General and host-associated bacteriophage indicators of faecal pollution

Bacteriophages as indicators of general and host-associated faecal pollution

1.0 General Characteristics of Bacteriophages (Phages)

Bacteriophages or viruses that infect bacteria are extremely abundant in nature, probably the most abundant life form on Earth. They outnumber bacteria in most studied habitats (Weinbauer, 2004), including human and animal –associated microbial communities (Letarov and Kulikov, 2009).

Phages which infect faecal bacteria are important virus indicators and viral surrogates for wastewater treatment efficacy. They are very useful in pollution assessment and control but like all phage carry on very specific activities in nature including in sewage and wastewater environments.

Phage infect and lyse bacteria thus they contribute to bacterial mortality, releasing organic compounds. They have an important impact on the cycling of organic matter in the biosphere (Suttle, 1994). On the other hand, they control microbial diversity by selecting for some types of bacteria that are resistant to their attack. As well, many phages can mobilize genetic material among different host bacteria in a process known as transduction. By transduction, genetic material can be introduced into a bacterium by a phage that has previously replicated in another bacterium, in which it packaged random DNA fragments (generalized transduction) or the DNA adjacent to the prophage attachment site (specialized transduction).

Phages can only replicate in metabolizing host cells. Once within a host bacterium, phages replicate by one of two ways, the lytic and the lysogenic cycles. In the lytic cycle, phages immediately after infection multiply, rapidly disrupting their host cells from within and then releasing the phage progeny that varies regularly between 10 and 1000 depending on the phage and the physiological status of the host cell. This process can be as short as 25-30 minutes in fast growing host bacteria. Phages that only follow the lytic cycle are known as virulent phages. In contrast, temperate, or lysogenic, phages can follow the lysogenic cycle, in which the genome of the temperate phage remains in the host bacteria, replicating along with it. At this stage, the phage is known as a prophage, which can be induced to follow the lytic cycle. Induction occurs either spontaneously or when stimulated by natural or anthropogenic inductors.

Bacteriophages can only replicate inside susceptible host bacteria. A given phage can only infect certain bacteria to the point that different strains of the same species differ in their susceptibility to phage attack (Muniesa et al., 2003; Thingstad et al., 2014). Receptor molecules on the surface of the bacteria mainly determine the host-specificity of phages. Phage receptors have been described in different parts of bacteria (capsule, cell wall, flagella and pili). Phages attached to the receptors located in the cell wall are typically known as somatic phages. Most of the known phages infect a limited number of strains.

In their extracellular phase, a phage consists of a genome of either RNA or DNA surrounded by a protein coat named capsid. These particles are known as virions. Many phages also contain additional structures such as tails and spikes. Much less frequently, they contain lipids. In addition, they display a range of nucleic acid structures consisting of either double stranded (ds) or single stranded (ss) RNA or DNA, but never both. Phages may readily be grouped on the basis of a few gross characteristics including host range, morphology, nucleic acid, strategies of infection, morphogenesis, phylogeny, serology, sensitivity to physical and chemical agents, and dependence on properties of hosts and environment. The present classification adopted by the International Committee on Taxonomy of Viruses is mostly based in phage morphology and characteristics of the nucleic acid (Fauquet and Fargette, 2005). Phages of particular interest in water quality assessment are included into the seven families whose characteristics are summarized in Figure1.

Figure 1. Most common morphological types in somatic coliphages and F-specific phages. Bar 50 nm.

Owing to their simple structure and composition, virions persist quite successfully in the environment and are moderately resistant to natural and anthropogenic stressors (Grabow, 2001). It is likely that phages infecting bacteria indigenous to a given habitat are less persistent than their bacterial host (Ogunseitan et al., 1990), but also that phages survive better than host bacteria in habitats where they are aliens (Grabow, 2001). This will be the case of phages infecting enteric bacteria once outside the gut.

Typically, infectious phages are detected by their effects, mostly lysis, on the host bacteria that they infect. The most important factor in defining a method for the detection of a given phage or group of phages is the bacterial host strain. Easy methods to detect their presence or absence in a given volume of sample are available. Phages are counted by direct quantitative plaque assays (Adams, 1959) (Figure 2). The plaque assay provides the results in plaque forming units (PFU). A plaque forming unit is an entity, usually a single virion but also e.g. a clump of virions that originates a single plaque. PFU are also denominated for example in the ISO standards, plaque forming particles (pfp). The presence of phages in a given volume of sample can also be determined by the qualitative presence-absence enrichment test (Adams, 1959) (Figure 3). Enrichment of multiple tube serial dilutions allows estimating numbers of phages by “quantal” methods, as for example the most probable number procedure.

Figure 2. Bacteriophage enumeration by double agar layer method

Figure 3. Detection of bacteriophage by the presence/absence method

Phages are used as faecal indicators to mimic enteric viruses better than any other group of indicators, and show moderate resistance and persistence in the water environment and through wastewater treatment (Armon and Kott, 1996; Grabow, 2001; IAWPRC Study group on health related water microbiology, 1991; Jofre, 2007). This is one of the reasons why different groups of phages infecting enteric bacteria have been proposed as either faecal or viral indicators. Phages used as indicators for faecal pollution and microbial source tracking, indicators as surrogated for pathogens (index) and indicators of public health risk have been assembled according to the detection methods. Coliphage are those viruses that infect the bacteria E.coli and somatic coliphages (IAWPRC Study group on health related water microbiology, 1991; Kott et al., 1974), F-specific RNA phages (Havelaar et al., 1990; IAWPRC Study group on health related water microbiology, 1991), and phages infecting Bacteroides (IAWPRC Study group on health related water microbiology, 1991; Tartera and Jofre, 1987) are those extensively studied regarding their potential as indicators of various aspects of water quality control. Reviews with a wide coverage of this potential application of phages as indexes, indicators or markers exist (Armon and Kott, 1996; Grabow, 2001; IAWPRC Study group on health related water microbiology, 1991; Jofre, 2007).

Other that the fact that they mimic various aspects of virus behaviour, the appeal of phages as indicators lies in the availability of feasible, fast and cost effective detection methods, and their abundance in wastewaters of human and animal origin. Moreover, samples can be kept at 4ºC for at least 48 hour without any significant change in the numbers of infectious phages(Mendez et al., 2002); and small volume samples can be kept for months at -20ºC or -80ºC after the addition of 10% v/v glycerol (Mooijman et al., 2005). Reference suspensions of phages needed for quality assurance are easily prepared and conserved (Mendez et al., 2002; Mooijman et al., 2005). Finally, phenomena such as stress, injury or reactivation that frequently lead to misinterpretation of environmental data on bacterial indicators are not applicable to phages.

1.1 Somatic Coliphages

The term somatic coliphages defines the phages infecting Escherichia coli through the cell wall. Most known somatic coliphages found in municipal wastewater belong to the Myoviridae, Siphoviridae, Podoviridae and Microviridae families (Muniesa et al., 1999; Rajala-Mustonen and Heinonen-Tanski, 1994). Somatic coliphages most frequently used as model organisms areϕX174, PDR1, T2 and T7.

Somatic coliphages attach to the bacterial cell wall and may lyse the host cell in 25-30 min producing between 100 and 1000 phages per infected cell depending on both the phage and the physiological state of the cell. They produce plaques (Figure 2) of widely different size and morphology.

Host strains of somatic coliphages include Escherichia coli and related species such asShigellaspp and Klebsiella spp. Some of these may occur in pristine waters, so exceptionally somatic coliphages can find hosts in these environments. Several sorts of studies have shown many factors limit somatic coliphage replication in water environments. 1) The rather narrow host range of phages infecting the host strains used in the standardized methods (Muniesa et al., 2003). 2) The high densities of host bacteria and phages needed to ensure phage replication (Muniesa and Jofre, 2004; Wiggins and Alexander, 1985). 3) The presence in water of numerous background bacterial flora and particulate material that interferes with coliphage replication (Muniesa and Jofre, 2004; Wiggins and Alexander, 1985). 4) The metabolic activity of the host needed for phage replication is too low in the water environment (Cornax et al., 1991). 5) The envelope stress response that affects E. coli when released into the environment that could trigger responses diminishing phage infection (Perez-Rodriguez et al., 2011; Raivio, 2011). Moreover, the contribution of lysogenic induction to the presence of free coliphages is not noticeable. Hernandez-Delgado and Toranzos (Hernández-Delgado and Toranzos, 1995) have shown that neither sewage isolates nor laboratory phage strains replicated in pristine river water in a tropical area. These and other data reviewed (Jofre, 2009) strongly suggest that the contribution of somatic coliphages replicated outside the gut to the numbers of somatic coliphages detected in water environments is negligible.

Different host strains of E. coli as well as different assay media count different numbers and types of somatic coliphages (Havelaar et al., 1990; Muniesa et al., 1999). Consequently, from now on in this chapter preferential attention is given to data obtained with the standardized methods indicated.

1.2 F-specific and F-specific RNA Bacteriophages

F-specific bacteriophages, also termed sexual coliphages or male-specific phages infect bacteria through the sex pili, which are coded by the F plasmid first detected in E. coli K12. F-specific RNA phages, a subgroup of F-specific phages consist of a simple capsid of cubic symmetry of 21-30 nm in diameter and contain a single-stranded RNA as the genome. The F-specific RNA phages group (Leviviridae) contains two genera (Levivirus and Allolevirus) and three minor unclassified groups (Fauquet and Fargette, 2005). Levivirus contains subgroups I and II, whereas Allolevirus contains subgroups III and IV. These four groups coincide with the serotypes first described by Furuse(Furuse, 1987). Posterior genomic characterization has allowed establishing that genogroups match with the serotypes (Beekwilder et al., 1996; Hsu et al., 1995), at least for practical purposes. Other F-specific phages are the rod-shaped DNA phages of the family Inoviridae (Fauquet and Fargette, 2005).

F-specific RNA phages used as model are MS2 and f2 belonging to genotype I, GA to genotype II, Qβ to genotype III and F1 to genotype IV. Sero and genogrouping of RNA F-specific phages has been used to differentiate subgroups in waters receiving faecal wastes, and the study of subgroups distribution may contribute to source tracking of faecal pollution. This is because groups I and IV are the predominant in waters contaminated with animal faecal residues whereas groups II and III are mostly associated with human pollution (Jofre et al., 2011). Long and collaborators have also studied whether typing of F-specific DNA phages could be useful for microbial source tracking (MST), but results seem to indicate that they are not (Long et al., 2005)

The F plasmid is transferable to a wide range of Gram-negative bacteria and so F-specific phages may have several hosts. Yet, it is thought that their main natural host is E. coli. The sexual pili are not synthesized below 32ºC. In spite of the temperature conditioned production of sexual pili, necessary for phage replication, conflicting reports on replication of F-specific phages in wastewater and groundwater exist (Havelaar and Pot-Hogeboom, 1988). Nevertheless, Woody and Cliver reviewed conditions affecting F-specific RNA phage replication other than pili formation in the host bacteria, and concluded that replication outside the gut cannot be excluded but also that it is very improbable (Woody and Cliver, 1997). Therefore, the influence of replication outside the gut in the numbers of these phages in the environment can be considered as negligible.

The infectious process of F-specific RNA phages is inhibited by the presence of RNase in the assay medium, which can be used to distinguish between the F-specific RNA phages and the other phages than can infect the host bacteria. These are the rod-shaped F-specific DNA phages of the family Inoviridae, which also infect the host cell through the sex pili as well as the somatic phages that can infect the bacterial host used.

1.3 Bacteriophages Infecting Bacteroides

Bacteriophages infecting strains of several species of Bacteroideshave been detected in faeces and wastewater contaminated with faecal wastes (Jofre et al., 2014). Replication of phages infecting Bacteroidesoutside of the gut of warm-blooded animals is even more unlikely than in the case of the coliphages referred to in previous sections. This is due to the additional stringent requirements of the host strain regarding anaerobiosis and nutrients that are unlikely to occur and coincide in natural water environments (Tartera and Jofre, 1987).

Phages infecting different Bacteroides species described so far are all tailed and the vast majority of them have the morphology corresponding to the family Siphoviridae (McLaughlin and Rose, 2006; Queralt et al., 2003; Tartera and Jofre, 1987). As well, the genome sequence of the few Bacteroidesinfecting phages studied corresponds to that of Siphoviridae (Hawkins et al., 2008; Ogilvie et al., 2013).

Phages infecting Bacteroides seem to infect the host through the cell wall and therefore are somatic phages(Puig et al., 2001) yet most phages infecting Bacteroideshave a quite narrow host range (Cooper et al., 1984; Kory and Booth, 1986; Tartera and Jofre, 1987).

In addition to the narrow host range of phages, strains ofBacteroides spp. differ in their ability to recover phages in faecal material, wastewaters and wastewater contaminated waters (Payán et al., 2005; Puig et al., 1999). Bacteroides strains differ also in their capability to detect phages in the faeces of different animal species, including humans, and hence in their ability to determine the origin of faecal contamination in a given sample. Some strains of B. fragilis, such as RYC2056 and VPI3625, detect phages both in human and non-human faecal wastes (Blanch et al., 2006; Kator and Rhodes, 1992; Puig et al., 1999). Others are able to discern the faecal source. Those detecting higher numbers of phages are reported in Table 1. However, this source specificity is not absolute and though seldom, host strains reported in Table 1 detect very low numbers of phages in the non-corresponding sources. The method proposed by Payán et al. (2005) allows isolating hosts appropriate for a given host and a given geographical area with reasonable success. The total cost associated with each attempt to isolate new strains did not exceed 1,000 euros including consumables and labour (approximately 1 month).

Strains RYC2056, GA17, GB124 and PG2166 were selected as those giving higher and more consistent counts in the matching wastewaters and are those that will be further discussed.

2.0 Detection and Concentration Methods

Generally phage are assayed without concentration directly with their respective host as plaque assay or in a presence absence format.

2.1 Somatic Coliphages

Methods for the detection and enumeration of somatic coliphages have been standardized by ISO 10705-2 (Anonymous, 2000), USEPA 1601 and 1602 (EPA, 2001a; 2001b)and Standard Methods (Rice, 2012), (Table 2).

Other than some details on media and assay conditions, the three methods use E. coli C as host strain. These are either E. coli ATCC 13706(Rice, 2012)or their nalidixic acid resistant clones E. coli CN13 (ATCC 700609) used in the USEPA 1601 and 1602 methods, or E. coli CN, more frequently referred as WG5, (ATCC 700078) used in the ISO 10705-2 method. Hosts resistant to nalidixic acid were introduced to minimize the growth of background bacteria that frequently interferes in the correct visualization of plaques in the plaque assay test. This allows quantifying phages in samples avoiding filtration of the sample to eliminate the background bacteria. Of course, filtration through 0.22 μm non-protein binding membrane filters such as those of polyvinylidene fluoride or polyether sulfone can also be applied to eliminate background bacteria present in the sample. In contrast, membrane filters of mixed cellulose and acetate membranes will adsorb and consequently retain phages. Strains ATCC13706, CN13 and WG5 detect similar numbers of phages in water matrixes when using the same method (Grabow et al., 1993; Guzmán et al., 2008). Other, non C E. coli host strains count lesser numbers of somatic coliphages (Grabow et al., 1993; Havelaar and Hogeboom, 1983; Stetler, 1984). One of these strains used frequently (Rose et al., 2004) is E. coli C3000 (ATCC 15597) that is an Hfr strain derived from K12, which is used to count simultaneously somatic coliphages and F-specific phages. Strain E. coli CB390 was tailored to detect the same numbers of somatic coliphages as WG5 and as many F-specific phages as host strains used in standardized procedures (Guzmán et al., 2008)and it detects similar number of F-specific phages but more somatic coliphages than C3000.

ISO-10705-2 (Anonymous, 2000) includes both the double agar layer (DAL) plaque assay method for the quantification of PFU (Figure 2) and the presence/absence test (Figure 3) than can also be adapted to a more probable number format. A simplified version of these methods can be found in Havelaar and Hogeboom (Havelaar and Hogeboom, 1983). Mooijman et al. (2005)proved that the implementation of the ISO standard method is feasible in routine microbiology laboratories without previous experience in phages.

USEPA Method 1601 (U.S. EPA, 2001a)deals with the presence absence method and EPA method 1602 (U.S. EPA, 2001b)with the single agar layer (SAL) plaque assay. A simplified version of these methods can be read in USEPA (U.S. EPA, 2001c). USEPA (U.S. EPA, 2003a, 2003b)has carried out interlaboratory validation tests of methods 1601 and 1602. A double agar layer (DAL) plaque assay method with strain C as the host is described in the 22nd edition of Standard Methods (Rice, 2012).

Results using these plaque assay methods can be obtained in 6 hours, though the habitual period of assay is 18 hours.

Detection and enumeration of somatic coliphages by the standardized methods is quite cost effective. The cost in material, media, reagents and labour is similar to the one for detection of faecal coliforms or E. coli. It can be done in routine microbiology laboratories. Additionally, the ISO procedure includes optional steps for laboratories with limited equipment.

Additionally to the host strains there are minor differences in the media between the standardized procedures. USEPA and ISO methods count similar numbers of somatic coliphages (Guzmán et al., 2008). As well, the present Standard Methods method is supposed to provide similar counts than ISO and USEPA. In contrast, the Standard Methods procedure described in previous editions has been reported to perform poorly (Green, 2000).

Regarding fast methods, no procedures based in serologic detection or PCR have been described that are applicable to the detection of the group somatic coliphages in waters. The diversity of phages included in this group makes this approach complex. In contrast, FAST PHAGE, which is applicable to the methods based on presence/absence of USEPA and ISO, has been described by Salter and Durbin (Salter and Durbin, 2012). This method is based on a rapid extracellular beta-galactosidase enzyme release and detection during the coliphage induced lysis.

Volumes of sample tested as described in the methods can be scaled up keeping the proportions of the mixtures of medium, host suspension and sample and using different size Petri dishes accordingly.

2.2 F-specific and F-specific RNA Bacteriophages

Host strains to detect sexual coliphages must produce sexual pili encoded by the F-plasmid. Hfr E. coli strains such as C3000 (ATCC 15597) were firstly used for this purpose, but these strains also detect high numbers of somatic coliphages. Later, strains Escherichia coli HS(E.coli Famp, ATCC 700891)(Debartolomeis and Cabelli, 1991)and Salmonella entericaserovar typhimurium (most frequently reported as Salmonella typhimurium) WG49 (NCTC 12484)(Havelaar and Hogeboom, 1984)were tailored to detect mainly F-specific phages, though both still detect a very small proportion of somatic phages infecting either E. coli or S. typhimurium. Strains HS and WG49 were selected as host strains in the standardized methods mentioned later on. The phages detected by these strains are counted as sexual or F-specific coliphages. The number of F-specific RNA phages is the difference between the numbers of phages counted in the absence and in the presence of RNase in the assay medium. More than 90-95 % of the phages detected in sewage by strains HS and WG49 are F-specific RNA phages (Debartolomeis and Cabelli, 1991; Havelaar and Hogeboom, 1984), but this percentage may be lower in receiving waters and treated wastewaters. Counts achieved by strains HS and WG49 are similar (Chung et al., 1996; Grabow et al., 1993). Because of the genetic markers (resistance to ampicillin and capacity to use lactose), the stability (and re-selection) of strain WG49 is easier to check than that of strain HS that has no explicit genetic markers.

Standardized methods for the detection and enumeration of both F-specific and F-specific RNA phages exist (Table 3). The ISO-10705-1 (Anonymous, 1995)standard method for the detection and enumeration of phages endorse Salmonella typhimurium WG49 as host strain and include a step with RNase in the assay medium. Consequently, it detects both F-specific and F-specific RNA phages. The ISO standard includes both the double layer plaque assay method and the presence-absence method. A simplified version can be found in Havelaar and Hogeboom (1984)and Standard Methods (Rice, 2012). Mooijman et al. (2005) proved that the implementation of the ISO standard method is feasible in routine microbiology laboratories without previous experience in phages.

EPA Method 1601 (U.S. EPA, 2001a)standard deals with the presence-absence method and EPA method 1602 (U.S. EPA, 2001b)with the single layer plaque assay (SAL) for the detection of F-specific phages. These methods use E.coli Famp as host strain. A simplified version can be found in USEPA (U.S. EPA, 2001c)and Standard Methods (Rice, 2012). USEPA (U.S. EPA, 2003a, 2003b)has carried out interlaboratory validation tests on methods 1601 and 1602.

Results using both plaque assay methods can be obtained in 18 hours.

The cost in material, media and reagents and labour for the detection of F-specific coliphages is similar to that for the detection of somatic coliphages. The cost for the detection of F-specific RNA phages is 10-15 % higher because of the need of RNase and the double number of plates. It can be done in routine microbiology laboratories.

Using any one of the methods, volumes tested as described in the basic standardized methods can be scaled up keeping the proportions of the mixtures of assay media and sample.

Molecular methods are available. However, they are for a specific phage or for a subgroup, not for the full group. The sum of genome copies (GC) obtained by RT-qPCR (Reverse transcription-quantitative polymerase chain reaction) of the different subgroups accounts for all F-specific RNA phages. These methods are referred to later on in the section on F-specific RNA phages for microbial source tracking section.

Different procedures for detecting F-RNA phage subgroups are available, some based on serological recognition and others based on nucleic acid sequence recognition (Table 3).

Among the first group of methods, there are the “neutralization” method first described (Furuse, 1987)and the culture latex agglutination and typing test (Love and Sobsey, 2007). The latex agglutination method is fast and can be applied in situ, though it may require pre-enrichment. However, nowadays accessibility of specific antisera is less practical than the availability of nucleic acid probes and primers.

Methods based either on plaque hybridization with specific probes and those based on RT-PCR seem more realistic. Nowadays RT-qPCR is commonly used.

Plaque hybridization is applied on plaques grown by either the ISO-10705-1 or the USEPA 1602 standards (Beekwilder et al., 1996; Hsu et al., 1995; Schaper and Jofre, 2000). This method implies the transfer of the plaques to four different N⁺ hybond membranes and posterior hybridization of each membrane with a probe specific for each one of the 4 subgroups. It has the advantage that many plaques can be studied at once.

Reverse transcription-polymerase chain reaction (RT-PCR) assays have been developed for MS2, the prototypical F-RNA phage of subgroup I (O'Connell et al., 2006), the different genogroups (Ogorzaly and Gantzer, 2006)and multiplex for all subgroups (Friedman et al., 2009; Kirs and Smith, 2007; Wolf et al., 2010). RT-PCR can be applied either to phages recovered from plaques obtained by the plaque assay method or directly (RT-qPCR) to determine the number of genome copies (GC) present in a given sample. The first approach allows apportioning the subgroups of the infectious phages, whereas the second does not provide insight in the infectiousness of the phages, but results can be obtained within a few hours.

The actual concentrations of phages belonging to each one of the subgroups in a given sample can be estimated by applying the percentages corresponding to each group to the concentration of F-specific RNA phages.

In contrast, direct RT-qPCR can provide directly the real numbers of GC. The numbers of GC usually exceed those of infectious phages determined by plaque assay. This difference is actually observed in faeces, and even varies from sample to sample (Hartard et al., 2015). In raw human and animal wastewaters, the GC numbers exceed the values of infectious phages by between 1.5 and 2.6 log₁₀units. However, this is not always the case, probably because of poor efficiency of the RT qPCR applied on certain types of samples or because of aggregation-disaggregation of phage particles. Thus, (Hata et al., 2013) detected more PFUs than GC in the influent of a wastewater treatment plant, but detected more GC in the effluent. Additionally, the difference in GC and PFUs will likely increase after inactivation in water environments and after water treatment, since it is well know that phage GC signals are more persistent in nature and more resistant to treatments than the plaque assay measuring infectious viruses.

2.3 Bacteriophages Infecting Bacteroides

The method used for the Bacteroides phages, namely ISO-10705-4 (Anonymous, 2001), includes both the double agar layer (DAL) plaque assay method for the quantification of PFU and the presence/absence test. A simplified version of these methods can be read in Araujo et al. (2001). These methods are similar to those described for coliphages, the only difference being that Bacteroides has to be grown under anaerobic conditions and that incubation times are longer. However, anaerobic jars and sachets are appropriate for cultures in Petri dishes and screw caped tubes completely filled with culture medium are enough for liquid cultures. Manipulation can be done on the open bench. The ISO methods are applicable to all the host strains referred to earlier. Mooijman et al. (2005) proved that the implementation of the ISO standard method is feasible in routine microbiology laboratories without previous experience with phages.

The cost in material, media, reagents and labour for the detection of phages infecting Bacteroides is similar to that for detection of somatic coliphages with an additional 10-15 % for anaerobic conditions. It can be done in routine microbiology laboratories. Additionally, the ISO procedure includes optional steps for laboratories with limited equipment.

2.4 Methods for Concentration

The presence/absence method allows testing of relatively large volumes (up to one litre) (Grabow, 2001), however concentration may be required either because greater volumes need to be tested or because quantification is required. Most methods described for concentrating animal viruses are not adequate for concentrating phages (Grabow, 2001). For volumes ranging from 10 to 1000 mL, two methods are recommended. For water with low turbidity, Sobsey et al. (1990) developed a simple, inexpensive and practical procedure for the recovery and detection of F-specific phages using mixed cellulose and acetate membrane filters with a diameter of 47 mm and a pore size of 0.45 μm after addition of salts and pH adjustment. This method was slightly modified by Mendez et al. (2004b) showing an excellent performance for up to 1 litre of sample for concentrating somatic coliphages, F-specific RNA phages and phages infecting Bacteroides. For samples with high turbidity, flocculation with magnesium hydroxide (Schulze and Lenk, 1983) is practicable for the three groups of phages(Contreras-Coll et al., 2002). Phage can also be concentrated by ultrafiltration as mentioned for other viruses.

3.0 Data on Occurrence in the Environment

3.1 Faeces

3.1.1 Somatic coliphages

Somatic coliphages have been isolated in variable percentages of human and animal stool samples (Table 4).

Reported percentages of positive human specimens range from 54 to 91%, whereas those of animal samples range from <1 to 100 %. The variability is such that some publications report the isolation of somatic coliphages in all animal species tested and in 100% of the samples tested (Havelaar et al., 1986) whereas other authors failed to isolate somatic coliphages from any sample of 50 % of species analysed (McMinn et al., 2014). As well, the concentrations reported are very variable and range from <1 to 7.3x10⁵/gram in humans and from <1 to 10⁸ in animals. Animal species studied (Table 4) include farm, domestic and wild species more likely contributing to faecal contamination of the environment.

These percentages and concentrations are very likely an underestimation. Such supposition is grounded on the facts that stool is a very complex matrix and no methods have been settled for defining the amounts of sample to be analysed or the need of applying extraction procedures. Regarding the amount of sample to be tested, if, as expected, coliphages replicate as lytic phages in the gut following the kill the winner model of phage bacteria population dynamics (Rodriguez-Valera et al., 2009), the number of phages will probably vary in form of waves of abundance in the gut content moving towards the cecum. This model involves a phage-bacteria population dynamics in which an increase in a host population (the winner) is followed by an increase in the rate at which that winner is killed by phages with the corresponding increase in the phage population. In fact, in measuring phages infecting Bacteroides, Tartera and Jofre (Tartera and Jofre, 1987) tested different stool specimens of the same individual and detected values ranging from <1 to >2.4x10⁸infectious phages/gram; very likely composite samples will provide a better view of the presence and concentrations of phages in the stool samples. Additionally, Jones and Johns (Jones and Johns, 2009) reported for F-specific phages that increasing the amount of sample increased the percentage of samples from which phages were isolated as well as extraction and polyethylene glycol precipitation improved the detection.

3.1.2 F-specific and F-specific RNA bacteriophages

F-specific RNA phages have been isolated in variable percentages of human and animal stool samples (Table5). The same consideration as for somatic coliphages regarding the great variability of results reported apply for F-specific RNA phages. Reported percentages of positive human specimens range from 6 to 73%, whereas those of animal samples of different species range from 0 to 100 %. Also, the concentrations reported are very variable and range from <1 to 1x10⁴PFU/gram in humans and from <1 to > 1.2x10⁶ in animals. Animal species studied (Table 5) include farm, domestic and wild species that most likely contribute to faecal contamination of the water environment.

The comments regarding the potential underestimation of the values made for somatic coliphages are applicable to F-specific RNA phages.

Several reports on the distribution of subgroups in human and animal faeces are available. All reports indicate that subgroups II and III are generally associated with human faecal material, whereas subgroups I and IV make up the majority in animal faecal wastes. This distribution is observed whatever the method used. Nevertheless, this association is not always 100 % exact. Thus, Havelaar et al. (1990) have detected subgroup II in pig faeces; Schaper et al (Schaper et al., 2002b) subgroup II in faeces of pigs, cattle and poultry; Hsu et al. (1995) subgroup II in pig faeces; Hartard et al. (2015) subgroup II in faeces of ducks and geese; and Cole and Sobsey (2003) subgroups II and III in faeces of cows.

3.1.3 Bacteriophages infecting Bacteroides

Phages infecting different strains of Bacteroides have been isolated in variable percentages of human and animal stool samples (Table 6). In this case, the host strain introduces the greater factor of variability since as indicated earlier Bacteroidesstrains differ in their capability to detect phages in different faecal sources. Bacteroides fragilis RYC2056, has been reported to recover phages from 28% of human stool specimens, although it also recovers phages from animal faeces (Puig et al., 1999), with the maximum incidence, 30%, in pigs. Bacteroides fragilis HSP40 phages have been isolated from 10-13 % of human stool samples, but never from animal faeces, with values ranging from >1 to 1.2x10⁴PFU/gram (Gantzer et al., 2002; Grabow et al., 1995; Tartera and Jofre, 1987). GB-124 has been reported in 4% of human samples and never in animal samples (Diston and Wicki, 2015; McMinn et al., 2014).

3.2 Raw Wastewater

3.2.1 Somatic coliphages

Somatic coliphages are the most abundant indicator phages in raw municipal and hospital wastewater. Their concentrations are usually less than one order of magnitude lower than those of faecal coliforms (or E. coli) wherever they have been counted (Table 7). Bacterial indicators and somatic coliphage concentrations in raw municipal wastewaters are comparable regardless of the geographical location and the income level of the country.

In Argentina, Lucena et al. (2003) studied samples of twenty-eight septic tanks and reported the detection of somatic coliphages in all samples tested with values ranging from 10⁶ to 10⁷ PFU/100 mL, which are similar to those reported for sewage in the area.

The reported values of somatic coliphages range from 10⁴-10⁵PFU/100 mL in manures to 4x10⁸ in abattoir wastewaters (Table 8). The proportion of these counts with counts of faecal coliforms and/or E. coli are similar to those found in municipal wastewater.

Importantly enough, at source of pollution, somatic coliphages do not show seasonal differences (Muniesa et al., 2012; Zhang and Farahbakhsh, 2007).

3.2.2 F-specific and F-specific RNA bacteriophages

F-specific phages and F-specific RNA phages are the second most abundant indicator phages in raw municipal and hospital wastewater with most frequently reported values ranging from 10⁴ to 10⁶PFU/100 mL. These values are usually between one and two orders of magnitude lower than the numbers of faecal coliforms (or E. coli) and somatic coliphages (Table 9). The numbers of F-specific and F-specific RNA phages in wastewaters are comparable wherever they have been determined regardless of the geographical location and the income (level of development) of the country (region/area). Numbers of F-specific phages in sewage do not show seasonal differences (Haramoto et al., 2015; Zhang and Farahbakhsh, 2007).

Lucena et al. (2003) enumerated F-specific RNA phages by the ISO methods of twenty-eight septic tank samples in Argentina and reported the detection of F-specific RNA phages in all samples tested with values ranging from 10⁵ to 10⁶ PFU/100 mL, which are similar to those reported for sewage in the area. Calci et al. (1998) using the USEPA methods detected F-specific phages in 10 of 17 samples of septic tank samples with values ranging from <10 to 10⁶ PFU/100 mL.

Also, F-specific and F-specific RNA phages are detected in substantial concentrations in abattoir wastewater and animal slurries and manures. The reported values range from 10⁴PFU/100 mL in animal waste slurries to 2x10⁸ in abattoir wastewaters (Table 10).

The subgroup percentages detected in raw wastewaters are reported in Table 11. As in the case of faeces, this distribution is not absolute. Most of the data reported in table 11 refer to phages enumerated by either the USEPA or the ISO methods and subgrouping of the phages in plaques. Wolf et al. (2010) and Hata et al. (2013) detected GC by multiplex RT-q-PCR and RT-q-PCR.

3.2.3 Bacteriophages infecting Bacteroides

In this case, the numbers detected depend on the host strains used since these differ in detection regarding the source of faecal contamination and there are also some geographical differences.

Concentration of phages detected by the non-discerning strain RYC2056 are relatively consistent everywhere around the world both in human and animal wastewaters with most frequent values ranging from 10⁴ to 10⁵ PFUs/100 mL. (Table 12).

Lucena et al. (2003) detected strain RYC2056 phages in 77 % of samples from twenty-eight septic tank samples in Argentina with counts ranging from 10² to 10³PFU/100 mL. The lower numbers of Bacteroidesstrain RYC2056 phages compared to coliphages in the source of contamination, and the anaerobic conditions required for their cultivation, tend to discourage the use of RYC2056 phages as general indicator, at least in temperate climates, despite attractive features.

Table 13 shows counts of phages in wastewater of human origin obtained by using Bacteroides host strain GA and GB124 which yielded the highest counts in assays on hosts specific for human Bacteroides phages.

Detection of phages by strains GA17, GB124 show a certain geographic discontinuity, which was previously appreciated for strain HSP40 (Kator and Rhodes, 1992; McLaughlin and Rose, 2006; Sirikanchana et al., 2014). In Southern Europe, Tunissia and Colombia, most phage concentrations detected by GA17 ranged between 5x10⁴ and 5x10⁵ PFUs/100 mL. McMinn et al. (2014) have reported values of phages detected by strain GB214 in primary sewage effluents studied in several states of USA with average values in some states well below 10³ to some states with average values nearby 10⁴/100 mL.

As it happens with somatic and F-specific coliphages, the presence of phages infecting strains RYC2056 and GA17 in sewage shows no seasonality (Muniesa et al., 2012).

3.3 Treated Wastewater

Detailed information on phages as treatment indicators as well as on removal during treatment will be given in Using indicators to assess microbial treatment and disinfection efficacywithin this section and in chapters within the technology section. Here, data on phages found in different kinds of wastewater treatment effluents categorized as secondary are provided.

3.3.1 Somatic coliphages

Primary sedimentation, up flow anaerobic sludge blanket processes, flocculation-aided sedimentation, activated sludge digestion, activated sludge digestion plus precipitation and trickling filters remove bacterial indicators and somatic coliphages in numbers ranging from 0.3 to 3.0 log₁₀. Differences in the elimination of faecal indicator bacteria and somatic coliphages are minor and vary slightly from treatment to treatment. Therefore, the values of somatic coliphages in secondary effluents range mostly from 10³ to 10⁵PFU/100 mL (Table 14).

Another point source is the effluents of waste stabilization ponds (or oxidation ponds, or lagoons) of very varied configuration and extent of treatment. Therefore, the effluents released after lagooning are quite diverse regarding the concentration of indicators (Alcalde et al., 2003; Campos et al., 2002; Gomila et al., 2008; Lucena et al., 2004). Moreover, counts of indicators in the effluents can vary between seasons in climates with marked seasonal variation.

DeBorde et al. (1998) studied the septic effluents of a high school in the USA and detected somatic coliphages in all 43 samples tested with values between 10⁴ and 10⁵PFU/100 mL, though in this case the host strains used were not those recommended in the standardized methods.

3.3.2 F-specific and F-specific RNA bacteriophages

Primary sedimentation, up flow anaerobic sludge blanket processes, flocculation-aided sedimentation, activated sludge digestion, activated sludge digestion plus precipitation and trickling filters remove bacterial indicators, somatic coliphages, F-specific and F-specific RNA phages in numbers ranging from 0.3 to 3.0 log₁₀. The reported values of F-specific and F-specific RNA phages in secondary effluents range mostly from 10² to 10⁴PFU/100 mL (Table 15). The relative proportions of the F-specific and F-specific RNA phages with other indicators remain similar to those in untreated wastewater.

As in the case of somatic coliphages, the concentrations of F-specific phages detected in effluents released by lagooning are quite variable. In climates with marked seasonal variation the numbers of pathogens and indicators in the effluents may vary between summer and winter in the same treatment facility (Alcalde et al., 2003; Gomila et al., 2008; Hill and Sobsey, 1998; Lucena et al., 2004).

DeBorde et al. (1998) studied the septic effluents of a high school (350 persons) in the USA and detected F-specific phages in all 43 samples tested with values ranging from 10⁴ to 10⁵/100 mL, though this number might be an overestimation since the host strain used in this study, C3000, detects also somatic coliphages.

In secondary effluents of municipal activated sludge plants, there is a very significant increase in the percentage of subgroup I and the practical disappearance of subgroup IV(Table 11). This has been observed after either subgrouping of plaques or by direct determination of GC copies.

3.3.3 Bacteriophages infecting Bacteroides

Numbers of Bacteroides phages in secondary effluents range mostly from 5x10 to 10⁴PFU/100 mL(Table 16). The relative proportions of the Bacteroides phages with other indicators remain similar to those in untreated wastewater.

As in the case of somatic coliphages, the numbers of Bacteroides phages in lagoon effluents are quite variable. In climates with marked seasonal variation the quality of the effluents depends on the season (Campos et al., 2002; Gomila et al., 2008; Lucena et al., 2004).

3.4 Untreated Sludge

3.4.1 Somatic coliphages

The numbers of reports about concentrations of somatic coliphages in untreated sludge (primary sludge, raw sludge (mix of primary and activated), activated sludge and thickened sludge) are relatively scarce as compared to reports on phages in treated sludge. Most of the reported results are summarized inTable 17. Somatic coliphages were detected in all samples tested. Though concentrations in the different reports are difficult to compare since the concentrations reported in the scientific literature are referred to different measures such as dry mass, wet mass and volumes, and since different extraction procedures have been used, their concentrations are quite high in all reports. The highest concentrations are reported in primary sludge and the lowest in activated sludge. Numbers above 10⁷PFU/gram of dry weight have been reported for primary and raw sludge. Their counts in sludge remain in proportion to those of bacterial faecal indicators and other phages in raw wastewater.

3.4.2 F-specific and F-specific RNA bacteriophages

The same general comments done for somatic coliphages in untreated sludge mentioned earlier apply for F-specific RNA phages. Most of the reported results regarding untreated sludge are summarized inTable 18. F-specific RNA phages were detected in all samples tested. As in raw and treated wastewater they rank second in abundance among the indicator phages. The highest numbers were reported in primary sludge and the lowest in activated sludge. Concentrations well over 10⁶PFU/gram of dry weight have been reported for primary and raw sludge. Their concentration in sludge and proportion to faecal bacteria and other phages remain similar to that in raw wastewater.

Phages in plaques obtained from municipal untreated sludge gave the following distribution: 56% of serogroup I, 38% of group II and 6% had a mixed population. This pattern resembles that of secondary effluents more closely than that of raw wastewater (Nappier et al., 2006).

3.4.3 Bacteriophages infecting Bacteroides

The same general comments made for somatic coliphages in non-treated sludges apply to Bacteroides phages. Most of the reported results regarding non-treated sludge are summarized inTable 19. As in raw and treated wastewater they rank third in abundance among the indicator phages. Bacteroides phages were detected in all samples tested when the matching (faecal source and geographic area) host strain was used. Highest counts were reported in primary sludge and the lowest in activated sludge. Concentrations well over 10⁴PFU/gram of dry weight have been reported for primary and raw sludge. Their numbers in sludge and proportion to faecal bacterial and other phages remained similar to that in raw wastewater.

4.0 Persistence

Persistence of pathogens and indicators in water environments depends on both external factors related to water condition and in the microbe itself. Each microbe behaves differently. In the case of viruses and phages even simple mutations can change the resistance to treatment and persistence in the environment (Wigginton and Kohn, 2012). Because of this, model experiments with viruses and phages adapted to grow in laboratory culture should be viewed cautiously.

Most available data on phage persistence determined in model experiments refer to laboratory grown model phages and will be reviewed in more detail in the persistence section.

4.1 General Indicator Bacteriophages

Coliphages are present in raw sewage in numbers high enough to carry out model inactivation experiments after diluting sewage in fresh, brackish or marine water. In contrast, concentrations of phages infecting Bacteroides in raw sewage are not high enough to perform in situ experiments like hose reported for coliphages and consequently phages grown in the laboratory have to be inoculated into the mixture. The short discussion in this chapter is predominantly on coliphages naturally occurring in sewage and laboratory grown phages infecting Bacteroides and correspond to experiments in which the persistence of the three groups of phages was studied at once. The results show that indicator phages persist in water environments for longer than E. coli does (Table 20).

Indirect information regarding comparative persistence is also provided by the changes of the ratios between the numbers of conventional bacterial indicators and phages in water environments with aged pollution. Different studies report a decrease of the ratio in both river and marine waters when compared to the ratios of sewage (Contreras-Coll et al., 2002; Ibarluzea et al., 2007; Lucena et al., 2003; Mocé-Llivina et al., 2005; Skraber et al., 2002). This diminution confirms the observation of the model experiments.

As a whole, it can be concluded that persistence of the three groups of indicator phages is intermediate and as reviewed by Verbyla and Mihelcic (Verbyla and Mihelcic, 2014) their persistence is in the range of that of human viruses.

We think that it is worth mentioning that the great majority of data regarding the presence and persistence of F-specific RNA phages as a group in water have been obtained in cold and temperate climates. But, evidences exist indicating that in environments with temperatures higher than 25ºC F-specific RNA phages persist shorter than the other phages and bacterial indicators (Agulló-Barceló et al., 2013; Alcalde et al., 2003; Cole et al., 2003; Durán et al., 2002; Guzmán et al., 2007; Mandilara et al., 2006; Mocé-Llivina et al., 2005). Geographic areas with surface sea and fresh water temperatures over 25ºC either all year round (between 40º latitude North and 40º latitude South) or during the warm seasons (many areas of Europe and USA) are quite extensive. Then, the persistence of F-specific and F-specific RNA phages needs some extra verification before rating their persistence and recommending their use in warm regions.

4.2 Host Specific Indicator Bacteriophages

Many experiments regarding persistence of F-specific RNA subgroups in the water environment indicate a significant difference in the persistence among the subgroups. Indeed, persistence studies with phage isolates belonging to the different subgroups (Cole et al., 2003; Schaper et al., 2002a) and with naturally occurring phages (Hartard et al., 2015; Muniesa et al., 2009) indicate differential persistence between subgroups. In all cases subgroup I being the more persistent followed by subgroups II, III and IV in this order. These diverse survival rates of subgroups of F-specific RNA phages to environmental stressors constitute a limitation to their suitability as trackers of the origin of faecal pollution. This fact is aggravated by a similar behaviour in advanced wastewater treatments. All together these features lead to a marked preponderance of subgroup I in surface waters (Brion et al., 2002; Cole et al., 2003; Stewart-Pullaro et al., 2006).

All the results obtained in the in situ experiments point out that phages infecting different host strains of Bacteroides, including those that allow tracking faecal sources, are more persistent than E. coli and are in the range of persistence of somatic coliphages. Regarding the ratios between numbers of different indicators, Muniesa et al. (2012) reported that the numerical ratios between somatic coliphages/phages infecting Bacteroides strains in wastewaters are maintained in river and marine water. This fact points out that the persistence of phages infecting Bacteroides is similar to that of somatic coliphages. Having this in mind, Muniesa et al. (2012) have suggested that the numerical ratios between the concentrations of somatic coliphages and other phages detected by strains able to identify host associated faecal contaminant are a good marker for microbial source tracking.

5.0 Short Review on the Application of Phage as Indicators

5.1 General Quality Indicators

Coliphages have been used in academia for many years as both faecal and viral indicators. The high concentrations found in raw wastewaters and in many other matrixes contaminated with faecal remains, the easy, fast and cost effective methods, the persistence in the water environment and the resistance to treatments (see chapters dealing with sanitation, disinfection and persistence), which resemble those of viruses, make the indicator phages good surrogate indicator candidates for various set-ups.

Additionally, regulatory authorities in different areas of the world are beginning to consider phages as indicators of water quality and for validating and verifying water treatment processes.

5.1.1 Quality indicators of recreational waters

A substantial amount of data about the presence and levels of indicator phages, mostly coliphages, in all type of surface waters and in very different climatic areas is available (Burbano-Rosero et al., 2011; Contreras-Coll et al., 2002; Ebdon et al., 2007; Haramoto et al., 2005; Ibarluzea et al., 2007; Jiang et al., 2007; Lucena et al., 2003; Rezaeinejad et al., 2014; Taylor et al., 2001). The general trend is that the three groups of phages keep the same proportions as in raw wastewater, but they have reduced the concentration gap with either E. coli or faecal coliform bacteria. This shift in concentrations is quite prominent in samples with values of either E. coli or faecal coliforms below 10²-10³ CFU/100 mL (Contreras-Coll et al., 2002; Ibarluzea et al., 2007; Lucena et al., 2003; Skraber et al., 2002). This is probably due to the greater persistence of phages. The potential application of coliphages in quality control of bathing and recreational surface waters is currently being studied by USA regulatory authorities (U.S. EPA, 2015), though at present, neither concentrations nor which coliphages should be considered have been defined.

5.1.2 Indicators for the efficiency of reclamation processes and the quality of reclaimed water

Many of the tertiary treatment processes applied for water reclamation have as the major aim the removal and inactivation of pathogens. The most frequently used processes include filtration and disinfection by chemicals or UV radiation or a combination. The tertiary effluents generally contain substantially reduced numbers of pathogens and indicators, but the proportions between different indicators found in raw wastewater and secondary effluents changed after treatment.

A substantial amount of information is available on the removal of somatic coliphages and F-specific coliphages by tertiary treatment processes and on the numbers of surviving phages in treated effluents (Costán-Longares et al., 2008; den Blanken, 1985; Gomila et al., 2008; Luther and Fujioka, 2004; Mandilara et al., 2006; Montemayor et al., 2008; Nieuwstad et al., 1988; Rose et al., 2004; Rubiano et al., 2012; Soriano et al., 2011; Stiegel et al., 2013; Zhang and Farahbakhsh, 2007).

Removal of both groups of phages varies for different treatment processes and counts of surviving phages may range between 0.5 and >4.0 log₁₀units. Typically counts of somatic and F-specific coliphages in many tertiary effluents (reclaimed waters) are higher than those of faecal indicator bacteria, or at least, the difference is much smaller than in secondary effluents. According to the only available report on Bacteroides phages, their behaviour is similar to that of coliphages.

Because of these kinds of results indicating that the resistance of phages to many treatment processes (see chapters on sanitation and disinfection)resembles that of viruses, coliphages have been included in some regulations regarding water reclamation. Among these are the states of Queensland, Australia (Queensland Government, 2005) and North Carolina, USA (North Carolina Administration, 2011). Both regulations specify a given reduction in numbers of coliphages by the treatment processes concerned as well as the maximum number of coliphages permitted in the reclaimed water.

5.1.3 Quality indicators of groundwater

The presence of indicator phages in groundwater has very frequently been reported as presence in 1 litre after concentration and in most studies the results are reported as presence/absence (Abbaszadegan et al., 2003; Borchardt et al., 2004; Jung et al., 2011; Locas et al., 2007; Lucena et al., 2006).

In summary, somatic coliphages were detected in 8.7% (Payment and Locas, 2011) to 22.4% (Lucena et al., 2006) of groundwater samples, and F-specific phages in 1% (Borchardt et al., 2004) to 18.3% of samples (Lucena et al., 2006).The only study including phages of Bacteroides reports 1.7% (Lucena et al., 2006). Thus again somatic coliphages are the most abundant ones, followed by F-specific coliphages and Bacteroides phages. In two of these studies, partial data on aquifers with no faecal contamination and aquifers heavily contaminated are reported, and the values of positive isolations of phages clearly match with the expected contamination. Thus, for somatic coliphages values range from 0% to 62 % (Lucena et al., 2006) and from 0% to 25 % (Locas et al., 2007).

In an analysis of 100 mL groundwater samples, Lucena et al. (2006) reported the presence of E. coli in 18.8%, somatic coliphages in 22.4% and F-specific RNA phages in 18.1% of samples, and in a similar analysis Lee et al. (2011)reported the same indicators to be present in 15%, 12% and 7.5% of samples.

The USEPA rule on groundwater considers coliphages as alternative indicators for quality assessment of groundwater (U.S. EPA, 2006)

5.1.4 Quality indicators of drinking water

This refers to disinfected water only. Since indicator phages are more resistant to treatment processes than bacterial indicators (except spores of Clostridium) it can be expected that coliphages and Bacteroides phages are detected in samples of drinking water treated to eliminate the bacterial faecal indicators. This was confirmed by the detection of coliphages in a meaningful percentage of drinking water samples with free residual chlorine and no, or very low, numbers of bacterial indicators in Canada, Egypt and Peru(El-Abagy et al., 1988; Palmateer et al., 1990; Ratto et al., 1989). Méndez et al. (2004a)detected neither E coli nor faecal coliforms in 427 samples of 100 mL of a metropolitan water network containing >0.1 mg/L of chlorine, but detected somatic coliphages, F-specific RNA phages and Bacteroides phages in 8.8, 10 and 4.5 % of 1000 mL samples, respectively. Likewise, Armon et al. (1997) detected faecal coliforms, somatic coliphages, F-specific coliphages and Bacteroides phages in 0.7, 5.6, 7.1 and 5 %, respectively, of 1536 100 mL-samples of disinfected water in Israel. Therefore, coliphages arise as attractive indicators for disinfected drinking water that will provide some more protection than the traditional bacterial indicators. We are not aware of any drinking water quality regulation that considers phages as indicators.

5.1.5 Quality indicators for treated sludge

In many countries, regulations require the treatment of sludge for certain destinations as for example in agriculture. As seen in a previous section, coliphages are quite abundant in untreated sludge. The most frequent treatments of sludge are thermophilic digestions, pasteurization, lime stabilization and composting. Anaerobic mesophilic digestion shows a poor hygienization of all the indicators including phages, with values of somatic coliphages as high as 4.4x10⁶ PFUs/gram (dry weight) of digested sludge (Guzmán et al., 2007). Thermophilic digestions, pasteurization and composting achieving very significant reductions in the number of bacterial indicators showed in-between reductions of F-specific RNA and very moderate reduction in the numbers of somatic coliphages (Astals et al., 2012; Guzmán et al., 2007). Also, storage in quick lime is much more efficient in reducing the numbers of faecal coliforms than those of somatic coliphages (Campos et al., 2002).

Therefore, phage indicators, particularly somatic coliphages, seem to be a serious candidate to be considered as an indicator of hygienization in sludge management. Western Australia (Western Australian Government, 2012) and Colombia (Republica de Colombia, 2014) have introduced coliphages in the regulations about the quality of biosolids (treated sludge) to be applied in agriculture.

5.2 Other Potential Uses

In order not to unreasonably enlarge this section it is only mentioned that among other applications related to water quality, phages are also considered as potential indicators for the microbiological quality of shellfish, as well as markers for the transport of microorganisms in soil, and to validate filters.

5.2.1 Indicators for microbial source tracking

Among the subgroups of F-specific RNA phages, subgroup I seems to predominate in surface waters. This is difficult to interpret because subgroup I also appear to be the most persistent. In contrast, the few existing data (Ebdon et al., 2007; Muniesa et al., 2012)seem to indicate that phages infecting Bacteroides and their ratio to somatic coliphages seem a good option for MST of surface waters where E. coli concentrations are equal to or greater than 10³ CFUs/100 mL.

Lee et al. (2011) found that 85% of 60 F-RNA phages isolated from groundwater being subgroup I. No sound data about phages infecting strains GA17 and GB124 in groundwater are available in order to evaluate their potential use for MST in groundwater.

Information on the presence and values of coliphages in solid matrixes is relatively abundant, mostly in untreated and treated sludge, in contrast it is scarce about phages infecting Bacteroides and non-existing for F-specific RNA phage subgroups. The few data available seem to indicate that the ratio between the numbers of somatic coliphages and phages infecting Bacteroides will remain in these matrixes similar to the ratio in sewage (Guzmán et al., 2007; Lasobras et al., 1999; Yahya et al., 2015).

5.2.2 Target enumeration to infer the occurrence of intestinal pathogens

Additionally to their worth as indicators of faecal contamination the indicator phages have been viewed as potential surrogates of human viruses. Consequently, the possible relationship between presence and levels of indicator phages and human viruses in waters has been studied, though with disparate results. However, these studies are quite varied with respect to the assessed parameters and the methods used. Phages have usually been evaluated by plaque assay by standardized (or equivalent methods) in small volumes, and without a concentration procedure. Human viruses have been evaluated after concentration of great volumes and posterior detection with very different methods, such as plaque assay on cell culture and different PCR approaches. Most of these counts, mostly in matrixes where the numbers of phages and viruses are low and mostly those of viruses have a great amount of uncertainty that can explain part of the heterogeneity of the results.

Some studies have failed to show any relationship in fresh surface waters (Hot et al., 2003; Jiang et al., 2007; Viau et al., 2011; Westrell et al., 2006), marine surface water (Boehm et al., 2009; Jiang et al., 2007) and groundwater (Abbaszadegan et al., 2003; Borchardt et al., 2004; Payment and Locas, 2011).

Others have found relationships between human viruses and phage. Jiang et al. (2001) found that the presence of human adenovirus detected by nested-PCR was significantly associated with F-specific coliphages in marine surface waters impacted by urban run-off.

Ballester et al. (2005) described that the presence of astroviruses and adenoviruses was significantly associated with the presence of both somatic and F-specific coliphages in coastal water impacted by WWTP. The presence of rotavirus and enterovirus was only linked to the presence of F-specific phages. Human viruses were detected by ICC-nPCR and ICC-RT-nPCR.

Love et al. (2014) described that the presence of F-specific coliphages was associated with the presence of genome copies of adenoviruses but not noroviruses in coastal waters.

Rezaeinejad et al (2014) found that in urbanized catchments of fresh water in tropical Singapore F-specific PFU numbers were associated to counts of genome copies of noroviruses but not with GC of adenoviruses, astroviruses and rotaviruses.

Mocé-Llivina et al. (2005) described a significant association between the presence of enteroviruses enumerated on cell culture and somatic coliphages in marine waters.

Skraber et al. (2004) revealed that in a French river the numbers of samples positive for infectious enterovirus as well as genome copies of enteroviruses and of noroviruses increased with increasing densities of somatic coliphages.

Even if results are far away from indicating a clear correlation between densities of indicator coliphages and human viruses in waters, there is evidence that somatic and F-specific coliphages are more strongly associated with pathogenic viruses that the traditional bacterial indicators.

5.2.3 Relation to risk

Several epidemiological studies conducted to evaluate the relationships between the presence of indicator phages in surface waters and swimming illnesses have been performed with disparate outcomes. However, overall the epidemiological evidence suggests a likely relationship between coliphages and human health.

Von Schirnding et al. (1992) conducted a prospective cohort study at 2 South African marine beaches.The numbers of coliphages detected were fairly low and no statistically significant relation was found between sickness and numbers of somatic coliphages.

In studies at freshwater lakes in Germany, Wiedenmann et al. (2006) found a significantly increased risk of gastroenteritis for persons bathing in water with concentrations of somatic coliphages greater than 10 PFU/100 mL versus non-bathing fellows.

In an epidemiological study to evaluate water quality and health effects for waters at a marine beach in Florida (Abdelzaher et al., 2011), the detection of somatic coliphages overlapped with the highest illness days. However, no significant correlation between health outcomes and somatic coliphages was observed.

Lee et al. (1997) found in Great Britain a statistically significant association between risk of gastrointestinal illness and density of F-specific RNA coliphages.

Van Asperen et al. (1998) did not find exposure-response association between F-specific RNA phages and gastrointestinal illness in a study performed in Holland.

In a study on swimmers at marine beaches in California,Colford et al. (2007)found skin rash and diarrhoea, but no other health symptoms, to correlate with the presence of F-specific phages but not with that of somatic coliphages.

In a comparison of swimmers to non-swimmers at marine beaches in Alabama, Mississippi, and Rhode Island, a significantly higher risk of gastrointestinal illness has been reported for days when coliphages were present (Wade et al., 2010). No epidemiological studies seem to be on record regarding a correlation of phages in drinking water to related disease.

An overlapping between a jaundice outbreak and a high incidence of somatic coliphages in potable water occurred in a municipality of West Bengala, India in 2014. Somatic coliphages ranging from 40 to 250 PFU/100 mL were detected in 16 out of 20 samples during the outbreak detected in December 2013, whereas E. coli was detected in only 2 of the samples. A few weeks after the outbreak, somatic coliphages were only detected in four out of twenty samples with positive values ranging from 10 to 50 PFU/100 mL, and E. coli in none (Mookerjee et al., 2014).