A 5-year survey (2007–2011) of enteric viruses in Korean aquatic environments and the use of coliforms as viral indicators


  • These two authors contributed equally to this work


Kyoung-Hee Baik, Water Analysis and Research Center, K-water, 560 Sintanjin-ro, Daedeok-gu, Daejeon 306-711, Republic of Korea. Tel: +82 42 629 2010; Fax: +82 42 629 2079; e-mail: bkhee@kwater.or.kr


Three hundred and thirty-nine water samples obtained from 90 locations in Korea from 2007 to 2011 were tested for the presence of enteric viruses (EV), total coliforms (TC), and fecal coliforms (FC). A total culturable virus assay revealed that 89 samples (26.3%) were positive for EVs, the average concentration being 5.8 most probable number (MPN)/100 L. The Han river basin exhibited the highest contamination by EVs (occurrence, 41.3%; average concentration, 24.0 MPN/100 L). EV contamination was found more frequently in river water (occurrence, 33.6%; concentration, 8.4 MPN/100 L) than in lake water or groundwater. The concentration of EVs was highest in spring (7.7 MPN/100 L), whereas it was found most frequently in winter (36.1%). The number of TCs ranged from 0 – 1.2 × 105 colony forming units (CFU)/100 mL and that of FCs from 0–6.2 × 103 CFU/100 mL per sample. Statistical analyses showed that the presence of EVs, TCs and FCs did not correlate significantly with temperature or turbidity. In addition, presence of TCs and FCs was not significantly correlated with presence of EVs. In conclusion, TCs and FCs may not be accurate microbial indicators of waterborne EVs in Korean aquatic environments.

Surface water, such as river and lake water, represents much of Korea's source water. The number of people that use tap water in Korea is about 48.39 million (94.1% of the total registered population of 51.43 million) [1]. Since surface water is widely used as a source of tap water, improperly treated tap water from contaminated surface water could pose a significant threat to public health. Thus, it is important to properly monitor and maintain the quality of source water.

Waterborne EVs typically proliferate within the internal organs of humans and animals and leave their bodies in the feces [2], [3]. These viruses include enteroviruses, adenoviruses, hepatitis A virus, reoviruses, and noroviruses [4]. Such waterborne EVs are transmitted through contaminated water, foods, or infected persons and may be fatal to older people as well as those with weaker immune systems, particularly children [5], [6]. EVs that have infected humans reach concentrations of 108–1010 microorganisms/g in their feces [7]. Contamination of water sources by feces containing such high concentrations of pathogenic microorganisms can be detrimental to human health. In order to address this problem, the Korean government has been performing nationwide surveillance of EV contamination of source water in water treatment plants since 2003.

Indicator microorganisms indirectly represent the degree of soil pollution and water quality and may be used to estimate the presence of pathogenic microorganisms, including viruses [8], [9]. An ideal indicator microorganism is one that is nonpathogenic and able to be rapidly detected and quantified. Its presence should also correlate closely with the presence of the subject microorganism (the one being estimated) [10].

Coliform bacteria, namely TCs and FCs, are commonly used as indicators of fecal contamination or microbial pathogens in water. However, these microorganisms reportedly have inadequacies as indicator microorganisms ([11]–[15]). Studies on alternative indicator microorganisms such as enterococci, bacteriophages, coliphages and F-specific RNA coliphages have also been performed [12], [16]. However, because relationships between indicator and subject microorganisms may vary according to weather and geographical conditions, large-scale surveys using many different samples from various distinct regions are vital.

Thus, we performed a nationwide surveillance of EVs, TCs and FCs in 339 samples from 90 sites used as source water in Korea from 2007–2011. In addition, we investigated the feasibility of using TCs and FCs as microbial indicators of EVs.


Samples and water sampling

Three hundred and thirty-nine water samples were collected from 90 sites that served as source water for Korean water treatment plants from 2007–2011 (Table 1). Water sampling for viruses was carried out based on the Water Treatment Rules of the Korean Ministry of Environment [17] and the USEPA manual [18] using a 1-MDS positively charged filter (Cuno, Meriden, CT, USA). The samples were approximately 200 L in volume. During the sampling process, the water pressure was controlled to below 2.11 × 104 Pa and the pH reduced to 6.5–7.5 with 0.1 M HCl, if the original pH values were more than 8.0. The maximum flow was also controlled to below 11.4 L/min.

Table 1. Detection of enteroviruses, total coliforms and fecal coliforms in various water basins and water sources
Water basinsWater sourceNo. of samplesEVs (MPN/100 L)Occurrence of EVs (%)TCs (CFU/100 mL)FCs (CFU/100 mL)
  1. G, groundwater; L, lake water; M, mixed water (R + L or R + G); ND, not detected; R, river water.
  2. aSubtotal, total number of samples or arithmetic averages for each water basin.
  3. K.S., Keum/Seom; N, Nakdong; H. Han; Y. Yeongsan.
K.SR682.232.67.3 × 1031.5 × 102
 L26002.1 × 1031.9 × 102
 aSubtotal941.623.45.9 × 1031.1 × 102
NR641.425.04.1 × 1038.4 × 101
 L280.417.92.2 × 1039.1 × 100
 Subtotal921.122.83.5 × 1036.1 × 101
HR4929.542.92.9 × 1037.7 × 101
 L125.541.72.9 × 1031.8 × 100
 G2008.3 × 101ND
 Subtotal6324.041.32.8 × 1036.0 × 101
YR2002.7 × 101ND
 M60.816.74.2 × 1022.5 × 100
 Subtotal80.612.53.2 × 1021.9 × 100
OthersR374.640.52.3 × 1033.9 × 101
 L9005.5 × 1020.8 × 100
 G300.033.33.9 × 1023.5 × 100
 M62.350.02.1 × 1033.2 × 101
 Subtotal822.323.21.4 × 1032.1 × 101
TotalR2208.433.64.5 × 1039.4 × 101
 L751.01.32.1 × 1031.0 × 101
 G320.033.13.8 × 1023.3 × 100

Elution and concentration of viruses from water samples

A sterilized solution of 1.5% beef extract solution (1 L; pH 9.5, 0.375% glycerin) (BD, Franklin Lakes, NJ, USA) was used to elute and concentrate viruses from water samples. The beef extract was pressurized to enable it to flow into a cartridge housing, which included a water sample filter. When the filter was completely soaked with solution, the pressure was removed and the process left to continue for 1 min. The elution process was repeated twice. To concentrate the extract solution, it was mixed by agitation at speed that did not produce bubbles; the pH values in the eluted solution were adjusted to 7.0–7.5 with a 1 M HCl solution. The volume of the eluted solution was then measured and the pH values adjusted to 3.5 ± 0.1. After mixing the eluted solution without agitation at room temperature for 30 mins, it was centrifuged (2500 g at 4°C) for 15 mins. The supernatant was discarded and the precipitate mixed at room temperature for 10 mins by readjusting the pH to 9.0 following suspension in 30 mL of 0.15 M sodium phosphate solution. After centrifuging the eluted solution (4000 g at 4°C) for 10 mins, the supernatant (including the viruses) was collected and the pH of the supernatant adjusted to 7.3 ± 0.1. The supernatant was filtered using 0.22 μm sterilizing filters that had been presoaked with 1.5% beef extract, divided into subsamples 1 and 2, and frozen at −70°C until further use.

Cells and viruses

Buffalo green monkey kidney cells from between passage numbers 140 and 150 (kindly provided by the NIER, Incheon, Korea) were used to detect EVs. The cells were cultured at 36.5 ± 0.5°C in a 5% CO2 incubator in minimal essential medium/L-15 medium (Invitrogen, Grand Island, NY, USA) supplemented with 5% FBS (Invitrogen). Poliovirus type 3 (kindly provided by the NIER) was used as a positive control for CPE.

Total culturable virus assay

The volume of subsamples and the amount of virus inoculation were calculated according to the USEPA manual [18] and a previously described method [19]. A 0.15 M sodium phosphate solution corresponding to the inoculation volume was inoculated as a negative control and poliovirus type 3 (20 plaque forming units) also inoculated as a positive control. BGMK cells were cultured in 10 T25 cell culture vessels (Nalge Nunc International, Rochester, NY, USA) for 3–4 days. To inoculate the samples, BGMK cells were washed with serum-free Dulbecco's Modified Eagle's Medium and the wash medium discarded. The samples were inoculated into the vessels, which were rocked every 15 mins during the adsorption period to prevent cell death from dehydration in the middle of the vessels. After inoculating the samples for 90 mins, maintenance medium (10 mL) was added to the cell culture vessels for 14 days, and the CPE observed by using a microscope. When CPE was evident in more than 75% of the BGMK monolayer, the culture vessels were frozen at − 70°C; all remaining cultures (including controls) were frozen after 14 days. In addition, second cultures to confirm the results of the first cultures were conducted. The first cultures were thawed and filtered with 0.22 μm sterilizing filters. The filtered cultures (1 mL) were inoculated and cultured for 14 days, as previously described (18,19). Samples that exhibited CPE in both the first and second cultures were considered EV-positive. Samples that exhibited CPE only in the second culture were subjected to a third culture; those samples that exhibited CPE in both the second and third cultures were also considered EV-positive. The concentration of EVs was calculated by the Information Collection Rule Most Probable Number calculator (version 2.0), which was downloaded from the USEPA.

Assessment of total coliforms

The sample water (100 mL) was filtered using a filtration membrane (pore size, 0.45 μm; diameter, 47 mm) (Advantec MFS, Tokyo, Japan) and then cultured on Difco m Endo Agar LES (BD) medium at 35 ± 0.5°C for 24 hrs. Counts were performed on plates that displayed 20–80 colonies with a dark red, metallic luster and are presented as the number of TC colonies per 100 mL (CFU/100 mL).

Assessment of fecal coliforms

The sample water (100 mL) was filtered in the same manner as that described for TCs and cultured in Difco mFC agar (BD) medium at 44.5 ± 0.2°C for 24 hrs. Counts were carried out on plates with 20–60 blue colonies and are presented as the number of FC colonies per 100 mLs (CFU/100 mL).

Statistical analysis

To investigate the use of TCs and FCs as microbial indicators of EVs and to explore the relationship between those waterborne microorganisms and physical-chemical environmental factors, SPSS 12.0 software for Windows was used to perform Pearson correlation and regression analyses. P values below 0.05 were considered statistically significant.


Detection of enteric viruses, total coliforms and fecal coliforms

Infectious EVs were detected in 89 of 339 samples (26.3%), the detection range and average concentration being 0–491.0 MPN/100 L and 5.8 MPN/100 L, respectively. Among the water systems studied, the Han river water system was the most frequently and most severely contaminated by viruses (average concentration, 24.0 MPN/100 L; incidence, 41.3%) (Table 1). River water (average concentration, 8.4 MPN/100 L; incidence, 33.6%) was more frequently and more severely contaminated than were lake water (average concentration, 1.0 MPN/100 L; incidence, 1.3%) and groundwater (average concentration, 0.03 MPN/100 L; incidence, 3.1%). Among the 339 samples in this study, four (from two different locations) had EV concentrations of over 100 MPN/100 L. However, four further analyses of treated water originating from these two water treatment plants within one year failed to detect EVs (data not shown). Seasonal virus occurrence was also investigated and revealed that, whereas the incidence of EVs was highest in winter (36.1%), their concentration during this season was relatively low (2.0 MPN/100 L) (Table 2). The geometric average concentration of EVs in winter (3.8 MPN/100 L) was also lower than that of EVs in spring (5.1 MPN/100 L) and summer (5.5 MPN/100 L) (data not shown). TCs were detected in the range of 0–1.2 × 105 CFU/100 mL, the average concentration being 3.5 × 103 CFU/100 mL (Table 2). FCs were detected in the range of 0–6.2 × 103 CFU/100 mL, the average concentration being 6.4 × 101CFU/100 mL (Table 2).

Table 2. Seasonal variations in physical-chemical factors and waterborne microbial pathogens
SeasonNTemp (°C)pHTurbidity (NTU)EVs (MPN/100 L)Occurrence of EVs (%)TCs (CFU/100 mL)FCs (CFU/100 mL)
  1. N, number of samples.
  2. Spring months: March, April, May; summer months: June, July, August; autumn months: September, October, November; winter months: December, January, February.
Spring12813.07.53.0207.725.02.0 × 1037.9 × 100
Summer8222.97.26.1855.724.45.1 × 1031.2 × 102
Autumn9318.67.54.1464.725.84.4 × 1031.2 × 102
Winter368.17.43.3842.036.12.3 × 1035.1 × 100
Total33916.47.44.1335.826.33.5 × 1036.4 × 101

Relationship between environmental factors

As shown in Figure 1a, concentrations of TCs and FCs varied according to water temperature. In particular, during June–September (high temperatures) TCs were detected at a concentration of 5.4 × 103 CFU/100 mL, which is ∼1.6 times higher than the average concentration detected: 3.4 × 103 CFU/100 mL. FCs were detected at a concentration of 1.6 × 102 CFU/100 mL, which was ∼2.5 times higher than the average concentration detected: 6.4 × 101 CFU/100 mL (1a). To determine whether there was a statistically significant relationship between presence of the indicator microorganisms and water temperature, Pearson correlation and regression analyses were performed. Both TCs and FCs failed to show statistically significant correlations with water temperature (Table 3). As shown in Figure 1b, no statistically significant correlations were found between presence of EVs and water temperature, even though there were relatively high detection rates (R = 0.082, R2 = 0.007, P = 0.000) in March and November (low temperatures). High concentrations of viruses have reportedly been detected at low water temperatures in Korea [20], whereas in the present study high concentrations of viruses attributable to increased turbidity were detected in July, August, and September (high water temperatures). In Korea, these months are rainy; the average rainfall on the sampling dates in July, August, and September (5.4, 7.8, and 8.9 mm/day, respectively) were higher than those (0.4–3.5 mm/day) of other months. In addition, it is also possible that turbid materials in environmental water were responsible for the less frequent detection of high concentrations of EVs in summer; high turbidity can result in impaired efficiency of recovery of EVs from water.

Table 3. Correlation coefficients between microbial indicators and environmental factors or waterborne enteric viruses
SamplesRR2P value
  1. R, Pearson correlation coefficient; R2, multiple correlation coefficient; P value < 0.05, except for asterisked result.
Temperature vs. TCs0.1460.0210.000
Temperature vs. FCs0.1390.019 0.221*
Temperature vs. EVs0.0820.0070.000
Turbidity vs. TCs0.1990.0390.000
Turbidity vs. FCs0.1500.0220.000
Turbidity vs. EVs0.0460.0020.039
TCs vs. EVs0.0250.0010.005
FCs vs. EVs0.0150.0000.007
Figure 1.

Effect of temperature on total coliforms, fecal coliforms and waterborne enteric viruses. (a) Relationship between concentrations of coliforms and temperature (closed circles). (b) Relationship between concentrations of enteric viruses and temperature (closed circles).

The relationship between turbidity and detection of indicator microorganisms and viruses was also investigated. The range of turbidity of the samples was between 0.06 NTU (minimum) and 59.20 NTU (maximum). Water sampled during July–September exhibited high turbidity, suggesting that the presence of detectable EVs, TCs, and FCs may be related to turbidity (Fig. 2). However, high concentrations of EVs were detected in the winter season, when turbidity was low; thus, the pattern for EVs differed from that of the indicator microorganisms (2b). The results of Pearson and regression analyses for determining general correlations (Table 3) showed that although there were very weak correlations between TCs and FCs and turbidity (TCs: R = 0.199, R2 = 0.039, P = 0.000; FCs: R = 0.150, R2 = 0.022, P = 0.000), there were no statistically significant correlations between EVs and turbidity (R = 0.046, R2 = 0.002, P = 0.039).

Figure 2.

Effect of turbidity on total coliforms, fecal coliforms and waterborne enteric viruses. (a) Relationship between concentrations of coliforms and turbidity (open circles). (b) Relationship between concentrations of enteric viruses and turbidity (open circles).

Total and fecal coliforms as indicator microorganisms

Whether detection of TCs and FCs was directly related to that of EVs was also investigated. A relationship between presence of TCs and EVs was observed in July (high water temperatures and turbidities, Fig. 3). However, correlation and regression analyses revealed no statistically significant correlations between presence of TCs and FCs and that of EVs (Table 3). Thus, TCs and FCs are not suitable indicator microorganisms for EVs in drinking source water.

Figure 3.

Relationship between waterborne enteric viruses and total coliforms and fecal coliforms Closed triangles indicate concentrations of waterborne enteric viruses.


We performed nationwide surveillance of EVs, TCs and FC in Korean aquatic environments for five years to investigate the feasibility of using TCs and FCs as microbial indicators of the presence of EVs. We showed that contamination of river water was more frequent and severe than that of lake water and groundwater. These differences may be because lake water has fewer sources of contamination and its volume is greater compared with river water. In addition, because contaminating substances are naturally removed from groundwater by filtration through the soil layer, EV contamination of river water is likely to be greater than that of groundwater,. However, river water constitutes approximately 46.2% of all source water in Korea [1], the Han river being the main source water for metropolitan areas. Thus, continuous monitoring and management of river water for EV contamination is important for public health.

According to the Korean Water Treatment Rules [17], if the concentration of EV in untreated source water is over 100 MPN/100 L, treated water should be assessed for EV contamination four times a year. In this study, although the EV concentrations of four samples were over 100 MPN/100 L, we detected no EVs in treated water. In this study, EVs were detected less often (26.3%) than was reported for the Busan area of Korea (61.9%) by Park et al. [7], but more often than in a study of national water sources between 2001 and 2005 (3.7%) performed by Kim et al. [21]. However, there were no significant differences between the occurrence of EVs in the present study and that identified by surveillance by Lee and Lee between 2003 and 2006 (32.8%) [20]. These differences in the reported occurrence of EVs may be because the studies investigated different regions and seasons. It is difficult to ascertain the reasons for these differences because few studies have investigated EVs on a nationwide scale using TCVA. However, there are reportedly some differences in the detection rate of EVs when nested RT-PCR or integrated cell culture RT-PCR methods are used to assess and noroviruses in groundwater ([22]–[25]).

The detection rates of TCs and FCs in this study also differed from those previously reported. Rhee et al. studied the detection rate of TCs and FCs in the Han river water system and reported differences in concentrations of TCs and FCs detected from year to year [26]. The arithmetical averages of TC concentrations were 3.7 × 103 CFU/mL, 2.7 × 104 CFU/mL and 4.7 × 102 CFU/mL in 2000, 2002 and 2004, respectively, and those of FCs were 3.7 × 102 CFU/mL, 2.7 × 103 CFU/mL and 3.7 × 101 CFU/mL in 2000, 2002 and 2004, respectively. The detection rate may vary significantly from year to the year depending on changes in sewage treatment facilities. Lee et al. could find no consistent correlations between detection rate and water quality in basins, timing of investigation, or the method (membrane filtering and MPN methods) used for detecting TCs and FCs [27]. Therefore, differences in detected concentrations of TCs and FCs may be due to differences in the degree of pollution in the target regions and the detection methods employed ([26]–[28]).

Various environmental factors may affect the presence of microorganisms in aquatic environments. Specifically, temperature and turbidity have significant effects on waterborne microorganisms. We therefore investigated the relationship between these factors and the presence of TCs, FCs and EVs and found no statistically significant correlations with water temperature. However, our findings did suggest that presence of TCs and FCs might correlate weakly with turbidity. Although in temperate climate regions EVs are most often detected in late summer to early fall, detection of EVs may vary widely due to regional characteristics. Indeed, in some regions viruses are reportedly detected more frequently in winter than in summer [7], [29]. In the present study, we also detected viruses more frequently in the winter season (low water temperatures) than in summer. We more frequently detected EVs when the turbidity was high; however, in winter, when there are low water temperatures and turbidities, we also detected EVs frequently (data not shown).

We have shown that TCs and FCs are not suitable indicator microorganisms for EVs. Studies on the feasibility of using TCs, FCs, fecal streptococcus, bacteriophages, and coliphages as indicator microorganisms for EVs have produced some negative results regarding whether these indicator microorganisms accurately represent the presence of pathogenic microorganisms [11], [12], [14], [30]–[33].

Since it is difficult to detect low concentrations of EVs in water accurately by using TCVA, studies on indicator microorganisms are important. Ideally, such indicator microorganisms could serve as indirect indexes of the presence of pathogenic microorganisms and the degree of pollution of soil and water [8].

Based on the results of the present study of Korean basins, we conclude that TCs and FCs are not suitable for use as indicator microorganisms for EVs. Recent developments in molecular biology-based virus detection methods include PCR, real-time PCR, nucleic acid sequence-based amplification analysis and DNA microarrays [34]–[39], as well as integrated cell culture-RT-PCR for faster detection of infectious EVs [39]. Therefore, further investigation of methods that can easily and quickly detect low concentrations of EVs in water is required, as well as on the use of various indicator microorganisms to estimate contamination by waterborne EVs.


No authors have any conflicts of interest to disclose.