|Home | About | Journals | Submit | Contact Us | Français|
We investigated coliphages from various fecal sources, including humans and animals, for microbial source tracking in South Korea. Both somatic and F+-specific coliphages were isolated from 43 fecal samples from farms, wild animal habitats, and human wastewater plants. Somatic coliphages were more prevalent and abundant than F+ coliphages in all of the tested fecal samples. We further characterized 311 F+ coliphage isolates using RNase sensitivity assays, PCR and reverse transcription-PCR, and nucleic acid sequencing. Phylogenetic analyses were performed based on the partial nucleic acid sequences of 311 F+ coliphages from various sources. F+ RNA coliphages were most prevalent among geese (95%) and were least prevalent in cows (5%). Among the genogroups of F+ RNA coliphages, most F+ coliphages isolated from animal fecal sources belonged to either group I or group IV, and most from human wastewater sources were in group II or III. Some of the group I coliphages were present in both human and animal source samples. F+ RNA coliphages isolated from various sources were divided into two main clusters. All F+ RNA coliphages isolated from human wastewater were grouped with Qβ-like phages, while phages isolated from most animal sources were grouped with MS2-like phages. UniFrac significance statistical analyses revealed significant differences between human and animal bacteriophages. In the principal coordinate analysis (PCoA), F+ RNA coliphages isolated from human waste were distinctively separate from those isolated from other animal sources. However, F+ DNA coliphages were not significantly different or separate in the PCoA. These results demonstrate that proper analysis of F+ RNA coliphages can effectively distinguish fecal sources.
Fecal contamination of various water resources poses a serious risk to human health through consumption of microorganisms that inhabit these environments, as well as through recreational exposure (20). To correctly assess and properly manage the human health risk associated with water contamination, it is necessary to have information regarding the source of fecal contamination (12, 19, 30, 31, 33). In previous studies, several microbiological methods have been applied to distinguish fecal contamination sources, particularly between human and animal sources (20, 30). For example, different types of antibiotic resistance patterns (14), molecular markers (2), identified animal viruses (26), and whole-genome patterns based on repetitive-element PCR or ribotyping (4, 28) have been applied to microbial source tracking (MST). However, these MST methods have not yet been fully evaluated and characterized.
Coliphages are viral indicators of fecal contamination in groundwater proposed by the U.S. Environmental Protection Agency (37) and could be useful target microorganisms for MST. Based on infectivity through the host's sex pili, coliphages can be categorized into somatic and F+-specific coliphages. In addition, coliphages can be categorized based on the type of nucleic acids (RNA versus DNA). At present, F+ coliphages consist of those from Leviviridae (icosahedral, single-stranded RNA phages), including Levivirus and Allolevivirus, and Inoviridae (filamentous, single-stranded DNA phages) (5). Levivirus contains both MS2-like (group I) and GA-like (group II) viruses, whereas Allolevivirus contains both Qβ-like (group III) and SP-like (group IV) viruses (32). These subgroups of coliphages were initially classified based on serological typing (18). Different genotypes of F+ RNA coliphages are associated with different types of fecal sources (30, 33). A number of previous studies have reported that groups II and III are isolated mainly from human feces and that groups I and IV are associated mainly with animal feces (10, 17, 18, 27). However, the specificity of this association can vary, because group I F+ RNA coliphages were isolated from human-waste-dominated municipal wastewater (13). In addition, the applicability of this MST method in different geographical regions, such as Asia, is still unclear. Furthermore, the application of DNA coliphages to MST has not been much investigated.
In several previous studies, F+ coliphages were isolated and analyzed using culture and serological methods (7, 22, 34). Nucleic acid sequence-based analyses of F+ RNA-specific genes have not been performed or are very limited (40). Stewart et al. (32) performed F+ RNA group III sequence analysis, and MST based on direct nucleic acid sequence was found to be more reliable than nucleic acid hybridization. However, a comprehensive analysis based on nucleic acid sequence has not been applied to coliphages. Thus, the objectives of this study were (i) to test the applicability of both RNA and DNA coliphage-based MST in South Korea and (ii) to apply nucleic acid sequence-based analyses of both RNA and DNA coliphages isolated from various animal and human fecal sources to MST.
Thirty-two fecal samples from pigs, cows, chickens, and wild animals were collected from different animal farms or wild animal habitats. All of the animal farms were located in rural areas near Seoul, South Korea. The sampling sites of wild animal habitats were located in the Chonnam area of the southwestern part of South Korea, where migrating birds from Siberia rest every winter. The fecal samples were collected using disposable polyvinyl bags, which were stored in polyethylene containers at 4°C until analysis. All animal fecal samples, except wild animal feces, were processed within 8 h of sampling and analyzed within 24 h. Since the sampling site for the wild animal feces was relatively far away, the wild animal feces were transported into the laboratory on dry ice and then analyzed within 24 h.
Municipal wastewater (containing human sewage and feces) samples were collected from the influents of sewage treatment plants in the metropolitan area of Seoul, South Korea. Since human feces and animal feces are treated separately at different waste treatment plants, we collected the samples from wastewater treatment plants only for human wastes. In total, 11 samples were collected from these wastewater treatment plants. Human wastewater samples were collected using autoclaved polyethylene bottles and stored at 4°C until analysis. Samples were analyzed within 24 h of collection.
Approximately 5 g of collected fecal samples was suspended in 20 ml phosphate-buffered saline (pH 7.4) and vortexed for 5 min. The fecal suspensions were centrifuged at 5,000 × g for 20 min, and the supernatant was carefully collected for further analysis. For human wastewater samples, each sample was centrifuged at 5,000 × g for 20 min to remove large debris, and the supernatant was carefully collected for further analysis. The resulting supernatants were analyzed using single-agar-layer methods (36). Briefly, to isolate somatic and male-specific (F+) coliphages from the supernatant, Escherichia coli CN13 (ATCC 700609) and E. coli Famp (ATCC 700891) were used, respectively. From the plaques observed on the single-agar-layer plate, approximately 20 coliphage isolates were selected using a sterile wooden toothpick, and each isolate was subjected to two-step enrichment procedures (35). The sterile toothpick used in the plaque assay was suspended in 5 ml tryptic soy broth containing 100 μl log phase of either E. coli CN13 or E. coli Famp. The tube was gently vortexed for 1 min, and the phage isolated from a single plaque was incubated at 37°C for 16 to 18 h at 150 rpm. After cultivation, chloroform was added, and the tube was vigorously vortexed for 5 min. The suspension was centrifuged at 5,000 × g for 20 min, and the single phage isolate in the resulting supernatant was removed and stored at −70°C until further analysis. For each experiment, either E. coli MS2 (ATCC 15597-B1) or X174 (ATCC 13706-B1) was used as a positive control of somatic or male-specific coliphages, respectively. Sterile phosphate-buffered saline was used as the negative control.
To distinguish F+ DNA and F+ RNA coliphages, each isolated F+ coliphage was subjected to an RNase sensitivity assay (18). Escherichia coli Famp was cultivated to exponential phase and poured into a 150-mm-diameter petri dish in 0.8% tryptic soy agar with and without RNase (100 μg/ml) (Sigma-Aldrich, St. Louis, MO). Serial dilutions (5 μl; undiluted, 10−2, 10−4, and 10−6) of the isolated coliphages were spotted onto both RNase-positive and RNase-negative plates. The plates were incubated for 12 to 18 h at 37°C. When the plaque was observed in the presence of RNase, the phage was considered to be a DNA coliphage. When a plaque was observed only on the RNase-negative plate, the bacteriophage was considered to be an RNA coliphage.
F+ coliphage strains were amplified using primers as previously described (38). Either DNA coliphages or RNA coliphages were heat released and diluted 1:50. The released nucleic acid was amplified using either PCR or reverse transcription-PCR (RT-PCR) assays as described in a previous study (38). Amplified PCR/RT-PCR products were purified using the QIAquick PCR purification kit (Qiagen, Valencia, CA) and sequenced at Cosmo Gene Tech (Seoul, South Korea).
Chromatographs of each sample sequence were checked and edited using SeqMan software (Lagergene 6.0; DNAStar, Inc., Madison, WI). Nucleic acid sequences with unclear chromatographs or shorter fragments were excluded from further data analysis. In addition, any possible clone from the same fecal sample was excluded as well. Each sequence of analyzed samples was exported to GenBank and confirmed using BLAST. All sequences with both F+ RNA and F+ DNA coliphages were imported into the ARB library (25). In addition, all sequences of Leviviridae and Inoviridae (M13-like phages) available in GenBank were downloaded and imported into the ARB software package. Sequences were aligned, and phylogenetic analysis was performed based on the neighbor-joining method.
To distinguish the nucleic acids of bacteriophages from both animal and human fecal sources, a principal coordinate analysis (PCoA) was performed using UniFrac software (23). The nucleic acid sequences of bacteriophages were aligned using Clustal W and exported to the UniFrac software for PCoA. PCoA was performed to determine whether bacteriophages from different fecal sources were distinctive using either UniFrac significance tests or the P test. F+ RNA and F+ DNA coliphages were analyzed separately.
The nucleic acid sequences of novel F+ RNA coliphage strains were deposited into GenBank and assigned GenBank accession numbers GQ268004 to GQ268012.
Table Table11 summarizes the prevalence of somatic and F+ coliphages from various fecal sources in Korea. In most fecal sources, somatic coliphages occurred with a high incidence (86 to 100%). However, the prevalence of somatic coliphages in cow feces was lower than in other fecal sources (58%). In the fecal samples from a pig septic tank and from chickens, the concentrations of somatic coliphages ranged from 380 to 13,700 PFU/g (median density of 680 PFU/g) and from 3 to 10,000 PFU/g (median density of 1,450 PFU/g), respectively. On the other hand, the concentrations of somatic coliphages in cow feces ranged from 1 to 25 PFU/g (median density of 11 PFU/g).
The prevalence of F+ coliphages was lower than that of somatic coliphages (Table (Table1).1). F+ coliphages were abundantly found (73 to 100%) in human wastewater, the pig septic tank, and goose feces. On the other hand, F+ coliphages had a lower prevalence in cow feces (25%). F+ coliphages in human wastewater ranged from <1 to 285 PFU/ml (median, 5 PFU/ml). The median number of coliphages found in pig fecal samples was slightly higher than in other fecal samples (Table (Table11).
To identify the type of nucleic acids (either DNA or RNA) in isolated coliphages, the sensitivity of 311 F+ coliphage isolates to RNase was tested using a spot plate assay (Table (Table2).2). Among the 311 bacteriophage samples that were tested, the numbers of F+ DNA and F+ RNA coliphages were 157 and 154, respectively. F+ DNA coliphages were more prevalent among cow (95%) and chicken (79%) feces. In both pig feces and the pig septic tank, F+ RNA coliphages were more prevalent (64% and 68%) than F+ DNA coliphages, and most bacteriophages from goose fecal samples were F+ RNA coliphages (95%). In human wastewater, the prevalences of F+ DNA and F+ RNA coliphages were 59% and 41%, respectively.
In total, 311 nucleic acid sequences from isolated F+ coliphages were confirmed and classified using both BLAST searches and phylogenetic analyses. Based on comparisons to the standard reference strains, each of the 133 RNA coliphage isolates from various fecal sources was classified into one of four groups (groups I to IV). Over half (52%) of the RNA coliphages were categorized into group I, 10% into group II, 12% into group III, and 19% into group IV (Table (Table3).3). Most coliphages isolated from animals were categorized into groups I and IV, except one isolate from a goose fecal sample. Most RNA coliphages isolated from human wastewater samples were categorized into groups II (32%) and III (42%). Ten RNA coliphage isolates (26%) were categorized into group I. Approximately nine isolates (7%) did not share any similarity with previously identified RNA coliphages and were classified into a novel group. Among the 131 F+ DNA coliphages, 109 isolates (83%) were identified as f1-like phages. Eighteen (13%) and four (3%) isolates were identified as fd-like and M13-like phages, respectively. Of the 64 F+ DNA coliphages isolated from human wastewater samples, 47 (73%) were identified as f1-like, 14 (22%) as fd-like, and 3 (5%) as M13-like bacteriophages. Most of the F+ DNA coliphages were grouped as either f1-like or fd-like bacteriophages.
To investigate the genetic diversity of isolated F+ coliphage sequences, phylogenetic analyses were performed using ARB software.
Partial regions of replicase genes (266 bp [Levivirus] and 229 bp [Allolevivirus]) were PCR or RT-PCR amplified and subsequently sequenced. Phylogenetic analyses were performed with all Leviviridae nucleic acid sequences from GenBank. In total, 133 isolated F+ RNA coliphages were grouped into two major clusters (Qβ-like and MS2-like). Among the isolated F+ RNA coliphages, the majority of the phages from human wastewater sources belonged to either GA-like (group II) or Qβ-like (group III) phages (Fig. (Fig.1A).1A). The RNA coliphages isolated from animal fecal sources belonged to either MS2 (group I)- or FI (group IV)-like phages. Bacteriophages isolated from pig septic tank fecal sources were identified as a novel separate group.
To investigate the genetic variability of F+ DNA coliphages, the nucleic acids of the gene IV region were sequenced (38). All F+ DNA coliphages from various sources and reference strains were categorized into two different major clusters (fd-like and M13-like). The majority of phages from human wastewater sources were identified as fd-like phages. In addition, 1 phage from chicken and 16 phages from swine septic tank sources were included as well in the fd-like phages. The phages from the other animal fecal sources, including chicken, cow, and pig, belonged to the M13- and f1-like phage groups (Fig. (Fig.1B1B).
To determine differences among F+ coliphages isolated from different fecal sources, we performed significance tests and PCoA based on the phylogenetic analysis. The genetic characteristics of F+ RNA bacteriophages isolated from human fecal samples were significantly different from those of the other samples, as determined by the P test (P value < 0.05), except the cow samples (Table (Table4).4). The significant P values indicated that F+ RNA coliphages isolated from various sources have unique genetic characteristics. Among the fecal samples from various sources, only the F+ RNA coliphages from humans were significantly different from F+ RNA coliphages in the other sources (P ≤ 0.001) (Table (Table4).4). These results indicated that the sequences from this sample are associated with unique branch lengths in the tree, suggesting that F+ RNA coliphages from human sources are distinctive from F+ RNA coliphages isolated from other fecal sources. However, when F+ DNA coliphages were analyzed using the same significance analysis, the P test result was not significant among the various source samples (data not shown). F+ RNA coliphage significance test observations might best be illustrated using PCoA (Fig. (Fig.2A).2A). When a scatter plot of the first two principal components (PCs) was analyzed, both PC1 and PC2 explained the major variation in the data (50.9 and 24.2%, respectively). PC2 clearly separated all fecal source isolates from human source isolates (Fig. (Fig.2A).2A). However, all fecal samples from animal or human sources were not clearly separated using PCoA of F+ DNA coliphages (Fig. (Fig.2B2B).
Currently, little information is available for describing the genetic characteristics of both F+ and somatic coliphages for MST. This study is the first to apply nucleic acid sequence-based genotyping methods for both F+ DNA and F+ RNA coliphages to MST. Bacteriophages have been considered useful targets in MST. However, MST is region specific and has not been studied much in Asia (11). Our study confirmed that F+ RNA coliphages isolated in Korea have genetic characteristics similar to those of coliphages isolated in other regions (30, 33). Genogroups I and IV were found mostly in animal fecal samples, and genogroups II and III were found mostly in human-waste-dominated wastewater samples. These results coincide with those of other, previous studies (10, 17, 27). However, 26% of F+ RNA coliphages isolated from human wastewater samples were identified as genogroup I. Thus, the identification of fecal sources solely on the basis of their F+ RNA coliphages belonging to genogroup I would be limited (33).
F+ coliphages were commonly detected in human and animal fecal samples in this study (Table (Table1).1). These results are consistent with those of other studies. Previous studies reported that the prevalences of F+ coliphages were 79%, 60%, and 100% in human waste, goose waste, and swine fecal wastewater, respectively (3, 5). Concentrations of F+ coliphages in pig fecal samples were slightly higher than those in other animal sources, such as chicken, cow, and goose feces. In previous studies, F+ RNA coliphages were rarely detected in fecal samples from adult chickens and cattle but were more often detected in pig fecal samples (17). In this study, a much lower prevalence (25%) was observed only in cow fecal samples. Another study reported that F+ coliphages were either not detected or not commonly detected in cow sources (16). Different microbial communities may be present in the cow gut because of different digestion mechanisms and ingested food (1). Further study should be focused on the different characteristics of microbial ecology in the guts of cows and other animals with respect to diet and other factors.
The ecological mechanism by which F+ coliphages became prevalent in human wastewater and septic tanks, but not in individual fecal samples, is not fully understood. In F+ coliphages, prolonged replication appears improbable because a minimum host density is required for replication (39). In both wastewater and septic tanks, host density could increase because feces from many different individuals are combined. The frequency of individual fecal samples containing F+ coliphages is not typically high (33). However, if any individual fecal sample contains F+ coliphages within a wastewater sample, the F+ coliphages may be able to effectively replicate in the wastewater, owing to the high density of the host population (33, 39).
The RNase test was applied to distinguish F+ RNA and F+ DNA coliphages. To confirm the results of RNase testing, either PCR or RT-PCR analysis for each of the bacteriophage isolates was also performed. These results indicated that the results of both the RNase test and PCR/RT-PCR assays were generally consistent with each other. However, the results were not consistent for bacteriophages isolated from some human wastewater samples and animal fecal samples. In human and animal wastewater samples, several F+ DNA coliphages that were identified by spot assay were then identified as F+ RNA coliphages using RT-PCR analysis. These results suggest that the RNase sensitivity assay could be inaccurate in some cases, resulting from its inability to completely inactivate F+ RNA bacteriophages. Havelaar (15) found that F+ RNA coliphages could become resistant to RNase, which could lead to a false-negative result. Hence, resistance to the RNase spot assay does not necessarily indicate the presence of F+ DNA coliphages (40). On the other hand, F+ RNA coliphages were identified as F+ DNA coliphages by PCR analysis in animal fecal samples. In this case, F+ DNA coliphages may be inactivated by RNase in some environments.
This study suggested that the ratio of F+ DNA coliphages to F+ RNA coliphages indicated the sources of the fecal samples. For example, F+ DNA coliphages were much more prevalent (95%) than F+ RNA coliphages (5%) in cow feces. On the other hand, F+ RNA coliphages were more prevalent (95%) than F+ DNA coliphages (5%) in goose feces. These results are consistent with those of a previous study (5). Cole et al. (5) reported a high ratio of F+ DNA coliphages (82%) to F+ RNA coliphages (18%) in cow feces and a low ratio of F+ DNA (0%) to F+ RNA (100%) coliphages in goose feces. For human and pig fecal samples, a mixed ratio of F+ DNA and RNA coliphages was observed, consistent with results of a previous study (22). This study indicated that both F+ DNA (59%) and F+ RNA (41%) coliphages were present in human wastewater and that, similarly, F+ DNA (32%) and F+ RNA (68%) coliphages were found together in pig fecal samples. A previous study also showed that F+ DNA (65%) and F+ RNA (35%) coliphages were present in human wastewater and that F+ DNA (34%) and F+ RNA (66%) coliphages were prevalent in a hog lagoon (22).
In this study, F+ RNA coliphages were grouped based on the nucleic acid sequences of a partial replicase gene. As indicated in other studies, most F+ RNA coliphages from animal fecal samples belonged to either group I or group IV, and the majority of F+ RNA coliphages isolated from human wastewater belonged to groups I, II, and III. Group I F+ RNA coliphages were thus found in both human wastewater and animal fecal samples. Many previous studies have reported that group I coliphages were predominantly found in human-waste-dominated wastewater (5, 6, 33). Schaper et al. (29) found that group I coliphages were more persistent than other groups of coliphages under various environmental conditions and stresses. This persistence is an important factor in their presence in wastewater.
Few studies have investigated the prevalence and ecology of F+ DNA coliphages (9), and one study suggested F+ DNA coliphages as a potential indicator for source tracking (22). Until now, little information has been available as to whether specific subgroups of F+ DNA coliphages are associated with specific fecal sources (38). Unlike the previous study that used the serotypes of F+ DNA coliphages (22), this study investigated F+ DNA coliphages as a potential target microorganism in MST using DNA sequence-based genotypes. Typically, DNA is more stable than RNA, which suggests that F+ DNA coliphages could be better targets for MST (38). However, our results suggest that F+ RNA coliphages are more strongly associated with the ability to discriminate between human and animal sources than F+ DNA coliphages are. Further studies should be conducted to investigate new types of F+ DNA bacteriophages and specific genetic markers to be used in MST.
Somatic coliphages were commonly found in both animal-waste- and human-waste-dominated wastewater samples. Concentrations of somatic coliphages varied from 30 PFU/ml to 3.8 × 103 PFU/ml in human-waste-dominated wastewater and from 1 PFU/g to 1.4 × 104 PFU/g in animal fecal samples (Table (Table1).1). These observations are similar to those found in other studies (22, 33). A previous study reported 22 PFU/ml to 3.6 × 103 PFU/ml in wastewater, 3.8 × 102 to 1.4 × 104 PFU/g in hog lagoons, 1 to 2 PFU/g from cow sources, and 1 × 104 to 4.7 × 106 PFU/g from chicken litter (33). Because somatic coliphages are a more diverse group than F+ coliphages, it is more difficult to identify target genetic markers for MST. However, if the right genetic marker were identified in somatic coliphages, they would be more attractive for MST because of their high prevalence in both individual fecal samples and the environment.
Phylogenetic analyses of isolated bacteriophages were performed using the ARB software program (25). These analyses showed that F+ RNA coliphages isolated from various sources were divided into two main clusters (Fig. (Fig.1A).1A). All F+ RNA coliphages isolated from human wastewater were grouped with the Qβ-like phages, while F+ RNA coliphages isolated from pig, chicken, and pig septic tank sources were grouped with the MS2-like phages. These results are encouraging because different phylogenetic characteristics are exhibited based on the types of human and animal fecal sources. The novel sequences were identified from collected animal fecal samples. However, when the phylogenetic analysis was performed for F+ DNA coliphages (Fig. (Fig.1B),1B), there was no distinction between human and animal fecal sources.
In this study, the UniFrac method was applied to distinguish between the fecal sources based on the phylogenetic analysis. UniFrac is a bioinformatics tool to test the statistical significance of phylogenetic distances among different samples. Lozupone and Knight demonstrated that UniFrac is robust even for very similar samples when the sample size is large (24). Recently, several studies applied UniFrac to compare samples from multiple communities (8, 21, 23). However, this tool has never been applied to MST. The results of the P test showed significant differences (P < 0.005) between F+ RNA coliphages isolated from human and animal fecal samples. This result indicates that RNA coliphages could be applied in MST as target microorganisms. To determine whether bacteriophages isolated from various sources were different based on a unique branch length in the phylogenetic tree, the UniFrac significance test was performed (23). Among various fecal source samples, bacteriophages isolated from human wastewater were significantly different from those isolated from other samples (P < 0.005) (Table (Table4).4). This result suggests that F+ RNA coliphages isolated from human wastewater were clearly distinctive from those from other animal fecal samples. Furthermore, a PCoA confirmed these results. The PCoA whose results are shown in Fig. Fig.2A2A showed that F+ RNA coliphages isolated from human wastewater were clearly separated from those from other animal sources. Together, these results indicate that F+ RNA coliphages from human sources were significantly different from those from other fecal sources. However, when F+ DNA coliphage was analyzed, no distinctive pattern was observed (Fig. (Fig.2B).2B). It also needs to be acknowledged that while phylogenetic analysis and PCoA were able to distinguish F+ RNA coliphage isolates from known fecal sources, analysis of an environmental water sample receiving mixed fecal inputs in a real-world setting would be more complicated. In conclusion, our study demonstrated that bacteriophages could be very useful targets in MST. Different criteria, such as genogroup of bacteriophages and PCoA, could be applied for identifying fecal sources in environmental samples.
This study was supported by the Ministry of Environment of the Republic of Korea as the Eco-Technopia 21 Project (900-20080010).
Published ahead of print on 18 September 2009.