Comparison of Planktonic and Biofilm Cultures of Pseudomonas fluorescens DSM 8341 Cells Grown on Fluoroacetate

Abstract

Comparisons between the physiological properties of Pseudomonas fluorescens biofilm cells grown in a tubular biofilm reactor and planktonic cells grown in a chemostat were performed. Fluoroacetate was the sole carbon source for all experiments. The performance of cells was assessed using cell cycle kinetics and by determining specific fluoroacetate utilization rates. Cell cycle kinetics were studied by flow cytometry in conjunction with the fluorescent stain propidium iodide. Determination of the DNA content of planktonic and biofilm cultures showed little difference between the two modes of growth. Cultures with comparable specific glycolate utilization rates had similar percentages of cells in the B phase of the cell cycle, indicating similar growth rates. Specific fluoroacetate utilization rates showed the performance of planktonic cells to be superior to that of biofilm cells, with more fluoroacetate utilized per cell at similar specific fluoroacetate loading rates. A consequence of this decreased biofilm performance was the accumulation of glycolate in the effluent of biofilm cultures. This accumulation of glycolate was not observed in the effluent of planktonic cultures. Spatial stratification of oxygen within the biofilm was identified as a possible explanation for the overflow metabolism of glycolate and the decreased performance of the biofilm cells.

There is a general consensus that the adhesion of microbes to a surface influences bacterial metabolism; however, the experimental results are often contradictory (34). Some studies have compared the physiological status of biofilm and planktonic cells by determining their growth rates, with some investigators (2, 9) reporting increased biofilm growth rates in comparison to planktonic growth rates while others (1) have reported the opposite. Other researchers have compared the influence of adhesion on biofilm metabolic activity to planktonic activity, and as with growth rate, conflicting observations have been reported (16, 31, 11). The objective of the present study was to characterize planktonic cells grown in a chemostat and biofilm cells grown in a tubular biofilm reactor (TBR) and compare their performance with respect to the degradation of the model xenobiotic compound fluoroacetate.

Halogenated compounds are extensively used in many applications (refrigeration, lubricants, pharmaceuticals, insecticides, and herbicides) and can be considered significant environmental contaminants. The biodegradation of many chlorinated compounds has been widely reported (5, 26, 28). However, in spite of the increased use of organofluorine compounds in the past 60 years, there is limited information on their degradation (17). Currently, a large fraction of wastewater streams containing fluorinated compounds are incinerated (12). Improved biological waste treatment processes require a deeper understanding of microbial degradation of fluorinated compounds.

Some previous studies have focused on the biodegradation of fluorinated aromatic compounds using biofilm reactors (3, 10); however, there have been no studies on the degradation of fluorinated aliphatic compounds in biofilm reactors. Thus, sodium fluoroacetate was chosen as the model xenobiotic to study the efficiency of aliphatic organofluorine degradation in biofilms. It was the first naturally occurring fluorinated compound to be isolated, obtained from the South African shrub Dichapetalum cymosum (23). Fluoroacetate is highly toxic to mammals and has found extensive use as a vertebrate pesticide, particularly in Australia and New Zealand. A number of studies have focused on the isolation and identification of microbial soil isolates with the ability to degrade fluoroacetate (14, 33, 35), and other studies have focused on the mechanism of defluorination (13, 15, 21). However, there has been no research on the degradation of fluoroacetate by biofilm cultures. Biofilm systems appear ideal for the degradation of xenobiotics because of the many reported advantages they have over planktonic cultures. Most microorganisms that have the ability to degrade xenobiotic compounds have comparatively slow growth rates, and biofilm reactors allow the enrichment of these microorganisms independent of hydraulic retention times (36). It has been shown in numerous studies that biofilms are less susceptible than suspended bacteria to changes in environmental conditions such as temperature and pH and the presence of metabolic products and toxic substances (8, 25, 27, 36). The high cell concentrations that can be achieved in biofilm systems in combination with high volumetric flow rates could potentially result in high volumetric productivities without the risk of cell washout.

The species Pseudomonas fluorescens has been extensively studied; it commonly exists as a biofilm in natural environments and is ubiquitous in industrial environments (6, 29, 30). The specific strain used here, P. fluorescens DSM 8341, was previously isolated from a soil sample in Western Australia, and in a study with 23 other microbial soil isolates, it was shown to be the most efficient degrader of fluoroacetate when fluoroacetate was the sole carbon source (37). The effect of the environmental factors, pH and temperature, on the biodefluorination of fluoroacetate by P. fluorescens was also determined (38); however, at present there are no reported planktonic growth kinetics established for this strain, nor has it previously been grown as a biofilm. In this context a TBR was employed to investigate the degradation of fluoroacetate by a P. fluorescens biofilm, in conjunction with chemostat studies that were conducted to determine the efficiency of planktonic degradation of the substrate. Specific utilization rates, flow cytometry, and fluorescent microscopy were employed to compare the performance and physiological status of biofilm and planktonic cells grown with fluoroacetate as the sole organic substrate.

MATERIALS AND METHODS

Medium and culture conditions.

P. fluorescens (DSM 8341) was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Germany). Brunner's minimal medium (medium 457; DSMZ) supplemented with sodium fluoroacetate (Sigma, United Kingdom) as the sole organic substrate was the growth medium for all experiments with the exception of the 50 mM biofilm experiments. In these experiments the concentration of KH₂PO₄ was increased from 1.52 to 15.2 g/liter in order to provide greater buffering capacity.

Biofilm reactor.

The biofilm was grown in the lumen of a silicone tube (Alteil, United Kingdom), referred to as a TBR, which is schematically presented in Fig. 1. The nominal inner dimension of the silicone tubing was 3 mm, and the wall thickness was 1 mm. The system consisted of a medium reservoir, peristaltic pump, a glass flow break, four 40-cm sections of silicone tubing separated by 5 sample ports, a glass flow break, and a spent medium reservoir. This reactor configuration allowed the determination of local metabolite concentrations and biofilm characteristics, and thus each section can be considered an independent reactor. Four separate TBR experiments were performed, with concentrations of 10, 20, and 50 mM fluoroacetate in the medium feed; the 20-mM experiment was repeated. Thus, 16 different specific fluoroacetate loading rates were examined (four reactors with four sections each). Cells were grown for 24 h at 30°C in batch culture prior to reactor inoculation to ensure that the cells were in the exponential phase of growth. The 24-h culture was adjusted to a turbidity of approximately 0.1 at 660 nm in phosphate-buffered saline (PBS), mixed with 10 ml of medium, and inoculated into the reactor. Following inoculation the system was operated in static mode for approximately 48 h at 30°C, after which time the flow of medium was initiated at a flow rate of 6 ml/h (velocity, 0.023 cm/s) and maintained throughout the experiments. The dilution rate was approximately 2.5 times greater than the maximum planktonic growth rate; this was chosen to ensure the washout of planktonic cells and encourage biofilm formation. Samples were taken from the reactor daily from sample port 5 and from all five ports 24 h prior to biofilm harvesting. The following parameters were analyzed: optical density, fluoroacetate, free fluoride ion, glycolate, DNA content of cells, and CFU count. The biofilm was allowed to develop for approximately 300 h, at which time the bioreactor was disassembled, and the biofilm was harvested for analysis.

FIG. 1.

Schematic of TBR. Each section consists of 40 cm of silicone tubing, with a total reactor length of 160 m.

Biofilm harvesting.

Prior to biofilm harvesting, the TBR was drained of bulk liquid to prevent any interference of detached cells. Biofilm cells were harvested from the inside of the tubing; the tubing was sectioned into 5-cm divisions, and thickness, dry weight, DNA content of cells, and number of CFU were measured. Biofilm was removed from the lumen of the tubing by pinching the outer surface of the tubing; the lumen was washed with PBS, and the disrupted biofilm was harvested.

Chemostat trials.

Chemostat studies were performed in duplicate in a 3-liter bioreactor (Applikon, The Netherlands) with a working volume of 1.5 liters, and the performance of planktonic cells was assessed at four fluoroacetate loading rates representing growth rates between 45 and 80% of the maximum growth rate. The reactor was equipped with controls for agitation, pH, temperature, and dissolved oxygen. For pH adjustment 1 M NaOH was used. The inlet gas flow rate was maintained at a constant rate of 1.5 liters/min by use of a mass flow controller. Dissolved oxygen was controlled by agitation and maintained at or above 40% air saturation throughout the experiments. Temperature was maintained at 30°C throughout the experiments. The reactor was inoculated with 15 ml of 24-h-old culture adjusted to an optical density of 0.1 at 660 nm. Samples were taken from the reactor periodically, and the following parameters were analyzed: optical density, fluoroacetate, free fluoride ion, glycolate, DNA content of cells, dry weight, and number of CFU. The reactor was operated under batch conditions during the exponential phase of growth; at the onset of the stationary phase of growth, the reactor was switched to continuous operation. The feed flow rate was measured by recording the mass of liquid effluent over a specified period of time using a balance (Mettler Toledo). The initial fluoroacetate concentration during the batch phases of growth was 20 mM.

Batch growth trials.

Planktonic growth rates were determined for initial fluoroacetate concentrations of 10, 20, and 50 mM in 250-ml conical flasks incubated at 30°C with shaking at 150 rpm. The flasks were operated with a working volume of 50 ml and inoculated with 0.5 ml of a 24-h-old culture adjusted to an optical density of 0.1 at 660 nm with PBS. The flasks were sampled periodically and analyzed for optical density and free fluoride ion.

Anaerobic batch culture experiments.

Anaerobic batch culture experiments were performed in triplicate with 20 mM fluoroacetate or 20 mM glycolate as the sole carbon source in 100-ml Schott Duran bottles with a working volume of 30 ml. An anaerobic environment was maintained by continuous subsurface sparging of nitrogen gas at a flow rate of 0.5 ml/min. This flow rate was found to be sufficient to ensure an anaerobic environment while also avoiding liquid losses due to evaporation over the 8-h duration of the experiment. The Duran bottles were inoculated by harvesting 30 ml of a 24-h-old culture grown on fluoroacetate; the harvested cells were then resuspended in 1 ml of fresh medium prior to inoculation. The bottles were incubated at 30°C with shaking at 150 rpm for 8 h, at which time they were sampled and analyzed for optical density, dissolved oxygen, and metabolite concentrations. There was no dissolved oxygen detected in any of the anaerobic experiments while the dissolved oxygen concentrations in the aerobic control bottles were found to be above 40% air saturation after the 8-h incubation period.

Epifluorescence microscopy.

A 5-cm section of tubing containing biofilm was washed with PBS (1 ml) and stained with 5-cyano-2,3-ditoyl tetrazolium chloride (CTC; 400 μg/ml) for 2 h at 30°C. The tubing was drained, washed with PBS (1 ml), and stained with 4′,6′-diamidino-2-phenylindole (DAPI; 100 μg/ml) for 30 min at ambient temperature. The tubing was washed with PBS and embedded with 5 ml of 22-oxyacalcitriol (OCT) histological cryoembedding medium (Tissue TEK OCT compound), which was gently injected into the tubing. The tubing was then placed in an embedding chamber for 10 min to solidify the OCT, and samples were stored at −80°C until sectioning. The embedded biofilm was sectioned (10-μm cross-sections) after the biofilm was removed from the silicone tubing, using a Microm HM 550 cryostat (Microm, Walldorf, Germany). The sections were placed on glass slides and examined with an Olympus BX51 epifluorescence microscope using a 4× objective. Photographs were obtained with an Olympus DP70 digital camera. The exposure time was 500 ms to acquire CTC images and 50 ms to acquire DAPI images.

Thickness measurements and image analysis.

Biofilm thickness was analyzed by Able Image Analyser software (Mu Labs, Slovenia). Thickness was measured as the distance from the membrane to the biofilm liquid interface. From each section of the reactor, six phase-contrast images were analyzed for thickness, and 25 measurements were taken from each image. These measurements were then averaged to give a final thickness. Depth profiles of respiratory activity (CTC) and biomass (DAPI) were measured perpendicular to the membrane in overlapped CTC and DAPI images acquired at the same point. The interface between the biofilm and the membrane was set at zero. Fifteen individual profiles were performed in six different images from each section. Three representative profiles were then averaged to produce a final radial profile of respiratory activity and biomass for each section.

Dry weight.

Biofilm dry weight was determined by drying a 5-cm section of tubing at 60°C for 48 h, at which point a constant weight was achieved. The tubing was immersed in warm water to hydrate the biofilm, and the biofilm was removed by pinching the silicone tubing, which was then washed vigorously with water to remove any residual biofilm. The silicone tubing was dried for a further 48 h and reweighed. The weight of the biofilm was calculated by subtracting the dry weight of the empty tubing from the weight of the tubing plus biofilm.

Specific utilization rate.

The specific utilization rate (q) is defined as the concentration of substrate degraded per unit time per CFU. The utilization of a carbon/energy substrate is separated into two fluxes (Fig. 2) corresponding to consumption of the substrate for incorporation into biomass (q_an) and utilization of substrate for energy, which can be further subdivided into the energy required for growth (q_en) and energy required for maintenance (q_m). In some cases it is possible that there is a rate-limiting step in the overall catabolic pathways, which leads to the accumulation of an intermediate metabolite and is described here as overflow metabolism. Fluoroacetate was the sole organic substrate available to the cells. It is initially degraded by the enzyme fluoroacetate dehalogenase (Fig. 3), yielding free fluoride ion and glycolate; glycolate is then utilized as the carbon/energy source. Figure 3 shows that carbon is conserved in the fluoroacetate dehalogenase reaction; thus, there are two specific utilization rates: q_f is the specific utilization rate of fluoroacetate, which is a single enzymatic reaction that results in the production of glycolate, and q_g is the specific utilization rate of glycolate consumed for the production of cell material (q_an) and ATP generation (q_en and q_m). Unconsumed glycolate (overflow metabolism) is transported across the cell membrane into the bulk liquid. In order to compare q_f and q_g directly, specific utilization rates were calculated as the concentration (in mM) of carbon.

FIG. 2.

The utilization of fluoroacetate involves the enzymatic cleavage of the carbon-fluorine bond with the production of glycolate. The specific utilization rate of glycolate, q_g, is a combination of q_an where carbon is incorporated into cell (anabolism), q_en, where carbon is used to provide energy for growth, and q_m, where carbon is used to provide energy for the maintenance of cellular function not associated with growth (maintenance of intracellular osmotic potential) but excludes any glycolate that is produced but not utilized (overflow metabolism [O_m]) and represents the total carbon utilized. The specific fluoroacetate utilization rate includes q_an, q_e, q_m, and O_m and represents the total fluoroacetate degraded.

FIG. 3.

Fluoroacetate defluorination via haloacetate dehalogenase.

Flow cytometry.

Cells were harvested from biofilm and planktonic cells as described above, diluted to a concentration of approximately 5 × 10⁶ cells per ml, and washed three times in PBS. This was performed by resuspending the biofilm in PBS, centrifuging for 5 min at 14,000 rpm, and discarding the supernatant. The washed cells were suspended in 0.1 ml of PBS and 1 ml of ice-cold 70% ethanol. The cells were kept at 4°C for 24 h to fix the cells prior to transfer to −20°C for storage. Prior to analysis, the cells were separated from the ethanol by centrifugation and resuspended in 1 ml of PBS containing propidium iodide (80 μM). Cells were then analyzed for DNA content using a Dako Cyan ADP flow cytometer (Dako, Glostrup, Denmark), and histograms showing fluorescence intensity against cell number were generated (Fig. 4). The different phases of cell cycle (Fig. 4, phases B, C, and D) were then determined from these histograms using Multicycle software (Phoenix Flow Systems).

FIG. 4.

Typical histogram of fluorescent intensity for planktonic cells. Multicycle software was used to determine the B, C, and D phases of the cell cycle. B-phase cells have a single copy of DNA, C-phase cells are synthesizing DNA, and D-phase cells have a double copy of DNA and are predivision.

Fluoroacetate, free fluoride ion, and glycolate analysis.

Free fluoride ion was measured using an ion selective fluoride combination electrode (Thermo Orion model 290). Fluoroacetate concentration was determined by fluorine-19 nuclear magnetic resonance spectroscopy. Samples were prepared by mixing 0.6 ml of culture fluid with 0.2 ml of D₂O (added to provide a lock signal), and analysis was performed using a Varian 400 MHz spectrometer. The known free fluoride concentration in each effluent sample was used as an internal standard. The ratio of free fluoride ion signal (−120 ppm) to fluoroacetate signal (−215 ppm) was used to calculate fluoroacetate concentration. Glycolate was measured by the method described by Lewis and Weinhouse (20).

RESULTS

Biofilm formation, growth, and structural characteristics.

The TBR system was employed as a method to characterize P. fluorescens biofilms grown at various carbon loading rates. The carbon loading rate is defined here as the total carbon available in a section of tubing; it includes both the carbon in the form of fluoroacetate and the carbon in the form of glycolate. It has been shown that fluid velocity is a critical parameter in determining biofilm-adhesive strength for P. fluorescens biofilms (4); accordingly, in order to maintain a constant velocity in all experiments, carbon loading rates were varied by altering the fluoroacetate inlet medium concentration and not the dilution rate. Experiments were performed with initial fluoroacetate concentrations of 10, 20, and 50 mM in the medium feed. Visual observation of initial biofilm formation occurred between 116 and 160 h after reactor inoculation. Biofilm appearance, as indicated by colony formation on the lumen of the silicone tube, depended on the carbon loading rate; in experiments employing 10 and 20 mM fluoroacetate, where carbon loading rates were between 0 and 0.190 mM/h, biofilm growth was apparent in a shorter period of time than in experiments with 50 mM fluoroacetate, where carbon loading rates were between 0.300 and 0.510 mM/h. Steady state, as indicated by fluoroacetate, glycolate, and free fluoride ion concentrations in the effluent, was achieved earlier in the experiments with lower carbon loading rates. In the reactors with initial carbon loading rates of less than 0.190 mM/h, complete carbon utilization had occurred by the time steady state had been reached. However, in the reactor with carbon loading rates of above 0.300 mM/h, approximately 3 mM fluoroacetate and 18 mM glycolate were still detectable at steady state.

Visual observations of the biofilm formation provides some qualitative information as to when initial biofilm formation occurs, and measured metabolite concentrations throughout the time course of the experiments indicate when steady-state conditions have been reached, but these data do not provide any quantitative measurement of biofilm structure or performance. Thus, biofilm characteristics of thickness, dry weight, cell number, and density were calculated at different sections of the reactor (Fig. 1) for different carbon loading rates after biofilm harvesting (Table 1). Cross-sectional biofilm thickness was calculated from phase-contrast images. These thickness measurements show little variation in the average biofilm thickness for all sections where the carbon loading rates were above 0.017 mM/h; the average thickness was 116.8 μm, with a standard deviation around the mean of ± 22.9 μm. An exception to this trend was the section with a carbon loading rate of 0.107 mM/h, where the average thickness was 130 μm thicker than that recorded in any other section. The same trends observed for average thickness were observed for dry weight, CFU count, and density measurements, with a few exceptions, most notably for the section with a carbon loading rate of 0.107 mM/h. Images taken from the section with this carbon loading rate suggest that the biofilm may have been in the process of sloughing, and this could explain the greater thickness and lower density values recorded in this section. The reactor experiment using 20 mM fluoroacetate was performed in duplicate, and single-factor analysis of variance showed that there was no significant difference between biofilm thickness, dry weight, cell number, and density values recorded in these two reactors (F = 0.15, F critical = 4.7, and P = 0.69). Biofilm was present in three individual sections of the reactors where the carbon loading rate was 0 mM/h as a result of carbon source depletion, and in these sections the presence of biofilm was probably caused by the presence of carbon during the early stages of the reactor operation.

TABLE 1.

Biofilm thickness, dry weight, CFU count, and density data after approximately 300 h of growth at various carbon loading rates^a

Carbon loading rate (mM/h)	Initial reactor fluoroacetate concn (mM)	Reactor section no.	Thickness (μm)	Dry wt (g/40-cm section)	CFU/40-cm section	Density (g/liter)
0.000	10	4	12	0.0000	1.12 × 109	0.00
0.000	10	3	22	0.0016	6.48 × 109	18.5
0.000	20	4	29	0.0120	6.20 × 109	108
0.007	20	4	32	0.0128	1.68 × 109	80.0
0.017	10	2	150	0.0952	4.40 × 1010	169
0.036	20	3	92	0.0624	1.91 × 1010	181
0.036	20	3	85	0.0496	1.52 × 1010	154
0.080	20	2	95	0.0728	8.81 × 109	202
0.081	20	2	136	0.0656	2.72 × 1010	128
0.107	10	1	281	0.0696	4.56 × 1010	65.0
0.168	20	1	120	0.0720	1.92 × 1010	159
0.187	20	1	120	0.0872	1.36 × 1010	192
0.304	50	4	134	0.0552	6.16 × 109	108
0.362	50	3	147	0.0768	4.40 × 109	138
0.412	50	2	96	0.0728	1.12 × 1010	201
0.509	50	1	110	0.0648	1.92 × 1010	156

^aCarbon loading rates are used here and not fluoroacetate loading rates to take into account that some sections of the reactors have both fluoroacetate and glycolate in the inlet feed.

Comparisons between planktonic and biofilm performance.

Comparisons between biofilm and planktonic cells are difficult to make due to the intrinsic difference between the two modes of growth, such as compositional differences in biomass. Typically, performance is determined by specific utilization rates (q), where q is defined as the concentration of substrate degraded per unit of time divided by the dry weight of biomass. However, dry weight measurements do not take into account extracellular polymeric substances, which account for a much higher fraction of biofilm biomass than of planktonic biomass. It has been demonstrated that the composition of P. fluorescens B52 extracellular polymeric substances is different for biofilm and planktonic cells (18). Thus, in this study q was calculated using the CFU count rather than dry weight. CFU measurements are advantageous in this situation as they are a measure of viable cells for both biofilm and planktonic systems. The reactor working volume for the chemostat was 1.5 liters while the reactor volume in each section of the TBR was only 0.0028 liters. To account for this difference in reactor volumes, specific fluoroacetate loading rates (mM/CFU/h) rather than fluoroacetate loading rates (mM/h) were used in conjunction with specific fluoroacetate and glycolate utilization rates to compare the performance of biofilm and planktonic cells (Fig. 5). For the range of loading rates investigated, the data clearly reflect higher utilization rates for planktonic cells than for biofilm cells. Linear regression analysis of fluoroacetate utilization rates at specific fluoroacetate loading rates between 0 and 14 × 10⁻¹² mM/CFU/h shows a linear relationship for both biofilm (R² = 0.91) and planktonic cells (R² = 0.90). Biofilm fluoroacetate loading rates above 14 × 10⁻¹² mM/CFU/h were not included in the regression analysis as the performance of planktonic cells was not assessed above this range. The performance of planktonic cells was superior to biofilm cells, with a higher dependency on specific fluoroacetate utilization recorded for planktonic cells (slope, 0.87) than for biofilm cells (slope, 0.57). Specific glycolate utilization rates for planktonic cells were the same as specific fluoroacetate utilization rates; thus, the fluoroacetate degradation/glycolate production rate was equal to the rate of glycolate utilization for planktonic cells at all specific fluoroacetate loading rates examined. However, there were significant differences between specific glycolate and specific fluoroacetate utilization rates for biofilm cells at all specific fluoroacetate loading rates above 2 × 10⁻¹² mM/CFU/h (t statistic = 1.43; P = 0.091, based on a one-tailed t test, assuming unequal variances). A result of this difference between specific fluoroacetate and glycolate utilization rates for biofilm cells was the accumulation of glycolate (overflow metabolism). Thus, the data indicate a rate-limiting step in the utilization of glycolate by biofilm cells that was not observed for planktonic cells.

FIG. 5.

Specific fluoroacetate (•) and glycolate (○; shown with the symbol for fluoroacetate superimposed) utilization rates for planktonic cells and specific fluoroacetate (▪) and glycolate (□) utilization rates for biofilm cells at various specific fluoroacetate loading rates. Regression analysis shows the superior performance of planktonic cells (slope = 0.87; R² = 0.90) in comparison to biofilm cells (slope = 0.57; R² = 0.91) at specific fluoroacetate loading rates between 2 × 10⁻¹² and 14 × 10⁻¹² mM/CFU/h. Dashed lines, 95% confidence intervals.

Oxygen limitation.

Depth profiles of respiratory activity as indicated by CTC and biomass concentration as indicated by DAPI show spatial stratification of respiratory activity within the biofilm (Fig. 6). The region of highest respiratory activity is located adjacent to the biofilm membrane interface where oxygen concentration is highest. Respiratory activity decreases toward the bulk liquid interface while biomass concentration does not, suggesting decreased oxygen availability in this location. While it was not possible to measure the oxygen profiles within the biofilm in this study, the penetration depth of oxygen into a biofilm is governed by a reaction-diffusion interaction. The relative rate of oxygen diffusion and oxygen consumption by the biofilm determines the depth of oxygen penetration, which can be determined theoretically from the following equation (32): a = (2D_eS_o/K_o)^1/2, where a is the penetration depth of the reacting solute, D_e is the effective diffusion coefficient of the solute in the biofilm, S_o is the solute concentration at the biofilm interface, and K_o is the volumetric reaction rate of the solute inside the biofilm. Estimates of each of these parameters were made to allow the calculation of oxygen penetration depth. The effective diffusion coefficient of oxygen in the biofilm at 30°C was taken as 57% of the diffusion coefficient of pure water at this temperature (32), or 1.28 × 10⁻⁵ cm² s⁻¹. The concentration of dissolved oxygen at the biofilm membrane interface was taken as 8.0 mg liter⁻¹, and the volumetric consumption rate of oxygen was determined experimentally to be 1.64 mg liter⁻¹ s⁻¹. From these estimates the calculated depth of oxygen penetration into the biofilm from the membrane was 111 μm. The average biofilm thickness in the sections of the reactors where overflow metabolism of glycolate occurred was 118 μm ± 21 μm. These oxygen depth calculations combined with biofilm thickness measurements and the spatial stratification of respiratory activity within the biofilm suggest that oxygen does not completely penetrate into the biofilm. A study by Donnelly and Murphy (7) demonstrated that the fluoroacetate dehalogenase enzyme isolated from P. fluorescens does not require the presence of oxygen to degrade fluoroacetate. Thus, a possible explanation for the overflow metabolism of glycolate is oxygen limitation. Anaerobic batch culture experiments were performed to test the hypothesis that oxygen limitation was the reason for the overflow metabolism of glycolate in the TBR. After 8 h of incubation, there was no growth, as indicated by optical density measurements, in anaerobic cultures containing either fluoroacetate or glycolate as the substrate, demonstrating that oxygen is required for biomass production. However, when the cells were incubated with fluoroacetate under anaerobic conditions, 1.2 ± 0.01 mM glycolate and 1.4 ± 0.05 mM free fluoride ion were detected in the effluent. Thus, fluoroacetate was degraded in the absence of oxygen but glycolate was not, thus supporting the suggestion that oxygen is the rate-limiting step in overall glycolate catabolism. Interestingly, when P. fluorescens was grown aerobically, there was greater biomass production when the cells were grown on glycolate as the sole carbon source (optical density, 0.811 ± 0.11) than when the sole carbon source was fluoroacetate (optical density, 0.411 ± 0.16). The cells grown on glycolate utilized more carbon (16.2 ± 0.3 mM) than cells grown on fluoroacetate (8.5 ± 0.45 mM) in the same time period, which is surprising since the cells were adapted to growth on fluoroacetate and not glycolate. As with the shake flask and chemostat experiments, no glycolate was detected in the effluent of the control fluoroacetate aerobic experiment, indicating that the fluoroacetate degradation rate was equal to the glycolate utilization rate when oxygen was not limiting.

FIG. 6.

Fluorescent images of biofilm stained with the respiratory indicator CTC (red) and the DNA binding dye DAPI (A) and the corresponding depth profiles of the location of respiratory activity (solid line) and biomass (dotted line) within the biofilm (B) show that respiratory activity decreases toward the biofilm-bulk liquid interface.

Comparisons between the DNA content of planktonic and biofilm cells.

Bacteria with growth rates between 3.3 and 1 h⁻¹ can have multiple copies of DNA, making cell cycle analysis difficult due to overlapping replication cycles. Batch culture planktonic growth trials were performed to determine the growth rates of P. fluorescens grown on 10, 20, and 50 mM fluoroacetate and were found to be 0.12, 0.15, and 0.14 h⁻¹, respectively. The B phase of the bacterial cell cycle can be described as the time between cell division and the initiation of a new round of replication, the C phase is the time between initiation and termination of chromosome replication, and the D phase is the time between chromosome termination and cell division. The C and D phases of growth can be considered constants. However, at growth rates of less than 1 h⁻¹ the B phase of bacterial growth is not constant but increases with a decreasing growth rate; thus, an increase in the percentage of cells in a population in the B phase can be expected in slowly growing cells (19). The distributions of biofilm and planktonic cells in the various phases of cell cycle are compared with specific glycolate utilization rates in Fig. 7. Cell cycle distributions were compared using specific glycolate utilization rates and not specific fluoroacetate utilization rates as the former is a measurement of the utilization of substrate for carbon and energy while the latter is a measurement of a single enzyme activity. Biofilm and planktonic cells that had specific glycolate utilization rates of zero (zero carbon loading rates), as expected had a very high percentage of cells in the B phase (approximately 90%), while there were very few cells in the D phase of cell cycle. The data suggest that there is no significant difference between the percentage of cells in the B phase of growth for planktonic and biofilm cells investigated at similar specific glycolate utilization rates between 4 × 10⁻¹² and 8 × 10⁻¹² mM/CFU/h (t statistic = 0.25 and P = 0.403, based on a one-tailed t test, assuming unequal variances), which indicates similar growth rates. Planktonic cells with specific glycolate utilization rates of between 4 × 10⁻¹² and 8 × 10⁻¹² mM/CFU/h were observed to have growth rates of between 0.07 and 0.08 h⁻¹; thus, biofilm cells with similar glycolate utilization rates were growing at 45 to 53% of the maximum planktonic growth rate. On the basis on a one-tailed t test, assuming unequal variances, the data show that, for biofilm cells, decreased specific glycolate utilization rates from between 4.5 × 10⁻¹² and 8 × 10⁻¹² mM/CFU/h to between 0.06 × 10⁻¹² and 2 × 10⁻¹² mM/CFU/h significantly increase the percentage of cells in the B phase (t = 3.98 and P = 0.00266), suggesting growth rates of less than 0.07 h⁻¹. There was a statistically significant decrease in the percentage of cells in the B phase of growth when planktonic specific glycolate utilization rates increased from between 4 × 10⁻¹² and 8 × 10⁻¹² mM/CFU/h to between 8 × 10⁻¹² and 12 × 10⁻¹² mM/CFU/h (t = 1.57 and P = 0.09, based on a one-tailed t test, assuming unequal variances). This decrease in the percentage of cells in the B phase was as expected as the cells with specific glycolate utilization rates of between 8 × 10⁻¹² and 12 × 10⁻¹² mM/CFU/h are utilizing more carbon, have higher growth rates between 0.09 and 0.12 h⁻¹, and consequently have fewer cells in the B phase of growth.

FIG. 7.

Cell cycle distributions for planktonic (A) and biofilm (B) cells at a number of specific glycolate utilization rates. Open bar, B phase; dark gray bar, C phase; and light gray bar, D phase. Standard deviations were determined from a minimum of three independent biological samples. Planktonic specific glycolate utilization rates between 4 to 8 and 8 to 12 mM/CFU/h correspond to growth rates of between 0.07 to 0.08 and 0.09 to 0.12 h⁻¹, respectively.

DISCUSSION

The increasing number of applications that rely on organofluorine compounds has resulted in their becoming universal environmental contaminants (24). However, the degradation of fluorinated xenobiotics is poorly understood in comparison to other halogenated compounds since research has tended to concentrate on chlorinated and brominated pollutants. Biodegradation pathways can result in the accumulation of intermediates that are potentially inhibitory or toxic (22), and because of the importance of diffusion in biofilms, the accumulation of such intermediates could potentially affect performance. The biodegradation of the model xenobiotic, fluoroacetate, by P. fluorescens was investigated, and the performance of planktonic and biofilm cells was compared.

Here, two methods have been used to determine the physiological status and activity of planktonic and biofilm cells. First, the metabolic activities of biofilm and planktonic cells were assessed in terms of specific utilization rates. It was found that the performance of planktonic cells was superior to that of biofilm cells, with planktonic cells mineralizing more fluoroacetate per CFU per hour than the biofilm cells. When the cells were grown under planktonic conditions, there was no difference between specific fluoroacetate and specific glycolate utilization rates at any of the specific fluoroacetate loading rates examined (Fig. 5). This was not the case for the biofilm cells, and specific glycolate utilization rates were lower than their corresponding specific fluoroacetate utilization rates at a number of specific fluoroacetate loading rates. Thus, the major differences between biofilm and planktonic cells were the performance of the biofilm cells in terms of fluoroacetate and glycolate utilization and the overflow metabolism of glycolate.

It has been shown previously (7) that the fluoroacetate dehalogenase enzyme isolated from P. fluorescens does not require the presence of oxygen to degrade fluoroacetate, and in this study it has been demonstrated that the utilization of glycolate via aerobic respiration requires oxygen as a terminal electron acceptor. Thus, oxygen limitation within the biofilm offers an explanation for the overflow metabolism of glycolate. Spatial stratification of respiratory activity, as indicated by CTC, supports the possibility of oxygen limitation within the biofilm. Respiratory activity is highest at the biofilm-membrane interface and decreases toward the biofilm-bulk liquid interface (Fig. 6). This spatial stratification of nutrients within the biofilm is a major difference between biofilm and planktonic cells. For planktonic cells grown in the chemostat system, the dissolved oxygen concentration was controlled at or above 2.8 mg/liter, providing sufficient oxygen for all cells. Oxygen was supplied to the biofilm by diffusion through the silicone membrane, which also acts as a support for the biofilm. As the biofilm grows, it increasingly acts to resist the mass transfer of nutrients to the inner regions of the biofilm. Calculations of oxygen penetration depth suggest that the biofilm was sufficiently thick to prevent the penetration of oxygen to the region of the biofilm located adjacent to the bulk liquid interface, and they support the possibility of oxygen limitation.

At similar specific fluoroacetate loading rates, the specific utilization rate of fluoroacetate was lower for biofilm cells than planktonic cells. A possible explanation for this result is that less energy is available to biofilm cells due to decreased glycolate utilization; thus, there is less energy produced for, among other requirements, the production of fluoroacetate dehalogenase and for the production of any permease necessary for the transport of fluoroacetate into the cell. Therefore, a reduced glycolate utilization rate due to oxygen limitation could ultimately affect fluoroacetate utilization. This suggestion is supported by the observation that under anaerobic conditions, resting cells degraded only 1.4 mM fluoroacetate compared with 8.5 mM in resting cells incubated aerobically.

Cell cycle distributions show that there was no statistical difference in the percentage of cells in the B phase of growth at specific glycolate utilization rates between 4 × 10⁻¹² and 8 × 10⁻¹² mM/CFU/h for both biofilm and planktonic cells, suggesting similar growth rates (Fig. 7). If growth rates of biofilm cells were decreased in comparison to planktonic cells at similar specific glycolate utilization rates, then an increase in the number of cells occupying the B phase of growth would be expected. This was not the case, and the cells can be considered to be in a similar physiological state. While some biofilm populations had increased percentages of cells in the B phase of the cell cycle, these increases were probably due to carbon limitation as a result of the low carbon loading rates (between 0.06 × 10⁻¹² and 2 × 10⁻¹² mM/CFU/h) in these sections and not as a result of any intrinsic differences between the two modes of growth. The planktonic system was not operated at these low carbon loading rates, and consequently it was not possible to make comparisons between the two modes of growth at these low carbon loading rates.

In conclusion, the TBR was found to be a versatile system for determining the performance of biofilm cells. The TBR allows the determination of local metabolite concentrations and other parameters such as dry weight, CFU count, and biofilm thickness, which can then be used to assess biofilm performance. While other commonly used biofilm reactor systems such as the rotating disk reactor, the capillary biofilm reactor, and the drip flow biofilm reactor allow the determination of some of these parameters, none allows the determination of so many parameters simultaneously. The TBR system does have its limitations; most significantly, it is labor-intensive, which makes replication time-consuming. The identification of the overflow metabolism in biofilm systems could have important implications in the treatment of wastewater streams containing fluorinated compounds. In this case the overflow product was glycolate. Glycolate subsequently acted as the carbon source and did not have any detrimental effect on performance. However, the degradation of other fluorinated compounds, such as fluoroaromatics, might result in the accumulation of toxic fluorometabolites, e.g., fluorocatechol, which could detrimentally affect the performance of biofilm cells. Planktonic cells were found to be superior to biofilm cells at degrading the xenobiotic, with higher fluoroacetate and glycolate utilization rates per CFU recorded at similar specific fluoroacetate loading rates. This difference in performance can be explained by oxygen limitation in the biofilm; however, the high free-fluoride concentrations recorded in the TBR may also have a negative impact on performance. These results show that while planktonic cells were more efficient at utilizing both fluoroacetate and the intermediate metabolite glycolate, the advantages of biofilm systems for the degradation of xenobiotics, such as the enrichment of slow-growing species, may outweigh the superior performance of planktonic cells observed here.