Indicators of microbial quality in recreational waters

... in determining microbial risks of recreational waters from aquatic facilities in Singapore Despite widespread hype of the potential of using molecular techniques in water quality monitoring, they... Attributes of an ideal indicator Since the beginnings of water quality research, many indicators have been proposed but all possess shortcomings in one way or another In 2004, several attributes of an... effects of waterborne disease outbreaks in coastal and inland recreational waters, including treated (spas, pools, interactive fountains) and untreated (reservoir) waters In 2007 and 2008, a total of

INDICATORS OF MICROBIOLOGICAL QUALITY IN RECREATIONAL WATERS HO DANLIANG (B.Sc. (Hons), NUS) A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF SCIENCE DEPARTMENT OF MICROBIOLOGY NATIONAL UNIVERSITY OF SINGAPORE 2014 i Declaration I hereby declare that the thesis is my original work and it has been written by me in its entirety. I have duly acknowledged all the sources of information which has been used in the thesis. This thesis has also not been subrnitted for any degree in any university previously. '4 1 Ho Danliang 06 lanuary 2014 Acknowledgements My heartfelt thanks goes to my supervisor, A/P Lee Yuan Kun for his invaluable guidance and kind encouragement throughout the course of my research. I thank him especially for challenging me to think further and deeper about my work, for engaging me in new ideas, for giving me intellectual freedom to conduct my own research, and on the overall for grooming me to become a better scientist. I would like to thank our collaborator on the other aspect of our project, Dr Karina Gin from the Department of Civil and Environmental Engineering, for her interest and support in our research, and her group members Dr Goh Shin Giek, Dr Saied Rezaeinejad, Mr Fang Haoming, Ms Liang Liang and Ms Genevieve Gabrielle Vergara for sharing their knowledge, equipment and reagents with us. My thanks goes to Ms Jennifer Yap and Ms Soh Liting from National Environmental Agency (NEA), Dr Lim Tien Tze and Ms Chang Siao Yun from Public Utilities Board (PUB) for their interest and input in this project, as well as for their generous sharing of experiences. My sincere gratitude goes to my mentor, Dr Koh Yiling, Eileen, for teaching me necessary laboratory and research skills ('tricks of the trade') in the first few months of my research candidature, and for providing guidance and advice in the nitty-gritty details of study design, benchwork as well as bioinformatics analysis. I would also like to thank her for her leadership, friendship and companionship; without which our laboratory would not be like what it is now. I am deeply indebted to Mr Low Chin Seng for his assistance and support in the logistic and technical aspects of conducting my experiments and fieldwork. I would like to thank him for being a fatherfigure to me in a laboratory setting, for his genuine care and concern (albeit under a rough exterior), and also for unselfishly imparting me his skills and knowledge. All results described in this thesis were accomplished with the help and support of members of the 'Water Group': Eileen Koh, Henry Teo, Siti Radiah. I would like to thank them for fighting this battle with me till the end, for helping me accomplish the impossible, and all in all for making my research stint an enjoyable and memorable experience. I would like to thank the other members of our laboratory for their wonderful guidance and constant friendship: the post-docs Dr Shen Hui and Dr Ng Yikai for their patience in answering all my queries, fellow graduate students Lin Huixin, Daphne Ng, Kelvin Koh, Zhao Ran, Yao Lina, Kenneth Tan for their comradeship, banter, and all the heated discussions on how my research should proceed, our ex and current honours students Chen Kang Ting, Huang Ningxin, Regina Ang, Angela Lim, Lee Huiting, Yow Kok Siang, Ang Sze Chien, Jolene Phua for their precious friendship and for generating much-needed laughter in an otherwise monotonous laboratory setting. Lastly but not least, my family members and my husband, Lim Suang, for their unwavering love and support. This work is dedicated to them. iii Table of contents DECLARATION ................................................................................................................................... ii ACKNOWLEDGEMENTS ................................................................................................................... iii TABLE OF CONTENTS ....................................................................................................................... iv LIST OF TABLES ............................................................................................................................... viii LIST OF FIGURES ............................................................................................................................... ix LIST OF ABBREVIATIONS ................................................................................................................ xi SUMMARY ......................................................................................................................................... xiv CHAPTER 1: INTRODUCTION ........................................................................................................... 1 1.1 Overview of recreational water quality..................................................................................... 1 1.2 Recreational water quality assessment...................................................................................... 2 1.3 Limitations of current methods of water quality monitoring ................................................. 3 1.4 Research objectives ..................................................................................................................... 3 1.5 Scope of study .............................................................................................................................. 4 CHAPTER 2: LITERATURE REVIEW ................................................................................................ 6 2.1 Waterborne diseases ................................................................................................................... 6 2.1.1 Overall prevalence and epidemiology of waterborne diseases ......................................... 6 2.1.2 Health criteria of waterborne pathogens ........................................................................... 7 2.1.2.1 Aeromonas hydrophilia..................................................................................... 7 2.1.2.2 Escherichia coli O157....................................................................................... 8 2.1.2.3 Legionella pneumophila.................................................................................... 9 2.1.2.4 Mycobacterium avium complex ...................................................................... 10 2.1.2.5 Pseudomonas aeruginosa ............................................................................... 11 2.1.2.6 Staphylococcus aureus .................................................................................... 12 2.2 Indicator approach ................................................................................................................... 13 2.2.1 Historical background and current status ........................................................................ 13 2.2.2 Attributes of an ideal indicator ....................................................................................... 15 2.2.2.1 Biological attributes ........................................................................................ 15 2.2.2.2 Method attributes ............................................................................................ 16 2.2.3 Limitations of traditional indicators ................................................................................ 16 2.2.4 Alternative indicators ...................................................................................................... 18 iv 2.2.4.1 Fecal anaerobes ............................................................................................... 18 2.2.4.2 Chemical compounds ...................................................................................... 19 2.2.5 Classical methods of bacterial indicator detection .......................................................... 19 2.2.5.1 Multiple-tube fermentation technique............................................................. 20 2.2.5.2 Membrane filtration technique ........................................................................ 21 2.2.6 Emerging methods in indicator detection ....................................................................... 21 2.2.6.1 Immunological methods.................................................................................. 22 2.2.6.2 Real-time PCR ................................................................................................ 22 2.2.7 Issues with molecular detection of indicators ................................................................. 23 2.2.7.1 Differentiating cell viability ............................................................................ 23 2.2.7.1 Overcoming PCR inhibition ........................................................................... 24 2.3 Direct pathogen monitoring ..................................................................................................... 25 2.3.1 Phylogenetic analysis using 16S ribosomal RNA gene .................................................. 26 2.3.2 DNA microarrays ............................................................................................................ 27 2.3.3 Sequencing studies .......................................................................................................... 28 2.3.3.1 Next-generation sequencing overview ............................................................ 28 2.3.3.2 Application to metagenomics and pathogen detection.................................... 29 2.3.3.3 Bioinformatics ................................................................................................ 30 CHAPTER 3: QUANTIFYING MOLECULAR MARKERS AS INDICATORS OF WATERBORNE PATHOGENS ................................................................................................................ 31 3.1 Introduction ............................................................................................................................... 31 3.2 Materials and methods ............................................................................................................. 32 3.2.1 Sample collection and processing ................................................................................... 32 3.2.1.1 Study sites ....................................................................................................... 32 3.2.1.2 Sampling procedures ....................................................................................... 35 3.2.1.3 Chemical tests ................................................................................................. 35 3.2.1.4 Membrane filtration ........................................................................................ 35 3.2.1.5 DNA extraction from water samples .............................................................. 36 3.2.2 Real-time polymerase chain reaction (qPCR) analysis ................................................... 36 3.2.2.1 Reference strains used in standards ................................................................ 36 3.2.2.2 Development of plasmid standards ................................................................. 37 3.2.2.3 Validation of primer set using end-point PCR ................................................ 38 3.2.2.4 Validation of primer set using real-time PCR ................................................. 41 3.2.2.5 Real-time PCR of samples .............................................................................. 42 3.2.2.6 Data analysis ................................................................................................... 43 3.2.3 Differentiation between live/dead bacteria ..................................................................... 43 3.2.3.1 Culture strain................................................................................................... 43 3.2.3.2 Preparation of dead cells ................................................................................. 43 3.2.3.3 Comparing PMA and EMA pre-treatment ...................................................... 43 3.2.3.4 Determining and validating optimum PMA concentration ............................. 44 3.2.4 PCR inhibition ................................................................................................................ 45 3.2.5 Field tests ........................................................................................................................ 45 3.2.5.1 Pathogen detection .......................................................................................... 45 3.2.5.2 Evaluating pathogen-indicator relationship .................................................... 46 v 3.3 Results ........................................................................................................................................ 46 3.3.1 Comparing the effectiveness of PMA and EMA pre-treatment ...................................... 46 3.3.2 PMA optimisation ........................................................................................................... 47 3.3.3 Validation of PMA pre-treatment in environmental water matrix .................................. 49 3.3.4 PCR inhibition ................................................................................................................ 51 3.3.5 Detection of selected potential pathogens in environmental samples ............................. 53 3.3.6 Examination of indicator-pathogen relationship in single-point studies......................... 54 3.4 Discussion................................................................................................................................... 57 3.4.1 Development of optimisation procedures for qPCR bacterial marker quantification ..... 57 3.4.2 Indicator-pathogen occurrence and relationships in recreational water samples ............ 60 3.5 Conclusion ................................................................................................................................. 63 CHAPTER 4: OCCURRENCE, DISTRIBUTION AND DECAY OF SELECTED MOLECULAR MARKERS .................................................................................................................... 64 4.1 Introduction ............................................................................................................................... 64 4.2 Materials and methods ............................................................................................................. 65 4.2.1 Study site description ...................................................................................................... 65 4.2.2 Sample collection and processing ................................................................................... 65 4.2.2.1 Sampling for spatial variability ...................................................................... 66 4.2.2.2 Intensive sampling ......................................................................................... 66 4.2.3 Culture analysis............................................................................................................... 67 4.2.4 Laboratory microcosm study .......................................................................................... 72 4.2.5 Statistical procedures ...................................................................................................... 73 4.3 Results ........................................................................................................................................ 73 4.3.1 Determination of spatial variability of bacterial markers ............................................... 73 4.3.2 Determination of temporal variability of bacterial markers ............................................ 75 4.3.3 Marker decay kinetics ..................................................................................................... 80 4.4 Discussion................................................................................................................................... 83 4.4.1 Spatial and temporal variability in bacterial marker distribution.................................... 83 4.4.2 Bacterial marker survival and persistence ...................................................................... 84 4.5 Conclusion ................................................................................................................................. 86 CHAPTER 5: DIRECT PATHOGEN DETECTION AND DISCOVERY USING SEQUENCING METHODS .................................................................................................................... 88 5.1 Introduction ............................................................................................................................... 88 5.2 Materials and methods ............................................................................................................. 89 5.2.1 Establishment of 16S rRNA clone library ...................................................................... 89 5.2.1.1 PCR amplification........................................................................................... 90 5.2.1.2 DNA ligation, transformation and plasmid extraction .................................... 91 5.2.1.3 Verification of insert ....................................................................................... 92 vi 5.2.1.4 Restriction fragment length polymorphism (RFLP) screening ....................... 92 5.2.1.5 Bioinformatics analysis ................................................................................... 93 5.2.2 454-sequencing ............................................................................................................... 93 5.2.2.1 PCR amplification and purification ................................................................ 94 5.2.2.2 Sample clean-up, quantification and quality check ........................................ 94 5.2.2.3 Sequence analysis and phylogenetics assignment .......................................... 95 5.2.2.4 Sequence identity calculation ......................................................................... 95 5.3 Results ........................................................................................................................................ 97 5.3.1 Phylogenetic characterisation using clone libraries ........................................................ 97 5.3.2 Taxonomic assignment of sequences from 454-sequencing ......................................... 103 5.3.3 Within-sample diversity analysis .................................................................................. 106 5.3.4 Comparison against pathogen database ........................................................................ 109 5.4 Discussion................................................................................................................................. 111 5.4.1 Overall bacterial community structure .......................................................................... 111 5.4.2 Pathogen discovery and detection ................................................................................. 113 5.5 Conclusion ............................................................................................................................... 115 CHAPTER 6: CONCLUSIONS ......................................................................................................... 116 REFERENCES ................................................................................................................................... 119 APPENDIX I ...................................................................................................................................... 132 APPENDIX II ..................................................................................................................................... 133 APPENDIX III .................................................................................................................................... 135 APPENDIX IV.................................................................................................................................... 137 vii List of tables Table 3.1 Location characteristics of sampled recreational water facilities Table 3.2 List of gene-targeted primers used in this study Table 3.3 Validation of amplification efficiency and lowest limit of detection of primer set Table 3.4 Mean, standard deviation, minimum and maximum qPCR counts obtained when BJ, IMM and VC samples were diluted 1X, 10X and 100X. Table 3.5 Regression analyses and correlations of pathogens and indicators Table 4.1 Summary of the type and purpose of each culture medium used to culture viable bacteria from collected environmental water samples Table 4.2 Regression analyses and correlations of indicator and potential pathogens Table 4.3 Variable loadings on the first five rotated principle components Table 5.1 Human pathogenic bacteria, disease and accession number of reference sequences Table 5.2 Identity and amount of sequences corresponding to potentially pathogenic bacteria viii List of figures Figure 3.1 Geographical locations of sampling sites Figure 3.2 Individual site images - (A) BJ, (B) CS, (C) IMM, (D) SB, (E) SC, (F) VC Figure 3.3 Example of gel used to check PCR product size and primer specificity Figure 3.4 Comparison between PMA and EMA treatment of viable and heat-killed E. coli mixtures Figure 3.5 Chart showing the relationship between different PMA concentrations and enumeration of E. coli spiked in pure water Figure 3.6 Chart showing the relationship between different PMA concentrations and enumeration of E. coli spiked in environmental water matrix Figure 3.7 Chart showing the final counts (in CFU/100ml) obtained from qPCR results for different dilutions of samples collected from three different locations Figure 3.8 Chart showing pathogen distribution within sampled interactive fountains Figure 3.9 Chart showing distribution of (A) fecal markers and (B) pathogens within sampled interactive fountains Figure 4.1 Illustrative diagram of fountain water recycling system Figure 4.2 Hierarchical chart depicting processing workflow for every liter of water sample Figure 4.3 Bacteria counts for (A) total heterotrophic plate count, (B) total/fecal coliforms, (C) Pseudomonas and (D) Streptococcus and fecal enterococci as obtained by culture Figure 4.4 Bacteria counts for (A) total bacterial load, (B) Enterobacteriaceae, (C) Aeromonas, (D) Pseudomonas, (E) Legionella and (F) Mycobacterium as obtained by qPCR Figure 4.5 Temporal patterns of distribution for bacteria markers enumerated using (A) culture and (B) qPCR Figure 4.6 Scree plot showing a sharp drop after PC1, indicating that most of the variance in the data had been explained by the first component Figure 4.7 Survival profiles of selected pathogens (A-D) and indicators (E-F) Figure 5.1 Flowchart summarizing the steps involved in the establishment of each clone library Figure 5.2 Illustration of the 16S rRNA gene Figure 5.3 Example of different RFLP patterns obtained from double digest of 12 clones within the Science Centre (SC) clone library Figure 5.4 Phylogenetic tree of sequences obtained from water samples collected from the following locations: (A) Bugis Junction, (B) IMM, (C), Science Centre and (D) VivoCity ix Figure 5.5 Proportions of unique sequences found in each taxon Figure 5.6 Taxonomic classification of bacteria at the phylum/class level within (A) Bugis Junction, (B) IMM, (C) Science Centre and (D) VivoCity Figure 5.7 Collector's curves obtained for samples from BJ, IMM, SC and VC using the (A) Chao1 and (B) Inverse Simpson diversity estimator Figure 5.8 Rarefaction analysis of observed richness in bacterial communities sampled from BJ, IMM, SC and VC Figure 5.9 Heatmap comparing relative abundance of potentially pathogenic sequences that occurred in the samples with red to green continuum representing high to low abundance respectively x List of abbreviations A. hydrophilia Aeromonas hydrophilia ANOVA Analysis of Variance APHA American Public Health Association ATCC American Type Culture Collection B. thetaiotaomicron Bacteroidetes thetaiotaomicron BCYE Buffered Charcoal Yeast Extract BD Becton, Dickinson and Company BJ Bugis Junction BLAST Basic Local Alignment Search Tool bp Basepair CFU Colony-forming units CS City Square CT Cycle threshold DNA Deoxyribonucleic acid E. coli Escherichia coli EC broth Escherichia coli broth ELISA Enzyme-linked immunosorbent assay EMA Ethidium monoazide FA/PCA Factor analysis/Principal components analysis FIB Fecal-indicator bacteria Gb Gigabyte GC-MS Gas chromatography-mass spectrometry GI Gastro-intestinal GM Geometric mean GVPC Glycine, Vancomycin, Polymyxin B sulphate, Cyclohexamide HUS Hemolytic-uremic syndrome IEA Immuno-enzyme assay xi IFA Immunofluorescence assay k Decay coefficient kB Kilobase L. pneumophila Legionella pneumophila LB Luria Bertani LB-A LB-ampicillin LOD Limit of detection M. avium Mycobacterium avium M. smithii Methanobrevibacter smithii MAC Mycobacterium avium complex MF Membrane filtration ML Maximum Likelihood MPN Most probable number MRSA Methicillin-resistant Staphylococcus aureus MTF Multiple-tube fermentation MUG 4-methylumbelliferyl-β-D-glucuronide NEA National Environmental Agency NGS Next-generation sequencing NRC National Research Council NTAC National Technical Advisory Committee NTC Non-template control P. aeruginosa Pseudomonas aeruginosa P. alcaligenes Pseudomonas alcaligenes PCR Polymerase chain reaction PMA Propidium monoazide PUB Public Utilities Board PVC Polyvinyl chloride qPCR Real-time polymerase chain reaction xii RDP Ribosomal Database Project RNA Ribonucleic acid rRNA Ribosomal RNA S. aureus Staphylococcus aureus SB Sembawang Shopping Centre SC Science Centre Singapore SMEWW Standard Methods for the Examination of Water and Wastewater  Kendall's tau correlation coefficient TDS Total dissolved solids U.K. United Kingdom U.S.A. United States of America USCDC United States Centers for Disease Control and Prevention USEPA United States Environmental Protection Agency USPHS United States Public Health Service UV Ultraviolet VBNC Viable-but-non-culturable VC VivoCity w/v Weight per volume WBDOSS Waterborne Disease and Outbreak Surveillance System WHO World Health Organisation xiii Summary Worldwide, recreational water quality criteria are based on the measurement of fecal indicator bacteria using culture techniques. However various limitations of using traditional fecal indicators in monitoring bathing water quality and predicting disease outbreak exist; these include low levels of correlation with pathogens, low detection sensitivity, ability to multiply outside the water column, as well as inability to predict pathogens of non-fecal sources, thus raising questions on the applicability of current measurement criteria on recreational water facilities. In the present study, molecular techniques were evaluated for the ability to detect and predict pathogen occurrence. Optimisation procedures including development of standards, sensitivity and specificity testing, determination of LOD and amplification efficiency for detection of selected indicators and pathogens using qPCR were performed. The differentiation of viable and dead bacteria using EMA and PMA pre-treatment, as well as the extent of PCR inhibition occurring within environmental samples was also evaluated. In comparison to EMA, PMA pre-treatment on samples spiked in pure water performed better with maximum dead cell exclusion and minimum live cell signal suppression, and 30µM was determined to be the optimum concentration. PCR inhibition was found to be present in one out of the three studied sites, highlighting the necessity of conducting such verification tests prior to any molecular analyses of environmental samples. The optimised qPCR assays were used to analyse indicator and pathogen count and distribution within six recreational water facilities in Singapore using a point-sampling approach. A more intensive sampling for spatial and temporal variability of bacterial markers were also conducted within a single recreational water facility. Using correlation analysis, linear regression and FA/PCA, it was discovered that Enterobacteriaceae and M. smithii were good predictors of pathogen presence in study sites subjected to point-sampling, but none of the measured indicators predicted pathogens well when temporal variability was considered. It was found that the bacterial indicators and pathogenic markers possessed similar transport characteristics but markedly different decay rates under the same conditions, possibly explaining for the observed lack of correlation. xiv The inability to establish any indicator-pathogen relationship necessitated the use of direct pathogen detection as an alternative method to water quality assessment. Direct pathogen detection and discovery was performed using the traditional Sanger sequencing approach to construct clone libraries, and 454-sequencing that allows for alpha and beta-diversity analyses as well as comparison against a database of known pathogens. Both methods revealed the presence of consistently dominant taxa which included Alpha, Beta and Gamma-proteobacteria, as well as Bacteroidetes. In addition, 454sequencing identified a number of rare phyla which included Acidobacteria, Chlamydia, Clostridia, Delta-proteobacteria and Verrucomicrobiae. On the other hand, the distribution of pathogenic genera were markedly different across samples, and several genera that were not known to be associated with an aqueous environment (Neisseria, Streptococcus, Chlamydia, Rickettsia) were discovered. Furthermore, a few potentially pathogenic genera (Pseudomonas and Mycobacterium) were found to persist in large amounts within all samples. The methods described in this study have been used to characterise the bacterial and pathogenic community present within recreational water and could also potentially be used to screen for indicators that better correlate with water-borne pathogens. xv Chapter 1: Introduction 1.1 Overview of recreational water quality Worldwide, there is an increase in the popularity of recreational activities that involve contact with water. In Singapore, government initiatives have resulted in a redesign of our waterways to incorporate various lifestyle attractions including kayaking and windsurfing activities (PUB, 2013). There is also a trend in the increasing use of indoor water recreation such as commercial spa pools, Jacuzzi and interactive water playgrounds. Such frequent contact with water can expose individuals to health hazards such as viral, bacterial or protozoan pathogenic microorganisms within the water environment, and increases the risk of spreading water-borne diseases. Epidemiological reports have demonstrated a link between adverse health effects and prior immersion in poor quality recreational waters (Craun et al., 2005; USCDC, 2008; USCDC, 2011), and these reports are mostly based on acute, self-limiting diseases such as gastro-intestinal (GI) illnesses, which are transmitted via the fecal-oral route. On the other hand, many studies have shown that pathogens associated with more severe outcomes or secondary health problems may also be spread by the use of contaminated water (Bartram and Rees, 2000; Pond, 2005). These pathogens, which include Legionella, Mycobacterium avium complex, Vibrio vulnificus, Naegeria fowleri, Hepatitis A and Hepatitis E virus, have been isolated from surface waters and their role in pathogenesis are well-documented (Bartram and Rees, 2000; Pond, 2005). However due to the possibilities of other transmission routes for the pathogens-in-question, epidemiological research demonstrating a causative link between recreational water usage and contracting such diseases have been scarce. It is equally plausible that other unknown pathogens could be transmitted by use of recreational water and this association has not been investigated. There is much that is not understood about the microbial risks present in recreational water settings. Therefore, from a research standpoint, it is important to investigate what is the nature of 1 the microbial community, including pathogens, that is present in our recreational waters. It is also integral to develop accurate, reliable and scientifically defensible methods that assess water quality, in order to protect public health. 1.2 Recreational water quality assessment The risks of developing GI illnesses from fecal-polluted recreational waters are well-studied and established; as such water quality assessment efforts have been focused on fecal contamination (Craun et al., 2005; USCDC, 2008; USCDC, 2011). However due to the low concentration of pathogens, direct detection is difficult to perform, hence bacteria that are correlated to pathogen levels have traditionally been used as surrogate measures. Fecal indicator bacteria (FIB) such as fecal coliforms, Enterococcus and Escherichia coli are often used for water quality assessment. Being commonly found in the lower intestines of humans and other warm-blooded mammals, FIB are thought to indicate the presence of fecal material, thus co-occur with and signal the presence of human pathogens. The extensive use of FIB to measure water quality is due to well-established and relatively inexpensive detection and enumeration methods, as well as several epidemiological studies supporting the use of FIB as predictors for gastro-intestinal illnesses among swimmers (Wade et al., 2003; Shah et al., 2011). First developed for drinking water standards, the use of FIB was later expanded and adopted for ambient and recreational water criteria. Currently, the United States Enviromental Protection Agency (USEPA)‟s recommended recreational water quality criteria, which is adopted by 95% of states within the United States of America (U.S.A.), suggests that within a 30-day interval, the rate of GI disease attributable to swimming should be 36 cases per 1000 recreators if the monthly geometric mean (GM) of Enterococcus is less than 35 colony-forming units (CFU)/100ml, or if the GM of E. coli is less than 126 CFU/100ml (USEPA, 2012). In Singapore, the National Environmental Agency (NEA) adopts guidelines from the World Health Organisation (WHO) 2 standards (2003) stating that „95% of the time, the Enterococcus counts should be less than or equal to 200 counts per 100ml of water‟ (NEA, 2013) . 1.3 Limitations of current methods of water quality monitoring Despite the widespread usage of fecal indicator bacteria (FIB) in water quality monitoring, the limitations of using the existing bacterial indicator approach to assess water quality are becoming increasingly evident (Section 2.2.3 and 2.2.4). There is little evidence to suggest that fecal contamination is the major source of pathogens in recreational water settings, however the actual pathogen diversity present and the ability of FIB to predict them has not been well studied. Furthermore, even in water bodies where fecal contamination is known to occur, research has shown that the strength of indicator/pathogen relationship is weak, raising questions on their suitability as predictors of disease (Griffin et al., 1999, Nobel and Fuhrman, 2001). Current monitoring methods are also based on traditional culturing techniques which are less accurate, sensitive and more time consuming as compared to molecular methods. Several methods have been proposed, including enzyme-linked immunosorbent assay (ELISA), microarrays and real-time polymerase chain reaction (qPCR), of which the latter is the most promising. However qPCR procedures for water quality assessment are currently unstandardised; they also suffer from several limitations: 1) the inability to differentiate live and dead bacteria and 2) PCR methods tend to be negatively affected by presence of inhibitor compounds. As such is unable to replace culture techniques as the gold standard yet. 1.4 Research objectives The suitability of FIB as markers of water quality assessment has been questioned. Furthermore, there is a paucity of studies performed on how applicable current methods of water quality assessment are in determining microbial risks of recreational waters from aquatic facilities in 3 Singapore. Despite widespread hype of the potential of using molecular techniques in water quality monitoring, they have not been fully optimised for use in recreational water settings. Hence, the objectives of this study are as follows: 1) To validate and optimize real-time quantitative detection methods for the detection of FIB in environmental waters, 2) To determine correlation and predictive powers of FIB against pathogens in environmental waters in spatial and temporal studies, and hence determine the applicability of current indicators in Singapore recreational waters, and 3) To develop alternative methods of assessing recreational water quality, should the current indicators be found to be unsuitable. 1.5 Scope of study Because the field of recreational water microbiology encompasses various aspects, prior to conducting this study, it is necessary to specify the scope of research based on the conducted literature review (Chapter 2). Unlike coastal waters and swimming pool bodies, interactive water fountains and spa pools currently do not possess specific microbiological quality guidelines. Due to logistic constraints it is difficult to obtain access to spa facilities for sampling purposes. As such this study will focus on interactive water fountains (splash pads or water play areas) as a subset of all recreational water facilities in Singapore. Furthermore, there is a general consensus that alternative pathways of pathogen transmission (other than fecal-oral pathway) occur frequently in the recreational water context, and that there is little research conducted on them (NRC, 2004; Pond, 2005). As such this study will focus more on pathogens endemic to the water environment and those spread via aerosols or skin contact. These pathogens are mostly of bacterial origin, such as Legionella spp. and Pseudomonas aeruginosa. Based on these set of 4 criteria, this study will attempt to characterize only the bacterial population found in interactive water fountains. 5 Chapter 2: Literature Review 2.1 Waterborne diseases Water is essential to life. However, water also serves as a medium for disease transmission, resulting in undesirable health outcomes in humans. This section briefly reports on the prevalence of waterborne diseases worldwide, as well as disease epidemiology in the United States. Subsequently, it will discuss the health criteria of selected waterborne bacterial pathogens including their characteristics and ecology, health effects, transmission routes and evidence for association with recreational water supplies. 2.1.1 Overall prevalence and epidemiology of waterborne diseases At the global scale, water-based infectious outcomes have a large impact on public health. Worldwide, at least 4.0% of deaths and 5.7% of disease burden worldwide is attributable to waterborne diseases which included diarrheal diseases, schistosomiasis and trachoma (Pruss et al., 2002; Kosek et al., 2003; Lewin et al., 2007). In total 1415 microorganisms are reported to be pathogenic, out of which 348 are known to be water-associated, causing 115 infectious diseases (Taylor et al., 2001). In the United States, disease outbreak data is collected by the Waterborne Disease and Outbreak Surveillance System (WBDOSS), which is a national surveillance system that monitors the nature, scope and health effects of waterborne disease outbreaks in coastal and inland recreational waters, including treated (spas, pools, interactive fountains) and untreated (reservoir) waters. In 2007 and 2008, a total of 134 outbreaks associated with recreational water was reported, resulting in at least 13,699 cases of illnesses, in which 40% of outbreaks comprised of dermatological illnesses, acute respiratory conditions as well as combined illnesses. The main etiological agents in these cases were of bacterial origin, which included Legionella spp. and P. aeruginosa (USCDC, 2011). 6 2.1.2 Health criteria of waterborne bacterial pathogens 2.1.2.1 Aeromonas hydrophilia Aeromonas spp. are gram negative rods belonging to the family Enterobacteriaceae which can cause a variety of gastro-intestinal and extra-intestinal diseases. Aeromonas hydrophilia is one of the most well-characterised species out of this group and is the most medically-relevant, as it has been isolated from wounds and can produce heat-labile enterotoxins, which result in the production of haemolysins and cytotoxins (Burke et al., 1981). Aeromonas spp. are ubiquitous in aquatic environments including surface waters, drinking and recreational waters. They grow over a wide temperature range with optimal growth at 30°C, and can remain viable for extended periods of time (Warburton, 2000). They may colonise water supplies as well as form biofilms which are resistant to disinfection (Holmes et al., 1996; Bomo et al., 2004). The primary health effect of Aeromonas infection is acute diarrheal disease; however wound infections may also occur following traumatic injury in an aqueous environment. Serious complications include meningitis, septicemia and pneumonia, and they have also been implicated in traveler's diarrhoea. Both healthy adults as well as immunocompromised individuals are susceptible to infection (Janda and Abbott, 1996; Joseph, 1996). There are few documented outbreaks of Aeromonas infections, out of which they are mostly associated with untreated drinking water supplies (Ghaenem et al., 1993). However superficial infections resulting from environmental exposure to recreational water are relatively common and has been described extensively in medical literature. The health risk of this pathogen is controversial: on one hand, there has been no reported outbreaks associated with drinking water supplies, yet the presence of several virulence factors resulting in the potential to cause disease in susceptible populations should warrant concern (Pavlov et al., 2004). As such it has been placed 7 in the USEPA Candidate Contaminant List and is subjected to further research, evaluation and risk assessment (USEPA, 1998). 2.1.2.2 Escherichia coli O157 Belonging to the family Enterobacteriaceae, pathogenic Escherichia coli has been found to be responsible for a variety of GI ailments, of which those caused by the O157 serogroup presents the most severe symptoms. E. coli O157 is enteric and mainly found in the intestines of humans and other mammals. Tolerance to low pH and cold temperatures aids the transmission of this bacterium. Furthermore, it has several virulence factors that contribute to pathogenesis: the production of Shiga-like toxins, the production of a haemolysin that enhances the toxicity of the Shiga toxins, as well as the ability to attach to and colonise intestinal surfaces (Kuntz and Kuntz, 1999). The primary disease symptom is abdominal cramping and bloody diarrhoea, while the most severe outcome is hemolytic-uremic symptom (HUS), which develops in around 2-8% of cases and is characterised by hemolytic anemia, acute kidney failure and low platelet counts (Rowe et al., 1991; Lansbury and Ludlam, 1997). Secondary health outcomes of patients suffering from HUS include abnormal kidney function requiring long-term dialysis, high blood pressure, seizures, blindness and paralysis (Sieglar et al., 1994). E. coli O157 is mainly transmitted by contaminated foods and water, as well as direct contact with animals and person-to-person spread. It is strongly associated with recreational water and there is much evidence linking outbreaks of E. coli O157 to bathing in contaminated waters, particularly swimming in freshwater ponds or wading pools (Brewster et al., 1994, Keene et al., 1994; Wang and Doyle, 1998; USCDC, 1998). These studies have identified the development of HUS in a subset of patients and have highlighted the severe clinical outcomes of infection by E. 8 coli O157. In most cases, outbreaks resulted from the use of pools that were not adequately chlorinated. 2.1.2.3 Legionella pneumophila The Legionella group of organisms belong to family Legionellaceae and genus Legionella, of which Legionella pneumophila is the organism most frequently associated with pathogical outcome, namely Legionnaire's disease. Various other Legionella species including L. anisa, L. micdadei and L. wadsworthii are also associated with disease (Hunter, 1998, Bartram et al., 2007). Legionella spp. occur in natural environments such as rivers and streams and have also been isolated from sewage-contaminated waters. Legionella are gram-negative rods that require Lcysteine and iron salts for growth in vitro. They are highly resistant to environmental factors and have been shown to multiply at high temperatures up to 60°C, as well as persist for long periods of time (up to 10 years) in water systems (Lawrence et al., 1999). They are intracellular pathogens and require the presence of other bacteria or protozoa as hosts for survival and growth. L. pneumophila causes Legionnaire's disease, of which common symptoms include high fever, severe headache, muscle and abdominal pain, diarrhoea and changes in mental status. The major cause of death is respiratory failure (Roig et al., 1993). Secondary health outcomes are numerous and include pulmonary complications, renal failure, cerebral abscesses, arthritis and seizures (Loveridge, 1981; Mamane et al, 1983; Smeal et al., 1985; Peliowski and Finer, 1986; Anderson and Sogaard, 1987). Legionnaire's disease is mostly community-acquired through direct inhalation of aerosols (Evenson, 1998). The infectivity dose is believed to be small as contaminated aerosols generated from a considerable distance were shown to be sufficient for infection. Susceptible populations include the elderly, people who are immuno-compromised, people with damaged respiratory tracts e.g. those with lung problems or smokers, as well as people with chronic illnesses such as 9 renal disease. It is thought to be one of the most frequent causes of community-acquired pneumonia (Thi Minh Chau and Muller, 1983) and outbreaks are likely to be under-reported due to misdiagnosis (Helms et al., 1980). Legionella is most strongly associated with spa waters such as hot tubs and saunas (Bornstein et al., 1989; Kuroki et al., 1998) but have also been isolated from fountains and wading pools (USCDC, 2011). In Singapore, Legionella spp. has been isolated from cooling towers and interactive water fountains (Lam et al., 2011), however there is little epidemiological evidence linking reported cases of Legionnaire's disease to these environmental isolates. 2.1.2.4 Mycobacterium avium complex Mycobacterium belong to the order Actinomycetales, family Mycobacteriaceae. Several species are both associated with water and cause disease in humans and animals. The organism of greatest public health concern is the Mycobacterium avium complex (MAC), which comprises of M. avium and M. intracellulare (Pedley et al., 2004). Mycobacterium can be found in various different environments including surface waters. MAC is highly resistant to antibiotics, heavy metals and disinfectants due to the presence of an impermeable cell wall, which allows it to grow at temperatures up to 45°C and at low oxygen concentrations (Rastogi et al., 1981; Miyamoto et al., 2000; Taylor et al., 2000). Primary infections caused by MAC include cough, production of sputum and chest pain. Severity of infections depends on the exposed persons' level of immunity: healthy adults suffer from mild cases of pneumonia while immunocompromised individuals could present with life-threatening respiratory illnesses. Besides pulmonary infections, Mycobacterium has also been associated with infections of the middle ear (Trupiano and Prayson, 2001). Lastly there is also research suggesting that MAC is one of the factors known to trigger Crohn's disease, a type of autoimmune disease primarily affecting the GI tract (Pickles et al., 2001). 10 MAC is primarily transmitted by environmental exposure such as inhalation of aerosols or ingestion of contaminated water. It is found endemic in natural waters as well as man-made facilities such as swimming pools, spas and footbaths. There is strong evidence linking exposure to environmental MAC to development of skin and soft tissue infections, and even pneumonia (Collins et al., 1984; Shelton et al., 1999). In most cases, disease outbreaks can be attributed to a high bather load, lack of disinfection, or inadequate maintenance of filters and pipes. 2.1.2.5 Pseudomonas aeruginosa The pseudomonads are a group of Gram-negative, curved rods classified into the Gammaproteobacterium phylum, genus Pseudomonas. They can grow in a wide range of habitats and are highly ubiquitous in the environment, contributing to food spoilage, degradation of petroleum and other materials, as well as plant and animal diseases. P. aeruginosa is of greatest medical importance as it is commonly associated with human illnesses and is potentially lethal. Pseudomonas are highly persistent and can survive for prolonged periods in filtered freshwater, possibly due to biofilm formation as well as the ability to grow at very low nutrient levels (van der Kooij et al., 1982; de Vicente et al., 1988). They are also reported to be strongly resistant to disinfectants such as quaternary ammonium compounds, phenolics and hypochlorite (Hollyoak et al., 1995). P. aeruginosa, ranked as the 5th most frequent cause of nosocomial infections, is by itself responsible for 10% of nosocomial pneumonia in hospital settings (Dembry et al., 1998), and has been documented to cause cerebrospinal infections and bacterimia (Fick, 1992). In healthy individuals, P. aeruginosa frequently causes ear and skin infections, and has also been linked to infections of the urinary-tract and gastro-intestinal tract. P. aeruginosa-induced infections are notoriously difficult to treat, associated with peculiar clinical aspects, may result in serious 11 complications and have a high morbidity rate (Pollack, 1995; Shigemura et al., 2006; Goldman et al., 2008). P. aeruginosa is naturally found in surface waters; however numbers have been reported to be increased in water contaminated by fecal material, stormwater discharge and high bather load, linking its presence to human activities. Other reported sources of P. aeruginosa is urban run-off and water leeching from soil (Reali and Rosati, 1994; Geldreich, 1996). The presence of P. aeruginosa in recreational water bodies such as spas, whirlpools and hydrotherapy pools is epidemiologically linked to outbreaks of outer ear and skin diseases (Ratnam et al., 1986; Aspinall and Graham, 1989; Craun et al., 2005). Factors include warm temperatures, water aeration and high bather load, as well as inadequate disinfection (Brett and du Vivier, 1985; Ratnam et al., 1986). Strong adherance of P. aeruginosa to polyvinyl chloride (PVC) piping and filters through biofilm formation makes removal difficult. 2.1.2.6 Staphylococcus aureus Staphylococcus are a group of gram-positive, cocci-shaped bacteria that are usually arranged in grape-like clusters. They belong to the Gamma-proteobacterium phylum, genus Staphylococcus, with at least 15 different species within the genus. Within the group, S. aureus, S. epidermidis and S. saprophyticus are associated with human diseases; however S. aureus is the most medically significant due to potential lethality in clinical settings (WHO, 2013). It is capable of growing aerobically or anaerobically, and is able to produce enterotoxins which are heat-stable and resistant to degrading enzymes, and can cause disease when ingested (Le Loir et al., 2003). Staphylococcus aureus infections include that of the skin, GI disease, septicaemia, endocarditis, pneumonia and bacterimia (Lowry, 1998). One of the most well-characterised clinical complication is toxic-shock syndrome, which manifests as a persistently high fever, low blood pressure and confusion, and is rapidly followed by coma and multiple organ failure (Lowry, 12 1998). The emergence of methicillin-resistant strains of Staphylococcus aureus (MRSA) further complicates treatment. Staphylococcus have a short duration of persistence in the environment and has been reported to be easily controlled by conventional disinfection procedures. On the other hand, being commonly found as part of the normal microflora of human skin, they are easily released into water environments and thus transmitted to susceptible populations. Substantial quantities of S. aureus and MRSA have been reported to be shedded by humans in bathing waters (Plano et al., 2011), and outbreaks of S. aureus infections have been associated with exposure to marine waters (Charoenca and Fujioka, 1993, 1995). 2.2 Indicator approach As awareness increased about disease transmission by water, there was a growing realisation of the importance of monitoring water supplies in order to protect public health. This section describes the history of recreational water monitoring practices including an explanation of the indicator approach and the guidelines that Singapore currently practices, followed by a description of the attributes of an ideal indicator, as well as limitations of traditional indicators, their detection methods and emerging trends in indicator detection and method development. 2.2.1 Historical background and current status The earliest outbreak that was linked to disease was documented by a physician, John Snow, who determined that drinking sewage-polluted water led to cholera in the 1800s (Paneth, 2004). Following his work, early efforts by researchers were focused on the prevention of diseases transmitted by the fecal-oral route. Because methods used for direct detection and monitoring of pathogens were not sensitive and hence not protective of public health, an index used to discern fecal-contaminated water based on E. coli was developed by Theobold Smith (Smith, 1891). 13 Subsequently the American Public Health Association (APHA) standardized methods of E. coli detection and published the results in the Standard Methods for the Examination of Water and Wastewater (SMEWW) (Wolfe, 1972). Following studies showing that E. coli displayed higher persistence and a strong positive correlation to several pathogenic organisms, E. coli as an indicator for fecal-contaminated waters came into widespread use (Kerr and Butterfield, 1943; Wattie and Butterfield, 1944). However, ongoing concerns about the suitability of E. coli led to the development of alternative indicators, notably Enterococcus and Clostridium perfringens (NRC, 2004). Concerns over the quality of recreational water followed the development of drinking water standards. The United States Public Health Service (USPHS) was the first governmental agency to study the association between usage of recreational water and enteric disease transmission, and unequivocally demonstrated a link between an epidemic of typhoid fever to “bathing in polluted waters” (Stokes, 1927a, b). Guidelines for recreational water usage first appeared in the 1950s; however different jurisdictions employed standards differing in terms of both the type of indicator (total coliform, Enterococcus or E. coli) as well as statistical reporting (geometric mean or median) (Garber, 1956). In an attempt to standardize recreational water criteria, the National Technical Advisory Committee (NTAC) proposed a series of guidelines which included the use of fecal coliforms in 1968. In 1986, the USEPA adopted and modified the proposed microbiological criteria, to include Enterococci and E. coli. These guidelines remain in use today and have only been revised slightly in the latest edition of recreational water quality criteria (USEPA, 2012). Based on the criteria promulgated by USEPA and other epidemiological studies, WHO also published a set of standards for coastal and fresh waters in 2003, and for swimming pools and other similar environments in 2006 (WHO, 2003, 2006). Following the promulgation of guidelines by USEPA and a growing awareness about the link between poor water quality and disease spread, governmental agencies in developed countries 14 such as the United Kingdom (U.K.), Australia and Singapore began to routinely employ methods that test water quality in bathing waters. In Singapore, NEA adopted the guidelines recommended by WHO in 2003 and established the following criteria in August 2008 (NEA, 2013): For recreational beaches, the parameters used to assess water quality are: 1. The Enterococcus counts should be less than or equal to 200 counts per 100ml of water 2. Susceptibility of the location to fecal influence 3. Only beaches classified as „Good‟ and above will be considered suitable for primary contact activities For freshwater bodies1, 95% of the time: 1. The Enterococcus counts should be less than or equal to 200 counts per 100ml of water 2. Chlorophyll-a‟s concentration should be less than or equal to 50mg/L of water 2.2.2 Attributes of an ideal indicator Since the beginnings of water quality research, many indicators have been proposed but all possess shortcomings in one way or another. In 2004, several attributes of an ideal biological indicator and its measurement criteria are listed as follows (NRC, 2004): 2.2.2.1 Biological attributes 1. It should possess a strong quantifiable relationship to the pathogen and the degree of public risk 1 Refers only to inland waters exposed to outdoor conditions such as reservoirs and swimming pools. 15 2. It should possess similar survival and persistence levels as pathogens, in reference to environmental hazards such as ultraviolet (UV) light exposure, temperature, salinity, predation, desiccation, freezing and response to disinfectants 3. It should have similar fate and transport mechanisms as pathogens, in terms of behavior during filtration, sedimentation, and particle adsorption 4. It should occur in greater numbers than pathogens and thus be easily detectable 5. It should be specific to the source of contamination 2.2.2.2 Method attributes 1. Ability to quantify 2. Specificity to desired target organism/group 3. Broad applicability to different settings 4. Adequate sensitivity 5. Results rapidly obtained 6. Able to measure viable/infective target organism 7. Logistical feasibility in terms of cost, manpower and training, as well as sampling and processing requirements 2.2.3 Limitations of traditional indicators Due to historical reasons as explained in Section 2.2.1, fecal indicator bacteria (FIB), which includes total and fecal coliforms, E. coli and Enterococcus are currently being used as monitoring tools to indicate water quality and predict the presence of pathogens. However, it is widely recognised that they are not ideal indicators and there are numerous limitations associated with their usage. This section attempts to discuss the shortcomings of FIB with respect to the ideal biological attributes as stated in Section 2.2.2. 16 There is an abundance of papers reporting on the lack of correlation between concentrations of FIB and that of various types of pathogens, including enteroviruses, bacteriophages, Campylobacter, Salmonella and Cryptosporidium, in several different water mediums such as sewage effluent, coastal waters, estuarine systems and inland surface waters (Carter et al., 1987; Ferguson et al., 1996; Baggi et al., 2001; Noble and Fuhrman, 2001; Lemarchand and Lebaron, 2003; Savichtcheva and Okabe, 2006). As the most important biological attribute that an indicator should possess is a strong association between indicator and pathogen numbers, such reports raises serious questions on the protective value of FIB and their suitability in predicting the degree of public health risk. FIB have been reported to possess differential survivability and persistence as compared to pathogens. FIB are less resistant to harsh environmental conditions, disinfection processes and survive for a shorter period of time as compared to pathogens which are more hardy such as the protozoan pathogens (Cryptosporidium and Giardia), parasitic or biofilm-forming bacteria (Legionella or Pseudomonas), as well as several types of viruses (McFeters, 1990; Hurst et al., 2002; Baggi et al., 2001; Hörman et al., 2004). Furthermore, FIB are shown to be able to regrow and multiply in the water column, which do not favour the survival of enteric pathogenic organisms (Davies et al., 1995; Byappanahalli and Fujioka, 2004; Byappanahalli et al., 2006). This leads to a decoupling of the pathogen-indicator relationship and further erodes their association. Models of bacterial fate and transport kinetics have shown that FIB are able to survive in the water column by attaching to and persisting in fine particles or sediments, especially in the presence of organic matter (Ferguson, 1994; Friesa et al., 2008). On the other hand pathogen transport and distribution is more complicated and depends on several factors including temperature, salinity, dissolved oxygen, pH, microbial size, growth, predation, metal or nutrient availability (Metge, 2013). In particular, pathogens are prone to loss through settling and 17 inactivation due to temperature, UV exposure and grazing (Brooks et al., 2004). Such differential transport mechanisms results in a changing of the relationship between pathogen and indicator at varying distance from the source contamination and at different time periods, making it difficult to assess water quality accurately. FIB are good predictors of GI diseases as they are strongly associated with pathogens which are enteric and transmitted via the fecal-oral route. However, they are not considered source-specific with regards to two criteria: 1) they do not allow distinction whether the fecal contamination source is from humans or from animals, and 2) the source input of various non-enteric pathogens such as Legionella and Mycobacterium is likely to be different because of different host ecology, and thus it is unlikely that the existing FIB approach can predict them accurately (Cappelluti et al., 2003; Lam et al., 2011). 2.2.4 Alternative indicators The vast majority of alternative indicators proposed have focused on markers for fecal contamination, and they include fecal anaerobes, chemical compounds and viral indicators. Of the three categories, only the first two pertain to the detection of bacterial pathogens which is of interest in this study, while viral indicators are mainly used as surrogate indicators for enteric viruses. As such this section focuses solely on the description of fecal anaerobes and chemical compounds. 2.2.4.1 Fecal anaerobes These group of organisms usually belong to the group of Bacteroides or Bifidobacterium and are strongly associated with fecal matter in warm-blooded animals. The primary advantage of using Bacteroides and Bifidobacterium as indicators is their inability to tolerate atmospheric oxygen and survive in aerobic conditions. As such, detection of these organisms in large amounts indicate recent and extensive fecal contamination (Savichtcheva and Okabe, 2006). Several Bacteroides 18 species are host-specific and as such, tracking methods can be developed to identify the source of fecal contamination (Kreader, 1995; Simpson et al., 2004). Due to difficulties in cultivation in anaerobic conditions, detection and enumeration is usually performed using molecular approaches such as end-point and qPCR. However, a lack of correlation to non-enteric pathogens, a lack of studies on the transport and geographical distribution of these markers, as well as fluctuating levels of persistence due to seasonal changes are all problems associated with measurement of these molecular indicators (Savichtcheva and Okabe, 2006). 2.2.4.2 Chemical compounds Chemical compounds such as fecal sterols (which include cholesterol and coprostanol) have been proposed as alternative indicators to address the shortcomings of most microbial markers, such as multiplication in the water column as well as non-source specificity. In particular, coprostanol is one of the major fecal sterol found in human and animal feces, and possesses a half-life of less than 10 days at 20°C (Isobe et al., 2002), as such detection indicates the presence of recent fecal contamination events. Fecal sterols can be detected and measured routinely using gas chromatography-mass spectrometry (GC-MS). However, a major limitation of fecal sterols as indicators is the absence of source specificity as a result of different transport mechanisms between fecal sterols and pathogens: coprostanol can associate with and be preserved within sediments therefore detection may imply old or remote fecal contamination. There is also a lack of studies that determine host specificity, detection sensitivity and correlation to pathogen levels, as such they are not commonly used as indicators of microbial water quality (Savichtcheva and Okabe, 2006). 2.2.5 Classical methods of bacterial indicator detection Traditionally, indicator bacteria such as coliforms have been detected using culture-dependent methods. These methods are regarded as the 'gold-standard' methods and are routinely performed 19 in monitoring agencies. However, they possess several limitations that detract them from being regarded as ideal indicators as described in Section 2.2.2. The major shortcoming is the lack of sensitivity, which could lead to an underestimation of numbers especially when aiming to detect stressed or starved bacterial cells. They are also labourious and time-consuming to perform, resulting in a time-lag between sample collection and obtaining of results, between which the microbial composition and numbers may have experienced drastic changes. Most importantly, several studies have shown that FIB levels exhibit extreme variability over short time scales (Leecaster and Weisberg, 2001; Boehm, 2007). This leads to the possibility that FIB concentrations may vary more quickly than the monitoring results obtained, such that by the time culture results are available, the water quality may have changed substantially. This creates problems for policy makers who have to rely on outdated and possibly erroneous information when deciding whether or not to issue swimming bans or advisories. 2.2.5.1 Multiple-tube fermentation technique Multiple-tube fermentation (MTF) involves serial dilution of the collected water sample, followed by inoculation of each dilution into a set of tubes containing lactose or lauryl tryptose broth. Tubes are monitored for coliform growth and production of gas following a 24hr incubation period at 37°C. The number of tubes with a positive, presumptive reaction is noted, and subjected to a second, confirmatory test using brilliant green lactose bile broth. This method can be modified to test for fecal coliforms and E. coli by using Escherichia coli broth (EC broth) and EC broth supplemented with 4-methylumbelliferyl-β-D-glucuronide (MUG) respectively. Results are expressed statistically in terms of the most probable number (MPN) of bacteria, or the mean number of organisms expected to be present in the sample considering the number of positive tubes and the dilution factor. This technique suffers from low precision and possible interference from non-coliform bacteria in the environmental water sample, resulting in an underestimation of 20 coliform numbers. Furthermore it is more time-consuming as compared to the membrane filtration technique. 2.2.5.2 Membrane filtration technique The membrane filtration (MF) technique involves the filtration of a known volume of water sample through a 0.45µm sterile nitrocellulose membrane. The membrane is retained and incubated on selective medium such as m-Endo or Tergitol-TTA agar; subsequently the number of total coliforms is enumerated via colony count on the membrane. This method is considered presumptive and typically confirmatory tests are conducted by inoculating individual colonies in brilliant green lactose bile broth, EC broth or EC-MUG broth for total coliforms, fecal coliforms and E. coli respectively. MF is widely accepted as a technique for water quality monitoring (APHA, 1998) and presents several advantages over MTF, which includes a shorter processing time, and the ability to examine larger quantities of water, thereby resulting in greater sensitivity as compared to MTF. However, questions have been raised about the problem of interference from competing bacteria, resulting in suppression of coliform growth and an underestimation of their numbers. Another problem is the inability to recover stressed or injured coliforms due to exposure to disinfection agents such as chlorine. Both problems result in lowered accuracy of this technique. 2.2.6 Emerging methods in indicator detection Culture-based methods for detecting indicator organisms, as elaborated upon in the earlier section, suffer from several limitations such as long processing times, low sensitivity, specificity, accuracy and reliability. Newly developed methods for the detection of indicators boast higher rapidity, sensitivity and specificity; these same attributes also make them suitable for the direct detection of low pathogen levels directly from the environmental sample. This section will cover immunological techniques, which are common techniques used for the detection of indicators, 21 and give an in-depth analysis of qPCR as a method that is able to detect and quantify both indicators and pathogens. 2.2.6.1 Immunological methods Immunological methods are based on specific antigen-antibody interactions to detect target bacteria. Generally either cell-capture using ELISA, cell-detection using immunofluorescence assay (IFA) or immuno-enzyme assay (IEA) can be performed, using polyclonal or monoclonal antibodies. All of these tests are simple, rapid and relatively sensitive, allowing for specific detection, identification and quantification, and as such attempts have been made at using immunological methods for detection of water indicators. However, the main limitation is that there are very low numbers of target cells in the sample, and the accuracy depends on the specificity of the antibody. Various studies found that these methods require pre-culture of the sample for 24hrs (Obst et al., 1989), demonstrated cross-reactivity with competitors of the target organism (Levasseur et al., 1992) as well as produced false positive results despite using monoclonal antibodies (Hubner et al., 1992). As such, immunological methods have been met with limited success. 2.2.6.2 Real-time PCR Real-time PCR is a molecular technique that enables quantification by detecting the amount of amplified products as measured by the emission of fluorescence. Several PCR protocols have been developed to detect and/or measure the levels of traditional fecal indicators (e.g. E. coli, Enterococcus, fecal coliforms) and pathogens (e.g. Salmonella) (Iqbal et al., 1997; Tsen et al., 1998; Waage et al., 1999). The high sensitivity makes it a technique that is not only suitable to monitor indicator count but also detect pathogen levels directly. By using primers that target phylogenetic markers (e.g. 16S rRNA, gyrB or rpoD) and/or virulence factors, it is possible to detect and identify waterborne protozoa, bacteria and viruses, and even determine whether the 22 targeted organism possesses the virulence marker, therefore indicating possible pathogenicity (Joly et al., 2006; Matsuda et al., 2009; Benitez and Winchell, 2013; Le Gall et al., 2013). qPCR not only possesses higher sensitivity and specificity as compared to microarray and end-point PCR, it also has a faster rate of detection as well as the ability to provide quantitative results. On the other hand, detection is limited to microorganisms with known sequences, and only one gene (or a few genes, if multiplex protocols are used) may be targeted at any one time. qPCR has been successfully demonstrated to quantify various pathogens including Giardia, Cryptosporidium, Vibrio parahaemolyticus, pathogenic Candida, Campylobacter, Legionella, Aeromonas and Salmonella in various different environmental matrixes including surface waters (tap, pond and river water), rainwater and sewage (Blackstone et al., 2003; Brinkman et al., 2003; Guy et al., 2003; Ahmed et al., 2008). However, its major limitation is the inability to differentiate nucleic acid material obtained from viable and dead cells, as well as PCR inhibition issues arising from the humic substances and colloid matter commonly present in environmental samples (Rompre et al., 2002). Nonetheless, with the multiple advantages over traditional culture and immunological methods, various research groups have recommended its use as a complementary tool or even total replacement of standard methods in the monitoring of recreational waters (Metcalf et al., 1995; Haugland et al. 2005; USEPA, 2012). 2.2.7 Issues with molecular detection of indicators 2.2.7.1 Differentiating cell viability Although molecular methods such as real time PCR offer great advantages in terms of detecting and enumerating indicator organisms, one major limitation is the inability to differentiate DNA contributed from live and dead cells. Studies have shown that DNA may persist for extended periods of time in the environment after cell death (Josephson et al., 1993; Masters et al., 1994), which will result in generation of false positive signals or an overestimation of the target 23 organism when molecular methods are used for detection and enumeration, resulting in an inaccurate measure of pathogen risk. While incorporation of microscopic visualization or flow cytometry techniques allows for live/dead cell discrimination, it is difficult to perform for water samples where pathogen load is low and is not feasible for routine monitoring purposes. In the last few years, propidium monoazide (PMA) and ethidium monoazide (EMA) have been reported to be a promising alternative method to distinguish between live and dead cells in environmental samples (Nocker et al., 2007). PMA and EMA are intercalating dyes that covalently binds to DNA to form a stable nitrogen-carbon bond, modifying DNA permanently. They cannot penetrate through intact membrane of live cells and thus can be used to selectively bind to DNA contributed from dead cells with compromised membrane structure while leaving DNA from live cells intact. Various studies have reported on the feasibility of using PMA/EMA to reduce signal from dead bacteria within selected environmental water matrixes (Nogva et al., 2003; Rudi et al., 2005; Nocker and Camper, 2006; Nocker et al., 2006). In addition, PMA combined with qPCR has been successfully applied for the detection and enumeration of fungi, protozoa, enteric viruses and bacteriophages (Vesper et al., 2008; Brescia et al., 2009; Fittipaldi et al., 2010; Parshionikar et al., 2010). However, as environmental matrixes are highly complex and contains a number of factors (e.g. turbidity) that may affect the effectiveness of PMA/EMA treatment, it is still necessary to optimize and evaluate this approach when applied to 'real-world' recreational water samples (Gedalanga and Olson, 2009). 2.2.7.2 Overcoming PCR inhibition PCR inhibitors are inhibitory molecules or particles that are usually found in clinical and environmental samples such as soil and fecal material, and inhibits PCR by mechanisms such as: interacting and inactivating DNA polymerase or DNA intercalating dyes such as SYBR Green I, interfering with probe/primer binding, degradation or capture of nucleic acids or interference with 24 the DNA extraction reaction through inhibition of the release of nucleic acid material (Wilson, 1997; Rådström et al., 2003). Because PCR inhibition can affect the yield and specificity of the reaction resulting in lowered amplification efficiency as well as an underestimation of the amount of gene target, it is important to determine their effect and optimize the PCR reaction when working with environmental samples (Katcher and Schwartz, 1994). In qPCR, the effect of PCR inhibitors is usually investigated by studying the amplification efficiency and the linear range of amplification, typically performed by doing a dilution series of DNA spiked in the environmental matrix in comparison with nuclease-free H2O. In the presence of PCR inhibition problems, it is then necessary to perform pretreatment of the samples to reduce or remove the effect of inhibitors, using techniques such as density gradient centrification (Lindqvist et al., 1997), dilution (Al-Soud and Rådström, 1998), DNA extraction (Dahlenborg et al., 2001) or filtration (Lantz et al., 1999). Alternatively, inhibitor-tolerant thermostable polymerases or inclusion of amplification facilitators which are able to enhance PCR efficiency in the presence of inhibitors could be used in the PCR reaction (Al-Soud and Rådström, 1998). 2.3 Direct pathogen monitoring There are many unanswered questions with regards to the severity and frequency of disease in recreational water settings, and it is highly plausible that various diseases arising from usage of contaminated water may have gone unnoticed. Most characterized illnesses are acute and of GI origin and therefore research in this field have mostly focused on developing methods to detect fecal contamination using the indicator approach. As described in the previous section, the indicator approach is flawed and possesses several limitations such as different contamination sources and differences in survival, persistence, fate, transport therefore resulting in a lack of correlation between indicator and pathogen levels. With the advent of molecular tools which boasts greater rapidity, sensitivity and specificity as compared to traditional culture-based 25 methods, it is now also possible to conduct direct detection and monitoring of pathogens. Not only a wider range of pathogens can be detected, novel pathogens can be discovered and studied for associations with waterborne diseases (Aw and Rose, 2012). Methods for pathogen discovery and detection commonly employ the 16S ribosomal RNA (16s rRNA) gene for phylogenetics characterisation and assignment of identity. This section first discusses the use of the 16S rRNA gene as a phylogenetic marker for characterising microorganisms, as well as describes the various molecular tools which are available for direct pathogen detection, including microarrays and sequencing techniques. 2.3.1 Phylogenetic analysis using 16S ribosomal RNA gene The 16S rRNA gene is a valuable tool for studying bacterial phylogeny and taxonomy due to a number of reasons. These reasons include 1) it is present and conserved in almost all known bacterial species, 2) sequence differences between bacterial groups are a result of random sequence changes which makes it a good measure of evolution, 3) the structure of this gene consists of conserved regions interspersed with regions prone to mutations (variable regions), allowing for the possibility of designing universal primers that target almost all bacterial groups, and 4) the large size (1500bp) of the gene is suitable for bioinformatics purposes. The ease in sequencing the 16S rRNA gene for bacterial identification has resulted in an explosion of the number of recognized taxa (13537 published names to date compared to 62 in 1980 for taxonomic category of Class and below), many of which belong to uncultured bacterial species (LPSN, 2013). To handle this data and make it accessible to the general public, as well as to provide tools for the analysis of 16S rRNA data, several online databases have been set up. Ribosomal Database Project (RDP) (http://rdp.cme.msu.edu/), Greengenes (http://greengenes.lbl.gov) and Silva (http://www.arb-silva.de/) are examples of repositories that are quality-checked, regularly updated and therefore of greater use to the scientific community. 26 The widespread availability of 16S rRNA data also makes possible the identification up to genus and species level bacterial isolates that are not culturable or that generate ambiguous biochemical profiles. In spite of the usefulness of 16S rRNA gene sequencing in classifying bacterial groups, it should be noted that due to a high degree of similarity (only a few nucleotide differences) between closely related species, the phylogenetic power is low at the species level and there may be an inability to discriminate between certain genus. For example the type strains of Edwardsiella species are 99.35 to 99.81% similar in DNA sequence and are not able to be differentiated by 16S rRNA analysis, yet they are distinguishable biochemically and by DNA homology. There is also currently no scientific consensus on the definition of species identification using 16S rRNA gene sequencing, and studies vary widely as to calling a species match (Janda and Abbott, 2007). Other issues include sequencing accuracy, isolate purity and DNA extraction methods, all of which may affect the final identification. 2.3.2 DNA microarrays DNA microarrays are two-dimensional ordered arrays that consist of a high density of immobilized nucleic acids, which serve as probes that hybridize to complementary sequences found in the sample, as such detecting thousands of genes in a single assay. Although theoretically useful to directly detect and analyse the pathogens present in environmental samples, this technology suffers from limitations such as the lack of pathogen sequence information, and lesser sensitivity and specificity as compared to qPCR (refer to Section 2.2.6.2), making it impractical for the routine monitoring and detection of low levels of pathogen in environmental matrixes. On the other hand, it has been useful in analyzing the microbial profile and determining the community structure in complex environmental samples such as soil and wastewater, when conducted in tandem with an enrichment procedure such as PCR amplification. To this end, 27 phylogenetic microarrays targeting the 16S rRNA gene (PhyloChip) have been applied to the analysis of microbial community structure of bacteria (DeSantis et al., 2007), and is useful as a research tool to discover and characterize novel pathogens within the environmental sample. 2.3.3 Sequencing studies Prior studies of bacteria evolution and phylogenetics, followed by the discovery of 16S rRNA gene sequences heralded a paradigm shift from culture to molecular-based in the field of pathogen identification. It was realised that targeted sequencing of the 16S rRNA gene could enable investigators to establish an identity of several microbial species that could not be identified by traditional culture methods (Pace et al., 1985). Initial studies were based on Sanger sequencing of the entire 1.5kB gene, however, this method was laborious and not suitable for routine use. As more work was conducted on the 16S rRNA gene, it was found that the gene consisted of hypervariable regions interspersed with conserved regions, and sequencing a number of the hypervariable regions (approximately 500bp) is sufficient for genus and species-level identification. The precise definition of sequence targets to produce shorter reads, coupled with the introduction of next-generation sequencing (NGS) platforms which allowed for a tremendous increase in output, resulted in an explosion of data on microbial discovery and characterisation. 2.3.3.1 Next-generation sequencing overview Also referred to as massively parallel sequencing, NGS allows for sequencing reactions to be carried out at a scale not possible to be achieved by traditional Sanger sequencing. In addition, it offers several advantages, such as the miniaturization of sequencing chemistries and higher throughput, leading to lower cost per sequenced base as well as reduction in time and manpower. The principle is based on segregation of target DNA strands (emulsified in lipid droplets as in Roche 454-sequencing or attached to flow cell surface as in Illumina sequencing), followed by controlled binding of nucleic acids and detection of bioluminescence to monitor pyrophosphate 28 release. Currently, Roche's GS-FLX+ system allows for a throughput of more than 1,000,000 reads in a single run within 23 hours while Illumina's Hiseq system produces 176Gb of data within 40 hours easily, with both platforms achieving more than 10X coverage when sequencing human genomes (Roche, 2013; Illumina, 2013). 2.3.3.2 Application to metagenomics and pathogen detection NGS is mostly applied to re-sequencing studies, such as the metagenomics analysis of bacterial and viral diversity in environmental samples (Bibby et al., 2010, 2011; Kristiansson et al., 2011; Ye and Zhang, 2011). This is achieved by extracting genomic DNA and converting it into a NGS library, followed by sequencing. The reads are aligned to known reference sequences for microorganisms that are thought to be present within the sample. Because of the relatively short read lengths (400-500bp) produced by sequencing as well as the availability of large, well-curated public databases of the 16S rRNA gene, most applications have focused on the sequencing of selected hypervariable regions within this gene. Examples of metagenomics studies include the sequencing of microbial diversity in oceans, soils, human gut microbiota and oral cavity (Sogin et al., 2006; Huber et al., 2007; Keijser et al., 2008; Turnbaugh et al., 2009; Urich et al., 2008). NGS can also be applied to detection of pathogens in environmental waters using similar procedures as those involved in the study of microbial ecology, which includes concentration of bacteria, viruses or protozoa, followed by DNA/RNA extraction, pyrosequencing and bioinformatics analysis. Using this approach, it is possible to detect novel pathogens in an unbiased, high-throughput manner that will otherwise not be discovered by PCR and microarray methods where investigators are limited by sequence information and must decide on which known pathogen groups to target (Aw and Rose, 2012). Studies that have utilised NGS techniques for pathogen discovery include the discovery of a novel arenavirus transmitted through solidorgan transplantation, 29 2.3.3.3 Bioinformatics To ensure the accurate classification and identification of pathogens, it is necessary to consider the challenges of downstream bioinformatics analysis of the data output. Starting from sequence collection, algorithms should be used to trim off primer sequences and filter off poor quality reads which have a higher chance of containing ambiguous bases or to be aberrantly long or short (Huse et al., 2007). Alignment of sequences can be conducted using alignment programs such as Clustal-W, MUSCLE or NAST (Thompson et al., 1997; DeSantis et al., 2003; Edgar, 2004). Chimeric-sequences may be generated as a result of PCR errors that couple individual DNA strands during amplification; as such they should be removed via tools such as Bellaphon (Huber et al., 2004) or online programs available in Greengenes and RDP. After the obtainment of highquality data, sequences may then be identified using classifiers such as the Bayesian Classifier in RDP or prokMSA in Greengenes, and then screened for the presence of pathogenic species. During metagenomics characterisation of previously unsequenced microbial habitats, the species diversity may be of interest to the researcher. As such there are multiple indices that have been developed for this purpose. These indices include the Chao and Ace estimates of species richness, the Simpson and Shannon indices that measure the abundance and evenness of species present, as well as rarefaction curves that accesses species diversity as a result of sampling and allows one to determine whether the amount of obtained reads from sequencing is sufficient to encompass the true diversity that is present. 30 Chapter 3: Quantifying molecular markers as indicators of waterborne pathogens 3.1 Introduction Traditional methods of water quality assessment relied on conventional culture techniques to enumerate indicators of fecal contamination such as total and fecal coliforms, Enterococcus sp. and E. coli. These methods are not necessarily indicative of pathogens that are introduced via extra-enteric pathways such as urban runoff and human bather load or that have different levels of persistence and transportation routes within the water system. There are also several flaws in the existing indicator approach such as a lack of a consistent strong relationship between the pathogen and the indicator, as well as a necessary delay between sample collection and results analysis because of culture requirements. In Singapore, assessment of recreational waters from aquatic facilities is based on such methods despite the lack of data documenting the applicability of these guidelines in our context. Quantifying bacterial markers using qPCR has proven to be a viable option against traditional molecular methods. Several research groups have proposed the usage of certain molecular markers such as Bacteroidetes thetaiotaomicron alpha-mannanase gene and Methanobrevibacter smithii nifH gene as indicators for fecal contamination (Yampara-Iquise et al., 2008; Johnston et al., 2010), however these methods have not been tested on recreational waters from aquatic facilities. Furthermore, qPCR suffers from the inability to discriminate between viable and dead cells, and are hence may not be indicative of the true risk of contracting diseases through exposure to polluted waters. Even though DNA intercalating dyes such as PMA and EMA are proposed as a solution by suppressing signals produced by dead cells within the sample, the extent of such suppression effects is unknown and can vary considering the complex nature of the environmental matrix. In addition, environmental samples typically contain PCR inhibitory 31 substances that may affect the amplification efficiency of qPCR, and any attempts at designing a qPCR assay should take this into account. This section will focus on the work that has been performed to assess the effectiveness of some proposed bacterial markers on interactive water fountains in Singapore. In particular, it will cover the development of real-time assays targeting several commonly used and alternative indicators, as well as pathogens that are believed to be often found in environmental waters. This is followed by assay optimisation tests including a comparison between the effects of PMA and EMA, the effects of using different concentrations of PMA on laboratory-grade water as well as environmental matrix, and the determination of possible inhibitory effects from the collected samples and the environmental matrix. Lastly, it will determine the correlation and predictive ability of tested indicators against known pathogens found in several interactive fountain water samples. 3.2 Materials and methods 3.2.1 Sample collection and processing 3.2.1.1 Study sites Water samples were collected from the following locations:  Bugis Junction (BJ) Mall Level 1 outdoor interactive fountain  City Square (CS) Mall Level 1 outdoor interactive fountain  IMM building Level 3 Garden Plaza playground  Sembawang Shopping Centre (SB) Level 3 outdoor playground  Science Centre Singapore (SC) Level 1 Waterworks  VivoCity (VC) Level 3 wading pool 32 The geographical distribution and individual site images are shown in Figures 3.1 and 3.2 respectively while Table 3.1 shows the characteristic of each recreational facility. Sampling conditions were kept consistent with regards to time of collection and weather condition, and was avoided on days during or after heavy rainfall. Figure 3.1: Geographical locations of sampling sites. Map data © 2014 Google, MapIT. Captions have been added by the author. 33 A B C D E F Figure 3.2: Individual site images - (A) BJ, (B) CS, (C) IMM, (D) SB, (E) SC, (F) VC. (B) has been adapted from kidsandparenting.com (Kids and Parenting, 2013); all other images are author's own. Table 3.1: Location characteristics of sampled recreational water facilities. Location Type of facility Water system Exposure status Bugis Junction (BJ) Fountain Recycled Open City Square (CS) Fountain Recycled Sheltered IMM Water playground Recycled Open Sembawang (SB) Water playground Recycled Sheltered Science Centre (SC) Water playground Recycled Sheltered Wading pool Recycled Open VivoCity (VC) 34 3.2.1.2 Sampling procedures Water samples were collected according to the recommended procedures in Standard Methods for the Examination of Water and Wastewater (SMEWW), Section 9060 (APHA, 1988). Preliminary results (data not included) had determined that 1L sampling volume was sufficient to capture most microorganisms without significant impact on water flow during membrane filtration; as such sampling volume was standardised to 1L for the duration of this study. Samples were collected either from the water reservoir or through water nozzle jets, depending on whether the facility was a fountain/wading pool, or an outdoor playground, respectively. Samples were collected using a pre-autoclaved 1L measuring beaker, and transferred into pre-autoclaved sterile plastic screw-cap bottles (Thermo Scientific, U.S.A) or disposable Whirl-Pak® Stand-up Bags (Nasco, U.S.A). Commercial sodium thiosulfate pills (Nasco, U.S.A) used to dechlorinate residual chlorine levels were added to water samples to give a final concentration of 120mg/L. Samples were placed on ice-packs and transported using cooler bags to the laboratory, and were processed within 6hrs of collection. 3.2.1.3 Chemical tests Temperature was measured on-site using a Multi-parameter PCSTestr™ 35 probe (Eutech Instruments, U.S.A). Measurement of other water parameters were conducted in the laboratory on duplicate 50ml water samples with no added sodium thiosulfate. Total Dissolved Solids (TDS), salinity, conductivity and pH were measured using the PCSTestr™ 35 probe (Eutech Instruments, U.S.A). In addition, specific colorimetric test strips (Waterworks, U.S.A) were used to test for hardness, total and free chlorine as well as alkalinity. All procedures were performed according to manufacturer‟s instructions. 3.2.1.4 Membrane filtration 35 Membrane filtration was carried out according to the recommended procedures in SMEWW Section 9222 (APHA, 1988). Sample filtration was carried out using a pre-autoclaved filter unit connected to a vacuum pump. Water samples were pooled and mixed thoroughly, and subsequently filtered through a 0.22µm nitrocellulose membrane (Merck Millipore, Germany) under aseptic conditions. A pre-filtration step was conducted for turbid water samples. This step was performed if the flow rate was less than 1L in 30min through a 0.22µm nitrocellulose membrane (arbitrary value chosen in consideration to the amount of water sample that can be processed within 6hrs of collection). Samples were first filtered aseptically through a 3.0µm cellulose nitrate filter (Sartorius, Germany) into a clean, autoclaved bottle to remove large debris and particulate matter, before the filtrate was again filtered through a 0.22µm nitrocellulose membrane. The 3.0µm membrane for pre-filtration membrane was discarded while the 0.22µm nitrocellulose membrane was retained for culture analysis or DNA extraction. 3.2.1.5 DNA extraction from water samples Membranes were stored in 5ml RapidWater™ Bead Tubes (MO BIO, U.S.A.) at -20°C until ready for DNA extraction. DNA extraction was performed using RapidWater™ DNA Isolation Kits (MO BIO, U.S.A.) according to manufacturer‟s instructions. 3.2.2 Real-time polymerase chain reaction (qPCR) analysis 3.2.2.1 Reference strains used in standards All bacterial strains used in this study were obtained from the American Type Culture Collection (ATCC, U.S.A.), with the exception of Aeromonas hydrophilia which was obtained from the Department's culture stock. A. hydrophilia and E. coli (ATCC 11775) were cultured in nutrient broth (Oxoid, U.S.A.) at 37°C. Enterococcus faecalis (ATCC 19433) was cultured in brain heart infusion broth (Oxoid, U.S.A.) at 37°C. Pseudomonas alcaligenes (ATCC 53877) and Staphylococcus aureus (ATCC 6538) were cultured in tryptic soy broth (Oxoid, U.S.A.) at 28°C. 36 Confirmation of culture identity was performed using the appropriate API test kits (bioMérieux SA, France) prior to standard development. Total bacterial counts were determined by serial dilution of each culture strain grown in the appropriate culture broth and conditions (as described above) overnight. 100µl portions of each dilution were spread on the following agar culture medium: nutrient agar (Oxoid, U.S.A.) for A. hydrophilia and E. coli, tryptic soy agar (Oxoid, U.S.A.) for P. alcaligenes, and Baird-Parker agar (Becton, Dickinson and Company [BD], U.S.A.) for S. aureus. Each dilution was performed in triplicates. Colonies were counted to determine the CFU counts per milliliter of culture. DNA extraction was performed using an in-house protocol detailed in Appendix III. 3.2.2.2 Development of plasmid standards Plasmid standards were developed for bacteria which were difficult to obtain or culture (Bacteroidetes thetaiotaomicron, Legionella pneumophila and Methanobrevibacter smithii). The targeted gene was checked using NCBI Basic Local Alignment Search Tool (BLAST) program (Altschul et al. 1990), and subsequently sent for synthesis (Integrated DNA Technologies, U.S.A.). E. coli DH5α competent cells were transformed with the 1:1000-diluted synthesized plasmid and plated on Luria Bertani (LB) agar supplemented with ampicillin as selection medium incubated at 37°C overnight (Appendix II). Colonies were picked and grown in LB-A broth at 37°C overnight in 10ml culture tubes (Greiner bio-one, Belgium). Cultures were harvested by transferring into 2ml microcentrifuge tubes (Axygen, Corning, U.S.A.) and subsequently pelleted down. Plasmid extraction was performed using the E.Z.N.A.® Plasmid Miniprep Kit (Omega Bio-Tek, U.S.A.) according to manufacturer‟s instructions. The purified plasmid was quantified by the Nanodrop™ 2000 Spectrophotometer (Thermo Scientific, U.S.A.). Plasmid copy number was estimated using an online calculator (Staroscik, 2004). The extracted plasmids were stored at -20°C until further use. 37 3.2.2.3 Validation of primer set using end-point PCR Selected primers for this study target common (Enterobacteriaceae and Enterococcus) and alternative (M. smithii and B. thetaiotaomicron) indicators of fecal pollution, as well as markers of potential pathogens which are known to be present in environmental water samples (Table 3.2). Primers were tested for specificity by submitting and checking the sequences using the Probe Match program in the RDP website (Cole et al., 2008). Primers were further tested for specificity using end-point PCR on the total genomic DNA of the following organisms: A. hydrophilia, E. faecalis, E. coli, P. alcaligenes and S. aureus. Each 10-µl PCR reaction consisted of the following components: 0.05µl TaKaRa Ex Taq (5 units/µl), 1µl 10X Ex Taq Buffer, 0.4µl dNTP Mixture (2.5mM each), 0.4µl each of forward primer TX9 (10µM) and reverse primer 1391R (10µM), 1µl template DNA (1ng/µl) and 6.75µl milliQ water. PCR reaction was accompanied with a negative control replacing template DNA with 1µl milliQ water. Thermocycling of each primer set proceeded as follows: one cycle at 95°C for 15s, 30 cycles of denaturation at 94°C for 20s, appropriate annealing temperature (Table 3.2) for 20s and extension at 72°C for 50s, and lastly one cycle of final extension at 72°C for 10min. PCR products were checked for size and specificity using a diagnostic 2% (w/v) agarose gel, for amplicon sizes less than 100bp, or a 1% (w/v) agarose gel, for amplicon sizes above 100bp (Table 3.2), and confirmed to have only a single band present at the target group (Figure 3.3). 38 Figure 3.3: Example of gel used to check PCR product size and primer specificity. Image shows amplification results of various bacteria groups using Leg16S primer pair, showing only a single band present in Lane 2 (L: L. pneumophila). Lanes 3-8 (Ec: E. coli, Ef: E. faecalis, S: S. aureus, A: A. hydrophilia, P: P. alcaligenes, V: Vibrio parahaemolyticus. Lane 8 represents negative control (-ve). Imaging faults in the machine could cause white areas to be visualised in some gels (such as that found in '-ve'); they are identified as artifacts based on shape (usually round or irregular) and size (usually a distance from expected band position). 39 Table 3.2: List of gene-targeted primers used in this study. Target group (gene) Pathogens Aeromonas (16S rRNA) Legionella (16S rRNA) Mycobacterium (Ag85) Pseudomonas (16S rRNA) Escherichia coli (uidA) Staphylococcus (16S rRNA) Indicators B. thetaoitaomicron (α-mannanase) Enterobacteriaceae (16S rRNA) Enterococcus (23s rRNA) M. smithii (nifH) Primer name Primer sequence (5’-3’) Aer66f Aer613r Leg16S1-A Leg16S2-A BB-1 BB-3 PSD7F PSD7R uidA1318F uidA1698R g-Staph-F g-Staph-R Bac-alpha-man-F Bac-alpha-man-R En-lsu-3f En-lsu-3r ECST784F ENC854R Mnif-H-F Mnif-H-R Annealing temp (°C) PCR product size (bp) Reference GCGGCAGCGGGAAAGTAG GCTTTCACATCTAACTTATCCAAC AGGGTTGATAGGTTAAGAGC CCAACAGCTAGTTGACATCG ATCAACACCCCGGCGTTCGAG CGGCAGCTCGCTGGTCAGGA CAAAACTACTGAGCTAGAGTACG TAAGATCTCAAGGATCCCAACGGCT CCGATCACCTGTGTCAATGT GTTACCGCCAACGCGCAATA TTTGGGCTACACACGTGCTACAATGGACAA AACAACTTTATGGGATTTGCWTGA 61 ~500 Yu et al. 2005 57 386 Joly et al. 2006 60 162 Fauville-Dufaux et al. 1992 60 215 Matsuda et al. 2009 60 ~400 Bower et al. 2005 60 79 Matsuda et al. 2009 CATCGTTCGTCAGCAGTAACA CCAAGAAAAAGGGACAGTGG TGCCGTAACTTCGGGAGAAGGCA TCAAGGACCAGTGTTCAGTGTC AGAAATTCCAAACGAACTTG CAGTGCTCTACCTCCATCATT GAAAGCGGAGGTCCTGAA ACTGAAAAACCTCCGCAAAC 60 63 Yampara-Iquise et al. 2008 60 428 Matsuda et al. 2009 60 ~100 Frahm and Obst 2003 57 151 Johnston et al. 2010 40 3.2.2.4 Validation of primer set using real-time PCR Before proceeding with absolute quantification of target DNA in samples, each primer set was tested for efficiency and sensitivity using qPCR. Standards were subjected to 10-fold dilutions and amplified by qPCR in duplicates under appropriate annealing temperature (Table 3.1). Meltcurve analysis as well as gel electrophoresis was used to validate the purity of PCR product. A standard curve was generated for each primer set based on Cycle Threshold (CT) values. Primer efficiency (%) was calculated according to the following formula: −1 𝐸 = (10 𝑚 ) − 1 × 100 where m refers to the slope of the standard curve. The sensitivity of primers were tested by determining the limit of detection (LOD). This was achieved by performing a dilution series, and obtaining the highest CT value at which the data point remains at the log-linear range of the standard curve. Amplification efficiency and limits of detection of each primer set are shown in Table 3.3. All primer pairs were verified to have good amplification efficiency (ranging between 90-110%). Sensitivity was shown to be slightly lower for primer pairs amplifying Staphylococcus, Enterobacteriaceae and Enterococcus. Considering that these bacteria groups are found in high densities in gut microbiota (Staphylococcus: 105 cells/g; Enterobacteriaceae and Enterococcus 106 to 109 cells/g), contamination of fecal material in water sources should introduce substantial amounts; as such primer sensitivity was not deemed as limiting and we proceeded to work with identified primer pairs (Matsuda et al., 2009). 41 Table 3.3: Validation of amplification efficiency and lowest limit of detection of primer set. Bacteria group (gene) Pathogens Aeromonas (16S rRNA) Legionella (16S rRNA) Mycobacterium (Ag85) Pseudomonas (16S rRNA) Escherichia coli (uidA) Staphylococcus (16S) Indicators B. thetaoitaomicron (α-mannanase) Enterobacteriaceae (16S rRNA) Enterococcus (23s rRNA) M. smithii (nifH) Primer name Amplification efficiency (%) Limit of detection (No. gene targets/µl) Aer66f Aer613r Leg16S1-A Leg16S2-A BB-1 BB-3 PSD7F PSD7R uidA1318F uidA1698R g-Staph-F g-Staph-R 90 2 94 5 96 8 99 2 98 4 94 158 Bac-alpha-man-F Bac-alpha-man-R En-lsu-3f En-lsu-3r ECST784F ENC854R Mnif-H-F Mnif-H-R 97 10 107 33 106 78 96 3 3.2.2.5 Real-time PCR of samples Real-time PCR was performed in 96-well optical plates (Scientific Specialities, Inc., U.S.A.) using an ABI PRISM 7500 sequence detection system (Applied Biosystems, U.S.A.). Each 20-µl reaction was composed of the following reagents: 10µl Maxima SYBR Green/ROX qPCR Master Mix (2X) (Thermo Scientific, U.S.A.), 0.4µl each of forward and reverse primer (10µM each), 2µl template DNA and 7.2µl milliQ water. Each run consisted of samples run in triplicates, serial dilutions of standard DNA in triplicates, as well as a single non-template control (NTC). Thermocycling conditions were as follows: one cycle at 50°C for 2min, one cycle of denaturation at 95°C for 10min, 40 cycles of denaturation at 95°C for 20s, appropriate annealing temperature (Table 3.2) for 20s and extension at 72°C for 50s, followed by melt-curve analysis to distinguish the target DNA from other non-specific PCR products. Samples were progressively denatured by 42 increasing temperatures from 75°C to 95°C in 0.2°C increments, with detection of fluorescence dye emission during each temperature increment, to generate melt curve profiles. 3.2.2.6 Data analysis Cycle thresholds (CT values) were automatically determined by the ABI PRISM 7500 operating system. The standard curve was derived from the serial dilutions of standards from each qPCR run. Quantification was performed by comparing the average CT value of each sample against the standard curve. Data analysis was performed using SPSS® Statistics 20 (IBM, U.S.A.). Spearman‟s correlational test for non-parametric data was used to determine the level of correlation between bacteria and water parameters. The significance level was set at 95% confidence level. 3.2.3 Differentiation between live/dead bacteria 3.2.3.1 Culture strain E. coli (ATCC 11775) was used for all optimization experiments involving PMA and EMA. E. coli was cultured in sterile 15ml Falcon tubes (Fisher Scientific, U.S.A.) containing 10ml of nutrient broth (Oxoid, U.S.A.) at 37°C overnight prior to experimentation. 3.2.3.2 Preparation of dead cells E. coli were heat-killed by heating the cells to 100°C in a boiling water bath for 5min. Lack of viability was confirmed by plating 100µl of heat-killed cells on nutrient agar (Oxoid, U.S.A.) and incubated at 37°C overnight. 3.2.3.3 Comparing PMA and EMA pre-treatment Following the manufacturer's instructions of using 20mM of PMA (Biotium, U.S.A) and the recommendations of Nocker and Camper (2006) of using 5mg/ml of EMA as working 43 concentrations, both were dissolved with nuclease-free water to their respective concentrations, filter-sterilised using a 0.45µm filter and aliquoted into a foil-wrapped microcentrifuge tube for storage at -20°C. E. coli was grown overnight and a portion heat-killed prior to spiking different concentrations of 'live' and 'dead' bacteria into ten sterile bottles each containing 100ml of autoclaved milliQ water. The spiked samples were filtered aseptically through a 0.22µm nitrocellulose membrane, and treated with either EMA or PMA. Treated membranes were placed aseptically on a sterile 45mm transparent petri dishes. They were first incubated in the dark at room temperature for five minutes, followed by exposure to a 500W halogen light source for an additional five minutes. To prevent degradation of sample due to heat generated from the light source, the samples were kept on ice during light exposure. After treatment, the membranes were stored in RapidWater™ Bead Tubes (MO BIO, U.S.A.) at -20°C until ready for DNA extraction. DNA extraction was performed using RapidWater™ DNA Isolation Kits (MO BIO, U.S.A.) according to manufacturer‟s instructions. The extracted sample was quantified using qPCR targeting the E. coli uidA gene. 3.2.3.4 Determining and validating optimum PMA concentration As per the previous experiment, PMA was dissolved with nuclease-free water to a final concentration of 1.779mM, filter-sterilised using a 0.45µm filter and aliquoted into a foilwrapped microcentrifuge tube for storage at -20°C. Different PMA concentrations (10µM, 20µM, 30µM, 40µM, 50µM) were added into individually foil-wrapped tubes containing 1ml nuclease free water. Heat-treated E. coli was mixed with live cultures in defined ratios, with live cells representing 0%, 25%, 50%, 75% and 100% of the total volume. These mixtures were spiked into 50ml autoclaved milliQ water, prior to filtering through a 0.22µm nitrocellulose membrane. PMA treatment, DNA extraction and qPCR analysis was performed as described in the previous section. In addition, the amount of viable bacteria was also determined through serial dilution followed by spread plating on nutrient agar. All samples were performed in duplicates. 44 To validate the chosen PMA concentration, the same experimental set-up was repeated using environmental matrix in replacement of milliQ water. Specifically, water was collected from two different interactive fountains (VC and BJ) within a 1-week period, filtered under aseptic conditions through a 0.22µm nitrocellulose filter and stored at 4°C. The water was combined and mixed to homogenisation before autoclaving at 121°C for 15min, and subjected to UV treatment for 15min. All other aspects of the experiment remain the same. 3.2.4 PCR inhibition From the literature survey, the easiest way to remove the effect of inhibitory substances is through dilution of the environmental samples. As such, point samples obtained from BJ, IMM and VC were filtered and extracted according to the procedures described above. The extracted samples were diluted 10X and 100X prior to quantification of the Enterobacteriaceae 16S rRNA gene by qPCR. This gene was chosen because the Enterobacteriaceae group of bacteria is known to be highly prevalent in environmental samples and the expected high numbers allows for visualisation of any inhibitory effects more easily (Gavini et al., 1985). Quantification was performed by comparing against a standard curve using known E. coli concentrations and any differences in counts obtained between different dilutions were calculated using statistical tools. 3.2.5 Field tests 3.2.5.1 Pathogen detection Samples were collected from each of the six targeted recreational water facilities (BJ, CS, IMM, SB, SC, VC). The sampling timeframe was kept within a period of two months to reduce the confounding effect of seasonal changes. Membrane filtration and DNA extraction was performed as reported previously. Based on existing literature, six potential pathogens (Aeromonas sp., E. coli, Legionella sp., Mycobacterium sp., Pseudomonas sp. and Staphylococcus sp.) were 45 shortlisted for detection and enumeration by qPCR, using standards and selective primers as reported in Section 3.2.2. 3.2.5.2 Evaluating pathogen-indicator relationship Samples were collected from each of the six targeted recreational water facilities (BJ, CS, IMM, SB, SC, VC) from January 2012 to April 2013. Membrane filtration, DNA extraction and chemical testing was performed as previously described. Real-time PCR was used to detect and quantify the targeted indicators (Enterobacteriaceae, Enterococcus, B. thetaiotaomicron, M. smithii) and pathogens (Aeromonas, Legionella, Mycobacterium, Pseudomonas). Correlations between indicator and pathogen were determined by calculating Kendall's tau correlation coefficient for non-parametric data. Possible indicator-pathogen correlations were further examined by multiple regression analysis with the indicator as a possible predictor of each pathogen, controlling for the effects of water parameters including temperature, TDS, salinity, pH, conductivity, total and free chlorine. 3.3 Results 3.3.1 Comparing the effectiveness of PMA and EMA pre-treatment Both PMA and EMA have been reported to be effective in differentiating between viable and non-viable cells and have been successfully applied in combination with qPCR to selectively quantify viable cells in mixed environmental samples. However studies have also reported that while these DNA intercalating dyes significantly suppresses counts obtained from dead cells, it also penetrates live cells and results in the generation of false negative signals in the presence of live cells (Nocker and Camper, 2006). In this study, PMA and EMA were evaluated for their potential application in differentiating between mixed cultures of live and dead E. coli cells. It was found that PMA-treated set-ups 46 showed a signal suppression of more than 6 log scales in the presence of dead cells (Figure 3.4: P2), demonstrating its effectiveness in reducing the counts obtained by non-viable bacteria. This is in comparison to EMA-treated set-ups where dead cell signal suppression was less than 4 log scales and amplification was observed even in the absence of both live and dead cells (Figure 3.4: E2). Accuracy of quantification of live cells was similar for PMA-treated and EMA-treated setups, which showed a signal suppression of approximately 1-2 log scales in comparison to the calculated amount of spiked live cells in both cases (Figure 3.4). Gene targets/100ml Comparison between PMA and EMA treatments 1E+09 10000000 10000000 1000000 100000 10000 1000 100 10 1 Dead Live qPCR count P1 P2 P3 P4 P5 E1 E2 E3 E4 E5 Figure 3.4: Comparison between PMA and EMA treatment of viable and heat-killed E. coli mixtures. P1-5 refers to 5 set-ups treated with PMA while E1-5 refers to 5 set-ups treated with EMA. 'Dead' and 'Live' refers to calculated amount of spiked live and dead cells respectively while 'qPCR count' is the count obtained from the CT value after qPCR quantification. 3.3.2 PMA optimisation Optimisation of PMA concentration was performed for both live and dead cultures at a concentration of 2.8x105 CFU/100ml (confirmed by plate counts), which more closely corresponds to that found in environmental water samples. Figure 3.5A shows that in live cells, increased concentrations of PMA up to 20µM resulted in a reduction of bacteria count by about 3-fold as reported from the PMA-qPCR method, before the effect starts to plateau. This is 47 supported by data published by Bae and Wuertz (2009) that there was significant variability in cell counts at low feces concentration, and suggests a relationship between PMA concentration and greater likelihood of PMA penetration into undamaged cells when there is lower target numbers for PMA-qPCR. In heat-killed cells, PMA treatment resulted in a substantial decrease of bacteria count irregardless of PMA concentration, possibly indicating that a lower PMA concentration is desirable when working with diluted samples. Figure 3.5B shows the level of discrimination between variable ratios of live to dead cell cultures at different PMA concentrations, in comparison to standard culture. Low PMA concentrations (10µM) were not able to suppress signals contributed by dead bacteria, while high PMA concentrations (40-50µM) showed evidence of reduced viable bacteria cell count, especially at higher live cell concentration. In fact, only a PMA concentration of 30µM closely matched the count that is obtained from standard culture methods, as there was very little deviation in the higher live cell concentration (50% and 75% live cells) and only a 2-fold decrease at a lower live cell concentration (25% live cells). A Live versus Dead bacteria counts 1.00E+07 CFU/100ml 1.00E+06 1.00E+05 Live 1.00E+04 Dead 1.00E+03 1.00E+02 1.00E+01 1.00E+00 Culture PMA 10uM PMA 20uM PMA 30uM PMA 40uM PMA 50uM 48 CFU/100ml B Profile of Live/dead mixtures 2.00E+05 1.80E+05 1.60E+05 1.40E+05 1.20E+05 1.00E+05 8.00E+04 6.00E+04 4.00E+04 2.00E+04 0.00E+00 25% live 50% live 75% live Culture PMA 10uM PMA 20uM PMA 30uM PMA 40uM PMA 50uM Figure 3.5: Chart showing the relationship between different PMA concentrations and enumeration of E. coli spiked in pure water. qPCR of the Enterobacteriaceae specific 16S rRNA gene as well as the standard culture method was performed for A) live and dead cells, B) Defined live/dead cell mixtures. 3.3.3 Validation of PMA pre-treatment in environmental water matrix It has been suggested that certain factors found in environmental samples present practical and theoretical limitations to the application of PMA in the detection of viable cells. For example, PMA may penetrate viable cells with reversibly damaged membranes, which is likely to be present in environmental samples. Additionally, turbidity resulting from high levels of suspended solids or biomass may result in shading and hence inhibition of the PMA-DNA cross-linkage step. As such, an additional experiment was performed to investigate the effect of factors present in environmental water samples on the results of the PMA-qPCR analysis method. Similar to the previous experiment performed on PMA optimisation using MilliQ water (Section 3.3.2), treatment of heat-killed cells using PMA resulted in substantial suppression of bacteria cell count irregardless of concentration, whereas amplification was observed for the non-treated heatkilled cells (Figure 3.6A), indicating the effectiveness of PMA treatment in removing signals contributed by dead cells. Counts obtained from the non-treated heat-killed cells were two logfold lower than the amount spiked, suggesting that other factors such as the manner of heat- 49 treatment or the extraction procedures could also have contributed to the lower amounts obtained. Live-cell counts peak in the mid-ranges of PMA concentration suggesting PMA concentration of 30uM is optimal. Interestingly, PMA treatment resulted in suppression of qPCR count (approximately two log-fold units) in environmental samples in comparison to the non-treated control, irregardless of PMA concentration (Figure 3.6B). These results indicate that certain factors within environmental samples could interfere with the PMA treatment or downstream DNA extraction procedure thus leading to anomalous bacterial counts. A Live versus Dead bacteria counts 1.00E+07 CFU/100ml 1.00E+06 1.00E+05 1.00E+04 Live 1.00E+03 Dead 1.00E+02 1.00E+01 1.00E+00 Culture PMA PMA PMA PMA PMA Not 10uM 20uM 30uM 40uM 50uM treated 50 B Profile of live/dead mixtures 1.00E+07 CFU/100ml 1.00E+06 1.00E+05 1.00E+04 25% live 1.00E+03 50% live 1.00E+02 75% live 1.00E+01 1.00E+00 Culture PMA PMA PMA PMA PMA Not 10uM 20uM 30uM 40uM 50uM treated Figure 3.6: Chart showing the relationship between different PMA concentrations and enumeration of E. coli spiked in environmental water matrix. qPCR of the Enterobacteriaceae specific 16S rRNA gene as well as the standard culture method was performed for A) live and dead cells, B) Defined live/dead cell mixtures. 3.3.4 PCR inhibition PCR detection methods are essential in this study to detect organisms that are not possible to culture, however results may be compromised by the presence of inhibitors, including humic acid, organic matter and clay in environmental water samples. Because the targeted recreational water sources for this study are mostly sheltered and are less exposed to environmental factors such as wind and rain, the extent of possible PCR inhibition is not understood. In this study, the possibility of PCR inhibition in samples collected from interactive fountains was investigated. A one-way Analysis of Variance test (ANOVA) was used to determine count difference among different dilutions of DNA extracts obtained from different interactive water fountain locations. Results showed that there was significant differences only in counts obtained for IMM: F(2,6) = 20.906, p 1 which explained for around 78% of the total variance, where PC1 took up 32% and the rest took up between 5 to 15%. The scree plot showed a sharp change of slope after the PC1 (Figure 4.6), indicating that most of the variance had been explained by variables in the PC1. Strengths of factor loadings were determined based the classification of factor loadings by Liu et al. (2003) as 'strong', 'moderate' and 'weak' corresponding to loading values > 0.75, 0.5-0.75 and 0.3-0.5 respectively. Varimax 76 rotation of the PCs showed moderate to strong positive factor loadings of each bacteria group in different VF, indicating that these bacteria groups do not cluster and are unlikely to be correlated with each other (Table 4.3). TDS, salinity, conductivity, total hardness and alkalinity were found to cluster together in PC1 indicating the presence of an underlying factor (e.g. physical water quality) governing their correlation. However, these variables only accounted for 32% of the variance in the data, implying that there were other factors responsible for the difference in bacterial counts. 77 A B Figure 4.5: Temporal patterns of distribution for bacteria markers enumerated using (A) culture and (B) qPCR. Red arrows refer to the date in which 'thorough cleaning' (defined as filter backwash and dosing of disinfectant) was performed, while green arrows refer to the date in which 'routine cleaning' (defined as clearing of leaf debris and change of fountain water) was done. The blue arrow indicated a time point meant for 'routine cleaning' which fell on a public holiday. 78 Table 4.2: Regression analyses and correlations of indicator and potential pathogens. Molecular methods (qPCR) Enterobacteriaceae (All time points) Aeromonas Legionella Mycobacterium Pseudomonas B 0.498 -1.744 2.097 -0.004 Linear Regression S.E. t 0.740 0.674 1.357 -1.286 6.001 0.349 0.025 -0.162 2.321 3.347 -0.377 -0.010 1.203 2.455 5.847 0.014 1.930 1.363 -0.064 -0.712 N.S. N.S. N.S. N.S. 0.444 0.444 0.001 0.056 N.S. N.S. N.S. N.S. 0.025 0.018 0.014 0.019 1.783 0.911 N.S. N.S. 0.418 0.242 p < 0.005 p < 0.005 p-value N.S. N.S. N.S. N.S. Kendall's correlation p-value  0.402 p < 0.005 0.038 N.S. 0.283 p < 0.05 0.223 N.S. (Partial time points) Aeromonas Legionella Mycobacterium Pseudomonas Culture methods P. aeruginosa Coliforms Enterococcus Figure 4.6: Scree plot showing a sharp drop after PC1, indicating that most of the variance in the data had been explained by the first component. 79 Table 4.3: Variable loadings on the first five rotated principle components Variables Enterobacteriaceae Pseudomonas Aeromonas Mycobacterium Legionella Temperature TDS Salinity pH Conductivity Hardness Total Chlorine Free Chlorine Total Alkalinity VF1 0.094 0.282 0.479 0.446 -0.552 0.214 0.941 0.941 -0.122 0.939 0.850 VF2 -0.370 0.211 0.152 -0.420 0.364 0.828 0.040 0.041 0.325 0.047 -0.164 -0.104 0.182 -0.201 0.648 0.671 -0.579 VF3 0.479 -0.040 0.724 0.004 0.184 0.033 0.218 0.221 -0.335 0.219 0.118 -0.179 -0.251 -0.064 VF4 0.028 0.764 0.004 -0.133 -0.351 0.139 0.151 0.147 -0.752 0.152 0.113 -0.123 -0.199 0.108 VF5 0.023 0.139 0.096 0.594 0.211 0.009 0.018 0.020 0.045 0.014 -0.099 -0.861 0.030 0.321 4.3.3 Marker decay kinetics Microcosms are often used to investigate environmental processes because they are simulations of the environmental condition while allowing the researcher to modify and thus study the effect of individual parameters (Klein et al., 2011; Schulz and Childers, 2011). Typically the data obtained is based on the measurement of bacteria marker deactivation when subjected to different forms of treatment, and the results act as an indication of marker behaviour and fate in the environment. Such comparison of pathogen and indicator decay rates could provide significant insights into their relationship under the influence of each studied variable. We suspected that currently tested indicators did not correlate well with pathogen distribution in the field because of marker persistence issues (differential bacterial die-off/marker decay rates). As such, a laboratory closed microcosm was set up and monitored over a period of 7 days without any input of biomass or nutrient source, in order to determine bacterial marker decay rates. There were two types of survival profiles for pathogens, with Pseudomonas and Aeromonas showing similar patterns of decay and a decrement of approximately 4 log units (k=1.45 and 1.63 respectively), while Legionella and Mycobacterium exhibited an absence of decay and the former 80 even showed a slight increment of 1 log units (k=-0.25 and 0 respectively). Indicators and pathogens exhibited different patterns of decay, with Enterobacteriaceae gene target counts showing an intermediate decay rate (k=0.88) in comparison to the two different pathogen profiles, while M. smithii appeared to be highly unstable and showed a drastic decline in count to undetectable levels after 1 day (k=5.28) (Figure 4.7 E-F). Overall, results suggested that there were variable patterns of decay among different pathogenic markers as well as between indicators and pathogens. Such differences in marker persistence could be a contributing reason to the conflicting results of pathogen-indicator relationships found in this study, as well as that reported in literature (Dorevitch et al., 2010). 81 Figure 4.7: Survival profiles of selected pathogens (A-D) and indicators (E-F). k represents the decay coefficient as calculated by Chick's Law. Error bars have been added to the charts but are too narrow to be visualised. . 82 4.4 Discussion 4.4.1 Spatial and temporal variability in bacterial marker distribution Environmental conditions are often complex and a variety of parameters could possibly introduce spatial and temporal variation of bacteria marker populations. Such factors include chemical and physical parameters such as temperature, use of disinfectant or filtering techniques, weather conditions as well as spatial differences between different contamination source and distribution pathways. These factors are often ignored in the development of indicators for fecal and non-fecal contamination despite the fact that they determine the reliability of any indicator-pathogen relationship observed. There were no observed difference in indicator and pathogenic marker counts irregardless of whether they were sampled from the nozzle, reservoir water or the filter region, suggesting that bacterial populations were well distributed through the entire water system and there was no significant localisation of any single microbial group. This showed that the measured bacteria groups were likely to possess similar transport mechanisms within the water system, which is one of the attributes of an ideal indicator (Section 2.2.2). Interestingly, the results also highlighted the possibility of the ineffectiveness of the sand filter at removing these indicators and pathogens from the water, or their constant introduction into the water supply either within the water system or from the surrounding environment. Considering that most of the measured bacteria were potentially biofilm-forming (Aeromonas, Pseudomonas, Mycobacterium) or could persist and regrow in sediments (Enterobacteriacea), it was likely that the contamination source originated from the water system, however the exact location is difficult to pinpoint without the aid of microbial source tracking techniques. Knowledge of such data will be useful from a regulatory standpoint but was less relevant for our purposes. 83 Temporal differences in bacterial marker distribution were present, however did not follow any significant trend corresponding to the disinfection and cleaning cycle, possibly explaining for the lack of correlation and predictive power of measured indicators towards pathogens. The absence of any relationship between the measured indicator and pathogenic markers was further confirmed through FA/PCA which showed that the measured indicator and pathogen variables were unable to be clustered and hence correlated together. The overall conclusion obtained from analysing results of bacteria marker distribution data over temporal scales using simple correlation, multiple regression and FA/PCA overturned that based on point sampling (Section 3.3.6). Results suggested that in our study, there was an absence of both physical and microbiological indicators that can act as reliable markers of pathogen occurrence in field settings when temporal fluctuations of bacterial population are considered, and the association between Enterobacteriaceae and Pseudomonas/Aeromonas as well as M. smithii/Legionella as observed from the single-point study was likely due to chance. To our knowledge, this is the first study that compares indicator-pathogen relationships using the same markers in both single-point and temporal studies, as such there is no literature to support the results that we have obtained. However it is reasonable to postulate that environmental variables such as sunlight and temperature affect bacterial physiology and growth kinetics differentially (different levels of persistence), and the combined effect is more significant than that contributed by routine filter backwash, cleaning and disinfection procedures. 4.4.2 Bacterial marker survival and persistence A laboratory microcosm decay study was established to document the survival profiles and measure the persistence of indicators and pathogenic markers under conditions mimicking that of the Bugis Junction water fountain. Pathogens generally have different survival profiles as indicators, implying that bacterial physiological response and growth is affected differentially 84 under the same conditions of temperature and light/dark cycle. In particular, the rapid decay of the M. smithii marker is probably a reflection of the bacterium's inability to tolerate aerobic environments, making it particularly unsuitable to indicate the presence of environmental pathogens that are aerobic and persist for extended periods in the water column such as those measured in this study. Enterobacteriaceae, though being shown to possess greater persistence than Pseudomonas and Aeromonas and less than Legionella and Mycobacterium, is also unsuitable as differential persistent rates results in uncoupling of the indicator-pathogen relationship and is likely a major contributing reason to their lack of correlation observed in the previous study. It is worrying that in our study Legionella and Mycobacterium are shown to remain viable and persist for extended periods of time (more than seven days). Both pathogens are highly resilient organisms that have been shown to colonise and persist in a variety of drinking water supplies even after treatment and disinfection, and their pathogenicity have been documented well in humans (Covert et al., 1999; Perola et al., 2003; Borella et al., 2005; Hilborn et al., 2006). Furthermore, both pathogens have been shown to enter and resuscitate from the viable-but-nonculturable (VBNC) state which impairs the ability to detect them using normal culture techniques currently employed by monitoring agencies. Their continued presence in recreational water fountains, and their extended viability and persistence in the water body represents a potential health risk as yet understudied by regulatory bodies. Although this study has shown differential persistence of bacterial indicators and pathogens in recreational water samples, it has not determined the environmental variables which are responsible. This is especially so for pathogens shown to possess high persistence as determination of the variables that affect/impede their numbers could lead to more efficient monitoring or even the development of measures that could suppress natural bacteria populations. Future work could include laboratory microcosm studies on the isolated effects of sunlight, 85 temperature, UV and predation on marker decay kinetics. Furthermore, the persistence of Legionella and Mycobacterium has not been well documented as it extends past the length of this study; it is suggested that future work should include a preliminary study to determine the timeframe required for these bacterial markers to decay prior to commencing any microcosm studies. As both organisms are biofilm-forming, it is also possible that selective localisation to PVP piping in aquatic water systems could protect them from sunlight and other environmental influences, as such accounting for differential persistence in comparison to indicator organisms; further research could also include the collection and analysis of swab samples for biofilm detection. Existing fecal indicators have been shown to be unsuitable for the prediction of the common pathogens present in recreational waters from aquatic facilities, however currently there are no microbiological indicators for non-fecal indicators available. Given that qPCR techniques are only able to quantify markers of known sequences selectively, screening and discovering novel indicators suitable for this purpose may prove to be a difficult and almost impossible task using this method. In fact, technological improvements have now provided us with the tools to detect and quantify pathogens directly, as such eliminating problems associated with establishing an indicator-pathogen relationship entirely. Chapter 5 will look into direct pathogen detection methods using sequencing techniques in selected interactive water fountain samples. 4.5 Conclusion Tested bacterial indicators and pathogenic markers have been shown to exhibit no statistical significance in spatial distribution showing that bacterial counts are homogeneous within the water system. However bacterial counts fluctuate independently of the effects of cleaning and disinfection schedules, implying either the differential effects they have on bacterial physiological response or the effect of other uncontrolled variables (possibly of environmental origin) on 86 bacterial count. As a result, no significant, strong and reliable correlations were observed between indicators and pathogens of concern. This lack of correlation could be due to the differential persistence of pathogens and indicators in the water column, with pathogens persisting for extended periods in comparison to indicators. The inability of fecal indicators to predict the presence of pathogens commonly found in interactive water fountains highlight the need for alternative measures of water quality in recreational water systems. 87 Chapter 5: Direct pathogen detection and discovery using sequencing methods 5.1 Introduction The indicator approach to signal the presence of pathogens was first developed in water quality monitoring because traditional culture methods did not possess the sensitivity to detect and identify low concentrations of pathogens directly from water samples. With the advent of molecular tools, it is now possible to detect a wide range of pathogens in water, and discover novel and potential pathogens including those that were previously difficult or impossible to cultivate. These molecular tools include DNA microarrays, real-time PCR and sequencing techniques, where the majority of methods developed focused on the use of large and wellcurated 16S rRNA databases to identify putative pathogens. DNA microarrays, while being able to simultaneously interrogate thousands of genes in a single assay, requires sequence information and does not possess the detection sensitivity required for detecting extremely dilute concentrations of pathogens commonly found in water samples. On the other hand, while realtime PCR possesses the sensitivity required for direct pathogen detection, it is also limited by sequence information as well as by the range of pathogens to be studied in a given assay, and is therefore useful mostly for routine monitoring of targeted pathogens. Chapters 3 and 4 have described work on testing indicator/pathogen relationships for selected bacterial groups, and concluded that the relationship did not hold up in field tests conducted on a recreational water facility. Furthermore, we have found that a number of microbial pathogens and indicators thought to be commonly found in water bodies were not present in our samples. It was recognised that there is much to learn about the true diversity of microorganisms, including that of the pathogens present in recreational waters (NRC, 2004; Pond, 2005). There is hence a necessity to look at alternative means of accessing water quality from an approach that allows the 88 researcher to investigate the entire spectrum of pathogens present in the water sample. Sequencing techniques including Sanger sequencing and high-throughput NGS present a solution to this problem. Traditional Sanger sequencing involves the cloning and sequencing of full or partial-length 16S rRNA gene in plasmid vectors and the subsequent construction of a clone library. NGS involves massive sequencing of short 16S rRNA reads, which can then be classified and identified according to taxonomy. Each method has its advantages and limitations as reviewed in Section 2.3.3. This section will focus on pathogen identification and discovery using both the Sanger sequencing and NGS approach. From each environmental sample, a defined region of the 16S rRNA gene will be amplified, sequenced and classified based on a total bacteria 16S rRNA reference database. The diversity and abundance of each sample will be compared and analyzed according to a reference pathogenic bacteria list. Sample preparation for 454-sequencing was performed by a fellow colleague, Dr Koh Yiling, Eileen; all other work described in this section are of author's own. 5.2 Materials and methods 5.2.1 Establishment of 16S rRNA clone library A clone library was established as detailed in the subsequent subsections, for the following locations: BJ, IMM, SC, VC. Each location was sampled twice, and membrane filtration and DNA extraction was conducted as detailed in Section 3.2. Approximately 100 clones from each environment were screened, sequenced and analysed to form the clone library. A summary of the methodology is shown in Figure 5.1. 89 16s rRNA (PCR) amplification and purification DNA ligation Transformation Verification of insert Plasmid extraction Picking and growing of white colonies RFLP screening Sequencing Phylogenetics analysis Figure 5.1: Flowchart summarizing the steps involved in the establishment of each clone library. 5.2.1.1 PCR amplification Partial length 16S rRNA genes were amplified in a standard PCR reaction using the universal primers TX9 (GGATTAGAWACCCBGGTAGTC) and 1391R (CCTATCCCCTGTGTGCCTTGGCAGTCTCAG), which amplified the V5-V8 variable region of the 16S rRNA gene (Figure 5.2). Reagents were obtained from TaKaRa Bio Inc. (Japan) except for primers and milliQ water. Each 10-µl PCR reaction consisted of the following components: 0.05µl TaKaRa Ex Taq (5 units/µl), 1µl 10X Ex Taq Buffer, 0.4µl dNTP Mixture (2.5mM each), 0.4µl forward primer TX9 (10µM), 0.4µl reverse primer 1391R (10µM), 1µl template DNA (1ng/µl) and 6.75µl milliQ water. PCR reaction was accompanied with a negative control replacing template DNA with 1µl milliQ water. Thermocycling proceeded with 94°C for 3min, followed by 30 cycles of 94°C for 20s, 52.7°C for 20s and 72°C for 45s, and final extension at 72°C for 3min. Size and purity of PCR product was evaluated on a diagnostic 1% (w/v) agarose gel. PCR fragment was purified using High Pure PCR Product Purification kit (Roche Applied Science, U.S.A.) according to manufacturer‟s instructions. 90 Figure 5.2: Illustration of the 16S rRNA gene. Shaded areas represent V1-V8 variable regions respectively. TX9 and 1391R primers flank V5-V8 variable regions and amplify a region of approximately 600bp. Image adapted from Patin et al. (2013). 5.2.1.2 DNA ligation, transformation and plasmid extraction All reagents necessary for performing the ligation reaction was purchased from Promega (pGEM® -T Easy Vector Systems, Promega, U.S.A.) PCR products were non-directionally ligated into the multiple cloning region of pGEM® -T Easy vectors by incubating the following ligation mix in a 16°C water bath overnight: 5µl 2X ligation buffer, T4 DNA ligase, 1µl pGEM® -T Easy (50ng/µl), 1µl T4 DNA ligase (3 Weiss units/µl) and 3µl template DNA. Ligation reaction was accompanied with a positive control replacing template DNA with 1µl Control Insert DNA and 2µl milliQ water. Preparation of reagents required for transformation, as well as the transformation procedure was carried out according to Appendix II. Ligation products were transformed into E. coli DH5α competent cells. Transformants were plated onto LB agar with ampicillin/IPTG/X-Gal (LBXI-A) plates for blue-white screening and incubated at 37°C overnight. For each sample (IMM, BJ, SC, VC), 100 white colonies were randomly selected, picked and grown in LB-ampicillin medium (Appendix II) at 37°C overnight using 10ml culture tubes (Greiner bio-one, Belgium). Cultures were harvested by decanting into 2ml microcentrifuge tubes (Axygen, Corning, U.S.A.), pelleted down and stored at -20°C. Plasmid extraction was then performed using the E.Z.N.A.® Plasmid Miniprep Kit (Omega Bio-Tek, U.S.A.) according to manufacturer‟s instructions. 91 5.2.1.3 Verification of insert Presence of insert within each plasmid extract was confirmed through restriction-enzyme (RE) digestion using EcoRI. All reagents required for RE digest were purchased from Promega (U.S.A.). The 5-µl reaction mix consisted of the following components: 0.5µl Buffer H 10X Buffer, 0.1µl EcoRI (12 units/µl), 0.1µl acetylated bovine serum albumin (BSA) (10µg/µl), 2.0µl plasmid DNA, 2.3µl milliQ water. RE digests were incubated at 37°C overnight. The presence of an approximately 600bp insert was verified by running the entire 10µl reaction mixture on a 1% agarose gel. 5.2.1.4 Restriction fragment length polymorphism (RFLP) screening Each plasmid extract was completely digested in double-digest reactions using TaqI and MspI restriction enzymes (Thermo Scientific, U.S.A.). Digestion using MspI (0.5µl 10X Buffer Tango, 0.1µl TaqI (10 units/µl), 2µl plasmid DNA, 2.4µl milliQ water) was carried out at 37°C for 1hr, followed by digestion using TaqI (additional 0.1µl TaqI into each tube) at 65°C for 1hr. DNA band patterns were resolved by electrophoresing the whole mixture on a 2% (w/v) agarose gel. Clones were grouped based on similarity of profile (Figure 5.3). From each cluster of plasmid extracts with similar profiles, a random sample was sent off to a commercial sequencing facility (AITbiotech, Singapore) for DNA sequencing of the 16S rRNA gene insert. 92 Figure 5.3: Example of different RFLP patterns obtained from double digest of 12 clones within the Science Centre (SC) clone library. Six unique banding patterns (A-F) were obtained for this particular sample. M: Marker lane. 5.2.1.5 Bioinformatics analysis The raw 16S rRNA gene sequence reads from the environmental water samples were visually checked, then trimmed to remove vector and primer sequences. They were then compared with sequences in the Ribosomal Database Program (RDP) database by using the Seqmatch online analysis tool (Cole et al., 2008) to determine the phylogenetics affiliation. The results provided a guide to determining which 16S rRNA gene sequences were chosen for phylogenetics analysis. Both the environmental and reference sequences were subjected to multiple sequence alignment using MUSCLE software (Edgar, 2004). The aligned sequences were used to generate a phylogenetic tree by employing the Maximum Likelihood (ML) approach using MEGA5 with the Kimura 2-parameter model and default settings for the other parameters (Tamura et al., 2011). Nodal support was estimated using the bootstrap method conducted with 1000 repetitions. Phylogenetic trees were visualised using TreeGraph2 software (Stöver and Müller, 2010). 5.2.2 454-sequencing The same samples analysed using the clone library approach were selected for high-throughput 454-sequencing. 93 5.2.2.1 PCR amplification and purification Sample preparation was according to instructions provided by the Waikato DNA Sequencing Facility (WDSF, http://bio.waikato.ac.nz/sequence/) at the University of Waikato, Hamilton, New Zealand. Universal primers TX9 (GGATTAGAWACCCBGGTAGTC) and 1391R (CCTATCCCCTGTGTGCCTTGGCAGTCTCAG), which amplified the V5-V8 variable region of the 16S rRNA gene were used for the PCR amplification step. Each 30-µl PCR reaction consisted of the following components: 0.3µl TaKaRa PrimeStar High-fidelity Enzyme (2.5U), 6.0µl 5X Ex Taq Buffer, 3.0µl dNTP Mixture (2.5mM each), 1.2µl forward primer TX9 (10µM), 1.2µl reverse primer 1391R (10µM), 10ng/µl template DNA and remaining volume made up with milliQ water. Thermocycling proceeded with 94°C for 3min, followed by 24 cycles of 94°C for 20s, 52°C for 20s and 72°C for 45s, and final extension at 72°C for 3min. The PCR product obtained (~605bp) was visualised using a 2% (w/v) TAE gel, and the band excised and purified using High Pure PCR Product Purification kit (Roche Applied Science, U.S.A.) according to manufacturer‟s instructions. A second round of PCR was performed under the same PCR conditions for 10 cycles using primers adapted with barcode sequences. 5.2.2.2 Sample clean-up, quantification and quality check Clean-up of the PCR product was performed using Agencourt AMPure XP beads (Beckman Coulter, U.S.A.) following manufacturer's instructions. DNA quantification was performed using Quant-iT™ PicoGreen (Invitrogen, U.S.A.) using the protocol described in Appendix IV. Quality verification was performed using a diagnostic 2% (w/v) TAE gel showing a single band (with no primer dimers) corresponding to the expected size. The final mixture was sent out to 1st BASE (Singapore) for pyrosequencing using the Roche 454 GS-FLX platform. 94 5.2.2.3 Sequence analysis and phylogenetics assignment The open-source program Mothur (Schloss et al., 2009) was used to perform all bioinformatics analysis. Sequences were processed by removing sequences containing more than 1 ambiguous base followed by trimming off barcodes and adaptor sequences. Sequences were aligned using the Silva reference alignment (http://www.mothur.org/wiki/Silva_reference_alignment) followed by screening and filtering of the alignment to remove gaps, missing data and produce a fixed alignment length. Sequencing error was further reduced by performing pre-clustering of sequence data to merge sequence counts within 2bp of a more abundant sequence. Chimera checking was performed using the chimera.uchime command, and sequences affiliated with chloroplasts, mitochondria, eukaryotes and unknowns (sequences that cannot be classified at the Kingdom level) were removed. A distance matrix based on the processed sequences was generated with a cutoff of 0.15, followed by clustering into operational taxonomical units (OTUs) and determination of the majority consensus taxonomy. Alpha-diversity of each sample was calculated by generating collector's curve of the Chao1 and Inverse Simpson richness estimator and rarefaction curves. 5.2.2.4 Sequence identity calculation To compare sequences with those of known pathogenic bacteria (Table 5.1), representative sequences were retrieved from Genbank and combined with sequences from this study into a fasta file. The web-based version of Clustal (hosted by The European Bioinformatics Institute, http://www.ebi.ac.uk/Tools/msa/clustalo/) was used to generate sequence alignments using the default settings, as well as to calculate the identities between the sequences in the alignment. The number of reads corresponding to each pathogenic bacteria sequence was represented by a heatmap generated using Microsoft Excel. 95 Table 5.1: Human pathogenic bacteria, disease and accession number of reference sequences. Genus Pathogenic species Disease Accession No. Aeromonas Aeromonas veronii Aeromonas hydrophilia Arcobacter butzleri Septicemia Gastroenteritis Diarrhea Abdominal pain Cutaneous anthrax Pulmonary anthrax Gastrointestinal anthrax Nausea, vomiting and diarrhea Whooping cough Lyme disease Brucellosis Acute enteritis Community-acquired respiratory infection Nongonococcal urethritis Lymphogranuloma venereum Trachoma Inclusion conjunctivitis of the newborn Psittacosis Botulism Pseudomembranous colitis Gas gangrene Acute food poisoning Anaerobic cellulitis Tetanus Diphtheria Opportunistic infections DQ029351.1 AB610604.1 FJ968634.1 Arcobacter Bacillus Bacillus anthracis Bacillus cereus Bordetella Borrelia Brucella Campylobacter Chlamydia Bordetella pertussis Borrelia burgdorferi Brucella abortus Campylobacter jejuni Chlamydia trachomatis Chlamydophilia Clostridium Chlamydophilia psittaci Clostridium botulinum Clostridium difficile Clostridium perfringens Clostridium tetani Corynebacterium Enterobacter Escherichia Corynebacterium diphtheriae Enterobacter aerogenes Enterobacter cloacae Enterococcus faecalis Enterococcus faecium Escherichia coli Francisella Haemophilus Francisella tularensis Haemophilus influenzae Helicobacter Helicobacter pylori Klebsiella Klebsiella pneumoniae Enterococcus Nosocomial infections Urinary tract infections Diarrhea Meningitis in infants Traveller's diarrhea Diarrhea in infants Hemorrhagic colitis Hemolytic-uremic syndrome Tularemia Bacterial meningitis Upper respiratory tract infections Pneumonia, bronchitis Peptic ulcer Risk factor for gastric carcinoma and gastric B-cell lymphoma Pneumonia Urinary tract infections Septicemia Ankylosing spondylytis Soft tissue infections AY643481.1 AJ629413.1 AF366576.1 AF477988.1 EF192471.1 AY628389.1 DQ019301.1 AB285329.1 X68315.1 X73450.1 DQ196135.1 X74770.1 X84248.1 AB004750.1 Y17665.1 AB154827.1 AY692451.1 EU118103.1 AY968238.1 AY613478.1 AY593986.1 U33121.1 96 Legionella Legionella pneumophila Leptospira Listeria Mycobacterium Leptospira interrogans Listeria monocytogenes Mycobacterium leprae Mycobacterium tuberculosis Mycobacterium ulcerans Mycoplasma pneumoniae Neisseria gonorrhoeae Neisseria meningtidis Mycoplasma Neisseria Pseudomonas Rickettsia Salmonella Pseudomonas aeruginosa Rickettsia rickettsii Salmonella enterica Serratia Shigella Staphylococcus Serratia marcescens Shigella sonnei Staphylococcus aureus Staphylococcus epidermidis Staphylococcus saprophyticus Streptococcus Streptococcus agalactiae Streptococcus pneumoniae Streptococcus pyogenes Treponema Treponema pallidum Vibrio Yersinia Vibrio cholerae Yersinia pestis Legionnaire's Disease Pontiac fever Leptospirosis Listeriosis Leprosy Tuberculosis Mycoplasma pneumonia Gonorrhea Ophthalmia neonatorum Septic arthritis Meningococcal disease Waterhouse-Friderichsen syndrome Pseudomonas infections Rocky mountain spotted fever Typhoid fever type salmonellosis (dysentery, colitis) gastroenteritis and enterocolitis Nosocosmial infections Bacillary dysentery Staphylococcus infections Infections of implanted prostheses Cystitis in women Meningitis and septicemia in infants Endometritis in postpartum women Opportunistic infections with septicemia and pneumonia Scarlet fever Rheumatic fever Necrotizing fasciitis Syphilis Congenital syphilis Cholera Bubonic plaque Pneumonia plaque AF129523.1 AY995728.1 AJ535697.1 X53999.1 X52917.1 X58954.1 AF132741.1 AM921674.1 AY735364.1 AY548952.1 DQ150682.1 EF579646.1 M59160.1 EF032687.1 EF463060.1 AY699287.1 L37596.1 AB112407.1 AJ617796.1 NR_028598.1 M88726.1 AY494843.1 AJ232223.1 Data obtained from Ye and Zhang (2011). 5.3 Results 5.3.1 Phylogenetic characterisation using clone libraries Clone libraries of DNA samples obtained from BJ, IMM, SC and VC were established to identify the diversity of bacterial communities found within, and visualised by phylogenetic trees corresponding to samples from each location. Sequences from the clone library revealed that 97 bacterial diversity were mostly homogeneous regardless of location, with most clones clustering near or within Alpha-proteobacteria, Beta-proteobacteria, Gamma-proteobacteria, Bacteroidetes or Actinobacteria (Figure 5.4). Proportions of unique sequences clustering within or near each taxon differed depending on location, but generally Proteobacteria were most well-represented, appearing in all locations sampled (Figure 5.5). There were several unclassified sequences within the IMM, SC and VC library, possibly because the assignment was based on the sequences in RDP curated database mostly consisting of cultured species, which could have omitted certain environmental isolates that were uncultured (Figure 5.4). Through a BLAST search, two unclassified clades corresponded to Domain Bacteria and were likely to be either Verrucomicrobia (clade A) or Proteobacteria (clade C). Clade B identified strongly with freshwater Eukaryotae while some species of clade D was suggested to be phototropic but the taxonomic status (prokaryotic or eukaryotic) could not be determined. In particular, there was a significant proportion of bacteria with sequences classifying into Clade B and D within the SC and VC libraries (clade B: 41.5% and 20.0% respectively; clade D: 21.5% and 16.7% respectively). 98 A Figure 5.4: Phylogenetic tree of sequences obtained from water samples collected from the following locations: (A) Bugis Junction, (B) IMM, (C), Science Centre and (D) VivoCity. Clones were created using amplification products from the 16S rRNA gene. Representatives were selected based on top hits from BLAST and clustered using a maximum likelihood approach. An archaea 16S rRNA gene from Haloarcula marismortui was used to root the trees. Bootstrap values are shown on each branch. 99 Figure 5.4 continued. B 100 Figure 5.4 continued C 101 Figure 5.4 continued D 102 100% 90% 80% Undetermined 70% Firmicutes 60% Cyanobacterium 50% Actinobacterium 40% Bacteroidetes 30% Gamma-proteobacterium Beta-proteobacterium 20% Alpha-proteobacterium 10% 0% BJ IMM SC VC Figure 5.5: Proportions of unique sequences found in each taxon. 5.3.2 Taxonomic assignment of sequences from 454-sequencing In total 87089 sequences were obtained from all four samples in a single 454-pyrosequencing run, where 15230 sequences were unique. After quality checking, alignment and taxonomy assignment, the number of sequences dropped to 57221 with 15127 from BJ, 22917 from IMM, 13853 from SC and 5324 from VC. Taxonomic assignment of sequences was performed at the phylum/class level by obtaining the top 50 OTUs and plotting the phylum/class against the number of reads. The bacterial community composition within most of the locations was characterised predominantly by Beta-proteobacteria, Alpha and Gammaproteobacteria as well as Bacteroidetes (Figure 5.6). The data also detected a number of rare phyla (Acidobacteria, Chlamydia, Clostridia, Delta-proteobacteria and Verrucomicrobiae) which was overlooked by the classic Sanger sequencing approach. On the other hand, the bacterial population within Science Centre water samples was markedly different from the other three locations with a large 103 proportion of unclassified sequences (80% versus 3-15% respectively). In addition, there was very few Alpha (0.7% versus 2-8%), Beta (0.3% versus 40-60%) and Gamma-proteobacteria (0.08% versus 1530%) in Science centre bacteria population as compared to that from other locations. 104 A B C D Figure 5.6: Taxonomic classification of bacteria at the phylum/class level within (A) Bugis Junction, (B) IMM, (C) Science Centre and (D) VivoCity. 105 5.3.3 Within-sample diversity analysis Collector's curves were generated to quantify the alpha-diversity (within-sample diversity). Chao1's estimate curves showed that indices of diversity and evenness differed significantly among locations. BJ and IMM samples had relatively even species richness with the presence of some rare species (as shown by the steep and continued increase in slope), while SC and VC samples was dominated by only a few species and repeated sampling only yielded few new species as shown by the leveling of the curves (Figure 5.7A). This could be attributed to the SC sample being dominated mostly by unclassified sequences (10995 out of 13853 reads) while VC sample was dominated by beta-proteobacteria classified up to the family level (3142 out of 5324 reads classified to Oxalobacteraceae). Plotting the Inverse Simpson's index against the number of sequences sampled showed that unequal sampling did not strongly influence the diversity estimate (Figure 5.7B). Rarefaction analysis revealed that BJ and IMM samples exceeded 1000 OTUs while SC and VC libraries barely approached 200 OTUs, indicative of the greater species diversity found in the former two samples (Figure 5.8). Individual libraries also appeared to vary in richness in concordance with collector's curve results. Rarefaction curves corresponding to BJ, IMM and VC samples suggested that these locations seemed not to have been sampled to completion, which is indicative of insufficient sequencing coverage. 106 A B Figure 5.7: Collector's curves obtained for samples from BJ, IMM, SC and VC using the (A) Chao1 and (B) Inverse Simpson diversity estimator. 107 Figure 5.8: Rarefaction analysis of observed richness in bacterial communities sampled from BJ, IMM, SC and VC. 108 5.3.4 Comparison against pathogen database To find out the diversity of pathogenic populations in the samples, sequences from this run were compared with representative 16S rRNA gene sequences from GenBank. In total 3772 sequences, accounting for 6.59% of total sequences, were found to correspond to known pathogenic bacteria with a large proportion of them (5.49% of total sequences) having identities of over 99% (Table 5.2). Out of the 34 genera that were targeted, 14 occurred in our water samples. Pseudomonas and Mycobacterium were present in all four locations in high abundance while Neisseria, Bacillus, Aeromonas, Chlamydia and Staphylococcus were found in BJ and IMM (Figure 5.9). Exposed recycled water systems (BJ and IMM) were associated with greater diversity as compared to sheltered location (SC) and open wading pool (VC). Among the four locations, BJ was observed to possess the greatest pathogenic diversity with 13 different pathogenic genera, where Pseudomonas was the dominant population accounting for 96.6% of pathogenic bacteria present in the sample. Pathogenic genera were more evenly distributed in IMM samples; in addition it was the only sample that harboured Rickettsia. Even though the nature of the water facility is different in SC and VC, the profile of harboured pathogenic genera were quite similar, suggesting the existence of factors that select for predominantly Pseudomonas and Mycobacteria in SC and VC. 109 Table 5.2: Identity and amount of sequences corresponding to potentially pathogenic bacteria. Identity to known pathogen sequences >99% 97-99% 80% unclassified sequences). In fact, most metagenomics analysis of environmental samples have uncovered significant proportions of unclassified microbial diversity, of which a fraction is reported to overlap with the rare biosphere where there is little understanding of their ecological and metabolic roles (Lynch et al., 2012). Phylogenetics analysis of the clone libraries recovered a number of novel clades of which comparison with top BLAST hit 111 results suggested among others, the existence of eukaryotic-like sequences (clade B). Although these results cannot be corroborated with those from the 454-sequencing libraries, it is reasonable to postulate the existence of freshwater eukaryotic microorganisms (e.g. unicellular algae and protozoa) in our samples as their presence in aquatic habitats have been reported in literature (Medinger et al., 2010; Kermarrec et al., 2013). However, the eukaryotic ribosome is not as well characterised as the prokaryotic ribosome, as a result sequence representation in public databases is not as extensive, making identification difficult. Despite the presence of similar bacteria phyla within all the sampled interactive water fountains, the relative abundances of dominant taxa were markedly different in the Science Center library as compared to the other target locations, with the SC library showing an obvious paucity of identified reads. Due to the small sample size of this study, it could not be determined whether this anomalous profile represents an outlier, or a true difference between water obtained from SC and from malls (BJ, IMM, VC). Nonetheless, it is postulated that this difference in community structure could be real as results from Chapter 3 (Figure 3.8A), which illustrated microbial counts obtained from repeated sampling, also showed an absence of most bacterial markers within the SC samples. Considering that the water supply to the Science Center water playground is the same as that supplied to the other recreational water facilities, differences in community structure could only be explained by environmental influences specific to this particular fountain. There are a myriad of putative explanations e.g. chlorine dosing, localised weather conditions, bather population; however all of these factors are shared by other measured recreational sites and any single factor is unlikely to result in the obvious disparity. In fact, the clone library approach suggested the majority of reads in the unclassified sector identified to eukaryotic-like and phototrophic (possibly algal) sequences, making it reasonable to wonder whether it was due to overgrowth of eukaryotic populations within this particular location. Eukaryotic/prokaryotic interactions in aquatic environments are substantially documented in literature and it is possible that competition or predator/prey relationships could have result in the observed community structure (Sherr and Sherr, 2002; Pernthaler, 2005). The potential presence of eukaryotic sequences should be considered when attempting to accurately characterise the 112 microbial community in samples obtained from similar environments. On the other hand, taking the study in this direction implies changes in methodology as the 16S rRNA marker has limited phylogenetic resolution for eukaryotes and alternative phylogenetic markers such as 18s or 28s rRNA should be explored (Pruesse et al., 2007). Furthermore, the DNA extraction method used in this study is optimised for bacterial cells and may not work as efficiently on eukaryotes which could possess cell walls. 5.4.2 Pathogen discovery and detection Using 454-sequencing and comparison against a database of known pathogens, almost half of the pathogenic genera targeted (14 out of 34) were found in our samples, where only three (Pseudomonas, Mycobacterium, Aeromonas) had been previously targeted using real-time PCR. In fact, several of the highlighted pathogenic groups are instead associated with human exposure (Neisseria, Streptococcus) or exist as intracellular parasites (Chlamydia, Rickettsia). These pathogens are not known to be waterborne, nor are they listed on the USEPA or WHO watchlists as microbial contaminants to be monitored within our drinking and recreational water supplies. As such, it is intriguing to discover these pathogenic genera in our water samples, and we believe that the risk of exposure to them have so far been understudied. Furthermore, the ubiquitous presence of Pseudomonas and Mycobacterium in our samples is testament of their abundance in the environment and resistance to a variety of environmental hazards including disinfection procedures. While it is not appropriate to suggest that all tested recreational water fountains have been compromised, their existence should raise some concerns as they could grow to present a health risk when provided with the correct set of conditions. It also suggests that current methods of disinfection and cleaning are insufficient in keeping our recreational water facilities free of these potentially pathogenic organisms. On the other hand, the limitations of this study should be considered in light of the results: 1) PMA treatment was not performed as the protocol has not been optimised for 454-sequencing, hence no inferences can be made about the viability of these pathogenic groups; 2) Not all species within a single pathogenic genera are pathogenic, hence presence of these pathogenic genera does not 113 necessarily translate to a higher risk of disease; 3) 454-sequencing is a semi-quantitative approach and the abundance of each pathogenic group can only be compared in relation to each other; no inferences can be made about the true number of pathogenic organisms present within the samples. This is especially in consideration of the fact that 454-sequencing has a lower sensitivity threshold in comparison to qPCR (possibly explaining for the absence of Aeromonas in the VC and SC libraries when 454-sequencing was performed). Accurate pathogen counts are important as the minimum infective dose should be accounted for when determining microbial risk; 4) the small sample size limits the generalisability of this study to only the few interactive fountains that were tested; different pathogen compositions and trends may emerge with an increased sample size. Despite being a preliminary method which requires further modification and optimisation, the approach we have used in this study is valuable in discovering and detecting pathogens directly from recreational water facilities, furthering our knowledge on the bacterial and pathogenic diversity present. However, we believe that there is still value in developing indicators to signal the presence of potential pathogens in our water supplies. This is because the high start-up cost and technical expertise required makes running NGS on a routine basis impractical for monitoring purposes, and it is not feasible to develop large numbers of pathogenic markers to target all identified pathogens, even by using multiplex procedures. On the other hand, we would like to suggest a departure from the traditional indicator-pathogen paradigm, as we believe that NGS technology could be used as screening tools to select for potential candidate indicator organisms based on a number of characteristics (e.g. resistance to disinfectants). These indicators could be normal bacteria, but could also be pathogens (e.g. Pseudomonas and Mycobacterium), as such removing the need to establish any indicator-pathogen relationships. In this case their presence will highlight problems in water quality management e.g. inadequate disinfection procedures, and not the prediction of any particular pathogen. Subsequently for routine monitoring purposes, real-time PCR should still be used to quantify these discovered indicator markers, and only in the presence of a hit, should direct pathogen detection be performed using a similar approach to our study in order to elucidate the causative pathogen identity. 114 5.5 Conclusion A preliminary overview of bacterial diversity using clone library construction and 454-sequencing uncovered the existence of a number of dominant phyla including Proteobacteria and Bacteroidetes within our samples, suggesting that the bacterial community composition of recreational water samples is similar to that found in most natural environments. Pathogenic diversity is also high in two of the four water samples tested, and several genera has been associated with human transmission and are of clinical relevance. The prevalence of Pseudomonas and Mycobacterium in recreational water samples could signal a potential health risk that has previously gone undetected; such data is useful from a regulatory standpoint when conducting risk assessments. The methods described in this study has been used to perform a preliminary characterisation of both the bacterial and pathogenic diversity present in recreational water samples, however more optimisation work is required to confirm the presence of viable, pathogenic organisms. 115 Chapter 6: Conclusions Microbial indicators have been widely used to indicate the occurrence of pathogens in contaminated water supplies. With technological advancements and increased understanding of microbiology at the molecular level, we are now able to measure existing indicators as well as develop novel indicators using faster and cheaper methods. However, there are still numerous issues that need to be investigated in order to rapidly and accurately identify recreational water that has been contaminated with pathogens; these issues include the detection and differentiation of viable bacteria from dead ones, problems in accurately enumerating cells using molecular techniques in the presence of inhibitory compounds, problems in establishing reliable indicator-pathogen relationships due to the need of both indicator and pathogen to possess similar persistence, fate and transport levels, etc. Despite knowing these limitations coupled with a lack of understanding of their suitability in recreational waters, regulatory bodies worldwide are still applying the indicator approach to monitor microbiological water quality. In Chapter 3, we have investigated and validated the method attributes of using molecular techniques in the development and use of indicator markers for recreational water monitoring. We have discovered that simple dilution is sufficient to reduce the effect of PCR inhibitors in our samples. We have determined an optimum PMA concentration as a pre-treatment to efficiently reduce the amplification of dead cells, however at the extent of suppressing some live cell counts in environmental matrices. It could pose problems in detection sensitivity when developing qPCR assays; however should be of less consequence in routine monitoring of indicators which should be present in large amounts within water samples. In Chapter 4, we have looked at the biological attributes of existing indicators and hence determined their applicability in the context of monitoring recreational waters from aquatic facilities. We have discovered that there seems to be an absence of most existing fecal indicator bacteria and associated fecal pathogens in recreational water systems. Instead, the pathogens which are detected are associated with the surrounding environment and could possibly be endemic within the water system. 116 These pathogens are highly resistant to environmental hazards including disinfection and cleaning procedures, and a number occur in all the locations that were analysed. They also exhibit prolonged persistence despite the lack of external nutrient input. The remaining fecal indicator bacteria were not able to predict these pathogens accurately even though they exhibit similar transport kinetics within the water system. This uncoupling of indicator-pathogen relationship not only likely reflects different source inputs, but could also be due to differential persistence levels. Considering the advances in molecular technology and the subsequent development of cost-effective solutions to identify bacterial markers directly, we have studied the feasibility of direct pathogen detection using sequencing techniques in Chapter 5. We have found that pathogen identification in recreational water samples using NGS is possible, and have in the process discovered several other potentially pathogenic genera that were previously untargeted in recreational water monitoring. We have also found that while the bacterial community composition is roughly consistent across most samples with a number of dominant phyla, the presence and distribution of pathogenic genera is uneven; however we are unsure of the factors that contribute to this effect. The results of this study is preliminary and should be considered together with other complex issues such as the need for investigations to confirm health risk, as well as source identification and mitigation issues within the entire monitoring framework. The current indicator system is indeed flawed; however there is still value in holding on to the indicator approach as direct pathogen detection using sequencing technology currently do not possess the speed, logistical feasibility and sensitivity required in routine screening and monitoring procedures. In fact, direct pathogen measurement could be undertaken as a confirmatory procedure only when screening indicators persist at high levels. NGS is a promising technique in the development of better indicators that correlates with pathogens in terms of contamination sources and survival or transport characteristics, however this field is still in its infancy and most research conducted thus far are of an exploratory nature. Furthermore, there is currently little research performed on strategic sampling and concentration techniques optimised for NGS protocols. Nonetheless, we are confident that the adaptation of NGS technology is a feasible approach both in terms of direct pathogen detection and 117 possibly also for indicator development. We foresee that it will change the current concept of the indicator-pathogen system, as well as revolutionise the entire field of water quality monitoring. 118 References Abdelzaher, A. M., Wright, M. E., Ortega, C., Solo-Gabriele, H. M., Miller, G., Elmir, S., Newman, X., Shih, P, Bonilla, J. A., Bonilla, T. D., Palmer, C. J., Scott, T., Lukasik, J., Harwood, V. J., McQuaig, S., Sinigalliano, C., Gidley, M., Plano, L. R. W., Zhu, X., Wang, J. D. and Fleming, L. E. (2013). Presence of Pathogens and Indicator Microbes at a Non-Point Source Subtropical Recreational Marine Beach. Applied and Environmental Microbiology, 79(24), 724-732. Ahmed, W., Huygens, F., Goonetilleke, A. and Gardner, T. (2008). Real-Time PCR Detection of Pathogenic Microorganisms in Roof-Harvested Rainwater in Southeast Queensland, Australia. Applied and Environmental Microbiology, 74(17), 5490-5496. Al-Soud, W. A., and Rådström, P. (1998). Capacity of nine thermostable DNA polymerases to mediate DNA amplification in the presence of PCR-inhibiting samples. Applied and Environmental Microbiology, 64, 3748-3753. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. and Lipman, D. J. (1990). Basic local alignment search tool. Journal of Molecular Biology, 215(3), 403-410. Anderson, B. B., and Sogaard, I. (1987). Legionnaires' disease and brain abscess. Neurology, 37(2), 333-334. APHA. (1998). Standard Methods for the Examination of Water and Wastewater (20th Edition ed.). Washington, D.C. Aravindraja, C., Viszwapriya, D. and Karutha Pandian, S. (2013). Ultradeep 16S rRNA Sequencing Analysis of Geographically Similar but Diverse Unexplored Marine Samples Reveal Varied Bacterial Community Composition. PLoS ONE, 8(10), e76724. doi: 10.1371/journal.pone.0076724 Arnal, C., Ferré-Aubineau, V., Besse, B., Mignotte, B., Schwartzbrod, L. and Billaudel, S. (1999). Comparison of seven RNA extraction methods on stool and shellfish samples prior to hepatitis A virus amplification. Journal of Virological Methods, 77(1), 17-26. Ashby, M. N., Rine, J., Mongodin, E. F., Nelson, K. E. and Dimster-Denk, D. (2007). Serial Analysis of rRNA Genes and the Unexpected Dominance of Rare Members of Microbial Communities. Applied and Environmental Microbiology, 73(14), 4532-4542. Aspinall, S. T., and Graham, R. (1989). Two sources of contamination of a hydrotherapy pool by environmental organisms. Journal of Hospital Infection, 14(4), 285-292. Aw, T. G., and Rose, J. B. (2012). Detection of pathogens in water: from phylochips to qPCR to pyrosequencing. Current Opinion in Biotechnology, 23, 422-430. Bae, S., and Wuertz, S. (2009). Discrimination of Viable and Dead FecalBacteroidales Bacteria by Quantitative PCR with Propidium Monoazide. Applied and Environmental Microbiology, 75(9), 2940-2944. Baggi, F., Demarta, A., and Peduzzi, R. (2001). Persistence of viral pathogens and bacteriophages during sewage treatment: lack of correlation with indicator bacteria. Research in Microbiology, 152(8), 743-751. Bartram, J., and Rees, G. (2000). Monitoring Bathing Waters - A Practical Guide to the Design and Implementation of Assessments and Monitoring Programmes. London and New York: E & FN Spon. Bartram, J., Chartier, Y., Lee, J. V., Pond, K. and Surman-Lee, S. (2007). Legionella and the prevention of legionellosis. Geneva, Switzerland: WHO Press. Benitez, A. J., and Winchell, J. M. (2013). Clinical application of a multiplex real-time PCR assay for simultaneous detection of Legionella species, Legionella pneumophila, and Legionella pneumophila serogroup 1. Journal of Clinical Microbiology, 51(1), 348-351. Bibby, K., Viau, E. and Peccia, J. (2010). Pyrosequencing of the 16S rRNA gene to reveal bacterial pathogen diversity in biosolids. Water Research, 44, 4252-4260. Bibby, K., Viau, E. and Peccia, J. (2011). Viral metagenome analysis to guide human pathogen monitoring in environmental samples. Letters in Applied Microbiology, 52, 386-392. 119 Blackstone, G. M., Nordstrom, J. L., Vickery, M. C. L., Bowen, M. D., Meyer, R. F. and DePaola, A. (2003). Detection of pathogenic Vibrio parahaemolyticus in oyster enrichments by real time PCR. Journal of Microbiological Methods, 53(2), 149-155. Boehm, A. B. (2007). Enterococci concentrations in diverse coastal environments exhibit extreme variability. Environmental Science and Technology, 41(24), 8227-8232. Boehm, A. B., Ashbolt, N. J., Colford, J. M., Dunbar, L. E., Fleming, L. E., Gold, M. A., Hansel, J., Hunter, P. R., Ichida, A. M., McGee, C. D., Soller, J. A. and Weisberg, S. B. (2009). A sea change ahead for recreational water quality criteria. Journal of Water and Health, 7(1), 9-20. Boehm, A. B., Fuhrman, J. A., Mrse, R. D. and Grant, S. B. (2003). A tiered approach for identification of a human fecal pollution source at a recreational beach: case study at Avalon Bay, Catalina Island, California, USA. Environmental Science and Technology, 37(4), 673-680. Bomo, A. M., Stevik, T. K., Hovi, I. and Hanssen, J. F. (2004). Bacterial removal and protozoan grazing in biological sand filters. Journal of Environmental Quality, 33(3), 1041-1047. Borella, P., Guerrieri, E., Marchesi, I., Bondi, M. and Messi, P. (2005). Water ecology of Legionella and protozoan: environmental and public health perspectives. Biotechnology Annual Review, 11, 355-380. Bornstein, N., Marmet, D., Surgot, M., Nowicki, M., Meugnier, H., Fleurette, J., Ageron, E., Grimont, F., Grimont, P. A. D., Thacker, W. L., Benson, R. F. and Brenner, D. J. (1989). Legionella gratina sp. nov. isolated from French spa water. Research in Microbiology, 140(8), 541-522. Bower, P. A., Scopel, C. O., Jensen, E. T., Depas, M. M. and McLellan, S. L. (2005). Detection of Genetic Markers of Fecal Indicator Bacteria in Lake Michigan and Determination of Their Relationship to Escherichia coli Densities Using Standard Microbiological Methods. Applied and Environmental Microbiology, 71(12), 8305-8313. Brescia, C. C., Griffin, S. M., Ware, M. W., Varughese, E. A., Egorov, A. I. and Villegas, E. N. (2009). Cryptosporidium propidium monoazide-PCR, a molecular biology-based technique for genotyping of viable Cryptosporidium oocysts. Applied and Environmental Microbiology, 75, 6856-6863. Brett, J., and du Vivier, A. (1985). Pseudomonas aeruginosa and whirlpools. British Medical Journal (Clinical Research ed.), 290(6474), 1024-1025. Brewster, D. H., Browne, M. I., Robertson, D., Houghton, G. L., Bimson, J. and Sharp, J. C. M. (1994). An outbreak of Escherichia coli O157 associated with a children's paddling pool. Epidemiological Infection, 112, 441-447. Brinkman, N. E., Haugland, R. A., Wymer, L. J., Byappanahalli, M., Whitman, R. L. and Vesper, S. J. (2003). Evaluation of a Rapid, Quantitative Real-Time PCR Method for Enumeration of Pathogenic Candida Cells in Water. Applied and Environmental Microbiology, 69(3), 17751782. Brookes, J. D., Antenucci, J., Hipsey, M., Burch, M. D., Ashbolt, N. J. and Ferguson, C. (2004). Fate and transport of pathogens in lakes and reservoirs. Environment International, 30, 741-759. Burke, V., Robinson, J., Atkinson, H. M., Dibley, M., Berry, R. J., and Gracey, M. (1981). Exotoxins of Aeromonas hydrophila. The Australian Journal of Experimental Biology and Medical Science, 59(Pt 6), 753-761. Byappanahalli M. N. and Fujioka, R. (2004). Indigeneous soil bacteria and low moisture may limit but allow faecal bacteria to multiply and become a minor population in tropical soils. Water Science and Technology, 50(1), 27-32. Byappanahalli, M. N., Whitman, R. L., Shively, D. A., Sadowsky, M. J., and Ishii, S. (2006). Population structure, persistence, and seasonality of autochthonous Escherichia coli in temperate, coastal forest soil from a Great Lakes watershed. Environmental Microbiology, 8(3), 504-513. Cappelluti, E., Fraire, A. E., and Schaefer, O. P. (2003). A case of "hot tub lung" due to Mycobacterium avium complex in an immunocompromised host. Archives of Internal Medicine, 163, 845-848. 120 Carter, A. M., Pacha, R. E., Clark, G. W. and Williams, E. A. (1987). Seasonal occurrence of Campylobacter spp. in surface waters and their correlation with standard indicator bacteria. Applied and Environmental Microbiology, 53(3), 523-526. Chang, B., Taguri, T., Sugiyama, K., Amemura-Maekawa, J., Kura, F. and Watanabe, H. (2010). Comparison of ethidium monoazide and propidium monoazide for the selective detection of viable Legionella cells. Japanese Journal of Infectious Diseases, 63, 119-123. Charoenca, N., and Fujioka, R. S. (1993). Assessment of staphylococcus bacteria in Hawaii marine recreational waters. Water Science and Technology, 27, 283-289. Charoenca, N., and Fujioka, R. S. (1995). Association of staphylococcal skin infections and swimming. Water Science and Technology, 31, 11-17. Chick, H. (1908). An Investigation of the Laws of Disinfection. Journal of Hygiene, 8, 92-158. Cole, J. R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R. J., Kulam-Syed-Mohideen, A. S., McGarrell, D. M., Marsh, T., Garrity, G. M. and Tiedje, J. M. (2008). The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acid Research, 37, D141-D145. Collins, C. H., Grange, J. M. and Yates, M. D. (1984). Mycobacteria in water. Journal of Applied Bacteriology, 57(2), 193-211. Conway de Macario, E., and Macario, A. J. L. (2009). Methanogenic archaea in health and disease: A novel paradigm of microbial pathogenesis. International Journal of Medical Microbiology, 299(2), 99-108. doi: http://dx.doi.org/10.1016/j.ijmm.2008.06.011 Covert, T. C., Rodgers, M. R., Reyes, A. L., and Stelma Jr., G. N. (1999). Occurrence of Nontuberculous Mycobacteria in Environmental Samples. Applied and Environmental Microbiology, 65(6), 2492-2496. Craun, G. F., Calderon, R. L., and Craun, M. F. (2005). Outbreaks associated with recreational water in the United States. International Journal of Environmental Health Research, 15(4), 243-262. Dahlenborg, M., Borch, E. and Rådström, P. (2001). Development of a combined selection and enrichment PCR procedure for Clostridium botulinum types B, E, and F and its use to determine prevalence in fecal samples from slaughtered pigs. Applied and Environmental Microbiology, 67, 4781-4788. Davies, C. M., Long, J. A., Donald, M., and Ashbolt, N. J. (1995). Survival of fecal microorganisms in marine and freshwater sediments. Applied Environmental Microbiology, 61(5), 1888-1896. de Vicente, A., Aviles, M., Borrego, J. J. and Romero, P. (1988). Die-off and survival of Pseudomonas aeruginosa in freshwater. Zentralblatt für Bakteriologie, Mikrobiologie und Hygiene, 185(6), 534-547. Dembry, L. M., Zervos, M. J., and Hierholzer, W. J., Jr. (1998). Nosocomial bacterial infections. In A. S. Evans, and Brachman, P. S. (Ed.), Bacterial infections of humans: epidemiology and control (3rd ed., pp. 501–509). New York: Plenum Medical Book Co. DeSantis, T. Z., Brodie, E. L., Moberg, J. P., Zubieta, I. X., Piceno, Y. M. and Andersen, G. L. (2007). High-density universal 16S rRNA microarray analysis reveals broader diversity than typical clone library when sampling the environment. Microbial Ecology, 53, 371-383. DeSantis, T. Z., Dubosarskiy, I., Murray, S. R. and Andersen, G. L. (2003). Comprehensive aligned sequence construction for automated design of effective probes (CASCADE-P) using 16S rDNA. Bioinformatics, 19, 1461-1468. Dorevitch, S., Ashbolt, N. J., Ferguson, C. M., Fujioka, R., McGee, C. D., Soller, J. A. and Whitman, R. L. (2010). Meeting Report: Knowledge and Gaps in Developing Microbial Criteria for Inland Recreational Waters. Environmental Health Perspectives, 118, 871-876. Edgar, R. C. (2004). MUSCLE: multiple sequence alignment with high accuracy and high thoroughput. Nucleic Acids Research, 32(5), 1792-1797. Efstratiou, M. F., Mavridou, A., Richardson, S. C. and Papadakis, J. A. (1998). Correlation of bacterial indicator organisms with Salmonella spp., Staphylococcus aureus and Candida albicans in sea water. Letters in Applied Microbiology, 26, 342-346. 121 Evenson, L. J. (1998). Legionnaires' disease. Primary Care Update Ob/Gyns, 5(6), 286-289. Fauville-Dufaux, M., Vanfleteren, B., De Wit, L., Vincke, J. P., Van Vooren, J. P., Yates, M. D., Serruys, E., Content, J. (1992). Rapid detection of tuberculous and non-tuberculous mycobacteria by polymerase chain reaction amplification of a 162 bp DNA fragment from antigen 85. European Journal of Clinical Microbiology and Infectious Diseases, 11(9), 797-803. Ferguson, C. M. (1994). Faecal indicators and pathogens in water and sediment of the Georges River. (MSc thesis), University of Technology, Sydney. Ferguson, C. M., Coote, B. G., Ashbolt, N. J. and Stevenson, I. M. (1996). Relationships between indicators, pathogens and water quality in an estuarine system. Water Research, 30(9), 2045-2054. Fick, R. B. J. (1992). Pseudomonas Aeruginosa the Opportunist : Pathogenesis and Disease. Florida: CRC Press, LLC Field, K. G., and Samadpour, M. (2007). Fecal source tracking, the indicator paradigm, and managing water quality. Water Research, 41, 3517-3538. Fiksdal, L., Vik, E. A., Mills, A. and Staley, T. (1982). Non-standard methods for enumerating bacteria in drinking water. Journal American Water Works Association, 74, 313-318. Fittipaldi, M., Pino Rodriguez, N. J., Codony, F., Adrados, B., Peñuela, G. A. and Morató, J. (2010). Discrimination of infectious bacteriophage T4 virus by propidium monoazide real-time PCR. Journal of Virological Methods, 168(1-2), 228-232. Frahm, E., and Obst, U. (2003). Application of the fluorogenic probe technique (TaqMan PCR) to the detection of Enterococcus spp. and Escherichia coli in water samples. Journal of Microbiological Methods, 52(1), 123-131. François, T., Hehemann, J-H., Etienne, R., Mirjam, C. and Gurvan, M. (2011). Environmental and Gut Bacteroidetes: The Food Connection. Frontiers in Microbiology, 2, 93. doi: 10.3389/fmicb.2011.00093 Friesa, J. S., Characklisb, G. W. and Noblea, R. T. (2008). Sediment–water exchange of Vibrio sp. and fecal indicator bacteria: Implications for persistence and transport in the Neuse River Estuary, North Carolina, USA. Water Research, 42(4-5), 941-950. Garber, W. F. (1956). Bacteriological standards for bathing waters. Sewage & Industrial Wastes, 28(6), 795-807. Gavini, F., Leclerc, H. and Mossel, D. A. A. (1985). Enterobacteriaceae of the “coliform group” in drinking water: identification and worldwide distribution. Systematic and Applied Microbiology, 6, 312-318. Gedalanga, P. B., and Olson, B. H. (2009). Development of a quantitative PCR method to differentiate between viable and nonviable bacteria in environmental water samples. Applied Microbiology and Biotechnology, 82(3), 587-596. Geldreich, E. E. (1996). Microbial Quality of Water Supply in Distribution Systems. Boca Raton, Florida: CRC Lewis. Ghanem, E. H., Mussa, M. E. and Eraki, H. M. (1993). Aeromonas-associated gastroenteritis in Egypt. Zentralblatt Fuer Mikrobiologie, 148(6), 441-447. Goh, S. G. (2012). Molecular detection and analysis of Enterococcus faecalis survival in tropical environmental waters (Ph.D thesis), Nanyang Technological University. Goldman, M., Rosenfeld-Yehoshua, N., Lerner-Geva, L., Lazarovitch, T., Schwartz, D. and GrisaruSoen, G. (2008). Clinical features of community-acquired Pseudomonas aeruginosa urinary tract infections in children. Pediatric Nephrology, 23(5), 765-768. Gomez-Alvarez, V., Revetta, R. P. and Santo Domingo, J. W. (2012). Metagenomic Analyses of Drinking Water Receiving Different Disinfection Treatments. Applied and Environmental Microbiology, 78(17), 6095-6102. Griffin, D. W., Gibson, C. J. III, Lipp, E. K., Riley, K., Paul, J. H. III and Rose, J. B. (1999). Detection of viral pathogens by reverse transcriptase PCR and of microbial indicators by standard 122 methods in the canals of the Florida Keys. Applied and Environmental Microbiology, 65, 4118-4125. Guy, R. A., Payment, P., Krull, U. J. and Horgen, P. A. (2003). Real-Time PCR for Quantification of Giardia and Cryptosporidium in Environmental Water Samples and Sewage. Applied and Environmental Microbiology, 69(9), 5178-5185. Haugland, R. A., Siefring, S. C., Wymer, L. J., Brenner, K. P. and Dufour, A. P. (2005). Comparison of Enterococcus measurements in freshwater at two recreational beaches by quantitative polymerase chain reaction and membrane filter culture analysis. Water Research, 39(4), 559568. Helms, C. M., Renner, E. D., Viner, J. P., Hierholzer, W. J. Jr., Winterneyer, L. A. and Johnson, W. (1980). Indirect immunofluorescence antibodies to Legionella pneumophila: frequency in a rural community. Journal of Clinical Microbiology, 12(3), 326-328. Hilborn, E. D., Covert, T. C., Yakrus, M. A., Harris, S. I., Donnelly, S. F., Rice, E. W., Toney, S., Bailey, S. A. and Stelma Jr., G. M. (2006). Persistence of Nontuberculous Mycobacteria in a Drinking Water System after Addition of Filtration Treatment. Applied and Environmental Microbiology, 72(9), 5864-5869. Hollyoak, V. A., and Freeman, R. (1995). Pseudomonas aeruginosa and whirlpool baths. The Lancet, 346(8975), 644-645. Holmes, P., Niccolls, L. M. and Sartory, D. P. (1996). The ecology of mesophilic Aeromonas in aquatic environment. In B. Austin, Altwegg, M., Gosling, P. and Joseph, S. W (Ed.), The Genus Aeromonas (pp. 39-76). New York: John Wiley & Sons. Hong, P. Y., Hwang, C., Ling, F., Andersen, G. L., LeChevallier, M. W. and Liu, W. (2010). Pyrosequencing Analysis of Bacterial Biofilm Communities in Water Meters of a Drinking Water Distribution System. Applied and Environmental Microbiology, 76(16), 5631-5635. Hörman, A., Rimhanen-Finne, R., Maunula, L., Bonsdorff, C. H., Torvela, N., Heikinheimo, A., and Hänninen, M. L. (2004). Campylobacter spp., Giardia spp., Cryptosporidium spp., Noroviruses, and Indicator Organisms in Surface Water in Southwestern Finland, 2000-2001. Applied Environmental Microbiology, 70(1), 87-95. Huber, J. A., Mark Welch, D. B., Morrison, H. G., Huse, S. M., Neal, P. R., Butterfield, D. A. and Sogin, M. L. (2007). Microbial population structures in the deep marine biosphere. Science, 318, 97100. Huber, T., Faulkner, G. and Hugenholtz, P. (2004). Bellerophon: a program to detect chimeric sequences in multiple sequence alignments. Bioinformatics, 20, 2317-2319. Hubner, I., Steinmetz, I., Obst, U., Giebel, D. and Bitter-Suermann, D. (1992). Rapid determination of members of the family Enterobacteriaceae in drinking water by an immunological assay using a monoclonal antibody against enterobacterial common antigen. Applied Environmental Microbiology, 58, 3187-3191. Hunter, P. R. (1998). Waterborne Disease Epidemiology and Ecology. Chichester, UK: Wiley & Sons. Hurst, C. J., Crawford, R. L., Knudsen, G. R., McInerney, M. J. and Stetzenbach, L. D. (2002). Manual of Environmental Microbiology, second ed. . Washington, DC: ASM Press. Huse, S. M., Huber, J. A., Morrison, H. G., Sogin, M. L. and Welch D. M. (2007). Accuracy and quality of massively parallel DNA pyrosequencing. Genome Biology, 8, R143. Huson, D. H., Mitra, S., Weber, N., Ruscheweyh, H. and Schuster, S. C. (2011). Integrative analysis of environmental sequences using MEGAN4. Genome Research, 21, 1552-1560. Illumina. (2013). Sequencing. Retrieved 22nd September, 2013, from http://www.illumina.com/applications/sequencing.ilmn Iqbal, S., Robinson, J., Deere, D., Saunders, J.R., Edwards, C. and Porter, J. (1997). Efficiency of the polymerase chain reaction amplification of the uid gene for detection of Escherichia coli in contaminated water. Letters in Applied Microbiology, 24, 498-502. Isobe, K. O., Tarao, M., Zakaria, M. P., Chiem, N. H., Minh, L. Y. and Takada, H. (2002). Quantitative application of fecal sterols using gas chromatography – mass spectrometry to investigate 123 fecal pollution in tropical waters: western Malaysia and Mekong Delta, Vietnam. Environmental Science and Technology, 36, 4497-4507. Janda, J. M., and Abbott, S. L. (1996). Human Pathogens. In B. Austin, Altwegg, M., Gosling, P. and Joseph, S. W. (Ed.), The Genus Aeromonas (pp. 39-76). New York: John Wiley & Sons. Janda, J. M., and Abbott, S. L. (2007). 16S rRNA Gene Sequencing for Bacterial Identification in the Diagnostic Laboratory: Pluses, Perils, and Pitfalls Journal of Clinical Microbiology, 45(9), 2761-2784. Jiang, S. C., and Chu, W. (2004). PCR detection of pathogenic viruses in Southern California urban rivers. Journal of Applied Microbiology, 97(1), 17-28. Johnston, C., Ufnar, J. A., Griffith, J. F., Gooch, J. A., Stewart, J. R. (2010). A real-time qPCR assay for the detection of the nifH gene of Methanobrevibacter smithii, a potential indicator of sewage pollution. Journal of Applied Microbiology, 109(6), 1946-1956. Joly, P., Falconnet, P-A., André, J., Weill, N., Reyrolle, M., Vandenesch, F., Maurin, M., Etienne, J. and Jarraud, S. (2006). Quantitative Real-Time Legionella PCR for Environmental Water Samples: Data Interpretation. Applied and Environmental Microbiology, 72(4), 2801-2808. Joseph, S. W. (1996). Aeromonas gastrointestinal disease: a case study in causation? In B. Austin, Altwegg, M., Gosling, P. and Joseph, S. W. (Ed.), The Genus Aeromonas (pp. 311-335). New York: John Wiley & Sons. Josephson, K. L., Gerba, C. P. and Pepper., I. L. (1993). Polymerase chain reaction detection of nonviable bacterial pathogens. Applied Environmental Microbiology, 59, 3513-3515. Katcher, H. L., and Schwartz, I. (1994). A distinctive property of Tth DNA polymerase: Enzymatic amplification in the presence of phenol. Biotechniques, 16, 84-92. Keene, W. E., McAnulty, J. M., Hoesly, F. C., Williams, L. P. Jr., Hedberg, K., Oxman, G. L., Barrett, T. J., Pfaller, M. A. and Fleming, D. W. (1994). A swimming associated outbreak of haemorrhagic colitis caused by Escherichia coli O157:H7 and Shigella sonnei. New England Journal of Medicine, 331, 579-584. Keijser, B. J., Zaura, E., Huse, S. M., van der Vossen, J. M., Schuren, F. H., Montijn, R. C., ten Cate, J. M. and Crielaard, W. (2008). Pyrosequencing analysis of the oral microflora of healthy adults. Journal of Dental Research, 87(11), 1016-1020. Kelly, A. J., Justice, C. A. and Nagy, L. A. (1983). Predominance of chlorine tolerant bacteria in drinking water systems. Abstracts of the Annual Meeting of the American Society for Microbiology 79th Meeting, Paper No. Q122. Kermarrec, L., Franc, A., Imet F. R., Chaumeil, P., Humbert, J. F. and Bouchez, A. (2013). Nextgeneration sequencing to inventory taxonomic diversity in eukaryotic communities: a test for freshwater diatoms. Molecular Ecology Resources, 13, 607-619. Kerr, R., and Butterfield, C. (1943). Notes on the relations between coliform organisms and pathogens. Public Health Reports, 58(15), 589-607. Kids&Parenting. (2013). Eco-learning and play at City Square Mall. Retrieved 6th January, 2014, from www.kidsandparenting.com/kidsandp/places-to-go/eco-learning-and-play-at-citysquare-mall Klein, M., Brown, L., Ashbolt, N. J., Stuetz, R. M., and Roser, D. J. (2011). Inactivation of indicators and pathogens in cattle feedlot manures and compost as determined by molecular and culture assays. . FEMS Microbiology Ecology, 77(1), 200-210. Kosek, M., Bern, C., and Guerrant, R.L. . (2003 ). The global burden of diarrhoeal disease, as estimated from studies published between 1992 and 2000. Bulletin of the World Health Organisation, 81, 197-204. Kreader, C. A. (1995). Design and evaluation of Bacteroides DNA probes for the specific detection of human fecal pollution. Applied and Environmental Microbiology, 61(4), 1171-1179. Kristiansson, E., Fick, J., Janzon, A., Grabic, R., Rutgersson, C., Weijdegard, B., Soderstrom, H. and Larsson, D. G. J. (2011). Pyrosequencing of antibiotic-contaminated river sediments reveals high levels of resistance and gene transfer elements. PLoS ONE, 6(2), e17038. 124 Kuntz, T. B., and Kuntz, S. T. (1999). Enterhaemorrhagic E. coli infection. Primary Care Update Ob/Gyns, 6(6), 192-195. Kuroki, T., Sata, S., Yamai, S., Yagita, K., Katsube, Y. and Endo, T. (1998). Occurrence of free-living amoebae and Legionella in whirlpool baths. The Journal of the Japanese Association for Infectious Diseases, 72(10), 1056-1063. Lam, M. C., Ang, L. W., Tan, A. L., James, L., and Goh, K. T. (2011). Epidemiology and control of Legionellosis, Singapore. Emerging Infectious Diseases, 17(7), 1209-1215. Lansbury, L. E., and Ludlam, H. (1997). O157:H7: Lessons from the past 15 years. Journal of Infection, 96(5), 391-397. Lantz, P. G., Stålhandske, F. I., Lundahl, K. and Rådström, P. (1999). Detection of yeast by PCR in sucrose solutions using a sample preparation method based on filtration. World Journal of Microbiology and Biotechnology, 15, 345-348. Lawrence, C., Reyrolle, M., Dubrou, S., Forey, F., Decludt, B., Goulvestre, C., Matisota-Bernard, P., Etienne, J. and Nauciel, C. (1999). Single clonal origin of a high proportion of Legionella pneumophila serotype 1 isolates from patients and the environment in the rea of Paris, France, over a 10-year-period. Journal of Clinical Microbiology, 37(8), 2652-2655. Le Gall, F., Le Berre, R., Rosec, S., Hardy, J., Gouriou, S., Boisramé-Gastrin, S., Vallet, S., Rault, G., Payan, C. and Héry-Arnaud, G. (2013). Proposal of a quantitative PCR-based protocol for an optimal Pseudomonas aeruginosa detection in patients with cystic fibrosis. BMC Microbiology, 21(13), 143. doi: 10.1186/1471-2180-13-143 Le Loir, Y., Baron, F. and Gautier, M. (2003). Staphylococcus aureus and food poisoning. Genetics and Molecular Research, 2, 63-76. LeChevallier, M. W., and Seidler, R. J. (1980). Staphylococcus aureus in rural drinking-water. Applied and Environmental Microbiology, 39, 739-742. Leecaster, M. K., and Weisberg, S. B. (2001). Effect of sampling frequency on shoreline microbiology assessments. Marine Pollution Bulletin, 42, 1150-1154. Lemarchand, K., and Lebaron, P. (2003). Occurrence of Salmonella spp. and Cryptosporidium spp. in a French coastal watershed: relationship with fecal indicators. FEMS Microbiology Letters, 218(1), 203-209. Lepp, P. W., Brinig, M. M., Ouverney, C. C., Palm, K., Armitage, G. C. and Relman, D. A. (2004). Methanogenic Archaea and human periodontal disease. Proceedings of the National Academy of Sciences of the United States of America, 101(16), 6176-6181. Levasseur, S., Husson, M.O., Leitz, R., Merlin, F., Laurent, F., Peladan, F., Drocourt, J.L., Leclerc, H. and Van Hoegaerden, M. (1992). Rapid detection of members of the family Enterobacteriaceae by a monoclonal antibody. Applied Environmental Microbiology, 58, 1524–1529. Levy, D., Bens, M. S., Craun, G. F., Calderon, R. L. and Herwaldt, B. L. (1998). Surveillance for waterborne disease outbreaks - United States, 1995-1996. Lewin, S., Norman, R., Nannan, N., Thomas, E., and Bradshaw, D. (2007 ). Estimating the burden of disease attributable to unsafe water and lack of sanitation and hygiene in South Africa in 2000. South African Medical Journal, 97, 755–762. Lindqvist, R., Norling, B. and Thisted Lambertz, S. (1997). A rapid sample preparation method for PCR detection of food pathogens based on buoyant density centrifugation. Letters in Applied Microbiology, 24, 306-310. Liu, C. W., Lin, K. H. and Kuo, Y. M. (2003). Application of factor analysis in the assessment of groundwater quality in a blackfoot disease area in Taiwan. Science of the Total Environment, 313, 77-89. Loveridge, P. (1981). Legionnaires' disease and arthritis. Canadian Medical Association Journal, 124(4), 366-367. Lowry, F. D. (1998). Staphylococcus aureus infections. New England Journal of Medicine, 339, 520532. 125 LPSN. (24 December 2013). List of Prokaryotic Names with Standing in Nomenclature. Retrieved 2nd January, 2014, from http://www.bacterio.net/-number.html#aboveclass Lynch, M. D. J., Bartram, A. K. and Neufeld, J. D. (2012). Targeted recovery of novel phylogenetic diversity from next-generation sequence data. ISME Journal, 6(11), 2067-2077. Mamane, M., Kasparian, P., Babo, P. and Accard, J. L. (1983). Legionnaires' disease with late pulmonary sequelae. Apropos of a case. Le Poumon et le Coeur, 39(6), 305-308. Martinez-Medina, M., Aldeguer, X., Lopez-Siles, M., González-Huix, F., López-Oliu, C., Dahbi, G., Blanco, J. E., Blanco, J., Garcia-Gil, L. J. and Darfeuille-Michaud, A. (2009). Molecular diversity of Escherichia coli in the human gut: New ecological evidence supporting the role of adherent-invasive E. coli (AIEC) in Crohn's disease. Inflammatory Bowel Diseases, 15, 872882. Matsuda, K., Tsuji, H., Asahara, T., Matsumoto, K., Takada, T. and Nomoto, K. (2009). Establishment of an Analytical System for the Human Fecal Microbiota, Based on Reverse TranscriptionQuantitative PCR Targeting of Multicopy rRNA Molecules. Applied and Environmental Microbiology, 75(7), 1961-1969. McFeters, G. A. (1990). Drinking Water Microbiology: Progress and Recent Developments. New York: Springer. Means, E. G., Hanami, L., Ridgway, H. F. and Olson, B. H. (1981). Evaluating mediums and plating techniques for enumerating bacteria in water distribution systems. Journal American Water Works Association, 53, 585-590. Medinger, R., Nolte, V., Pandey, R. V., Jost, S., Ottenwalder, B., Schlotterer, C. and Boenigk, J. (2010). Diversity in a hidden world: potential and limitation of next-generation sequencing for surveys of molecular diversity of eukaryotic microorganisms. Molecular Ecology Resources, 19(S1), 32-40. Metcalf, T. G., Melnick, J. L. and Estes, M. K. (1995). Environmental Virology: From Detection of Virus in Sewage and Water by Isolation to Identification by Molecular Biology-A Trip of Over 50 Years. Annual Review of Microbiology, 49, 461-487. Metge, D. (2013). Fate And Transport of Bacterial, Viral, and Protozoan Pathogens During ASR Operations- What Microorganisms Do We Need To Worry About And Why? U.S. Geological Survey. Retrieved 8th September, 2013, from http://water.usgs.gov/ogw/pubs/ofr0289/dm_fatetransport.htm Mittelman, M. W., Habash, M., Lacroix, J.-M., Khoury, A.E. and Krajden, M. (1997). Rapid detection of Enterobacteriaceae in urine by fluorescent 16S rRNA in situ hybridization on membrane filters. Journal of Microbiological Methods, 30, 153–160. Miyamoto, M., Yamaguchi, Y. and Sasatsu, M. (2000). Disinfectant effects of hot water, ultraviolet light, silver ions and chlorine on strains of Legionella and nontuberculous mycobacteria. Microbios, 101, 7-13. Moriñigo, M. A., Muñoz, M. A., Cornax, R., Martinez-Manzanares, E. and Borrego, J. J. (1992). Presence of indicators and Salmonella in natural waters affected by outfall wastewater discharges. Water Science and Technology, 25(9), 1-8. NEA. (2013). Recreational Water Quality. Retrieved 11th August, 2013, from http://app2.nea.gov.sg/anti-pollution-radiation-protection/water-pollution/recreationalwater-quality Noble, R. T., and Fuhrman, J. A. (2001). Enteroviruses detected by reverse transcriptase polymerase chain reaction from the coastal waters of Santa Monica Bay, California: low correlation to bacterial indicator levels. Hydrobiologia, 460(1-3), 175-184. Noble, R. T., and Weisberg, S. B. (2005). A review of technologies for rapid detection of bacteria in recreational waters. Journal of Water and Health, 03(4), 381-392. Nocker, A., and Camper, A. K. (2006). Selective Removal of DNA from Dead Cells of Mixed Bacterial Communities by Use of Ethidium Monoazide. Applied and Environmental Microbiology, 72(3), 1997-2004. 126 Nocker, A., Cheung, C. Y. and Camper, A. K. (2006). Comparison of propidium monoazide with ethidium monoazide for differentiation of live vs. dead bacteria by selective removal of DNA from dead cells. Journal of Microbiological Methods, 67(2), 310-320. Nocker, A., Sossa-Fernandez, P., Burr, M. D. and Camper, A. K. (2007). Use of Propidium Monoazide for Live/Dead Distinction in Microbial Ecology. Applied and Environmental Microbiology, 73(16), 5111-5117. Nogva, H. K., Dromtorp, S. M., Nissen, H. and Rudi, K. (2003). Ethidium monoazide for DNA-based differentiation of viable and dead bacteria by 5 '-nuclease PCR. Biotechniques, 34(4), 804813. NRC. (2004). Committee on indicators for waterborne pathogens. Washington: The National Academy Press. Obst, U., Hubner, I., Wecker, M. and Bitter-Suermann, D. (1989). Immunological method using monoclonal antibodies to detect Enterobacteriaceae in drinking water. Aqua, 38, 136-142. Pace, N., Stahl, D., Olsen, G. and Lane, D. (1985). Analyzing natural microbial populations by rRNA sequences. American Society for Microbiology News, 51, 4-12. Pan, Y., and Breidt, F. (2007). Enumeration of viable Listeria monocytogenes cells by real-time PCR with propidium monoazide and ethidium monoazide in the presence of dead cells. Applied and Environmental Microbiology, 73, 8028-8031. Paneth, N. (2004). Assessing the contributions of John Snow to Epidemiology - 150 years after removal of the broad street pump handle. Epidemiology, 15(5), 514-516. Parshionikar, S., Laseke, I. and Fout, G. S. (2010). Use of propidium monoazide in reverse transcriptase PCR to distinguish between infectious and noninfectious enteric viruses in water samples. Applied and Environmental Microbiology, 76, 4318-4326. Patin, N. V., Kunin, V., Lidstrom, U. and Ashby, M. N. . (2013). Effects of OTU Clustering and PCR Artefacts on Microbial Diversity Estimates. Microbial Ecology, 65, 709-719. Pavlov, D., de Wet, C. M., Grabow, W. O. and Ehlers, M. M. . (2004). Potentially pathogenic features of heterotrophic plate count bacteria isolated from treated and untreated drinking water. International Journal of Food Microbiology, 92(3), 275-287. Pedley, S., Bartram, J, Rees, G., Dufour, A and Contruvo, J. A. (Ed.). (2004). Pathogenic mycobacteria in water. A guide to public health consequences, monitoring and management. London, UK: IWA Publishing. Peliowski, A., and Finer, N. N. (1986). Intractable seizures in Legionnaires' disease. The Journal of Pediatrics, 109(4), 657-658. Pernthaler, J. (2005). Predation on prokaryotes in the water column and its ecological implications. Nature Reviews Microbiology, 3, 537-546. Perola, O., Kauppinen, J., Kusnetsov, J., Heikkinen, J., Jokinen, C. and Katila, M. L. (2003). Nosocomial Legionella pneumophila serogroup 5 outbreak associated with persistent colonization of a hospital water system. Acta Pathologica, Microbiologica et Immunologica Scandinavica, 110(12), 863-868. Pickles, J., Feakins, R. M., Hansen, J., Sheaff, M. and Barnes, N. (2001). A case of Mycobacterium avium-intracellulare pulmonary disease and Crohn's disease. Grand Rounds, 2, 24-28. Plano, L. R. W., Garza, A. C., Shibata, T., Elmir, S. M., Kish, J., Sinigalliano, C. D., Gidley, M. L., Miller, G., Withum, K., Fleming, L. E. and Solo-Gabriele. (2011). Shedding of Staphylococcus aureus and methicillin-resistant Staphylococcus aureus from adult and pediatric bathers in marine waters. BMC Microbiology, 11(5), 1-10. Pollack, M., Koles, N. L., Preston, M. J., Brown, B. J. and Pier, G. B. (1995). Functional properties of isotype-switched immunoglobulin M (IgM) and IgG monoclonal antibodies to Pseudomonas aeruginosa lipopolysaccharide. Infection and Immunity, 63(11), 4481-4488. Pond, K. (2005). Recreational Water and Disease. London, UK: IWA Publishing. 127 Pruesse, E., Quast, C., Knittel, K., Fuchs, B. M., Ludwig, W., Peplies, J. and Glöckner, F. O. (2007). SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Research, 35(21), 7188-7196. Pruss, A., Kay, D., Fewtrell, L., and Bartram J. . (2002). Estimating the burden of disease from water, sanitation, and hygiene at a global level. Environmental Health Perspectives, 110, 537–542. PUB. (23rd January 2013). Active, Beautiful, Clean. Retrieved 10th August, 2013, from http://www.pub.gov.sg/abcwaters/Pages/default.aspx Rådström, P., Löfström, C., Lövenklev, M., Knutsson, R. and Wolffs, P. (2003). Strategies for Overcoming PCR Inhibition. In D. a. Dveksler (Ed.), PCR Primer (2nd ed.). New York: Cold Spring Harbor Laboratory Press. Rastogi, N., Frehel, C., Ryter, A, Ohayon, H., Lesourd, M. and David, H. L. (1981). Multiple drug resistance in Mycobacterium avium: is the wall architecture responsible for the exclusion of antimicrobial agents? Antimicrobial Agents and Chemotherapy, 20, 666-667. Ratnam, S., Hogan, K., March, S. B. and Butler, R. W. (1986). Whirlpool-associated folliculitis caused by Pseudomonas aeruginosa: report of an outbreak and review. Journal of Clinical Microbiology, 23(3), 655-659. Reali, D., and Rosati, S. (1994). Antibiotic susceptibility and serotyping of Pseudomonas aeruginosa strains isolated from surface waters, thermomineral waters and clinical specimens. Zentralblatt für Hygiene und Umweltmedizin, 196(1), 75-80. Regnault, B., Martin-Delautre, S., Lejay-Collin, M., Lefevre, M. and Grimont, P. A. D. (2000). Oligonucleotide probe for the visualization of Escherichia coli/Escherichia fergusonii cells by in situ hybridization: specificity and potential application. Research in Microbiology, 151, 521-533. Roche. (2013). 454-sequencing. Retrieved 22nd September, 2013, from http://www.454.com/ Roig, J., Carrerres, A. and Domingo, C. (1993). Treatment of Legionnaires' disease. Current recommendations. Drugs, 46(1), 63-79. Rompre, A., Servais, P., Baudart, J., de-Roubin, M. R. and Laurent, P. (2002). Detection and enumeration of coliforms in drinking water: current methods and emerging approaches. Journal of Microbiological Methods, 49, 31-54. Rowe, P. C., Orbine, E., Wells, G. A. and McLaine, P. N. (1991). Epidemiology of haemolytic-uraemic syndrome in Canadian children from 1986 to 1988. The Journal of Pediatrics, 119(2), 218-224. Rudi, K., Moen, B., Dromtorp, S. M. and Holck, A. L. (2005). Use of Ethidium Monoazide and PCR in Combination for Quantification of Viable and Dead Cells in Complex Samples. Applied and Environmental Microbiology, 71(2), 1018-1024. Santo Domingo, J. W., Bambic, D. G., Edge, T. A., and Wuertz, S. (2007). Quo vadis source tracking? Towards a strategic framework for environmental monitoring of fecal pollution. Water Research, 41, 3539-3552. Savichtcheva, O., and Okabe, S. (2006). Alternative indicators of fecal pollution: Relations with pathogens and conventional indicators, current methodologies for direct pathogen monitoring and future application perspectives. Water Research, 40, 2463-2476. Schloss, P. D., Westcott, S. L., Ryabin, T., Hall, J. R., Hartmann, M., Hollister, E. B., Lesniewski, R. A., Oakley, B. B., Parks, D. H., Robinson, C. J., Sahl, J. W., Stres, B., Thallinger, G. G., Van Horn, D. J., Weber, C. F. (2009). Introducing mothur: Open-source, platform-independent, community-supported software for describing and comparing microbial communities. Applied and Environmental Microbiology, 75(23), 7537-7541. Schulz, C. J., and Childers, G. W. (2011). Fecal Bacteroidales Diversity and Decay in Response to Variations in Temperature and Salinity. Applied and Environmental Microbiology, 77, 25632572. Shah, A. H., Abdelzaher, A. M., Phillips, M., Hernandez, R., Solo-Gabriele, H. M., Kish, J., Scorzetti, G., Fell, J. W., Diaz, M. R., Scott, T. M., Lukasik, J., Harwood, V. J., McQuaig, S., Sinigalliano, C. D., Gidley, M. L., Wanless, D., Ager, A., Lui, J., Stewart, J. R., Plano, L. R. W., and Fleming, L. E. 128 (2011). Indicator microbes correlate with pathogenic bacteria, yeasts, and helminthes in sand at a subtropic recreational beach site. Journal of Applied Microbiology, 110, 1571-1583. Shelton, B. J., Flanders, W. D. and Morris, G. K. (1999). Mycobacterium sp. as a possible cause of hypersensitive pneumonitis in machine workers. Emerging Infectious Diseases, 5(2), 270-273. Sherr, E. B., and Sherr, B. F. (2002). Significance of predation by protists in aquatic microbial food webs. Antonie van Leeuwenhoek, 81, 293-301. Shigemura, K., Arakawa, S., Sakai, Y., Kinoshita, S., Tanaka, K., & Fujisawa, M. (2006). Complicated urinary tract infection caused by Pseudomonas aeruginosa in a single institution (1999-2003). International Journal of Urology, 13(5), 538-542. Sieglar, R., Pavia, A., Christoferrerson, R. and Milligan, M. (1994). A 20-year population-based study of postdiarrhoeal haemolytic uraemic syndrome in Utah. Pediatrics, 96, 35-40. Simpson, J. M., Santo Domingo, J. W., & Reasoner, D. J. (2004). Assessment of equine fecal contamination: the search for alternative bacterial source-tracking targets. FEMS Microbiology Ecology, 47(1), 65-75. Smeal, W. E., Scehfeld, L. A. and Hauger, W. (1985). Legionella causing rhabdomyolysis and renal failure. Postgraduate Medicine, 78(2), 42-44. Smith, T. (1891). A New Method for Determining Quantitatively the Pollution of Water by Fecal Bacteria. 13th Annual Report of the New York State Board of Health. Albany, New York. Sogin, M. L., Morrison, H. G., Huber, J. A., Welch, D. M., Huse, S. M., Neal, P. R., Arrieta, J. M. and Herndl, G. J. (2006). Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proceedings of the National Academy of Sciences of the United States of America, 103(32), 12115-12120. SPSS. (2011). IBM SPSS Statistics Base 20. SPSS Inc., Chicago, IL. Staroscik, A. (2004). Calculator for determining the number of copies of a template. Retrieved 9th August, 2013, from http://cels.uri.edu/gsc/cndna.html Stokes, W. R. (1927a). Report of the committee on bathing places. American Journal of Public Health, 17, 121. Stokes, W. R. (1927b). A search for pathogenic bacteria in swimming pools. American Journal of Public Health, 17, 334. Stöver, B. C., and Müller, K. F. (2010). TreeGraph 2: Combining and visualizing evidence from different phylogenetic analyses. BMC Bioinformatics, 11(7). doi: 10.1186/1471-2105-11-7 Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., and Kumar, S. (2011). MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution, 28(10), 2731-2739. Taylor, L. H., Falkinham, J. O., III, Norton, C. D. and LeChavallier, M. W. (2000). Chlorine, chloramine, chlorine dioxide, and ozone susceptibility of Mycobacterium avium. Applied Environmental Microbiology, 66, 1702-1705. Taylor, L. H., Latham, S.M., and Woolhouse, M.E. (2001). Risk factors for human disease emergence. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 356, 983– 989. Thi Minh Chau, L. Y., and Muller, H. E. (1983). Incidence of Legionnaires' disease. Deutsche Medizinishche Wochenschrift, 108(40), 1508-1511. Thomas, F., Hehemann, J. H., Rebuffet, E., Czjzek, M. and Michel,G. (2011). Environmental and gut bacteroidetes: the food connection. Frontiers in Microbiology, 2, 93. doi: 10.3389/fmicb.2011.00093 Thompson, J. D., Gibson,T. .J., Plewniak, F., Jeanmougin, F., and Higgins, D. G. . (1997). The CLUSTAL X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acid Research, 25(24), 4876-4882. Trupiano, J. K., and Prayson, R. A. (2001). Mycobacterium avium intracellulare otitis media. Annals of Diagnostic Pathology, 5(6), 350-353. 129 Tsen, H. Y., Lin, C. K. and Chi,W. R. (1998). Development and use of 16S rRNA gene targeted PCR primers for the identification of Escherichia coli cells in water. Journal of Applied Microbiology, 85, 554-560. Turnbaugh, P. J., Hamady, M., Yatsunenko, T., Cantarel, B. L., Duncan, A., Ley, R. E., Sogin, M. L., Jones, W. J., Roe, B. A., Affourtit, J. P., Egholm, M., Henrissat, B., Heath, A. C., Knight, R. and Gordon, J. I. (2009). A core gut microbiome in obese and lean twins. Nature, 457(7228), 480484. Ufnar, J. A., Wang, S. Y, Christiansen, J. M., Yampara-Iquise, H., Carson, C. A. and Ellender, R. D. (2006). Detection of the nifH gene of Methanobrevibacter smithii: a potential tool to identify sewage pollution in recreational waters. Journal of Applied Microbiology, 101(1), 44-52. Urich, T., Lanzen, A., Qi, J., Huson, D. H., Schleper, C. and Schuster, S. C. . (2008). Simultaneous assessment of soil microbial community structure and function through analysis of the metatranscriptome. PLoS ONE, 3, e2527. USCDC. (1998). Surveillance for waterborne disease outbreaks - United States, 1995-1996 Surveillance Summaries (Vol. MMWR 47, pp. 1-34). USCDC. (2008). Surveillance for waterborne disease and outbreaks associated with recreational water use and other aquatic facility-associated health events - United States, 2005-2006 Surveillance Summaries (Vol. MMWR 57, pp. 1-38). USCDC. (2011). Surveillance for waterborne disease outbreaks and other health events associated with recreational water - United States, 2007-2008 Surveillance Summaries (Vol. MMWR, pp. 1-37). USEPA. (1998). Announcement of the Drinking Water Contaminant Candidate List: Notice. Federal Register, 63(40), 10274-10287. USEPA. (2012). Recreational water quality criteria. Retrieved 11th August, 2013, from http://water.epa.gov/scitech/swguidance/standards/criteria/health/recreation/index.cfm van der Kooij, D., Oranje, J. P. and Hijnen, W. A. (1982). Growth of Pseudomonas aeruginosa in tap water in relation to utilization of substrates at concentrations of a few micrograms per liter. Applied and Environmental Microbiology, 44(5), 1086-1995. Varma, M., Field, R., Stinson, M., Rukovets, B., Wymer, L. and Haugland, R. (2009). Quantitative realtime PCR analysis of total and propidium monoazide-resistant fecal indicator bacteria in wastewater. Water Research, 43(19), 4790-4801. Vesper, S., McKinstry, C., Hartmann, C., Neace, M., Yoder, S. and Vesper, A. (2008). Quantifying fungal viability in air and water samples using quantitative PCR after treatment with propidium monoazide (PMA). Journal of Microbiological Methods, 72, 180-184. Waage, A. S., Vardund, T., Lund, V. and Kapperud, G. (1999). Detection of low numbers of Salmonella in environmental water, sewage and food samples by a nested polymerase chain reaction assay. Journal of Applied Microbiology, 87(3), 418–428. Wade, T. J., Calderon, R. L., Sams, E., Beach, M., Brenner, K. P., Williams, A. H., and Dufour, A. P. (2006). Rapidly Measured Indicators of Recreational Water Quality Are Predictive of Swimming-Associated Gastrointestinal Illness. Environmental Health Perspectives, 114, 24-28. Wade, T. J., Pai, N., Eisenberg, J. N. S., and Colford, J. M., Jr. (2003). Recreational water quality and GI illness. Environment Health Perspectives, 111(8), 1102-1109. Wagner, A. O., Malin, C., Knapp, B. A. and Illmer, P. (2008). Removal of free extracellular DNA from environmental samples by ethidium monoazide and propidium monoazide. Applied and Environmental Microbiology, 74(8), 2537-2539. Wang, G., and Doyle, M. P. (1998). Survival of enterohaemmorhagic Escherichia coli O157:H7 in water. Journal of Food Protection, 61(6), 662-667. Wang, L., Li, Y. and Mustapha, A. (2009). Detection of viable Escherichia coli O157:H7 by ethidium monoazide real-time PCR. Journal of Applied Microbiology, 107, 1719-1928. 130 Wang, M. Y., Olson, B. and Chang, J. S. (2007). Improving PCR and qPCR detection of hydrogenase A associated with Clostridia in pure cultures and environmental sludges using bovine serum albumin. Applied Microbiology and Biotechnology, 77(3), 645-656. Warburton, D. W. (2000). Methodology for screening bottled water for the presence of indicator and pathogenic bacteria. Food Microbiology, 17(1), 3-12. doi: http://dx.doi.org/10.1006/fmic.1999.0283 Wattie, E., and Butterfield, C. (1944). Relative resistance of Escherichia coli and Eberthella typhosa to chlorine and chloramines. Public Health Reports, 59(52), 1661-1671. WHO. (2003). Guidelines for safe recreational water environments: Coastal and fresh waters. Geneva, Switzerland: WHO Press. WHO. (2006). Guidelines for safe recreational water environments: Swimming pools and similar environments. Geneva, Switzerland: WHO Press. WHO. (2013). Microbial fact sheets. Retrieved 26th August, 2013, from http://www.who.int/water_sanitation_health/dwq/gdwq3_11.pdf Wilson, I. G. (1997). Inhibition and facilitation of nucleic acid amplification. Applied and Environmental Microbiology, 63, 3741-3751. Wolfe, H. (1972). The coliform count as a measure of water quality. In R. Mitchell (Ed.), Water Pollution Microbiology (pp. 333-346). New York: Wiley-Interscience. Yampara-Iquise, H., Zheng, G., Jones, J. E. and Carson, C. A. (2008). Use of a Bacteroides thetaiotaomicron-specific a-1-6, mannanase quantitative PCR to detect human faecal pollution in water. Journal of Applied Microbiology, 105, 1686-1693. Yampara-Iquise, H., Zheng, G., Jones, J. E., Carson, C. A. (2008). Use of a Bacteroides thetaiotaomicron-specific alpha-1-6, mannanase quantitative PCR to detect human faecal pollution in water. Journal of Applied Microbiology, 105(5), 1686-1693. Ye, L., and Zhang, T. (2011). Pathogenic bacteria in sewage treatment plants as revealed by 454 pyrosequencing. Environmental Science and Technology, 45, 7173-7179. Yu, H. B., Zhang, Y. L., Lau, Y. L., Yao, F., Vilches, S., Merino, S., Tomas, J. M., Howard, S. P. and Leung, K. Y. (2005). Identification and Characterization of Putative Virulence Genes and Gene Clusters in Aeromonas hydrophila PPD134/91. Applied and Environmental Microbiology, 71(8), 4469-4477. 131 Appendix I KCl-HCl acid buffer (10X): A. 2.0M KCl (14.9g in 100ml distilled water) B. 2.0M HCl (16.7ml conc. HCl in 100ml distilled water) Mix 18 parts of (A) with 1 part of (B) to prepare 10X acid buffer. Store at room temperature. Acid treatment procedure: 1) In a 2.0ml micro-centrifuge tube, add 10X acid buffer to water sample in a 1:10 ratio (25µl buffer to 250µl sample) and mix by vortexing. 2) Incubate the acidified suspension at room temperature for 15min, before plating onto BCYE and BCYE-GVPC agar. 132 Appendix II Transformation Luria-Bertani (LB) broth: 5g Yeast extract 10g Tryptone 10g NaCl 1L Deionized water Autoclaved at 121°C for 15mins. LB agar: Additional inclusion of 15g granulated agar in 1L LB broth and autoclaved at 121°C for 15mins. LB-A agar: A. 100mg/ml Ampicillin (100mg in 1ml distilled water, filter-sterilised) To every 200ml LB agar, autoclaved and cooled to 55°C, add 200µl (A) into the agar. Swirl to mix and pour into petri dishes. Plates can be stored up to 1 week. LB agar with ampicillin/IPTG/X-Gal (LBXI-A): A. 100mg/ml Ampicillin (100mg in 1ml distilled water, filter-sterilised) B. 100mg/ml isopropyl-1-thio-β-D-galactopyranoside (IPTG) (100mg in 1ml distilled water, filter-sterilised) C. 50mg/ml 5-bromo-4-chloro-3-inodlyl-β-D-galactopyranoside (X-gal) (50mg in 1ml dimethyl sulfoxide [DMSO]) 133 To every 200ml LB agar, autoclaved and cooled to 55°C, add 200µl (A), 200µl (B) and 100µl (C) into the agar. Swirl to mix and pour into petri dishes. Plates can be stored up to 1 week. Procedure: 1) To 50µl E. coli DH5α competent cells in a 15ml Falcon tube, add 1µl of ligation reaction. Incubate on ice for 20min. 2) Heat shock the tubes for 45s using a water bath set to 42°C. Return the tubes to ice for 2min. 3) Add 900µl LB broth and incubate at 37°C on a shaking incubator for 1hr. 4) Centrifuge tube and remove 800µl supernatant. Resuspend transformants in remaining supernatant. Plate 100µl of transformants on LBXI-A plates and incubate plates overnight at 37°C. Colony growing: LB broth with ampicillin: To each culture tube containing 3ml LB broth, autoclaved and cooled to 4°C, add 6µl of 100mg/ml filter-sterilised Ampicillin. Procedure: 1) Using a sterile autoclaved wooden toothpick (available at departmental stores), touch the tip of the toothpick to a white colony. 2) Drop the toothpick into culture tube containing 3ml LB-ampicillin broth. The tip of the toothpick in contact with the white colony should be fully immersed in the broth. 3) Incubate the culture tube at 37°C on a shaking incubator overnight. 134 Appendix III Genomic DNA extraction Reagents: - 1.5ml screw-capped tubes containing 0.3g autoclaved glass beads (Ø 0.1mm) - Tris-SDS (prepared by mixing 250ml of [200mM Tris-HCl, 80mM EDTA adjusted to pH 9.0] and 50ml of 10% SDS) - Chlorofoam/isoamyl alcohol/ethanol (prepared in the ratio of 80:4:16) - 3M Sodium acetate adjusted to pH 5.2 - Isopropanol (100%) - Ethanol (70%) - Tris-EDTA (TE) buffer adjusted to pH 8.0 - 1ml of culture (containing 108 CFU/ml) to be extracted Procedure: 1) Pellet the cells by centrifugation and decant the supernatant. 2) Add 600µl Tris-SDS and resuspend the cells. Transfer the suspension into a 1.5ml screwcapped tube with pre-added glass beads. 3) Disrupt the bacteria cells by shaking the tube vigorously using a Mini Beadbeater (BioSpec Products, U.S.A.) for 30s. Centrifuge at 4°C for 5min. 4) Transfer the supernatant into a new 1.5ml screw-capped tube. Add an equal volume of chlorofoam/isoamyl alcohol/ethanol. Vortex at high speed for 45s. Centrifuge at 4°C for 5min. 135 5) Transfer 500µl of the top aqueous portion into a new 2.0 microcentrifuge tube. Add 50µl of sodium acetate, followed by 600µl of isopropanol. Mix by inversion. Centrifuge at 4°C for 5min. 6) Decant the supernatant. Add 1ml of ethanol (70%). Centrifuge at 4°C for 5min. 7) Decant the supernatant. Air-dry the DNA pellet at room temperature overnight. 8) Resuspend the dried pellet with 200µl TE buffer. 9) Aliquot the DNA extract into 10µl volumes and store at -20°C until use. 136 Appendix IV Standards preparation The standard curve consisted of 100ng/ml, 50ng/ml, 25ng/ml, 12.5ng/ml, 6.25ng/ml, 3.13ng/ml, 1.56ng/ml, 0625pg/ml, 312.5pg/ml and 0ng/ml standards. DNA stock (100µg/ml) was first diluted to working concentration (2µg/ml), which was in turn diluted down to form the first standard (100ng/ml) using 1X TE buffer as diluent. Subsequent standards were obtained by 2-fold serial dilutions using 1X TE buffer as diluent. Sample analysis 1X Picogreen was prepared from the stock Picogreen reagent through a 200X dilution using 1X TE buffer. Samples were also diluted 100X using 1X TE buffer. 100µl of each DNA template (standard or sample) was loaded into each well of a 96-well flat bottom transparent UV microplate (Corning, U.S.A.) pre-loaded with 100µl 1X Picogreen solution. Each standard/sample was analyzed in duplicates. Measurement DNA fluorescence was measured using a fluorescent microplate reader at 480nm excitation and 520nm emission wavelengths. Fluorescence value of the reagent blank was subtracted from each of the samples. The standard curve was generated by plotting fluorescence against standard concentration, and sample concentration interpolated from the best-fit line. The assay was repeated if the R2 value of the standard curve was lower than 0.98. 137 [...]... correlation and predictive powers of FIB against pathogens in environmental waters in spatial and temporal studies, and hence determine the applicability of current indicators in Singapore recreational waters, and 3) To develop alternative methods of assessing recreational water quality, should the current indicators be found to be unsuitable 1.5 Scope of study Because the field of recreational water microbiology... increase in the popularity of recreational activities that involve contact with water In Singapore, government initiatives have resulted in a redesign of our waterways to incorporate various lifestyle attractions including kayaking and windsurfing activities (PUB, 2013) There is also a trend in the increasing use of indoor water recreation such as commercial spa pools, Jacuzzi and interactive water playgrounds... pneumonia in hospital settings (Dembry et al., 1998), and has been documented to cause cerebrospinal infections and bacterimia (Fick, 1992) In healthy individuals, P aeruginosa frequently causes ear and skin infections, and has also been linked to infections of the urinary-tract and gastro-intestinal tract P aeruginosa-induced infections are notoriously difficult to treat, associated with peculiar clinical... pools, interactive fountains) and untreated (reservoir) waters In 2007 and 2008, a total of 134 outbreaks associated with recreational water was reported, resulting in at least 13,699 cases of illnesses, in which 40% of outbreaks comprised of dermatological illnesses, acute respiratory conditions as well as combined illnesses The main etiological agents in these cases were of bacterial origin, which included... marine waters (Charoenca and Fujioka, 1993, 1995) 2.2 Indicator approach As awareness increased about disease transmission by water, there was a growing realisation of the importance of monitoring water supplies in order to protect public health This section describes the history of recreational water monitoring practices including an explanation of the indicator approach and the guidelines that Singapore... Alternative indicators The vast majority of alternative indicators proposed have focused on markers for fecal contamination, and they include fecal anaerobes, chemical compounds and viral indicators Of the three categories, only the first two pertain to the detection of bacterial pathogens which is of interest in this study, while viral indicators are mainly used as surrogate indicators for enteric viruses As... transmitted by use of recreational water and this association has not been investigated There is much that is not understood about the microbial risks present in recreational water settings Therefore, from a research standpoint, it is important to investigate what is the nature of 1 the microbial community, including pathogens, that is present in our recreational waters It is also integral to develop... Worldwide, recreational water quality criteria are based on the measurement of fecal indicator bacteria using culture techniques However various limitations of using traditional fecal indicators in monitoring bathing water quality and predicting disease outbreak exist; these include low levels of correlation with pathogens, low detection sensitivity, ability to multiply outside the water column, as well as inability... evidence linking exposure to environmental MAC to development of skin and soft tissue infections, and even pneumonia (Collins et al., 1984; Shelton et al., 1999) In most cases, disease outbreaks can be attributed to a high bather load, lack of disinfection, or inadequate maintenance of filters and pipes 2.1.2.5 Pseudomonas aeruginosa The pseudomonads are a group of Gram-negative, curved rods classified into... point-sampling approach A more intensive sampling for spatial and temporal variability of bacterial markers were also conducted within a single recreational water facility Using correlation analysis, linear regression and FA/PCA, it was discovered that Enterobacteriaceae and M smithii were good predictors of pathogen presence in study sites subjected to point-sampling, but none of the measured indicators