinally pregrown K-12 was not retained by Nevertheless, short-range attractive forces might exist for either biobarrier material. RRF and CRF trends are consistent. the more hydrophobic strains (e.g., for UCFL -71, 167, 263, For example, respectively, UCFL-263 and UCFL-94 were 348) especially on the Pyrax and dolomite materials. Hence, consistently the most and least retained strains across all it is likely that overall interaction was unfavorable. conditions examined (Figure 2 and Supporting Information Although the different biobarrier sphericities (glass < Figure S1). dolomite < Pyrax) would favor wedging and straining in the Analysis of Factors Contributing to Retention. Analysis latter (7, 27), such processes were likely of limited significance of variance indicated that all three factors (strain type, here. With typical cell size between 0.4 to 0.8 µm (Supporting cultivation conditions, and column material) contributed very Information Table S1), and minimum grain size diameters significantly (p < 0.01) to cell retention and cell recovery. In of 425 µm, the computed colloid/grain ratios (0.08-0.2%) addition, all the pairwise and threewise interactions were are below those where straining is considered significant (7). significant or close to the significance level (p always
