Competition of Escherichia coli O157 with a drinking water bacterial community at low nutrient concentrations

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Competition of Escherichia coli O157 with a drinking water bacterial community at low nutrient concentrations Marius Vital a,b, Frederik Hammes a, Thomas Egli a,b,* a

Eawag, Swiss Federal Institute of Aquatic Science and Technology, Department for Environmental Microbiology, U¨eberlandstrasse 133, P.O. Box 611, CH-8600 Du¨bendorf, Switzerland b Institute of Biogeochemistry and Pollutant Dynamics, ETH Zu¨rich, CH-8092 Zu¨rich, Switzerland

article info

abstract

Article history:

In contrast to studies on (long-term) survival of enteric pathogens in the environment,

Received 2 May 2012

investigations on the principles of their growth and competition with autochthonous

Received in revised form

aquatic bacteria are rare and unexplored. Hence, improved basic knowledge is crucial for

9 August 2012

an adequate risk assessment and for understanding (and avoiding) the spreading of

Accepted 27 August 2012

waterborne diseases. Therefore, the pathogen Escherichia coli O157 was grown in compe-

Available online 14 September 2012

tition with a drinking water bacterial community on natural assimilable organic carbon (AOC) originating from diluted wastewater, in both batch and continuous culture. Growth

Keywords:

was monitored by flow cytometry enabling enumeration of total cell concentration as well

Competition

as specific E. coli O157 detection using fluorescently-labelled antibodies. An enhanced

Escherichia coli O157

competitive fitness of E. coli O157 with higher AOC concentrations, higher temperatures

EHEC

and increased dilution rates (continuous culture) was observed. A classical “opportunist”

Risk assessment

versus “gleaner” relationship, where E. coli O157 is the “opportunist”, specialised for growth

Pathogen growth

at high nutrient concentrations (mmax: 0.87 h
1 and Ks: 489 mg consumed DOC L
1), and the

Assimilable organic carbon (AOC)

bacterial community is the “gleaner” adapted to nutrient-poor environments (mmax: 0.33 h
1 and Ks: 7.4 mg consumed DOC L
1) was found. The obtained competition results can be explained by the growth properties of the two competitors determined in pure cultures and it was possible to model many of the observed dynamics based on Monod kinetics. The study provides new insights into the principles governing competition of an enteric pathogen with autochthonous aquatic bacteria. ª 2012 Elsevier Ltd. All rights reserved.

1.

Introduction

Enteric pathogens are commonly assumed to die-off after they are shed from the host into the natural environment. However, a few reports indicate that some of them, such as (pathogenic) Escherichia coli, are able to survive for a long time or even grow under certain conditions in water and soil (e.g.,

Camper et al., 1991; Ishii et al., 2006, 2007; Vital et al., 2008). There are numerous abiotic and biotic factors controlling the ability of enteric pathogens to grow or survive but our knowledge is substantially limited for most of those factors (Winfield and Groisman, 2003). A good example is competition. In the aquatic and terrestrial environment, enteric pathogens have to compete with other microbes for available

* Corresponding author. Eawag, Swiss Federal Institute of Aquatic Science and Technology, Department for Environmental Microbiology, ¨ eberlandstrasse 133, P.O. Box 611, CH-8600 Du¨bendorf, Switzerland. Tel.: þ41 44 823 51 58; fax: þ41 44 823 55 47. U E-mail address: [email protected] (T. Egli). URL: http://www.eawag.ch 0043-1354/$ e see front matter ª 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.watres.2012.08.043

6280

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

nutrients (Gottschal, 1993; Torsvik et al., 2002; Konopka, 2009). However, controlled in vitro experiments are scarce and, hence, the basic understanding of the principles governing competitive growth of heterotrophs in nature is poor. Microbiologists still rely on concepts established in the period from 1950e1970, where the focus was on “pure and simple competition” under defined conditions, i.e., where a single growth-limiting substrate (usually an organic carbon compound) is available for all competitors and competition for this substrate is the only interaction between the competitors (see Powell, 1958; Fredrickson and Stephanopoulos, 1981). Basic Monod kinetics was used to explain the outcome of competition experiments in such simple systems, despite that no complete experimental kinetic data sets were available (Jannasch, 1967; Harder and Veldkamp, 1971; Veldkamp and Jannasch, 1972; Hansen and Hubbell, 1980). Because of the difficulty to obtain experimental stoichiometric and kinetic data in real environmental systems much of the subsequent research on competition of heterotrophic bacteria was done in silico where different scenarios were theoretically investigated (e.g., Hsu, 1980; Hale and Somolinos, 1983; Gottschal, 1993; Dukan et al., 1996). Growth of heterotrophic bacteria in the aquatic environment is usually carbon-limited (Morita, 1997). Typically, most of the dissolved organic carbon is present as polymers and only a (usually small) fraction of the DOC pool e the so-called assimilable organic carbon (AOC) e is readily available for microbial growth. AOC is not composed of a single substrate but is a mixture of many different individual compounds, all present at very low concentrations (Mu¨nster, 1993). Experimental data from many different model systems suggest that under such conditions heterotrophs perform “mixed substrate growth”, i.e., they use a range of carbon compounds simultaneously (Harder and Dijkhuizen, 1976; Egli, 1995). Hence, growth is not controlled by a single carbon source but is limited by several “homologous” substrates simultaneously (Harder and Dijkhuizen, 1976). Performing “mixed substrate growth” provides the cell with a number of kinetic and physiological advantages (Egli, 2010). Note that this phenomenon should not be confused with simultaneous limitation by non-homologous nutrients such as combinations of nitrogen, phosphorous or carbon, each satisfying a different physiological function (see e.g., Egli, 1991; Huisman and Weissing, 2001; Grover, 2004). Although growth of pure microbial cultures, including pathogens, on natural AOC has been investigated (e.g., van der Kooij et al., 1982; Pomeroy and Wiebe, 2001; Vital et al., 2008; Kirschner et al., 2008) we are not aware of reports specifically focussing on their competitive behaviour. This lack of data is at least partly due to methodological limitations for adequately and quantitatively approach this topic in the laboratory. Particularly the quantification of the growth of natural microbial flora was fraught with difficulties because most of these cells cannot be cultivated on agar plates. New cultivationindependent methods, particularly flow cytometry-based techniques, allow us to follow bacterial growth at very low cell concentrations in environmental samples and to combine this with the specific detection of target organisms (Tanaka et al., 2000; Vital et al., 2007), which makes them ideal for investigating bacterial competition on natural AOC in vitro.

The aim of this work was to improve the knowledge of the principles governing competition of enteric pathogens when competing with a natural bacterial flora; such information is crucial for microbial risk assessment and for understanding the spreading of diseases by enteric pathogens. For this purpose, E. coli O157 was grown in competition with a bacterial community derived from drinking water on natural AOC in both batch and continuous culture. This approach represents a classical “opportunist” versus “gleaner” relationship, where E. coli O157 is the “opportunist”, specialised for growth at high nutrient concentrations, and the bacterial community is the “gleaner” adapted to nutrient-poor environments (Veldkamp and Jannasch, 1972; Grover, 1990). The influence of nutrient concentration, temperature and dilution rate on the outcome of the competition was studied using flow cytometric-based methods.

2.

Material and methods

2.1.

Bacterial strains and pre-cultivation

The verotoxin-negative E. coli O157 (Nent 2540-04) was stored at
80  C. The cryo-culture was streaked onto a Tryptic soy agar plate (Biorad, Reinach, Switzerland) and incubated for 24 h at 37  C. Cells from a single colony were transferred with a loop into ten-times diluted Luria-Bertani (LB) broth and were incubated overnight at 37  C. Subsequently, cells from this overnight culture were transferred into 10,000-times diluted LB medium (starting concentration 5  103 cells mL
1) and incubated for four days at 30  C. The cells were then inoculated into sterile, 100-times diluted wastewater (starting concentration 5  103 cells mL
1) and grown for four days at 30  C before being used as an inoculum. To prepare the drinking water bacterial community (BC), indigenous bacterial communities from both bottled mineral water and unchlorinated tap water (Du¨bendorf, Switzerland) were directly inoculated (initial concentration w3  103 cells mL
1 from each water) and grown together in 100-times diluted wastewater for four days at 30  C. This BC was kept at room temperature and served as a stock culture for all experiments.

2.2.

Preparation of carbon-free materials

Carbon-free glassware (bottles and vials) was prepared as described in Hammes and Egli (2005). In short: all glassware was first washed with a common detergent and thereafter rinsed three-times with deionised water. Then it was submerged overnight in 0.2 N HCl and subsequently rinsed with deionised water again and air-dried. Finally, the bottles and vials were heated in a muffle furnace to 500  C for at least 6 h to burn-off residual carbon compounds. Teflon-coated screw caps for the glassware were washed and treated identically with acid (0.2 N HCl). Caps were then soaked in a 10% sodium persulphate solution (60  C, 1 h), rinsed three-times with deionised water and finally air-dried.

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

2.3.

Flow cytometry

For total cell counts, an aliquot of 10 mL of SYBRgreen (Molecular Probes, Basel, Switzerland), 100-times diluted in dimethylsulfoxid (Fluka Chemie AG, Buchs, Switzerland), was added to 1 mL of a bacterial suspension and incubated for 15 min at room temperature in the dark before analysis. For better permeabilization of the outer membrane EDTA (pH 8) was added (5 mM final concentration) to the sample together with the stain (Berney et al., 2007). If a sample contained more than 106 cells mL
1, it was appropriately diluted before staining. Direct detection of E. coli O157 was achieved using FITC-labelled antibodies (0.5 mg mL
1, KPL, MD, USA) in combination with flow cytometry. Ten-times diluted antibody solution (1 mL) was added to 1 mL of sample. Before analysis, the suspension was incubated for 20 min at room temperature in the dark. No cross-reaction was observed with cells present in the BC. All samples were measured on a CyFlow Space flow cytometer (Partec, Mu¨nster, Germany) equipped with a 200 mW argon laser emitting at a fixed wavelength of 488 nm and equipped with volumetric counting hardware. The trigger was set on the green fluorescence (520 nm) channel and signals for total cell counting were collected on the combined 520 nm/630 nm (red fluorescence from SYBR Green) dot plot (see Hammes et al., 2008). For specific E. coli O157 detection, the signals were collected on the combined 520 nm/sideward scatter (SSC) dot plot. The quantification limit of the instrument was about 1000 cells mL
1 with an average standard deviation of less than 5% (Hammes et al., 2008).

2.4. Determination of the assimilable organic carbon (AOC) and of consumable dissolved organic carbon (cDOC) concentration AOC determination was based on the method published by Hammes and Egli (2005). As described in the original method, the BC was grown into late stationary phase (reached after incubation for four days at 30  C) and the final cell number was measured using flow cytometry. We also determined the DOC concentration after growth of the BC and without growth (sterile control). The consumed DOC, here referred to as “consumable DOC” (cDOC), corresponds to the AOC. This procedure was only applied on 40-fold diluted wastewater; cDOC concentrations in other dilutions of wastewater were calculated from the dilution factor. As E. coli O157 can use only a fraction of the carbon compounds of the BC, identical measurements were performed for the enterobacterium. In our analyses and simulations, the consumed DOC of E. coli O157 was considered to be a part of the total cDOC determined with the BC.

2.5.

Preparation of diluted wastewater

Untreated wastewater (instead of river water or drinking water) was used because the high fraction of AOC in wastewater allowed investigating bacterial growth over a broad range of organic carbon concentrations of the same quality and, at the same time, to quantify the carbon consumed for growth with DOC analysis. Undiluted untreated wastewater (Du¨bendorf, Switzerland) was sterilized by 20 kDa filtration

6281

(Fresenius Medical Care, Bad Homburg, Germany), kept sterile at 4  C and served as a stock for all performed experiments. The DOC after filtration was 51.7 mg L
1 and the cDOC (¼ cDOCBC ¼ AOC) was 36.4 mg L
1, respectively. To prepare the different dilutions, the stored wastewater was diluted with pasteurized (60  C for 30 min) still mineral water (Evian, France) containing no (<10 mg L
1) cDOC. The obtained solutions were then directly filtered (0.22 mm) into sterile 40 mL carbon-free glass vials. From the stock the following diluted wastewater media were prepared (calculated cDOC concentrations in mg L
1 are included in parentheses): 10-fold (3644), 40-fold (911), 100-fold (364), 200-fold (182), 500-fold (73) and 1000-fold (36 mg L
1 cDOC). All continuous culture experiments were performed with diluted wastewater containing 820 mg L
1 cDOC. In order to check whether the wastewater was indeed carbon-limited, an experiment was performed where inorganic nutrients (Ihssen and Egli, 2004) were added to 40-times diluted sterile wastewater in triplicate samples and the solutions were inoculated either with E. coli O157 or the BC. The final cell concentrations reached in batch cultivation in both cultures were not significantly different from control cultures with no added nutrients (data not shown).

2.6.

Non-competitive growth in diluted wastewater

2.6.1.

Batch cultivation

Cells used for inoculation were harvested from stationary phase cultures (see above) and inoculated into carbon-free 40 mL glass vials containing 30 mL of different dilutions of sterile wastewater (starting concentration 3  103 cells mL
1). The cultures were incubated at 30  C and the final cell concentration was enumerated with flow cytometry as described above. All results are given as net growth (in cell number), i.e., after deducting the inoculum concentration. For growth curve experiments bacterial samples (100 mLe1 mL) were collected throughout the growth cycle at different time points until stationary phase was reached. The specific growth rate (m) based on cell number increase was calculated from the initial phase of growth as described earlier (Vital et al., 2007). For investigating growth at 12, 15, 20, 25 and 30  C, respectively, the inocula were cultivated in the corresponding media and at corresponding temperatures. All temperature experiments were performed with sterile diluted wastewater with a cDOC concentration of 364 mg L
1.

2.6.2.

Continuous cultivation

For growth experiments in continuous culture, three reactors (200 mL Schott flasks) containing 100 mL of diluted wastewater were simultaneously fed from the same medium reservoir (5 L Schott flask). The reactors were operated at 30  C for 40e110 h. Experiments were performed at three different dilution rates (D ¼ 0.1 h
1, D ¼ 0.2 h
1 and D ¼ 0.3 h
1, respectively). The reactors were inoculated with stationary phase-grown cells (starting concentration 3  103 cells mL
1), run in the batch-mode to the beginning of the stationary growth phase and then switched to the continuous culture mode. The steady-state, defined by a stable cell concentration over time, was achieved after two to four volume changes. Due to the low cell density no active aeration was necessary and the medium exchange in and out of the reactors was

6282

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

achieved by a peristaltic pumping system. No substantial wall growth was observed that might have affected our experiments.

2.7.

Competition in batch culture

E. coli O157 and the BC were separately pre-grown at corresponding dilutions of sterile wastewater and temperatures, respectively, into exponential growth phase. Cells were then re-inoculated into a new pre-warmed medium at a ratio of 1:1 (3  103 cells mL
1 of each competition partner). The increase in total cell concentration and of E. coli O157 was then determined directly after inoculation and further monitored over the whole competition period using flow cytometry (see above).

2.8.

Competition in continuous culture

To investigate the competition during continuous culture growth, E. coli O157 and the BC were separately pre-grown in

continuous culture at identical dilution rates into steadystate. Then 50 mL of each culture was combined with the competition partner and the culture was monitored as a function of time using flow cytometry (see above). Specific growth rates (m) were calculated from three adjacent points as described earlier (Berney et al., 2006).

2.9.

Simulations

Equations used for simulations are shown in Table 1. All simulations are based on single substrate saturation kinetics established by Monod (1949). However, here the single growth-limiting substrate (s) was replaced by the amount of consumable DOC available for a bacterial culture, namely for E. coli O157 (cDOCE) and for the BC (cDOCBC), where it was assumed that all carbon compounds available for E. coli O157 were also accessed by the BC. Furthermore, the Monod saturation constants (Ks), referred to as KBC and KE, respectively, were adjusted to the amount of DOC available for each culture (i.e., cDOCBC and cDOCE, respectively). No experiments were

Table 1 e Equations used for simulations. The measured parameters are displayed in Table 2 and Fig. 2, respectively. Equation Batch growth and competition mBC ¼ mmaxBC * cDOCBC/(cDOCBC þ KBC) mE ¼ mmaxE * cDOCE/(cDOCE þ KE) DcDOCBC/Dt ¼ cDOCini
((BC#
BC#ini) * YBC þ (E#
E#ini) * YE) DcDOCE/Dt ¼ 0.53 * cDOCini
(0.53 * ((BC#
BC#ini) * YBC) þ (E#
E#ini) * YE) DBC#/Dt ¼ BC# þ (BC# * emBC * Dt
BC#) DE#/Dt ¼ E# þ (E# * emE * Dt
E#)

Equation number 1 2 3 4 5 6

Effect of temperature YBC ¼ 21000 * ( C) þ 9 * 106 YE ¼
13000 * ( C)2 þ 738000 * ( C)
5 * 106 YE was set as YE at 25  C and at temperatures above 25  C. mmaxBC ¼ 0.0134 * ( C)
0.085 mmaxE ¼ 0.0426 * ( C)
0.439

9 10

Continuous culture growth and competition DBC#/Dt ¼ mBC * BC#
D * BC# DE#/Dt ¼ mE * E#
D * E# DcDOCBC/Dt ¼ D * 820
mE * E#/YE
mBC * BC#/YBC
D * cDOCBC DcDOCE/Dt ¼ D * 820
mE * E#/YE e 0.53 * mBC * BC#/YBC
D * cDOCE 2 a ststcDOC ¼ 8300 * D
1580 * D þ 115

11 12 13 14 15

7 8

mBC: specific growth rate for the BC. mE: specific growth rate for E. coli O157. mmaxBC: maximum specific growth rate for the BC. mmaxE: maximum specific growth rate for E. coli O157. KBC: Monod saturation constant for the BC. KE: Monod saturation constant for E. coli O157. YBC: Yield for the BC. YE: Yield for E. coli O157. BC#: cell concentration of the BC. BC#ini: initial cell concentration of the BC. E#: cell concentration of E. coli O157. E#ini: initial cell concentration of E. coli O157. cDOCBC: cDOC for the BC. cDOCE: cDOC for E. coli O157. cDOCini: initial concentration of cDOC. ststcDOC: estimated residual cDOC concentration in second phase during competition. D: dilution rate. a The parameter ststcDOC was not derived from the data presented in Fig. 1, but was estimated (see main text). By definition 53% of the total ststcDOC was considered to be available for E. coli O157.

6283

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

1.2x107

A cells formed (µg cDOC)-1

performed on the influence of the temperature on Ks values of the two competitors because previous reports suggest constant Ks values with changing temperatures (Kovarova et al., 1996; Hall et al., 2008). Hence, the values measured at 30  C were applied for simulations at other temperatures. For simulating the competition in continuous culture in experiments during the second phase where a stable wash-out of E. coli O157 was observed (Figs. 7 and 8), it was assumed that a constant fraction of that pool, namely 53%, was available for E. coli O157, referred to “cDOC steady-state” (ststcDOC). At each dilution rate (D ¼ 0.1 h
1, D ¼ 0.2 h
1 and D ¼ 0.3 h
1) a parameter estimation analysis for ststcDOC was performed yielding the best fit with the experimentally determined washout-rate of E. coli O157. The three estimated values were subjected to regression analysis, resulting in the ststcDOC subsequently used for simulations (equation (15)).

107 8.0x106 6.0x106 4.0x106 2.0x106 0 10

3.1. Determination of stoichiometric and kinetic parameters In order to determine the key growth parameters of E. coli O157 and the drinking water bacterial community (BC), respectively (Table 2), the two competitors were separately grown in batch culture at different dilutions of wastewater at 30  C. The BC was able to utilize 70% of the DOC present, referred to here as

25

30

25

30

25

30

0.40

specific growth rate (h -1)

Results

20 temperature (°C)

B

0.35

3.

15

0.30 0.25 0.20 0.15 0.10 0.05 0.00 10

15

20 temperature (°C)

maximum specific growth rate (h-1)

1.0

specific growth rate (h-1)

0.8

0.6

0.4

0.2

C 0.8

0.6

0.4

0.2

0.0 10

0.0 0

1000

2000

3000

4000

cDOC (µg L-1)

0

530

1060

1590

2120

cDOC (µg L-1) available for E. coli O157

Fig. 1 e “The opportunist” versus “the gleaner” kinetic relationship between E. coli O157 (C) and the drinking water bacterial community (,) at 30  C. The specific growth rates in correlation with the consumable DOC (cDOC) are displayed. For E. coli O157 only a fraction (53%) of the total cDOC is available (for explanation see text). The lines represent the simulated correlation based on Monod kinetics. Corresponding maximum specific growth rates (mmax) and saturation constants (Ks) are presented in Table 2. The error bars represent the standard deviation on triplicate samples.

15

20 temperature (°C)

Fig. 2 e Effect of temperature (12e30  C) on the yield (A) and the specific growth rate (B) of E. coli O157 (C; solid line) and the drinking water bacterial community (,; dashed line) during batch growth with diluted wastewater (100-times, cDOC [ 364 mg L-1). The error bars represent the standard deviation on triplicate samples. Panel C shows the corresponding extrapolated maximum specific growth rates of the two cultures (for explanation see text).

the “consumable DOC” (cDOC, or, when referring to the community also as cDOCBC; for definitions see Materials and Methods). 37.7% of the DOC, equivalent to 53% of cDOC, was available for E. coli O157. Furthermore, the numerical cell yield of the pathogen was only half of that of the BC (Table 2). This is

6284

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

107

A

30

cell concentration (mL-1)

E. coli O157 (% of total cells mL-1)

35

25 20 15 10

106

105

104

5 103

0 0

200

400

600

800

0

1000

20

cDOC (µg L-1)

consistent with previous data, where it was shown that cells of E. coli O157 are bigger than average cells of a freshwater bacterial community and, hence, needed more carbon to produce a new cell (Vital et al., 2008). The modelled kinetic Monod constants, mmax and Ks, are based on the consumable DOC as the growth limiting carbon substrate and demonstrates a clear “opportunist” versus “gleaner” relationship (Fig. 1). The opportunist E. coli O157 displayed a high mmax together with a high half-saturation constant, KE, whereas the BC, originating from nutrient-poor environments, exhibited both, a lower mmax and a lower KBC. The results presented in Fig. 1 suggest a crossing-over of the two curves at a cDOC concentration of 590 mg L
1. As a next step, the effect of temperature on specific growth rate and yield was investigated during batch cultivation for the two cultures in 100-times diluted wastewater containing 364 mg cDOC L
1 (Fig. 2). Whereas E. coli O157 produced considerably fewer cells per mg of consumed DOC with decreasing temperatures, the yield of the BC was hardly affected (Fig. 2A). Both competitors showed a similar increase of their specific growth rate (m) with increasing temperature (Fig. 2B). This suggests that the effect of temperature on specific growth rate has little influence on the competition. However, the experiment was performed at a cDOC concentration where the specific growth rate of the BC at 30  C equalled its mmax, whereas that of E. coli O157 was still below mmax/2 (compare Fig. 1). Hence, the recorded influence of

80

90

40

80

90

100 E. coli O157 (% of total cells mL-1)

Fig. 3 e Batch culture competition between E. coli O157 and a drinking water bacterial community (BC) at different concentrations of cDOC at 30  C. The measured final cell concentrations of E. coli O157 as a percentage of the total cell concentrations (C) are shown. The error bars represent the standard deviation on triplicate samples. The lines represent in silico predictions on competition outcome at different starting cell number ratios of E. coli O157 and the BC (ratios between competitors in parenthesis); solid line (1:1) e original model; short-dashed line (7:3) and longdashed line (3:7). The open circles (B) illustrate the modelled results of individual competition experiments taking into account the experimentally measured initial ratios of the two competitors after inoculation.

40 time (h)

B 80

60

40

20

0 0

20 time (h)

Fig. 4 e Batch culture competition between E. coli O157 (C) and the drinking water bacterial community (,) at a cDOC concentration of 911 mg LL1 and a temperature of 30  C (A) is displayed. Panel B shows the corresponding percentage of E. coli O157 (C) of the total cell concentration. The lines represent the modelled results based on Monod kinetics (compare Table 1). Error bars indicate the standard deviation on triplicate samples.

temperature on m (slope in the experiment) does also reflect the actual influence on mmax of the BC, whereas it had to be mathematically extrapolated for the mmax of E. coli O157 (Fig. 2C). This extrapolation suggests that, based on its mmax, E. coli O157 should have an increasing competitive advantage with rising temperatures.

3.2. Competition in batch culture e influence of the nutrient concentration When inoculating the same cell number into batch cultures, elevated nutrient concentrations enhanced growth of E. coli O157 when competing with the BC (Fig. 3). The higher the cDOC concentration, the higher was the percentage of E. coli O157 of the total cell number at the end of competition experiments (Fig. 3 and A1). At a cDOC concentration of 73 mg L
1 only 0.5% of all cells were detected as E. coli O157, whereas the pathogen comprised 20.5% of all cells at 911 mg cDOC L
1. This corresponds to a 40-fold difference. Note that

6285

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

108

1 10 8 6 4

2

7

10

106 500 105

400 300

104

200 100

103

2

0 0

10

30

40

time (h)

0 15

20

25

30

temperature (°C)

Fig. 5 e The influence of temperature on batch culture competition between E. coli O157 and a drinking water bacterial community (BC) is displayed. The measured final E. coli O157 cell concentrations as a percentage of the total cell concentrations (C) are shown. The dashed line indicates the prediction of the model, whereas the open circles represent the simulated results considering the measured starting ratios of the two competitors at the beginning of each individual competition experiment (see Fig. 3). Diluted wastewater (100-times, cDOC [ 364 mg LL1) was used for the experiment. The error bars represent the standard deviation on triplicate samples.

although the starting cell ratio of the two competitors was always around 1:1, E. coli O157 never represented more than 20.5% of the total cell concentration in stationary phase at all cDOC concentrations tested. According to Fig. 1, one would expect that the pathogen was growing faster than the BC above a cDOC concentration of 590 mg L
1 and, hence, competition experiments above this value should have

107

Fig. 7 e Competition of E. coli O157 (C) and the drinking water bacterial community (BC, ,) in diluted wastewater (cDOC [ 820 mg LL1) at 30  C in continuous culture at a dilution rate of D [ 0.2 hL1. The lines represent the modelled results (solid line: E. coli O157; short dashed line: BC; long dashed line: residual cDOC; dotted-dashed line: residual cDOC available for E. coli O157 and dotted line: theoretical total washout (mE [ 0 hL1)). The Figure is separated into two parts indicating the biphasic behaviour of E. coli O157 during competition with the BC in continuous culture (see text). The error bars represent the range on duplicate samples.

140 specific growth rate (% to D)

10

steady state cell concentration (mL-1)

20

residual cDOC (µg L-1)

12

cell concentration (mL-1)

E. coli O157 (% of total cells mL-1)

14

1

120

2

100 80 60 40 20 0 0

106

2

4

6

8

10

Volume changes

105 0.10

0.15

0.20

0.25

0.30

dilution rate (h-1)

Fig. 6 e Steady-state cell concentrations of either E. coli O157 (C) or the drinking water bacterial community (BC; ,) during growth in continuous culture with diluted wastewater (cDOC concentration [ 820 mg LL1) at different dilution rates. The error bars represent the standard deviation on triplicate samples.

Fig. 8 e Specific growth rates m relative to the dilution rate of E. coli O157 growing in competition with the drinking water bacterial community in continuous culture at 30  C in diluted wastewater (cDOC [ 820 mg LL1). For calculations of m see Materials and methods. The results are based on competition experiments at three different dilution rates (D [ 0.3 hL1: ; D [ 0.2 hL1: C and D [ 0.1 hL1: D). For cell numbers see Fig. 7, S3 and S4. The lines represent the modelled results (dotted line: D [ 0.3L1; dashed line: D [ 0.2L1 and solid line: D [ 0.1L1). The Figure is separated into two parts indicating the biphasic competition dynamics of E. coli O157 and the BC (see text). The error bars represent the range on duplicate samples.

6286

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

Table 2 e Basic stoichiometric and modelled kinetic parameters of the two cultures at 30  C. Parameter
1 a

DOC consumed (mg L ) DOC consumed (% of total DOC)a Cells formed ( 106 mL
1)a Yield (cells formed  106 (mg cDOC
1))a mmax (h
1)b KE and KBC (mg cDOC L
1)b

E. coli O157

BC

486.5  19.8 37.7  1.5 2.78  0.22 5.72  0.68 0.87 489

911  4.9 70.5  0.4 10  0.7 11  0.52 0.33 7.4

BC: Drinking water bacterial community. mmax: maximum specific growth rate. KE and KBC: saturation constants for E. coli O157 (KE) and the BC (KBC); the values were calculated considering only the fraction of DOC available for growth. a Values obtained from 40-fold diluted wastewater (DOC ¼ 1.29 mg L
1). b Parameters based on simulations using Monod kinetics.

favoured growth of E. coli O157. Indeed, at a concentration of 911 mg cDOC L
1, E. coli O157 exhibited a higher m than the BC and after 12 h the enterobacterium consequently accounted for 85% of the total cell concentration (Fig. 4). However, since only 53% of the cDOC was available for the pathogen, the BC continued to grow when growth of E. coli O157 had already stopped. Thus, the proportion of E. coli O157 of the total cell concentration was finally reduced to around 20%. Based on the measured parameters of the competitors in pure culture (Table 2), it was possible to predict the influence of cDOC concentration on competition outcome; the model allowed to explain the experimentally derived data (solid lines in Fig. 3 and A1). Simulations on the influence of different initial cell number ratios of the two competitors at the start of a batch competition suggests a considerable influence of this parameter on the outcome of a competition experiment (dashed lines in Fig. 3). Since, in the performed experiments the starting ratios indeed varied (they were not always exactly 1:1), additional simulations for individual competition experiment were performed, which took the measured initial ratios of competitors into account. This improved in silico predictions (open circles in Fig. 3 and A1).

3.3. Competition in batch culture e influence of temperature Increasing temperature had a positive influence on the competitive ability of E. coli O157 in batch culture (Fig. 5). At 12  C the pathogen contributed to only 0.5% of the total cells in stationary phase, whereas 10.9% of the total population were detected as E. coli O157 at 30  C. This corresponds to a 20-fold difference (see also Fig. A2). The established model was able to describe the measured competition outcome.

3.4.

Growth in continuous culture

Steady-state cell concentrations of the two competitors growing separately at different dilution rates are displayed in Fig. 6. E. coli O157 established lower steady-state cell concentrations than the BC due to both its lower yield and the ability to consume only 53% of the cDOC (see above). Steady-state cell

concentrations of both cultures correlated negatively with the dilution rate (D). This observation can partly be explained by Monod kinetics shown in Fig. 1 and partly by the substrate utilization pattern during growth in complex media. With respect to the former aspect, Monod kinetics imply that the residual concentration of cDOC in steady-state is increasing with increasing D. Consequently, less carbon is available for biomass production at a high D than at lower growth rates (also compare modelled residual cDOC concentrations at the different dilution rates from Fig. 7, A3 and A4). This effect should be particularly pronounced in the culture of E. coli O157 because KE was considerably higher than KBC. Higher resulting biomass concentrations at lower dilution rates can also be attributed to the complex composition of the organic carbon pool in wastewater where an increasing number of carbon sources becomes available at reduced dilution rates. In other words, one expects that slowly growing cells have access to a broader cDOC spectrum in comparison to fast growing ones (Egli et al., 1986).

3.5.

Competition in continuous culture

During competition experiments in continuous culture the BC was expected to control the residual cDOC concentration in the reactor because of its kinetic properties (Fig. 1). Fig. 7 presents the outcome of the competition at D ¼ 0.2 h
1. The competitors were grown separately to steady-state and then mixed at a ratio of 50 mL:50 mL. It should be pointed out that the cell concentration of E. coli O157 was lower than that of the BC at the beginning of the competition, because the bacterium produced a lower cell concentration in steady-state than the BC (compare Fig. 6). Two distinctly different phases were recognized. At the beginning of the competition, the cell concentration of E. coli O157 was only slightly affected and this was followed by a phase where stable washout of E. coli O157 occurred (Fig. 7). This behaviour of E. coli O157 can be explained in the following way: at the start of the experiments half of the liquid in the reactor stemmed from the E. coli O157 culture, which contained an elevated residual cDOC concentration; the residual cDOC concentration was, thus, higher than that required by the BC to grow at this D. In addition, the cell concentration of the BC was reduced to 50% of its original steady-state cell concentration. This low initial BC cell concentration together with the high residual cDOC concentration allowed stable growth of E. coli O157 with a m almost equal to D at the beginning of the competition experiments. The increase of the BC cell concentration that followed resulted in a decrease of the residual cDOC concentration approaching steady-state concentration of the BC, which consequently promoted washout of E. coli O157 at a constant rate (Fig. 7). The washout-rate of E. coli O157 was lower than theoretical wash-out considering zero growth (mE ¼ 0 h
1); this suggests that also when competing with the BC the pathogen was still able to consume cDOC and grow slowly (however, lower than D). The effect of the dilution rate on the competition between E. coli O157 and the BC is shown in Fig. 8. At all dilution rates investigated (D ¼ 0.1 h
1, D ¼ 0.2 h
1 and D ¼ 0.3 h
1, respectively), the biphasic competition dynamics was observed (as discussed above). Furthermore, a correlation

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

between dilution rate and the competitive fitness of E. coli O157 was found. The higher the D, the better E. coli O157 was able to compete. The model was only partly able to predict the results observed. At D ¼ 0.2 h
1, the model predicted the experimental results (Fig. 7), whereas the simulation at D ¼ 0.1 h
1 underestimated the slow washout of E. coli O157 in the first phase (Fig. 7; A3). At D ¼ 0.3 h
1 the model failed to predict the experimentally derived data (Fig. 7 and A4). Over the whole period of the experiment the simulation predicted a slow increase in the E. coli O157 cell concentration at this dilution rate (mE slightly higher than D), which was in contradiction to the observed washout of the pathogen. The dilution rate of 0.3 h
1 is approaching mmax of the BC (0.33 h
1) and one would, hence, predict the initial increase in cell concentration to be very slow, which should result in a high residual cDOC concentration over a long time, promoting considerable growth of E. coli O157 (Fig. A4). However, the observed initial increase in cell concentration of the BC from the experimental data was faster (m ¼ 0.38 h
1; Fig. A4), which might explain the discrepancy between the prediction of the simulation and the experimental results. In this respect it should be pointed out that the composition of the BC is probably not stable but has to be considered as a changing entity, which is continuously adapting by selection to prevailing growth conditions. Bacterial communities, cultivated under similar conditions as done in this study, were displaying considerably higher specific growth rates (m) than the mmax established here (Vital et al., 2007; Wang et al., 2009). Hence, the fact that the BC displayed transiently a higher m at D ¼ 0.3 h
1 than the predicted mmax, can be ascribed to adaptation/selection processes during cultivation at a high specific growth rate.

4.

Discussion

This study provides substantial new insights into the principles of competitive growth between enteric pathogens and autochthonous aquatic bacteria. The results demonstrate that E. coli O157 is not only able to proliferate at low substrate concentrations in pure culture, as reported earlier (Vital et al., 2008), but that it can also grow in competition with bacteria originating from nutrient-poor environments. However, growth of E. coli O157 was drastically restricted in the presence of competing bacteria. A number of earlier studies reported that autochthonous bacteria considerably limit growth of enteric bacteria in the environment (e.g., Camper et al., 1985; van Elsas et al., 2007). However, the factors involved, specifically the growth properties of competitors were only vaguely considered leaving the underlying competition dynamics in the dark (van Elsas et al., 2011). This study demonstrates that a main drawback of E. coli O157 was its inferior kinetic properties for competing at low carbon concentrations with a natural bacterial community (Fig. 1). The established kinetic relationship between the competitors, also known as the “opportunist” versus the “gleaner” relationship, was often proposed for autochthonous microorganisms and enteric bacteria (e.g., Jannasch, 1967; Grover, 1990), but has to our knowledge not yet been verified in the laboratory. This relationship implies that competitive fitness of E. coli O157 is correlated with both increasing

6287

nutrient concentrations and increasing growth rates, which was indeed observed in our experiments (Figs. 3 and 8). Several in situ observations confirm this finding and have indicated a positive effect of elevated nutrient concentrations on growth of Enterobacteriaceae in water. For example, both LeChevallier et al. (1996) and Camper et al. (2000) reported a positive correlation between the assimilable organic carbon (AOC) concentration and the occurrence of coliform bacteria in different drinking water systems during a broad survey across North America. Next to nutrients, temperature is a key factor governing heterotrophic growth and competition in the environment. It is a well-known fact that growth of E. coli is positively influenced by enhanced temperatures (e.g., Raghubeer and Matches, 1990) and it was even suggested that this enterobacterium is part of the indigenous microbial flora in the tropics due to elevated average water temperatures (Winfield and Groisman, 2003). However, in this study it was shown for the first time that elevated temperatures are directly enhancing growth of an enteric pathogen by increasing its competitive fitness; the results can be explained by the influence of temperature on both the yield and mmax of competitors (Fig. 4). The results support the suggestion that rising annual water temperatures enhance growth of enteric pathogens in aquatic ecosystems (IPCC, 2007). However, it should be pointed out that the BC used here originated from moderately tempered environments (<20  C; Wang et al., 2008; Hammes et al., 2010). Although it was adapted to the individual temperatures (one batch cultivation cycle is equivalent to w10 generations), long-term adaptation processes might increase its competitive fitness at high temperatures. We demonstrated that competition dynamics under environmental nutritional conditions can be described by the measured growth properties of the individual competitors in pure culture, namely (1) the availability of DOC in batch and continuous culture, respectively, (2) the yield and (3) the kinetic parameters. It was possible to mathematically describe most of the observed competition experiments using conventional single substrate kinetics based on Monod (Monod, 1949). This is astonishing, because Monod based his concepts on experiments performed with pure cultures growing with a single limiting carbon/energy source. In contrast, the nutrient pool used in this study was composed of many different carbon substrates simultaneously limiting growth (Egli, 1995). We ignored these complexities and simply replaced the concentration of the single growth-limiting substrate (s) in Monod’s model with the concentration of consumable DOC for the competitors. Our results suggest that mathematical modelling combined with kinetic and stoichiometric growth parameters obtained under realistic growth conditions can, hence, provide an opportunity to postulate different scenarios on competitive growth in the environment. An interesting aspect to explore in silico would be the influence of varying nutrient supply during competition in continuous culture (see e.g., Hsu, 1980; Hale and Somolinos, 1983). The experiments in this study were performed by continuously feeding the reactors with an identical concentration of nutrients. However, in the environment a discontinuous supply of nutrients is the rule rather than the exception, which considerably alters

6288

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

competition dynamics. It was suggested that nutrient oscillations under certain conditions lead to stable growth of “opportunists” competing with “gleaners” (Grover, 1990; Gottschal, 1993). Such observations suggest that discontinuous nutrient supply may enhance growth of enteric pathogens. Furthermore, it would be interesting to investigate how adaptive and selective processes influence the competitive fitness of E. coli O157. For example, it was shown for a population of E. coli K12 that extended growth in a glucose-limited continuous culture system resulted in a reduction of Ks from 400 to 500 mg glucose L
1 down to 30 mg glucose L
1 by selection of high-glucose-affinity mutants (Senn et al., 1994; Wick et al., 2002). Implementation of an equivalent adaptation process in the here presented simulation would change the Ks of E. coli from 489 mg consumed DOC L
1 to a value approaching that of the BC. As a result the crossing point of the two competitors would be considerable lowered (Fig. 1) and the outcome of competition experiments would be dramatically altered in favour of E. coli O157. However, it is unclear whether E. coli has the same adaptation potential during growth in the environment, where probably a multitude of substrates are limiting growth simultaneously and, thus, the selection pressure of an individual carbon compound is considerably reduced. So far environmental isolates of E. coli from fresh- and drinkingwater exhibited very similar Ks values for glucose as the unadapted strain of E. coli K12 indicating no adaptation to or selection of high-affinity mutants under environmental conditions (Ihssen et al., 2007). The results presented indicate considerable competitive fitness of the enteropathogen E. coli O157 when provided with carbon of natural origin. However, for drinking water conditions, our experiments represent probably a ‘worst case scenario’. In our experimental set-up we started with an initial cellular ratio of 1:1 of the two competitors. In natural environments or during drinking water treatment and distribution this is rather unlikely; here, pathogens are usually outnumbered by far by non-pathogenic bacterial cells, even during pollution events. Therefore, for batch culture conditions we investigated the influence of more realistic starting ratios in silico by altering the initial concentration of the bacterial community (BC) over a broad range of three log units (Fig. A5). As expected, the simulations suggest that an increase in the initial BC cell concentration would lead to considerably reduced growth of the pathogen (Fig. A5). The results imply that in waters containing a well-adapted microbial community acting as a “placeholder”, i.e., of a community that occupies all nutritional niches, growth of enteric pathogens should either be severely hampered, or should even not occur. However, disturbing events such as a transient increase in nutrient concentration, a significant change in nutrient composition (e.g., due to pollution), or a reduction or structural alteration of the “placeholder community” (e.g., as a result of disinfection measures) might generate a risk for pathogen growth by opening a nutritional niche. Improved competitive behaviour of E. coli O157 due to transiently changing environmental conditions was not only recorded under batch culture conditions but also during competitive growth with a BC in continuous culture. Whereas a steady wash-out of E. coli O157 occurred in the second phase of the experiments, representing a stable ecosystem dominated by

a well-adapted community acting as a placeholder, no or only a slight wash-out of the pathogen was observed in the first phase, which can be considered as an unstable environment because residual cDOC concentrations were not at steady state yet (Fig. 7, A3 and A4). Such considerations are specifically relevant for drinking water treatment and distribution. Often, the strategy prevails to get rid-off all bacteria present in treated drinking water through a final disinfection step. However, by killing or removing all “placeholders” (and leaving behind all unused growth-supporting carbon compounds), uncontrolled regrowth can create the opportunity for unwanted bacteria such as pathogens to proliferate. Interestingly, a surprisingly high concentration of E. coli was reported in biofilms of a drinking water distribution system that uses a final chlorination step during treatment and greatly reduces autochthonous bacteria in the water right before leaving the treatment plant; the treated drinking water contained high DOC concentrations, which led to a lack of disinfectant residuals with increasing residence time of the water during distribution, and hence, an open niche for growth of E. coli (Juhna et al., 2007). Similarly, Wang et al. (2008) showed that industrial-scale ultrafiltration of groundwater during drinking water production removed most, but not all, indigenous bacteria resulting in a subsequent uncontrolled re-growth after filtration; this led to a drastically altered community structure from its original pattern and, hence, also introduced a potential risk for the growth of unwanted microbes. The production of so-called “biologically stable” water includes not only best possible elimination of carbonaceous nutrients (AOC) but at the same time ensures the presence of a viable autochthonous microbial community acting as “placeholders” (van der Kooij, 1992; Hammes et al., 2010), which might be the combination of choice to close the niche for pathogen growth. However, this does not propose the omission of disinfection steps during drinking water treatment, but suggests that the subsequent, unavoidable re-growth on organic carbon should happen in a controlled manner.

Acknowledgements Many thanks to Friedhelm Steinhilber for help with the simulations. This study was funded by TECHNEAU (6th European Framework Program, 018320).

Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.watres.2012.08.043.

references

Berney, M., Weilenmann, H.U., Ihssen, J., Bassin, C., Egli, T., 2006. Specific growth rate determines the sensitivity of Escherichia coli to thermal, UVA, and solar disinfection. Applied and Environmental Microbiology 72 (4), 2586e2593.

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

Berney, M., Hammes, F., Bosshard, F., Weilenmann, H.U., Egli, T., 2007. Assessment and interpretation of bacterial viability by using the LIVE/DEAD BacLight kit in combination with flow cytometry. Applied and Environmental Microbiology 73 (10), 3283e3290. Camper, A.K., LeChevallier, M.W., Broadaway, S.C., McFeters, G.A., 1985. Growth and persistence of pathogens on granular activated carbon filters. Applied and Environmental Microbiology 50 (6), 1378e1382. Camper, A.K., McFeters, G.A., Characklis, W.G., Jones, W.L., 1991. Growth kinetics of coliform bacteria under conditions relevant to drinking water distribution systems. Applied and Environmental Microbiology 57 (8), 2233e2239. Camper, A.K., Butterfield, P.W., Ellis, B.D., Jones, W.L., Huck, P.M., Anderson, W.B., Volk, C., Welch, N.J., LeChevallier, M.W., 2000. Investigation of the Biological Stability of Water in Treatment Plants and Distribution Systems. American Water Works Association Research Foundation. Dukan, S., Levi, Y., Piriou, P., Guyon, F., Villon, P., 1996. Dynamic modelling of bacterial growth in drinking water networks. Water Research 30 (9), 1991e2002. Egli, T., Bosshard, C., Hamer, G., 1986. Simultaneous utilization of methanol-glucose mixtures by Hansenula polymorpha in chemostat: influence of dilution rate and mixture composition on utilization pattern. Biotechnology and Bioengineering 28 (11), 1735e1741. Egli, T., 1991. On multiple-nutrient-limited growth of microorganisms, with special reference to dual limitation by carbon and nitrogen substrates. Antoine van Leeuwenhoek 60 (3e4), 225e234. Egli, T., 1995. The ecological and physiological significance of the growth of heterotrophic microorganisms with mixtures of substrates. Advances in Microbial Ecology 14, 305e386. Egli, T., 2010. How to live at very low substrate concentrations. Water Research 44 (7), 4826e4837. Fredrickson, A.G., Stephanopoulos, G., 1981. Microbial competition. Science 213 (4511), 972e979. Gottschal, J.C., 1993. Growth kinetics and competition e some contemporary comments. Antoine van Leeuwenhoek 63 (3e4), 299e313. Grover, J.P., 1990. Resource competition in a variable environment, phytoplankton growing according to Monod’s model. American Naturalist 136 (6), 771e789. Grover, J.P., 2004. Predation, competition, and nutrient recycling, a stoichiometric approach with multiple nutrients. Journal of Theoretical Biology 229 (1), 31e43. Hall, E.K., Neuhauser, C., Cotner, J.B., 2008. Toward a mechanistic understanding of how natural bacterial communities respond to changes in temperature in aquatic ecosystems. The ISME Journal 2 (5), 471e481. Hale, J.K., Somolinos, A.S., 1983. Competition for fluctuating nutrient. Journal of Mathematical Biology 18 (3), 255e280. Hammes, F.A., Egli, T., 2005. New method for assimilable organic carbon determination using flow-cytometric enumeration and a natural microbial consortium as inoculum. Environmental Science and Technology 39 (9), 3289e3294. Hammes, F., Berney, M., Wang, Y., Vital, M., Koster, O., Egli, T., 2008. Flow-cytometric total bacterial cell counts as a descriptive microbiological parameter for drinking water treatment processes. Water Research 42 (1e2), 269e277. Hammes, F., Berger, C., Ko¨ster, O., Egli, T., 2010. Assessing biological stability of drinking water without disinfectant residuals in a full-scale water supply system. Journal of Water Supply: Research and Technology e AQUA 59 (1), 31e40. Hansen, S.R., Hubbell, S.P., 1980. Single-nutrient microbial competition, qualitative agreement between experimental and theoretically forecast outcomes. Science 207 (4438), 1491e1493.

6289

Harder, W., Veldkamp, H., 1971. Competition of marine psychrophilic bacteria at low temperatures. Antoine van Leeuwenhoek 37 (1), 51e63. Harder, W., Dijkhuizen, L., 1976. Mixed substrate utilization. In: Dean, A.C.R., Ellwood, D.C., Evans, G.G.T., Melling, J. (Eds.), Continuous Culture 6, Applications and New Fields. Ellis Horwood, Chichester, pp. 297e314. Huisman, J., Weissing, F.J., 2001. Fundamental unpredictability in multispecies competition. American Naturalist 157 (5), 488e494. Hsu, S.B., 1980. A competition model for a seasonally fluctuating nutrient. Journal of Mathematical Biology 9 (2), 115e132. Ihssen, J., Egli, T., 2004. Specific growth rate and not cell density controls the general stress response in Escherichia coli. Microbiology SGM 150 (6), 1637e1648. Ihssen, J., Grasselli, E., Bassin, C., Francoise, P., Piffaretti, J.C., Ko¨ster, W., Schrenzel, J., Egli, T., 2007. Comparative genomic hybridization and physiological characterization of environmental isolates indicate that significant ecophysiological properties are highly conserved in the species Escherichia coli. Microbiology SGM 153 (7), 2052e2066. IPCC, 2007. Climate change 2007, the physical science basis. In: Solomon, S., Quin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University, Cambridge UK. Ishii, S., Ksoll, W.B., Hicks, R.E., Sadowsky, M.J., 2006. Presence and growth of naturalized Escherichia coli in temperate soils from Lake Superior watersheds. Applied and Environmental Microbiology 72 (1), 612e621. Ishii, S., Hansen, D.L., Hicks, R.E., Sadowsky, M.J., 2007. Beach sand and sediments are temporal sinks and sources of Escherichia coli in Lake Superior. Environmental Science and Technology 41 (7), 2203e2209. Jannasch, H.W., 1967. Enrichments of aquatic bacteria in continuous culture. Archives in Microbiology 59 (1e3), 165e173. Juhna, T., Birzniece, D., Larsson, S., Zulenkovs, D., Sharipo, A., Azevedo, N.F., Menard-Szczebara, F., Castagnet, S., Feliers, C., Keevil, C.W., 2007. Detection of Escherichia coli in biofilms from pipe samples and coupons in drinking water distribution networks. Applied and Environmental Microbiology 73 (22), 7456e7464. Kirschner, A.K., Schlesinger, J., Farnleitner, A.H., Hornek, R., Suss, B., Golda, B., Herzig, A., Reitner, B., 2008. Rapid growth of planktonic Vibrio cholerae non-O1/non-O139 strains in a large alkaline lake in Austria, dependence on temperature and dissolved organic carbon quality. Applied and Environmental Microbiology 74 (7), 2004e2015. Konopka, A., 2009. What is microbial community ecology? The ISME Journal 3 (11), 1223e1230. Kovarova, K., Zehnder, A.J., Egli, T., 1996. Temperature-dependent growth kinetics of Escherichia coli ML 30 in glucose-limited continuous culture. Journal of Bacteriology 178 (15), 4530e4539. LeChevallier, M.W., Welch, N.J., Smith, D.B., 1996. Full-scale studies of factors related to coliform regrowth in drinking water. Applied and Environmental Microbiology 62 (7), 2201e2211. Monod, J., 1949. The growth of bacterial cultures. Annual Reviews of Microbiology 3, 371e394. Morita, R.Y., 1997. Bacteria in Oligotrophic Environments. Chapman and Hall, New York. Mu¨nster, U., 1993. Concentrations and fluxes of organic carbon substrates in the aquatic environment. Antoine van Leeuwenhoek 63 (3e4), 243e274. Pomeroy, L.R., Wiebe, W.J., 2001. Temperature and substrates as interactive limiting factors for marine heterotrophic bacteria. Aquatic Microbial Ecology 23 (2), 187e204.

6290

w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 6 2 7 9 e6 2 9 0

Powell, E.O., 1958. Criteria for the growth of contaminants and mutants in continuous culture. Journal of General Microbiology 18 (1), 259e268. Raghubeer, E.V., Matches, J.R., 1990. Temperature range for growth of Escherichia coli stereotype O157, H7 and selected coliforms in E. coli medium. Journal of Clinical Microbiology 28 (4), 803e805. Senn, H., Lendenmann, U., Snozzi, M., Hamer, G., Egli, T., 1994. The growth of Escherichia coli in glucose-limited chemostat cultures, a re-examination of the kinetics. Biochimica et Biophysica Acta 2101 (3), 424e436. Tanaka, Y., Yamaguchi, N., Nasu, M., 2000. Viability of Escherichia coli O157:H7 in natural river water determined by the use of flow cytometry. Journal of Applied Microbiology 88 (2), 228e236. Torsvik, V., Ovreas, L., Thingstad, T.F., 2002. Prokaryotic diversity e magnitude, dynamics, and controlling factors. Science 296 (5570), 1064e1066. van der Kooij, D., Oranje, J.P., Hijnen, W.A., 1982. Growth of Pseudomonas aeruginosa in tap water in relation to utilization of substrates at concentrations of a few micrograms per liter. Applied and Environmental Microbiology 44 (5), 1086e1095. van der Kooij, D., 1992. Assimilable organic carbon as indicator for bacterial regrowth. American Water Works Association 84 (2), 57e65. van Elsas, J.D., Hill, P., Chronakova, A., Grekova, M., Topalova, Y., Elhottova, D., Kristufek, V., 2007. Survival of genetically

marked Escherichia coli O157:H7 in soil as affected by soil microbial community shifts. The ISME Journal 1 (3), 204e214. van Elsas, J.D., Alexander, V.S., Rodrigo, C., Jack, T.T., 2011. Survival of Escherichia coli in the environment: fundamental and public health aspects. The ISME Journal 5 (2), 173e183. Veldkamp, H., Jannasch, H.W., 1972. Mixed culture studies with the chemostat. Journal of Applied Chemistry and Biotechnology 22 (1), 105e123. Vital, M., Fu¨chslin, H.P., Hammes, F., Egli, T., 2007. Growth of Vibrio cholerae O1 Ogawa Eltor in freshwater. Microbiology SGM 153 (7), 1993e2001. Vital, M., Hammes, F., Egli, T., 2008. Escherichia coli O157 can grow in natural freshwater at low carbon concentrations. Environmental Microbiology 10 (9), 2387e2396. Wang, Y., Hammes, F., Egli, T., 2008. The impact of industrialscale cartridge filtration on the native microbial communities from groundwater. Water Research 42 (16), 4319e4326. Wang, Y., Hammes, F., Boon, N., Chami, M., Egli, T., 2009. Isolation and characterization of low nucleic acid LNA-content bacteria. The ISME Journal 3 (8), 889e902. Wick, L.M., Weilenmann, H.U., Egli, T., 2002. The apparent clocklike evolution of Escherichia coli in glucose-limited chemostats is reproducible at large but not at small population sizes and can be explained with Monod kinetics. Microbiology SGM 148 (9), 2889e2902. Winfield, M.D., Groisman, E.A., 2003. Role of non-host environments in the lifestyles of Salmonella and Escherichia coli. Applied and Environmental Microbiology 69 (7), 3687e3694.