ARTICLE IN PRESS

Water Research 38 (2004) 26142618

Rapid enumeration of virus-like particles in drinking water samples using SYBR green I-staining
Johanna M. Rinta-Kantoa,1, Markku J. Lehtolaa,*, Terttu Vartiainena,b, Pertti J. Martikainenb
b

Department of Environmental Health, National Public Health Institute, P.O. Box 95, FIN-70701 Kuopio, Finland Department of Environmental Sciences, University of Kuopio, Bioteknia 2, P.O. Box 1627, FIN-70211 Kuopio, Finland Received 22 May 2003; received in revised form 23 February 2004; accepted 5 March 2004

Abstract We studied the suitability of SYBR green I-staining for determining total counts of virus-like particles and bacteria in drinking water. Low background uorescence and lack of unspecic staining made drinking water samples an excellent matrix for SYBR green I-staining. Direct microscopic count method is a rapid and economical tool for assessing the total number of virus-like particles in aquatic samples, compared to culture-dependent or molecular biology methods. We applied this method to show the efciency of a large-scale drinking water purication process in the removal of virus-like particles and bacteria from lake water. r 2004 Elsevier Ltd. All rights reserved.
Keywords: Drinking water; Viruses; Bacteriophage; SYBR green I; Bacteria

1. Introduction Despite the extremely oligotrophic character of the habitat, drinking waters harbour a diverse and active microbial community, consisting of bacteria, fungi, yeasts and protozoa [1]. Natural waters and drinking water distribution systems are physically linked to each other when surface water from lakes or rivers is used as the raw water for drinking water purication. Despite the multi-step process, a small fraction of microbes pass through the drinking water purication and become established in the microbial community of a distribution network causing hygienic, aesthetic and technical problems. To solve and understand these problems, drinking water purication processes, drinking water

*Corresponding author. Tel.: +358-17-201371; fax: +35817-201155. E-mail address: markku.lehtola@ktl. (M.J. Lehtola). 1 Present address: Department of Microbiology, University of Tennessee, Knoxville, TN 37996, USA.

chemistry and microbiology and their effects on microbial growth have been intensively studied. The occurrence of viruses in drinking water is commonly considered to be a result of a failure in the water treatment process or in the distribution network, causing a potential risk to public health. The use of culture-based detection methods, and a focus only on viruses pathogenic to higher forms of life, have led to the presumption that viruses are rarely present in drinking water. Studies of viruses in drinking waters have mainly concentrated on screening the water for potential pathogens and studying the survival potential of pathogenic viruses in the drinking water distribution system. In several studies, selected bacteriophages have been used as indicators of fecal contamination of drinking water, or as model viruses for enteric viruses because of their similar properties in transport and survival [2,3]. In lake and sea water environments, bacteriophages are estimated to be the most numerous component of plankton, with abundance ranging from 104 to 108 viruses ml1 [4]. Enumeration of virus particles has

0043-1354/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2004.03.008

ARTICLE IN PRESS
J.M. Rinta-Kanto et al. / Water Research 38 (2004) 26142618 2615

become an essential part of studies in the eld of aquatic microbial ecology during the 1990s. The abundance of viruses in natural water samples was rst assessed solely with transmission electron microscopy (TEM) [5,6]. Soon it was realized that it was possible to visualize virus particles with techniques combining bright nucleic acid stains and epiuorescence microscopy [710]. Enumeration of virus-like particles (VLPs) by SYBR green I-staining and epiuorescence microscopy enables rapid analysis of bacterial and VLP numbers in water samples [11,12]. Studies have shown that counts of SYBR green I stained bacteria and VLPs strongly correlate with results obtained by TEM or other uorescent staining techniques [9,11]. Thus SYBR green I-method has become a widely accepted method among aquatic microbial ecologists for obtaining accurate total counts of bacteria and VLPs. However, SYBR green Istaining is not able to give any information about viability or identity of viruses, which requires culturing methods or use of PCR-based methods. The aim of this study was, rstly, to apply SYBR green I-staining method to drinking water samples and evaluate its suitability in this application. Secondly, we assessed the abundance of virus-like particles and bacteria in the different steps of a water purication process, using SYBR green I-staining and direct counting by epiuorescence microscopy. We monitored the total count of bacteria and VLPs at various stages of the water treatment process in a surface waterworks in central Finland, which utilizes bank ltration followed by a conventional water purication process.

sampling and processed immediately after arrival to the laboratory. Water samples were preserved by adding 37% formaldehyde to samples to a nal concentration of 2%. Formaldehyde was ltered through 0.22-mmpore-size lter before use. VLPs and bacteria were stained with SYBR green I (Sigma, St. Louis, USA) nucleic acid stain as described earlier [12]. Briey, 0.5 ml of lake water or 1.5 ml water from the other water samples was ltered on 0.02-mm-pore-size aluminium oxide lter (Anodisc 25, Whatman Ltd., Maidstone, England). VLPs and bacteria collected on the lters were stained with SYBR green I (Sigma) nucleic acid stain and enumerated using 1000 magnication under blue excitation using an Olympus BH-2 epiuorescence microscope equipped with an ocular grid. Twenty randomly selected elds of view or a total of 200 viruses and 200 bacteria were counted from each sample. Bacteria and VLPs were distinguished based on their dimensions and/or their relative brightness. The formaldehyde used to preserve samples was checked before use by making a sample of sterile water containing 2% nal concentration of ltered formaldehyde; the control sample was ltered and stained with SYBR green I stain using the same protocol as for the other water samples. The control sample contained no uorescently stained particles. Pearson correlation coefcients were calculated with SPSS for Windows version 10.1.3 (SPSS Inc.) program. Regression line was calculated with Microsoft Excel 97 program.

3. Results 2. Materials and methods All pieces of glassware used in collecting water samples were washed with phosphate-free detergent (Deconex, Borer Chemie AG, Zuchwil, Switzerland), then immersed in HCl for 2 h, subsequently rinsed with deionized water (Millipore, Molsheim, France) and then heated for 6 h at 550 C. Water samples were collected from the lake water and the subsequent purication steps at Kuopio Waterworks located in central Finland. The waterworks utilizes bank ltrated lake water as raw water for a conventional water purication process involving chemical coagulation, rapid sand ltration, and disinfection. Samples from each sampling point were taken at ve different times of the year; in January 2002, April 2002, May 2002, June 2002 and August 2003. In April 2002 the bank ltration step was temporarily out of use due to annual maintenance. Water samples were collected from lake water, after bank ltration, after chemical coagulation and rapid sand ltration, from disinfected drinking water and from tap water at 12 km distance from the waterworks. Water samples were kept cool during VLPs were found from all samples (Fig. 1). The changes in abundance of VLPs and bacteria in water samples are presented in Figs. 2a and b. The numbers of VLPs and bacteria in the lake water varied by season, both being lowest in mid-winter in January and highest in May and August. Using lake water temperature data, the spring turnover of the lake water was estimated to have taken place around the rst week of May (data not shown). In August 2003 lake waters were very warm (24 C) which increased microbial concentrations in the lake. Bank ltration of the lake water reduced the numbers of VLPs by 5881% and bacteria by 4488%. Chemical coagulation of the bank ltrated water further decreased the number of remaining VLPs by 4469% and of bacteria by 5486%. Bank ltration and chemical coagulation together removed a total of 8592% of VLPs and 8895% of bacteria from the numbers present in the lake water. In April 2002, during a temporary shutdown of bank ltration, the lake water was directed straight to chemical coagulation. Sixty eight percent of VLPs and 67% of bacteria were removed from the lake

ARTICLE IN PRESS
2616 J.M. Rinta-Kanto et al. / Water Research 38 (2004) 26142618
700 600

VLPs/ml x 105

500 400 300 200 100 0 0 20 40 60 80 100

y = 6.4906x + 8.8326
R2 = 0.9707

bacteria/ml x
Other data points

105

April 2002 data points

Linear (Other data points)

Fig. 1. Epiuorescense photomicrograph of viruses (small green dots) and bacteria (larger dots and rods) ltered onto 0.02-mm-pore-size membrane. Sample was obtained from tap water in August 2003. Photographed with Olympus BX51 microscope and Soft Imaging System CC-12 digital CCD camera.

Fig. 3. Correlation between bacterial and VLP abundance in water samples taken from different steps of water purication process (lake water, after bank ltration, after chemical coagulation, after disinfection and tap water at 12 km distance from the waterworks). Numbers of bacteria and VLPs from January, May, June and August have been included when calculating the equation for the regression line. Data points from April samples were not included because the water treatment process was temporarily altered at the time of sampling.

Fig. 2. Abundance of (a) VLPs and (b) bacteria at various stages of the drinking water purication process and the distribution pipe line according to season. Lake water (LW), bank ltrated lake water (BF), chemically coagulated water (CC), drinking water (DW), and tap water (TW) at 12 km distance from the waterworks. The arrow symbol (k) indicates the point in the water purication process where bank ltration was temporarily out of use in April 2002.

water, which indicated a steady removal efciency of particles during this step. When the complete water purication at the waterworks was in use (January, May, June and August), the abundance of bacteria and

VLPs found in the tap water samples remained at an almost constant level at each time of the year and also in respect to levels of bacteria and VLPs found in drinking water leaving from the waterworks. In August 2003 drinking waters were warm and concentration of bacteria increased in distribution networks. The abundance of VLPs exceeded bacterial abundance in all samples. Positive correlation was found between numbers of bacteria and VLPs among all samples analyzed in this study. There was a positive and strong correlation (r 0:99; p 0:000) between the numbers of bacteria and VLPs in the water samples collected in January, May, June and August (Fig. 3). In these samples the virus-to-bacterium ratio (VBR) throughout the process was 6.5 viruses per bacterium, as determined from the slope of the linear regression line (Fig. 3). In the lake water sample collected in April 2002, the VBR was approximately 4 times higher than in the lake water samples collected at other times of the year (Table 1). At the same time, bank ltration was temporarily omitted due to annual maintenance and cleaning at the waterworks. Therefore the lake water was used as raw water for the drinking water purication process. A high count of VLPs was detected after chemical coagulation and rapid sand ltration as well as after disinfection, which was also reected by the high VBR (Table 1). In tap water sample of April 2002, counts of bacteria and VLPs were greater than was found at other times of the year (Fig. 2).

ARTICLE IN PRESS
J.M. Rinta-Kanto et al. / Water Research 38 (2004) 26142618 2617 Table 1 Ratio of viruses to bacteria (VLPs per bacterium) in the water samples taken from different steps of drinking water treatment process and from tap water at 12 km distance from the waterworks Date 21.1.2002 11.4.2002 13.5.2002 14.6.2002 12.8.2003 Lake water 6.69 29.30 7.66 6.08 6.11 Bank ltrated lake water 3.69 8.15 6.26 9.69 Chemically coagulated water 5.30 28.41 14.05 15.75 11.81 Drinking water 4.26 37.12 6.73 11.25 18.71 Tap water 4.28 8.93 6.94 4.32 5.36

4. Discussion SYBR green I-staining combined with direct counting provides a rapid and inexpensive method for studying total number of bacteria and viruses in drinking water. Drinking water is an almost ideal sample matrix for SYBR green staining; when compared to natural water samples, background uorescence caused by dissolved material or other unspecic staining is very low. Evenly distributed, uniformly stained particles and lack of interfering background uorescence allowed for easy distinction between bacteria and VLPs and obtaining of precise counting results. Although fading of the SYBR green I stain has been recognized to be a problem when counting samples with a high particle density [11], the particle density in drinking water is usually moderate or low and fading of the stain did not interfere with counting. However, one disadvantage in direct counting of VLPs is that it does not indicate the viability of the viruses; after chlorination, for instance, part of the viruses may be inactivated. Wommack et al. [13] have presented data suggesting that direct counts are not capable of indicating whether VLPs are infective and thus in some cases direct counts of VLPs may overestimate the number of infective bacteriophages. In water, turbidity can be used as an indicator of microbial contamination [14]. In terms of viral contamination, however, good correlation between turbidity and bacteriophages has not been found [14,15]. Some correlation can be seen as a result of particle association of coliphages [15]. However, SYBR green I-staining method is more accurate for detecting viruses and has a much lower detection limit than could ever be achieved through the analysis of turbidity. The VBR was >1 in all samples enumerated in this study. Data from several previous studies, reviewed in Ref. [4], indicates that in natural waters the VBR is generally ranging from 3 to 10 viruses per bacterium. In all untreated lake water samples bacterial and viral abundance showed seasonal variation, and the VBR in samples taken in January, May, June and August fell into the range presented earlier in the literature. According to our results, the bank ltration and chemical coagulation efciently retained bacteria and

VLPs from the raw water. In chlorination there were little change in numbers of VLPs; however, part of the VLPs may have been inactivated but not destructed. Higher numbers of VLPs and bacteria found in drinking water and tap water samples reected the temporary shut down of bank ltration. Correlation between viral and bacterial numbers indicated a steady removal of particles in the drinking water treatment process. Total numbers of bacteria and VLPs from this study represent abundance of VLPs and bacteria in planktonic samples at certain time points. In natural waters the virus population will decline unless new virus particles are produced through bacterial lysis [16,17]. As our data, especially the April samples, suggest, virus particles in the drinking water distribution system can originate directly from the raw water as free particles. Previous studies have indicated that certain bacteriophages are able to survive at least 23 days in drinking water and survive better than fecal bacteria [18,19]. The residence time of water in the distribution system studied here has been previously estimated to be from 3 to 3.5 days (data not shown), which probably is short enough for survival of bacteriophages. Further investigation of the role of bacteriophages may reveal a new aspect of drinking water microbial diversity and the regulation of microbial growth in drinking water microbial biolms. Previous laboratorybased studies have revealed that bacteriophages are able to use biolm bacteria as their hosts [20,21]; that biolms formed by E. coli are susceptible to phage attack; and that biolms can act as reservoirs of virus particles [22,23].

5. Conclusions SYBR green I-staining is a suitable method for analysing the effects of water coagulation and ltration on VLP abundances, but probably not for disinfection efciency. A rapid and accurate method for evaluating the total viral load in public water supplies will also have applications for public health. Monitoring the removal of very small size, virus-like particles could serve as an indicator for the efciency of the drinking water

ARTICLE IN PRESS
2618 J.M. Rinta-Kanto et al. / Water Research 38 (2004) 26142618 [10] Proctor LM, Fuhrman JA. Mortality of marine bacteria in response to enrichments of the virus size fraction from sea water. Marine EcolProg Ser 1992;87:28393. [11] Bettarel Y, Sime-Ngando T, Amblard C, Laveran H. A comparison of methods for counting viruses in aquatic systems. Appl Environ Microbiol 2000;66:22839. [12] Noble RT. Enumeration of viruses. Method Microbiol 2001;30:4351. [13] Wommack KE, Hill RT, Muller TA, Colwell RR. Effects of sunlight on bacteriophage viability and structure. Appl Environ Microbiol 1996;62:133641. [14] Ferguson CM, Coote BG, Ashbolt NJ, Stevenson IM. Relationships between indicators, pathogens and water quality in an estuarine system. Water Res 1996;30(9): 204554. [15] Lipp EK, Farrah SA, Rose JB. Assessment and impact of microbial fecal pollution and human enteric pathogens in a coastal community. Marine Pollut Bull 2001;42(4):28693. [16] Bratbak G, Thingstad F, Heldal M. Viruses and the microbial loop. Microbial Ecol 1994;28:20921. [17] Weinbauer MG, Suttle CA. Potential signicance of lysogeny to bacteriophage production and bacterial mortality in coastal waters of the Gulf of Mexico. Appl Environ Microbiol 1996;62:437480. [18] Quignon F, Kiene L, Levi Y, Sardin M, Schwartzbrod L. Virus behaviour within a distribution system. Water Sci & Technol 1997;35:3118. [19] Momba MNB, Kaleni P. Regrowth and survival of indicator microorganisms on the surfaces of household containers used for the storage of drinking water in rural communities of South Africa. Water Res 2002;36: 30238. [20] Doolittle MM, Cooney JJ, Caldwell DE. Tracing the interaction of bacteriophage with bacterial biolms using uorescent and chromogenic probes. J Ind Microbiol 1996;16:33141. [21] Hughes KA, Sutherland IW, Jones MV. Biolm susceptibility to bacteriophage attack: the role of phage-borne polysaccharide depolymerase. Microbiol (Reading) 1998;45:303947. [22] Corbin BD, McLean RJC, Aron GM. Bacteriophage T4 multiplication in a glucose limited Escherichia coli biolm. Can J Microbiol 2001;47:6804. [23] Storey MV, Ashbolt NJ. Persistence of two model enteric viruses (B40-8 and MS2-bacteriophages) in water distribution pipe biolms. Water Sci Technol 2001;43:1338.

purication process in the removal of viruses, including those potentially pathogenic to humans.

Acknowledgements We thank Kuopio Waterworks for cooperation during this study. We give special thanks to the staff of the Laboratory of Environmental Microbiology at the National Public Health Institute. We also thank Tiia Myllykangas for providing the lake water temperature data.

References
[1] Szewzyk U, Szewzyk R, Manz W, Schleifer K. Microbiological safety of drinking water. Annu Rev Microbiol 2000;54:81127. [2] Havelaar AH. Bacteriophages as model organisms in water treatment. Microbiol Sci 1987;12:3624. [3] Jofre J, Oll! E, Ribas F, Vidal A, Lucena F. Potential e usefulness of bacteriophages that infect Bacteroides fragilis as model organisms for monitoring virus removal in drinking water treatment plants. Appl Environ Microbiol 1995;61:322731. [4] Wommack KE, Colwell RR. Virioplankton: viruses in aquatic ecosystems. Microbiol Mol Biol Rev 2000;64: 69114. [5] Bergh , Brsheim KY, Bratbak G, Heldal M. High abundance of viruses found in aquatic environments. Nature 1989;340:4678. [6] Proctor LM, Fuhrman JA. Viral mortality of marine bacteria and cyanobacteria. Nature 1990;343:602. [7] Hara S, Terauchi K, Koike I. Abundance of viruses in marine waters: assesment by epiuorescence and transmission electron microscopy. Appl Environ Microbiol 1991; 57:27314. [8] Hennes KP, Suttle CA. Direct counts of viruses in natural waters and laboratory cultures by epiuorescence microscopy. Limnol Oceanogr 1995;40:10505. [9] Noble RT, Fuhrman JA. Use of SYBR green I for rapid epiuorescence counts of marine viruses and bacteria. Aquat Microb Ecol 1998;14:1138.