This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rees, H. C.
Right arrow Articles by Lerner, D. N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Rees, H. C.
Right arrow Articles by Lerner, D. N.
Agricola
Right arrow Articles by Rees, H. C.
Right arrow Articles by Lerner, D. N.

 Previous Article  |  Next Article 

Applied and Environmental Microbiology, June 2007, p. 3865-3876, Vol. 73, No. 12
0099-2240/07/$08.00+0     doi:10.1128/AEM.02933-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Biodegradation Processes in a Laboratory-Scale Groundwater Contaminant Plume Assessed by Fluorescence Imaging and Microbial Analysis{triangledown}

Helen C. Rees,1* Sascha E. Oswald,2 Steven A. Banwart,1 Roger W. Pickup,3 and David N. Lerner1

Department of Civil and Structural Engineering, Kroto Research Institute, North Campus, University of Sheffield, Broad Lane, Sheffield S3 7HQ, United Kingdom,1 UFZ, Helmholtz Centre for Environmental Research, Department of Hydrogeology, Permoserstrasse 15, 04318 Leipzig, Germany,2 Centre for Ecology and Hydrology, Lancaster Environment Centre, Library Avenue, Bailrigg, Lancaster LA1 4AP, United Kingdom3

Received 19 December 2006/ Accepted 19 April 2007


arrow
ABSTRACT
 
Flow reactors containing quartz sand colonized with biofilm were set up as physical model aquifers to allow degrading plumes of acetate or phenol to be formed from a point source. A noninvasive fluorescent tracer technique was combined with chemical and biological sampling in order to quantify transport and biodegradation processes. Chemical analysis of samples showed a substantial decrease in carbon concentration between the injection and outflow resulting primarily from dilution but also from biodegradation. Two-dimensional imaging of the aqueous oxygen [O2(aq)] concentration field quantified the depletion of O2(aq) within the contaminant plume and provided evidence for microbial respiration associated with biodegradation of the carbon source. Combined microbiological, chemical, and O2(aq) imaging data indicated that biodegradation was greatest at the plume fringe. DNA profiles of bacterial communities were assessed by temperature gradient gel electrophoresis, which revealed that diversity was limited and that community changes observed depended on the carbon source used. Spatial variation in activity within the plume could be quantitatively accounted for by the changes observed in active cell numbers rather than differences in community structure, the total biomass present, or the increased enzyme activity of individual cells. Numerical simulations and comparisons with the experimental data were used to test conceptual models of plume processes. Results demonstrated that plume behavior was best described by growth and decay of active biomass as a single functional group of organisms represented by active cell counts.


arrow
INTRODUCTION
 
Groundwater contamination by organic pollutants is a legacy of historic industrial activity and a threat to future resources. The duration, cost, and uncertain success of pumping out dissolved contamination is a daunting obstacle, both to the protection of groundwater and to the recycling and redevelopment of industrially contaminated land. Monitored natural attenuation (MNA) is an alternative remediation strategy. MNA reduces the chemical risk of the contamination through naturally occurring, in situ processes. Most importantly, biodegradation through the activity of indigenous microbial populations reduces risk by destroying contamination mass in the aquifer.

MNA is a much more sustainable approach with far less reliance on continuous energy input for pumping or other removal (5). The in situ destruction of contamination reduces human exposure to the contamination and eliminates the need for subsequent disposal or treatment of the contaminated water as a hazardous material. Conceptually, MNA is an application of environmental biotechnology that considers the groundwater aquifer an in situ biological reactor. Mathematical modeling of the reactor system is thus required to aid in interpreting field data sets that describe reactor behavior and for predictions of future reactor performance to meet specific remediation objectives.

At a process level, biological oxidation of organic pollutants is limited by the supply of electron acceptors (EAs) in the subsurface (7, 23) and is strongly influenced by their concentration distribution within the plume. Spatial patterns of EA concentrations usually differ from those of the pristine aquifer (6). A plume of dissolved organic contaminants leaking from a point source will develop strongly reducing conditions, with depletion of EAs near the source and within the plume core downstream of the source area. EA concentrations along transverse gradients at the plume fringes will increase from depleted levels in the core and approach those of the pristine aquifer. In this fringe region, conditions for biodegradation are particularly favorable as electron acceptors are constantly being replenished by dispersive mixing from outside the plume, mirrored by a dispersive mixing of contaminants from the zones of higher concentration in the plume core (17).

Interpretation of field observations and evaluation of plume development require knowledge of the redox conditions, i.e., identification of EAs, organic carbon solutes, their concentration patterns, resulting net oxidation capacities, and relative rates of supply. Underpinning this is a requirement to understand transverse mixing of EAs in relation to microbial processes. This is because rates of biodegradation through microbial respiration, and thus, MNA performance, will be limited by the rate of electron acceptor supply at the plume fringe (33). Dilution of contaminants in this region can also increase biodegradation rates for organic compounds that exhibit substrate toxicity to the microbial populations (37).

The use of field data to develop improved conceptual understanding and mathematical models of coupled transport and microbial processes has inherent limitations. This is due primarily to the high cost of field experimentation and to the generally sparse data sets that result when set against the enormous complexity of the natural aquifer system. This complexity includes the spatial variability in physical and chemical conditions, unknown microbial species and their physiology, and the complex ecological interactions between microorganisms and the response to in situ conditions. Due to these challenges, advances in understanding microbial processes require laboratory-based methods that are controlled, reproducible, and representative of key system behaviors at the field scale. This study presents for the first time microbiological observations from a physical-scale model aquifer that demonstrates high reproducibility in flow, transport, and biological processes. This is demonstrated through observation of the organic carbon source and the electron acceptor concentration fields and quantitative assessment of microbiological activity.

Solute transport in porous media has been studied using tracer imaging (8, 16) including UV-excited fluorescent solutes (15). To determine values for physical parameters, such as dispersivity, Huang et al. (14) investigated biodegradation during solute transport in a laboratory plume with limited O2 supply, using a noninvasive imaging technique with a tracer whose fluorescence under UV excitation is quenched by dissolved O2. In this study, we have applied this fluorescent tracer technique for laboratory-scale physical model contaminant plumes in order to address several challenges to the application of MNA at field scale. These challenges are (i) to assess the mass balance for the degrading organic substrate by using sparse point estimates of concentration within a plume; (ii) to improve current conceptualization of biodegradation processes by observing patterns of biomass, respiration activity, and relative biodiversity; (iii) to constrain mathematical models of plumes with electron acceptor (O2) concentrations obtained at high spatial resolution by fluorescent imaging; and (iv) to test conceptualizations of biodegradation processes and their mathematical descriptions by comparing measurements with numerical simulation results.

The novelty in this approach is the primary emphasis on spatial patterns of microbial biomass and activity. Furthermore, due to methodological constraints, microbiology in many contexts has been viewed as a black box (9, 25, 35). This is no longer true, as molecular microbiological techniques now permit the structure of microbial communities to be described along with their activities (9). This is demonstrated in a civil engineering context, particularly in studies of bioreactors, bioremediation, and wastewater treatment (31, 32, 36). The development of laboratory bench-scale reactors, described here as model aquifers, is an initial step for the microbiologist to provide meaningful data to the groundwater engineer. Only when this is combined with essential physical and chemical information will contaminant plume behavior be fully described and understood. The overarching aim of this work is, thus, to more clearly identify the essential data required to better predict plume behavior in the performance assessment of MNA in a microbiological, hydrological, and chemical perspective.


arrow
MATERIALS AND METHODS
 
Growth medium and inoculum preparation.
The basal mineral medium used in all experiments consisted of 10 ml of a pH 7.0 buffer solution containing 10 mM NaH2PO4 and 10 mM Na2HPO4; NaCl, 0.017 g liter–1; NaHCO3, 0.014 g liter–1; MgSO4, 0.013 g liter–1; NaNO3 0.069 g liter–1 dissolved in 1 liter of ultra-high-quality water and autoclaved at 121°C for 20 min. Inocula for the experimental systems were obtained from the Four Ashes site (United Kingdom) and comprised phenol-contaminated groundwater pumped from borehole MW7 (16). MW7 borehole water (1.5 ml) was inoculated into 150 ml mineral medium supplemented with 250 µg liter–1 acetate or 200 µg liter–1 phenol and incubated for approximately 65 h at room temperature with shaking (200 rpm). These experiments required a microbial community to be established on the porous matrix of high-quality quartz sand (212- to 300-µm grain size) which was sterilized by autoclaving before use. This was achieved as described by Li et al. (18), with the cultured inocula from either the acetate or phenol enrichment in a sterile glass column filled with the sterile quartz sand/mineral medium slurry.

Setup and operation of the flow cell.
The flow cell was based on that described previously by Huang et al. (14, 15) and modified as described below (Fig. 1A and B). The flow cell was operated in the dark to minimize interference by other light sources and was sterilized with ethanol prior to use. A Perspex flow cell (internal dimensions of 200 by 100 by 5 mm) with a quartz glass plate on the front side for visualization was filled with the inoculum established as a quartz sand/mineral medium slurry through an inlet pipe located on the back of the flow cell. The slurry was mixed inside the flow cell to prevent layering. A sterile stainless steel mesh filter prevented the loss of the quartz sand through the outlet pipes. The average porosity of the matrix was determined from the reactor volume and density and mass of quartz sand to be 0.45. A point source injection pipe (internal diameter of 3 mm) with perforations in the flow direction was located centrally through the internal space of the flow cell, 20 mm from the top. Septa ports were arranged in the back of the flow cell to allow for aqueous sampling both across a transect and along the plume in the flow direction.


Figure 1
View larger version (16K):
[in this window]
[in a new window]

  		FIG. 1. Design and operation of the flow cell. (A) Schematic of the flow cell showing the locations of various ports in the back plate. (B) Experimental setup. Dotted arrows indicate the direction of flow, solid arrows indicate the UV illumination, and dashed arrows represent transmitted light from the fluorescent tracer dye in the flow cell. Images are recorded by a charge-coupled-device (CCD) camera and shown on a computer screen.

A constant flow rate of 0.13 ml min–1 was maintained by using a peristaltic pump and a syringe pump (model 2400003; Harvard Apparatus, Kent, United Kingdom) to inject the mineral medium plus carbon source at a constant flow rate of 8.6 µl min–1. Oxygen concentrations were visualized using the complex ruthenium(II)-dichlorotris(1,10-phenanthroline) [Ru(phen)3Cl2] (Sigma, Dorset, United Kingdom) at a concentration of 1.0 x 10–4 M in the mineral medium. Under UV illumination, the dissolved complex emits fluorescent light (12), which is a function of the concentration of dissolved O2 (4).

Two sets of experiments were performed in triplicate flow cells. The first experiment used acetate (335 mg liter–1) as the carbon source and was run for 24 h, and the second experiment used phenol as the carbon source (164 mg liter–1) as an environmental contaminant, to which the source inoculum had been previously exposed, and was run for 24 h. A further experiment to test nonreactive transport used the inert tracer fluorescein to image the flow field and the fluorescein plume formation in sterile control experiments.

Imaging of plume formation.
Time-lapse images of the plume were taken using a charge-coupled-device camera (color KP-D581; Hitachi Kokusai, Leeds, United Kingdom) with a UV long-pass filter. Images were visualized using Global Lab Imaging software version 2 (Data Translation, Marlboro, MA) via an image grabber board (Data Translation, Marlboro, MA). These images were then transferred into Image J (public domain software) for analysis.

Dissolved O2 calibration.
Mineral medium solutions containing Ru(phen)3Cl2 and different concentrations of dissolved O2 (0, 2, 4, 6, and 8.5 mg/liter) were passed through the flow cells prior to experiments and were imaged as described above. The average fluorescence intensity was obtained for each image using Image J software and was plotted as I0/I (where I0 is the initial pixel intensity at zero O2 over I, the pixel intensity at the standard concentration) versus standard O2 concentration, with the resulting linear equation generated used to calibrate O2 concentration against tracer fluorescence intensity, as described by Huang et al. (14).

Chemical analyses.
Aqueous samples (100 µl) were taken via the septa ports in the back of the flow cells, diluted (1 in 25 for acetate and 1 in 15 for phenol), and then filtered. Acetate was measured by ion chromatography using a Dionex DX-120 ion chromatograph (Dionex, Sunnyvale, CA), and phenol was measured by high-performance liquid chromatography (HPLC) using a Perkin Elmer series 200 HPLC system (Perkin Elmer, Wellesley, MA). The O2 and pH levels in the inflow, outflow, and injection solutions were measured using a calibrated O2 probe (model O20; WPA, Cambridge, United Kingdom) and a pH meter with combined glass electrode (model HI9321; Hannah Instruments, Leighton Buzzard, United Kingdom), respectively.

Biological analysis.
Destructive sampling was used to take five samples from a transect of the plume and a further sample from the source area close to the injection pipe (Fig. 1A). Metabolically active cells and total counts were visualized using 5-cyano-2,3-ditolyl tetrazolium chloride (CTC) staining (30) and 4'6-diamidino-2-phenylindole (DAPI) staining (27), respectively. Samples removed from the flow cell comprised 0.2 g of sand that was added to 0.2 ml of physiological saline solution, sonicated for 40 min (Decon FS100 sonic bath; Ultrasonics, Hove, United Kingdom), with the resulting suspension removed and 10 µl of a 10 mM CTC solution added before incubation overnight in the dark.

Samples were diluted as necessary to visualize single cells, and 10-µl aliquots were spotted onto diagnostic slides, mounted with Vecta Shield containing DAPI (Vector Laboratories Ltd., Peterborough, United Kingdom). At least 10 fields of view per sample were counted with a Zeiss Axioplan 2 imaging epifluorescence microscope. CTC- or DAPI-stained cells were visualized in the same frame by changing the wavelength of excitation of the fluorescence microscope. A further 0.2 g of each sample was used for total protein analysis using a modified Lowry assay (20), as described by the manufacturer (Sigma-Aldrich, Dorset, United Kingdom). The sand recovered after sonication was analyzed for protein content to assess the efficiency of sonication to disperse attached cells into the liquid supernatant.

PCR-temperature gradient gel electrophoresis.
Sand removed from the flow cell (0.2 g) was used for PCR-temperature gradient gel electrophoresis (TGGE) analysis. 16S rRNA gene amplicons were generated with a Biometra PCR thermocycler (Biometra, Goettingen, Germany) following two rounds of PCR. The first round, using primers GM5F (Escherichia coli nucleotide sequence positions 341 to 357 [21]) and rP1 (E. coli, positions 1512 to 1492 [29]), was carried out as follows: 96°C for 5 min, 30 cycles of 1 min at 96°C, 1 min at 55°C, and 1 min at 72°C, and a final extension of 10 min at 72°C. The second-round PCR was carried out using the primers GM5F-gC, consisting of the primer GM5F with a 5'-end 40-bp GC clamp (21), and rD2 (E. coli, positions 536 to 519 [29]), using "touchdown" PCR (10) where the annealing temperature of the program above was dropped by 1°C for every 2 cycles, from 65 to 55°C, with a further 10 cycles at 55°C. PCR products were purified and concentrated using QIAGEN quick-spin columns (QIAGEN Ltd., Crawley, United Kingdom) and analyzed by TGGE on a 12% acrylamide gel with a temperature gradient of 45 to 60°C at a constant voltage of 120V for 3 h. Gels were stained with ethidium bromide in running buffer, and bands were visualized under ultraviolet illumination with a Canon G3 digital camera, followed by computer-based analysis performed by a BioDocAnalyse gel documentation system (Whatman Biometra, Goettingen, Germany).

Reactive transport modeling.
The flow cell experiments were modeled mathematically. Conceptualizing the inlet pipe as a line source (14), the following analytical solution, based on the error function erf, was applied to calculate the concentration of nonreactive tracers (i.e., transport in the experiments with sterile conditions).

Formula 1(1)

Formula 2(2)
Here x and y are the Cartesian coordinates (in millimeters); C is the concentration and C0 is the source concentration (mg liter–1); Y is the length of the source in the transverse y direction, centered at the origin (mm); v is the pore water velocity oriented in the x direction as a uniform flow field (mm s–1); DT is the transverse dispersion coefficients (mm2 s–1); aT is transverse dispersivity (mm); and D* is the effective molecular diffusion coefficient (mm2 s–1). This formula was calculated for a number of x and y positions, using Microsoft Excel, inserting the different values of D* (0.0003 mm2 s–1 for fluorescein) (Table 1, phenol and acetate values).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Summary of transport and kinetic biodegradation parameters

Aerobic biodegradation kinetics were described by a dual Monod model, which has been shown to be more appropriate than an instantaneous degradation model for describing O2 consumption in the type of flow cell experiments described here (14). The reaction rate term in the advection-dispersion equation for transport is then

Formula 3(3)

Formula 4(4)

Formula 5(5)
where O is oxygen concentration (mg liter–1), X is the total microbial concentration (mg dry cell weight liter–1), µm is the maximum specific growth rate (s–1), YX/C is the growth yield for the carbon source (mg dry cell weight gained/mg substrate consumed), F is the stoichiometric ratio of oxygen to substrate in the degradation reaction (mg oxygen consumed/mg substrate consumed), KED is the half-saturation constant for the carbon source (mg substrate liter–1), KO is the half-saturation constant for oxygen (mg oxygen liter–1), and d is the death rate for the microbial population X (s–1). The numerical modeling was carried out using Visual Modflow 3.0, applying RT3D2.5 as a module to simulate the reactive transport using the dual Monod equations as defined above. The numerical solution method was based on a mixed method of characteristics/hybrid method of characteristics scheme with implicit solver, and the grid size in the central part of the plume was chosen as 1 mm by 0.5 mm. It should be noted that RT3D does not require the maximum specific growth rate, µm, as the input value but uses a maximum substrate utilization rate, kmax, as given at the right sides of equations 3 to 5.

Accompanying batch and flow cell experiments.
Batch experiments were performed according to Huang et al. (14) to determine independently the biological parameters (for µm, KO, KC, F, and Yx/c) needed for numerical simulations for biodegradation of both acetate and phenol as carbon sources (see Table 1 and equations 3 to 5). Furthermore, additional flow cell experiments under sterile conditions were used to assess nonreactive transport behavior and to obtain the relevant transport parameters as previously described (14, 15). These experiments used fluorescein, acetate, or phenol as a tracer in the same flow cell but without microbes present. In this case a steady-state plume was established after 24 h (equivalent to six flow cell volume changes).


arrow
RESULTS AND DISCUSSION
 
Impact of microbial activity. (i) Microbial analysis.
Results from protein analysis carried out on sand samples (Table 2) illustrate the variation in total biomass across the flow cells at time zero. There was no significant change in biomass with location or time through the experiments, within the precision of measurements of replicate reactors. Total (DAPI-stained) and active (CTC-stained) cell counts showed that active cells typically comprised less than 5% of the total cell counts (Table 2). There were significantly higher activities (CTC cell counts) within the source and fringe areas of the acetate plume compared to that of the plume core. These results for acetate quantitatively support the current conceptual model of plume biodegradation in that the activity is greatest where contaminants and electron acceptors coexist in sufficient supply (at the plume fringes and in the source area).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Summary of protein concentration and active cell counts as a percentage of the total number of cells

For the phenol plume, the point estimates of activity also reflected this spatial pattern of increased activity at the fringe and source, but the observed differences with the plume core were not statistically significant. For all plumes there was no significant difference between activity in the source area and that in the plume fringe compared to background activity outside the plume (Table 2). The activity levels found were similar to levels found in low-nutrient environments, such as groundwater and seawater (30).

(ii) Substrate mass loss.
Chemical analysis of fluid samples taken from flow cells at the end of the experimental run showed that there was a substantial decrease in concentration of carbon between the injection well and the outflow (Table 3). This decrease results from both biodegradation and dilution across the outflow chamber. A less pronounced decrease was observed for sterile experiments, due only to dilution. However, the differences between the active and sterile systems were not statistically significant and, thus, do not offer direct evidence of biodegradation. Even under the controlled and reproducible conditions of these laboratory experiments, the extent of biodegradation is not possible to conclusively demonstrate by analysis of data for organic carbon loss alone. Numerical simulations demonstrated that after 24 h, only about 10% of the injected carbon source will be consumed, in our experiments (data not shown), indicating that the determination of a degradation rate would require an extremely accurate carbon mass balance. This result also demonstrates challenges to obtaining accurate carbon mass balance at the field scale where the organic carbon load is much higher than the oxidant load of aqueous O2 and other soluble EAs in the groundwater. This does not reflect a low potential for MNA because as a plume grows, the interfacial area for dispersive mixing of organic solutes and EAs grows until the contaminant mass loss by degradation is equal to the contaminant loading at the source. However, at any discrete sampling location, the decrease in organic carbon concentration compared to that in the absence of biodegradation may be within the uncertainty of sampling and measurement, thus requiring other analysis such as electron balances, as described below (3).


View this table:
[in this window]
[in a new window]

  		TABLE 3. Summary of experimental parameters in MW7 water-inoculated quartz sand under saturated conditionsa

The relatively small mass loss of substrate and the estimation of contaminant mass loss must rely on the electron balance. This approach infers organic carbon mass loss by analyzing the corresponding loss of EAs. In this study, strong evidence for biodegradation exists, since even small amounts of organic substrate oxidation result in a significant depletion of dissolved O2. This is because gaseous O2 is relatively insoluble in water and exhibits low concentration of electron equivalents compared to that in the organic substrate. At field sites (3), decreases in EAs, such as NO3– and SO42–, can also be observed, as well as increases in dissolved Fe(II), Mn(II), and CH4, as evidence of other respiration processes. For our flow reactors, the biodegradation process is reflected by the consumption of dissolved O2 as the sole electron acceptor present.

(iii) O2 distribution in the flow cell.
Dissolved O2 data, which also defined the extent of the reactive plume, was obtained using the fluorescent Ru(phen)3Cl2 O2 tracer. Fluorescence intensity was converted quantitatively to O2 concentration by using the I0/I calibration plot for each flow cell. All flow cells challenged with a carbon source showed O2 depletion between the inflow and outflow solutions (Table 3), with a clear depression of O2 concentration in the plume core. However, none of the flow cells showed complete O2 depletion in the core of the plume. This could be due to a number of reasons, including uncertain values measured within the calibration error (note error bars in Fig. 2), limited microbial growth, or competitive inhibition of microaerophilic populations by aerobes during preparation of growth culture inoculum.


Figure 2
View larger version (24K):
[in this window]
[in a new window]

  		FIG. 2. Experimental oxygen transect profile plots for all three flow cells from the 24-h phenol experiment. The positions of all three plumes are similar. Error bars represent the standard deviations of 10 transects from each of the flow cells between 98 and 102.5 mm from the carbon source injection point, which increase away from the plume core. Also shown are the equivalent numerical simulations with the most advanced conceptual model.

Figure 2 shows the O2 profiles across the plumes in the three repeats of the 24-h phenol experiment. The O2 distribution data for transects across the plume were compared for triplicate runs of the two 24-h experiments. The 24-h acetate flow cell shows a similar profile (data not presented). The positions of the plumes were similar but not precisely identical in the triplicate flow cells. In all cases, the O2 concentration increased from the core to the fringe of the plume. These results are similar to those previously obtained in other laboratory physical-scale models of biodegrading groundwater contaminant plumes (14, 15). Most importantly, the depletion of O2 within the contaminant plume provides strong evidence of microbial respiration associated with biodegradation of the organic carbon substrate.

(iv) Numerical simulations.
Batch experiments and the modeling of the sterile flow cell experiments were carried out as described in references 14 and 15. A numerical model was generated that represented the reactive flow cell experiments. The modeling parameters, except for the yield coefficients, were determined experimentally, either from the flow cell data, independently in batch experiments, or from the sterile flow cell runs (Table 1). A broad range of yield coefficients based on published values was considered (13, 28), as the values are known to be variable between populations and also during degradation processes. The choice of yield coefficient value also affects the stoichiometric ratio F (equation 4) for oxygen consumption. This is because it is subject to variation, as carbon is often converted to biomass rather than being subjected to complete mineralization. Variable yield coefficients thus correspond to the variable consumption of O2, with smaller values of F (equation 4) for larger yield coefficients. In the case of phenol, the stoichiometric ratio was calculated based on the theoretical reaction equation; for example, taking the average yield coefficient, F is 1.74 mg of O2 consumed per mg of phenol degraded (see equation 6).

Formula 6(6)
For the acetate degradation, however, the experimentally determined stoichiometric ratio from batch growth experiments was used for all simulations.

(a) Nonreactive transport of the organic substrate.
Based on sterile imaging experiments with fluorescein tracer, the value for transverse dispersivity was determined by matching calculated and measured solute tracer concentrations on various transects, using equations 1 and 2 (Fig. 3). The resulting value of 0.04 mm compares well with a previous estimate of 0.05 mm in a similar scale model plume (15). This value implies that transverse mixing in the experiment is driven by molecular diffusion and by mechanical transverse dispersion in roughly equal proportions. The concentration field resulting from nonreactive transport of acetate and phenol was calculated using the appropriate effective diffusion coefficients in equation 2 (Fig. 3). Due to their larger diffusion coefficients, the transverse spreading of organic substrates is slightly more pronounced, and the maximum concentrations, on the centerline, are slightly reduced relative to that of fluorescein.


Figure 3
View larger version (12K):
[in this window]
[in a new window]

  		FIG. 3. Transverse concentration profile of the conservative fluorescent tracer. The transverse concentration profile of conservative fluorescent tracer (fluorescein) at x = 78 mm down gradient of the tracer injection point; comparison of concentrations imaged and concentration calculated via analytical solution (cf. equation 2) using a transverse dispersivity of 0.04 mm. Also shown are the calculated concentration profiles of phenol and acetate under nonreactive conditions and their concentrations measured in septa on this profile.

(b) Dissolved O2 consumption by biodegradation.
Four conceptual models (CM) of biodegradation were simulated to compare with the flow cell experiments, starting with the simplest one, published previously by Huang et al. (14). Model complexity was then increased stepwise to reflect the patterns observed for the microbial activity reported above. In each case, all parameter values for microbial growth, death, and substrate utilization were taken as determined independently in batch experiments and were not adjusted. (i) In the Total Biomass CM, the total of the measured biomass concentration participates in the degradation process and was taken to be constant. (ii) In the Active Biomass CM, the measured activity (Table 2, time zero values) was used to quantitatively identify the part of the total microbial biomass that actually participates in the degradation and was taken to be constant. (iii) In the Active-with-Growth CM, the growth of the active microbial population was simulated based on the maximum growth rate modified by the dual Monod terms expressing the limited availability of oxygen and substrate. (iv) In the Growth-and-Decay CM, both growth and decay of the active microbial population were simulated. Decay was described by a first-order death rate, d, as presented in equation 5. This rate was set to be 1/10 of the maximum growth rate as an approximation (e.g., reported in reference 24) for Pseudomonas putida growing on toluene or as used in simulations (34).

The results of the total biomass CM (Fig. 4A) show an increasing depletion of O2 along the centerline until complete removal at about 80 mm of travel distance for both substrates. This does not match the observed rapid decrease of O2 concentrations close to the injection or the incomplete removal of O2 further down gradient (Fig. 2). Decreasing the biomass to the active fraction has a dramatic effect in which degradation and consumption of O2 are severely limited, as can be seen in the Active Biomass CM (Fig. 4B). Incorporating growth of the active biomass, the Active-with-Growth model results (Fig. 4C) in an O2 distribution that resembles the measured distribution and predicts the observed differences between phenol and acetate. Finally, the Growth-and-Decay CM (Fig. 4D) was able to further differentiate the behaviors of phenol from those of acetate, reproducing the slightly higher O2 concentrations throughout the phenol plume core.


Figure 4
View larger version (85K):
[in this window]
[in a new window]

  		FIG. 4. Simulated oxygen distribution from phenol and acetate flow cells. Simulated oxygen distribution (cut in the centerline) for phenol and acetate entering as substrates at x = 0. The underlying conceptual approaches for representing biomass of degraders were: CM Total Biomass, constant total biomass; CM Active Biomass, constant active biomass; CM Active with Growth, active biomass with growth; and CM Growth and Decay, addition of a death rate. Simulations were performed with RT3D2.5 and visualized with Tecplot 10. The flow direction is along the x axis; units for x and y are given in mm and in mg/liter for dissolved oxygen concentration (Oxygen).

These results demonstrate that simulations of biodegradation by a microbial population within the plume best reproduce experimental observations by including both the growth and the decay of active biomass. Furthermore, the initial biomass of microbes degrading the substrate had to be set as the active fraction of the total biomass, based on the experimental values for active cell counts (Table 1 and 2).

Figure 5 further illustrates several strengths and weaknesses in the mathematical representation. This is done by comparing dissolved O2 concentrations along the full length of the plume centerline. In the case of phenol (Fig. 5A), a comparison of results for the range of likely yield coefficients shows O2 concentrations generally to be within the range of variability for the triplicate runs. The exception is the area close to the injection where O2 is consumed less than was observed. For acetate (Fig. 5B), a more rapid consumption of O2 was predicted closer to the injection than was observed. Thus, while the simulations only qualitatively reproduce the observed behavior close to the source, they do capture quantitatively the key differences in O2 consumption by phenol (less) versus acetate (greater) in the plume core.


Figure 5
View larger version (24K):
[in this window]
[in a new window]

  		FIG. 5. Measured oxygen concentration at the centerline of the substrate plume in the three replicate flow cells (A) for phenol and (B) for acetate, compared to centerline oxygen concentration simulated numerically with consideration of growth and decay of microorganisms. Simulations are shown for an average yield, a minimum yield, and a maximum yield coefficient.

A comparison along lateral plume transects revealed that the observed depletion of oxygen spreads further outward than was simulated (Fig. 2). However, the shape of the profile was reproduced correctly. A closer match could be achieved by accounting for additional dispersive mixing (Fig. 2), as may be introduced by the microbial biofilm, which has already been previously suggested (14).

Microbial population dynamics.
The observed differences between phenol and acetate substrates becomes more obvious when studying the development of the biomass in the simulations with time (Fig. 6). The biomass in the phenol flow cell is initially higher than that in acetate but does not increase as rapidly. The more rapid growth in the acetate case results in a substantial localized increase in biomass close to the injection, which reflects the rapid oxygen consumption in that location. This emphasizes the importance of the maximum growth rate with respect to population dynamics and competitiveness, because although the rate is only two times higher for acetate than for phenol, it results in a much higher number of acetate degraders after a relatively short time despite a smaller initial activity. An increase in microbial activity close to the source is indicated by the extent of simulated O2 depletion. The observed O2 depletion is somewhat smaller than that simulated. The simulations also show that the biomass had not reached steady state after 24 h. Further simulations indicated that biomass development should reach a plateau after 1 to 2 weeks (not shown).


Figure 6
View larger version (69K):
[in this window]
[in a new window]

  		FIG. 6. Increased active cells relative to initial activity according to the Growth-and-Decay CM, simulated with RT3D2.5 and visualized with Tecplot 10. The comparison indicates a lower reaction of the phenol-supplemented community than that using acetate as the substrate, as noted by the scale for phenol. Notably, the scale of the z axis plotting active biomass is 10-fold in the case of the acetate-degrading population, and the 24-h phenol value had to be cut off because it was 1,000-fold more than the initial active biomass.

The modeling results provide three essential insights into the conceptual approach. (i) Since both substrate and EA were observed to be present at the plume fringe, a conceptual approach allowing such overlap, e.g., the dual Monod approach used, is required to capture the plume behavior correctly. (ii) It is important to account for the fact that only a small fraction of the biomass detected to be present is actually active and thus participating in the degradation process. (iii) Growth and decay of this (active) biomass is critical to the temporal evolution of the plume.

The significant increase in active biomass simulated in Fig. 6 is conceptualized and mathematically represented as an increase of the existing functional activity within the prevailing microbial community. However, the necessary increase in active biomass could alternatively arise from strong selection of specific functional groups that result in a significant shift in community structure. These alternative scenarios are explored in the following section.

Microbial community structure.
TGGE was used to monitor the microbial community structure generated from 16S rRNA gene amplicons amplified from (i) the subsamples from the flow cells, (ii) the growth culture used as an inoculum, and (iii) the original MW7 borehole water source. A diagrammatic representation of the community profiles for the two 24-h experiments is presented in Fig. 7. Cultivation of MW7 water with acetate or phenol generated changes in the community structure, and the changes were different depending on the carbon source present. Within the resolution limits of the technique, MW7 water contained eight distinct amplicons (Fig. 7, bands A to H).


Figure 7
View larger version (9K):
[in this window]
[in a new window]

  		FIG. 7. Diagrammatic representation of TGGE gels from a 24-h acetate experiment (A) and a 24-h phenol experiment (B). The amplicons present in MW7 water used as the original growth inocula (lane MW7) and the growth culture used to create biofilm (lane Growth) are shown. The amplicons present at time zero (T = 0) in six replicate samples taken randomly from the flow cell (replicates 1 to 6) are shown. For T = 24 h, the amplicons present in flow cells 1 to 3 (flow cell shown in parentheses in the figure) at six sampling points are shown; the samples were taken across a plume transect (where samples 1 and 5 represent background samples, samples 2 and 4 were from fringe areas, and sample 3 is a sample of the core) and in the carbon source injection area (sample 6). Gels were run at a gradient of 45 to 60°C.

The limited diversity in the MW7 groundwater used as an inoculum was not expected as a previous study (26) had shown that groundwater samples taken from a contaminated part of a phenol plume had a diverse microbial community when studied using PCR-TGGE. This limited diversity could in part be due to a limited capacity for band resolution (19, 21, 29), which has been seen in density gradient gel electrophoresis (DGGE), and the fact that bands can contain multiple amplicons as illustrated by sequencing (29). Exposure to acetate in the growth culture resulted in the loss of three bands (Fig. 7, bands D, G, and H) and the generation of at least three new bands (X', Y', and Z') enriched from species numerically below the level of sensitivity of the MW7 PCR. Similarly, exposure to phenol resulted in the loss of different bands (Fig. 7, bands D, G, and I) and the generation of one new band common to acetate growth (Fig. 7, X').

At time zero (T = 0), the structure of the acetate-grown community appeared to be similar, with the loss of two to three bands. Although some variation was observed at 24 h, the acetate-grown community appeared to be stable, with the retention of 75% of its community profile in all areas of the plume that were examined. Conversely, the phenol-grown community retained only 50% of bands at T = 0 (overall band loss compared to that of the MW7 of 63%) compared to the inoculum community structure. At T = 24 h, the phenol community appeared to be stable, with only a few samples showing the loss of band G. At T = 24 h, the communities in the three replicate flow cells (Fig. 7, T = 24 h, flow cells 1 to 3) for each carbon source were highly conserved, and no changes specifically related to position in the plume were observed (e.g., Fig. 7, T = 24 h, flow cell 1, sampling points 1 to 6.

Microbial communities generally respond to environmental stress or disturbance with a decrease in diversity because the ecological balance of the community is upset by the changes (1). A study has shown that when chemical pollutants were applied to microbial communities, the diversity decreased, and out of those left, the dominant populations had seemingly adapted to the conditions with enhanced tolerances and substrate utilization capabilities (2). In the current study, the results of the TGGE suggested that the main changes in community structure occurred during the development of the inoculum, and once it was established on the quartz matrix, the community structure was generally stable and unaffected by the conditions within the cell. More subtle changes would be masked by the low sensitivity of the technique. However, these experiments were run for only 24 h, and more change may be apparent if the duration of the experiment was increased to allow further evolution of the community within the plume.

These results suggest that despite nutrient limitation in the background zones of the plume and oxygen limitation in the core of the plume, the prevailing microbial community had largely adapted to the chemical conditions created by the formation of the degrading plume. This is an important result and provides a relatively simple conceptual basis with which to tackle field-scale microbial diversity and its variation across the gradients at the fringe of anaerobic plumes in otherwise oxic aquifers (a common case at contaminated sites). Our model infers that any observed change in microbial activity across a plume fringe is not necessarily accompanied by dramatic shifts in community structure. This in turn suggests that changes in activity occur through increased or decreased expression of function within the existing community rather than dramatic structural shifts in specific functional groups within the community. However, longer experimental times may generate more significant structural community changes through selection. Further changes may also occur for plume fringes that exhibit greater EA depletion and greater extremes in redox conditions. In our experiments, the plume core does not develop strongly reducing conditions. The reactor conditions described above are also analogous to that of a field site where the plume has existed for some time (years to decades) and the functional populations for biodegradation are highly selected. A further point of interest that is not addressed in this study but is potentially interesting for future research is the rate of ecological progression and shift in functional groups within the indigenous microbial community upon initial exposure of a pristine aquifer to contamination.

Comparison of the laboratory physical model with field systems.
A critical assessment of sampling and measurement methods in the laboratory flow reactors demonstrates some of the challenges for assessing MNA at field scale. Carbon mass balance and microbial activity measurements relied on discrete sampling of fluid via the septa and collection of grains for cell counts, respectively. For the active cell counts, it is important to note the plume volume sampled in these measurements. Sampling was done at greater than the grain scale in order to ensure sufficient grain and cell mass for the measurements to be carried out. Samples typically had a volume of the fringe with a lateral dimension of 3 to 4 mm. This affects the accuracy of assigning a sample measurement to the center of sampled area. Thus, reported active counts may be somewhat lower than the expected local maximum, since they reflect an average over the fringe area. The relatively lower observed activity in the source zone compared to the simulation results may likewise be due to the same effect of spatial averaging that can substantially underestimate local peak concentrations in plume fringe biomass. This same problem arises in field investigations where borehole fluid samples represent mixed water that arrives in the borehole from different locations within the aquifer catchment volume during pumping for sample collection. Analysis results for the mixed sample, however, are usually assigned single values that represent the precise location of the borehole. This averaging process can thus underestimate peak values that occur in situ.

Accurate interpretation of the spatial pattern in substrate concentration in fluid sampled via the septa also proved challenging for the reactor experiments. An observed random, quite small, lateral shift in the plume position between or during experiments meant the centerline of the plume and expected maximum substrate concentration did not remain at the fixed port location.

These results serve to illustrate one of the greatest challenges in demonstrating MNA performance at field scale. Although significant biodegradation may be occurring over time, a survey of field borehole fluid samples may not provide sufficient precision due to the large contaminant mass compared to the rate of biodegradation. This is further complicated by the possibility of dynamic behavior where the flow field or the source flux of contaminant changes over time and results in fluctuating concentrations at fixed sampling locations downstream.

Although in vitro cell activity measurement was not within the scope of our experiments, recent studies suggest that it could be possible and thus give a more direct measure of activity in real time. A system has been set up to allow lux bioreporter activity to be monitored in porous media in a stainless steel column using a novel fiber optic detection system (11). lux has also been used in a flow cell system similar to the one used here to monitor microbial transport (22). The authors of that study also observed the possibilities of combining the use of the lux bioreporter and the O2 tracer used in our study. Future studies using these methods could thus provide a more accurate representation of spatial patterns of activity at (theoretically) pixel-scale resolution. This would eliminate a significant source of error to the activity measurements, reported here, i.e., spatial averaging over 3- to 4-mm transects across the gradients at the plume fringe. Within the constraints of the experimental methodology, this study demonstrated that biodegradation with O2 as a sole terminal electron acceptor is greatest at the plume fringe and in the source area. This effect has been observed in a large-scale aquifer which had been subject to a substantially higher degree of phenol contamination (than used in the flow reactors) but was thought to arise from substrate toxicity and contaminant dilution at the plume fringe (16). Analysis of the laboratory results showed spatial variability in biodegradation rates and implied that it was linked mainly to variability in active cell numbers rather than differences in community. The variation of activity in time and space, results from relative rates of growth/death and differential expression within the microbial community, i.e., macroscopic biodegradation activity does not arise from a "resting" community or from growth alone.

An important conclusion is that even for very well-constrained and highly characterized flow regimens and uniform velocity fields in model plumes, the concentration distribution for substrate and EA alone are not sufficient to conceptualize the biodegradation processes appropriately. This potentially results in incorrect selection of mathematical descriptions for these processes and incorrect prediction of plume behavior. Only when quantitative microbiological data are combined with essential physical and chemical information is plume behavior correctly described.

In conclusion, the contribution of the microbiologist to a groundwater engineering scenario generally comprises noncontinuous sampling, often separated by months or years, and is mainly due to inaccessibility of the contaminant plume and cost of sampling (17, 26, 38). The development of scale model bench systems has many advantages, particularly as the behavior of the microbial community can be monitored in real time and the observed changes related the prevailing chemical conditions and to the activities observed. This is particularly important as biodegradation in a contaminant plume is one of the most important processes in successful natural attenuation of impacted groundwater systems (3). Furthermore, conditions in laboratory models can be manipulated to test microbial responses. The information gained will allow insights into the microbial black box, which has, up until now, generally been excluded from engineering considerations.


arrow
ACKNOWLEDGMENTS
 
H.C.R. and the experimental work were fully supported by a United Kingdom Biotechnology and Biological Sciences Research Council grant, number E15832.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Department of Biology, University of Leicester, University Road, Leicester LE1 7RH, United Kingdom. Phone: 44 (0)116 2525366. Fax: 44 (0)116 2523330. E-mail: hcr4{at}le.ac.uk Back

{triangledown} Published ahead of print on 27 April 2007. Back


arrow
REFERENCES
 
    1
  1. Atlas, R. M. 1984. Diversity of microbial communities. Adv. Microb. Ecol. 7:1-47.
    2. 2
  2. Atlas, R. M., A. Horowitz, M. Krichevsky, and A. K. Bej. 1991. Response of microbial-populations to environmental disturbance. Microb. Ecol. 22:249-256.[Medline]
    3. 3
  3. Banwart, S. A., and S. F. Thornton. 2003. The geochemistry and hydrology of groundwater bioremediation by natural attenuation, p. 93-138. In I. M. Head, I. Singleton, and M. Milner (ed.), Bioremediation: a critical review. Horizon Scientific Press, Wymondham, United Kingdom.
    4. 4
  4. Carraway, E. R., J. N. Demas, B. A. Degraff, and J. R. Bacon. 1991. Photophysics and photochemistry of oxygen sensors based on luminescent transition-metal complexes. Anal. Chem. 63:337-342.
    5. 5
  5. Chapelle, F. H. 1999. Bioremediation of petroleum hydrocarbon-contaminated ground water: the perspectives of history and hydrology. Ground Water 37:122-132.[CrossRef]
    6. 6
  6. Christensen, T. H., P. L. Bjerg, S. A. Banwart, R. Jakobsen, G. Heron, and H.-J. Albrechtsen. 2000. Characterisation of redox conditions in groundwater contaminant plumes. J. Contam. Hydrol. 45:165-241.[CrossRef]
    7. 7
  7. Cirpka, O. A., E. O. Frind, and R. Helmig. 1999. Numerical simulation of biodegradation controlled by transverse mixing. J. Contam. Hydrol. 40:159-182.[CrossRef]
    8. 8
  8. Corapcioglu, M. Y., and P. Fedirchuk. 1999. Glass bead micromodel study of solute transport. J. Contam. Hydrol. 36:209-230.[CrossRef]
    9. 9
  9. Curtis, T. P., W. T. Sloan, and J. W. Scannell. 2002. Estimating prokaryotic diversity and its limits. Proc. Natl. Acad. Sci. USA 99:10494-10499.[Abstract/Free Full Text]
    10. 10
  10. Don, R. H., P. T. Cox, B. J. Wainwright, K. Baker, and J. S. Mattick. 1991. Touchdown Pcr to circumvent spurious priming during gene amplification. Nucleic Acids Res. 19:4008.[Free Full Text]
    11. 11
  11. Dorn, J. G., M. K. Mahal, M. L. Brusseau, and R. M. Maier. 2004. Employing a novel fiber optic detection system to monitor the dynamics of in situ lux bioreporter activity in porous media: system performance update. Anal. Chim. Acta 525:63-74.[CrossRef]
    12. 12
  12. Garcia-Fresnadillo, D., M. D. Marazuela, M. C. Moreno-Bondi, and G. Orellana. 1999. Luminescent Nafion membranes dyed with ruthenium(II) complexes as sensing materials for dissolved oxygen. Langmuir 15:6451-6459.[CrossRef]
    13. 13
  13. Goel, R., T. Mino, H. Satoh, and T. Matsuo. 1998. Intracellular storage compounds, oxygen uptake rates and biomass yield with readily and slowly degradable substrates. Water Sci. Technol. 38:85-93.[CrossRef]
    14. 14
  14. Huang, W. E., S. E. Oswald, D. N. Lerner, C. C. Smith, and C. M. Zheng. 2003. Dissolved oxygen imaging in a porous medium to investigate biodegradation in a plume with limited electron acceptor supply. Environ. Sci. Technol. 37:1905-1911.[Medline]
    15. 15
  15. Huang, W. E., C. C. Smith, D. N. Lerner, S. F. Thornton, and A. Oram. 2002. Physical modelling of solute transport in porous media: evaluation of an imaging technique using UV excited fluorescent dye. Water Res. 36:1843-1853.[Medline]
    16. 16
  16. Jia, C., K. Shing, and Y. C. Yortsos. 1999. Visualisation and simulation of non-aqueous liquids solubilization in pore networks. J. Contam. Hydrol. 35:363-387.[CrossRef]
    17. 17
  17. Lerner, D. N., S. F. Thornton, M. J. Spence, S. A. Banwart, S. H. Bottrell, J. Higgo, H. E. H. Mallinson, R. W. Pickup, and G. M. Williams. 2000. Ineffective natural attenuation of degradable organic compounds in a phenol-contaminated aquifer. Ground Water 38:922-928.[CrossRef]
    18. 18
  18. Li, G., W. E. Huang, D. N. Lerner, and X. Zhang. 2000. Enrichment of degrading microbes and bioremediation of petrochemical contaminants in polluted soil. Water Res. 34:3845-3853.
    19. 19
  19. Lindstrom, E. S. 1998. Bacterioplankton community composition in a boreal forest lake. FEMS Microbiol. Ecol. 27:163-174.
    20. 20
  20. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265.[Free Full Text]
    21. 21
  21. Muyzer, G., E. C. Dewaal, and A. G. Uitterlinden. 1993. Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S ribosomal RNA. Appl. Environ. Microbiol. 59:695-700.[Abstract/Free Full Text]
    22. 22
  22. Oates, P. M., C. Castenson, C. F. Harvey, M. Polz, and P. Culligan. 2005. Illuminating reactive microbial transport in saturated porous media: demonstration of a visualization method and conceptual transport model. J. Contam. Hydrol. 77:233-245.[CrossRef][Medline]
    23. 23
  23. Oya, S., and A. J. Valocchi. 1998. Transport and biodegradation of solutes in stratified aquifers under enhanced in situ bioremediation conditions. Water Resour. Res. 34:3323-3334.[CrossRef]
    24. 24
  24. Pedersen, A. R., S. Moller, S. Molin, and E. Arvin. 1997. Activity of toluene-degrading Pseudomonas putida in early growth phase of a biofilm for waste gas treatment. Biotechnol. Bioeng. 54:131-141.[Medline]
    25. 25
  25. Pickup, R. W. 1995. Sampling and detecting bacterial populations in natural environments. Soc. Gen. Microbiol. Symp. Ser. 52:298-315.
    26. 26
  26. Pickup, R. W., G. Rhodes, M. L. Alamillo, H. E. H. Mallinson, S. F. Thornton, and D. N. Lerner. 2001. Microbiological analysis of multi-level borehole samples from a contaminated groundwater system. J. Contam. Hydrol. 53:269-284.[CrossRef][Medline]
    27. 27
  27. Porter, K. G., and Y. S. Feig. 1980. The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanogr. 25:943-948.
    28. 28
  28. Reardon, K. F., D. C. Mosteller, and J. D. B. Rogers. 2000. Biodegradation kinetics of benzene, toluene, and phenol as single and mixed substrates for Pseudomonas putida F1. Biotechnol. Bioeng 69:385-400.[CrossRef][Medline]
    29. 29
  29. Rees, H. C., W. D. Grant, B. E. Jones, and S. Heaphy. 2004. Diversity of Kenyan soda lake alkaliphiles assessed by molecular methods. Extremophiles 8:63-71.[CrossRef][Medline]
    30. 30
  30. Rodriguez, G. G., D. Phipps, K. Ishiguro, and H. F. Ridgway. 1992. Use of a fluorescent redox probe for direct visualization of actively respiring bacteria. Appl. Environ. Microbiol. 58:1801-1808.[Abstract/Free Full Text]
    31. 31
  31. Rowan, A. K., J. R. Snape, D. Fearnside, M. R. Barer, T. P. Curtis, and I. M. Head. 2003. Composition and diversity of ammonia-oxidising bacterial communities in wastewater treatment reactors of different design treating identical wastewater. FEMS Microbiol. Ecol. 43:195-206.[Medline]
    32. 32
  32. Thompson, I. P., C. J. van der Gast., L. Ciric, and A. C. Singer. 2005. Bioaugmentation for bioremediation: the challenge of strain selection. Environ. Microbiol. 7:909-915.[CrossRef][Medline]
    33. 33
  33. Thornton, S. F., D. N. Lerner, and S. A. Banwart. 2001. Assessing the natural attenuation of organic contaminants in aquifers using plume-scale electron and carbon balances: model development with analysis of uncertainty and parameter sensitivity. J. Contam. Hydrol. 53:199-232.[CrossRef][Medline]
    34. 34
  34. Thullner, M., M. H. Schroth, J. Zeyer, and W. Kinzelbach. 2004. Modeling of a microbial growth experiment with bioclogging in a two-dimensional saturated porous media flow field. J. Contam. Hydrol. 70:37-62.[CrossRef][Medline]
    35. 35
  35. Tiedje, J. M., K. Assuming-Brempong, K. Nusslein, T. L. Marsh, and S. J. Flynn. 1999. Opening the black box of soil microbial diversity. Appl. Soil Ecol. 13:109-122.[CrossRef]
    36. 36
  36. van der Gast, C. J., B. Jefferson, E. Reid, T. Robinson, M. J. Bailey, S. J. Judd, and I. P. Thompson. 2006. Bacterial diversity is determined by volume in membrane bioreactors. Environ. Microbiol. 8:1048-1055.[CrossRef][Medline]
    37. 37
  37. Watson, I. A., S. E. Oswald, S. A. Banwart, R. S. Croucj, and S. F. Thornton. 2005. Modelling the dynamics of fermentation and respiratory processes in a groundwater plume of phenolic contaminants interpreted from laboratory- to field-scale. Environ. Sci. Technol. 39:8829-8839.[Medline]
    38. 38
  38. Williams, G. M., R. W. Pickup, S. F. Thornton, H. E. H. Mallinson, Y. Moore, and C. White. 2001. Biochemical characterisation of a coal tar distillate plume. J. Contam. Hydrol. 53:175-198.[CrossRef][Medline]

Applied and Environmental Microbiology, June 2007, p. 3865-3876, Vol. 73, No. 12
0099-2240/07/$08.00+0     doi:10.1128/AEM.02933-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rees, H. C.
Right arrow Articles by Lerner, D. N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Rees, H. C.
Right arrow Articles by Lerner, D. N.
Agricola
Right arrow Articles by Rees, H. C.
Right arrow Articles by Lerner, D. N.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»