REPORT NO: 928/1/03

February 2003


1. Background and Motivation

Water is the most important single natural resource limiting growth and development in South Africa. Management of the quality of this precious resource is, therefore, of special importance. In addition, water-borne diseases are one of the most important causes of morbidity and mortality world-wide. International bodies such as the World Health Organization and the World Bank base their strategic planning and budgets on estimates of some 50 000 deaths per day due to water related diseases. The devastating public health impact of water-borne diseases in South Africa is illustrated by the ongoing cholera epidemic in some parts of the country. Estimates on the socio- economic implications of water-borne diseases in South Africa run into billions of Rands.

Viruses feature prominently among the wide spectrum of pathogens associated with water-borne transmission. Unfortunately the detection of human viruses in water requires relatively sophisticated laboratory facilities and advanced expertise. Direct monitoring of water for the presence of viruses is, therefore, expensive and not within the capabilities of many routine laboratories. The same applies to direct assessment of the efficiency of water treatment and disinfection processes with regard to the removal and inactivation of viruses.

Monitoring of the virological quality of water, and assessment of the efficiency of treatment and disinfection processes does, therefore, heavily depend upon more practical indirect methods. Commonly used indirect methods include indicator organisms. The application of bacterial indicators of faecal pollution for this purpose, notably coliform bacteria, became well established over the past century. Although coliforms and other faecal bacteria serve a most valuable purpose in the assessment of faecal pollution and the potential presence of micro-organisms excreted by humans and animals, they do have certain definite shortcomings. These shortcomings refer in particular to the value of faecal bacteria as indicators for both the incidence and behaviour of pathogens such as viruses and the cysts and oocysts of protozoa in water environments and water treatment processes. These shortcomings are related to basic differences such as:

The shortcomings of faecal bacteria as indicators for pathogens such as viruses and protozoan cysts and oocysts has been confirmed in many studies. Evidence includes the detection of viruses in drinking water supplies which meet all specifications for treatment, disinfection and faecal indicator counts. These results of virological analysis are confirmed by epidemiological data which associate viral infections with drinking water supplies which meet internationally accepted specifications for treatment, disinfection and counts of faecal indicator bacteria. The latest data on viruses in at least some drinking water supplies which meet quality specifications for treatment, disinfection and indicators, indicate that these supplies may exceed acceptable risks for viral infections.

In endeavours to overcome the shortcomings of commonly used faecal indicator bacteria, attention has world-wide been given to alternative indicators which may share more properties with viruses than faecal bacteria. Bacteriophages (phages) proved strong candidates in this regard. Somatic coliphages have long been used to supplement faecal bacteria as indicators for viruses in water environments and water treatment processes. Phages are attractive for this purpose because they closely resemble viruses in terms of size, structure and composition, they are excreted by humans and warm-blooded animals, their survival in water environments and treatment processes resembles that of viruses more closely than bacteria, they rarely if ever multiply in water environments, and they are detectable by simple, rapid, and economic techniques which yield results within 24 h.

F-RNA (male-specific) coliphages have even more attractive features as indicators for viruses than somatic coliphages. This is because of their structure, size, morphology and composition which are almost identical to that of typical entero viruses. Under the electron microscope F-RNA coliphages are hardly distinguishable from polio, coxsackie and other entero viruses. In addition, F-RNA coliphages fail to multiply in water environments, and there is a substantial body of evidence that their behaviour in water environments and resistance to water treatment and disinfection processes closely resembles that of viruses. Another attractive feature of F-RNA coliphages is that they can be classified into four sero- or genotypes which seem to be specific for humans or animals. This implies that typing of F-RNA coliphages detected in raw and treated water supplies may be used to distinguish between faecal pollution of human and animal origin. Faecal bacteria such as coliforms, including E coli, can not be used to distinguish between faecal pollution of human and animal origin. One disadvantage of F-RNA coliphages is that reliable detection in water is not as easy as for somatic coliphages, and typing of F-RNA coliphage sero- or genotypes requires relatively advanced molecular technology.

2. Objectives

In view of the above considerations, this project was undertaken with the following objectives:

2.1. To develop practical techniques for the detection of F-RNA coliphages in water environments, which includes recovery techniques and plaque assays for small numbers of phages in large volumes of water.
2.2. To develop techniques to confirm the identity of F-RNA coliphages, which includes the selection of host bacteria, testing of RNase sensitivity and electron microscopy.
2.3. To develop gene probe techniques for genotyping F-RNA coliphage serogroups.
2.4. To compare various labels for probes in order to develop simple and practical hybridisation techniques to distinguish between F-RNA coliphage genotypes.
2.5. To investigate the excretion of F-RNA coliphage genotypes by animals and humans.
2.6. To compare the incidence of F-RNA coliphages in selected water environments to that of enteric viruses.

Basically the purpose of this study was to investigate the extent to which F-RNA coliphages in South Africa can be divided into the same genotypes described in other parts of the world, to assess the application of these phages to distinguish between faecal pollution of human and animal origin, and to assess the reliability of F -RNA coliphages as indicators for human viruses in raw and treated water sources.

A study of this kind has not yet been carried out in South Africa. The results were expected to have meaningful benefits for practical technology and expertise on water quality assessment. In addition, the project offered valuable opportunities for the education and training of manpower essential to the water industry, and to make a meaningful contribution to capacity building and technology transfer in a field of major national importance.

3. Literature Review

A comprehensive literature review has been included in the project report. Further details have been recorded by Uys (1999). Essential features of the literature review, together with conclusions and recommendations regarding water quality analysis, have been published (Ashbolt et al, 2001; Grabow, 2001; Schaper et al, 2001).

4. Results

The following is a summary of results obtained:

4.1. Isolation of F-RNA coliphages from water
Reliable and accurate methods were developed for the detection of F-RNA coliphages in small and large volumes of water. The techniques are based on double agar layer plaque assays using small and large petri dishes for different volumes of water and different densities of F-RNA coliphages in test samples. A highly sensitive procedure for the detection of small numbers of F- RNA coliphages in large volumes of water has been established. The qualitative presence- absence test on 500 ml volumes of water is well suited for the detection of F-RNA coliphages in drinking water.
4.2. Detection of F -RNA coliphages
In an evaluation of potential host strains evidence has been presented that the genetically engineered Salmonella typhimurium strain WG49 is the host of choice for the detection of F- RNA coliphages in raw and treated water sources. The reliability of F-RNA coliphage detection was confirmed by electron microscopic analysis of isolates, confirmation of the RNase sensitivity of isolates, and finally by genetic hybridisation with known nucleotide sequences. The importance of using the host in optimal logarithmic growth phase for plaque assays has been confirmed. This unfortunately requires carefully controlled growth conditions and monitoring of host numbers which renders the test more complicated than for somatic coliphages.
4.3. Genotyping of F-RNA coliphages
Molecular techniques for the typing of F-RNA phage genotypes have been optimised and established into a procedure suitable for routine laboratory testing. Appropriate genetic probes for the four genotypes of F-RNA coliphages were kindly supplied by co-workers in the USA, Netherlands and Spain. Best results were obtained with low hybridisation stringency for maximum probe-target interaction, followed by high washing stringency which yielded less background without non-specific signals. A direct hybridisation technique which yielded genotyping results within 2-3 days after sample collection has been developed. This reliable and practical approach to typing has not previously been described.
4.4. Host specificity of F-RNA coliphage genotypes
Analysis of many stool samples from humans and a variety of animals confirmed that F-RNA genotypes II and III were highly specific for humans, and genotypes I and IV for animals. One exception to the rule was genotype II phages which were also detected in 28 % of pig faecal samples, but not in the faeces of any other animals. These results are in line with earlier findings and confirm the potential value of F-RNA genotypes to distinguish between faecal pollution of human and animal origin.
4.5. Incidence of F-RNA coliphage genotypes in water environments
A variety of waste waters and raw and treated water sources were analysed for the presence of F- RNA phage genotypes. Data on the incidence of genotypes in waste waters predominantly of human or animal origin, such as abattoir and hospital discharges, confirmed the specificity of genotypes for humans and animals. Genotype II phages predominated in sewage and river water, followed by genotype I phages. Genotype III phages predominated in hospital waste water, and genotype IV phages were rarely detected. The findings support the potential value of F-RNA phage genotypes to distinguish between faecal pollution of human and animal origin.
4.6. Value of F-RNA coliphages as indicators for human viruses
Studies on a wide variety of waste waters, as well as raw and treated water sources, confirmed that F-RNA coliphages generally tend to outnumber enteric viruses by approximately tenfold. This meets basic requirements for indicators. However, the ratio in numbers tended to decrease for treated drinking water supplies. In a meaningful number of drinking water supplies enteric viruses were detected but no F-RNA coliphages even using the sensitive presence-absence test on 500 ml samples. The latter findings may be due to one or both of two reasons. Enteric viruses may be more resistant to drinking water treatment and disinfection processes than F-RNA coliphages, or the detection techniques for enteric viruses were more sensitive than for F-RNA phages. The latter seems possible because F-RNA tests were on 500 ml samples while enteric viruses were recovered from more than 100 litre samples. This implies that F-RNA coliphages may have been detected like enteric viruses if detection techniques of comparable sensitivity were to be used.
4.7. International study on F-RNA coliphage genotypes
Data on F-RNA coliphages obtained in this study were combined with data obtained in related studies in Spain to assess the value of F-RNA coliphage genotypes as a tool to distinguish between faecal pollution of human and animal origin. A manuscript on the findings of this joint study has been submitted for publication (Schaper et al., 2001). Findings of this study may be summarised as follows:
The occurrence of F-specific RNA bacteriophages and their genotypes in human and animal faeces, and in sewage predominantly associated with human or animal wastes, has been investigated in South Africa and Spain. The bacteriophages were detected in 10% of human, 45% of bovine, 60% of porcine and 70% of poultry faecal specimens. The percentage of samples in which F-RNA bacteriophages were detected increased for pools of faecal specimens, and approached 100% for animal feedlot slurries. Only genotypes II and III were detected in human stools. Animal faeces contained predominantly, but not exclusively, genotypes I and IV. Hospital wastewater and domestic sewage contained predominantly genotypes II and III. Abattoir waste water contained predominantly genotypes I and IV. The predominance of genotypes II and III in raw domestic and urban sewage tended to shift towards predominance by genotypes I and II in settled sewage, secondary treated sewage and non-point diffuse effluents from developing communities, suggesting higher resistance of genotypes I and II to unfavourable environmental conditions. Despite differences between Spain and South Africa such as geographic location, climate and population, there was no statistical significance in the occurrence of F-RNA bacteriophages and their genotype in comparable samples of faeces and waste water. Although the association of genotypes II and III with human excreta and genotypes I and IV with animal excreta was statistically significant, the results for a variety of samples suggest that the association cannot be used for absolute distinction between faecal pollution of human and animal origin in raw and treated waste water or environmental water sources. However, in combination with other indicators the association may prove useful for determining the origin of faecal pollution.

5. Cost estimates

The cost of microbiological analyses varies substantially subject to a number off actors. A survey of tariffs quoted by leading service providers in the South African market indicates that on average the cost tests per sample for bacterial indicators of faecal pollution amounts to approximately R 100, somatic coliphages R 120, F -RNA coliphages R 140 and genotyping of F- RNA coliphages R 300, compared to enteric viruses R 850 and Cryptosporidium-Giardia R 1400. Features of the tests which need to be taken into account are that tests for many bacteria and phages yield results within one to two days, genotyping of F-RNA takes about four days, and tests for enteric viruses take two to four weeks.

6. Capacity building

One MSc degree has been awarded for work carried out as part of this project (Uys, 1999). A number of other postgraduate students in our Department gained experience in the field as part of contributions to the project and part of their education and training. These include Mr EE Muller who is registered for a MSc degree on bacteriophages associated with toxins in Escherichia coli.
Mr A Sundram, Umgeni Water, Pietermaritzburg, is registered for a MSc degree in our Department for work following on this project.
Ms TS Ndou is registered for a MSc degree at the University of Venda for work which was initiated by this project.

7. Technology transfer

Research staff and students from the following establishments visited our Department for training and education in techniques for using phages in water quality analysis: University of Venda:
Two members of research staff and three post-graduate students University of Fort Hare: One senior lecturer
Free State Technikon:
One senior lecturer and two post-graduate students
Department of Microbiology and Plant Pathology at the University of Pretoria: One postgraduate student Umgeni Water:
One post-graduate student
Rand Water:
Two members of research staff City Council of Windhoek:
One member of research staff

The detection of F-RNA coliphages has been established as integral part of the routine monitoring of several drinking water supplies, including reclaimed drinking water in Windhoek.

The value of the work carried out in this project has been outlined in an international editorial (Editorial, 1999). The findings and experience of this project have been included in water quality guidelines of the World Health Organization (Ashbolt et al., 2001) and specifications of the International Organization for Standardization.

8. Conclusions

All the objectives of the project have been accomplished.
New techniques for the detection and genotyping of F-RNA coliphages in water have been established.
Evidence has been presented that these techniques can make a valuable contribution to water quality assessment, overcoming at least some of the shortcomings of commonly used faecal indicator bacteria.
F-RNA coliphage tests yield valuable information on the virological quality of water at relatively low cost within a short period of time.

9. Future Research

The following areas of research warrant further attention:
9.1. Observations that F-RNA coliphage genotype II may be excreted by more animals than previously thought (Schaper et al., 2001) should be investigated to confirm the ability of F- RNA phage genotypes to distinguish between faecal pollution of human and animal origin.
9.2. The possibility that the survival of different genotypes of F-RNA coliphages in water environments may differ (Schaper et al, 2001) should be confirmed and the implications of this phenomenon for the indicator value of F-RNA coliphages should be investigated.
9.3. Efforts should be made to increase the sensitivity of F-RNA detection methods to levels comparable to those used for the detection of enteric viruses in drinking water.
9.4. Compare F-RNA coliphage genotypes to other determinands such as specific faecal sterols to distinguish between faecal pollution of human and animal origin. Best results may be obtained by using a combination of different methods.
9.5. Study the application of F-RNA tests as practical and cost effective tools in Hazard Analysis Critical Control Point (HACCP) systems for drinking water treatment processes. This would form an integral and fundamentally most important component of water quality control strategies currently pursued by the World Health Organization.
9.6. Follow up on technology transfer activities and capacity building aimed at introducing the new technology in water quality analysis.
9.7. Study the practical application of F-RNA coliphage indicators in more details and formulate strategies for optimal application of the indicators in routine monitoring programmes.

10. Publications emanating from the Project

Ashbolt N J, Grabow WOK, Snozzi M (2001) Chapter 13: Indicators of microbial water quality. In: Water Quality Guidelines: Guidelines, Standards and Health. Editors Fewtrell L and Bartram J. World Health Organization Water Series. IWA Publishing, London. Pp 289-315.

Editorial (1999) Phages gain ground as water quality indicators. Water 21 (Magazine of the International Water Association) November-December, 36-37. Grabow WOK (2001) Bacteriophages: Update on application as models for viruses in water. Water SA 27, 251-268.

Müller, E E (1997) An evaluation of methods for the detection of F-RNA coliphages and Bacteroides fragilis HSP40 phages in large volumes of water. BSc(Hons) Dissertation, Department of Medical Virology, University of Pretoria.

Schaper M, Jofre J, Uys M, Grabow WOK (2001) Distribution of genotypes of F-specific RNA bacteriophages in human and non-human sources of faecal pollution in South Africa and Spain.
Journal of Applied Microbiology (in press).

Uys M (1999) Molecular Characterisation F-Specific RNA Phages in South Africa. MSc Dissertation, Department of Medical Virology, University of Pretoria.

11. Conference poster

Müller, EE; Clay, CG and Grabow, WOK (2000). Detection and isolation of Escherichia coli O157:H7 from sewage and environmental waters using immunomagnetic separation. Poster: Biennial Conference of the Water Institute of Southern Africa, Sun City, 28 May to 1 June.