Flow-cytometric total bacterial cell counts as a descriptive microbiological parameter for drinking water treatment processes

ARTICLE IN PRESS WAT E R R E S E A R C H

42 (2008) 269 – 277

Available at www.sciencedirect.com

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

Flow-cytometric total bacterial cell counts as a descriptive microbiological parameter for drinking water treatment processes Frederik Hammesa, Michael Berneya, Yingying Wanga,c, Marius Vitala,c, Oliver Ko¨sterb, Thomas Eglia,c, a

Swiss Federal Institute of Aquatic Science and Technology (Eawag), P.O. Box 611, U¨berlandstr. 133, CH-8600 Du¨bendorf, Switzerland Wasserversorgung Zurich (WVZ), Hardhof 9, Postfach, CH-8023 Zu¨rich, Switzerland c Institute of Biogeochemistry and Pollutant Dynamics, ETH Zurich, CH-8092 Zu¨rich, Switzerland b

art i cle info

ab st rac t

Article history:

There are significantly more microbial cells in drinking water than what can be cultured on

Received 27 April 2007

synthetic growth media. Nonetheless, cultivation-based heterotrophic plate counts (HPCs)

Received in revised form

are used worldwide as a general microbial quality parameter in drinking water treatment

4 July 2007

and distribution. Total bacterial cell concentrations are normally not considered during

Accepted 10 July 2007

drinking water treatment as a design, operative or legislative parameters. This is mainly

Available online 14 July 2007

because easy and rapid methods for quantification of total bacterial cell concentrations

Keywords: Adenosine tri-phosphate (ATP) Drinking water Flow cytometry Heterotrophic plate counts (HPCs) Total cell concentration

have, up to now, not been available. As a consequence, the existing lack of data does not allow demonstrating the practical value of this parameter. In this study, we have used fluorescence staining of microbial cells with the nucleic acid stain SYBRs Green I together with quantitative flow cytometry (FCM) to analyse total cell concentrations in water samples from a drinking water pilot plant. The plant treats surface water (Lake Zu¨rich) through sequential ozonation, granular active carbon (GAC) filtration and membrane ultrafiltration (UF). The data were compared with adenosine tri-phosphate (ATP) measurements and conventional HPCs performed on the same water samples. We demonstrated that the impact of all three major treatment steps on the microbiology in the system could accurately be described with total cell counting: (1) ozonation caused chemical destruction of the bacterial cells; (2) GAC filtration facilitated significant regrowth of the microbial community; and (3) membrane UF physically removed the bacterial cells from the water. FCM typically detected 1–2 log units more than HPC, while ATP measurements were prone to interference from extracellular ATP released during the ozonation step in the treatment train. We have shown that total cell concentration measured with FCM is a rapid, easy, sensitive and importantly, a descriptive parameter of several widely applied drinking water treatment processes. & 2007 Elsevier Ltd. All rights reserved.

Corresponding author. Tel.: +41 44 823 5158; fax: +41 44 823 5547.

E-mail address: [email protected] (T. Egli). URL: http://www.eawag.ch 0043-1354/$ - see front matter & 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2007.07.009

ARTICLE IN PRESS 270

1.

WA T E R R E S E A R C H

Introduction

Bacteria are autochthonous in drinking water. High concentrations of planktonic bacterial cells (41  104 cells mL1) of broadly diverse populations are commonly found in both bottled mineral water and tap water on consumption (Leclerc and Moreau, 2002; Hoefel et al., 2003; Rinta-Kanto et al., 2004; Berry et al., 2006). These free-living bacterial communities found in water comprise microorganisms that have proliferated predominantly in biofilms, either during the drinking water treatment processes itself or in pipes during distribution of the water. The primary objectives of drinking water treatment (from a microbiological perspective) are to ensure the absence of any pathogenic bacteria in the finished product and to limit any uncontrolled regrowth during distribution of the water. To this end, treatment systems often make use of multiple hygienic barriers (e.g., ozonation, pre-chlorination, membrane filtration or UV disinfection) during treatment, while in several countries disinfection residuals (e.g., chlorine) are additionally added to the water before distribution (Von Gunten, 2003). However, some European countries, mainly The Netherlands, Germany and Switzerland, aim towards the production of high-quality drinking water where distribution without disinfection residuals can be done, primarily through control of growthlimiting substrates (Van der Kooij, 2002). Irrespective of the different disinfection processes, it is common for bacteria to regrow during treatment and distribution, and concentrations in the range of 104–105 cells mL1 of diverse microbial populations are normal in drinking water (Hoefel et al., 2003; Rinta-Kanto et al., 2004). It is extremely important to understand (and thus be able to control) general bacterial growth during drinking water treatment and distribution. Uncontrolled and excessive growth can lead to a deterioration of aesthetic water quality such as undesirable tastes, odours and visual turbidity, and can also lead to process malfunctioning such as clogging of filters, bio-fouling and biocorrosion (Lee et al., 1980). A proper understanding of microbial survival and growth during drinking water treatment and distribution starts with the ability to quantify all the microorganisms accurately and rapidly. For this purpose, the cultivation-dependant method of heterotrophic plate counts (HPCs) has been introduced more than 100 years ago (reviewed in, e.g., Bartram et al. (2003) and Sartory (2004)). HPC still remains the primary parameter for assessment of the general microbiological quality of drinking water, even though it has become dogged by questions about what the method actually measures, how the method should be performed and how the results should be interpreted (Allen et al., 2004; Sartory, 2004). Most critically, in what became commonly known as ‘‘the great plate count anomaly’’, HPC detects only a small fraction of the total bacterial cells in aquatic environments (Staley and Konopka, 1985; Allen et al., 2004). Even though the immense discrepancy between total bacterial concentrations and cultivable bacterial cell concentrations (HPCs) in aquatic samples has been shown numerous times (Staley and Konopka, 1985), the former is still not included in drinking water legislation, guidelines or everyday

42 (2008) 269– 277

operational decision-making. This could be attributed to the fact that, up to now, total cell counting was mainly restricted to microscopy (Greenberg et al., 1993; Rinta-Kanto et al., 2004). Despite recent advances in epi-fluorescence microscopes, instrument automation, multi-colour dyes and digital image analysis, microscopy remains a time- and labour-consuming method. Flow cytometry (FCM) has tremendous potential as an alternative tool for the analysis of bacteria in drinking water. This method can be used for direct enumeration of the total cell concentrations in water (Hammes and Egli, 2005; Lebaron et al., 1998), staining and detection of specific cellular features such as viability (Hoefel et al., 2003; Phe et al., 2005; Berney et al., 2006), or staining and enumeration of specifically targeted cells with antibodies (Clarke and Pinder, 1998). A lucrative feature of FCM is that it is fast, accurate and quantitative. In this study, we have used fluorescent labelling of microbial nucleic acids coupled with quantitative FCM to assess the total cell concentrations after different treatment processes in a drinking water pilot plant that processes lake water. We have compared the data obtained with FCM with conventional methods (HPC) and also with adenosine triphosphate (ATP) concentrations in the samples. The aim was to assess the use of FCM to obtain total microbial cell concentration data from drinking water samples rapidly, and to evaluate these data within the context of describing the major steps involving disinfection and regrowth in the treatment plant.

2.

Materials and methods

2.1.

Preparation of sampling glassware

Borosilicate glass sampling bottles (250 mL) with glass caps were used for sampling. Sterile, carbon-free glassware was prepared as described previously (Greenberg et al., 1993; Hammes and Egli, 2005).

2.2.

Basic configuration of the pilot plant and sampling

The pilot plant investigated in this study treated surface water from Lake Zurich (Switzerland), taken at a depth of 30 m, with the following basic composition during the sampling period: temperature ¼ 6.1 1C; pH ¼ 7.9–8.0; dissolved organic carbon (DOC) ¼ 1.2–1.4 mg L1; alkalinity ¼ 2.6 mmol L1. The treatment train consisted of the following processes (Fig. 1): an ozonation reactor (4.4 m3, operated at 0.3 g m3 residual ozone), a granular active carbon (GAC) filtration reactor (effective volume: 1.5 m3; flow rate: 6 m3 h1) and a membrane reactor (0.02 mm, Zeeweed 1000, Zenon). The plant processes about 190 m3 per day, and has been in operation for 100 days at the time of this study. Samples (250 mL) were taken once per week during 3 weeks (January 2007) before and after each treatment step, and on four levels in the GAC reactor (145, 115, 75 and 35 cm) (Fig. 1). The uppermost sampling point (145 cm) was located in such a manner that it was relative to the top 5–10 cm of the GAC filter bed. The samples were processed within 2 h of sampling.

ARTICLE IN PRESS WAT E R R E S E A R C H

Raw water

Lake Zürich

20 µm disc-filter

Pre-filtration

Ozonation 1

Ozonation 2

GAC filtration 1

GAC filtration 2

Vreactor: 4.4 m3 Contact time: 40 min

145 cm 115 cm 75 cm

Vreactor: 2.7 m3 VGAC: 1.5 m3 Contact time: 17 min Flow rate: 6 m3 h-1

42 (2008) 269 – 277

271

cence channel (FL1). The collection of data as FL1/FL3 dot plots allowed for optimal distinction between the stained microbial cells and instrument noise or sample background (Hammes and Egli, 2005). Samples were measured in triplicate, and selected samples were controlled with epifluorescence microscopy to confirm the bacterial nature of stained particles. The detection limit for drinking water bacteria was determined experimentally by diluting the indigenous microbial community of bottled mineral water (EVIAN) with cellfree water (0.1 mm filtered EVIAN), followed by staining and measurement as described above. Each dilution was done in triplicate and a total of nine dilutions ranging from 0.1% to 100% were made. The standard instrument error on FCM measurements was always below 5%.

2.4.

Heterotrophic plate counts

35 cm

Membrane filtration

0.02 µm Zeeweed (Zenon)

Treated water

Fig. 1 – Schematic presentation of the pilot plant used for this study. Dotted lines (- - -) indicate parallel reactors not analysed in this study. Water samples were taken before and after every treatment step and, additionally, from four levels of the GAC reactor. The distance values next to the GAC reactor denote the distance of the sampling points from the bottom of the reactor.

2.3.

Fluorescence staining and flow cytometry

Bacterial cells were stained with 10 mL mL1 SYBRs Green I (1:100 dilution in DMSO; Molecular Probes), and incubated in the dark for at least 15 min before measurement. Where necessary, samples were diluted just before measurement in filtered (0.22 mm; Millexs-GP, Millipore) bottled mineral water (EVIAN, France), so that the concentration measured with the FCM was always less that 2  105 cells mL1. FCM was performed using a PASIII FCM (Partec, Hamburg, Germany) equipped with a 25 mW solid-state laser, emitting at a fixed wavelength of 488 nm, and volumetric counting hardware. SYBRs Green I has excitation/emission maxima at 497/ 520 nm, respectively, but due to the emission spectrum of the dye (see www.probes.invitrogen.com), emission is also detectable above 600 nm (Hammes and Egli, 2005). Green fluorescence was collected in the FL1 channel (520720 nm), and red fluorescence in the FL3 channel (4615 nm) and all data were processed with the Flowmax software (Partec), and electronic gating with the software was used to separate positive signals from noise. The specific instrumental gain settings for these measurements were as follows: FL1 ¼ 500, FL3 ¼ 700, speed ¼ 3 (implying an event rate never exceeding 1000 events s1). All samples were collected as logarithmic (3 decades) signals and were triggered on the green fluores-

The HPC method was performed according to the Swiss guidelines for drinking water (SLMB, 2000, 56, E.1). In short, 1 mL of the water sample was transferred to a sterile Petri dish and mixed with about 15 mL plate count agar (Oxoid). The agar was kept at 46 1C before plating. Samples were incubated at 30 1C for 72 h.

2.5.

Adenosine tri-phosphate (ATP)

Total ATP was determined as described in Berney et al. (2006) using the BacTiter-GloTM reagent (Promega Corporation, Madison, WI, USA) and a luminometer (Glomax, Turner Biosystems, Sunnyvale, CA, USA). In short, a water sample (100 mL) was warmed to 30 1C in a sterile Eppendorf tube, while the ATP reagent (100 mL) was simultaneously warmed. The sample and the reagent were combined after 2 min at 30 1C and then the luminescence was measured directly. The data were collected as relative light units (RLU) and converted to ATP (nM) by means of a calibration curve made with a known rATP standard (Promega). We furthermore distinguished between cellular ATP and free ATP by filtering each sample through a 0.1 mm sterile syringe filter (Millexs-GP, Millipore), and then repeating the analysis described above. ATP was measured in triplicate for all samples, and the detection limit of the measurement was about 0.01 nM ATP with a standard error of 8–10%.

3.

Results and discussion

3.1.

Autochthonous bacteria in drinking water

It is common for diverse microbial populations to grow during drinking water treatment and distribution (Berry et al., 2006). Although most of the bacterial biomass in the distribution system resides and proliferates in biofilms, high concentrations of planktonic bacteria are found in the water at the time of consumption (Hoefel et al., 2003; Rinta-Kanto et al., 2004). The aim of this study was to assess changes in the total planktonic bacterial concentrations in a large-scale drinking water pilot plant (Fig. 1) with respect to both disinfection and regrowth processes, using fluorescent staining and FCM. This was done in comparison with a conventional microbial

ARTICLE IN PRESS 272

WA T E R R E S E A R C H

parameter (HPC) and with ATP measurements, specifically with the view to evaluate total cell concentration as measured with FCM as a general descriptive process parameter.

3.2.

42 (2008) 269– 277

light (SSC), but we have shown that the unique emission spectrum of SYBRs Green I provides optimal discrimination of bacteria from background in drinking water (Hammes and Egli, 2005).

Flow cytometry on planktonic drinking water bacteria

The need for rapid detection methods for planktonic bacteria in drinking water has been highlighted numerous times in the past (Boulos et al., 1999; Lepeuple et al., 2004; Rinta-Kanto et al., 2004). FCM is a sensitive method that can be used to enumerate the bacteria in a water sample within 20 min (inclusive of the required staining time) and it provides a high degree of reproducibility (Hammes and Egli, 2005). Fig. 2 shows the detection limit of the instrument used in this study (ca. 200 cells mL1), as determined with a serial dilution of the autochthonous bacterial community in bottled mineral water (EVIAN). The average error arising from staining and (where required) dilution, was always o5%, irrespective of the concentration of the cells. We have also tested other previously described dyes, e.g., SYTOs9 (Lebaron et al., 1998; Berney et al., 2007), other staining protocols such as the addition of EDTA or lysis buffer (Hammes and Egli, 2005), as well as variations in the stain concentrations and staining times (data not shown). We found that for this study, the direct method with SYBRs Green I described herein produced the optimal results with respect to fluorescence intensity, ease of use and reproducibility. In this study, we have used dot plots of green fluorescence (520 nm) and red fluorescence (4615 nm) to distinguish between background noise and positively stained microbial cells. Previous studies often employed dot plots of green fluorescence against scattered

3.3. Total cell concentration as a descriptive parameter of drinking water treatment Changes in the total planktonic cell concentrations were particularly descriptive of the specific treatment processes that were investigated in this study (Fig. 3). The raw water (Lake Zurich) contained about 1  106 cells mL1, which was consistent over a prolonged period of time before the particular study period (data not shown) and which is typical for local surface water (data not shown). In the first process step (ozonation), the raw water bacterial cell concentration was reduced significantly (3 log reduction). We have shown previously that algal cells can pass through the ozonation process structurally intact, even though they are no longer alive (Hammes et al., 2007). In contrast, as shown in the FCM dot plots (Fig. 4a and b), the ozonation process applied here destroyed the bacterial cells completely, leaving only a few detectable bacterial cells (ca. 1  103 cells mL1) along with cellular debris. Importantly, the FCM method used gives no information as such on the viability of the cells that pass the ozonation process intact, and, as a result, these few detected cells may well be inactive or even dead. Since ozonation should serve from a hygienic perspective as an absolute disinfection process (von Gunten, 2003), it is necessary to complement total cell concentration measurements in the

Total cell concentration (cells mL-1 x 104)

10

8

6 6000 5000

4

4000 3000 2000

2

1000 0 0 0

20

40

1

60

2 80

3

4

5

100

% of bottled mineral water Fig. 2 – Precision and detection limit of flow cytometry for measuring total bacterial cell concentration, as determined with a real drinking water sample. Samples from bottled mineral water (EVIAN) were diluted with filtered (0.1 lm) water from the same bottle. Bacterial cells were stained with SYBRs Green I and enumerated with flow cytometry. R2 of the trend line is 0.99 (n ¼ 27) and the inlay amplifies the lowest bacterial concentrations. Error bars indicate standard deviation on triplicate samples.

ARTICLE IN PRESS WAT E R R E S E A R C H

Total cell concentration (cells mL-1)

10 7 week 1 week 2 week 3

10 6

10 5

10 4

10 3 FCM detection limit 10 2

GAC effluent

After UF filtration

Treatment steps

GAC 35 cm

GAC 75 cm

GAC 115 cm

GAC 145 cm

After ozonation

Raw water

10 1

Fig. 3 – Changes in the total planktonic bacterial cell concentrations during different stages of drinking water treatment over a 3-week period in a pilot plant. Cell concentrations were enumerated with SYBRs Green I staining and flow cytometry. Error bars indicate the standard deviation on triplicate measurements.

future with viability assessments (Hoefel et al., 2003; Phe et al., 2005; Berney et al., 2006, 2007). In the second process step (GAC filtration), regrowth of bacterial cells in the GAC filtration reactor was clearly observed from an increase in planktonic cell concentrations over different levels of the filter (Figs. 3 and 4c and d). Note that the sample in Fig. 4c was processed undiluted, and the sample in Fig. 4d was processed after 10-fold dilution, hence the visual impression of more cells in Fig. 4c. The biological component of GAC reactors has been described previously (Magic-Knezev and van der Kooij, 2004; Hammes et al., 2006; Velten et al., 2007), although the extent to which microbial regrowth in these reactors contributes to the biological load in drinking water has, to our knowledge, not been quantified in detail before. The bacteria grow on an array of low molecular weight organic carbon compounds that are present in the raw water and/or are produced during the ozonation reaction; they are collectively described as assimilable organic carbon (Hammes et al., 2006, 2007; van der Kooij, 2002). In this respect, the bacteria replace the originally intended function of the GAC (adsorption) and continue to remove organics years after the sorption capacity of the carbon has been depleted (Jekel, 1982), thereby fulfilling a vital positive role in the treatment process. It is important to note that the FCM measurements were done on the liquid phase (planktonic cells) only, and an eventual comparison with the biomass attached to the GAC (Magic-Knezev and van der Kooij, 2004; Velten et al., 2007) would be required for an overall mass balance of bacterial growth in the reactor. Still, the regrowth was clearly quantified, with most of the growth (52%) occurring at the top of the reactor between 10 and 45 cm (Figs. 1 and 3). The regrowth is also clearly illustrated in the

42 (2008) 269 – 277

273

FCM dot plots (Fig. 4c and d). At the very top of the reactor, only a single bacterial cluster is evident (Fig. 4c), while at all the levels lower down, two defined clusters can be observed (Fig. 4d). These two clusters allude to the so-called low and high nucleic acid bacteria (Gasol et al., 1999; Lebaron et al., 2001), which represent two classes of bacteria typically observed in seawater and freshwater with FCM. These two classes are distinguished in size and fluorescence intensity, even though the composition and function of the two groups remain to be elucidated (Lebaron et al., 2001). As a final treatment step, membrane filtration has been introduced as an absolute barrier to particles including microorganisms at the end of the treatment train just before distribution, even though questions arise on the degree to which this treatment may actually stimulate unchecked regrowth during distribution (Hagen, 1998). The total cell concentration data suggest that this was indeed the case, as cell concentrations were reduced to or below the detection limit of the instrument (o200 cells mL1). Fig. 4e shows that the dot plots were indeed devoid of any cell signals. In this case, no viability assessment would be required, since the exclusion of SYBRs Green I-stained particles is sufficient proof of the successful operation of the membranes.

3.4. Comparison between heterotrophic plate counts (HPCs) and total cell concentrations HPCs, irrespective of the specific method used, are inadequate to enumerate the actual concentration of (heterotrophic) mesophilic bacteria in water (Allen et al., 2004). This is because relative to the carbon concentration in drinking water, all HPC media contain excessive nutrients. For comparison, R2A medium, which is often preferred in drinking water studies because it is considered to represent a low carbon concentration medium (Reasoner and Geldreich, 1985; Allen et al., 2004), has a DOC concentration of about 800 mg L1, which is 800 times in excess of that of the water studied here. With this difference, one can reasonably expect that some cells are just not cultivable under the restrictive environment imposed by synthetic media and selective incubation conditions (temperature, oxygen, etc.). Fig. 5 shows that between 0.001% and 6.5% of the cells were cultivable with the used method, with most of the culturability observed in the GAC reactor (0.8–6.5%), and least in the raw water (0.001%). With the exception of the raw water, the pattern observed with the HPC method was similar to that observed with the total cell counts. The high culturability of bacterial cells in the GAC reactor probably stems from the fact that the GAC reactor has not been back-flushed since the operation commenced. The HPC detected after ultrafiltration (UF) is ascribed to biofilm formation on the materials used after the filter, and not to cells escaping the filtrations process. From a practical perspective, HPC results had a standard error of 430% compared with FCM results with a standard error ofo5%. From a statistical perspective, the HPC method detects between 30 and 300 events in order to obtain a result, while FCM analysis detects between 40 and 20,000 events to obtain a result without any dilution required (Hammes and Egli, 2005). HPC takes 3–7 days to obtain results, while FCM

ARTICLE IN PRESS 274

WA T E R R E S E A R C H

42 (2008) 269– 277

Fig. 4 – Flow cytometry dot plots of water samples from the main stations of the pilot plant. All samples were stained with SYBRs Green I and analysed with flow cytometry. (a) Raw water from Lake Zu¨rich, (b) after ozonation, (c) top (145 cm) of the GAC filter, (d) after GAC filtration and (e) after membrane ultrafiltration. FL1 denotes green fluorescence signals (520 nm) and FL3 denotes red fluorescence signals (615 nm). Electronic gates (- - -) were used to separate bacterial cells from background.

requires 20 min. In short, this allows for a significantly faster reaction time if, for example, the UF membrane would incur physical damage or the ozonation reactor would malfunction. The speed of the measurements also allows for significantly

more samples (and replicates) to be processed than what is usually possible with HPC. However, one should also recognise that total cell concentration is an insufficient parameter to describe the efficiency of some processes such as UV

ARTICLE IN PRESS WAT E R R E S E A R C H

275

42 (2008) 269 – 277

106

106

103

102

102

101

101

100

100 GAC effluent

GAC 75 cm

After UF filtration

103

GAC 35 cm

104

GAC 115 cm

104

GAC 145 cm

105

After ozonation

105

Total cell concentration (cells mL-1)

107 heterotrophic plate counts total cell concentration

Raw water

Heterotrophic plate counts (cfu mL-1)

107

Treatment steps

Fig. 5 – Changes in heterotrophic plate counts (R2A medium, 30 1C, 3 days) during different stages of drinking water treatment (average of a 3-week period) in the pilot plant, in comparison with the average total cell concentration determined with FCM (from Fig. 3) recorded during the same time period. Error bars indicate standard deviation of average values over the 3-week period. 0.4

0.2

0.1

GAC effluent

After UF filtration

Treatment steps

GAC 35 cm

GAC 75 cm

GAC 115 cm

GAC 145 cm

0.0 After ozonation

ATP is another rapid method that has previously been promoted for its applicability in drinking water (Delahaye et al., 2003; Deininger and Lee, 2006). ATP is used as primary energy currency for all bacteria, and therefore it is a parameter suited for the quantification of the active biomass (Velten et al., 2007). An ATP measurement requires 5 min per sample, and has additionally the advantages that the equipment is simple to use and relatively affordable. Both Delahaye and co-workers (Delahaye et al., 2003) and Deininger and Lee (2006) suggested some correlation between ATP and HPC (specifically R2A media), but conclusive data were not provided and, to our knowledge, are still not available. Also, significant differences can exist between the cellular ATP content of different bacteria cultivated under different conditions or in different physiological states (LeChevallier et al., 1993; Schneider and Gourse, 2004). Fig. 6 shows the changes in total ATP concentrations through the treatment plant as a composite of cell-bound ATP and free ATP. Evidently, the pattern portrayed by the total cell concentration measurements (Fig. 3) was not repeated in total ATP

free ATP 0.3

Raw water

3.5. Comparison with adenosine tri-phosphate (ATP) as alternative parameter

cell-bound ATP A TP concentrations (nM)

disinfection, which does not physically destroy the cells (Berney et al., 2006). For this purpose, more emphasis should be placed on viability assessment using specific stains (Hoefel et al., 2003, Phe et al., 2005; Berney et al., 2006; Berney et al., 2007) and/or ATP (Delahaye et al., 2003) in combination with the total cell counts.

Fig. 6 – Average changes in the cell-bound and free ATP concentrations during different stages of drinking water treatment over a 3-week period in a pilot plant. ATP was determined with the luciferin–luciferase assay. For determination of free ATP, samples were pre-filtered with 0.1 lm membrane filters. The standard deviation in average ATP concentrations during the 3 weeks was between 2% and 36% (n ¼ 3).

measurements (Fig. 6), specifically with regard to the samples after ozonation and at the top (145 cm) of the GAC filtration reactor. Since both the total cell concentration with FCM

ARTICLE IN PRESS 276

WA T E R R E S E A R C H

Table 1 – Average values for the percentage (%) of extracellular ATP relative to the total ATP in solution and the ATP-per-cell concentrations (nmol cell1) of planktonic bacteria in the different stages (see Fig. 1) of the drinking water pilot plant (n ¼ 3) Extracellular ATP (%)

Raw water After ozone GAC 145 cm GAC 115 cm GAC 75 cm GAC 35 cm GAC effluent After UF

19 106 98 46 26 19 12 103

(74.9) (79.2) (73.3) (78.3) (73.7) (75.3) (73.2) (77.0)

ATP-per-cell concentrations (nmol cell1) 1.7  1010 n.a. 1.5  1010 4.8  1010 4.1  1010 4.3  1010 3.4  1010 n.a.

(70.26  1010) (72.1  1010) (70.84  1010) (70.18  1010) (70.23  1010) (70.55  1010)

n.a. ¼ not applicable.

42 (2008) 269– 277





measurements were particularly affected by extracellular ATP, which was released from microbial cells during ozonation. As such, the results suggest that total cell concentration would be a useful parameter for monitoring of general water quality during treatment and distribution, as well as for the design and optimisation of specific treatment steps. Though these are positive steps, further work on rapid detection of cellular viability and activity would increase confidence and acceptance of this approach and will form part of our future work.

Acknowledgements The authors acknowledge financial assistance from the Eawag Wave21 project and the 6th EU framework project TECHNEAU (018320), as well as technical and scientific contributions of I. Hu¨lshoff, J. Traber and S. Meylan. R E F E R E N C E S

(Figs. 3 and 4b) and the HPCs (Fig. 5) suggested that very few cells survive the ozonation process intact, the high total ATP concentration measured in this sample was unexpected. The measurement of free ATP concentrations (after 0.1 mm filtration of the sample) revealed that this must be ascribed to extracellular ATP, most probably released from algae and bacteria during the ozonation process. In the raw water, the free ATP constituted about 19% of the total ATP, in the ozonated water 100%, in the water just before UF filtration 12% and in the water after UF filtration again 100% (Table 1). Hence, ATP measurements should be used with extreme care in treatment processes such as ozonation and membrane filtration. The results show that cell-bound ATP is descriptive of the pattern shown by both the total cell concentration measurements and the HPC measurements. However, the requirement of constantly measuring free ATP adds both labour and consumables costs, while also increasing the potential error on the measurements caused by sample contamination. Additionally, more work is required for relating ATP concentrations to approximate cell concentrations. When only the cell-bound ATP was considered in our results together with the flow cytometric total cell concentrations, the calculated ATP per cell was between 1.5 and 4.8  1010 nmol cell1 (Table 1). This corresponds with previous samples from our group (data not shown) and also with results presented by Velten et al. (2007) and Magic-Knezev and van der Kooij (2004) for environmental aquatic bacteria.

4.

Conclusions

 The results show that total cell concentration is descrip-



tive of the main treatment processes that were investigated in this particular study. As a general parameter, total cell concentrations measured with FCM displayed marked advantages over two other microbiological parameters (HPC and ATP). With regard to HPC, this method is significantly faster and detects all cells irrespective of their culturability. In this study, ATP

Allen, M.J., Edberg, S.C., Reasoner, D.J., 2004. Heterotrophic plate count bacteria—what is their significance in drinking water? Int. J. Food Microbiol. 92, 265–274. Bartram, J., Cotruvo, J., Exner, M., Fricker, C., Glasmacher, A., 2003. Heterotrophic Plate Counts and Drinking-Water Safety. IWA Publishing on behalf of the World Health Organization, London, UK. Berney, M., Weilenmann, H.U., Egli, T., 2006. Flow-cytometric study of vital cellular functions in Escherichia coli during solar disinfection (SODIS). Microbiology 152, 1719–1729. Berney, M., Hammes, F., Bosshard, F., Weilenmann, H.-U., Egli, T., 2007. Assessment and interpretation of bacterial viability using LIVE/DEADs BacLightTM kit in combination with flow cytometry. Appl. Environ. Microbiol. 73, 3283–3290. Berry, D., Xi, C., Raskin, L., 2006. Microbial ecology of drinking water distribution systems. Curr. Opin. Biotechnol. 17, 297–302. Boulos, L., Prevost, M., Barbeau, B., Coallier, J., Desjardins, R., 1999. LIVE/DEAD BacLight: application of a new rapid staining method for direct enumeration of viable and total bacteria in drinking water. J. Microbiol. Methods 37, 77–86. Clarke, R.G., Pinder, A.C., 1998. Improved detection of bacteria by flow cytometry using a combination of antibody and viability markers. J. Appl. Microbiol. 84 (4), 577–584. Deininger, R.A., Lee, J., 2006. Rapid detection of bacteria in drinking water. In: Omelchenko, A., Pivovarov, A.A., Swindall, W.J. (Eds.), Modern Tools and Methods of Water Treatment for Improving Living Standards. Proceedings of the NATO Advanced Research Workshop on Modern Tools and Methods of Water Treatment for Improving Living Standards Dnepropetrovsk, Ukraine, 19–22 November 2003. Springer, Netherlands, pp. 71–78. Delahaye, E., Welte, B., Levi, Y., Leblon, G., Montiel, A., 2003. An ATP-based method for monitoring the microbiological drinking water quality in a distribution network. Water Res. 37, 3689–3696. Gasol, J.M., Zweifel, U.L., Peters, F., Fuhrman, J.A., Hagstrom, F., 1999. Significance of size and nucleic acid content heterogeneity as measured by flow cytometry in natural planktonic bacteria. Appl. Environ. Microbiol. 65, 4475–4483.

ARTICLE IN PRESS WAT E R R E S E A R C H

Greenberg, A.E., Clesceri, L.S., Eaton, A.D. (Eds.), 1993. Standard Methods for the Examination of Water and Wastewater, 18th ed. American Public Health Association, Washington, DC. Hammes, F., Salhi, E., Koster, O., Kaiser, H.P., Egli, T., von Gunten, U., 2006. Mechanistic and kinetic evaluation of organic disinfection by-product and assimilable organic carbon (AOC) formation during the ozonation of drinking water. Water Res. 40, 2275–2286. Hammes, F.A., Egli, T., 2005. New method for assimilable organic carbon determination using flow-cytometric enumeration and a natural microbial consortium as inoculum. Environ. Sci. Technol. 39, 3289–3294. Hammes, F.A., Meylan, S., Salhi, E., Ko¨ster, O., Egli, T., von Gunten, U., 2007. Formation of assimilable organic carbon (AOC) and specific natural organic matter (NOM) fractions during ozonation of phytoplankton. Water Res. 41, 1447–1454. Hagen, K., 1998. Removal of particles, bacteria and parasites with ultrafiltration for drinking water treatment. Desalination 119, 85–91. Hoefel, D., Grooby, W.L., Monis, P.T., Andrews, S., Saint, C.P., 2003. Enumeration of water-borne bacteria using viability assays and flow cytometry: a comparison to culture-based techniques. J. Microbiol. Methods 55, 585–597. Jekel, M.R., 1982. The DOHNE treatment plant in Mu¨lheim (Ruhr): the Mu¨lheim process. In: Masschelein, W.J. (Ed.), Ozonation Manual for Water and Wastewater treatment. Wiley, New York, pp. 306–308. Lebaron, P., Parthuisot, N., Catala, P., 1998. Comparison of blue nucleic acid dyes for flow cytometric enumeration of bacteria in aquatic systems. Appl. Environ. Microbiol. 64, 1725–1730. Lebaron, P., Servais, P., Agogue´, H., Courties, C., Joux, F., 2001. Does the high nucleic acid content of individual cells allow us to discriminate between active cells and inactive cells in aquatic systems? Appl. Environ. Microbiol. 67, 1775–1782. LeChevallier, M.W., Shaw, N.E., Kaplan, L.A., Bott, T.L., 1993. Development of a rapid assimilable organic carbon method for water. Appl. Environ. Microbiol. 59 (5), 1526–1531. Leclerc, H., Moreau, M., 2002. Microbiological safety of natural mineral water. FEMS Microbiol. Rev. 26 (2), 207–222.

42 (2008) 269 – 277

277

Lee, S.H., O’Connor, T.L., Banerji, S.K., 1980. Biologically mediated corrosion and its effects on water quality in distribution systems. J. Am. Water Works Assoc. 72 (11), 636–645. Lepeuple, A.-S., Gilouppe, S., Pierlot, E., de Roubin, M.-R., 2004. Rapid and automated detection of fluorescent total bacteria in water. Int. J. Food Microbiol. 92, 327–332. Magic-Knezev, A., van der Kooij, D., 2004. Optimisation and significance of ATP analysis for measuring active biomass in granular activated carbon filters used in water treatment. Water Res. 38, 3971–3979. Phe, M.H., Dossot, M., Guilloteau, H., Block, J.C., 2005. Nucleic acid fluorochromes and flow cytometry prove useful in assessing the effect of chlorination on drinking water bacteria. Water Res. 39, 3618–3628. Reasoner, D.J., Geldreich, E.E., 1985. A new medium for the enumeration and subculture of bacteria from potable water. Appl. Environ. Microbiol. 49, 1–7. Rinta-Kanto, J.M., Lehtola, M.J., Vartiainen, T., Martikainen, P.J., 2004. Rapid enumeration of virus-like particles in drinking water samples using SYBR green I-staining. Water Res. 38, 2614–2618. Sartory, D.P., 2004. Heterotrophic plate count monitoring of treated drinking water in the UK: a useful operational tool. Int. J. Food Microbiol. 92, 297–306. Staley, J.T., Konopka, A., 1985. Measurement of in situ activities of nonphotosynthetic microorganisms in aquatic and terrestrial habitats. Annu. Rev. Microbiol. 39, 321–346. Schneider, D.A., Gourse, R.L., 2004. Relationship between growth rate and ATP concentration in Escherichia coli: a bioassay for available cellular ATP. J. Biol. Chem. 279, 8262–8268. van der Kooij, D., 2002. Assimilable organic carbon (AOC) in treated water: determination and significance. In: Bitton, G. (Ed.), Encyclopedia of Environmental Microbiology. Wiley, Hoboken, NJ, USA, pp. 312–327. Velten, S., Hammes, F., Boller, M., Egli, T., 2007. Rapid and direct estimation of active biomass on granular-activated carbon through adenosine tri-phosphate (ATP) determination. Water Res. 41, 1973–1983. Von Gunten, U., 2003. Ozonation of drinking water. Part I. Oxidation kinetics and product formation. Water Res. 37, 1443–1467.