Idling Time of Motile Bacteria Contributes to Retardation and

Apr 1, 2011 - tion behaviors, motile bacteria appeared to idle in place after encountering the solid surfaces and this idling behavior persisted...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/est

Idling Time of Motile Bacteria Contributes to Retardation and Dispersion in Sand Porous Medium Jun Liu,† Roseanne M. Ford,†,* and James A. Smith‡ †

Department of Chemical Engineering, ‡Department of Civil and Environmental Engineering, University of Virginia, Charlottesville, Virginia 22904, United States

bS Supporting Information ABSTRACT: The motility of microorganisms affects their transport in natural systems by altering their interactions with the solid phase of the soil matrix. To assess the effect of these interactions on transport parameters, a series of breakthrough curves (BTCs) for motile and nonmotile bacteria, including E. coli and P. putida species, were measured from a homogeneously packed sand column under three different interstitial velocities of 1 m/d, 5 m/d, and 10 m/d. BTCs for the nonmotile bacteria were nearly identical for all three flow rates, except that the recovery percentage at 1 m/d was reduced by 5% compared to the higher flow rates. In contrast, for the motile bacteria, the recovery percentages were not affected by flow rate, but their BTCs exhibited a higher degree of retardation and dispersion as the flow velocity decreased, which was consistent with increased idling times of the motile strains. The smooth-swimming mutant E. coli HCB437, which is unable to change its swimming direction after encountering the solid surfaces and thus has the largest idling time, also exhibited the greatest degree of retardation and dispersion. All of the experimental observations were compared to results from an advection-dispersion transport model with three fitting parameters: retardation factor (R), longitudinal dispersivity (RL), and attachment rate coefficient (katt). In addition, the single-collector efficiency (η0) and collision efficiency (R) were calculated according to the colloid filtration theory (CFT), and confirmed that motile bacteria had lower collision efficiencies than nonmotile bacteria. This is consistent with previously reported observations that motile bacteria can avoid attachment to a solid surface by their active swimming capabilities. By quantifying the effect of bacterial motility on various transport parameters, more robust fate and transport models can be developed for decision-making related to environmental remediation strategies and risk assessment.

’ INTRODUCTION To cleanup groundwater contamination by implementing bioremediation requires improved understanding and control of bacterial surface interactions within the soil matrix.1 In order to remove dense nonaqueous-phase liquid (DNAPL) contaminants such as trichloroethene (TCE) and tetrachloroethene (PCE), which tend to settle and accumulate at the bottom of confined aquifers due to their high density and immiscibility,2 it is desirable to efficiently deliver contaminant-degrading bacteria to the contaminated region by minimizing their adsorption to the soil matrix. On the other hand, to attenuate mobile contaminants, it is beneficial to create an appropriate reaction zone, where bacteria become immobilized and form biofilms capable of degrading contaminants that pass through the reaction zone via the groundwater flow. Bacterial attachment to grain surfaces in the soil matrix is the first critical step3 in biofilm formation. In groundwater environments, the structure of the soil matrix interferes with the diffusive-like random walk exhibited by swimming bacteria in bulk aqueous solution.4 Several researchers 5,6 observed that, except for executing normal run and reorientation behaviors, motile bacteria appeared to idle in place after r 2011 American Chemical Society

encountering the solid surfaces and this idling behavior persisted until a bacterium reoriented itself to swim away from the surface. A 3-D tracking microscope was used to record individual bacterial swimming trajectories adjacent to a solid planar surface,7,8 and quantify bacterial idling time (tI),6,9,10 the time that bacteria spent on associating with the solid surfaces7 before returning to the bulk fluid. Thus, it was speculated that, under certain circumstances, bacteria do not physically attach to the solid surface they approach, but because they are unable to penetrate the surface, remain associated with the surface during this idling period.6,9,10 The proposed bacterial surface association mechanism,6,10 which was mathematically represented as a sorption/desorption term, was able to account for the accumulation of bacteria at an interface between bulk fluid and a packed bed of particulates, that diffusion alone could not explain.10,11 The addition of sorption-like terms to a diffusion model also Received: December 2, 2010 Accepted: March 17, 2011 Revised: March 8, 2011 Published: April 01, 2011 3945

dx.doi.org/10.1021/es104041t | Environ. Sci. Technol. 2011, 45, 3945–3951

Environmental Science & Technology

ARTICLE

successfully captured the highly impeded bacterial migration behavior that was observed in one packed column experiment 12,13 in which a step change in bacterial distribution gradually relaxed over time as bacteria diffused through the porous medium by their motility. Convective flow within the soil matrix complicates the interactions between bacteria and solid surfaces. In addition, and at typical groundwater flow rates, bacterial motility14,15 is another influential factor that must be accounted for in terms of the DLVO theory16 along with the effects contributed by hydrodynamic forces and electrostatic interactions.17,18 Colloid filtration theory (CFT)19 is one widely applied theoretical approach used to characterize the degree of bacterial interaction with the solid surface. Under an idealized clean-bed assumption, the bacterial attachment rate coefficient (katt) can be expressed as19 katt ¼

3 1ε Vf Rη0 2 dc

ð1Þ

in which, dc is the spherical collector diameter, Vf is the interstitial velocity (cm/s), ε is porous media porosity, R is collision efficiency, defined as the number of the contacts which succeed in producing adhesion divided by the number of collisions which occur between suspended particles and the filter media;19 η0 is the single-collector efficiency, defined as the rate at which particles strike the collector divided by the rate at which particles flow toward the collector.19 Note that, when clean-bed conditions are not satisfied or in the presence of repulsive electrostatic interactions, previous reports20,21 have used a distribution of collision efficiency (R) values to capture the experimental observations. It has been reported that bacterial idling time (tI) plays an important role in bacterial transport within static granular media.10 As most groundwater environments are subject to advective flow, this study was targeted to investigate the impact of tI on bacterial transport within a sand-packed column under flow, and the influence of different flow rates on bacterial association with encountered solid surfaces.

’ MATERIALS AND METHODS Bacteria and Culture Conditions. Experiments were conducted with Escherichia coli K12 derivatives, provided by Dr. Howard C. Berg,22 and soil-inhabiting bacterial strain Pseudomonas putida F1, provided by Dr. Caroline S. Harwood.23 More specifically, E. coli species included a motile wild-type strain E. coli HCB1, with full run and tumble ability; a motile smoothswimming mutant E. coli HCB437, deficient of tumbling capability; a nonmotile flagella-less mutant E. coli HCB137; a nonmotile tumbly mutant E. coli HCB359, and a nonmotile paralyzed mutant E. coli HCB136. Figure 4 provides a graphical depiction of the differences in swimming properties for the wild type and various mutant strains. Some physical characteristics (size, zeta potential and swimming properties) of the bacterial strains are described in the Supporting Information. The growth media for these E. coli species and P. putida F1 were Tryptone Broth6 and modified Hutner’s Mineral Medium,24 respectively. A 100 μL aliquot of the bacteria from frozen stock was used to inoculate 50 mL of sterile growth medium in a 250 mL baffled shake flask. The shake flask was incubated in a MaxQ4000 Shaker (Thermo Scientific, Dubuque, IA) set to approximately 30 ( 1 C and 150 rpm. Bacteria were harvested in midexponential growth phase when their absorbance reached about OD590 = 0.8,

measured by UV/vis Spectrophotometer DU730 (Beckman Coulter, Fullerton, CA). Before performing experiments, the growth substrate was removed by rinsing the bacteria three times with random motility buffer (0.029 g ethylenediaminetetraacetic acid (EDTA), 11.4 g K2HPO4, and 4.8 g KH2PO4 per 1 L deionized (DI) water) on 0.22 μm cellulose filters (Millipore Inc., Billerica, MA), and then the bacteria were resuspended in 10% strength random motility buffer, which had pH around 7 and ionic strength about 0.02 M, at a concentration of about 1.0  108 cells/mL, which was sufficiently dilute to prevent interaction among the bacteria. In addition, bacteria were inspected visually under the microscope to check for motility and consistency from experiment to experiment. Column Packing and Experimental Procedures. A 1.5 cm diameter, 6.8 cm long Omnifit glass chromatography column (Diba Industries, Danbury, CT) was used for this study. In order to create an even flow distribution across the radial dimension, each end of the column was equipped with a 1.5 cm diameter polystyrene disk with 20 μm pore diameter (Kimble Chase Life Science, Vineland, NJ), (See Figure S1 in the Supporting Information). Quartz sand (Acros Organics, Geel, Belgium), with 40100 mesh dimensions (280 μm mean diameter), was wet-packed into the glass chromatography column.26 Prior to packing, the sand was washed by 10% (v/v) hydrochloric acid, 0.5 M NaOH, and DI water three times to remove the metal ion contaminants on the sand surfaces. After rinsing thoroughly with DI water, the sand was dried in an Isotemp Oven (Fisher Scientific, Pittsburgh, PA) at 80 C. The column packed with clean sand was oriented vertically (see Figure S2 in the Supporting Information), and the influent bacterial suspension, with 0.3 mM sodium nitrate serving as the conservative tracer, was continuously injected into the column from the bottom. The interstitial velocity ranged from 1 m/d to 5 m/d and 10 m/d, for which, the volumetric flow rates controlled by an HPLC pump, Isocratic LC Pump 250 (PerkinElmer, Waltham, MA), were 0.04 mL/min, 0.2 mL/min, and 0.4 mL/min, respectively. Effluent samples were continuously collected from the top of the column in 1 mL aliquots. Both the bacterial and nitrate concentrations were quantified in terms of their optical density values measured by a spectrophotometer with settings at 590 and 220 nm wavelengths, respectively. Note that, prior to nitrate concentration measurement, bacteria in the samples were removed by filtration. The BTCs of effluent concentration profiles, which were normalized to influent bacterial concentration, were plotted versus eluted pore volume and compared with the results predicted by mathematical models. In order to avoid contamination of samples by residual bacteria inside the column, each experiment was conducted with new sand that was freshly prepared. 1-D Mathematical Model. Although the underlying mechanism governing bacterial surface association is different from reversible attachment, the phenomena as observed macroscopically appear similar; an individual bacterium spends a certain amount of time near the encountered solid surfaces. To mathematically describe these two bacterial surface interaction phenomena, an assumption of linear equilibrium between bacteria in the aqueous phase and bacteria associated with solid surfaces was adopted.27 A one-dimensional advection-dispersion transport equation21,27 R 3946

DC D2 C DC ¼ D 2  Vf  katt C Dt Dx Dx

ð2Þ

dx.doi.org/10.1021/es104041t |Environ. Sci. Technol. 2011, 45, 3945–3951

Environmental Science & Technology

ARTICLE

was applied to analyze the experimental data, in which, C is the normalized bacterial or tracer concentration, katt is the attachment rate coefficient (hr1), R is retardation factor, t is time (s), x is distance (cm), Vf is the interstitial velocity (cm/s), and D is the dispersion coefficient (cm2/s) as defined by μ0, ef f D ¼ RL V f þ ð3Þ ε where RL is the longitudinal dispersivity (cm), and μ0,eff/ε is the effective bacterial motility coefficient within porous media (cm2/s). The normalized bacterial and tracer concentrations were assumed to be unity at position x = 0 cm over the entire simulation time period. The model was solved with a finite difference numerical method by using MATLAB R2008a (The Mathworks, Inc., Natick, MA). The best fitted values for katt, R, and RL were determined by visual inspection and minimization of the sum of the squared errors of all data. According to colloid filtration theory (CFT), the singlecollector contact efficiency (η0) is a summation of three contributing factors: diffusion (ηD), interception (ηI), and sedimentation (ηG)19 according to η0 ¼ ηD þ ηI þ ηG

ð4Þ

in which, each component can be further expressed as follows:28 0:715 0:052 ηD ¼ 2:4AS NR0:081 NPe NvdW

ð5Þ

ηI ¼ 0:55AS NR1:675 NA0:125

ð6Þ

0:053 ηG ¼ 0:22NR0:24 NG1:11 NvdW

ð7Þ

1=3

where, AS = 2(1  γ5)/(2  3γ þ 3γ5  2γ6) = 1.55 is a porositydependent parameter where γ = (1  ε)3; NR is the aspect ratio of particle (bacterium) diameter (dp = 2rp = 1.5  106 m) to collector (sand grain) diameter (dc = 2rc = 2.8  104 m); NPe = Udc/D¥ is the Peclet number characterizing the ratio of convective transport to diffusive transport, in which, U represents fluid superficial velocity (m/s), D¥ represents the diffusion (motility) coefficient in an unbounded medium (m2/s), and corresponds to Brownian diffusion for a nonmotile species; NvdW= A/(kT) is the van der Waals number characterizing ratio of van der Waals interaction energy to the particle’s thermal energy, in which, A is the Hamaker constant (7  1021 J),29 k is the Boltzmann constant (1.3805  1023 J/K) and T is fluid absolute temperature (295K); NA= A/(12πμrp2U) is the attraction number representing combined influence of van der Waals attraction forces and fluid velocity on particle deposition rate due to interception, and μ is the viscosity of fluid (1  103 kg 3 m1s1); NG = 2rp2(Fp  Ff)g/(9μU) is the gravity number, the ratio of Stokes particle settling velocity to the approach velocity of the fluid, in which, Fp is bacterial density (1.1  103 kg/m3),29 Ff is fluid density (1.0  103 kg/m3), and g is the gravitational acceleration (9.81 m2/s). With the known physicochemical conditions of the sand column, the total single-collector contact efficiency (η0) was calculated, and the collision efficiency (R) was determined by the fitted irreversible attachment rate constant (katt) according to eq 1.

’ RESULTS AND DISCUSSION Impact of Flow Rate on Bacterial BTCs. The porosity of the sand packed column was 38% and the geometrical tortuosity (τp)

was 1.4. These values are similar to values reported in the literature from other column studies, ranging from 1.4 to 2.0.30 In Figure 1, experimental data from breakthrough curves (BTCs) for a conservative tracer and nonmotile bacteria are presented. Figure 2 shows the experimental data of BTCs for motile bacteria. The fitted transport parameters that were used as input to generate the model BTCs in Figures 1 and 2 are summarized in Table 1. As the BTCs displayed in Figure 1a show, the conservative tracer, sodium nitrate, achieved 100% recovery, and its BTCs were nearly identical over several different flow rates. An elastic collision was assumed for the interaction between tracer molecules and sand grains. Thus, the attachment rate coefficient (katt) was set to 0 h1. The resulting fitted retardation factor (R) and longitudinal dispersivity values (RL) were in the range of 0.96 to 1 and 0.05 to 0.08 cm, respectively. As each experiment was conducted in a newly packed sand column, this small variability in values was expected, given the inherent difficulty in reproducing the exact structure of the packed sand for each experiment. For the nonmotile flagella-less mutant E. coli HCB137 (Figure 1b), under 1 m/d interstitial velocity, its recovery reached 85% and increased to nearly 90% as expected at the faster flow rates of 5 m/d and 10 m/d. According to colloid filtration theory (CFT)19 and several colloid transport experiments,14,35 colloid recovery is expected to increase with increased fluid velocity as convection becomes the dominant factor for transport. As nonmotile bacteria may be modeled as colloidal particles, their transport in porous media was consistent with the predictions from CFT. Note that the fitted retardation factors (R) for E. coli HCB137 were smaller than the fitted values for conservative tracer. Unlike the conservative tracer molecule, which has an infinitely small size relative to the pore dimensions, a single bacterium is typically 12 μm in size. The larger size of a bacterium allows it to preferentially occupy the central regions of the pore where the fluid velocities are the greatest, whereas the smaller solute is able to diffuse close to the pore walls where the fluid velocities are slower. This phenomenon, described as hydrodynamic chromatography, can lead to the early breakthrough of bacteria.36,37 Although the nonmotile mutant E. coli HCB137 does not have flagella, the existence of flagella is not critical for initial attachment to solid surfaces.15 Thus, the transport patterns of nonmotile bacteria with flagella were expected to be the same as E. coli HCB137. The results in Figure 1c confirmed this expectation, as the BTCs of nonmotile tumbly mutant E. coli HCB359 and nonmotile paralyzed mutant E. coli HCB136 were in agreement with the model results generated for nonmotile flagella-less mutant E. coli HCB137 presented in Figure 1b. For both tracer and nonmotile bacteria, there was no obvious change in retardation or dispersion associated with their breakthrough curves over different flow velocities. However, this was not the case for motile bacteria. Consider the motile wild-type bacterial strain E. coli HCB1 (Figure 2a) for example, when the interstitial flow rate (Vf) was reduced from 10 m/d to 1 m/d, its BTCs gradually became more retarded and dispersed. Thus, as shown in Table 1, the best-fitted retardation (R) and longitudinal dispersivity (RL) values, increased from 1.08 to 1.26, and 0.18 to 0.47 cm, respectively. This trend appeared to be even more significant for the motile smooth-swimming mutant E. coli HCB437 (Figure 2b), as it has a longer bacterial idling time (tI) than the motile wild-type bacterial strain E. coli HCB1.7,10 3947

dx.doi.org/10.1021/es104041t |Environ. Sci. Technol. 2011, 45, 3945–3951

Environmental Science & Technology

Figure 1. (a) Conservative tracer, sodium nitrate, (b) nonmotile flagella-less mutant E. coli HCB137, (c) nonmotile tumbly mutant E. coli HCB359 and nonmotile paralyzed mutant E. coli HCB136 BTCs (exp) with 1-D transport fitting curves under 10 m/d (dotted line), 5 m/d (dashed line), and 1 m/d (solid line) interstitial flow rates. In panel c, the fitting curves are same as those in panel b. The error bars represent the range of repeated experimental results.

ARTICLE

Figure 2. (a) Motile wild-type E. coli HCB1, (b) motile smoothswimming mutant E. coli HCB437, and (c) motile wild-type P. putida F1 BTCs (exp data) with 1-D transport fitting curves under 10 m/d (dotted line), 5 m/d (dashed line), and 1 m/d (solid line) interstitial flow rates. The error bars represent the range of repeated experimental results. The larger error bars associated with the motile bacteria arise because of the variation in swimming speed and turn-angle distribution across individuals within the population and to some extent the variation from batch to batch of cultured bacteria. In panel (a) the differences between data points for 1 m/d and 5 m/d are statistically significant at pore volumes 1.55 (p < 0.1) and 1.77 (p < 0.05). 3948

dx.doi.org/10.1021/es104041t |Environ. Sci. Technol. 2011, 45, 3945–3951

Environmental Science & Technology

ARTICLE

Table 1. Transport Parameters for BTCs Obtained from Column Experiments E. coli HCB137

E. coli HCB1

E. coli HCB437

P. putida F1

3.2  105 a

2.0  108 b

3.8  106 c

2.0  105 d

3.2  106 e

Vf (m/d)

D0 or μ0 (cm2/s)

10

R

0.99

0.95

1.08

1.09

1.01

RL(cm)

0.08

0.11

0.18

0.24

0.05

katt(hr1)

0

0.49

5

1

a

tracer

0.49

0.54

0.70

η0

9.83  103

0.37

1.20

0.32

R

3.67  102

1.07  103

4.26  104

1.11  103

R

1

0.95

1.11

1.23

1.05

RL(cm)

0.07

0.11

0.17

0.40

0.10

katt(hr1)

0

0.23

0.29

0.27

0.29

η0

1.65  102

0.60

1.94

0.53

R

2.56  102

6.58  104

2.17  104

6.29  104

R

0.96

0.92

1.26

1.69

1.01

RL(cm) katt(hr1)

0.05 0

0.05 0.10

0.47 0.05

0.77 0.04

0.05 0.04

η0

6.11  102

1.95

6.36

1.73

R

1.28  102

1.85  104

5.21  105

1.92  104

Ref 31. b Ref 32. c Ref 33. d Ref 34. e See Supporting Information.

However, for P. putida F1, the differences among its BTCs under the three different flow rates were not as obvious. This result was consistent with the expected trend, because P. putida F1 exhibits relatively fast association and dissociation with solid surfaces,34 which means a smaller tI value compared to wild-type and smooth-swimming E. coli. A schematic to illustrate different bacterial swimming behaviors, which result in different tIvalues, is shown in Figure 4. Impact of Bacterial Idling Time on BTCs under Different Flow Rates. In order to emphasize the impact of bacterial idling time (tI) on bacterial migration profiles over different flow rates, in Figure 3ac, the bacterial BTCs from Figure 1 and Figure 2 are grouped according to the different experimental flow rates. As depicted in Figure 4, because the motile smooth-swimming mutant E. coli HCB437 is unable to change its swimming direction efficiently, it has the largest bacterial idling time (tI). For motile wild-type bacteria E. coli HCB1, which can easily reorient by executing a tumble, its tI is less than that of E. coli HCB437.7,10 Motile wild-type bacteria P. putida F138 can readily reverse direction by swimming backward, yielding an even smaller tI value. As the nonmotile bacterium is unable to swim, it does not associate with the solid surface for an extended period of time, and its tI is assumed to be zero. As shown in Figure 3, differences in the BTCs among the various bacterial strains were amplified as the flow rate decreased from 10 m/d to 5 m/d and to 1 m/d. When Vf was decreased to 5 m/d (Figure 3b), the motile smooth-swimming mutant E. coli HCB437, with the largest bacterial idling time (tI), displayed the most retarded and dispersed BTC. For species with lesser idling times such as motile wild-type E. coli HCB1 and P. putida F1, the degree of retardation and dispersion in their BTCs was correspondingly reduced, and the nonmotile flagella-less mutant E. coli HCB137, whose tI is zero, exhibited a minimal degree of retardation and dispersion. When Vf was decreased further to 1 m/d, motile bacterial BTCs exhibited even greater retardation and dispersion. The fitted retardation (R) and dispersivity (RL)

values were positively correlated with tI. Finally, if Vf was hypothetically reduced to zero, the highly impeded bacterial migration behavior similar to that observed in static granular porous media12,30 would be expected. The recovery reached approximately 90% for motile bacterial strains at all fluid velocities over the duration of the column experiments. At the lowest velocity of 1 m/d, the nonmotile strain was retained to a greater degree than the motile strains (85% recovery). A similar trend was also reported in a packed soil column study conducted by Camesano and Logan,14 in which, with the decrease of interstitial velocity from 120 m/d to 0.56 m/d, the fractional retention of nonmotile Pseudomonas fluorescens P17 increased by more than 800%, while for motile P17 the fractional retention decreased by 65%. This phenomenon implied that bacterial motility was one factor to prevent motile bacteria sticking to the solid surface. Fitted collision efficiency values (Table 1) decreased with decreasing fluid velocity for all species. Wild-type E. coli HCB1 and P. putida F1 exhibited nearly the same values for collision efficiencies, while values for the smooth-swimming mutant E. coli HCB437 were 24 times less than the wild-type over the range of velocities. In addition, collision efficiencies of all motile strains were smaller than those of their nonmotile counterpart, which indicated that there were less encounters between motile bacteria and the sand grains that resulted in collisions. A similar trend was also observed in the study by Camesano and Logan;14 at an interstitial velocity of 0.56 m/d, the collision efficiency for motile Pseudomonas fluorescens P17 (R = 0.003) was significantly smaller than that of nonmotile P17 (R = 0.018). McClaine and Ford15 also found that motile bacteria were less likely than nonmotile bacteria to attach to a surface under slow fluid velocities. Values for the single-collector contact efficiency (η0) (Table 6.1) exceeded unity for some motile bacteria, in particular for the smooth-swimming mutant E. coli HCB437. We believe this results from the extension of colloid filtration theory (CFT), which was originally intended for nonmotile colloids, to analyze 3949

dx.doi.org/10.1021/es104041t |Environ. Sci. Technol. 2011, 45, 3945–3951

Environmental Science & Technology

ARTICLE

Figure 4. Schematic illustration of (a) E. coli HCB137, (b) P. putida F1, (c) E. coli HCB1, and (c) E. coli HCB437 bacterial swimming behaviors in order of increasing bacterial idling time (tI).

In conclusion, bacterial idling time (tI) is one key factor contributing to greater retardation and dispersion in breakthrough curves of motile bacteria as the velocity approaches rates typical of groundwater flow. For the fast flow rates, when bacterial migration was dominated by convection over individual swimming, bacteria were most likely to travel along with the convective flow. However, under slow flow rates, especially when comparable with the average bacterial swimming speed, motile bacteria were able to approach the solid surfaces via its own motility. In those cases, the longer those bacteria were associated with the solid surfaces, the more retarded and dispersed their migration profile became. In comparison, nonmotile bacteria, which were unable to actively swim, did not associate with the solid surface even under the slow flow rate. As a result, the BTCs of nonmotile bacteria did not exhibit the same degree of retardation and dispersion. Although this investigation was conducted at the laboratory scale in a homogeneous system under well-controlled conditions, the findings from it suggest a plausible explanation for an interesting result that was reported previously from a field-scale study of bacterial migration through fractured crystalline bedrock.39 Becker et al.39 observed that nonmotile bacteria from the injectate cloud always transported to the recovery well location in advance of motile bacteria. Their result is consistent with our observations that nonmotile bacteria, which do not experience idling time near surfaces, migrate more readily through sand-packed columns with smaller retardation coefficients.

’ ASSOCIATED CONTENT

bS

Supporting Information. Some physical characteristics (size, zeta potential and swimming properties) of the bacterial strains are described for comparison. The column experimental system setup is provided in Figure S1, and the determination of P. putida F1 bacterial random motility coefficient is given in Figure S2. This information is available free of charge via the Internet at http://pubs.acs.org

’ AUTHOR INFORMATION Corresponding Author Figure 3. Different bacterial BTCs with 1-D transport fitting curves under different interstitial flow rates of (a) 10 m/d, (b) 5 m/d, and (c) 1 m/d. The error bars represent the range of repeated experimental results. In panel (c) differences between HCB1 and HCB137 are statistically significant at pore volumes of 1.11, 1.33, 1.55, and 1.77.

the behavior of motile bacteria. Swimming bacteria have diffusion coefficients that are about 3 orders of magnitude greater than diffusion coefficients for nonmotile bacteria, which impacts the calculation of the collector efficiency.

*Phone: (434) 924-6283; fax: (434) 982-2865; e-mail: rmf3f@ virginia.edu.

’ ACKNOWLEDGMENT We gratefully acknowledge support from the National Science Foundation Hydrological Sciences Program (EAR 0711377). The assistance from Ms. Sarah Triolo with sample collection and analysis for the column experiments is greatly appreciated. We also appreciate the reviewers’ constructive comments, which improved the manuscript. 3950

dx.doi.org/10.1021/es104041t |Environ. Sci. Technol. 2011, 45, 3945–3951

Environmental Science & Technology

’ REFERENCES (1) Scow, K. M.; Hicks, K. A. Natural attenuation and enhanced bioremediation of organic contaminants in groundwater. Curr. Opin. Biotechnol. 2005, 16, 246–253. (2) Christ, J. A.; Ramsburg, C. A.; Abriola, L. M.; Pennell, K. D.; Loffler, F. E. Coupling aggressive mass removal with microbial reductive dechlorination for remediation of DNAPL source zones: a review and assessment. Environ. Health Perspect. 2005, 113, 465–477. (3) Costerton, J. W.; Cheng, K. J.; Geesey, G. G.; Ladd, T. I.; Nickel, J. C.; Dasgupta, M.; Marrie, T. J. Bacterial biofilms in nature and disease. Annu. Rev. Microbiol. 1987, 41, 435–464. (4) Ford, R. M.; Harvey, R. W. Role of chemotaxis in the transport of bacteria through saturated porous media. Adv. Water Resour. 2007, 30, 1608–1617. (5) Wolfe, A. J.; Berg, H. C. Migration of bacteria in semisolid agar. Proc. Natl. Acad. Sci. U. S. A. 1989, 86, 6973–6977. (6) Kusy, K.; Ford, R. M. Monte carlo simulations derived from direct observations of individual bacteria inform macroscopic migration models at granular porous media interfaces. Environ. Sci. Technol. 2007, 41, 6403–6409. (7) Vigeant, M. A. S.; Ford, R. M. Interactions between motile Escherichia coli and glass in media with various ionic strengths, as observed with a three-dimensional-tracking microscope. Appl. Environ. Microbiol. 1997, 63, 3474–3479. (8) Frymier, P. D.; Ford, R. M.; Berg, H. C.; Cummings, P. T. Threedimensional tracking of motile bacteria near a solid planar surface. Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 6195–6199. (9) Narayanaswamy, K.; Ford, R. M.; Smith, J. A.; Fernandez, E. J. (2009), Surface association of motile bacteria and apparent tortuosity values in packed column experiments. Water Resour. Res., 2009, 45, DOI: 10.1029/2008WR006851. (10) Liu, J.; Ford, R. M. Idling time of swimming bacteria near particulate surfaces contributes to apparent adsorption coefficients at the macroscopic scale under static conditions. Environ. Sci. Technol. 2009, 43, 8874–8880. (11) Kusy, K.; Ford, R. M. Surface association of motile bacteria at granular porous media interfaces. Environ. Sci. Technol. 2009, 43, 3712–3719. (12) Sherwood, J. L.; Sung, J. C.; Ford, R. M.; Fernandez, E. J.; Maneval, J. E.; Smith, J. A. Analysis of bacterial random motility in a porous medium using magnetic resonance imaging and immunomagnetic labeling. Environ. Sci. Technol. 2003, 37, 781–785. (13) Olson, M. S.; Ford, R. M.; Smith, J. A.; Fernandez, E. J. Ouantification of bacterial chemotaxis in porous media using magnetic resonance imaging. Environ. Sci. Technol. 2004, 38, 3864–3870. (14) Camesano, T. A.; Logan, B. E. Influence of fluid velocity and cell concentration on the transport of motile and nonmotile bacteria in porous media. Environ. Sci. Technol. 1998, 32, 1699–1708. (15) McClaine, J. W.; Ford, R. M. Characterizing the adhesion of motile and nonmotile escherichia coli to a glass surface using a parallelplate flow chamber. Biotechnol. Bioeng. 2002, 78, 179–189. (16) Marshall, K. C.; Stout, R.; Mitchell, R. Mechanism of the initial events in the sorption of marine bacteria to surfaces. J. Gen. Microbiol. 1971, 68, 337–348. (17) Vigeant, M. A. S.; Ford, R. M.; Wagner, M.; Tamm, L. K. Reversible and irreversible adhesion of motile escherichia coli cells analyzed by total internal reflection aqueous fluorescence microscopy. Appl. Environ. Microbiol. 2002, 68, 2794–2801. (18) Berke, A. P.; Turner, L.; Berg, H. C.; Lauga, E. Hydrodynamic attraction of swimming microorganisms by surfaces. Phys. Rev. Lett. 2008, 101. (19) Yao, K. M.; Habibian, M. T.; Omelia, C. R. Water and waste water filtration: Concepts and applications. Environ. Sci. Technol. 1971, 5, 1105–1112. (20) Brown, D. G.; Abramson, A. Collision efficiency distribution of a bacterial suspension flowing through porous media and implications for field-scale transport. Water Res. 2006, 40, 1591–1598.

ARTICLE

(21) Tufenkji, N. Modeling microbial transport in porous media: Traditional approaches and recent developments. Adv. Water Resour. 2007, 30, 1455–1469. (22) Wolfe, A. J.; Conley, M. P.; Kramer, T. J.; Berg, H. C. Reconstitution of signaling in bacterial chemotaxis. J. Bacteriol. 1987, 169, 1878–1885. (23) Parales, R. E.; Ditty, J. L.; Harwood, C. S. Toluene-degrading bacteria are chemotactic towards the environmental pollutants benzene, toluene, and trichloroethylene. Appl. Environ. Microbiol. 2000, 66, 4098–4104. (24) Barton, J. W.; Ford, R. M. Determination of effective transport coefficients for bacterial migration in sand columns. Appl. Environ. Microbiol. 1995, 61, 3329–3335. (25) Valdes-Parada, F. J.; Porter, M. L.; Narayanaswamy, K.; Ford, R. M.; Wood, B. D. Upscaling microbial chemotaxis in porous media. Adv. Water Resour. 2009, 32, 1413–1428. (26) Wang, M. Bacterial chemotaxis in porous media. Ph.D. dissertation, University of Virginia: Charlottesville, VA, 2009. (27) Harvey, R. W.; Garabedian, S. P. Use of colloid filtration theory in modeling movement of bacteria through a contaminated sandy aquifer. Environ. Sci. Technol. 1991, 25, 178–185. (28) Tufenkji, N.; Elimelech, M. Correlation equation for predicting single-collector efficiency in physicochemical filtration in saturated porous media. Environ. Sci. Technol. 2004, 38, 529–536. (29) Simoni, S. F.; Harms, H.; Bosma, T. N. P.; Zehnder, A. J. B. Population heterogeneity affects transport of bacteria through sand columns at low flow rates. Environ. Sci. Technol. 1998, 32, 2100–2105. (30) Olson, M. S.; Ford, R. M.; Smith, J. A.; Fernandez, E. J. Analysis of column tortuosity for MnCl2 and bacterial diffusion using magnetic resonance imaging. Environ. Sci. Technol. 2005, 39, 149–154. (31) Cussler, E. L. Diffusion: Mass Transfer in Fluid Systems; 2nd ed.; Cambridge University Press: New York, NY, 1997. (32) Liu, J. Bacterial Surface Association in Porous Media. M.S. Thesis, University of Virginia: Charlottesville, VA, 2008. (33) Wu, M. M.; Roberts, J. W.; Kim, S.; Koch, D. L.; DeLisa, M. P. Collective bacterial dynamics revealed using a three-dimensional population-scale defocused particle tracking technique. Appl. Environ. Microbiol. 2006, 72, 4987–4994. (34) Kusy, K. Bacterial Migration within Heterogeneous Porous Media. Ph.D. dissertation, University of Virginia: Charlottesville, VA, 2005. (35) Close, M. E.; Pang, L.; Flintoft, M. J.; Sinton, L. W. Distance and flow effects on microsphere transport in a large gravel column. J. Environ. Qual. 2006, 35, 1204–1212. (36) Harvey, R. W.; George, L. H.; Smith, R. L.; Leblanc, D. R. Transport of microspheres and indigenous bacteria through a sandy aquifer: results of natural- and forced-gradient tracer experiments. Environ. Sci. Technol. 1989, 23, 51–56. (37) Camper, A. K.; Hayes, J. T.; Sturman, P. J.; Jones, W. L.; Cunningham, A. B. Effects of motility and adsorption rate coefficient on transport of bacteria through saturated porous media. Appl. Environ. Microbiol. 1993, 59, 3455–3462. (38) Harwood, C. S.; Fosnaugh, K.; Dispensa, M. Flagellation of Pseudomonas putida and analysis of its motile behavior. J. Bacteriol. 1989, 171, 4063–4066. (39) Becker, M. W.; Metge, D. W.; Collins, S. A.; Shapiro, A. M.; Harvey, R. W. Bacterial transport experiments in fractured crystalline bedrock. Ground Water 2003, 41, 682–689.

3951

dx.doi.org/10.1021/es104041t |Environ. Sci. Technol. 2011, 45, 3945–3951