AN IMPROVED VIRAL INDICATOR OF FECAL CONTAMINATION AND TREATMENT PROCESS EFFICIENCY TO MEET NEW EPA DRINKING WATER REGULATIONS

Transcription

1 AN IMPROVED VIRAL INDICATOR OF FECAL CONTAMINATION AND TREATMENT PROCESS EFFICIENCY TO MEET NEW EPA DRINKING WATER REGULATIONS Mark D. Sobsey and Thomas R. Handzel University of North Carolina at Chapel Hill Department of Environmental Sciences and Engineering CB #7400, Rosenau Hall, Chapel Hill, NC The research on which this report is based was financed in part by the United States Department of the Interior, Geological Survey, through the N.C. Water Resources Research Institute. Contents of this publication do not necessarily reflect the views and policies of the United States Department of the Interior nor does mention of trade names or commercial products constitute their endorsement or recommendation for use by the United States government. Agreement No G USGS Project No. 07 (FY 90) WRRl Project No

2 One hundred fifty copies of this report were printed at a cost of $1, or $7.61 per copy.

3 ACKNOWLEDGEMENTS We thank the many North Carolina water treatment plants and the Virginia water treatment plant that were kind enough to allow us to obtain water samples and to provide water quality data in support of this study. Their cooperation and assistance was invaluable and 'this study could not have been done without their support. We also thank the staff of the North Carolina Division of Environmental Management for their assistance in obtaining information about North Carolina water treatment plants and wastewater dischargers. We acknowledge Hyenmi Chung and Greg Lovelace for collection and analysis of wastewater samples for bacteria and phages. iii

4

5 ABSTRACT This study field-tested a new membrane filter method (MF) for the enumeration of F-specific RNA coliphages in source waters and treated drinking waters in an effort to assess the potential use of these organisms as a viral indicator. Simultaneous testing for F-specific coliphages by the Single Agar Layer (SAL) method as well as for E. coli and enterococci permitted comparisons between the two phage detection methods as well as the evaluation of F-specific coliphages as an indicator of fecal pollution. Mean concentrations of coliphages enumerated by the SAL and MF methods were comparable and correlation analysis resulted in a Pearson's correlation coefficient of F- specific phages were reasonably well correlated with both bacterial indicators as well as with the presence or absence of upstream, permitted wastewater discharges, suggesting the usefulness of these coliphages in assessing fecal contamination of natural waters. The reduction of F-specific coliphages by water treatment processes was observed at four drinking water treatment facilities. No phages were detected in any of the treated, or partially treated samples. Coliphage reductions of up to 5 logs were observed with an average reduction of >3.3 log,,. Using the S. typhimurium WG49 host, source waters were found to contain somatic Salmonella phages as well as F-specific coliphages. The presence of these Salmonella phages interfered with determining the relationship between the F-specific phages and other contaminants. Methods were subsequently found to eliminate or inhibit Salmonella phage growth. This study demonstrated the applicability of the membrane filter method for F-specific coliphage detection in natural waters and the usefulness of F-specific coliphages as indicators of fecal contamination. Modifications in the assay system to prevent interference by somatic Salmonella phages will further improve the specificity of the method. Further studies are needed to determine if F-specific coliphage serogrouping will distinguish between human and animal fecal contamination and if F-specific coliphages will indeed predict the presence of human enteric viruses in water.

6

7 Acknowledgements TABLE OF CONTENTS Page... iii Abstract Table of Contents List of Figures List of Tables... v Summary and Conclusions Recommendations Introduction Objectives Background and Review of the Literature Drinking Water and Health Human EntericViruses Infectious Hepatitis Rotaviruses Norwalk and Related Viruses Occurrence of Human Enteric Viruses in Drinking Water Indicator Organisms for Enteric Viruses Bacteriophages Structure and Classification Coliphages as Indicators Fecal Indicator Enteric Bacterial Indicator Viral Indicator FRNA Coliphages Serotyping FRNA Phages Resistance to Disinfection FRNA Coliphage Detection in Drinking Water as Treatment Indicators METHODS Introduction Host Bacteria F-specific Coliphage Assays Single Agar Layer (SAL) Method Membrane Filtration (MF) Method Bacteriological Analyses Collection of Samples vii ix xi xiii xv

8 4.5.1 Raw 'Water Samples Water Quality Data Reductions Through Water Treatment Processes Seleclion of Water Treatment Plants Collection of Samples Confirmation of Phage Plaques RNase Testing Wastewater Discharges Impacting Water Treatment Plants Phage Reductions Through Wastewater Treatment Plants RESULTS Survey of Water Treatment Plant Source Waters Raw Water Phage Levels and Wastewater Impacts Tests of Normality for Indicator Data Correlation Analysis Comparison Between MF and SAL Methods Relationships of Other Water Quality Factors to Bacterial and Phage Indicators Effect of Sample Volume on F-Specific Coliphage Recovery by the MF Method Microbial Reductions Through Water Treatment Plants Tests for RNase Sensitivity and Salmonella Phages F-specific Coliphages in Wastewater DISCUSSION REFERENCES viii

9 LIST OF FIGURES Page 3-1 Waterborne Outbreaks of Typhoid Fever. Gastroenteritis and Shigellosis ( ) Major Families of Bacteriophages Flow Diagram for the Fayetteville Water Treatnment Plant Flow Diagram for the Pittsboro Water Treatment Plant Flow Diagram for the Dunn Water Treatment Plant Flow Diagram for the Rocky Mount Water Treatment Plant Normal Probability Plots for E. coli and Enterococci Normal Probability Plots for F-specific Coliphages by MF and SAL Methods Correlation Plots of F-specific Coliphages by MF Method vs. E. coli and Enterococci as Log,, Values. with Pearson Correlation Coefficients Correlation Plots of F-specific Coliphages by SAL Method vs. E. coli and Enterococci as Log,, Values. with Pearson Correlation Coefficients Correlation Plot of E. coli and Enterococci as Loglo Values. with Pearson Correlation Coefficient Correlation Plot of F-specific Coliphages by SAL Method vs. MF Method... 48

10

11 LIST OF TABLES 3-1 Fecally Excreted Human Enteric Viruses Criteria for an Ideal Indicator Organism Water Treatment Plants and their Water Sources Levels of F-specific Coliphages. E. cob and Enterococci in Source Waters a Summary Statistics for Bacterial and Viral Indicators in Raw Water Samples b Summary Statistics for Water Quality Parameters Wastewater Effluent Discharges Impacting Raw Water Sources: Average Monthly Flow. Average Fecal Coliforms and Approximate Distances to Downstream Water Treatment Plants Mean Values for lndicators and Water Quality Parameters in Water Sources Categorized as High and Low Impact by Upstream Sewage Discharges Kolmogorov-Smirnov One-Sample Test Using Standard Distribution: Lillieffors Probabilities Tests of Association between Different Indicators a Paired Sample T Tests b Wilcoxon Signed Rank Tests a Pearson's Correlation Cofficients for E. coli. Enterococci and F-specific Coliphages by SAL and MF Methods with Other Water Quality Parameters b Probabilities Associated with Correlation Coefficients Indicator Turbidity in... Source Waters.~g~~ Reductions Water Treatment Processes: Composite of Samples from Four Treatment Plants Levels of lndicator Organisms and Log,, Reductions in Water Treatment Plants: Fayetteville Plant Levels of lndicator Organisms and Loglo Reductions in Water Treatment Plants: Rocky Mount Plant Levels of lndicator Organisms and Log., Reductions in Water Treatment Plants: DunnPlant Levels of lndicator Organisms and Log., Reductions in Water Treatment Plants: Pittsboro Plant... 56

12 Page Results of Parallel Assays of Surface Source Waters for F-specific Coliphages with and without RNase Present Proportions of F.Specific. FRNA and Somatic Salmonella Phages in Various Surface Source Waters F-specific Coliphages. Somatic Salmonella Phages and Indicator Bacteria in Raw Sewage and Treated Effluent of the Morehead City Sewage Treatment Plant Characterization of Phage isolates from Raw Sewage and Treated Sewage Effluent on Host S. typhimurium WG xii

13 SUMMARY AND CONCLUSIONS This study has generated some important information regarding the use of F-specific RNA coliphages as fecal and perhaps viral indicators of the quality of source and drinking waters and the use of a membrane filter (MF) technique to detect and enumerate them. The key findings are summarized below. 1. The MF method for the concentration of bacteriophages is applicable to a variety of source and drinking waters. The mean concentrations detected by this method are comparable to the single agar layer (SAL) method. Furthermore, the MF method permits much larger volumes to be examined. In addition, the detection of F-specific coliphages has potential as a rapid detection system because results can be obtained in 6 hours. 2. Somatic Salmonella phages can interfere with F-specific coliphage detection on host S. typhimurium WG49. Because this interference confuses possible relationships between the F- specific coliphages and other microbes, it may be necessary to prevent somatic Salmonella phage detection by adding an inhibitory agent to the assay system when using the WG49 host. Alternatively, a host that is resistant to somatic Salmonella phages but sensitive to F- specific coliphages could be developed. 3. The presence of F-specific coliphages enumerated via the MF method are an adequate indicator for fecal contamination in source waters. Levels of these phages are well correlated with two specific bacterial indicators of fecal contamination, E. coli and enterococci, and they are associated with the presence of known wastewater treatment plants discharging into the surface waters studied. F-specific RNA coliphages also have the potential to differentiate between fecal wastes of human and non-human origin, although this aspect of their indicator potential was not investigated in this study. 4. The MF procedure is simple enough to be used to monitor for F-specific coliphages on a periodic or routine basis to demonstrate compliance with the Surface Water Treatment Rule, which requires a minimum 4 log,, reduction of viruses through treatment processes. Phage reductions on the order of two to five orders of magnitude (2-5 log,,) can be followed through drinking water treatment plants by using the MF method. The actual phage reductions that can be followed depend on initial phage concentrations in raw water and the volume of finished water analyzed. Water treatment plants with lower raw water concentrations of phages could be followed for 3 to 4 log,, reductions by treatment, if larger water volumes were analyzed. This is possible using larger diameter filters or more replicate filters per sample. xiii

14

15 RECOMMENDATIONS Important issues to be resolved in the use of F-specific coliphages as indicators of fecal contamination concern the relative prevalence of F-specific RNA coliphages and F-specific DNA coliphages and ability of F-specific RNA coliphages to distinguish between human and animal fecal contamination on the basis of which serogroups are prevalent in the samples. F-specific RNA phage isolates collected during this study and in future samples should be serotyped according to the method described by Furuse (1987) to determine if they are from the serogroups most frequently associated with human sewage (groups II and Ill) or animal wastes (groups I and IV). Improvement in the MF method should continue with the development of a larger diameter filter holder which would accommodate larger diameter membrane filters. This would enable larger volumes of water to be filtered at faster rates as well as reduce the interfering effect of the accumulation of particulates on the membrane surface. Finally, this method should be used to compare the level of FRNA phages in source waters with the target pathogens, human enteric viruses. Parallel studies of source and treated waters for the presence of human enteric viruses and F-specific RNA coliphages should provide the most accurate evaluation of the usefulness of this method as an indicator system for the human enteric viruses.

16

17 1.O INTRODUCTION Documented waterborne outbreaks of gastroenteritis caused by viruses such as Norwalk Virus and infectious hepatitis caused by hepatitis A virus, have led to concern regarding the virological quality of drinking waters. With respect to microbial contaminants, traditional standards for finished water include turbidity, coliform bacteria and residual chlorine. Viruses, however, have been shown to be more resistant to treatment processes than the coliform bacteria and have been detected in finished waters meeting drinking water standards (Payment et al. 1985, Hejkal et al. 1982). The Surface Water Treatment Rule (SWTR) implemented by the United States Environmental Protection Agency is, in part, a response to the potential for virus transmission via drinking water (U.S. Environmental Protection Agency 1989). Under this rule water treatment facilities are required to filter and disinfect source waters in order to achieve a minimum 99.99% (4 log,,) reduction of viable human enteric viruses as well as a 99.9% (3 log,,) reduction of the protozoan Giardia lamblia. The SWTR created interest in the development of an indicator system, similar to the coliform bacteria, which would be more appropriate to the human enteric viruses. The direct quantitation of enteric viruses is difficult, expensive and impractical for testing on a routine basis. Due to the low infectious dose of viruses, large volumes of water must be tested in order to insure virus free drinking water, and virus assays require primate cell cultures that are technically demanding and expensive to maintain. A specific group of coliphages (viruses infecting E. col~) have shown promise as potential indicators of enteric viruses. The F-specific RNA coliphages are physically similar to the human enteric viruses and only infect male strains of E. coli. These bacteria produce the surface appendages, called pili, that are the receptor sites for these coliphages only when grown at elevated temperatures. Presumably, such bacteria originate only in the intestines of warm blooded animals. Furthermore, several of the coliphages in this group (MS2 and f2) have been studied extensively and have been shown to be similar to the enteric viruses in terms of their persistence in the environment and resistance to treatment processes. Most studies of F-specific RNA coliphages have examined sewage, human or animal feces, or heavily contaminated waters. These coliphages are known to be found in high numbers in sewage, relatively resistant to water and wastewater treatment processes and relatively persistent in the environment. They have thus been considered potential indicators of sewage pollution. Less is known about the ecology of F-specific coliphages in natural waters. In particular, there is little information regarding their relative presence in waters of various qualities and their relationship to other indicators of fecal pollution. This lack of information is due, in part, to the absence of a simple concentration method which would permit the routine sampling of a variety of waters. The simplest techniques available for the enumeration of coliphages are appropriate only to small sample volumes (generally no more than 100 ml). Concentration methods for larger sample volumes are available but are not appropriate for routine analyses.

18 The recent development of a membrane filter method for the concentration and enumeration of these coliphages was aimed at overcoming this obstacle. This method, which directly applies the membrane filter to an agar medium containing the host bacterium is simple, quick and inexpensive to use on a routine basis. Laboratory studies have demonstrated adequaterecoveries using spiked samples of tap and source waters (Sobsey et al. 1990). If applicable to field conditions this method could be a useful means of determining the quality of source waters used by drinking water treatment plants and as a means of estimating the levels of virus reductions achieved by these facilities.

19 2.0 OBJECTIVES The specific objectives of this study were as follows: 1. Evaluate the Membrane Filter (MF) method under field conditions by comparing recoveries via this method with an alternative method, The Single Agar Layer (SAL) method (Grabow and Coubrough 1986). 2. Compare the levels of F-specific coliphages, detected by both methods, with two specific bacterial indicators of fecal pollution, E. coli and enterococci, and with other water quality parameters. 3. Determine the effects of relevant water quality parameters such as temperature and turbidity on the recovery of F-specif ic coliphages. 4. Follow levels of F-specific coliphages and indicator bacteria through selected water treatment plants to determine whether 99.99% (4 log,,) reductions are achieved.

20

21 3.0 BACKGROUND AND REVIEW OF LITERATURE 3.1 DRINKING WATER AND HEALTH The potential of fecally contaminated water to transmit disease has been well established. Worldwide more than 250 million new cases of waterborne disease are reported each year resulting in more than 10 million deaths (Hazen 1988). Nearly 75% of these cases occur in tropical and subtropical areas of the developing world. The number one disease in the world today is diarrheal disease, which can cause extreme dehydration and death in very young children. In many developing nations drinking and bathing waters remain a primary transmission route for the bacterial, viral and protozoan agents responsible for diarrheal diseases. Here in the United States, and in many of the industrialized countries, the situation is much different. The realization of the importance of potable water to public health led to large investments in facilities to physically and chemically treat drinking waters and protect them from contamination with fecal material. The development of sanitary engineering and water bacteriology was greatly responsible for the reductions in waterborne outbreaks in the early part of this century. Outbreaks of enteric bacterial diseases such as cholera and typhoid fever, which were not uncommon in the last century, are extremely rare in the United States today. The role of water in the transmission of enteric diseases was first demonstrated by John Snow during an outbreak of cholera in London in 1854 (Snow 1936). Although the agent responsible for the outbreak, Vibrio cholera, was not identified until 30 years later, Snow deduced that contaminated water was responsible for the outbreak by comparing the incidence of disease among two communities with different sources of water. The community which drew its water from a polluted portion of the Thames River had a much higher incidence of cholera than a similar community which relied on water taken upstream from the point of contamination. In the United States, typhoid fever, caused by the bacterium Salmonella typi, was of great concern and was responsible for the early work in the development of water treatment processes. Experiments on water filtration were first carried out in the United States in the late 1800s, and the chemical coagulation-filtration process was patented in A landmark development in water treatment was the introduction of chlorine disinfection in 1908 in Jersey City, N.J. By 1914 most of the cities in the United States were using some type of chemical disinfectant. Thus, the standard treatment scheme used today, filtration and disinfection, was basically established by the early part of this century. The effects of these developments were dramatic. The death rate from typhoid fever decreased from an average of 30 per 100,000 to less than 0.1 per 100,000 (Fair et al. 1967). Similar reductions were seen for other waterborne diseases such as cholera, Shigellosis and Salmonellosis. Although the number of cholera and typhoid outbreaks have decreased over the last 60 years, the overall number of outbreaks has not. Figure 3-1 shows the number of waterborne outbreaks in the United States over the period The increase in outbreaks since

22 1965 is most likely due to better surveillance and reporting procedures as well as the increased number of individuals being served by community and non-community systems. Although the number of outbreaks has not changed over this time period, the etiologic agents responsible for these outbreaks have changed. Salmonella typhi, responsible for typhoid fever, was the pathogen most commonly identified in waterborne outbreaks prior to From hepatitis A virus and Shigella were the most frequently identified, and from 1975 onward Giardia lamblia was most often cited. Overall the disease most often reported is acute gastroenteritis of undefined etiology (Craun 1988). Thus, while the numbers of outbreaks due to bacterial pathogens has been decreasing, there have been increasing numbers of outbreaks caused by other pathogens, most important of which are the human enteric viruses and the protozoans G. lamblia and Cryptosporidium parvum. During the period there were 50 outbreaks resulting in 25,846 cases. Of these, only 7 (14%) of the outbreaks and 3050 (1 1.8%) of the cases were due to bacterial agents (Salmonella, Shigella and Campylobacter). Acute gastro-intestinal illness of unknown etiology was responsible for 48% of the outbreaks and 11.5% of the cases (Craun 1988). According to Levine et al. (1990), the incubation period, duration of illness and symptoms for many of these outbreaks, suggests that they were caused by 27nm diameter Norwalk-like viruses. The realization that many outbreaks of waterborne disease of unknown etiology may be due to enteric viruses created interest in determining the role of drinking and bathing waters in the transmission of these agents. Concern about the waterborne transmission of viruses stems from several factors: (1) virus laden wastes may contaminate both surface and ground waters, which may ultimately be the source water for a community's water supply; (2) enteric viruses are often more resistant to conventional water treatment processes such as filtration and disinfection; (3) the infectious dose of viruses for humans may be as small as one cell culture infectious unit; (4) enteric viruses can produce a variety of diseases, not all of which are readily and causally identified as arising from contaminated drinking water (Rao and Melnick 1986). Raw sewage is known to contain large numbers of enteric viruses. Reported recoveries of viruses in sewage vary widely, depending on the method used and many other factors, but an average of 7000lL has been estimated (Clarke et al. 1964, Melnick 1987). The number may be 100 to 1000 times higher in less developed countries (Melnick and Gerba 1980). In regions where sewage is discharged directly to receiving waters, or applied to agricultural land, large numbers of viruses may be released into the environment. Wastewater treatment plants are less successful in reducing viruses than bacteria. Typical virus reductions by conventional primary and secondary treatment followed by disinfection are only 90-99% (Rao and Melnick 1986). Hence, conventionally treated wastewater effluent, even with disinfection, is likely to contain appreciable levels of enteric viruses that are discharged into receiving waters.

23 FIGURE 3-1 : WATERBORNE OUTBREAKS OF TYPHOID FEVER, GASTROENTERITIS AND SHlGELLOSlS ( ) A typhoid Adapted from Craun 1986a

24 3.2 HUMAN ENTERIC VIRUSES Over 100 types of enteric viruses are excreted by infected humans in their feces and some are also excreted in their respiratory secretions, urine and other exudates. These viruses may thus find their way into sewage treatment plants and eventually surface waters (Melnick and Gerba 1980). Table 3-1 lists the families of these viruses and the symptoms or diseases they cause. The enteroviruses include polioviruses, coxsackieviruses A and B, enteric cytopathogenic human orphan viruses (echoviruses) and enteroviruses Hepatitis A virus, which is morphologically similar to the enteroviruses, was formerly classified as enterovirus 72, but is now in a separate genus. These viruses, belonging to the Picornaviridae family, are small (27nm), non-enveloped, icosahedral shaped and contain single stranded RNA. Poliovirus causes a severe disease, paralytic poliomyelitis, although this clinical disease occurs in only 0.1-1% of infections (Rao and Melnick 1986). In industrialized countries such as the United States, where a live, oral polio vaccine is used, the polioviruses excreted into the environment are usually vaccine derived and non-pathogenic. Table 3-1 : Fecally Excreted Human Enteric Viruses Virus Group Diseases or symptoms caused Enteroviruses: -Poliovirus -Echovirus -Coxsackie A -Coxsackie B -Enterovirus types Hepatitis A virus Norwalk virus and other small, round, viruses (Caliciviruses, Astroviruses Parvoviruses, etc.) Reoviruses Rotavirus Adenoviruses Epidemic non A, non B hepatitis virus (hepatitis E virus; HEV) Coronaviruses Polio, meningitis, fever Meningitis,fever,gastroenteritis Herpangina, respiratory disease, meningitis, fever, hand, foot and mouth disease Myocarditis, pleurodynia, fever, meningitis Meningitis,encephalitis, respiratory disease, fever, acute hemorrhagic conjunctivitis lnfectious hepatitis Acute adult gastroenteritis (epidemic vomiting and diarrhea) Respiratory and enteric infections Childhood gastroenteritis; diarrhea Respiratory disease; childhood gastroenteritis lnfectious hepatitis Respiratory illness; enteric infections Adapted from: Rao and Melnick 1986 and Havelaar 1986.

25 3.2.1 Infectious Hepatitis. Hepatitis A virus (HAV) has been demonstrated to cause waterborne outbreaks of infectious hepatitis. Although personal contact and foodborne transmission are responsible for the majority of hepatitis outbreaks, water has been clearly shown to be a vehicle for numerous outbreaks. It is also assumed that many cases of waterborne HAV go unreported due to the difficulty in determining the source of infection. Between 1958 and 1972, 83 waterborne epidemics of HAV were documented to have occurred in the United States (Mosley 1967, Goldfield 1976). Cliver (1 983) listed 1 17 outbreaks of infectious hepatitis due to contaminated drinking water. In the United States from 1971 to 1983, 22 outbreaks occurred causing 730 cases (Craun 1986a, ). In Georgetown, Texas, a waterborne outbreak occurred in 1980 in which 36 cases were reported. A few weeks before the HAV outbreak, an outbreak of gastrointestinal illness occurred involving 79% of the 10,000 residents of the community (Hejkal et al. 1982). Coxsackie virus B2 and B3 were isolated from the well water and Coxsackie B3 was also isolated from a sample of chlorinated tap water. Recently a new causative agent of infectious hepatitis has been recognized. Waterborne non- A, non-b hepatitis (hepatitis E) has been reported to be responsible for at least 5 epidemics in lndia and others in Nepal and the Soviet Union (Rao and Melnick 1986). It is suspected to be the causative agent for the single largest documented waterborne outbreak of infectious hepatitis which resulted in 30,000 cases in New Delhi, lndia in (Khuroo 1980). It is also suspected as the agent responsible for a hepatitis outbreak in Algeria resulting in 788 cases (Belabbes et al. 1985). The virus is clearly different from HAV and can infect individuals previously infected with HAV. It is now known that hepatitis E virus is a calicivirus, which puts it in the same taxonomic group as Norwalk virus and other Norwalk-like viruses of acute gastroenteritis. The caliciviruses are about nm in diameter, and they consist of positive sense, single-stranded RNA, about 7.5 kilobases long, which is surrounded by an outer capsid containing a single major protein. The most serious illnesses caused by hepatitis E virus occur in pregnant women where the case fatality rate averages 20% (Rao and Melnick 1 986) Rotaviruses. Rotaviruses are known to be a major cause of infantile gastroenteritis throughout the world. Approximately half of the hospitalized cases of acute diarrhea in very young children in temperate climates are caused by this agent (Blacklow and Cukor 1981). Rotavirus has also caused outbreaks among adults and is a major cause of travellers diarrhea (Halvorsrud and Orstavik, 1980). Rotaviruses are about 80 nm in diameter, contain I I unique segments of double-stranded RNA and have a double capsid layer giving them a wheel-like appearance. Rotavirus has been suspected as the causative agent in several waterborne outbreaks of gastroenteritis. These include an outbreak in Sweden in 1977 resulting in over 3000 cases and one in Colorado in 1981 in which over 1700 cases of gastroenteritis were reported (Lycke -- et al. 1978, Harris et al. 1983) Norwalk and Related Viruses. The Norwalk agent and related viruses are now believed to be a major cause of waterborne outbreaks of gastroenteritis in the United States. Although Norwalk virus is difficult to detect, due to the inability to propagate the virus in vitro, it has been detected by electron microscopy and immunoassays in fecal specimens from affected individuals of several waterborne outbreaks. Because of its characteristic symptomology (diarrhea, vomiting, abdominal pain, etc.), short incubation period (1-2 days),

26 and short duration of illness (1-2 days), it has been linked to a number of outbreaks of acute gastroenteritis of unknown,etiology in which these symptoms were prevalent. Norwalk virus is a calicivirus and it is antigenically and genetically related to a number of other small, round, structured viruses (SRSVs) that cause similar gastrointestinal illnesses, such as Snow Mountain agent, Hawaii agent and many others. These viruses have a characteristic appearance marked by cup shaped surface depressions. In 1977, commercially manufactured ice, contaminated with Norwalk-like viruses, was responsible for approximately 5000 cases of illness over several northeastern states (Levine -- et al. 1990). A study by the Centers for Disease Control estimated that 42% of the acute nonbacterial gastroenteritis outbreaks occurring from were caused by this agent (Kaplan et al. 1982). From , 24 of 50 reported waterborne outbreaks were of unknown etiology. Case histories from many of these outbreaks also suggest a Norwalk-like virus (Levine et al. 1990). 3.3 OCCURRENCE OF HUMAN ENTERIC VIRUSES IN DRINKING WATER The presence of enteric viruses in water treatment plant source waters and their resistance to normal treatment processes indicate that small numbers of human enteric viruses may be found in tap water meeting drinking water standards. Various researchers have, in fact, demonstrated the presence of enteric viruses in finished water (Payment et al. 1985, Keswick -- et al. 1984, Deetz et al. 1984). Because of the low infectious dose for most viruses it may, therefore, be possible that low levels of gastrointestinal illness may occur in communities served by such waters. At the present time drinking water is clearly demonstrated as a vehicle primarily in defined, common-source, and sometimes relatively large waterborne outbreaks. It is currently difficult to demonstrate low level or endemic transmission via drinking water for several reasons. First, many of the infections may be either mild or asymptomatic and may not be reported. Second, due to the low numbers and sporadic distribution throughout the community, drinking water may not be suspected, especially if the tap water is consistently free of coliform bacteria and maintains a residual concentration of chlorine. Third, the diversity of symptoms that may become manifest during viral illness may not indicate a common source. A significant role of drinking water in community-wide, endemic gastrointestinal illness has been suggested by a recent study (Payment et al. 1991). This community-based epidemiological study compared rates of gastrointestinal illness in households drinking conventionally treated tapwater derived from a contaminated surface source to households in which this water was further treated by reverse osmosis. About 30% more gastroenteritis occurred in the households drinking tapwater, and the symptoms of illness were most compatible with a viral etiology. Recognition of the role of contaminated surface water in causing drinking waterborne outbreaks, especially unfiltered supplies, led the United States Environmental Protection Agency to create more stringent standards for drinking water treatment. The Surface Water Treatment Rule (SWTR) of the National Primary Drinking Water Regulations (NPDWR), issued under the Safe Drinking Water Act (SDWA), are in response to concerns regarding the

27 potential adverse health effects of viruses, Giardia lamblia and Legionella (U.S. Environmental Protection Agency 1 989). Under these regulations water treatment plants are required to achieve a 99.99% (4 log,,) reduction in viruses and a 99.9% (3 log,,) reduction in G. lamblia throughout the treatment process. As no effective means of monitoring actual virus levels in raw and finished waters is now available, the SWTR of the SDWA allows for specific treatment processes to count towards the 99.99% goal for viruses. Therefore, water treatment plants using rapid filtration preceded by optimized coagulation and sedimentation are granted a 99%-99.9% (2-3 log,,) virus reduction. Then, they are required to apply disinfectant sufficient to reach a further 90-99% (1-2 log,,) reduction. The promulgation of the SWTR of the SDWA has created interest in developing a simple and reliable monitoring system for the presence of human enteric viruses in source and drinking waters. This has, in turn, promoted interest in developing an indicator system, similar to the coliform bacteria, which would provide a more accurate means of estimating the viral quality of source and drinking waters. 3.4 INDICATOR ORGANISMS FOR ENTERIC VIRUSES lndicator organisms have been proposed for several functions including the detection of fecal pollution (index function) and as model organisms to monitor the efficiency of treatment processes (indicator function). Ideally, an indicator would be present when pathogens are present and absent when pathogens are absent under all types of water and environmental conditions. Table 3-2 lists those characteristics of the ideal indicator. Table 3-2: Criteria for an Ideal lndicator Organism Should be present when pathogens are present and absent when pathogens are absent. Persistence and growth characteristics of the indicator should be similar to pathogens. lndicator should not multiply in the environment. Ratio between indicator and pathogens should be constant. lndicator should be present in greater concentrations than pathogens in contaminated waters. lndicator should be as resistant to adverse environmental factors and disinfection as pathogens. lndicator should be non-pathogenic and easy to quantify. Tests for the indicator should be easy and applicable to all types of water. Adapted from Goyal, S.M. 1983

28 Since 1914, coliform bacteria have been used to determine the sanitary quality of drinking waters (Hazen 1988). Although they do not meet all of the requirements listed in Table 3-2, they have traditionally been considered a useful indicator, especially for the classical bacterial pathogens, such as Salmonella, Shigella and Vibrio cholerae. Many shortcomings have become apparent. Not all of the coliform bacteria detected in water are necessarily of fecal origin, resulting, therefore, in occasional false positives for fecal contamination. This is especially a problem in tropical waters where total and fecal coliforms have been demonstrated not only to persist for long periods in the environment but also to multiply. Fujioka (1 988) found between 100 and 10,000 colony forming units of fecal coliforms per 100 ml in streams in Hawaii, including streams with no known fecal input. Rivera et al. (1988) found E. coli in epiphytes 15m above the ground in forests of Puerto Rico. Hazen and Toranzos (1990) suggest that E. coli may, in fact, be a naturally occurring bacterium in tropical rain forest watersheds. In tropical climates coliforms may not reliably estimate the sanitary quality of source waters and tend to overestimate the levels of pathogens. Recent advances in the understanding of the role of viruses and protozoans in waterborne disease outbreaks have also drawn attention to other problems of coliform bacteria. Compared to coliforms, viruses and protozoan cysts are known to be more persistent in the environment and more resistant to treatment processes. Therefore, drinking water or source waters free of coliform bacteria may not necessarily be free of viruses. The isolation of human enteric viruses in finished waters meeting coliform standards demonstrates the inadequacy of coliform bacteria to insure virus free water. Payment et al. (1 985) found viruses in 11 of 155 finished water samples, including poliovirus, coxsackie B3, B4, 85 and echo 7. Keswick et al. (1984) found 19 of 23 samples of finished drinking water contained viruses. Seven of nine dry season samples met turbidity, total coliform and chlorine residual standards and four of these seven samples contained viruses. Deetz et al. (1984) detected rotavirus in 10 of 10 drinking water samples in Mexico, some of which met total coliform standards. The inability of coliform bacteria to insure virus free waters has promoted interest in developing an indicator system more appropriate to the human enteric viruses. A logical means of insuring virus free drinking water is to analyze water samples directly for enteric viruses. This process is, however, technically difficult, time consuming and expensive. Large volumes of water need to be processed for virus concentration, and cell culture capabilities are required for virus detection and assay. The methods used are insufficiently sensitive due to limitations in the techniques, and some viruses, notably HAV and the Norwalk-type viruses, are undetectable because they do not grow well or at all in cell cultures. It is currently not feasible to monitor water supplies through the direct detection of human enteric viruses on a routine basis. A promising alternative to the direct detection of enteric viruses is the use of bacteriophages (bacterial viruses) as model organisms (IAWPRC Study Group 1991). Of special interest are the coliphages, the bacteriophages infecting E. coli and perhaps other coliform bacteria. As viruses, bacteriophages are physically more closely related to human enteric viruses and show similar characteristics such as persistence in the environment and resistance to treatment processes (Bell 1976, Snead et al. 1980, Kott 1981, Wentsel 1982).

29 3.5 BACTERIOPHAGES Structure and Classification. Bacteriophages are viruses which use bacterial cells as their replicative host. The non-enveloped bacterial viruses consist of a protein coat, or capsid, surrounding the nucleic acid. The nucleic acid may be either RNA or DNA and either single or double stranded. Six morphological groups of bacteriophages have been described by Bradley (1967), and they are shown in Figure 3-2. The first three groups (groups A-C) comprise phages that contain a capsid of cubic symmetry as well as a tail of various lengths. These tailed phages all contain double stranded DNA and infect the host cell via the cell wall (somatic phages). Group D consists of somatic, non-tailed phages with cubic symmetry and single stranded DNA as the nucleic acid. Group E phages exhibit cubic symmetry and contain single stranded RNA. Group F phages are long filamentous (helical) phages containing single stranded DNA. Phages belonging to Groups E and F differ from the other 4 groups in their point of attachment to the host cell. These phages adsorb to the hairlike projections, called pili, that are produced only by male strains of bacteria. In E. coli and similar bacteria the genetic information for production of these pili is encoded by a gene on the sex or F-plasmid, an extrachromosomal genetic element found in so-called male strains of E. coli. This is a transmissible plasmid that can be transferred by bacterial conjugation from male (F-plus) cells to female or F-minus cells. F+ E. coli and related bacteria possessing F-pili can be infected by pilus-dependent (male-specific or F+) coliphages. They can also be infected by somatic coliphages that attach directly to the bacterial outer envelope or cell wall. Bacteria without F- pili (female or F-minus strains) can be infected only by somatic phages. The F+ coliphages of group E are about 25 nm in diameter, icosahedral and contain single-stranded RNA. Those belong to group F are filamentous and contain double stranded DNA (Figure 3-2). Plasmids encoding for F-pili production have been successfully transferred to other cells of related species (Birge 1988). Havelaar and Hogeboom (1 984) successfully transferred the F plasmid from a male strain of E. coli K-12 to Salmonella typhimurium (strain WG45) which subsequently produced F-pili. These cells (designated S. typhimurium strain WG49) are thus vulnerable to infection by somatic Salmonella phages as well as by F-specific RNA and DNA phages. Others have transferred the F-plasmid to species of Shigella and Proteus (Birge, 1 988). 3.6 COLIPHAGES AS INDICATORS Coliphages have been considered as possible alternative or additional indicators to the traditional coliform bacteria. Gerba (1 987) suggests three possible indicator functions of coliphages: (1) indicators of fecal contamination, (2) indicators of enteric bacteria and (3) indicators of enteric viruses. Coliphages have been considered as potential indicators of enteric viruses in fecally contaminated water for several reasons. They are found in relatively high numbers in sewage and fecally contaminated source waters, under most circumstances they do not multiply in the environment, they are relatively persistent in the environment and resistant to treatment processes, and enumeration methods for them are simple and rapid Fecal Indicator. Coliphages are considered potential indicators of fecal pollution because they are found in high numbers in human and nonhuman feces and in sewage, are persistent in the environment, and supposedly do not multiply in the environment. In addition

30 FIGURE 3-2: MAJOR FAMILIES OF BACTERIOPHAGES (Adapted from Havelaar 1986) A. Myoviridae ds DNA. cell wall B. Styloviridae ds DNA cell wall Z. Podovi r idae ds DNA cell wall D. Microviridae ss DNA 0 cellwall 30 nm E. Leviviridae - F Inoviridae ' ssrna 0 sexpilus 24 nm MS2 fd ss DNA sex pilus 810x6 nm Each box shows: morphological group, family name, morphology and name of type species, type of nucleic acid, situation of receptor in host-cell, dimensions of capsid (head in case of tailed bacteriophages) of type species.

31 results of phage assays can be obtained in as little as six hours. Dhillon et al. (1976) investigated the levels of coliphages in humans, cows and pigs and found numbers ranging from 10' to I o7 plaque forming units (PFU)/g of feces. Counts varied among species and tended to be highest in the pig samples. Osawa et al. (1981) found similar ranges of phages, with human feces providing the lowest numbers, usually less than 1000 PFUIg. Havelaar gt - PFUIg in other mammalian species. al. (1986) reported coliphage counts ranging from greater than lo4 PFUIg in humans up to lo7 Despite the inconsistencies in coliphage levels in human fecal samples, raw sewage has been shown to be a consistent source of coliphages. Dhillon et al. (1970) found up to lo4 PFUIml in Hong Kong sewage. Osawa et al. (1981) and Furuse (1987) found up to lo6 PFUlml in Japan. Kott et al. (1974) detected between lo4 and lo5 PFUlml while the number of enteric viruses remained less than 1 PFUlml. Because viruses are obligate intracellular parasites, they require a host cell for replication and cannot otherwise multiply. For bacteriophages this means the presence of high enough concentrations of host bacteria for adsorption and infection to occur. These concentrations do occur in raw sewage, and raw sewage may be a possible site of limited F+ coliphage multiplication (Havelaar et al. 1984a). However, such multiplication appears to occur only in raw sewage and not elsewhere in the environment. This is probably because the required concentrations of susceptible host cells as well as the required temperatures for phage infection (>30 C) are both too low in environmental waters (Havelaar and Pot-Hogeboom 1 988). Previous studies have attempted to compare levels of coliphages to other indicators of fecal pollution. Differences in the host bacterium used for phage enumeration (usually E. coli C for somatic coliphages or E. coli K12 for F+ coliphages) as well as the bacterial indicator used (total coliforms, fecal coliforms and fecal streptococci) make comparisons between studies somewhat difficult. Toranzos et al. (1988) found somatic coliphages in samples of all Puerto Rican source waters known to be contaminated with sewage, but none were found in surface waters which were considered pristine. Suan et al. (1988) found somatic coliphages to be highly correlated with fecal coliforms in southeast Asian surface waters. Castillo et al. (1988), in Chile, found lower correlations between somatic coliphages and fecal coliforms (r=0.387) and a very weak correlation with total coliforms (r=0.029). Bell (1976) found that the coliform to coliphage ratio in sewage, secondary effluent and river water varied widely. Kenard and Valentine (1 974) showed consistent ratios between fecal coliforms and coliphages in sewage and river water of varied levels of contamination. Wentsel et al. (1982) found high correlations between fecal coliforms and coliphages in sewage, source and treated waters (r=0.69). Loh et al. (1988), examining surface and ground waters, reported a high correlation between fecal coliforms and total coliphages (r=0.77). Seeley and Primrose (1980) divided coliphages into three groups based on incubation temperatures: low temperature (LT) 15-30' C, mid temperature (MT) 15-45' C and high temperature (HT) 25-45' C. The authors suggest that the low temperature phages are not related to fecal pollution whereas the HT and many of the MT phages are so related. Parry gt - al. (1981) demonstrated the multiplication of LT phages in river water to which E. coli cells

32 were added. As a result of these findings, Seely and Primrose (1980) have suggested that the HT coliphages are specific to fecal contamination and might be useful as indicators of fecal contamination Enteric Bacterial Indicator. As mentioned above, coliphages have been demonstrated to be correlated with the bacterial indicators which are themselves indicators of enteric bacterial pathogens. Wentsel (1 982) also detected phages in treated water samples in the absence of indicator bacteria. Gerba (1 987) suggests coliphages may be good indicators of source waters but not treated waters due to the level of false positives as a result of the greater resistance of coliphages to treatment processes Viral Indicator. Coliphages have gained a great deal of attention as indicators of enteric viruses due to physical similarities, greater densities in sewage and feces than enteric viruses, and their simplicity of detection. Simkova and Cervenka (1 981) compared coliphages to enteric viruses in four types of water: sewage, irrigation channels, river waters and wells. They found general agreement in terms of presencelabsence and noted that seasonal variations for coliphages and enteric viruses followed the same pattern. Kott et al. (1 974) found the ratio of coliphages to enteric viruses ranged from 1 :I to 1000:l and also noted similar seasonal variations. However, in eight samples positive for coliphages only two were positive for enteroviruses. Vaughn and Metcalf (1 975) studying estuarine waters, found that the ratios between phage and enteric viruses varied widely. They also noted that 63% of the samples positive for enteric viruses were negative for coliphages, although this may be due, in part, to the recovery method used (Gerba 1987). Stetler et al. (1984) followed the reduction of coliphages, enteric viruses and indicator bacteria through a model drinking water treatment plant. He found a stronger correlation (r=0.54) between coliphages and enteric viruses than between the enteric viruses and the bacterial indicators. He also noted a seasonal fluctuation with both phages and viruses following a similar pattern. Coliphages outnumbered the human enteric viruses at each treatment step and stepwise reductions were similar for each. Geldenhuys and Pretorius (1 989) compared somatic coliphages and three bacterial indicators to enteric virus levels in polluted source waters in South Africa. They found the strongest correlation (r=0.42) was between the coliphages and enteric viruses. The correlation coefficients between coliphages and total and fecal coliforms were much lower (r=0.15 and r= respectively). The detection of coliphages in treated drinking water (Keswick et al. 1984, Sim and Dutka 1987, El-Abagy et al. 1988, Castillo et al. 1988, Ratto et al. 1989) is an indication of their resistance to treatment and potential usefulness as process indicators in water and wastewater treatment. The potential of coliphages as model organisms for enteric viruses may depend on several factors such as the choice of the host bacterium and an acceptable concentration method that can used for large volumes of water. Ideally, a host bacterium would detect only phages related to fecal pollution. Havelaar (1984) suggests that the HT phages described by Seely and Primrose (1 980) may be the most useful indicators. His research has, in fact, suggested that a particular group of HT phages, the F- specific RNA coliphages termed "FRNA phages" which infect only F+ cells grown at elevated

33 temperatures, may prove to be a useful indicator of enteric viruses (Havelaar, 1984, 1987, 1993; Havelaar et al., 1993). 3.7 FRNA COLIPHAGES F-specific RNA (FRNA) coliphages are a relatively homogeneous group of small, icosahedral coliphages from the group E Leviviridae previously described (Figure 3-2). Together with the filamentous F-specific DNA (FDNA) phages they comprise the F-specific coliphages, or coliphages which infect via the F-pili. As they are physically similar to the enteroviruses (small, icosahedral, single stranded RNA), the FRNA phages have generated interest as possible model organisms for the human enteric viruses. FRNA phages infect via the sex pili, which are produced by the host cell only at certain temperatures. Maximum production occurs between 37" and 42" C (Havelaar 1986). At temperatures below 30" C much fewer pili are produced and below 25" C none are produced. (Havelaar and Pot-Hogeboom 1988). Thus, it is believed that FRNA phages can multiply only in cells that were originally produced at elevated temperatures. Presumably, this elevated temperature condition occurs mainly in the intestines of warm blooded animals, and hence, only these cells would possess the F-pili. Extensive studies of fecal specimens from humans and other animals show wide discrepancies in the levels of FRNA phages. Havelaar et al. (1986) found humans to be a poor source of FRNA phages although high levels were found in pigs, sheep, calves and especially broiler chickens. Somatic coliphages were found much more frequently in humans although, again, at lower numbers than in the other animals. Osawa et al. (1981) examined 597 samples of human feces and found 2.3% contained FRNA phages. Furuse et al. (1983a) also found low levels of these phages. In sewage, however, FRNA phages are usually found in relatively high numbers. Havelaar (1 986) found FRNA phage levels to be approximately 1000 to 10,000 per ml of raw sewage. Furuse et al. (1983b) found variable levels of FRNA phage in sewage and fecal samples in Senegal, Ghana and Madagascar. In several raw sewage samples the total coliphage counts exceeded lo5 per gram yet no FRNA phages were detected. Discrepancies such as these may be due to differences in the choice of host bacterium and the methods of detection. The low levels of FRNA phages in human feces and high levels in raw sewage has led some to conclude that multiplication must occur in the environment (Havelaar, 1984). Although F-pili are produced only at elevated temperatures, cells grown at temperatures above 30" C can remain piliated when released into the environment. Havelaar and Pot-Hogeboom (1988) found that FRNA coliphage GA could multiply at 20 C if the host cells were grown at a higher temperature of 37 C and, therefore had pili. These results support the suggestion of Havelaar -- et al. (1986) that multiplication may occur in sewage where host cells originating in human intestines are continually added to the system. Wiggins and Alexander (1 985) have estimated that a concentration of lo4 colony forming units of host cells per ml are required in order for phage replication to occur. In the case of FRNA phages this means that lo4 piliated cells per ml would have to be present in the sewage environment for replication to occur. Raw sewage appears to be an ecological niche for

34 FRNA coliphages, thus making them a useful indicator of sewage pollution (Havelaar and Pot-Hogeboom 1986) Serotyping FRNA Coliphages. FRNA coliphages can be divided into four antigenic groups (I-IV) based upon serotyping. Studies by Furuse et al. (1 978, 1981, 1983a,b) and Havelaar et al. (1 986) have shown that the predominant strains in sewage belong to groups II and Ill whereas those in animal feces generally belong to group I. Group IV contains both human and non-human FRNA coliphages. Furuse et al. (1983b) has also demonstrated differences in the geographical distribution of specific FRNA serotypes. This categorization of FRNA phages by serogroup permits a greater degree of specificity in determining the origin of these coliphages. Theoretically, one can serotype plaques isolated from fecally contaminated waters and determine if the contamination is predominately human or animal in nature Resistance to Disinfection. Several studies have examined the inactivation of two prototype FRNA coliphages, f2 and MS2, by disinfection processes. Most report equal or greater resistance to chlorine disinfection by these phages as compared to the enteroviruses but often greater sensitivity to chlorine dioxide and ozone (Sobsey 1989). The RNA phage f2 was reported to be more resistant to chlorine than poliovirus type 1 (Shah and McCamish 1972, Kott et al. 1974) poliovirus type 3 (Cramer et al. 1976) phage T4 (Shah and McCamish 1972), and coliform bacteria (Longley et al. 1974). MS2 was often, but not always as resistant to free chlorine as hepatitis A virus in demand-free waters but was significantly more resistant than HAV in tap or effluent waters (Grabow et al. 1983). Coliphage f2 was reported to be more sensitive to chlorine dioxide than poliovirus type 1 (Olivieri et al. 1985, Tifft et al. 1977) and very sensitive to ozone (Rosen et al. 1975). Finch and Fairbairn (1991) reported that MS2 was more sensitive than poliovirus 3 to ozone. However, this comparison may be flawed by differences in the methods of virus purification that probably resulted in poliovirus 3 being more highly aggregated than MS2 and hence more resistant to disinfection. Hall et al. (1993) compared the inactivation of coliphage MS2 and hepatitis A virus by ozone and by mixtures of ozone and hydrogen peroxide. Both viruses were inactivated completely ( log,,) within 5 seconds by 0.4 mgll ozone. In mixtures of ozone and hydrogen peroxide both HAV and MS2 were completely inactivated within 5 seconds at ph 6-8, but MS2 survived longer than HAV at ph 10. Hence, MS2 was considered a good model for HAV inactivation by ozone and mixtures of ozone and hydrogen peroxide. It appears that FRNA coliphages are suitable indicators of enteric viruses for chlorine disinfection of water and wastewater. However, they may be more sensitive than human enteric viruses to inactivation by ozone or chlorine dioxide, which would make them unreliable indicators of enteric virus inactivation by these disinfectants. Because the literature on this subject is inconsistent, further studies are needed FRNA Coliphage Detection in Drinking Water as Treatment Indicators. The relative resistance of FRNA phages and their relationship to sewage or fecal pollution are the primary reasons for considering these phages as potential model organisms of human enteric viruses.

35 Of particular interest is the use of these coliphages in water and wastewater treatment plants to model the reductions of enteric viruses. Two problems concerning the enumeration of FRNA phages are the interference by somatic phages and the establishment of an acceptable method for concentrating the large volumes of water required at water treatment facilities to insure virus free drinking waters. As male strains of E. coli can be infected by F-specific RNA and DNA phages, as well as somatic phages, cell lysis or plaque formation may be a result of infection by any of these types of coliphages. Havelaar (1986) has, in fact, suggested that the majority of phages detected on these hosts are somatic phages. One solution is to perform parallel assays on male and female hosts and subtract the somatic phages detected on the female host from the total number detected on the male host. Havelaar and Hogeboom (1 984) attempted to overcome this problem by devising a modified bacterial host strain more specific to FRNA coliphages. The F-plasmid responsible for the production of F-pili was inserted from a male strain of E. coli into a strain of Salmonella typhimurium, which thereby causes it to produce F-pili. This bacterium is therefore susceptible to infection by F-specific RNA and DNA coliphages and also somatic Salmonella phages. As Salmonella tend to be less numerous than coliforms in the environment and in sewage, the majority of phages would, theoretically, be F-specific coliphages. Samples of raw sewage examined with this host have, in fact, shown a predominance of FRNA coliphages. Somatic Salmonella phages were generally a small percentage of total phages detected (Havelaar, 1984). Furthermore, by using RNase in the assay, which will inactivate RNA phages, one can easily differentiate FRNA and FDNA phages. This host bacterium has proven useful in the enumeration of FRNA coliphages in sewage. Results of different studies on the reduction of FRNA and other coliphages by sewage treatment processes and their occurrence in treated effluents tend to vary. In general, reductions of FRNA and other coliphages are similar to those of enteric viruses. Both enteric virus pathogens and coliphage indicator viruses are reduced less extensively than are enteric bacteria, such as coliforms, E. coli and fecal streptococci, especially if the treated effluent is disinfected (IAWPRC Study Group 1991, Havelaar, 1993). Recently, Havelaar et al. (1993) reported a good correlation between FRNA coliphages and culturable enteroviruses in different types of treated and untreated wastewater and fresh surface waters in the Netherlands. If found in sufficient numbers in source waters, FRNA coliphages could be used to monitor the effectiveness of water treatment processes with respect to the reduction of enteric viruses. Thus, FRNA coliphages would serve as indicators of treatment effectiveness for enteric virus reductions, assuming that FRNA coliphages and human enteric viruses are reduced similarly by the water treatment processes employed. The routine testing of source and drinking waters for F-specific coliphages has been hampered by the lack of a simple concentration and enumeration method. The recent development of a simple membrane filtration technique for the direct enumeration of coliphages facilitates the testing of larger volumes of source and drinking water for FRNA coliphages (Sobsey et al. 1990). The membrane filter method, when combined with a suitable host bacterium, may provide a simple and efficient means of determining the viral quality of

36 source and treated drinking waters, if FRNA coliphages are indeed reliable indicators of human enteric virus presence and reductions by treatment processes.

37 4.0 METHODS 4.1 INTRODUCTION In order to determine the levels of F-specific coliphages in source waters, samples of intake water were collected from 21 municipal water treatment plants over a 15 month period. The treatment plants and their water sources are listed in Table 4-1. Water samples were tested for F-specific coliphages and two groups of indicator bacteria considered specific for fecal contamination, E. coli and enterococci. Comparisons were made between the indicators to evaluate the relationship between F-specific coliphages and the two bacterial indicators of fecal pollution. F-specific coliphages were tested by two methods: a new membrane filtration (MF) technique, and a single agar layer (SAL) method described by Grabow and Coubrough (1986). The two methods were also compared with each other to determine their relative efficiency of F- specific coliphage recoveries. Water quality data were also collected from each facility at the time of sample collection. Each parameter was analyzed using correlation analysis to determine its relationship to the recovery of F-specific coliphages. Information regarding waste water discharges into North Carolina surface waters was collected from the North Carolina Department of Environmental Management. This was used to select water treatment plants for the survey as well as to compare FRNA coliphage levels with respect to the location of waste water discharges. Several water treatment plants were chosen from the list of 21 plants in order to conduct more intensive analysis of F-specific coliphage and indicator bacteria reductions by conventional water treatment processes. Levels of phages in source water and at each stage of treatment could then be used to determine phage reductions by treatment and to attempt to determine if a 99.99% (4 log,,) reduction, as required for viruses by the Surface Water Treatment Rule, was achieved. Additional tests examined phage levels through waste water treatment plants in order to estimate their contribution to FRNA coliphage levels in surface waters. Other tests attempted to differentiate RNA and DNA F-specific coliphages as well as somatic Salmonella phages. 4.2 HOST BACTERIA F-specific phages were assayed with a nalidixic acid and kanamycin resistant host bacterium, Salmonella typhimurium strain WG49, which contains an E. coli plasmid (F' 42 lac :: Tn5) responsible for the production of F pili. This host is susceptible to infection by F-specific RNA and DNA coliphages as well as somatic Salmonella phages. Salmonella typhimurium type WG45, which does not contain a plasmid for F-pili production, was also used in several experiments to detect the levels of somatic Salmonella phages. E. coli C3000, which is susceptible to F-specific coliphages as well as somatic coliphages, was also used in the differentiation of phage isolates.

38 Table 4-1 : Water Treatment Plants and their Water Sources Treatment Plant 1. Orange Water and Sewer Authority (OWASA) 2. Pittsboro 3. Durham 4. Raleigh 5. Burlington 6. Winston Salem 7. High Point 8. Greensboro 9. Smithfield 10. Dunn 11. Fayetteville 12. Sanford 13. Kinston 14. Kinston 15. Rocky Mount 16. Tarboro 17. Greenville 18. Eden 19. Fieldcrest 20. Charlotte 21. Danville, VA. Water source* University Lake Haw River Lake Michie Falls Lake Lake Burlington Yadkin River Oak Hollow Lake and City Lake Lake Higgins and Lake Brandt Neuse River Cape Fear River Cape Fear River Cape Fear River Deep Well #9, Black Creek Aquifer Deep Well #I 1, Black Creek Aquifer Tar River Tar River Tar River Dan River Smith River Lake Norman, Catawba River Dan River ' All sources are surface waters, except Kinston, which uses wells. The host bacteria were stored at -70 C prior to use. Overnight cultures were grown to log phase in host media at 37OC on the day of testing. Cultures of host bacteria were stored refrigerated at 4OC or on ice until the samples were analyzed. S. typhimurium type WG49 was grown and assayed in media containing the following ingredients: 1% tryptone, 0.8% NaCI, 0.15% yeast extract, 0.1 % glucose, 0.03% calcium chloride, 0.015% magnesium sulfate, 100 mgll naladixic acid, 100 mgll kanamycin sulfate. Bottom agar for the membrane filtration technique was made by adding 0.9% agar, 0.03% tetrazolium violet and 0.3% polysorbate 80 to the above medium. Double strength agar (2X) was used for the SAL method. S. typhimurium type WG45 and E. coli C3000 were grown in nutrient broth and assayed on nutrient agar, both containing 0.5% NaCI. 4.3 F-SPECIFIC COLIPHAGE ASSAYS Each raw water sample was tested for F-specific coliphages by the membrane filter method (MF) and single agar layer method (SAL) (Grabow and Coubrough 1986). Sample volumes of 100 ml were analyzed by the SAL procedure. Sample volumes ranging from 2 ml to 100 ml

39 were filtered, in triplicate, for the MF procedure. An outline of the basic protocol for the two methods follows Single Agar Layer (SAL) Method. Water sample (100 ml) placed in 37OC water bath for 10 min. Five ml of overnight host bacterium added to sample, mixed and returned to water bath for 3 min. Sample placed in 47' water bath for 3 min. Sample added to equal volume of 2X agar medium, mixed gently, and poured into mm diameter plates. After hardening, plates inverted and incubated at 37 C for 18 hours. Clear plaques counted as F-specific coliphages Membrane Filtration (MF) Method. Two ml of overnight host culture was added to I Oml of molten bottom agar medium and poured into 60 mm diameter plates (4-5 mllplate). MgCI, added to water sample to 0.05 M final concentration. Sample (volume 2-2,000 ml) filtered through 47 mm diameter filters with 0.45 pm pore size (Type HA, Millipore, Bedford, MA.). Filtration rate was 7-10 minutes per 100 ml of sample. Filter removed from holder and placed, face down, on prepoured bottom agar plates containing host bacteria. Plates inverted and incubated at 37OC for 18 hours. Clear plaques counted against violet background. 4.4 BACTERIOLOGICAL ANALYSES The membrane filtration method was used for the detection of E. coli and enterococci in source and treated waters. Testing for E. coli was performed on m-tec medium with confirmation by urease test (Standard Methods 1989). ME agar supplemented with indoxyl B- D-glucoside was used for enumerating enterococci (Dufour 1980). Mixed cellulose nitrate and acetate membrane filters were used for all samples (47 mm diameter, 0.45 pm pore size, GN6, Gelman). Volumes of 5 ml to 100 ml were filtered in triplicate for each raw water sample. E. coli plates were pre-incubated at 37OC for 90 minutes and subsequently placed at 44.5OC for 18 to 22 hours. Filters were then placed on sterile pads containing 2 ml of urea substrate for 15 minutes to score for urease activity. Enterococci plates were incubated at 4I0C for 48 hours. 4.5 COLLECTION OF SAMPLES Raw Water Samples. Samples of two liters were collected from nineteen surface water sources in eastern and central North Carolina, one groundwater source in eastern North Carolina and one surface water source in southern Virginia between July, 1989 and August, These sites included nine municipal water treatment plants in central NC (OWASA,

40 Raleigh, Durham, Burlington, Winston-Salem, Greensboro, Charlotte, High Point and Pittsboro). The remaining sites were chosen after locating wastewater discharge points from maps provided by the North Carolina Department of Environmental Management. Ten of these sites (Fayetteville, Dunn, Smithfield, Rocky Mount, Tarboro, Eden, Fieldcrest, Greenville, Danville and Sanford) draw their water from sources that are, to some degree, impacted by wastewater discharges upstream of the raw water intake. Two additional samples were collected from deep wells, drawing from the Black Creek aquifer, in Kinston, North Carolina. All samples were collected before the addition of disinfectant. At each site two liters of raw water were collected in sterile, one liter polypropylene bottles. Samples were taken from within the treatment plant's laboratory when possible. If permangenate was being added as a pretreatment step, the water was collected before the addition of this chemical. All samples were kept on ice and transported to the investigators' laboratories at the University of North Carolina-Chapel Hill. Each sample was tested within 24 hours of the time of collection Water Quality Data. Additional water quality data were obtained from each water treatment plant for the dates of sample collection. These data, which consisted of turbidity, water temperature, total coliforms, ph and color, were obtained from the monthly reports submitted to the state Department of Environmental Management. Some reports were not available for specific dates and others contained incomplete information. 4.6 REDUCTIONS THROUGH WATER TREATMENT PROCESSES Selection of Water Treatment Plants. Four water treatment plants were chosen from the twenty-one sampled in order to follow the reduction of F-specific coliphages and indicator bacteria through conventional treatment processes. The criteria used for selecting these plants were the levels of F-specific coliphages in the raw water samples and the type of treatment used in the plant. All four of the plants showed consistently high levels of FRNA coliphages in the raw water and no raw water samples were absent of FRNA coliphages. Each plant used a somewhat different scheme in treating its water, but all practiced pretreatment, filtration and disinfection. The four plants chosen were the Hoffer Water Treatment Plant in Fayetteville, the Pittsboro Water Treatment Plant, the Dunn Water Treatment Plant located in Erwin, and the Tar River Plant in Rocky Mount, North Carolina. The Fayetteville plant is a 24 million gallon per day (MGD) facility using the Cape Fear River as its raw water source. The plant uses a conventional treatment scheme consisting of: chemical pre-treatment with alum and powdered activated carbon (PAC) followed by coagulation-flocculation-sedimentation, disinfection, filtration through mixed media filters, storage and post-disinfection with chlorine. The Pittsboro plant is a 1.2 MGD facility which draws its water directly from the Haw River. At the time of this study, the normal treatment process began with pre-disinfection with chlorine followed by alum coagulation-sedimentation. This plant also used mixed media filters and usually post disinfected the filtered water with chlorine before it entered the clearwell. The Dunn Water Treatment Plant also uses the Cape Fear River and normally produces 8 MGD. Raw water is pre-treated with chlorine dioxide prior to alum coagulation. This plant also uses rapid granular media filtration and chlorine as a final disinfectant.

41 The Tar River Water Treatment Plant in Rocky Mount draws from an impoundment on the Tar River. Raw water is first ozonated prior to coagulation with alum and the settled water is disinfected with chlorine before filtration. A second dose of chlorine is added to the finished water before it enters the distribution system. The treatment processes for each of these facilities are shown schematically in Figures 4-1 to Collection of Samples. Samples were collected at four points along the treatment train: raw (intake), settled, filtered and finished water. The volumes analyzed were increased along the treatment train in order to compensate for the reduction in F-specific coliphages and indicator bacteria. Generally one liter of raw water, 15 liters of settled water, and 30 liters of filtered and finished water were collected. This allowed at least 5 liters of settled water for E. coli, enterococci and F-specific coliphage analyses by membrane filtration. Generally, 10 liters of water were tested for each analyte in both the filtered and finished samples. The volume of finished water collected was increased to 70 liters for the last set of samples from three water treatment plants (Pittsboro, Dunn and Fayetteville) so that 50 liters of each could be analyzed for F-specific coliphages. With this sample volume the rate of filtration for F-specific coliphage analysis by the MF method was increased to approximately 3 minutes per 100 ml. Samples were collected on a timed sequence basis such that to some degree the same 'batch' of water was sampled at each point in the plant. Average retention times were calculated from the size of the reactor basins and the flow rate at the time of collection of the raw water sample. These retention times were then used to determine the appropriate times to collect the settled, filtered and finished water samples. Turbidity, ph and chlorine residual (Hach DPD portable kit) data were also recorded as each sample was collected. At each sampling point water was collected in sterile polypropylene bottles to which sodium thiosulfate was previously added (final concentration of 50 mg/l). Samples were kept on ice and transported to the laboratory in Chapel Hill. Analysis was conducted within 24 hours. 4.7 CONFIRMATION OF PHAGE PLAQUES All questionable plaques, by either the MF or SAL method, were confirmed by spotting the picked plaques on plates containing host bacteria. Plaques were removed initially from the agar medium by aspiration with a micro-pipette and resuspended in 0.5 ml of phosphate buffered saline (PBS). The sample was vortex mixed and stored at -70 C. These suspensions were later thawed and 10 pi volumes were spotted onto plates containing bottom agar plus top agar lawns of S. typhimurium WG49 host. These plates were then incubated overnight and the presence or absence of lysis in the sample spots was recorded. This same approach was used to differentiate F-specific coliphages from somatic Salmonella phages. In this case the picked plaques were spotted on nutrient agar plates containing S. typhimurium type WG45. Lysis on these plates was considered positive for somatic Salmonella phages. Lack of lysis on these plates was indicative of an F-specific coliphage. 4.8 RNASE TESTING RNase tests were conducted to determine if phage isolates had either RNA or DNA genomes. Testing was done in two ways: during direct plating of water samples and by spotting previously picked plaques. For the direct plating method, parallel samples of source water

42

43

44 FIGURE 4-3: FLOW DIAGRAM FOR THE DUNN WATER TREATMENT PLANT ClO, Alum NaOH Chlorine NaOH Fluoride N 06 Raw Water F Coagulation I and I I Flocculation I Sediment ation Points Sampled

45

46 were analyzed by either the SAL or MF procedures both with and without RNase in the medium (10 mg RNase per 100 ml double strength agar). RNase was added to the molten agar just prior to mixing with the water sample. The difference between the two counts could be thus attributed to FRNA coliphages, because these phages would grow on plates without RNase and would not grow on plates with RNase. Phage isolates confirmed as F-specific coliphages (negative on plates containing S. typhimurium WG45 and positive on WG49 plates) also were also tested to determine the type of nucleic acid. Previously frozen plaque isolates were thawed and spotted (10 pi each) onto replicate, pre-poured bottom agar plates containing a host bacterial lawn of S. typhimurium WG49. One-half of the plates contained RNase (2.5 mg per 150 mm diameter plate). No lysis in spots on plates with RNAse was considered confirmation of an FRNA coliphage. Lysis on these plates was considered to be due to DNA phages. Isolates giving lysis on plates containing RNase were also spotted onto parent Salmonella typhimurium strain WG45, to differentiate FDNA coliphages from somatic Salmonella phages. Only somatic Salmonella phages and not FDNA coliphages would grow on this host. F-specific RNA phages would be expected to form plaques on S. typhimurium WG49 and on E. coli C3000 in the absence of RNase, because both are male (F+) hosts. F-specific DNA phages should form plaques on S. typhimurium WG49 with and without RNase but not on S. typhimurium WG45, as the latter does not contain the plasmid for sex (F) pili formation. Salmonella phages should form plaques on S. typhimurium WG49 as well as on S. typhimurium WG45 but not on E. coli C WASTEWATER DISCHARGES IMPACTING WATER TREATMENT PLANTS Maps showing the location of all permitted wastewater treatment plant discharges were obtained from the North Carolina Department of Environmental Management (NCDEM). The distance between these discharges and water treatment plant intakes was estimated from these maps. Any animal production or processing facilities with discharge permits were also noted. Average monthly flows as well as monthly mean fecal coliform levels were obtained from reports submitted to the NCDEM by each facility. This information was used to estimate which water treatment plants used source waters impacted by wastewater effluents as well as to compare the levels of F-specific coliphages with respect to the quantity and quality of the upstream discharges. As there are many factors which determine the effect of these discharges (e.g., volume and quality of effluent, size of receiving stream, etc.), a quantitative analysis was determined to be beyond the scope of this investigation. The following criteria were used to categorize water treatment plants as subject to either high or low impact: the number of upstream wastewater discharges, the volume of treated wastewater effluent discharged, the distance to the raw water intake, and if an impoundment is located upstream of the intake. This information was then used to separate the source waters into two categories: those with high wastewater impact and those with low wastewater impact. These two groups were compared for their levels of coliphages and indicator bacteria in source waters.

47 4.10 PHAGE REDUCTIONS THROUGH WASTEWATER TREATMENT PLANTS Studies were done to estimate the effect of sewage treatment on F-specific coliphages and the effects of treated sewage effluent discharges on receiving waters. Samples were collected from the Morehead City wastewater treatment plant. Samples were analyzed for E. coli and enterococci (by membrane filtration), F-specific coliphages and somatic Salmonella bacteriophages. Sodium thiosulfate was added to each sample containing residual chlorine. Several researchers have shown that raw sewage contains high numbers of bacteriophages, with FRNA coliphages greatly outnumbering Salmonella phages (Havelaar 1986, IAWPRC Study Group 1991). In order to evaluate the effectiveness of these phages as a sewage indicator, data was sought regarding the relative reduction of these phages by wastewater treatment processes and the concentration of FRNA coliphages and Salmonella phages in the effluent. Samples were collected at two points in the Morehead City wastewater treatment facility: raw sewage and final effluent. Each sample was tested for F-specific coliphages and somatic Salmonella phages. Raw sewage was tested by a standard double agar layer (DAL) method (Adams 1959). Serial dilutions of the sewage were added in 0.1 ml amounts to 5 ml tubes of molten soft agar (0.75% agar) to which the host bacterium was previously added. The tubes were then poured onto bottom agar plates (0.9% agar) and allowed to harden. Plates were incubated overnight at 37OC. Final effluent samples were tested by the DAL or SAL methods, as previously described. For the SAL method, sample volumes of 100 ml were added to equal volumes of 2X agar medium. Another 100 ml sample, containing 5 ml of S. typhimurium WG45, was added to 2X agar medium to test for somatic Salmonella phages. In order to estimate the levels of FRNA coliphages and Salmonella phages in the sewage, plaques were picked and spotted onto agar medium plates containing S. typhimurium WG45 as well as on FRNA coliphage hosts with and without RNAse in the medium, as previously described.

48

49 5.0 RESULTS 5.1 SURVEY OF WATER TREATMENT PLANT SOURCE WATERS The survey of water treatment plant source waters for the presence of F-specific coliphages was conducted between May, 1989 and August, Raw water samples were collected from a total of 20 water treatment plants in North Carolina and one in southern Virginia. These 21 sites included the principal municipalities in the Piedmont region of North Carolina as well as smaller water treatment plants in eastern North Carolina. Nineteen of the twenty-one sites use surface water and the remaining two were groundwater samples from the same utility but different aquifers (Kinston N.C.). A total of fifty-four samples were analyzed for F-specific phages by the membrane filtration (MF) method and the Single Agar Layer (SAL) method. Each sample was also tested for E. coli and enterococci by membrane filtration. The results of these analyses and the summary statistics are presented in Tables 5-1 and 5-2. Eighteen of the nineteen surface water samples had F-specific coliphage concentrations ranging from 0.3 plaque forming units (PFU)/100 ml to 1267 PFUll 00 ml. One of the nineteen surface water samples (Greensboro, ) as well as both ground water samples (Kinston #9 and #I 1, ) were negative for phages as well as for the bacterial indicators. Generally, F-specific coliphages were found where the bacterial indicators were found and absent in samples containing no E. coli or enterococci. In only one sample (Greenville, ) were F-specific coliphages recovered in the absence of both bacterial indicators (12.5 PFUllOO ml by MF method). Three samples were positive for bacterial indicators but negative for F-specific coliphages (< CFUIml for E. coli and CFU1100 ml for enterococci). As seen in Table 5.2, the mean concentration of E. coli was highest (geometric mean = 27.2 CFUII 00ml) followed by enterococci (mean = CFUll 00 ml) and F-specific coliphages (mean SAL = 6.8 PFUII OOmI, MF = 6.1 PFUII 00 ml). This survey of municipal water supplies demonstrated that F-specific coliphages can be found in a variety of surface waters, sometimes reaching relatively high concentrations when compared to E. coli and enterococci. Furthermore, the presence of these phages appears to coincide with the presence of these two bacterial indicators of fecal pollution. Ground waters drawn from deep wells were free of F-specific coliphages as well as the bacterial indicators. 5.2 RAW WATER PHAGE LEVELS AND WASTEWATER IMPACTS Although it was not possible to determine the sources of F-specific coliphages detected at each sampling site, an attempt was made to examine the levels of phages in source waters in relation to the locations of wastewater treatment plants and other major non-municipal discharges. Other researchers have reported relatively high concentrations of F-specific phages in raw sewage as well as in the feces of some mammals. The North Carolina Department of Environmental Management maintains data pertaining to all permitted wastewater discharges into receiving streams. These data were used to compare, in aqualitative manner, the levels of F-specific coliphages in source waters relative to impacts from upstream wastewater discharges.

50 Table 5-1 : Levels of F-specific coliphages, E. coli and Enterococci in Source Waters. Treatment Bacteria (CFU1100ml) Coliphage (PFUII 00ml) Plant Date E. coli Enterococci SAL' M F~ OWASA OWASA OWASA Pittsboro Pittsboro Pittsboro Pittsboro Pittsboro Pittsboro Pittsboro Pittsboro Pittsboro Durham Raleigh Burlington winston-salem3 Winston-Salem4 winston-salem4 Winston-Salem4 High Point Greensboro Smithfield Smithfield Smithfield Smithfield Dunn Dunn Dunn Dunn Dunn Sanford Sanford Kinston #9 Kinston #I 1 Fayetteville Fayetteville Fayetteville Rocky Mount Rocky Mount Rocky Mount Rocky Mount Eden

51 Table 5.1 : Continued Treatment Plant Bacteria (CFU1100ml) Coliphage (PFUII 00ml) Date E. coli Enterococci SAL' M F~ Eden Eden Fieldcrest Fieldcrest Fieldcrest Charlotte Tarboro Tarboro Greenville G reenville Danville Danville 'SAL = Single Agar Layer Method 2~~ = Membrane Filtration Method 3~homas WTP 4~ielson WTP Table 5-2a: Summary Statistics for Bacterial and Vira I Indicators in Raw Water Samples. F-specific coliphages E. coli Enterococci SAL MF TC* (concentrations per 100 ml) mean (geom.) min c0.33 c0.33 <I.O c0.33 ~0133 max SD No. of Samples * Total coliforms obtained from treatment plant records.

52 Table 5-2b: Summary Statistics for Water Quality Parameters* Temp Turb PH Color ("C) (NW (Units) Mean Minimum Maximum Std. Dev No. of samples * All data collected from water treatment plant records. Table 5-3 shows the permitted wastewater discharges, their average monthly flow, average monthly fecal coliform level, and estimated distance to downstream water treatment plant intakes for the time period August, 1989 through July, Based on this information, water treatment facilities were divided into high and low impact categories depending on the estimated impact of upstream discharges. A more detailed evaluation was not attempted as there are many other factors which would need to be included. In particular, smaller discharges such as from schools and trailer parks, unpermitted discharges and agricultural runoff were not uniformly included in the identification of sources, yet they could impact source water quality. Several other factors such as stream size and flow, rainfall, temperature and sunlight would also need to considered in a more comprehensive evaluation. These and other factors may influence the impact of sewage effluent and other discharges on the microbial quality of receiving waters. Table 5-4 shows the mean values for each of the indicators, as well as turbidity, temperature and ph values for source waters categorized as either high or low impact. Water treatment plants with known effluents from wastewater treatment plants or other sources show higher levels of phages and bacteria (mean concentration for E. coli and F-specific phages = 45.7 CFU1100 ml and 10.0 PFU1100 mi, respectively) than water treatment plants categorized as low impact (mean concentrations of E. coli and F-specific coliphages = 1.45 CFUII 00 ml and 0.7 PFUllOO ml, respectively). For each microbial indicator, the differences between the means for high and low impact sources were statistically significant at the level (oneway ANOVA). 5.3 TESTS OF NORMALITY FOR INDICATOR DATA Prior to conducting other statistical tests between the analytes, the bacterial and phage data were subjected to tests for normality in order to determine whether statistical tests which assume normal distributions could be employed. Each analyte was converted to log,, values

53 Table 5-3: Wastewater Effluent Discharges Impacting Raw Water Sources: Average Monthly Flow, Average Fecal Coliforms and Approximate Distances to Downstream Water Treatment Plants Flow Fecal Col. Approx. MGD (CFUII 00ml) Dist. PllTSBORO - Haw River Burlington South WWTP Burlington East WWTP Graham WWTP Mebane WWTP Greensboro (N.Buffalo) Greensboro (S.Buff alo) Reidsville Sanford - Cape Fear River Sanford WWTP Siler City WWTP Golden Poultry Robbins WWTP Bynum WWTP Pittsboro WWTP Dunn - Cape Fear River Buies Creek WWTP Lillington WWTP Broadway WWTP Fuqua Varina Holly Springs Fayetteville - Cape Fear River Dunn WWTP Erwin WWTP Dominion Textiles Spring Lake WWTP Fort Bragg WWTP Kelly Springfield Inc. Raleigh - Falls Lake Durham Eno. WWTP Durham L Lick WWTP Durham Northside WWTP 40 mi. 30 mi. 40 mi.

54 Table 5-3: Continued Flow Fecal Col. Approx. MD (CFU1100ml) Dist. Smithfield - Neuse River (impoundment) Raleigh Neuse R. WWTP Cary North WWTP Wake Forest WWTP Broadwell Farms River Dell Farms Danville - Dan River Eden Dry Creek WWTP Eden Mebane WWTP Reidsville WWTP Fieldcrest - Smith River Martinsville WWTP Rocky Mount - Tar River Cedar Creek WWTP Louisburg WWTP Bunn WWTP Perdue Inc Spring Hope WWTP (impoundment) NA NA NA NA NA NA NA NA Tarboro - Tar River Rocky Mount WWTP Enfield WWTP Scotland Neck WWTP Greenville - Tar River (impoundment) Tarboro WWTP Pine Tops WWTP Winston Salem - Yadkin River Boonville WWTP Yadkinville WWTP Dobson WWTP El kin WWTP Jonesville WWTP Chatham MFG. - Elkin King Sanitary District Wayne Poultry Purdue Inc.

55 Table 5-3: Continued Flow Fecal Col. Approx. MD (CFU1100ml) Dist. Eden - Dan River (impoundment) Madison WWTP mi No Permitted Upstream Discharges: OWASA - University Lake Burlington - Lake Cammack Charlotte - Lake Norman Kinston - Well Numbers 9 and 1 1 Durham - Lake Michie High Point - Oak Hollow Lake & City Lake Greensboro - Lake Higgins & Lake Brandt Table 5-4: Mean Values For Indicators and Water Quality Parameters in Water Sources Categorized as High and Low Impact by Upstream Sewage ~ischarges* Coliphages Total Turb. Temp E.Coli Enterococci SAL MF colif. (NW ("C) PH High Low Prob.** <.001 c.001 c.001 c.001 c '~igh Impact: Pittsboro, Winston-Salem, Smithfield, Dunn, Sanford, Fayetteville, Rocky Mount, Fieldcrest, Tarboro, Greenville, Danville. Low Impact: OWASA, Durham, Raleigh, Burlington, High Point, Greensboro, Kinston #9 and #I 1, Charlotte, Eden. Categorized as low impact if no upstream discharges. Raleigh categorized as low impact as it draws its raw water from a large lake, has a storage reservoir at the water plant and usually has low levels of coliforms in the raw water. Eden similarly categorized due to a pre-treatment reservoir. "one Way ANOVA F-Statistic

56 and subsequently tested for normality by two methods: graphically, using probability plots, and statistically with a Kolmogorov-Smirnov One-Sample test with a Lilliefors option. The graphs of the probability plots appear in Figures 5-1 and 5-2. The relatively linear appearance of each of the plots supports the assumption of normality. The Kolmogorov-Smirnov test with a Lilliefors option tests for normality without assuming a particular mean or standard deviation for the distribution. This test standardizes the variables and tests whether the standardized versions are normally distributed. Results of this test, with associated probabilities, are shown in Table 5-5. Four indicators, E. coli, enterococci, F- specific phages by MF and F-specific phages by SAL, were shown to be normally distributed with Lilliefors probabilities ranging from to Total coliforms were not normally distributed by this test (p = 0.001). Interestingly, turbidity was found to be normally distributed when transformed into log,, values. Therefore, when comparing turbidity to concentrations of bacteria (E. coli and enterococci) or phages, the log,, turbidity values were used. 5.4 CORRELATION ANALYSIS An important part of this study was to determine the effectiveness of F-specific coliphages as a potential indicator system analogous to the manner in which coliform and other indicator bacteria are presently used. Therefore, the levels of F-specific coliphages were compared to the levels of E. coli and enterococci in North Carolina drinking water sources. At the most basic level, an acceptable indicator should always be present when pathogens are present and absent when pathogens are absent. Although viral pathogens were not analyzed in this study, the two bacterial indicators, E. coli and enterococci, were measured as indicators of fecal contamination and hence the possible presence of pathogens. These two bacterial indicators are considered relatively specific as indicators of fecal contamination but unreliable as indicators of viral pathogens, because they may be less persistent than viruses. However, when these bacteria are present in fecally contaminated water, the presence of viruses is expected. In sewage or fecally contaminated waters we would expect to find high numbers of E. coli, enterococci and F-specific phages if they are indeed good indicators. Conversely, in pristine waters we would expect to find low numbers of these indicators, both bacteria and phages. Thus, the relationship between E. coli, enterococci and F-specific phages should be an important factor in evaluating F-specific coliphages as a potential indicator of fecal pollution. One means of analyzing this relationship is via correlation analysis. The microbial indicators (E. coli, enterococci and F-specific coliphages by MF and SAL methods) were subjected to correlation analysis using log,, values. Plots of these correlations, along with the Pearson's correlation coefficient, appear in Figures 5-3 to 5-5. All correlations were significant at the p = level. F-specific coliphages, as measured by the MF method were well correlated with E. coli (r=0.68) and enterococci (r=o.65). Similar correlation values were observed for F-specific coliphages analyzed by the SAL method (r=0.63 for E. coli and r=0.62 for enterococci). As expected E. coli and enterococci were strongly correlated (r=0.92).

57 Figure 5-1 : NORMAL PROBABILITY PLOTS FOR E. coli and ENTEROCOCCI

58 NCV m N

59 Figure 5-3: CORRELATION PLOTS OF F-SPECIFIC COLIPHAGES BY MF METHOD VS. E. coli AND ENTEROCOCCI AS LOG,, VALUES, WITH PEARSON CORRELATION COEFFICIENTS E. coli (log 0) Enterococci (log 0)

60 FIGURE 5-4: CORRELATION PLOTS OF F-SPECIFIC COLIPHAGES BY SAL METHOD VS. E. coli AND ENTEROCOCCI AS LOG,, VALUES, WITH PEARSON.- CORRELATION COEFFICIENTS Enterococci (Log 0) 44

61 FIGURE 5-5: CORRELATION PLOT OF E. coli VS. ENTEROCOCCI AS LOG,, VALUES, WITH PEARSON CORRELATION COEFFICIENT

62 Table 5-5: Kolmogorov-Smirnov One-Sample Test Using Standard Distribution: Lilliefors Probabilities. Parameter N Probability E. coli Ente rococci SAL Coliphage MF Coliphage Total Coliforms Log,, NTU These results indicate a significant association between F-specific coliphages, by SAL and MF methods, and two of the most specific bacterial indicators of fecal pollution, E. coli and enterococci. This supports the hypothesis that F-specific phages are found typically in surface waters impacted by fecal (human or non-human) wastes, and they are not found typically in the absence of fecal contamination. 5.5 COMPARISON BETWEEN MF AND SAL METHODS As each raw water sample was tested for F-specific coliphages by both the SAL and MF methods, comparisons can also be made between the two methods. The SAL method mixes 100 ml volumes of water sample plus host bacteria with an equal volume of molten agar medium. The MF method adsorbs phages to a membrane filter and the filter is then applied to pre-poured agar medium lawns of the host bacterium. In both methods the phages produce discrete zones of lysis (plaques) in the host cell lawns in the agar medium, which is somewhat analogous to the development of colonies of bacteria. As shown previously in Table 5-2, enumeration of phages via the two methods resulted in similar mean concentrations. The mean concentration using the SAL method was 6.8 PFU1100 ml and for the MF method it was 6.2 PFUI100 ml. Differences between the means were analyzed using two tests: one parametric (paired-sample t test) and one non-parametric (Wilcoxon signed-rank test). As shown in Tables 5-6a and 5-6b, the differences in the means are not statistically significant with either test (p=0.210 by paired t test and p=0.476 by Wilcoxon). By comparison, each of the bacterial indicators tested (E. coli, enterococci and total coliforms) was significantly different from each other and from the coliphages, with p values c0.05 (Tables 5-6a and 5-6b). Correlation analysis between the two phage plating methods resulted in a Pearson's correlation coefficient of 0.86, which was significant at the Pc0.001 level (Figure 5-6).

63 Table 5-6: Tests of Association between Different Indicators Table 5.6a: Paired Sample T-Tests E. coli Enterococci F-specific Coliphages SAL MF En te rococci Coliphage SAL Coliphage MF Total <0.001 Coliforms Table 5.6b: Wilcoxon Signed Rank Tests Enterococci Coliphage SAL Coliphage MF Total Coliforms E. coli Enterococci F-specific Coli p hages SAL MF

64 FIGURE 5-6: CORRELATION PLOT OF F-SPECIFIC COLIPHAGES BY SAL METHOD VS. MF METHOD AS LOG,, VALUES, WITH PEARSON CORRELATION COEFFICIENT5 SAL F-specific Phage (log 0)

65 Correlation analysis also showed that F-specific coliphages detected by the MF method were correlated slightly higher with E. coli (r=0.68) than those detected by the SAL method (r=0.63).similarly, F-specific coliphages by MF were correlated somewhat better with enterococci (r=0.65) than were coliphages by the SAL method (r=0.62). However, these differences are too small to be considered meaningful. These results lead to several observations about F-specific coliphages as fecal indicators and the methods for their detection. First, although recoveries are slightly higher using the SAL method than the MF method, the mean values are not significantly different. Second, the two methods are comparable in terms of recovery efficiency when examining water samples of varying degrees of fecal pollution. The high correlation between the two methods, as well as the small difference in overall means, indicates that the two methods are similar in sensitivity. Both F-specific coliphage plating methods have certain advantages and disadvantages. The SAL method is advantageous because the water sample is mixed directly with the agar medium, so recovery of viable phages is absolute. In contrast, the MF method requires that phages be adsorbed from the filtered water to the membrane filter, which relies on the efficiency of phage adsorption to the filter. In addition, these adsorbed phages must transfer successfully from the filter surface to the agar medium containing the host bacteria. If this transfer is not efficient, phages will not infect host cells. The MF method may be adversely affected by high turbidity waters, in that the turbidity could interfere with phage transfer from the filters to the agar medium host cell lawn. The MF method is advantageous in its simplicity in terms of materials and methodology, especially when large volumes of water are analyzed. This is especially relevant when testing waters with low concentrations of phages, which may require large sample volumes of perhaps hundreds to thousands of milliliters. The method also consumes less agar medium than does the SAL method. The SAL method requires as much medium as the volume of the sample. The MF method requires as little as one plate of medium containing only 4-5 ml, regardless of the sample volume filtered. 5.6 RELATIONSHIPS OF OTHER WATER QUALITY FACTORS TO BACTERIAL AND PHAGE INDICATORS Data for other water quality parameters (turbidity, ph, temperature and color) were collected in order to examine the relationship between these parameters and the ecology of F-specific phages, and also to determine their effect on the recovery of these phages. Data for parameters utilized in this analysis were obtained from the water treatment plant's records or measured directly when the sample was collected. Not all treatment plants record each parameter and some records were not available at the Department of Environmental Management. Correlation analysis was used to determine whether any of these parameters were significantly associated with the levels of each microbial indicator. These results are shown in Table 5-7. Turbidity as log,, NTU was found to be positively correlated with E. coli (r=0.76, pe0.001), enterococci (r=0.64, peo.oo1) and F-specific phages (r=0.38 p=0.015 for SAL, and r=0.40, p=0.01 for MF). Temperature, ph, and color were not significantly correlated with any of the indicators (p>0.05). Further attempts were made to discern associations between these parameters and the microbial indicators.

66 Table 5-7a: Pearson's Correlation Coefficients for E.coli, Enterococci, and F-specific Coliphages by SAL and MF Methods with Other Water Quality Parameters* Log,, Turb Color E. coli 0.76 En terococci 0.64 Coliphage SAL 0.38 Coliphage MF 0.48? ' 40 values for turbidity and temperature, 37 for ph and 18 for color. 5-7b: Probabilities Associated with Correlation Coefficients LOG Turb Temp.(OC) ph Color E. coli c.001 Enterococci <.001 Coliphage SAL Coliphage MF 0.01 Temperature was shown to be negatively correlated with E. coli (r=-0.28)' enterococci (r=- 0.27) and F-specific coliphages by both the SAL (r=-0.04) and MF (r=-0.20) methods although none of the correlations was significant at the 0.05 level. Because of the possibility that temperature may be significantly associated with microbial indicator densities in a relationship that is not a simple linear one, an alternative approach was used. The temperature data were divided into two groups: 'high' and 'low' based on the mean temperature for all samples. These data were then analyzed by a one-way analysis of variance. Temperature was shown to be significant with E. coli (p=0.02) and marginally significant with enterococci (p=0.09) but was not significant with F-specific coliphages by either of the two recovery methods (p=0.87 for SAL and p=0.35 for MF).

67 Virtually no correlation was observed between ph and any of the indicators. Color was positively correlated with each of the indicators although none of the correlations was significant at the p=0.05 level. No further analysis was performed on these parameters. It thus appears that of the four water quality parameters analyzed in this study only turbidity, as log,, NTU, was significantly correlated with each of the indicators. Temperature was negatively correlated as expected, but was less of a factor than turbidity and was significant only with E. coli when the temperature data were dichotomized as high and low values. 5.7 EFFECT OF SAMPLE VOLUME ON F-SPECIFIC COLIPHAGE RECOVERY BY THE MF METHOD For each sample collected and analyzed for F-specific coliphages, the water was filtered in incremental volumes through separate replicate filters in order to obtain plaques within a countable range. In general five-fold different sample volume increments were used (i.e., triplicate 100 ml volumes and triplicate 20 ml volumes). Average phage concentrations were calculated for each volume filtered. In order to evaluate the effect of sample volume on phage recovery, the mean values for F- specific phages were categorized as either large volume or small volume for each sample analyzed in which two separate volumes had countable numbers of phages. The 'large' volume was compared to the 'small' volume to determine if the volume of water filtered had any effect on the recovery of F-specific coliphages. Generally, average phage concentrations were higher when the lower filtered volume was used (data not shown). Mean recovery was 23.1 PFUI100 ml for the small volumes and 13.3 PFUI 100 ml for the large volume. If the phage concentrations from the SAL method are used as the 'gold standard' for F-specific coliphage concentrations (mean = 31.8 PFU1100 ml), then the relative recoveries of the MF method compared to the SAL method were 72.6% for the smaller volume and 41.8% for the larger volume. The reduced phage recovery by the MF method with increasing sample volume suggests that there may be interference due to the accumulation of particulate matter on the filters that prevents phages from being detected. The filters from the larger sample volumes contain more particulate matter, thus increasing the likelihood that phages adsorbed to or covered by particles do not come into contact with the bacteria on the surface of the agar media. In this study the smaller volume was generally used as the source of data for the MF method, as long as none of the three replicates were absent of phage. This suggests that the countable range often used for bacterial analysis (20 to 80 CFUIplate) may not always be applicable to the MF method for the recovery of F-specific coliphages. 5.8 MICROBIAL REDUCTIONS THROUGH WATER TREATMENT PLANTS Four water treatment plants were chosen in order to examine the comparative reductions of each analyte through typical treatment processes. At each plant four samples were collected at different points in the treatment schemes. A total of 11 sets of samples were collected from the four treatment plants: three each from Pittsboro, Fayetteville and Dunn and two sets of samples from the Rocky Mount Water Treatment Plant.

68 Table 5-8 is a composite of the results from all four water treatment plants, and Tables 5-9 through 5-12 summarize these results for each individual water treatment plant. Raw water samples averaged (geometric mean) 845 CFUll for E. coli, 402 CFUll for enterococci and 323 PFUll for F-specific phage by MF. Arithmetic mean raw water turbidity was 34.5 NTU. Following pre-disinfection (8 of 11 samples), coagulation, flocculation and sedimentation, concentrations of E. coli and F-specific phage were below the detection limit in each of the eleven samples using sample volumes of 1 to 10 liters. Enterococci were detected in three of the samples (mean = 1.9 CFUlI). Filtered and finished water samples were negative for E. coli and phages in all eleven samples but positive for enterococci in one sample at 0.2 CFUII. In eight of the eleven samples raw water is disinfected (chlorine, chlorine dioxide or ozone) prior to coagulation. The three samples from Fayetteville were not disinfected until the end of the sedimentation basins and just prior to filtration. Table 5-8: Indicator and Turbidity Levels in Source Waters and Log,, Reductions by Water Treatment Processes: Composite of 11 Samples from Four Treatment Plants. Turb (NTU) E. Coli Enterococci Coliphages Raw Arith. Mean ,122 1,725 1,518 Water Geom. Mean Logln Reductions Turb (NTU) E. coli Enterococci Coliphages Settled 1.06 >2.92 >2.52 >2.72 Filtered >3.22 >3.00 >3.10 Finished >3.30 >3.04 >3.30 As can be seen from the results in Tables 5-8 to 5-12, reductions of all analytes at each treatment plant exceeded the reductions in turbidity. The results summarized in Table 5-8 indicate average reductions of E. coli and enterococci of 3.3-log,, and 3.0-log,,, respectively, from raw to finished water. F-specific coliphage reductions averaged 3.3-log,,. Overall turbidity reductions averaged 2.15 log,, over the eleven samples. These represent conservative estimates expressed as "greater than" values, because in nearly all cases the extent of reduction was based on the detection limit of the sample, which

69 Table 5-9: Levels Of Indicator Organisms and Log,, Reductions in Water Treatment Plants: Fayetteville Plant Turb. Phage E. coli Entero FAC* Treatment NTU PFUIL (CFUIL) (CFUIL) Ow-) Raw Log,, Reductions Settled >2.7 >4.3 > Filtered >3.4 >5.0 > Finished >3.7 > Raw Settled >3.0 > Filtered >3.0 >4.0 > Finished >3.0 >4.0 > Raw Settled >3.7 >4.0 >4.3 Filtered >3.7 >4.0 > Finished >4.5 >4.0 > Composite for Fayetteville Plant Raw Settled >3.1 >4.0 >3.7 Filtered >3.3 >4.3 >4.1 Finished >3.7 >4.4 >4.3 *FAC - Free Available Chlorine Residual Concentration

70 Table 5-10: Levels Of Indicator Organisms and Log,, Reductions in Water Treatment Plants: Rocky Mount Plant Turb. Phage E. coli Entero FAC* Treatment NTU PFUIL (C FUIL) (C FUIL) (mg/l) Raw Logln Reductions Settled >2.3 >4.0 > Filtered >3.0 >5.0 > Finished >3.3 >4.5 >3.7 1.O Raw Settled >2.4 >2.3 > Filtered >2.7 >2.5 > Finished >2.7 >2.5 > Composite for Rocky Mount Plant Raw Settled >2.35 >3.15 >2.75 Filtered >2.85 >3.25 >2.90 Finished >3.0 >3.5 >3.05 "FAC - Free Available Chlorine Residual Concentration

71 Table : Levels Of Indicator Organisms and Log,, Reductions in Water Treatment Plants: Dunn Plant Turb. Phage E. coli Entero FAG* Treatment NTU PFUIL (CFUIL) (CFUIL) (mg/l) Raw Log,, Reductions Settled >3.2 >3.5 > Filtered >3.2 >3.5 > Finished >3.5 >3.7 >3.3 1.O Raw 21.O Settled >3.0 >2.9 > Filtered >3.4 >3.0 > Finished >3.4 >3.0 > Raw Settled >2.4 >2.7 > Filtered >2.8 >3.4 > Finished >3.5 >3.4 > Composite for Dunn Plant Raw Settled >2.90 >3.0 >2.6 Filtered >3.1 >3.3 >2.8 Finished >3.4 >3.4 >2.9 *FAC - Free Available Chlorine Residual Concentration

72 Table 5-12: Levels Of Indicator Organisms and Log,, Reductions in Water Treatment Plants: Pittsboro Plant Turb. Phage E. coli Entero FAC* Treatment NTU PFUIL (C FUIL) (CFUIL) (mgw Raw ND Log,, Reductions Settled ND ND >3.2 > Filtered ND ND >3.7 >3.2 > Finished ND ND >3.7 >3.4 > Raw Settled >4.7 >4.5 > Filtered >5.0 >4.7 > Finished >5.0 >4.7 > Raw Settled >2.5 >3.5 > Filtered >2.8 >3.7 > Finished >3.5 >3.7 > Composite for Pittsboro Plant Raw Settled >3.4 >3.4 >2.9 Filtered >3.8 >3.9 >3.6 Finished >4.0 >3.9 >3.6 *FAC - Free Available Chlorine Residual Concentration

73 was a function of the volume filtered. Had larger volumes been filtered, the reductions would likely be greater than those shown in the tables. F-specific phages were reduced by greater than 4 logs,, in two of the eleven samples and greater than 5 logs,, in one of the samples. For E. coli and enterococci a 4-log,, reduction was achieved in and cases, respectively. Once again, these calculated log,, reductions are limited by the volume filtered. In each of the last three sets of samples (third set from Fayetteville, Pittsboro and Dunn) the volume of finished water that was analyzed was increased from 10 liters to 50 liters for F- specific coliphages. This larger volume resulted in a 4.5 log,, reduction of F-specific coliphages for the Fayetteville sample. Levels of F-specific coliphage reductions were not as great for the other two plants, despite the filtration of larger sample volumes. The Pittsboro and Dunn samples were collected during the summer months, after relatively dry, warm weather. Under these conditions the raw waters contained low levels of phages and, consequently, four log,, reductions were not detectable. When examining larger volumes of finished water, it becomes possible in some water supplies to demonstrate a 99.99% (4 log,,) F-specific coliphage reduction between raw and finished waters. If 50 liters of finished water are analyzed, a 4-log reduction can be demonstrated if the raw water concentration of F-specific coliphages is at least 200 PFUII. Of the eleven samples included in this study, 7 had raw water concentrations of at least 200 PFUIL. When all source water samples of this study are considered, 33% (18154) are shown to have this level of F-specific coliphages. If larger sized filters are used for membrane filtration of F- specific coliphages it becomes feasible to further increase the volume of finished water analyzed, and thereby demonstrate 4-log,, reductions for raw source waters having concentrations less than 20 PFUII. Results from analyses of these eleven water treatment plants indicate that F-specific phages are efficiently removed by conventional pre-treatment, filtration and disinfection processes. The absence of detectable F-specific coliphages in treated water samples indicates the efficacy of the coagulation-flocculation-sedimentation process in reducing viruses, especially when combined with a pre-disinfection step. Even in the three Fayetteville samples, with no pre-disinfection, >99.9% phage reductions were achieved by the coagulation-sedimentation process alone, while turbidity reductions were <99%. The average turbidity reduction by coagulation-sedimentation for this plant (96.8%), was greater than that of the three other plants (85.0%, 92.8% and 90.9% for Pittsboro, Rocky Mount and Dunn respectively). The addition of powdered activated carbon may also have contributed to the efficiency of phage removal by the coagulation-sedimentation process at the Fayetteville plant. It should be noted that water treatment facilities in the study area treat their raw waters with a relatively high concentration of alum, which may be responsible for the large phage reductions during the coagulation-sedimentation process. The presence of enterococci in three samples of settled water, in the absence of the other analytes, indicates that this bacterium may not be removed as efficiently by the coagulationsedimentation process as either E. coli or F-specific coliphages. The reasons for this apparently greater persistence were not investigated in this study. It is possible that enterococci may be coagulated less readily by alum than either E. coli or F-specific coliphages

74 and thus be reduced to a lesser extent. In addition, the presence of enterococci in at least one finished water sample may indicate a greater resistance to chlorine disinfection under the prevailing conditions in this water. 5.9 TESTS FOR RNASE SENSITIVITY AND SALMONELLA PHAGES The host bacterium used in this study for the detection of F-specific coliphages, S. typhimurium type WG49, is potentially susceptible to three types of phages: F-specific RNA coliphages, F-specific DNA coliphages and somatic Salmonella phages. Of these three types of phages, FRNA coliphages generally predominate in raw sewage, as previously mentioned. Less is known, however, regarding the levels of each of these types of phages in wastewater effluents and surface waters. In order to determine the proportion of F-specific phages that are FRNA phages, parallel assays were performed with and without RNase. Plaques forming in the absence of RNase could be FRNA phages, FDNA phages, or somatic Salmonella phages. Plaque formation in the presence of RNase could be either FDNA or Salmonella phages but not FRNA phages. Table 5-13 shows the results of initial tests on phage characterization by RNase testing for phages in 8 water samples obtained from five different source waters and at different times of the year. The proportion of FRNA coliphages varied greatly. In 5 of 8 samples a majority of the phages appear to be FRNA coliphages, and in the other 3 samples only 4174, 27.2% and 0% of the total phages were FRNA coliphages. In a subsequent series of experiments, plaque isolates from both SAL and MF plates were resuspended in PBS and spotted onto plates containing 1X agar medium with one of the following bacterial hosts and RNase treatments: S. typhimurium WG49 with RNase, WG49 without RNase and S. typhimurium WG45 (not susceptible to F-specific coliphages). Positive and negative controls were also included on each plate. As shown by the results of these tests (Table 5-14), the proportions of the different phages varied among water sources. In several samples a majority of the phages are FRNA coliphages, but in 9 of 12 samples Salmonella phages predominate and there are few if any FRNA coliphages. Some samples had mixtures of all three types of phages: FRNA coliphages, FDNA coliphages and somatic Salmonella phages. Thus, in contrast to raw sewage, where the majority of phages are FRNA coliphages, in natural surface waters there are variable concentrations and proportions of both F-specific RNA and DNA coliphages, as well as somatic Salmonella phages F-SPECIFIC COLIPHAGES IN WASTEWATER Wastewater treatment plant discharges may contribute appreciable numbers of coliphages to receiving waters that are subsequently used as source water for drinking water treatment plants, even if they meet coliform or fecal coliform bacteria discharge requirements. Therefore, information was sought on the quantities and types of phages in the discharges. Several experiments were done to isolate, assay and characterize phages in samples from the Morehead City wastewater treatment plant. The plant employs conventional primary and secondary (activated sludge and trickling filtration) treatment and chlorination. Data on the levels of F+ coliphages, somatic Salmonella phages and indicator bacteria in samples of raw sewage and treated effluent collected monthly from the plant for a year are summarized in Table F+ coliphages were detected in raw sewage at high concentrations, with a

75 Table 5-13: Results of Parallel Assays of Surface Source Waters for F-Specific Coliphages with and without ~Nase* Present Coliphages (PFUII 00ml) %RNase No RNase + RNase Sensitive Water Plant Date SAL MF SAL MF SAL MF Pittsboro G reenville Fayetteville Dunn Pittsboro Dunn Pittsboro Danville '10 mg RNase added to 100 ml 2X agar for SAL method 10 mg RNase added to 100 ml 1 X agar (0.5 mglplate) for MF method. Table 5-1 4: Proportions of F-specific, FRNA and Somatic Salmonella Phages in Various Surface Source Waters Source No. of No. F- No. (%) No. (%) Plaques specif. FRNA* Salmonellatt Rocky Mountrrarboro Multiple sources TarborolG reenville Multiple sources Pittsboro Fayetteville Dunn Pittsboro Smithfield Danville, Fieldcrest Pittsboro Pittsboro TOTALS (71) '~rew on S. typhimurium WG49 but not in presence of RNAse. "~rew on S. typhimurium WG45 and in the presence of RNAse

76 geometric mean concentration of 2.0 x 1 o6 PFU per 100 ml. F+ coliphages also were readily detected in treated sewage effluents, with an average concentration of 1.3 x lo4 PFUllOO ml. Therefore, the average F+ coliphage reduction by sewage treatment was 99.35% or about 2.2 log,,. Somatic Salmonella phage levels in raw sewage and treated effluent were geometric means of 5,600 and 54, respectively, both of which are considerably lower than the levels of F+ coliphages. The mean reduction of somatic Salmonella phages by sewage treatment was 99% or 2.0 log,,, which is somewhat lower than the reduction of F+ coliphages. The bacterial indicators of fecal coliforms, E. coli and enterococci were reduced much more extensively by sewage treatment ( log,, or %). These results support previous findings showing that F+ coliphages are more resistant to conventional primary-secondary sewage treatment (plus chlorination) than are indicator bacteria. Table 5-15: F-Specific Coliphages, Somatic Salmonella Phages and Indicator Bacteria in Raw Sewage and Treated Effluent of the Morehead City Sewage Treatment Plant Organism Geom. Mean Range F+ Coliphages Somatic Salmonella Phages Fecal Coliforms E. coli Enterococci Raw Sewaqe F+ Coliphages Somatic Salmonella Phages Fecal Coliforms E. coli Enterococci Treated Sewaqe Effluent In order to determine what proportion of phages detected on S. typhimurium WG49 were F- specific RNA coliphages, a selected number of phage plaques from Morehead City raw sewage and treated sewage effluent were isolated from plates and cross-plated on S. typhimurium WG45 to determine if they were somatic Salmonella phages. Those that were truly F-specific coliphages were tested for RNAse sensitivity. As shown by the results summarized in Table 5-16, 89% of the phages in raw sewage and 99% of the phages in treated effluent were confirmed F+ coliphages. Of those that were F+ coliphages, 74% of those in raw sewage were FRNA coliphages and 97% of those in secondary effluent were FRNA coliphages. Thus, most of the phages in raw sewage and treated effluent that were detected on host S. typhimurium WG49 were FRNA coliphages.

77 Table 5-16: characterization of Phage Isolates from Raw Sewage and Treated Sewage Effluent on Host S. fyphimurium WG Sample No. Tested % F+ Coliphages %FRNA' Raw Sewage Treated Effl. Percent of total F+ coliphages that were RNAse sensitive.

78

79 6.0 DISCUSSION The evaluation and possible use of F-specific coliphages as indicators of fecal pollution and as model organisms for human enteric viruses in source and drinking waters have been hindered by the lack of a simple methodology appropriate to larger volumes of water and the lack of a suitable host bacterium specific only to these phages. The modified Salmonella typhimurium strain WG49, developed by Havelaar and Hogeboom (1984), has been shown to be an acceptable host for the detection of F-specific coliphages in sewage and feces. Less is known, however, regarding the applicability of this host to natural and treated waters, but limited studies have provided somewhat encouraging results in both fresh and saline waters (Handzel et al. 1993, Havelaar et al. 1993, IAWPRC Study Group 1991, Sobsey et al. 1990). The development of a direct MF method for concentrating and enumerating F+ coliphages permits the routine sampling of natural and treated waters for them (Sobsey et al. 1990). This study was designed to assess the appropriateness of the MF method for the examination of source waters and treated drinking water, as well as to evaluate the applicability of F-specific coliphages as an indicator of fecal and possibly viral pollution. Field testing of source waters from a variety of locations, by both the MF and SAL methods, resulted in comparable recoveries for both methods. All samples positive by the SAL method were also positive by the MF method. Three samples negative by SAL were positive by MF and each of these was positive for both types of indicator bacteria. This finding is likely a result of the larger volume analyzed by the MF method (300 ml) than by the SAL method (1 00 ml). Mean F-specific coliphage levels for all 54 samples were comparable by the two methods. When analyzed by parametric and non-parametric statistics, the two methods did not differ significantly in the detection of phages. The two methods were also highly correlated with each other on the basis of phage concentrations (r=0.86). When compared with the bacterial indicators, E. coli and enterococci, both the MF and SAL methods showed positive correlations, although the MF method showed slightly higher correlation coefficients than did the SAL method. The sample volume filtered in the MF method appeared to influence phage recovery efficiency. The filtration of lower sample volumes per filter tended to give higher mean recoveries than the larger sample volumes. This may be a consequence of the accumulation of particulate matter on the filters, which prevents contact between the adsorbed phage particles and the bacterial lawn. This particulate effect could be reduced by using larger diameter filters or replicate filtration of smaller sample volumes. The MF method appears to be applicable to the assessment of fecal contamination of source and drinking waters of various qualities. It permits the analysis of larger volumes than possible by the SAL method and is simple enough for routine use in modestly equipped laboratories. Up to 2 liters of finished water was filtered through 47 mm diameter filters with little discoloration or clogging of the filter. In addition to the plating methods, the validity of F+ coliphage assay procedures to evaluate source waters for fecal contamination depends on the host bacterium that is employed. This study used a modified strain of S. typhimurium which is susceptible to infection by the F-

80 specific coliphages as well as somatic Salmonella phages. The reliability of this host bacterium to detect F-specific coliphages as indicators of fecal contamination was compared with two specific bacterial 'indicators of fecal contamination, E. coli and enterococci. Most of the 54 source water samples tested were either positive for both types of indicators, bacteria and phages, or negative for both. Only one sample was negative for both bacterial indicators and positive for phages. Three samples were negative for bacteriophages and positive for one of the bacterial indicators. The good correlations found between the phages and either E. coli (r=0.68) or enterococci (r=0.63) suggest that these phages are an adequate indicator of fecal contamination for a variety of source waters. Furthermore, greater numbers of F-specific coliphages were recovered from source waters known to be impacted by wastewater effluent discharges than in source waters with few or no known discharges. An important advantage of using phages as an indicator of fecal pollution is that results can be seen in as little as six hours as compared to 24 hours for E. coli and coliforms and 48 hours for enterococci. Another possible advantage is that F-specific RNA phages can be serotyped and, based on serotype prevalence, it may be possible to determine whether they originated in human or non-human feces (Furuse 1987). However, this aspect of FRNA coliphages as indicators was not investigated in this study, and further research is needed. A practical method of detecting F-specific coliphages in large volumes of natural waters could have wide applicability if there is verification that F-specific coliphages are reliable model organisms for the enteric viruses. Previous research has shown that they fulfill many of the characteristics of a virus indicator organism such as: presence in greater numbers than enteric viruses, similar or greater persistence in the environment than enteric viruses, and similar or greater resistance to treatment processes than enteric viruses (Havelaar et al. 1993, IAWPRC Study Group 1991, Payment and Franco 1993). The results of this study demonstrated that the reduction of these phages through drinking water treatment plants can be followed for at least several orders of magnitude of virus reduction, depending on the initial level of viruses in the raw water. Of the four water treatment plants studied, greater than 4 log,, reduction of phages was observed in two of the plants, and greater than 3 log,, reduction was seen in the two other plants. Average log,, reductions of organisms for all eleven trials at the four water treatment plants, were 3.3 for E. coli, 3.0 for enterococci and 3.3 for F+ coliphages. All three microbial indicators (F-specific coliphages, E. coli and enterococci) surpassed reductions of turbidity, which averaged log,, reduction. These results agree with those of previous studies, which show reductions of bacteriophages surpassing that of turbidity (York and Drewry 1974). Guy and Mclver (1 977) reported reductions of 92.6% for indigenous somatic coliphages and 99.7% for added T4 (somatic) coliphages. Other researchers have reported reductions of the F-specific coliphage MS-2 by chemical coagulation and sedimentation to be 2-3 log,, (Chaudhuri and Engelbrecht 1970). In our study, only the Fayetteville treatment plant did not use a pre-disinfection step. All three samples, nonetheless, achieved at least a 99% level of reduction for F-specific coliphages. This plant does add powdered activated carbon to the raw water, however, and usually

81 achieves significant turbidity reduction. Previous studies have shown that activated carbon efficiently adsorbs viruses.(gerba et al. 1975). In order to demonstrate a 4-log,, reduction of bacteriophages, an initial concentration of 100 PFUllOO ml is required, if 10 liters of finished water is tested. Two of eleven samples in this study had raw water concentrations of at least 100 PFUllOO ml. If 50 liters of finished water is analyzed, the initial concentration required for documenting a 4-log,, reduction is ml. In this study seven of eleven raw water samples equaled or surpassed this concentration of F+ coliphages. Many of the sites sampled in this study were chosen for their perceived higher levels of fecal indicators. This implies that the levels found in source waters in this study are not necessarily representative of all source waters in this region. Thus, the levels of phage reductions seen in these water treatment plants may not be achievable in other facilities where the levels of fecal contamination and hence fecal indicators in the source waters are much lower. This study does indicate, however, that a significant number of surface waters, especially those drawing directly from streams and rivers, have levels of F-specific coliphages high enough to document a phage reduction of at least several orders of magnitude (log,,). Turbidity as expressed as a log,, value was shown to be positively correlated with F-specific coliphages. Higher turbidity waters were generally seen following periods of high rainfall and subsequent runoff. This observation suggests that appreciable quantities of these phages originated in sources such as animal waste runoff from nearby farms and increased inputs from wastewater treatment facilities that were either bypassing sewage or treating it less effectively. Most of the water plants sampled have watersheds that include agricultural land and are thus affected by farm runoff, and many of the plants were impacted by municipal wastewater effluent discharges. During periods of heavy rainfall the efficiency of wastewater treatment plants may be reduced due to the influx of storm waters. Under these circumstances bacteriophages as well as enteric viruses may be released into receiving waters at higher levels. At least one water treatment plant included in this study (Tarboro) is impacted by a wastewater plant (Rocky Mount WWTP) which, under heavy flow conditions, bypasses a portion of the wastewater directly into the receiving stream. In addition, it is possible that phages entering source waters at higher levels during periods of heavy rainfall are subsequently protected from adverse environmental effects by their association with particulates. Bacteriophages adsorbed to particulates are more resistant to environmental factors such as temperature, ultraviolet and visible light, inorganic ions and metabolic products of other microorganisms (Farrah 1987, Gerba and Schaiberger 1975). During low flow, low turbidity periods, viruses entering receiving waters are exposed to sunlight to a greater extent and are more likely to be inactivated. Similarly, attachment to particulates may stabilize the virus against temperature and other adverse effects. The affinity of viruses for particulate matter may also mean that during normal or low flow conditions, particle associated viruses settle out and accumulate in bottom sediments. During high flow periods these settled viruses may be resuspended by the increased flow and carried downstream to water treatment plant intakes. Solids-associated viruses in waste water effluent can accumulate in stream bottom sediments and reach concentrations of 10 to 10,000 times that of the water (Melnick 1987).

82 Temperature is considered to be the most important factor influencing the survival of viruses in the environment (Bitton 1987). Higher temperatures tend to increase inactivation of viruses. In this study temperature did not appear to be a significant factor in the detection of F-specific bacteriophages. Temperature was statistically significant only in its association with E. coli when dichotomized as a variable into either high or low temperature. Although the data were not analyzed for seasonality, repeat samples from the same plant do show lower recoveries during the summer months of June, July and August than during other months of the year. This may be due to higher temperatures during these months, lower rainfall or a combination of the two. Other researchers have shown similar seasonal distributions of viruses and bacteriophages, with higher recoveries generally during the fall and winter months (Simkova and Cervenka 1981, Kott et al. 1974, Stetler et al. 1984). The presence of readily detectable levels of somatic Salmonella phages in these waters was unexpected and complicates the applicability of the S. fyphimurium WG49 host bacterium to specifically monitor for F-specific coliphage levels in source waters. As the sources of these phages are uncertain, their presence may not be related to the presence of the human enteric viruses and may interfere with efforts to establish the association between FRNA coliphages and human enteric viruses. The relatively low numbers of Salmonella phages in the waste water effluent samples examined suggest that most of those detected in surface waters may not always be a result of sewage treatment plant discharges. Two other plausible sources of these phages are: (i) The fecal wastes of agricultural production facilities that reach receiving waters during storm runoff, and (ii) fecal wastes associated with wildlife sources, such as feral mammals and waterfowl. Most of the source waters sampled in this study are located in rural agricultural regions and are impacted by wastes from animal production facilities as well as wildlife. Rhodes and Kator (1991) found relatively high proportions of somatic Salmonella phages in stream and estuarine waters and sediments when they used host S. typhimurim WG49. They attributed these phages to non-point sources. The FRNA coliphages have been characterized on the basis of morphology, composition, biochemistry, antigenicity and genetics (genomic organization and nucleotide sequences). Superficially, they resemble some of the human enteric viruses in terms of their general features and their resistance to environmental factors and treatment processes, such as disinfection. Little is known, in this regard, about the somatic Salmonella phages. Thus, they may over- or under-estimate the actual level of human enteric viruses and give a false estimation of the viral quality of source and drinking waters. The presence of Salmonella phages also raises the possibility of the presence of Salmonella bacteria in these waters, which were not analyzed in this study. It is also possible that these somatic bacteriophages are not Salmonella phages specifically, but instead originate in host species of related enterobacteriaceae. Bradley (1 967) has shown that cross reactivity occurs between phages of closely related species such as Shigella and E. coli. These could, in fact, be somatic phages of enterobacteriaceae that are reactive with this particular strain of Salmonella typhimurium. Several of the phage isolates that were further characterized by spot testing on different host bacteria did, in fact, form plaques on both S. typhimurium type WG45 and E. coli C3000. Further examination of other water sources could determine whether somatic Salmonella phages are found in all surface waters during all seasons of the year. If so, this interference by somatic Salmonella phages would complicate the applicability of the S. typhimurium WG49

83 host to detect F-specific coliphages as model organisms for enteric viruses. A Salmonella strain that is less receptive to somatic phages may be necessary. Mutant strains of Salmonella that have alterations in the lipopolysaccharide portion of the cell wall may be candidates as an improved host. Similarly, mutant strains could be selected from surviving cells after repeated exposure to these somatic phages. Another approach to controlling the growth of somatic Salmonella phages during the isolation of F-specific coliphages is to add heat-killed Salmonella typhimurium cells or their isolated lipopolysaccharide (LPS) to the sample or the medium during analysis of the samples. This results in the somatic Salmonella phages adsorbing to the heat-killed cells or LPS, thereby preventing infection of the's. typhimurium WG49 host. We have reported on the successful development and application of this approach (Handzel et al. 1993). Our studies on the examination of raw and treated sewage provided results similar to those of other researchers, with FRNA coliphage levels of approximately lo3 to lo4 PFUIml and much lower somatic Salmonella phage levels of about PFUlml. These results agree with studies by Havelaar (1986), Ketratanakul and Ohgaki (1989) and Dhillon and Dhillon (1 974). F+ coliphages and somatic Salmonella phages were reduced by conventional primary and secondary sewage treatment plus chlorine disinfection. These combined treatment processes produced typical reductions of approximately 2 log,, or 99% as compared to reductions of indicator bacteria (fecal coliforms, E. coli and enterococci) of log,, or %. The concentrations of F-specific coliphages remaining in sewage effluents to be discharged into receiving waters varied somewhat, but they were generally in the range of a few hundred to a few thousand per I00 ml. These data suggest that sewage treatment plants can contribute F-specific coliphages to receiving waters that are used as source waters by drinking water treatment plants. It is likely that phages released during high flow conditions in sewage treatment plants also coincide with high turbidities in receiving waters, which may offer viruses greater protection from inactivation. This may also help explain the association between turbidity of source waters and the occurrence of coliphages. Under these same circumstances, the contribution of phages from land runoff would also be highest, which may explain the variable counts of somatic Salmonella phages. The rather consistent ratio of F-specific coliphages to Salmonella phages found in sewage samples does not seem to occur in natural waters in North Carolina. The variable but sometimes high levels of Salmonella phages present in source waters may interfere with the ability to establish an association between F-specific RNA coliphages and the human enteric viruses. However, methods to control somatic Salmonella phage detection during F-specific coliphage analysis have been developed, so this problem can be controlled. F-specific coliphages appear to have considerable potential to serve as indicators of fecal contamination and perhaps as indicators of human enteric viruses in water. Considerable quantities of F-specific coliphages could be detected regularly in surface sources of drinking water supplies in North Carolina. This establishes that F-specific coliphages occur at sufficient frequency and concentration to be useful indicators of the microbial quality of drinking water, if the other essential criteria of a fecal and enteric virus indicator can be met.

84

85 7.0 REFERENCES Adams, M.H ~acteriophages. Interscience, New York. Belabbes, E., A. Bouguermouh, A. Benatallah, and G. Illoul Epidemic non-a, non-b Viral Hepatitis in Algeria: Strong Evidence for its Spreading by Water. Journal of Medical Virology 16: Bell, R.G The Limitation of the Ratio of Fecal Coliforms to Total Coliphage as a Water Pollution Index. Water Research 10: Birge, E.A Bacterial and Bacteriophage Genetics. Springer-Verlag, New York. Bitton, G Fate of Bacteriophages in Water and Wastewater Treatment Plants. In: Phage Ecology. Goyal, S., Gerba, C.P., Bitton, G., Eds., New York, John Wiley and Sons. Blacklow, N.R., and G. Cukor Viral Gastroenteritis, New England Journal of Medicine 305:397. Bradley, D.E Ultrastructure of Bacteriophages and Bacteriocins. Bacteriological Review 31 :230. Castillo, G.R., R. Thiers, B.J. Dutka, and A.H. El-Shaarawi Coliphage Association with Coliform Indicators: A Case Study in Chile. Toxicity Assessment: An International Journal 3: Chaudhuri, M., and R.S. Engelbrecht Removal of Viruses from Water by Chemical Coagulation and Flocculation. Journal of the American Water Works Association 62:563. Clarke, N.A., G. Berg, P.W. Kabler, and S.L. Chang Human Enteric Viruses in Water: Source, Survival and Removability. In: Advances in Water Pollution Research, Jenkins, S.H., Ed., vol. 2, pp. 523, New York, Pergamon Press. Cliver, D.O Manual on Food Virologv. World Health Organization, Geneva. Cramer, W.N., K. Kawata, and C.W. Kruse Chlorination and lodination of Poliovirus and f2. Journal of the Water Pollution Control Federation 48:61. Craun, G.F. 1986a. Recent Statistics of Waterborne Disease Out-breaks ( ). In: Waterborne Diseases in the United States. Craun, G.F., Ed., Ch. 6, Boca Raton, FL, CRC Press. Craun, G.F Statistics of Waterborne Disease Outbreaks in the U.S. ( ). In: Waterborne Diseases in the United States. Craun, G.F., Ed., Ch. 5, Boca Raton, FL, CRC Press. Craun, G.F., February Surface Water Supplies and Health. Journal American Water Works Association 80(2):40.

86 Deetz, T.R., E.M. Smith, S.M. Goyal, C.P. Gerba, J.J. Vollet Ill, L. Tsai, H.L. Dupont, and B.H. Keswick Occurrence of Rota- and Enteroviruses in Drinking and Environmental Water in a eve loping Nation. Water Research 18(5): Dhillon, E.K.S., and T.S. Dhillon Synthesis of Indicator Strains and Density of Ribonucleic Acid Containing Coliphages in Sewage. Applied Microbiology 27:640. Dhillon, T.S., Y.S. Chang, S.M. Sun, and W.S. Chau Distribution of Coliphages in Hong Kong Sewage. Applied Microbiology 20: Dhillon, T.S., E.K.S. Dhillon, H.C. Chau, W.K. Li, and A.H. Tsang Studies on Bacteriophage Distribution: Virulent and Temperate Bacteriophage Content of Mammalian Feces. Applied and Environmental Microbiology 32: Dufour, A.P A 24-Hour Membrane Filter Procedure for Enumerating Enterococci. Abstracts of the Annual Meetinq. American Society for Microbiology. El-Abagy, M.M., B.J. Dutka, and M. Kamel Incidence of Coliphage in Potable Water Supplies. Applied and Environmental Microbiology 54(6): Fair, G.M., J.C. Geyer, and D.A. Okun Water and Wastewater Engineering. John Wiley and Sons, Inc., New York. Farrah, S.R Ecology of Phage in Freshwater Environments. In: Phage Ecology. Goyal, S.M., Gerba, C.P., Bitton, G., New York, John Wiley and Sons. Finch, G.R. and N. Fairbairn Comparative Inactivation of Poliovirus Type 3 and MS2 Coliphage in Demand-free Phosphate Buffer by Using Ozone. Applied and Environmental Microbiology 57: Fujioka, R.S Naturally Occurring Fecal Coliform and Fecal Streptococci in Hawaii's Freshwater Streams. Toxicity Assessment 3: Furuse, K., A. Ando, S. Osawa, and P. Watanabe Distribution of Ribonucleic Acid Coliphages in Raw Sewage from Treatment Plants in Japan. Applied and Environmental Microbiology 41 :I 139. Furuse, K., S. Osawa, J. Kawashiro, R. Tanaka, A. Ozawa, S. Sawamura, Y. Yanagawa, T. Nagao, and I. Watanabe. 1983a. Bacteriophage Distribution in human Feces: Continuous Survey of Health Subjects and Patients with Internal and Leukemic Diseases. Journal of General Virology 64: Furuse, K., T. Sakurai, A. Hirashima, M. Katsuki, A. Ando, and I. Watanabe Distribution of Ribonucleic Acid Coliphages in South and East Asia. Applied and Environmental Microbiology Furuse, K., T. Sakurai, Y. Inokushi, H. Inoko, A. Ando, and I. Watanabe. 1983b. Distribution of RNA Coliphages in Senegal, Ghana and Madagascar. Microbiology and Immunology 27(4):

87 Furuse, K Distribution of Coliphages in the Environment: General considerations. In: Phage Ecoloqv. Goyal, S.M., Gerba, C.P., Bitton, G. New York, John Wiley and Sons. Geldenhuys, J.C., and P.D. Pretorius The Occurrence of Enteric Viruses in Polluted Water, Correlation to lndicator Organisms and Factors Influencing their Numbers. Water Science and Technology 21 (3): Gerba, C.P Phage as lndicators of Fecal Pollution. In: Phage Ecoloav. Goyal, S.M., Gerba, C.P., Bitton, G. New York, John Wiley and Sons. Gerba, C.P., and G.E. Schaiberger Effect of Particulates on Virus Survival in Seawater. Journal Water Pollution Control Federation 47:93. Gerba, C.P., M.D. Sobsey, C. Wallis, and J.L. Melnick Factors influencing the adsorption of poliovirus onto activated carbon in wastewater. Environ. Sci. Technol. 9: Goldfield, M Epidemiological lndicators for Transmission of Viruses by Water. In: Viruses in Water. Berg, G, Bodily, H.L., Lennette, E.H., Melnick, J.L., Metcalf, T.G., Eds., pp. 256, Washington, DC, American Public Health Association. Goyal, S.M lndicators of Viruses. In: Viral Pollution of the Environment. Berg, Gerald. Boca Raton, FL, CRC Press. Grabow, W.O.K., and P. Coubrough Practical Direct Plaque Assay for Coliphages in 100-ml Samples of Drinking Water. Applied and Environmental Microbiology W(3) Grabow, W.O.K., V. Gauss-Muller, O.W. Prozesky, and F. Deinhardt lnactivation of Hepatitis A Virus and lndicator Organisms in Water by Free Chlorine Residuals. Applied and Environmental Microbiology 46(3): Guy, M.D., and J.D. Mclver The Removal of Virus By a Pilot treatment Plant. Water Research 1 1 : Hall, R.M., and M. Sobsey Inactivation of Hepatitis A Virus and MS2 by Ozone and Ozone-Hydrogen Peroxide in Buffered Water. Proceedings of International Association on Water Pollution Research and Control, Washington, DC, May Morris, R.W., Grabow, W.O.K. and Dufour, A.P. Editors. Halvorsrud, J. and I. Orstavik An epidemic of rotavirus-associated gastroenteritis in a nursing home for the elderly. Scandinavian Journal of lnfectious Disease 12:161. Handzel, T.R., R.M. Green, C. Sanchez, H. Chung, and M.D. Sobsey Improved specificity in detecting F-specific coliphages in environmental samples by suppression of somatic phages. Water Science and Technology 27: Harris, J.R., M.C. Cohen, and E.C. Lippey Water Related Disease Outbreaks in the United States. Journal of lnfectious Disease 148:759.

88 Havelaar, A.H Bacteriophages as Model Organisms in Water Treatment. Microbial Science 4: Havelaar, A.H Specific RNA Bacteriophages in Sewage: Methodology and Occurrence. Water Science and Technology 17: Havelaar, A.H F-Specific RNA Bacteriophages as Model Viruses in Water Treatment Processes. PhD Thesis. State University of Utrecht, Netherlands, 240 pp. Havelaar, A.H Bacteriophages as Models of Human Enteric Viruses in the Environment. ASM News 59: Havelaar, A.H., K. Furuse, and W.M. Hogeboom Bacteriophages and indicator Bacteria in Human and Animal Feces. Journal of Applied Bacteriology 60: Havelaar, A.H., and W.M. Hogeboom Method for the Enumeration of Male-Specific Bacteriophages in Sewage. Journal of Applied Bacteriology 56: Havelaar, A.H., and W.M. Pot-Hogeboom F-Specific RNA-Bacteriophages as Model Viruses in Water Hygiene. Ecological Aspects. Water Science and Technology 20(11/12) Havelaar, A.H., M. van Olphen, and Y.C. Drost F-specific RNA-bacteriophages are adequate model organisms for enteric viruses in fresh water. Appl. Environ. Microbiol Hazen, T Fecal Coliforms as Indicators in Tropical Waters: A Review. Toxicity Assessment 3: Hazen, T.C. and G.A Toranzos Tropical Source Water, pp In: Drinkinn Water Microbiology, G.A. McFeters (ed.). Springer-Verlag, New York. Hejkal, T., B. Keswick, R. LaBelle, C.P. Gerba, Y. Sanchez, G. Dreesman, B. Hafkin, and J.L. Melnick Viruses in a Community Water Supply Associated with an Outbreak of Gastroenteritis and Infectious Hepatitis. Journal of American Water Works 74:318. IAWPRC Study Group Bacteriophages as model viruses in water quality control. Wat. Res. 25: Kaplan, J.E., G.W. Gary, R.C. Baron, N. Singh, L.B. Schonberger, R. Feldman, and H.B. Greenberg Epidemiology of Norwalk Gastroenteritis and the Role of Norwalk Virus in Outbreaks of Acute Non Bacterial Gastroenteritis. Annals of Internal Medicine, 96(part 1): Kenard, R.P., and R.S. Valentine Rapid Determination of the Presence of Enteric Bacteria in Water. Applied Microbiology 27:484. Keswick, B.H., C.P. Gerba, H.L. Dupont, and J.B. Rose Detection of Enteric Viruses in Treated Drinking Water. Applied and Environmental Microbiology 47:

89 Ketratanakul, A., and S. Ohgaki Indigenous Coliphages and RNA- F-Specific Coliphages Associated with Suspended Solids in the Activated Sludge Process. Water Science and Technology 21 (3): Khuroo, M.S Study of an Epidemic of Non-A, Non-B Hepatitis. American Journal of Medicine 68:818. Kott, K Viruses and Bacteriophages. Science of the Total Environment 18: Kott, Y., N. Roze, S. Sperber, and N. Betzer Bacteriophages as Viral Pollution Indicators. Water Research 8: Levine, W.C., W.T. Stephenson, and G.F. Craun. March Waterborne Disease Outbreaks, Center for Disease Control Morbidity and Mortality Weekly Report 39(SS-1):l. Loh, C.L., M.S. Jangi, Y.C. Ho, Y.F. Ngeow, and C.W. Wang Development of a Field Test Kit for Coliphage Detection in Natural Waters. Toxicity Assessment 3:451. Longley, K.E., V.P. Olivieri, C. W. Kruse, and K. Kawata Enhancement of Terminal Disinfection of a Wastewater Treatment System. In: Virus Survival in WAter and Wastewater Systems. Malina, J.F. and B.P. Sagik, Editors. Center for Research in Water Resources, Univ. of Texas at Austin. Lycke, E., J. Bloomberg, G. Berg, A. Erikson, and L. Madsen Epidemic Acute Diarrhea in Adults Associated with Infantile Gastroenteritis. Lancet 2:1056. Melnick, J.L Human Enteric Viruses in Sediments, Sludges and Soils: An Overview. In: Human Viruses in Sediments, Sludges and Soils. Rao, V.C., Melnick, J.L., Eds., Boca Raton, FL, CRC Press. Melnick, J.L., and C.P. Gerba The Ecology of Enteroviruses in Natural Waters. CRC Critical Reviews in Environmental Control, 10: Mosley, J.W Transmission of Viral Diseases by Drinking Water. In: Transmission of Viruses by the Water Route. Ed. Berg, G., pp. 5-23, New York, John Wiley and Sons. Olivieri, V.P., F.H. Hauchman, C. Noss, and R. Vasi Mode of Action of Chlorine Dioxide on Selected Viruses. In: Water Chlorination: Chemistry, Environmental Impact and Health Effects. Jolley, J.L., Ed., pp. 651, Chelsea, MI, Lewis Publishers. Osawa, S., K. Furuse, and I. Watanabe Distribution of Ribonucleic Acid Coliphages in Animals. Applied and Environmental Microbiology 41 :I Parry, O.T., J.A. Whitehead, and L.T. Dowling Temperature Sensitive Coliphage in the Environment. Viruses in Wastewater Treatment, pp , Oxford, Pergamon Press. Payment, P. and E. Franco Clostridium perfrngens and Somatic Coliphages as Indicators of the Efficiency of Drinking Water Treatment for Viruses and Protozoan Cysts. Applied and Environmental Microbiology 59: