Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Comparing the distribution of pathogenic bacteria and common indicator microorganisms in biofilms on.. Maal-Bared, Rasha 2008

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
24-ubc_2008_fall_maalbared_rasha.pdf [ 2.87MB ]
Metadata
JSON: 24-1.0066404.json
JSON-LD: 24-1.0066404-ld.json
RDF/XML (Pretty): 24-1.0066404-rdf.xml
RDF/JSON: 24-1.0066404-rdf.json
Turtle: 24-1.0066404-turtle.txt
N-Triples: 24-1.0066404-rdf-ntriples.txt
Original Record: 24-1.0066404-source.json
Full Text
24-1.0066404-fulltext.txt
Citation
24-1.0066404.ris

Full Text

Comparing the Distribution of Pathogenic Bacteria and Common Indicator Microorganisms in Biofilms on Different Surface Types in an Agricultural Watershed in British Columbia (Canada) by Rasha Maal-Bared  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Resource Management and Environmental Studies)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) May 2008 © Rasha Maal-Bared, 2008  Abstract Little is known about the distribution of bacterial indicators and pathogens in biofilms on different surface types in natural aquatic environments. This study was conducted to examine the distribution of pathogens and indicator bacteria in biofilms in an agricultural watershed. The study particularly focused on whether biofilms can act as sinks for pathogenic organisms and could be monitored to protect public health. To do so, we monitored the presence of faecal contamination indicators (heterotrophic plate counts, faecal coliforms, enterococci, and E. coli) and particular pathogens (E. coli 0157, Campylobacter sp. and Salmonella sp.) in water, sediment, and in biofilms on river and slate rock, wood, sandpaper, and LexanTM in Elk Creek (British Columbia, Canada) from December 2005 to April 2007. Faecal indicator concentrations and pathogen presence were evaluated using standard culturing and isolation methods. The results showed that both faecal indicators and pathogens were present at the headwaters and that the use of water column grab samples underestimated faecal indicator numbers. Also, water column grab samples during the dry season were not representative of pathogens present in the creek. This indicates that biofilms might be the main reservoir of Salmonella sp. and pathogenic E. coli O157 in the summer when rainfall (which results in flow changes and sloughing) is limited. Campylobacter sp. was not retrieved in the dry season. Campylobacter on sediment, slate rock and wood showed high correlations with nitrates and enterococci, which could be used as faecal contamination surrogates. Numbers of indicator organisms and pathogens in one-month biofilms were compared to those in long-term biofilms (colonized 12 and 24 weeks) and short-term biofilms (colonized one to three weeks). The comparison showed that surface type, colonization period and water quality all affected the concentration of indicator organisms and pathogens present in biofilms. Finally, results showed high levels of phenotypic antibiotic resistance of E. coli and pathogenic E. coli O157 isolated from the watershed (even at the headwaters), particularly to tetracycline, ampicillin and streptomycin. This study highlights the potential biofilms could play in prediction of water quality changes, the improvement of sampling methods, and the study of aquatic environments.  ii  Table of Contents Abstract ............................................................................................................................................... ii Table of Contents................................................................................................................................ iii List of Tables.......................................................................................................................................vi List of Figures................................................................................................................................... viii Acknowledgements...............................................................................................................................x Co-Authorship statement......................................................................................................................xi 1 Introduction..................................................................................................................................1 1.1 Monitoring microbial water quality: Are we doing the best job we can?.............................. 1 1.2 Effects of agriculture on microbial water quality................................................................. 1 1.2.1 Sources of pathogens in agricultural watersheds ...........................................................1 1.2.2 Pathogen transport and storm events ............................................................................2 1.2.3 Monitoring microbial water quality ..............................................................................3 1.3 Biofilms .............................................................................................................................. 4 1.3.1 What are Biofilms? ......................................................................................................4 1.3.2 Biofilm Characteristics.................................................................................................7 1.3.3 Maturation and succession in biofilms..........................................................................8 1.3.3.1 Formation ............................................................................................................... 8 1.3.3.2 Detachment .......................................................................................................... 10 1.3.4 Can biofilms be used as monitoring tools ...................................................................11 1.3.4.1 Presence of pathogens in biofilms ......................................................................... 11 1.3.4.2 Could substratum choice affect the climax community in a biofilm?...................... 12 1.4 Research objectives and hypotheses................................................................................... 13 1.4.1 Main research objective .............................................................................................13 1.4.2 Research questions and hypotheses ............................................................................13 1.5 References ........................................................................................................................ 20 2 Seasonal and spatial distribution of indicators and pathogens in biofilms in agricultural watersheds: is monitoring water column grab samples sufficient?........................................................30 2.1 Introduction ...................................................................................................................... 30 2.2 Methods ............................................................................................................................ 32 2.2.1 Sampling location- Elk Creek.....................................................................................32 2.2.2 Biofilm samplers........................................................................................................34 2.2.3 Sample collection.......................................................................................................35 2.2.4 Microbial analyses .....................................................................................................36 2.2.4.1 Bacterial indicator organisms and E. coli O157 analysis........................................ 36 2.2.4.2 Salmonella spp. analysis ....................................................................................... 36 2.2.4.3 Campylobacter spp. analysis ................................................................................. 37 2.2.4.4 Controls................................................................................................................ 37 2.2.5 Standardizing across different substrata......................................................................37 2.2.6 Statistical Analyses ....................................................................................................38 2.3 Results .............................................................................................................................. 39 2.3.1 Comparing indicator organism numbers and pathogen frequencies recovered from water column, sediment and biofilms samples at each site...........................................................39 2.3.2 Indicator concentrations .............................................................................................41 2.3.3 Pathogen presence......................................................................................................42 2.4 Discussion......................................................................................................................... 46 2.5 Conclusion ........................................................................................................................ 52 2.6 Acknowledgements ........................................................................................................... 52 2.7 References ........................................................................................................................ 54  iii  3 Effects of surface types, water quality indicators and colonization period on the distribution of indicator bacteria and pathogenic organisms in biofilms in an agricultural watershed ..........................59 3.1 Introduction ...................................................................................................................... 59 3.2 Methods ............................................................................................................................ 61 3.2.1 Sampling location- Elk Creek.....................................................................................61 3.2.2 Biofilm samplers........................................................................................................62 3.2.3 Sample collection.......................................................................................................62 3.2.4 Microbial analysis......................................................................................................63 3.2.5 Standardizing across different substrata......................................................................63 3.2.6 Other water quality parameters...................................................................................63 3.2.7 Statistical analyses .....................................................................................................65 3.3 Results .............................................................................................................................. 65 3.3.1 Water quality in Elk Creek.........................................................................................65 3.3.2 Variations in one-month biofilm samplers analyzed by substratum type and site.........69 3.3.3 Long- and short-term biofilms....................................................................................73 3.3.4 Relationships between water quality variables, indicator organisms and pathogens in biofilms in short and long-term biofilms .....................................................................................76 3.4 Discussion......................................................................................................................... 78 3.5 Conclusions ...................................................................................................................... 84 3.6 Acknowledgements ........................................................................................................... 85 3.7 References ........................................................................................................................ 86 4 Campylobacter spp. distribution in Elk Creek British Columbia: improving sampling techniques 90 4.1 Introduction ...................................................................................................................... 90 4.2 Methods ............................................................................................................................ 92 4.2.1 Evaluating substrata for Campylobacter spp. monitoring purposes..............................92 4.2.2 Sampling location- Elk Creek.....................................................................................93 4.2.3 Biofilm samplers........................................................................................................93 4.2.4 Sample collection.......................................................................................................94 4.2.5 Campylobacter spp. analysis ......................................................................................94 4.2.6 Sample analysis .........................................................................................................95 4.2.7 Standardizing across different substrata......................................................................95 4.2.8 Water quality parameters ...........................................................................................95 4.2.9 Statistical Analyses ....................................................................................................95 4.3 Results and Discussion ...................................................................................................... 95 4.3.1 Campylobacter prevalence on different substrata........................................................95 4.3.2 Campylobacter correlation with other microbial water quality indicator bacteria.........98 4.3.3 Correlating Campylobacter prevalence in different biofilms with physical and chemical water quality characteristics ......................................................................................................100 4.3.4 Substrata evaluation for Campylobacter spp. monitoring ..........................................104 4.4 Conclusion ...................................................................................................................... 104 4.5 Acknowledgements ......................................................................................................... 106 4.6 References ...................................................................................................................... 107 5 Distribution and patterns of phenotypic antibiotic resistant Escherichia coli isolates from an agricultural watershed (Elk Creek, British Columbia)........................................................................111 5.1 Introduction .................................................................................................................... 111 5.2 Methods................................................................................................................................. 112 5.2.1 Sample collection.............................................................................................................112 5.2.2 Sample preparation ..........................................................................................................113 5.2.3 Broth microdilution MIC..................................................................................................113 5.2.4 Water quality data............................................................................................................114 5.2.5 Statistical analyses ...........................................................................................................114 5.3 Results ................................................................................................................................... 114 5.3.1 General phenotypic antibiotic resistance patterns in the watershed based on broth microdilution MIC results.........................................................................................................114  iv  5.3.2 Differences in phenotypic antibiotic resistance by site ......................................................117 5.3.3 Differences in phenotypic antibiotic resistance by substratum...........................................121 5.3.4 Pathogenic E. coli O157 antibiotic resistance ...................................................................121 5.3.5 Relationships between antibiotic resistance and other water quality parameters in the watershed 122 5.4 Discussion ............................................................................................................................. 126 5.5 Acknowledgements................................................................................................................ 132 5.6 References ............................................................................................................................. 133 6 Discussion ................................................................................................................................139 6.1 Summary of results and status of working hypotheses ............................................................ 139 6.2 Significance of the research.................................................................................................... 144 6.3 Strengths and limitations of the research................................................................................. 145 6.4 Future research....................................................................................................................... 147 6.5 References ............................................................................................................................. 149 Appendix A Elk Creek Pictures .................................................................................................. 153 Appendix B Biofilm Sampler Pictures ........................................................................................ 157 Appendix C Protozoa numbers within biofilms colonizing different substrata in Elk Creek ......... 159  v  List of Tables Table 2.1 Descriptive statistics and Kruskal-Wallis test results for indicator organism numbers (heterotrophic plate counts (HPC CFU/mg), faecal coliforms (FC MPN/mg), E. coli (EC MPN/mg), and enterococci (ENT MPN/mg)) in biofilms on different substrata in the wet season (n=11, total number of analyzed samples for all substrata=132) Descriptive statistics are shown as means, standard error in means (SE), medians, and variances (σ). Kruskal-Wallis analysis of variance test results are shown in terms of χ2 statistics, and p-values, and statistically significant results are bolded....................................................................................................................... 43 Table 2.2 Descriptive statistics and Kruskal-Wallis test results for indicator organism numbers (heterotrophic plate counts (HPC CFU/mg), faecal coliforms (FC MPN/mg), E. coli (EC MPN/mg), and enterococci (ENT MPN/mg)) in biofilms on different substrata in the dry season (n=6, total number of analyzed samples=72). Descriptive statistics are shown as means, standard error in means (SE), medians, and variances (σ). Kruskal-Wallis analysis of variance test results are shown in terms of χ2 statistics and p-values, and statistically significant results are bolded.... 44 Table 2.3 Percentage of samples positive for Campylobacter sp., Salmonella sp. and pathogenic E. coli O157 in biofilms, water and sediment on different surface types in the wet and the dry seasons. Table also presents number of samples tested in both seasons (nwet and nDry) and results of χ2 test. .................................................................................................................................................. 45 Table 3.1 Physical water quality variables measured at four sites in Elk Creek (BC) between December 2005 and April 2007 presented as means, medians, standard deviations, and sample sizes (n)..... 67 Table 3.2 Chemical water quality variables measured at four sites in Elk Creek (BC) between December 2005 and April 2007 presented as means, medians, standard deviations, and sample sizes (n). .................................................................................................................................... 68 Table 3.3 Biological water quality variables measured at four sites in Elk Creek (BC) between December 2005 and April 2007 presented as means, medians, standard deviations, and sample sizes (n). .................................................................................................................................... 69 Table 3.4 Results of Spearman correlations between indicators organisms numbers (Heterotrophic plate counts (HPC), faecal coliforms (FC), E. coli (EC) and enterococci (ENT)), pathogen frequencies (Campylobacter sp. (Campy), Salmonella sp. (Sal), and pathogenic E. coli O157 (EC O157)), ashfree dry weights (AFDW) in water, sediment and biofilms colonized for one month and water quality variables at different sites in Elk Creek (BC). A positive sign (+) indicates a positive relationship, while a negative sign (-) indicates a negative relationship....................................... 71 Table 3.5 Trends and variations in the concentrations of indicator organisms and presence of pathogens in biofilms collected from short- and long-term samplers across different sites........................... 73  vi  Table 3.6 Comparing water quality variables relationships with indicator bacteria concentrations and pathogen presence in biofilms on different surfaces, which were colonized for short- and longterm periods............................................................................................................................... 77 Table 4.1 Prevalence of Campylobacter sp. during the wet and the dry season in water and in biofilms on different surface types in Elk Creek, British Columbia. ......................................................... 97 Table 4.2 Spearman coefficients and p-values for associations between Campylobacter sp. (Camp) presence in water, slate rock, wood and sediment and the average numbers of heterotrophic plate counts (HPC), faecal coliforms (FC), E. coli (EC) and enterococci (ENT) found in the same media between December 2005 and December 2006. Statistically significant relationships (p<0.05) are bolded. ...................................................................................................................................... 98 Table 4.3 Spearman coefficients and p-values for associations between Campylobacter sp. presence in water, slate rock, wood and sediment and water quality characteristics (velocity, temperature (Temp), dissolved oxygen (DO), nitrate + nitrite (NOx), ortho-phosphate (PO4), ammonia (NH3), dissolved organic carbon (DOC), rainfall, and protozoa counts (Prtz counts)). Statistically significant relationships (p<0.05) are bolded. ........................................................................... 101 Table 4.4 Evaluating substrata for use as Campylobacter sp. monitoring tools in natural aquatic systems and agricultural watersheds...................................................................................................... 105 Table 5.1 Total number of E. coli isolates from Elk Creek (British Columbia), which were susceptible, moderately susceptible or resistant to different antibiotics tested based on broth microdilution MICs (n=214).......................................................................................................................... 115 Table 5.2 Distribution of susceptible, moderately susceptible, and resistant E. coli isolated from four different sites in Elk Creek (British Columbia) for all tested antibiotics. Site 1 is the control site and is located in the headwaters, and sites 2, 3, and 4 are situated in the agricultural reach. ...... 117 Table 5.3 Distribution of susceptible, moderately susceptible, and resistant E. coli in water, sediment and in biofilms on river rock, slate rock, wood, Lexan, and sandpaper isolated from all sites at Elk Creek British Columbia ........................................................................................................... 121 Table 5.4 Total number of pathogenic E. coli O157 isolates from Elk Creek (British Columbia), which were susceptible, moderately susceptible or resistant to different antibiotics tested based on broth microdilution MICs (n=27) ...................................................................................................... 122 Table 5.5 Results of logistic regression models highlighting relationships between different antibiotic resistant E. coli isolates (measured by MICs) and water quality parameters in Elk Creek. The positive and negative signs next to the antibiotic name indicate whether relationships were positive or negative. Statistically significant relationships are bolded, while strong trends that are not statistically significant are not bolded................................................................................. 124  vii  List of Figures Figure 1.1 Effects of agricultural activities on microbial water quality in different aquatic compartments of interest to this study............................................................................................................... 19 Figure 2.1 Map of Elk Creek in Chilliwack, British Columbia, and location of sites 1 to 4 (marked by the triangles) (Scale about 1: 90,000, 16.7 x 11.7 km). Arrow indicates north............................. 33 Figure 2.2 Biofilm samplers used to accumulate biofilms on a variety of substratum surface types. .... 35 Figure 2.3 Mean number of indicator organism per 100 mL of water and total number of samples positive for pathogens recovered from through water column grab samples at sites 1 to 4 (n=5, total number of processed samples for all sites= 40). The order of the legend from top to bottom corresponds with the bars from left to right. ............................................................................... 40 Figure 2.4 Mean number of indicator organism per 100 mL of sediment-PBS suspension and total number of samples positive for pathogens recovered from sediment samples at sites 1 to 4 (n=5, total number of processed samples for all sites= 40). The order of the legend from top to bottom corresponds with the bars from left to right. ............................................................................... 40 Figure 2.5 Mean number of indicator organism per 100 mL of biofilm-PBS suspension and total number of samples positive for pathogens recovered from all tested artificial substrata (river rock, slate rock, wood, Lexan and sandpaper) at sites 1 to 4 (n=25, total number of processed samples for all sites= 200). The order of the legend from top to bottom corresponds with the bars from left to right....................................................................................................................................... 41 Figure 2.6 Total number of samples that tested positive for Campylobacter (Campylobacter +), Salmonella (Salmonella +) and pathogenic E. coli O157 (E. coli O157 +) in biofilms on different substrata surface types, water and sediment in the wet season and the dry season. The order of the legend from top to bottom corresponds with the bars from left to right. ...................................... 46 Figure 3.1 Map of Elk Creek in Chilliwack, British Columbia, and location of sites 1 to 4 (marked by the triangles) (Scale about 1: 90,000, 16.7 x 11.7 km). Arrow indicates north............................. 62 Figure 3.2 Effect of different substratum colonization periods (1 week (n=2), 2 weeks (n=2), 3 weeks (n=2), 4 weeks (n=2), 12 weeks (n=6) and 24 weeks (n=4)) on indicator bacteria concentration, pathogen presence and ash-free dry weight (OM Weight) in water and river and slate rock. ....... 74 Figure 4.1 Map of Elk Creek in Chilliwack, British Columbia, and location of sites 1 to 4 (marked by the triangles) (Scale about 1: 90,000, 16.7 x 11.7 km). Arrow indicates north............................. 93 Figure 4.2 Comparing mean indicator numbers per mg of ash-free dry weight extracted from biofilm (n=11) and percentage of substrata positive for Campylobacter (n=22) across different surface types in the wet season............................................................................................................... 99  viii  Figure 4.3 Cumulative frequencies of positive Campylobacter sp. samples (Campy +) and water quality variables (rainfall (mm), temperature in oC, dissolved oxygen (DO), nitrate + nitrite (NOX), orthophosphate (PO4), ammonia (NH3), and dissolved organic carbon (DOC)) in Elk Creek (BC) between December 2005 and December 2006. Frequencies include Campylobacters recovered from water, biofilms and sediment. Dotted lines distinguish the dry from the wet season.......... 103 Figure 5.1 Minimum inhibitory concentration distribution of E. coli isolates from Elk Creek (British Columbia) to Ampicillin, Cefotaxime, Ciprofloxacin, Nalidixic acid, Streptomycin, and Tetracycline............................................................................................................................. 116 Figure 5.2 Total number of E. coli isolates resistant to Ampicillin, Cefotaxime, and Ciprofloxacin at the stated minimum inhibitory concentration (MIC 50%) at sites 1 to 4 in Elk Creek (British Columbia). .............................................................................................................................. 119 Figure 5.3 Minimum Inhibitory Concentrations of E. coli isolates (in µg/mL) and rainfall data (in mm) from Elk Creek (British Columbia) from December 2005 to April 2007................................... 125 Figure 6.1 Summary of research questions and results found in this study. Double lined boxes indicate major research questions asked during the study. ..................................................................... 140  ix  Acknowledgements First and foremost, I would like to acknowledge my fantastic supervisor Prof. Karen Bartlett for her help and support throughout this degree. Karen was always available for advice, ideas, and feedback. She was also a true mentor by all means of the word. If it were not for her, this work would not have been completed. I would also like to thank my committee members Prof. William Bowie, Prof. Eric Hall and Prof. Ken Hall for their contributions to the research design and the editing of my thesis. Their feedback has not only improved the final quality of my thesis, but has also taught me a lot. I would like to thank the CIHR Strategic Training Program in Public Health and the Agricultural Rural Ecosystem (PHARE) and Partner Institutes including the Institute of Cancer Research, Institute of Circulatory and Respiratory Health, Institute of Infection and Immunology, and the Institute of Population and Public Health for funding this study. I would like to acknowledge the Environmental Engineering laboratories at the University of British Columbia, particularly Ms. Paula Parkinson, for running the nutrient analyses. I would also like to thank the School of Environmental Health (SOEH), where all my laboratory work was completed. The wonderful people at SOEH never made me feel out of place and provided me with much needed intellectual and social enrichment when lab work was not going the way I had planned it. That includes Tim Ma, Tom Barnjak, Winnie Chu, Andrea Lam, and anyone else who has heard me complain about writing my thesis on a weekly basis. Many thanks to my amazing and lovely parents for their patience with my never ending academic pursuits, and to my siblings, Haya, Laith and Gheith, for letting me talk about biofilms on random occasions (more than any person ever ought to), and pretending they loved it. Also many thanks to my lovely parents-in-law, Rita and Keith, for their generosity, support, and the occasional drive out to Chilliwack or Vancouver to help a scientist in distress. Last, but not least, I would like to thank my wonderful, brilliant and astonishing husband Michael Zelmer. Michael, you know you are my rock! You came out to Chilliwack with me many a time and survived the occasional dangerous encounter with farm dogs, ditch water, liquid manure spreaders, and the pot-on-a-stick. You were there for me every second of the way, whether I needed support, feedback, edits or just someone to whine to. I could not have done it without you!  x  Co-Authorship statement For all of the submitted papers within this thesis, I was responsible for the preliminary research and sampling design, performing of field and laboratory work, data analysis and manuscript preparation. My supervisor, Dr. Karen Bartlett, was instrumental in designing the sampling method, the appropriate laboratory techniques and aiding with any sampling or analysis problems that arose. Dr. Bartlett also was vital to manuscript preparation and editing. The remainder of my co-authors, Dr. William Bowie, Dr. Eric Hall, and Dr. Ken Hall, took part in research design, troubleshooting and manuscript editing as well. The only component of the laboratory work not completed by myself was the analyses of nitrate, phosphate, ammonia and dissolved organic carbon, which was performed at the Environmental Engineering laboratories at the University of British Columbia by Paula Parkinson.  xi  1 1.1  Introduction Monitoring microbial water quality: Are we doing the best job we can?  Monitoring water quality is an essential component of the delivery of safe drinking water. However, current microbial water quality testing techniques are neither time-efficient nor costeffective. Standard techniques monitor for innocuous indicator bacteria, which do not always correlate with pathogen concentrations. Also, bacterial attachment and survival in biofilms are rarely examined. This problem is compounded in agricultural watersheds where faecal contamination of water sources is common and thus potential for pathogen survival in protected biofilm communities in aquatic environments is high. The contamination of aquatic ecosystems by agriculture and the role of biofilms in the retention of pathogens will be discussed in detail in the following sections of the introduction.  1.2 Effects of agriculture on microbial water quality 1.2.1 Sources of pathogens in agricultural watersheds The degradation of freshwater ecosystems through non-point pollution sources is a global concern (Sabater et al., 2002). Agricultural runoff carries pesticides, heavy metals, antibiotics, fertilizers and faecal coliforms, which can pose a potential risk to ecosystem, wildlife, aquatic and human health (Gilpin et al., 2002). The increase in number of water-borne pathogens has been often linked to manure spreading in agricultural watersheds around the world (Trevisan et al., 2002; Gilpin et al., 2002; Christesen, 2002; Hunter et al., 1999; Arienzo et al, 2001). However, many other sources of faecal contamination exist in agricultural watersheds, including sewage sludge, animal faeces deposited on the land by grazing animals, discharges from septic tank systems, and defecation by indigenous fauna (Tyrrel and Quinton, 2003). For example, researchers have found that faecal coliform loads on ungrazed grasslands compared to cattle grazing areas were quite high (Doran and Linn, 1979). Also, Staphylococcus aureus has been found in the vicinity of poultry processing plants, especially in defeathering machines. It can increase up to 1000-fold on carcasses during processing. The most commonly used methods (water immersion chilling and spray-washing) to clean up the machinery are not sufficient to kill the pathogen and usually just result in contaminated runoff (Dodd et al., 1993). Livestock faecal waste contains pathogenic organisms such as E. coli O157:H7, Salmonella and Campylobacter. In 1998, a study showed that approximately 89% of cattle in the UK shed  1  Campylobacter in feces, for example (Stanley et al., 1998). The number of pathogens shed by an animal is usually influenced by the animal age, diet, stress and season (Gilpin et al., 2002). Then, the probability of pathogen transport on the soil surface to a water body becomes dependent on duration and conditions of storage prior to manure spreading or rainfall (Tyrrel and Quinton, 2003). Temperature has been identified as the most important factor influencing pathogen survival in the environment. Other factors include pH, desiccation and competition (Tyrrel and Quinton, 2003). Research has shown that E. coli has the ability to survive composting even if manure-handling guidelines are followed. Thus, it is not surprising that flood irrigation with water contaminated with cattle or poultry feces is a major vehicle for pathogen introduction to surface water (Solomon, 2002). What makes the presence of pathogens in agricultural watersheds a greater public health concern is that there has been an increase in the antibiotic resistance of microbial populations in those ecosystems (Rooklidge, 2004). Antibiotics are used in agriculture to prevent and treat disease, and promote growth (Rooklidge, 2004). Consequently, animal waste can potentially contain: a) numerous bacteria that may be resistant to one or multiple antibiotics due to exposure in the host’s gut (Diarra et al., 2007; Witte, 2000), and b) some concentration of antimicrobial agents that, which once in the environment can affect selective pressures acting on bacteria thereby promoting resistance (Richardson and Bowron, 1985; Hirsch et al., 1999).  1.2.2 Pathogen transport and storm events Most pathogens that cause waterborne diseases leave the host via defecation either in a protected, encased form (cyst, oocyst or spore) making survival in a harsh external environmental more likely; or are shed in high numbers to increase the probability of completing a successful lifecycle (Rosen, 2000). Once microorganisms are on the soil surface, they can exist in three possible states: a) attached to soil particles, b) attached to waste particles, and c) free unattached aggregates (Tyrell and Quinton, 2003). Through the process of overland flow, microorganisms may be transported to an aquatic environment. This can occur either when rainfall intensity exceeds the soil infiltration rate (mm/hr), thus causing the water to flow over ground to the nearest river, stream or bank, or it can occur when the soil is being watered to a saturation point so that no more water can enter the soil. While shallow subsurface flow might contribute to transport, filtration, adhesion and biogeochemical processes limit the amount of transported particles to water bodies (Burt, 1997).  2  Once these microbial populations reach the surface water, several scenarios can occur. The bacteria can remain suspended in the water column as planktonic single cells or as aggregates. Bacteria can also colonize surfaces to form biofilms. Lock (1994) divides colonisable surfaces in natural environments into: a) soft surfaces, such as the sediment, or b) hard surfaces, such as stones, plants, or other solid debris in the water. However, Lock’s dichotomous characterization of colonisable surfaces, particularly sediment, is probably too simplified due to the highly variable nature of this substratum type. Researchers have hypothesized that an undisturbed section of sediment should be completely filled with bacterial coatings and extracellular polymeric substance (EPS) from biofilms. But the sediment surface is still expected to have different types of bacterial colonizers than the deeper parts due to the different environmental conditions (Lock, 1994). Studies estimating the number of bacteria in different biofilm compartments in the natural environment have been limited, although such studies have been conducted in water distribution systems (Morin et al., 1999). Since sediments are filled with microbial populations coating the grains, studies have shown that they serve as reservoirs for faecal pollution. Disturbances of sediments, resulting from storm flows for example, lead to the suspension of that sediment, the impacts of which can be found up to 1 km downstream (Crabill et al., 1999). Many authors have shown that the concentration of faecal coliforms in sediment can be several orders of magnitude higher than concentrations in the water column. The survival in the sediment is enhanced by the presence of soluble organics, which result in higher heterotrophic activity (Crabill et al., 1999; Obiri-Danso and Jones, 2000). Once the sediment is suspended, it causes the scattering of light by fine particles, giving it the cloudy appearance we refer to as turbidity (Davies-Colley and Smith, 2001). Analyses of hospital admissions and water turbidity readings have shown correlations, but have not described any specific time lag relationships (Payment, 2003).  1.2.3 Monitoring microbial water quality In the water column, the majority of pathogens die off due to environmental stressors, such as temperature, freeze-thaw cycle, predation, competition, disinfection, pollution, and so on (Rosen 2000). Pathogen die-off thus results in low pathogen concentrations in aquatic environments that may still cause disease if water consumption occurs directly (e.g. drinking, recreation) or indirectly (e.g. consumption of vegetables irrigated or washed with contaminated water). As mentioned previously, most waterborne pathogens we are concerned about are primarily derived from faecal contamination (human, avian and/or animal). Thus, the monitoring of water for faecal  3  indicator bacteria has been widely accepted for years as a quick and inexpensive solution to protect public health from faecally-derived pathogens (Bartram and Howard, 2003). The first systematic definition of indicator organisms was proposed by Bonde in 1966 (Yates, 2007). Bonde’s definition included biological attributes of the indicator, as well as methodological issues in culturing, isolating and identifying the organism. Among the biological criteria were the presence of the indicator when the pathogen is present; the inability of the indicator to proliferate in the natural environment, its ability to survive longer than the pathogen and in higher numbers; and be randomly distributed in the medium to be examined. The technical criteria were related to the ability to grow the indicators on simple media (without inhibition of growth when other organisms are present) and the ability to clearly identify the organism. In 2004, the United States’ National Research Council developed a new definition and stated that the very first criterion should be an indication of potential risk to public health (National Research Council, 2004). Among commonly used indicators are faecal coliforms (FCs), E. coli (EC), and enterococci (ENT). The most sensitive indicator of faecal contamination is still E. coli. The large numbers of E. coli present in the gut of warm-blooded animals and the limited presence of E. coli in external environments support its continued use as the best indicator of faecal contamination available (Edberg et al., 2000). However, there have been some concerns raised since E. coli has been shown to be able to proliferate in temperate and tropical non-host environments (Savichtcheva and Okabe, 2006). Unfortunately, research since the 1970s has repeatedly shown that faecal indicators do not always correlate with pathogen concentrations (Glasner and McKee, 2002; Warrington and Johnson, 2001; Harwood et al., 2005; Lemarchand and Lebaron, 2003). It is also unlikely that any indicator system will be useful for all types of pathogens (Yates, 2007). There are also several pathogens that no current or proposed indicators have been found for yet. Thus, there are cases where direct testing for pathogens may give us the most relevant public health information needed to make a decision (Yates, 2007).  1.3  Biofilms  1.3.1 What are Biofilms? Biofilms are accumulations of microorganisms at any surface (also called substratum) submerged in an aquatic environment (Decho, 2000). The cell attachment results in the excretion of an extracellular polymeric substance (EPS), which results in the observed 3-D gelatinous matrix (Gilbert et al., 1993). Within the EPS, the microbial community can grow and reproduce in a  4  protected environment. One of the most commonly used biofilm definitions is that of Characklis and Marshall (1990, p.4), which states: “A biofilm is a surface accumulation, which is not necessarily uniform in space or time, that comprises cells immobilized at a substratum and frequently embedded in an organic polymer matrix of microbial origin.” The existence of biofilms was first noticed in 1943 by a scientist named Zobell, who studied the attachment of bacterial cells in layers to bottle walls. He described the attachment process as time-dependant, and linked it to the corrosion or fouling of submerged materials in seawater (Zobell, 1943). Progress made afterwards was relatively slow until the 1970s. This was when researchers realized how ubiquitous the association of bacteria with surfaces actually was (Bryers, 2000). Major advances in the biofilm field were made after William Characklis provided a two-volume extensive literature review on the fundamentals and applications of “microbial slimes” in 1973 (Characklis 1973a; Characklis, 1973b). His work contributed to biofilm research and hence understanding of microbial activity at interfaces. Since the scientific community became interested in biofilms, biofilms have been found in every aqueous system (Wimpenny et al., 2000; Decho, 2000). They can provide uniform coverage to the colonized surface or they can be patchy, resulting in very different 3-D structures. They can also range in structure from monolayers of scattered single cells to thick mucous structures about 300-400 mm in thickness, such as algal mats (Characklis and Marshall, 1990). Microbial mats and suspended microbial flocs have many important features in common with biofilms and are therefore included in the definition. While microbial mats are formed on most sediments (Karsten and Kühl, 1996), flocs are fragile suspended structures found in river or seawater that typically occur after increased nutrient availability in the water (Decho, 2000). Wimpenny and Colasanti (1997) suggested that biofilm communities take on one of three structural types: the heterogeneous mosaic biofilm, the penetrated water channel biofilm, or the densely packed biofilm. Regardless of type, biofilms are very adsorptive and porous structures, which are more than 95% water. Water can carry large amounts of adsorbed or entrapped materials such as solutes and inorganic particles. Therefore, a biofilm is often composed of large fractions of inorganic and abiotic materials bound together by a biotic matrix (Characklis and Marshall, 1990). As the thickness of a biofilm increases, more microenvironments are created for microbial growth. Since the bulk liquid surrounding the biofilm carries a range of nutrients and  5  gases, transport and diffusion processes become very important. A thick biofilm will contain anaerobic, as well as aerobic regions as a result of oxygen diffusion limitations and respiration processes. An algal mat, on the other hand, will have diurnally changing anaerobic regions based on the production of oxygen by the algal population during daylight hours (De Beer and Stoodley, 2000). Researchers only recently recognized the importance of biofilm growth from a microbiological point of view. For years, planktonic bacterial existence was thought to have benefits such as faster growth and quicker access to nutritionally rich environments (van Loosdrecht et al., 1990). In reality, in dilute and hostile environments, attached growth on surfaces in large communities in a thick extracellular polysaccharide confers far more advantages. The following are some of the advantages discussed in the literature. Biofilm growth has been hypothesized to increase protection from phagocytosis by grazing protozoa in aquatic environments (Gilbert et al., 1993; Matz et al., 2004). Biofilm growth allows resident organisms to tolerate much higher concentrations of biocides and antibiotics in comparison with their planktonic counterparts (Flemming and Wingender, 2001b; Witte, 2000; Kümmerer, 2004). Biofilm growth improves growth prospects through the concentration of nutrients and extracellular products (Bryers, 2000). Biofilm growth allows continual growth at a favourable location, providing organic molecules, nutrition and optimal pH, without being washed away (Korber et al., 1995). Biofilm growth allows cooperative mobilization of nutrients by different organisms in mixed cultures (Gilbert et al., 1993). Biofilm growth allows modulation of the physical and chemical environment (especially the pH and the electropotential gradient across the biofilm) (Wimpenny et al., 1989). Biofilm growth facilitates genetic exchange through cell proximity (Bale et al., 1988). Biofilm growth facilitates the retention of genetic material, cellular debris and exoenzymes (Flemming and Wingender, 2001a). Biofilm growth protects resident organisms from desiccation and thereby increases the chances of maintaining a community for a long period of time compared to planktonic forms of life (Flemming and Wingender, 2001b; Lewandowski and Beyenal, 2003).  6  1.3.2 Biofilm Characteristics Biofilms have many characteristics that separate them from planktonic bacteria. Here are some of these features. A biofilm mainly consists of cells and surrounding EPS. The EPS consists of polysaccharides, proteins, polyuronic acids, nucleic acids and lipids (Schmidt and Ahring, 1994; Flemming and Wingender, 2001a; Sutherland, 2001). If the biofilm were to be described in terms of its transport and thermodynamic properties, the biofilm system would be described in terms of compartments or phases, such as: the substratum, the base biofilm, the surface biofilm, the bulk liquid, and the gas phase (Bryers, 2000). Biofilm formation seems to trigger cell signalling that causes physiological changes between planktonic and attached cells (Stoodley et al., 2002). Within a biofilm, bacterial density is relatively high and can reach 1010 cells per ml of hydrated biofilm (Rittman et al., 2000). Bacterial density of a biofilm is highly dependent on its 3-D structure, and the flow regime the biofilm was exposed to during colonization. For biodegradable substrates, the potential to consume the substrate is relatively high per unit volume of the aggregate (De Beer and Stoodley, 2004). The presence of an EPS matrix provides mechanical stability. Three types of bonds facilitate this interaction (Flemming et al., 1999): London dispersion forces, electrostatic forces, and hydrogen bonds Mass transfer is the limiting factor, because molecular diffusion is the main transport mechanism (Characklis and Marshall, 1990). Wide ranges of metabolic, physical and chemical heterogeneities occur within biofilms. For example, substrate concentration gradients form within the EPS matrix of the biofilm. When substrate removal kinetics in the biofilm are low and biofilms are thin, the relative concentration of substrates in the bulk liquid phase (or the water column) is much lower than that in the biofilm (Rittmann, 2004). Also, the dense aggregation of microorganisms can lead to physical and environmental changes within micrometers of the same biofilm through metabolic activities thus creating microniches. The creation of these microniches allows for the formation of a very complex and heterogeneous microbial community (Flemming et al., 1999).  7  Population density also controls gene expression, which regulates other behaviours such as sporulation, hydrolytic enzyme releases, and biofilm development (van Loosdrecht et al., 1990).  1.3.3 Maturation and succession in biofilms 1.3.3.1 Formation Although many factors affect the characteristics of the mature biofilm community, the process of biofilm formation and persistence can be broken down into six stages that are reported in the literature. 1. Development of a surface-conditioning film: In aqueous environments, any surface attracts dissolved organic molecules and other macromolecules that adsorb to it from the surrounding liquid (Bryers and Fletcher, 2000). The adsorption process of this organic film is very rapid in comparison to other biofilm processes. Scientists have shown that surface characteristics (wettability, surface tension, electrophoric mobility) had very little effect on the speed of adsorption even in waters with very low organic concentrations (Percival et al., 2000). Until recently it was believed that specific substrata had little effect on the composition of the conditioning film, but conditioning layer amounts, uniformity and bacterial compositions were still being investigated (Geesey et al, 2000). In natural aquatic environments, surface films seem to not be affected by substratum characteristics, and these organic molecules mostly tend to be polysaccharides or glycoproteins. The adsorption of molecules leads to several alterations in substratum properties. A decrease in hydrophobicity immediately results and a net negative charge is gained. Contact potentials and surface tensions are either decreased or increased based on the initial surface energy. Adsorption of a conditioning film decreases the surface energy of a clean high-energy surface, but has little effect on low-energy surfaces, such as silicone (Baier, 1975). The energy of a surface could be defined in terms of the surface’s hydrophobicity (low energy surfaces are often hydrophobic surfaces), or in terms of the surface’s Gibbs free energy. 2. A series of events that brings microorganisms closer to the surface: These commonly include mass transport of bulk fluid, thermal effects (Brownian motion and molecular diffusion) and gravity effects (differential settling and sedimentation) (Percival et al., 2000). Chemotaxis and motility can also bring microorganisms closer to a surface and have important role in biofilm formation (Merritt et al., 2007).  8  3. Adhesion of bacterial cells to the surface: Since both the conditioned substratum and the bacteria carry a net negative charge in the natural environment under normal pH conditions, we might assume that this creates a problem. This problem is generally avoided by the fact that the body of the bacterium does not come into direct contact with the surface. The bacterium utilizes extracellular structures, such as pili, fimbriae, flagella between itself and the substratum (Characklis and Marshall, 1990). Many of these extensions are equipped with adhesins that facilitate the attachment process (Wimpenny, 2000). This initial adhesion can be reversible or irreversible. 4. Desorption of reversibly adsorbed cells and irreversible adsorption of bacterial cells at a surface: Reversible adhesion refers to the association of bacteria with a surface using weak forces that can easily be broken. Under those conditions, the bacterium continues to exhibit some Brownian motion that could result in detachment. A bacterium bound by an irreversible adhesion, on the other hand, no longer exhibits Brownian motion and cannot be removed by weak shear forces (Bryers and Fletcher, 2000). There are two schools of thought on what explains forces allowing attachment. The first approach is a colloidal approach and considers adsorption in terms of “interfacial forces”. Long range forces involve the interaction between electrical double layer repulsion forces and van der Waals attraction forces and result in reversible adsorption; while short range forces involve chemical dipole bonding and hydrophobic interactions between extracellular components of the bacteria and the substratum resulting in irreversible binding (Korber et al., 1995). The other approach explains the attachment through a Gibbs free energy system. Here, adsorption is considered in terms of attraction between the bacterial surface free energy, the surface free energy of the substratum, and the surface tension of the liquid (Decho, 2000; Sommer et al., 1999). 5. Growth and division of the microorganisms, resulting in surface colonization, microcolony formation, and biofilm formation: Although biofilms can range from simple aggregates of bacteria to more complex communities, most biofilms in the aquatic environment are highly structured multi-species communities (Rickard et al., 2003). Microbial aggregates are kept together through the EPS. While some bacteria use the formation of filamentous growth as a form a reproduction, the most common method of reproduction remains binary fission. Bacteria growing in biofilms are copiotrophs that require relatively high levels of energy to reproduce. Thus, in oligotrophic environments they tend to starve. This results in a reduction in size (0.2 µm- 0.5 µm), as well as in endogenous metabolism (Bryers and Fletcher, 2000). Starved bacteria also tend to adhere  9  more strongly to surfaces than well-fed cultures and are capable of using energy substrates bound to surface for reproduction and growth (Morita, 1982). 6. Transport of substrates to and within biofilm: The bulk liquid surrounding the biofilm carries a lot of the nutrients and gases that the microcolonies require to persist. The transport in and out of the biofilm results in the formation of a strong diffusion gradient (Rittmann, 2004). The gradient is also maintained via physiological activity that assists diffusion and thus recycling of materials in the matrix (Sutherland, 2001). It has recently been shown that this resource concentration affects not only the morphology of the biofilm (stalked irregular branching, mushroom-shaped, or homogenous), but also the degree of porosity (Wolfaardt et al., 2000).  1.3.3.2 Detachment There are three kinds of factors that lead to the detachment of bacteria from a biofilm: biological, chemical, or physical (Brading et al., 1995). Rittman (1989) divided the physical factors affecting detachment into three main processes: shear removal, sloughing and abrasion events. Shearing (the continuous removal of biofilm in small sections) is mostly dependent on fluid dynamic conditions. As biofilm thickness and fluid shear stress at the biofilm-fluid interface increase, shear removal increases too. Sloughing, however, causes quick, massive losses in comparison. It occurs most commonly in nutrient rich environments and thick biofilms (Ohashi and Harada, 1996). Also, while shear removal seems to mostly be uniform erosion on the outside surface, sloughing is more of a random, sporadic process that causes the loss of large chunks or complete biofilms. It has been hypothesized that sloughing occurs due to depletion of oxygen or nutrients deep inside the biofilm or due to dramatic changes in the outer environment. So, as a biofilm becomes thicker, less oxygen diffuses into the biofilm matrix. In the anoxic regions of that biofilm, facultative anaerobic bacteria commence converting organic substrates into volatile fatty acids and insoluble gases, both of which may weaken the structure of the biofilm (Bryers, 2000). Finally, abrasion is the loss of biofilm due to the repeated collision of substratum particles. The detachment rate coefficient of a biofilm is highly affected by the physiological changes brought about by abrasion. This detachment coefficient tends to be lower in thin, dense biofilms that formed over time in high abrasion environments (Moore et al, 2000). Chemical processes are also very important. For example, the surrounding medium can have a large effect on the strength of association between cells and the biofilm. If the bulk liquid lacks calcium or magnesium, the association can become weaker (Decho, 2000). Other chemical  10  factors include cation exchange, substrate changes, substratum changes, and charge accumulation (Brading et al., 1995). Many biological factors can be considered when thinking about biofilm detachment. One of the most obvious biological factors in natural systems is predation and grazing by protozoa (Huws et al., 2005) and metazoa (Hunt and Parry, 1998). Also during replication, daughter cells can be detached from the mass by production of unattached daughter cells (Characklis, 1990). Lack of oxygen and increased concentration of waste products in the vicinity of attached cells can also affect biologically active cells (Rittmann, 1989). Some cells are more likely to detach during certain parts of their life cycles. Acinetobacter calcoaceticus, for example, produces a hydrophilic polysaccharide capsule during its stationary phase that causes the bacterium to detach. In general, it has been observed that rapidly dividing bacteria are not very good at attaching to biofilms (Brading et al., 1995). Variations in bacterial species may also contribute to detachment, especially in mixed biofilms. The detachment process affects the species distribution and might explain why succession in species occurs as a biofilm develops. Also, mixed cultures are more susceptible to sloughing. This could either be due to the fact that bacteria depolymerize the EPS of other species, making them less stable, or because the EPS and the microbial surfaces are incompatible, causing a weakness in binding (Brading et al., 1995). Bacteria are more likely to detach from biofilms in nutrient rich environments (Sawyer and Hermanowicz, 2000; Ohashi and Harada, 1996). In some cases, organisms have been known to produce enzymes to detach from a biofilm with less desirable qualities. Pseudomonas aeruginosa produces alginate to enhance its attachment and alginate lyase to induce sloughing under starvation conditions (Moore et al., 2000). Finally, the type of adhering bacteria can affect the rate of detachment. For example, rods and coccoids have low surface roughness in comparison with filamentous bacteria.  1.3.4 Can biofilms be used as monitoring tools 1.3.4.1 Presence of pathogens in biofilms Biofilms can serve as passive samplers in aquatic environments, by giving an idea of what the environment has previously (historically) been exposed to (Sabater et al, 2007). Much work has been completed on the retention, mineralization, degradation and transport of organic and inorganic contaminants in biofilms (Wolfaardt, 2000; Fuchs et al., 1996; Flemming and Wingender, 2001b; Schorer and Eisele, 1997; Ebise and Inoue, 2002). However, not much work  11  has focussed on the role biofilms could play in the study and prediction of microbial water quality. Since bacterial concentrations tend to be higher in biofilms than in the water column (previously referred to as the bulk liquid), the growth of opportunistic pathogens within biofilms is of particular interest to public health protection. Biofilms allow some pathogens to proliferate in the external environment outside the body of the host and increase their numbers to what might be an infective dose (Barbeau et al., 1998). Pathogen association with biofilms in drinking water distribution systems and potable water sources has been demonstrated (Keevil, 1999; Walker et al., 1995). Flanders and Yildiz (2004) pointed out that the protected, nutrient-rich environment of the biofilm might be ideal for pathogen retention. Pathogens, such as E. coli, Campylobacter jejuni and Pseudomonas aeruginosa have been shown to survive and proliferate in biofilms (Rittmann, 2004). Other examples of pathogens utilizing the heterogeneity of biofilms to survive are Helicobacter pylori and Legionella pneumophila. Cooperation between such microaerophilic pathogens and heterotrophs in low redox zones of the biofilm has been reported (Keevil, 2003). As for studies in natural aquatic systems, a study by Muirhead et al. (2004) simulated an artificial in-stream flood in the absence of rainfall and compared the contribution made by river rock biofilms and sediments to the loading of E. coli in a stream under dry conditions and found that sediment made the major contribution. It should be noted, however, that the study only looked at the concentration of indicator E. coli in the stream.  1.3.4.2 Could substratum choice affect the climax community in a biofilm? Biofilm studies investigating natural aquatic systems have frequently observed that substratum choice influenced bacterial biomass (Hunt and Parry, 1998), species richness (Baldy et al., 1995; Barbiero, 2000), species succession and colonization patterns (Tank and Dodds, 2003), heterotrophic activity (Romani and Sabater, 2000), and pollutant concentrations in the EPS (Kroepfl et al., 2006); all of which could influence pathogen and indicator survival in a biofilm. Algae, which are predominant in periphyton, can enhance bacterial growth in a biofilm through the release of extracellular dissolved organic compounds, but may inhibit bacterial growth by releasing algal exudate toxins (Olapade and Leff, 2006). Hunt and Parry (1998) found that the differences in bacterial biomass on a surface, which result from differences in chemical and physical characteristics, may even influence later predation patterns. These observed differences could be related to a variety of factors already addressed in this introduction, such as varying  12  levels of hydrophobicity (Cerca et al., 2005), surface tension (Characklis and Cooksey, 1983), and surface roughness (Hunt and Parry, 1998).  1.4 Research objectives and hypotheses 1.4.1 Main research objective The main objective of this study was to improve our understanding of the presence and ecology of heterotrophic activity indicators (heterotrophic bacteria), faecal indicator bacteria (faecal coliforms, E. coli and enterococci) and pathogens (E. coli O157, Salmonella sp., and Campylobacter sp.) in an agricultural stream to improve water quality monitoring practices for these organisms. The study focused on examining the differences in distribution of these organisms in biofilms on different substratum surface types (river rocks, slate rocks, Alder wood, Lexan ™, and sandpaper), in sediment and in the commonly examined water column. The results were used to assess the potential role biofilms could play as microbial sinks in the natural aquatic environment and in the prediction of water quality. The study was also investigating substrata that may improve sampling for indicator bacteria and pathogens in raw water sources.  1.4.2 Research questions and hypotheses The following research questions were the basis for this work. 1. Is there a difference among the concentrations of faecal indicator bacteria and prevalences of pathogens in biofilms on different substratum surface types, sediment, and in the water column? Ha: The concentrations of faecal indicator bacteria and the prevalences of pathogens in biofilms on different substratum surface types and in sediment, are higher than those in the water column at different sites in the watershed. Hb: The concentrations of faecal indicator bacteria and the prevalences of pathogens in biofilms on different substratum surface types, sediment, and in the water column in the dry season will be different from those in the wet season. Hc: Changes in water quality parameters (physical, chemical and biological) will result in changes in the concentrations of faecal indicator bacteria and the prevalences of pathogens in biofilms on different surface types, sediment, and in the water column.  13  2. Is there a difference among the concentrations of faecal indicator bacteria and prevalences of pathogens in biofilms on different substratum surface types based on biofilm age (or substratum colonization period)? Ha: Biofilms colonizing different substratum surface types for a short period of time (one to four weeks) will have different concentrations of faecal indicator bacteria and prevalences of pathogens than biofilms colonized for either one month or for longer periods of time (12 to 24 weeks). Hb: Changes in water quality parameters (physical, chemical and biological) will result in different changes in the concentrations of indicator bacteria and pathogens in biofilms on different substratum surface types that were colonized for different periods of time. 3. Is there a difference in the prevalences of antibiotic resistant E. coli and pathogenic E. coli O157 in biofilms on different substratum surface types, sediment and in the water column? Ha: The prevalences of antibiotic resistant E. coli and pathogenic E. coli O157 sampled from different sites and from different surfaces (biofilms on different surface types, sediment and in the water column) will be different. Hb: Changes in water quality parameters (physical and chemical) may affect the prevalences of antibiotic resistant E. coli isolated from biofilms on different surface types, sediment and in the water column. These research questions and hypotheses were based on several assumptions and previous research (the combination of which was used to develop the conceptual framework presented in Figure 1.1). One of the first assumptions in this study was that the agricultural stretch of Elk Creek would have higher inputs of faecal indicators and pathogens as a result of agricultural practices and thus higher concentrations of both indicators and pathogens may be retrieved at agriculturally impacted sites. It was also hypothesized that despite wildlife impacts, the headwaters would be relatively pristine in comparison with agriculturally impacted sites due to the absence of agriculture, and thus lower concentrations of indicators and pathogens were expected to be retrieved from the designated control site. These hypotheses were based on what was known about this watershed from previous investigations.  14  The Elk Creek Watershed is located in the Fraser Valley in British Columbia (Canada). It has an area of about 28 km2 and the watercourse is about 12 km long (Rood and Hamilton, 1995). It is mainly agricultural in nature, although it also has forested headwaters located in the mountains, and limited urban residential influences. The watercourse experiences an increasing pollution gradient as it flows north and drains into the Fraser River at Hope Slough. The agricultural watershed is utilized for dairy production and livestock rearing (2.46 animal units per hectare), field crops, greenhouses and pastureland. Manure and commercial fertilizer are applied to more than half of the total agricultural area (58%), while herbicides and insecticides are applied to 29% and 12% of the total agricultural land area; respectively (Vingarzan et al., 2002). The water in Elk Creek was designated for drinking purposes, protection of aquatic life, wildlife, livestock, irrigation, and recreation (Swain and Holms, 1985). After a Cryptosporidium parvum outbreak in April 1998, the water is no longer used for drinking purposes (Newman et al., 2003). The watershed was chosen for several reasons. Although the stream is no longer used for drinking water purposes, the control site in the headwaters was still a good surrogate for other small community drinking water systems in rural Canada, and the presence of pathogens in the water could still affect the health of livestock, farm animals, wildlife, fish, ecosystem health and human health (through contamination of irrigated vegetables, recreation, or occasional consumption). Elk Creek was also chosen for this study because the agriculturally-impacted sites downstream offered a variety of nutrient concentrations and water quality variables that needed to be inspected to establish some idea of the effects those conditions could have on pathogens and indicators in biofilms. Previous studies of Elk Creek (Schendel et al., 2004; Derksen et al., 2004), as well as the intensive agricultural nature and manure spreading behaviour in the watershed indicated a high likelihood of faecal contamination of the water body. This study was field-based for several reasons. While laboratory-based studies allow more control over variables and facilitate the drawing of conclusions, this study was designed to make observations about pathogen and indicator concentrations in biofilms in natural systems that may be relevant for raw drinking water in other watersheds. If this objective was to be met in the laboratory, we would have needed to simulate a variety of variables for which we may not be able to predict relevant concentrations or levels. For example, the concentrations of indicators and pathogens that actually reach the water body exhibit large variations under different environmental conditions. Many of these variations are the result of agricultural practices, as well as natural geological, biological and chemical characteristics of the aquatic system. Also, a large number of variables have been associated with pathogen survival in laboratory-based studies. The  15  control of many such variables in a microcosm in a manner that would make our results relevant to natural systems would be very difficult, especially for factors such as rainfall and synergistic nutrient changes. Also, growth-based methods were chosen for the isolation and identification of pathogenic organisms for two reasons: a) to focus on organisms that we know are associated with infection, since the role of viable non-culturable bacteria is still debated (Thomas et al., 1999), and b) to make our results replicable in settings where molecular-based methods are not available to public health inspectors, water managers, or scientists. Another assumption of this study was that rainfall, and thus seasonality, would have some effects on the concentration of faecal indictor bacteria and the prevalence of pathogens in the watershed due to the major role agricultural runoff plays in: a) carrying both faecal contamination bacteria and nutrients to the watershed; and b) altering the flow patterns of the creek. Previous studies have shown that peak runoff in the Elk Creek watershed occurs in late spring during snowmelt, and the flow becomes very low during late summer and early autumn. The estimated mean annual flow of this watershed is 0.61 m3/s (Rood and Hamilton, 1995). The dry season with limited rainfall typically begins in April and ends in September, while the wet season usually begins in October and ends in March. During the wet season, 70% of the annual rainfall of approximately 1500 mm falls in the watershed. Biofilms were hypothesized to provide faecal bacteria and pathogens with protected niches, nutrients and possibilities for extracellular exchanges. Therefore, survival in biofilms was hypothesized to be a preferred form of existence and concentrations of indicator organisms and pathogens were expected to be higher in biofilms and sediment than in the water column. If that is the case, then sampling the water column using grab samples to detect the presence of indicators and pathogens may underestimate the concentrations of bacteria in the aquatic environment. The pathogens investigated in this study were pathogenic E. coli O157, Salmonella sp, and Campylobacter sp. These are pathogens that have commonly been associated with agricultural ecosystems and are considered emerging waterborne pathogens of risk to public health (Huffman et al., 2003). All three pathogens use the intestinal tract of many warm-blooded animals as their host environment and thus end up in the manure of those host organisms (Pell, 1997; Winfield and Groisman, 2003; Hutchinson et al., 2004). These pathogens are also commonly found in freshwater and in biofilms (Buswell et al., 1999; Reeser et al., 2007; Zimmer et al., 2003; Winfield and Groisman, 2003; Armon et al., 1997; LeJeune et al., 2001). The study also monitored a selection of commonly used faecal indicator bacteria (faecal coliforms, E. coli, and enterococci) and an indicator of heterotrophic activity (heterotrophic plate counts) to improve our  16  understanding of the ecology of indicators in natural aquatic systems, and in biofilms in particular. The investigation of indicators and pathogens in biofilms was of interest because the need exists for microbial water quality studies to define under what circumstances and to what extent indicators correlate with pathogens (Savichtcheva and Okabe, 2006). Perhaps biofilms allow for a closer association or a better correlation. Another hypothesis was that substratum type affects a variety of biofilm characteristics (EPS mechanical strength and chemical characteristics, biofilm community composition, biofilm 3-D structure, etc). Substratum type was therefore thought to affect concentrations of indicator organisms, pathogen prevalence, as well as the role water quality variables would play in determining indicator and pathogen concentrations on different surface types. Biofilm age was also hypothesized to influence concentrations of indicator bacteria and the prevalence of pathogens on different surfaces since biofilms go through different stages of succession and thus competition. Microniche development may still be in progress in younger biofilms, thus increasing die-off rates through competition; while EPS mechanical strength, oxygen diffusion rates, and development of anoxic zones in older biofilms may drive sloughing, thus resulting in lower concentrations. Mature (one-month) old biofilms were hypothesized to have the highest concentrations of faecal indicator bacteria and prevalence of pathogens. Due to the differences in biofilm characteristics at different stages of substratum colonization, water quality variables may have different effects on biofilms based on their age. Finally, antibiotic resistance was hypothesized to be highest in the agricultural stretch of the watershed due to the high levels of resistant bacterial inputs, and as well as high selective pressures. Antibiotic resistant E.coli at the headwaters were hypothesized to be rare, if present at all. These assumptions were based on previous studies associating agriculture and antibiotic resistance (Kümmerer, 2003). E. coli resistance to antibiotics has frequently been reported in isolates from aquatic ecosystems (Sayah et al., 2005; Edge and Hill, 2005; Schwartz et al., 2003; Watkinson et al., 2007). Also, some studies were able to isolate resistant E. coli O157 from multiple use watersheds (Hamelin et al., 2006; Hamelin et al., 2007). Figure 1.1 illustrates a variety of inputs and processes in agricultural watersheds that may result in the contamination of aquatic environments with antibiotic resistant or susceptible faecally-derived bacteria and pathogens after a rainfall or irrigation event. The figure highlights the main compartments where faecal indicators and pathogens may partition or adsorb: the water column, sediment or biofilms.  17  This thesis is divided into six chapters (the first being the introduction). In the second chapter, the distribution of indicator organisms and pathogens on different substratum surface types and in water and sediment is compared across different sites and different seasons. In Chapter 3, the relationships among the numbers of indicator bacteria and pathogens frequencies on different substrata and different variables, which were thought to affect accumulation or survival in biofilms, are analyzed. Among these variables were physical, chemical and biological water quality variables and substratum colonization period (short-term versus long-term). The fourth chapter focuses on the distribution of Campylobacter sp. on different surfaces and demonstrates a method through which the effectiveness of different substrata as Campylobacter sp. monitoring tools can be determined. Chapter 5 discusses the presence and distribution of antibiotic resistant E. coli in the stream and explores the correlation of water quality variables to resistance to a variety of commonly used antibiotics. The final chapter is a summary of the findings, their significance and limitations.  18  Agriculture in an intensive agricultural watershed  Dairy Production  Manure or fertilizer on land  Crop production  Pastureland  Manure contaminated with pathogens  Manure contaminated with faecal bacteria  Water column  Sediment  Biofilms on rock  Concentration in different aquatic environment compartments that could affect livestock, wildlife, human and ecosystem health  Rainfall or irrigation results in runoff  Results of concern to our study  Animals defecate on land or into water  Manure spread on land for fertilization or waste disposal purposes  Manure contaminated with resistant faecal bacteria and pathogens  Wildlife  Process  Livestock and Poultry rearing  Pesticides  Inputs  Antibiotics for disease treatment, prevention and growth promotion  Figure 1.1 Effects of agricultural activities on microbial water quality in different aquatic compartments of interest to this study.  19  1.5 References Arienzo, M., Adamo, P., Bianco, M.R., Violante, P., 2001. Impact of land use and urban runoff on the contamination of the Sanro River Basin in Southwestern Italy. Water, Air and Soil Pollution 131, 349-366. Armon, R., Starosvetzky, J., Arbel, T., Green, M., 1997. Survival of Legionella pneumophila and Salmonella typhimurium in biofilm systems. Water, Science and Technology 35 (11), 293-300. Baier, R.E., 1975. Applied chemistry at interfaces. In: Baier, R.E. (Ed.). Applied Chemistry at Protein Interfaces, Advances in Chemistry Series 145 Washington, American Chemistry Society, 1-25. Baldy, V., Gessner, M.O., Chauvet, E., 1995. Bacteria, fungi and the breakdown of leaf litter in a large river. Oikos 74 (1), 93-102. Bale, M.J., Fry, J.C., Day, M.J., 1988. Transfer and occurrence of large mercury resistant plasmids in river epithelion. Applied and Environmental Microbiology 54, 972-978. Barbeau, J., Gaulthier, C., Payment, P., 1998. Biofilms, infectious agents and dental unit waterlines: a review. Canadian Journal of Microbiology 44, 1019-1028. Barbiero, R.P., 2000. A multi-lake comparison of epilithic diatom communities on natural and artificial substrates. Hydrobiologia 438, 157-170. Bartram, J., Howard, G., 2003. Drinking water standards for the developing world. Emerging waterborne pathogens. In: Mara, D., Horan, N. (Eds.) Handbook for Water and Wastewater Microbiology. Elsevier Press, London, 221-241. Brading, M.G., Jass, J., Lappin-Scott, H.M., 1995. Dynamics of bacterial biofilm formation. In: Microbial Biofilms. Cambridge, University Press, 46-63. Bryers, J.D., 2000. Biofilms: An introduction. In: Bryers, J.D. (Ed.) Biofilms II Process Analysis and Applications. New York Wiley-Liss, 3-12. Bryers, J.D., Fletcher, M., 2000. Biofilm formation and persistence. In: Bryers, J.D. (Ed.) Biofilms II: Process Analysis and Applications. New York, Wiley-Liss, 48-89. Buswell, C.M., Herlihy, Y.M., Keevil, C.W., Leach, S.A., 1999. The effect of aquatic biofilm structure on the survival of enteric pathogen Campylobacter jejuni in water. In: Keevil, C.W. Godfree, A., Holt, D., Dow, C. (Eds.) Biofilms in the Aquatic Environment. Royal Society of Chemistry, Cambridge, 153-159.  20  Burt, T.P., 1997. The hydrological role of floodplains within the drainage basin system. In: Haycock, N., Burt, T., Goulding, K., Pinay, G. (Eds.) Buffer Zones: Their Processes and Potential in Water Protection. Harpenden, UK: Quest Environmental, 21-32. Cerca, N., Pier, G.B., Vilanova, M., Oliveira, R., Azeredo, J., 2005. Quantitative Analysis of adhesion and biofilm formation on hydrophilic and hydrophobic surfaces of clinical isolates of Staphylococcus epidermiditis. Research in Microbiology 156, 506-514. Characklis, W.G., 1973a. Attached microbial growth: I. Attachment and Growth. Water Research 7, 1113-1127. Characklis, W.G., 1973b. Attached microbial growth: II. Frictional losses due to microbial slimes. Water Research 7, 1249-1258. Characklis, W.G., Marshall, K.C., 1990. Biofilms: A basis for an interdisciplinary approach. In: Characklis, W.G., Marshall, K.C. (Eds.) Biofilms. Wiley-Interdisciplinary Publication, New York, 3-15. Christesen, L., 2002. Dairy farming and river condition: investigating the sustainable use of water resources in an agricultural area. Water, Science and Technology 45 (11), 21-28. Crabill, C., Donald, R., Snelling, J., Foust, R., Southam, G., 1999. The impact of sediment faecal coliform reservoirs on seasonal water quality in Oak Creek, Arizona. Water Research 33 (9), 2163-2171. Davies-Colley, R.J., Smith, D.G., 2001. Turbidity, suspended sediment, and water clarity: A review. Journal of American Water Research Association 37(5), 1085-1101. De Beer, D., Stoodley, P., 2004. Microbial biofilms. Springer Verlag, New York. Decho, A.W., 2000. Microbial biofilms in intertidal systems: An overview. Continental Shelf Research 20, 1257-1273. Derksen, G., Schreier, H., Bestbier, R., Brown, S., 2004. Community watershed information system: Elk Creek watershed case study. Environment Canada & IRE-UBC (Multi-Media CD-ROM). Diarra, M.S., Silversides, F.G., Diarrassouba, F., Pritchard, J., Masson, L., Brousseau, R., Bonnet C., Delaquis, P., Bach, S., Skura, B.J., Topp, E., 2007. Impact of feed supplementation with antimicrobial agents on growth performance of broiler chickens, Clostridium perfringens and Enterococcus number, antibiotic resistant phenotype, and distribution of antimicrobial resistance determinants in Escherichia coli. Applied and Environmental Microbiology 73 (20), 6566-6576. Dodd, C.E.R., Chaffey, B.J., Daems, K., Waites, W.M., 1993. Adherent growth of Staphylococcus aureus from poultry processing plants. In: Denyer, S.P., Gorman, S.P.,  21  Sussman, M. (Eds.) Microbial Biofilms: Formation and Control. Blackwell Scientific Publications, London, 285-292. Doran, J.W., Linn, D.M., 1979. Bacteriological quality of runoff water from pastureland. Applied and Environmental Microbiology 37, 985-991. Ebise, S., Inoue, T., 2002. Runoff characteristics of pesticides from paddy fields and reduction of risk to the aquatic environment. Water, Science and Technology 45 (9), 127-131. Edberg, S.C., Rice, E.W., Karlin, R.J., Allen, M.J., 2000. Escherichia coli: the best biological drinking water indicator for public health protection. Journal of Applied Microbiology Symposium Supplement 88, 106S-116S. Edge , T.A., Hill, S., 2005. Occurrence of antibiotic resistance in Escherichia coli from surface waters and faecal pollution sources near Hamilton Ontario. Canadian Journal of Microbiology 51, 501-505. Flanders, J.R., Yildiz, F.H., 2004. Biofilms as reservoirs for disease. In: Ghannoum, M., O’Toole, G.A. (Eds.) Microbial Biofilms. ASM Press, Washington DC, 314-31. Flemming, H.C., Wingender, J., Moritz, R., Borchard, W., Mayer, C., 1999. Physico-chemical structure of biofilms – A short review. In: Keevil, C.W. Godfree, A., Holt, D., Dow, C. (Eds.) Biofilms in the Aquatic Environment. Royal Society of Chemistry, Cambridge, 1-12. Flemming, H.C., Wingender J., 2001a. Relevance of microbial extracellular polymeric substances (EPSs) - Part I: Structural and ecological aspects. Water, Science and Technology 43 (6), 1-8. Flemming, H.C., Wingender J., 2001b. Relevance of microbial extracellular polymeric substances (EPSs) - Part II: Technical aspects. Water, Science and Technology 43 (6), 9-16. Fuchs, S., Haritopoulou, T., Wilhelmi, M., 1996. Biofilms in freshwater ecosystems and their use as a pollutant monitor. Water, Science and Technology 34 (7-8), 137-140. Geesey, G.G., Beech, I., Bremer, P.J., Webster, B.J., Wells, D.B., 2000. Biocorrosion. In: Bryers, J.D. (Ed.) Biofilms II: Process Analysis and Applications. New York Wiley-Liss, 281325. Gilbert, P., Evans, J.D., Brown, M.R.W., 1993. Formation and dispersal of bacterial biofilms in vivo and in situ. Journal of Applied Bacteriology Symposium Supplement 74, 67S-78S. Gilpin, B.J., Gregor, J.E., Savill, M.G., 2002. Identification of the source of faecal pollution in contaminated rivers. Water, Science and Technology 46(3), 9-15. Glasner, A., McKee, L., 2002. Pathogen occurrence and analysis in relation to water quality attainment in San Francisco Bay area watersheds. San Francisco Estuary Institute  22  Report, San Francisco, USA. Retrieved January 29th, 2008. Document available at: http://www.sfei.org/watersheds/reports/Pathogens/Pathogenlitrevv6.pdf Hamelin, K., Bruant, G., El-Shaarawi, A., Hill, S., Edge, T. A., Bekal, S., Fairbrother, J. M., Harel, J., Maynard, C., Masson, L., Brousseau, R., 2006. A virulence and antimicrobial resistance DNA microarray detects a high frequency of virulence genes in Escherichia coli isolates from Great Lakes recreational waters. Applied and Environmental Microbiology 72, 4200-4206. Hamelin, K., Bruant, G., El-Shaarawi, A., Hill, S., Edge, T.A., Fairbrother, J., Harel, J., Maynard, C., Masson, L., Brousseau, R., 2007. Occurrence of virulence and antimicrobial resistance genes in Escherichia coli isolates from different aquatic ecosystems within the St. Clair River and Detroit River areas. Applied and Environmental Microbiology 73, 477-484. Harwood, V.J., Levine, A.D., Scott, T.M., Chivukula, V., Lukasik, J., Farrah, S.R., Rose, J.B., 2005. Validity of the indicator organism paradigm for pathogen reduction in reclaimed water and public health protection. Applied and Environmental Microbiology 71(6), 3163–3170. Hirsch, R., Ternes, T., Haberer, K., Kratz, K.L., 1999. Occurrence of antibiotics in the aquatic environment. Science of the Total Environment 225, 109-118. Huffman, D.E., Quinter-Betancourt, W., Rose, J., 2003. Emerging waterborne pathogens. In: Mara, D., Horan, N. (Eds.) The Handbook of Water and Wastewater Microbiology. London, Elsevier, 193-207. Hunt, A.P., Parry, J.D., 1998. The effect of substratum roughness and river flow rate on the development of a freshwater biofilm community. Biofouling 12 (4), 287-303. Hunter, C., Perkins, J., Tranter, J., Gunn, J., 1999. Agricultural land-use effects on the indicator bacterial quality of an upland stream in the Derbyshire Peak district in the United Kingdom. Water Research 33 (17), 3577-3586. Hutchison, M., Walters, L.D., Moore, T., Crookes, K.M., Avery, S.M., 2004. Effect of length of time before incorporation on survival of pathogenic bacteria present in livestock wastes applied to agricultural soil. Applied and Environmental Microbiology 70, 5111–5118. Huws, A.J., McBain, P., Gilbert, P., 2005. Protozoan grazing and its impact upon population dynamics in biofilm communities. Journal of Applied Microbiology 98 (1), 238–244. Jackson, C.R., 2003. Changes in community properties during microbial succession. Oikos 101 (2), 444-448.  23  Karsten, U., Kühl, M., 1996. Die Mikrobenmatte—das kleinste Ökosystem der Welt. Biologie Unserer Zeit 26(1), 16–26. Keevil, C.W., 1999. Pathogens and metabolites associated with biofilms. In: Keevil, C.W. Godfree, A., Holt, D., Dow, C. (Eds.) Biofilms in the Aquatic Environment. Royal Society of Chemistry, Cambridge, 145-152. Keevil, C.W., 2003. Rapid detection of biofilms and adherent pathogens using scanning confocal laser microscopy and episcopic differential interference contrast microscopy. Water, Science and Technology 47(5), 105-116. Korber, D.R., Lawrence, J.R., Lappin-Scott, H.M., Costerton, J.W., 1995. Growth of microorganisms on surfaces. In: Lapin-Scott, H.M., Costerton, W.J. (Eds.) Microbial Biofilms. Cambridge, University Press, 15-45. Kroepfl, K., Vladar, P., Szabo, K., Acs, E., Borsodi, A.K., Szikora, S., Caroli, S., Zaray, G., 2006. Chemical and biological characterization of biofilms formed on different substrata in Tisza river (Hungary). Environmental Pollution 144 (2), 626-31. Kümmerer, K., 2004. Resistance in the environment. Journal of Antimicrobial Chemotherapy 54, 311-320. LeJeune, J.T., Besser, T.E., Hancock, D.D., 2001. Cattle water troughs as reservoirs of Escherichia coli O157. Applied and Environmental Microbiology 67, 3053-3057. Lemarchand, K., Lebaron, P., 2003. Occurrence of Salmonella spp. and Cryptosporidium spp. in a French coastal watershed: relationship with fecal indicators. FEMS Microbiological Letters 218 (1), 203-209. Lewandowski, Z., Beyenal, H., 2003. Biofilm monitoring: a perfect solution in search of a problem. Water, Science and Technology 47 (5), 9-18. Lock, M.A., 1994. Dynamics of particulates and dissolved organic matter. In: Wotton, R.S. (Ed.) The Biology of Particles in Aquatic Systems, 2nd edition. Boca Raton, Lewis Publishers, 137-260. Matz, C., Bergfeld, T., Rice, S.A., Kjelleberg, S., 2004. Microcolonies, quorum sensing and cytotoxicity determine the survival of Pseudomonas aeruginosa biofilms exposed to protozoan grazing. Environmental Microbiology 6, 218–226. Merritt, P.M., Danhorn, T., Fuqua, C., 2007. Motility and chemotaxis in Agrobacterium tumefaciens surface attachment and biofilm formation. Journal of Bacteriology 189 (22), 8005–8014. Moore, G.F., Dunsmore, B.C., Jones, S.M., Smejkal, C.W., Jass, J., Stoodley, P., Lappin-Scott, H.M., 2000. Microbial detachment from biofilms. In: Allison, D.G., Gilbert, P.,  24  Lappin-Scott, H.M., Wilson, M. (Eds.) Community Structure and Cooperation in Biofilms. Cambridge University Press, 107-127. Morita, R.Y., 1982. Starvation survival of heterotrophs in marine environments. Advanced Microbial Ecology 6, 171-198. Morin, P., Gauthier, V., Saby, S., Block, J.C., 1999. Bacterial resistance to chlorine through attachment to particles and pipe surfaces in drinking water distribution systems. In: Keevil, C.W., Godfree, A., Holt, D., Dow, C. (Eds.) Biofilms in the Aquatic Environment. Royal Society of Chemistry, Cambridge, 171-190. Muirhead, R.W., Davies-Colley, R.J., Donnison, A.M., Nagels, J.W., 2004. Faecal bacteria yields in artificial flood events: quantifying in-stream stores. Water Research 38, 1215-1224. National Research Council, 2004. Indicators for waterborne pathogens. National Academies Press, Washington DC. Newman, R.F., Hooper, T.D., Powell, G.W., Njenga, F.M., 2003. The influence of range practices on waterborne disease organisms in surface water of British Columbia: a problem analysis. Resources Branch, B.C. Ministry of Forestry, Victoria, B.C. Tech. Rep. 008. Retrieved January 29th, 2008. Available at: http://www.for.gov.bc.ca/hfd/pubs/Docs/Tr/Tr008.htm Obiri-Danso, K., Jones. K., 2000. Intertidal sediments as reservoirs for hippurate negative Campylobacters, Salmonellae and faecal indicators in three EU recognised bathing waters in North West England. Water Research 34(2), 519- 527. Ohashi, A., Harada, H., 1996. A novel concept for evaluation of biofilm adhesion strength by applying tensile force and shear force. Water, Science and Technology 34 (5-6), 201211. Payment, P., 2003. Health effects of water consumption and water quality. In: Mara, D., Horan, N. (Eds.) The Handbook of Water and Wastewater Microbiology. London, Elsevier, 209-219. Pell, A. N., 1997. Manure and microbes: public and animal health problem? Journal of Dairy Science 80, 2673-2681. Percival, S.L., Walker, J.T., Hunter, P.R., 2000. Microbiological Aspects of Biofilms and Drinking Water. CRC Press, Boca Raton. Reeser, R.J., Medler, R.T., Billington, S.J., Jost, B.H., Joens, L.A., 2007. Characterization of Campylobacter jejuni biofilms under defined growth conditions. Applied and Environmental Microbiology 73, 1908-1913.  25  Richardson, M.L, Bowron, J.M., 1985. The fate of pharmaceutical chemicals in the aquatic environment. Journal of Pharmacy and Pharmacology 37, 1–12. Rickard, A.H., Gilbert, P., High, P.G., Kolenbrander, P.E., Handley, P.S., 2003. Bacterial coaggregation: an integral process in the development of multi-species biofilms. Trends in Microbiology 11 (2), 94-100. Rittmann, B.E., 1989. Detachment from biofilms. In: Characklis, W.G., Wilderer, P.A. (Eds.) Structure and Function in Biofilms. New York, Wiley, 49-58. Rittmann, B.E., Schwarz, A.O., Saez, P.B., 2000. Biofilms applied to hazardous waste treatment. In: Bryers, J.D. (Ed.) Biofilms II: Process Analysis and Applications. New York WileyLiss, 207-234. Rittmann, B.E., 2004. Biofilms in the water industry. In: Ghannoum, M., O’Toole, G.A. (Eds.) Microbial Biofilms. ASM Press, Washington, DC, 359-78. Romani, A.M., Sabater, S., 2000. Variability of heterotrophic activity in Mediterranean stream biofilms: A multivariate analysis of physical-chemical and biological factors. Aquatic Sciences 62 (3), 205-215. Rood, K.M., Hamilton, R.E., 1995. Hydrology and water use for salmon streams in the Chilliwack/Lower Fraser habitat management area, British Columbia. Prepared for the Fraser River Action Plan by Northwest Hydraulic Consultants Ltd., North Vancouver, BC. Rooklidge, S.J., 2004. Environmental antimicrobial contamination from terraccumulation and diffuse pollution pathways. Science of the Total Environment 325, 1-13. Rosen, B.H., 2000. Waterborne Pathogens in Agricultural Watersheds. United States Department of Agriculture, Burlington. Sabater, S., Guasch, H., Romani, A., Munoz, I., 2002. The effect of biological factors on the efficiency of river biofilms in improving water quality. Hydrobiologia 469, 149-156. Sabater, S., Guasch, H., Ricart, M., Romaní, A., Vidal, G., Klünder, C., Schmitt-Jansen, M., 2007. Monitoring the effect of chemicals on biological communities: the biofilm as an interface. Analytical and Bioanalytical Chemistry 387 (4), 1425-34. Savichtcheva, O., 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. Sawyer, L.K., Hermanowicz, S.W., 2000. Detachment of Aeromonas hydrophila and Aeromonas aeruginosa due to variations in nutrient supply. Water, Science and Technology 41 (5), 139-145.  26  Schendel, E.K., Schreier, H., Lavkulich, L.M., 2004. Linkages between phosphorus index estimates and environmental quality variables. Soil and Water Conservation Journal 59, 243-251. Schmidt, J. E., Ahring, B. K., 1994. Extracellular polymers in granular sludge from different upflow anaerobic sludge blanket (UASB) reactor. Applied Microbiology and Biotechnology 42, 457–462 Schorer, M., Eisele, M., 1997. Biofilms in the aquatic environment. Water, Air and Soil Pollution 99, 651-659. Sayah, R.S., Kaneene, J.B., Johnson, Y., Miller, R.A., 2005. Patterns of antimicrobial resistance observed in Escherichia coli isolates obtained from domestic- and wild-animal fecal samples, human sewage, and surface water. Applied and Environmental Microbiology 71, 1394–1404. Schwartz, T., Kohnen, T., Jansen, B., Obst, U., 2003. Detection of antibiotic-resistant bacteria and their resistance genes in wastewater, surface water, and drinking water biofilms. FEMS Microbiology Ecology 43, 325–35 Solomon, E.B., Yaron, S., Matthews, K.R., 2002. Transmission of Escherichia coli O157:H7 from contaminated manure and irrigation water to lettuce plant tissue and its subsequent internalization. Applied and Environmental Microbiology 68 (1), 397-400. Sommer, P., Martin-Rouas, C., Mettler, E., 1999. Influence of the adherent population level on biofilm population, structure and resistance to chlorination. Food Microbiology 16, 503-515. Stanley, K.N., Wallace, J.S., Currie, J.E., Diggle, P.J., Jones, K., 1998. The seasonal variation of thermophilic Campylobacter in beef, dairy, cattle, and calves. Journal of Applied Microbiology 85, 472-480. Stoodley, P., Sauer, K., Davies, D.G., Costerton, J.W. 2002. Biofilms as complex differentiated communities. Annual Review of Microbiology 56, 187-209. Sutherland, I.W., 2001. Exopolysaccharides in biofilms, flocs and related structures. Water, Science and Technology 43 (6), 77-86. Swain, L.G., Holms, G.B., 1985. Ambient water quality objectives for the Fraser River Sub-basin from Hop to Kanaka Creek. BC Ministry of Environment: Waste Management Branch. Retrieved January 29th, 2008. Available at: http://www.env.gov.bc.ca/wat/wq/objectives/fraserhope/fraserhope.html Tank, J.L., Dodds, W.K., 2003. Nutrient limitation of epilithic and epixylic biofilms in ten North American streams. Freshwater Biology 48, 1031-49.  27  Thomas, C., Hill, D.J., Mabey, M., 1999. Evaluation of the effect of temperature and nutrients on the survival of Campylobacter spp. in water microcosms. Journal of Applied Microbiology 86 (6), 1024–1032. Trevisan, D., Vansteelant, J.Y., Dorioz, J.M., 2002. Survival and leaching of fecal bacteria after slurry spreading on mountain hay meadows: consequences for the management of water contamination risk. Water Research 36, 275-283. Tyrrel, S.F., Quinton, J.N., 2003. Overland flow transport of pathogens from agricultural land receiving faecal waste. Journal of Applied Microbiology 94, 87S-93S. van Loosdrecht, M.G.M., Lyklema, J., Norde, W., Zehnder, A.J.B., 1990. Influence of interface on microbial activity. Microbiological Reviews 54, 75-87. Vingarzan, R., Belzer, W., Thompson, B., 2002. Nutrient levels in the atmosphere of the Elk Creek Watershed, Chilliwack, BC (1999-2000). Environment Canada technical report. Aquatic and atmospheric sciences division, Vancouver, Canada. Walker, J.T., Mackerness, C.W., Rogers, J., Keevil, C.W., 1995. Heterogeneous mosaic biofilm – A haven for waterborne pathogens. In: Lapin-Scott, H.M., Costerton, W.J. (Eds.) Microbial Biofilms. University Press, Cambridge, 196-206. Warrington, P.D., Johnson, T.R., 2001. Water quality criteria for microbiological indicators: an overview report. Prepared for the British Columbia Ministry of Environment (Water, Air and Climate Change Branch). Document retrieved on January 6, 2007. Retrieved January 29th, 2008. Available at: http://www.env.gov.bc.ca/wat/wq/BCguidelines/microbiology/microbiology.html Watkinson, A.J., Micalizzi, G.B., Graham, G.M., Bates, J.B., Costanzo, S.D., 2007. Antibioticresistant Escherichia coli in wastewaters, surface waters, and oysters from an urban riverine system. Applied and Environmental Microbiology 73 (17), 5667-5670. Wimpenny, J.W.T., 2000. An overview of biofilms. In: Allison, D.G., Gilbert, P., Lappin-Scott, H.M., Wilson, M. (Eds.) Community Structure and Cooperation in Biofilms. Cambridge, University Press. Wimpenny, J.W.T., Peters, A., Scourfield, M., 1989. Modeling spatial gradients. In: Characklis, W.G., Wilderer, P.A. (Eds.) Structure and Function of Biofilms. Chichester, John Wiley, 111-127. Wimpenny, J.W.T., Colasanti, R., 1997. A unifying hypothesis for the structure of microbial biofilms based on cellular automaton models. FEMS Microbiology Ecology 22 (1), 1–16.  28  Winfield, M.D., Groisman, E.A., 2003. Role of non-host environments in the lifestyles of Salmonella and Escherichia coli. Applied and Environmental Microbiology 69, 36873694. Witte, W., 2000. Ecological impact of antibiotic use in animals on different complex microflora: environment. International Journal of Antimicrobial Agents 14, 321-325. Wolfaardt, G.M., Korber, D.R., Karthikeyan, S., Caldwell, D.E., 2000. Biodegradation by biofilm communities. In: Allison, D.G., Gilbert, P., Lappin-Scott, H.M., Wilson, M. (Eds.) Community Structure and Cooperation in Biofilms. Cambridge University Press, 278294. Yates, M.V., 2007. Classical indicators in the 21st century-far and beyond the coliform. Water Environment Research 79 (3), 279-86. Zimmer, M., Barnhart, H., Idris, U., Lee, M.D., 2003. Detection of Campylobacter jejuni strains in the water lines of a commercial broiler house and their relationship to the strains that colonized the chickens. Avian Disease 47, 101-107. Zobell, C.E., 1943. The effect of solid surfaces upon bacterial activity. Journal of Bacteriology 46, 39-56.  29  2  2.1  Seasonal and spatial distribution of indicators and pathogens in biofilms in agricultural watersheds: is monitoring water column grab samples sufficient?1  Introduction  Water monitoring is an essential component in the delivery of safe drinking water to the public. Several outbreaks in Canada (such as the incidences in Walkerton, North Battleford and the Kashechewan First Nation Reserve in Ontario) have put the efficacy of Canadian water monitoring and treatment methods into question (Holme, 2003; Stirling et al., 2001). Obviously, community waterborne disease outbreaks are not a new issue to the water delivery community, but these events have forced the issue of safe water into the public eye and warranted many studies to investigate the weaknesses of our water delivery systems (Holme, 2003; Hrudey and Hrudey, 2007). Two issues often raised in these studies are the failure of current indicator bacteria systems and the non-representative microbial nature of water column grab samples compared to other surfaces in the aquatic environment (such as sediment) when monitoring a water source. Indicator organisms have been an inexpensive and quick way to monitor microbial water quality. This seemed like a reasonable practice since most pathogens are faecally derived, low in concentration in aquatic environments and expensive to test for (Bartram and Howard, 2003). In the United States, the use of coliforms to ascertain water safety dates back to 1914 (Yates, 2007). Since the 1970s, many studies have pointed out the lack of correlation between faecal contamination indicator bacteria and pathogen concentrations (Glasner and McKee, 2002; Bartram and Howard, 2003; Yates, 2007). The other problem, non-representative sample collection, can easily be explained through the concept of biofilms. In aquatic environments, the majority of bacteria and other microbes attach to surfaces instead of remaining suspended in the water column (Decho, 2000). Thus water column grab sampling only provides partial understanding of the aquatic microbial population that water consumers may be exposed to.  1  A version of this chapter has been published. Maal-Bared, R., Bartlett, K.H., Bowie, W.R., Hall, E.R., Hall, K.J. 2007. Seasonal and spatial distribution of indicators and pathogens in biofilms in an agricultural watershed. Annual Conference Proceedings of the American Waterworks Association, Toronto 2007. Copyright©2007, American Water Works Association.  30  As far as we know, there are no studies investigating the presence and concentrations of pathogens, as well as their commonly used indicator organisms, in biofilms on different surface types in natural aquatic environments. A study by Muirhead et al. (2004) compared the contribution sediment made to E. coli loading within a creek in the dry season, to that made by river rock biofilms. The study did not find that rock biofilms made a significant contribution to in-stream E. coli concentrations. A study by Rahim et al. (1985) investigated the potential use of invertebrates as passive samplers. They compared the concentrations of faecal coliforms in water, sediment and in bivalves (Lamellidens marginalis), and found that faecal coliform concentrations in the bivalves were higher than concentrations in water and sediment. The role of biofilms as early organic and inorganic chemical monitoring systems has become widely accepted (Sabater et al., 2007), but using biofilms for microbiological monitoring in raw water sources has not been considered to our knowledge. Biofilms reveal information regarding previous pollution events (historical data). They could also be used as predictors of future water quality, since attached microbes resuspend into the water column following environmental changes. The detachment of pathogens from biofilms into the water column to contribute to pathogen loading has not been shown in situ in natural aquatic environments, but it has previously been shown in water distribution systems (Stoodley et al., 2001). The purpose of the present study was to increase our understanding of indicator organism concentrations and pathogens in biofilms in potential drinking water sources in an effort to improve current monitoring protocols. To do so, a variety of natural and artificial substrata were investigated to determine each substratum’s ability to accumulate indicator bacteria and pathogens compared to a water column grab sample collected at the same site. The chosen substrata were: river rock, slate rock, wood, Lexan®, sandpaper, and sediment. Some of these surfaces were used due to their natural presence in the environment (river rock, wood, and sediment), their hydrophobicity (Lexan®), or their large, smooth and granular surface area (slate rock and fine grit sandpaper). The successful use of sandpaper coupons for biofilm monitoring has been demonstrated by Hunt and Parry (1998), while the enhanced attachment of bacteria to hydrophobic surfaces has been demonstrated by Cerca et al. (2005). Our research questions were the following. Is there any difference between the distribution of indicators and pathogens in water and the distribution of indicators and pathogens in biofilms in an agriculturally impacted stream?  31  Are there any differences in the concentration of heterotrophic plate counts (HPC), faecal coliforms (FC), E. coli (EC) and enterococci (ENT) on river rock, slate rock, wood, Lexan, sandpaper, and sediment in an agriculturally impacted stream? Are there any differences in pathogen presence (E. coli O157, Campylobacter sp. and Salmonella sp.) amongst the different surface types and water in an agriculturally impacted stream? How does seasonality (rainfall in particular) affect the distribution of indicator bacteria and pathogens on different substratum surface types? Our hypothesis was that due to the higher bacterial densities found in biofilms as opposed to water, and due to the increased availability of nutrients and the protection a biofilm confers, the concentrations of indicator bacteria and pathogens in different biofilms will be different from those in water. Also, different surfaces will exhibit varying amounts of surface tension, hydrophobicity, porosity, and/or surface area, which will all affect biofilm thickness, chemical characteristics, strength of biofilm attachment and sloughing rate. Finally, different seasons will result in different ambient temperatures, rain patterns, agricultural activities, and water disturbances that might affect biofilm detachment rates and faecal contamination input concentrations. Therefore, different seasons will exhibit differences in indicator bacteria and pathogen distributions on different surfaces.  2.2  Methods  2.2.1 Sampling location- Elk Creek The Elk Creek Watershed is located in the Fraser Valley to the east of Chilliwack, British Columbia, and encompasses about 28 km2 of agricultural urban and mountainous terrain. Its headwaters are located in Elk Mountain, Mount Thurston, and the Eastern Hillsides. Its watercourse flows north and drains into the Fraser River at Hope Slough (see Figure 2.1). The watercourse is about 12 km long (Rood and Hamilton, 1995). Peak runoff occurs in late spring during snowmelt, and the flow becomes very low during late summer and early autumn. The estimated mean annual flow of this watershed is 0.61 m3/s (Rood and Hamilton, 1995). The dry season typically begins in April and ends in September, while the wet season usually begins in October and ends in March.  32  Hope Slough  Site 4 Site 3  Nevin Creek  Site 2 Dunville Big Ditch  Creek  Ford Creek  Elk Creek  Site 1  North  Figure 2.1 Map of Elk Creek in Chilliwack, British Columbia, and location of sites 1 to 4 (marked by the triangles) (Scale about 1: 90,000, 16.7 x 11.7 km). Arrow indicates north.  The Elk Creek Watershed is mainly agricultural in nature and it encompasses a variety of greenhouses and improved pasturelands. Among the main agricultural activities in the watershed are the productions of dairy and field crops. The density of livestock is estimated at 2.46 animal units per hectare. Manure is applied to 58% of the total agricultural land area, while commercial fertilizers, herbicides and insecticides are applied to 58%, 29%, and 12% of the land, respectively (Vingarzan et al, 2002). The water in Elk Creek was originally designated for drinking purposes, protection of aquatic life, wildlife, livestock, irrigation, and recreation (Swain and Holms, 1985). After a Cryptosporidium parvum outbreak in April 1998, the water was no longer used for drinking purposes (Newman et al., 2003). Elk Creek exhibits an increasing pollution gradient as the stream flows towards Hope Slough (Schendel et al, 2004; Derksen et al., 2004). Four sites were located on the stream for this study. The first site (Latitude 49ο 08’ 26.5” N, Longitude 121ο 50’ 0.34” W) was located in the headwaters as a reference site. Sites 2 (Latitude 49ο 09’ 42.4” N, Longitude 121ο 50’ 52.1” W), 3 (Latitude 49ο 09’ 51.5” N, Longitude 121ο 51’ 7.8” W) and 4 (Latitude 49ο 10’ 36.3” N, Longitude 121ο 51’ 7.0” W) were located in the agricultural area. For photographs of the watershed and sites, see Appendix A.  33  Work by Derksen et al. (2004) has previously divided the Elk Creek watershed into four areas that are impacted by land use practices: the upper Elk, mid Elk, lower Elk and the mouth of Elk Creek. Derksen et al. (2004) also used GIS to describe the major land use impacts that affect those four regions of the watershed. At site 1, the majority of the land use is forestry-related (1200 ha of forest). In the mid Elk Creek Watershed, the majority of land is used for grass production (approximately 90 ha). Grass is either used for harvesting or as pastureland. Also, about 10 ha are dedicated to corn production and less than 5 ha are used for recreational purposes. In the lower Elk Creek Watershed, about 275 ha are used for grass production, 125 ha for corn production, 15 ha for fruit production and less than 5 ha for horticulture. Finally, at the Elk Creek mouth about 70 ha are used for horticulture, 25 ha for grass production, and 10 ha for corn production. While we can say for certain that site 1 is located in the upper Elk, and site 4 at the mouth of Elk Creek, sites 2 and 3 are probably impacted by the land uses that affect both the mid and lower Elk Creek Watershed.  2.2.2 Biofilm samplers Materials for use as biofilm substrata were purchased from a local hardware store and a landscaping retailer. The slate rocks were already cut and thus all other substrata (Lexan, sandpaper (fine grit # 500), and alder wood) were cut to a similar size (10 cm x 10 cm). The thickness of a substratum was not always adjustable. The river rocks chosen were similarly sized and shaped. Samples of substrata were anchored to 1.52 m (5 foot) long galvanized steel bars. Autoclaved substrata were attached in duplicate to the steel bar using three plastic ties, which are also called cable ties or zap straps (Cable Ties, Marr®, 6/6 Nylon, Thomas & Betts, Memphis, TN), per substratum. These “zap strap cages” were built so that very minimal contact occurred between the substrata, the straps and the steel bar. Finally, one bar at a time was placed at each site on the surface of the sediment and attached to a fixed object (bridge, tree, etc.) using a thin metal wire to prevent sampler loss. The samplers were placed so that the lengths of the galvanized steel bars were parallel to stream flow, perpendicular to stream bank width, and were not hanging in the water column. Figure 2.2 illustrates the structure of a biofilm sampler used in this study. Also, photographs of the samplers can be seen in Appendix B.  34  Plastic ties attach each substratum and are passed through the holes of the steel bar. Three plastic ties are used to attach each substratum.  Substrata (river rocks, slate rock, wood, Lexan, and sandpaper) are randomly distributed on the steel bar.  Galvanized steel bar with holes (length = 1.52 m) Figure 2.2 Biofilm samplers used to accumulate biofilms on a variety of substratum surface types.  2.2.3 Sample collection Samples were collected between December 2005 and December 2006, and only one site at a time was visited. In the wet season (October to April), 11 samplers were retrieved, while only six samplers were left in place and retrieved in the dry season (May- September). Every sampler was left at a site for four weeks before collection and analysis, since previous investigations have shown that a mature biofilm community requires somewhere between 2 weeks and one month for formation (Hunt and Parry, 1998). At the time of sampler collection, duplicate water column grab samples of river water were collected into sterilized 500 mL bottles, as well as duplicate 100 g surface sediment samples from the same site. The surface sediment grab samples contained the top 10-15 cm of the sediment and were collected using a standard Teflon pot that was attached to an extendable pole. The samples were kept on ice packs in a cooler for transport to the laboratory. Other chemical and physical water characteristics were collected and analyzed, but are reported elsewhere (Maal-Bared, 2008). In the laboratory, to remove biofilms from substrata, substrata were placed in 500 ml of sterile PBS (PBS, pH=7.2, 9 g NaCl/L, 0.0067M PO4) in a sterile beaker, which was placed into a sonication bath for 30 minutes. The bath was continuously cooled using ice packs. Sonication conditions were tested and determined during previous work to minimize sample loss. The PBS-biofilm suspension was used for analysis. Fifty grams of sediment samples were blended in 450 mL of sterile PBS for 3 minutes at low power in an autoclaved stainless steel blender, and then transferred to a sterile bottle. Prior to analysis, the  35  sediment suspension was shaken for 20 seconds to allow for resuspension and homogenization. Water was analyzed directly.  2.2.4 Microbial analyses 2.2.4.1 Bacterial indicator organisms and E. coli O157 analysis Standard Method procedures were used for the enumeration and isolation of indicator organisms (APHA et al., 1998). Heterotrophic plate counts (HPCs) were conducted using plate count agar (Difco™ BD Microbiology Systems, Sparks, MD). Spread plates were plated in triplicate using four ten-fold dilutions of the original sample or suspension and incubated at 35oC for 48 hours. Most probable numbers (MPNs) for faecal coliforms (FCs) were completed using A1 broth (Difco™ BD Microbiology Systems, Sparks, MD). E. coli (EC) MPNs were completed using EC broth (Oxoid™ Basingstoke, Hampshire, UK). Both FC and EC MPNs were incubated at 44.5oC and checked for gas production at 24 hours and 48 hours. Standard tables were used for the determination of MPNs. Positive test tubes for both FC and EC MPNs were confirmed by selective isolation onto Levine’s EMB plates (BBL™ BD Microbiological Systems, Sparks, MD). Colonies from EMB plates that were plated using a positive EC broth test tube were purified onto tryptic soy agar (Difco™ BD Microbiology Systems, Sparks, MD) and tested using the API® 20E kit (bioMerieux ™, Hazelwood, MO) to ascertain the identity of the isolated bacteria. Positive E. coli colonies were also tested using the pathogenic E. coli O157 Oxoid® Dryspot test kit (Oxoid™, Basingstoke, Hampshire, UK, DR0120). To complete the enterococci (ENT) MPN tests, azide dextrose broth (Difco™ BD Microbiology Systems, Sparks, MD) was used and test tubes were incubated at 35oC for 48 hours and then checked for turbidity. The positive test tubes were confirmed on bile esculin plates (Oxoid™ Basingstoke, Hampshire, UK). Growth from the bile esculin plates was tested again in a brain heart infusion broth (BBL™ BD Microbiological Systems, Sparks, MD) with 6.5% NaCl, which was incubated at 42oC for 24 hours and checked for turbidity.  2.2.4.2 Salmonella spp. analysis To identify Salmonella sp., 40 mL of PBS suspension was tempered in peptone broth (Oxoid™, Basingstoke, Hampshire, UK) for 4-5 hours at 37oC (1:9 ratio). Then a 1 mL sample of the tempered medium was added to tetrathionate enrichment broth (Difco™ BD Microbiology Systems, Sparks, MD) for 8 hours or overnight at 37oC. The enriched suspension was plated onto  36  XLD plates (Difco™ BD Microbiology Systems, Sparks, MD) and/or BGA plates (Oxoid™ Basingstoke, Hampshire, UK) in triplicate depending on the type of competing flora on the plate. Finally, the grown colonies were tested using the Oxoid® Salmonella Latex test (Oxoid™ Basingstoke, Hampshire, UK, FT0203) to confirm Salmonella sp. isolation.  2.2.4.3 Campylobacter spp. analysis To determine presence of Campylobacter sp., 40 mL of PBS suspension was tempered in peptone water for 4-5 hours at 37oC (1:9 ratio). Then a 10 mL sample of the tempered medium was added to an enrichment broth for 8 hours or overnight at 37oC (1:1 enrichment ratio). The enrichment broth was made up of 475 mL of Brucella broth (BBL™ BD Microbiological Systems, Sparks, MD) with one vial of Preston Campylobacter Supplement (Oxoid™, Basingstoke, Hampshire, UK, SR0117) and 25 mL of defibrinated horse blood (PML Microbiologicals, Wilsonville, OR). Finally, the enriched suspension was plated onto Campylobacter isolation plates (475 ml Campylobacter base agar (Oxoid™ Basingstoke, Hampshire, UK) with 25 mL defibrinated horse blood and one vial of Preston Supplement). The plates were incubated at 42.5oC for 48 hours in microaerophilic conditions using a BBL™ Campy Pak without the palladium catalyst (BBL™ BD Microbiology Systems, Sparks, MD). Colonies conforming to Campylobacter morphology were tested using the Oxoid™ DrySpot test (Dryspot Campylobacter Test Kit DR0150), which identifies five species of the Campylobacter genus (C. jejuni, C. lari, C. coli, C. fetus, and C. upsaliensis). It does not separate the species from each other.  2.2.4.4 Controls Positive and negative controls were used. Sterile distilled water was used as a negative control. ATCC control cultures E. coli (ATCC 25922), Campylobacter jejuni (ATCC 29428), Salmonella enterica (ATCC 29058), and Enterococcus faecalis (ATCC 19433) were used as positive controls.  2.2.5 Standardizing across different substrata To be able to compare the concentration of indicators in biofilms on different surface types, ashfree dry weight per mL of biofilm suspension extracted from each substratum was approximated using a “Loss on Ignition” (LOI) method. To do so, we found the dry weight for a fixed volume (60 mL) for each suspension, as well as water, by filtering that volume through a 4.7 cm ash-free filter (Gelman® A/E filter 61631). The filter was placed into a preweighed ceramic crucible and  37  both were placed into a muffle furnace at 100oC until a constant weight (the dry weight) was reached. The crucible and filter were then weighed again and placed back in the muffle furnace at 500oC for one hour for incineration to occur. The weight of the inorganic remains (the ash) was then determined at room temperature. Finally, an estimate of the organic matter that was vaporized (the AFDW) was determined by subtracting the post-incineration weight from the dry weight. Since the aim of the LOI method was to be able to compare indicator organism concentrations across different substrata and since the same volume of PBS was used to make each suspension, the percentage of PBS salt contribution to the AFDW was not estimated in the calculation.  2.2.6 Statistical Analyses All statistical analyses were completed using JMP 7.0 (SAS Institute, NC, USA). No statistical comparison was used to compare the numbers of faecal indicators and prevalence of pathogens at different sites due to the use of different units for analysis. However, non-parametric tests were used to examine differences amongst indicator numbers and pathogens frequencies on different substrata because variances were not equal and the distribution could not be transformed to fit a normal Gaussian distribution. Different statistical tests were used to determine statistical significance in the case of indicators and pathogens due to the differences in measurement types. In the case of indicator organisms, the tests used allowed the estimation of a number (CFU/mg AFDW or MPN/mg AFDW). The tests used to determine pathogen frequencies were presence or absence tests and the final results were presented as percentages of samples that tested positive for a particular pathogen. To determine whether there were statistically significant differences in indicator numbers between different substratum types, a Kruskal-Wallis analysis of variance test was used. To determine whether there were statistically significant differences in indicator numbers for any substratum types during different seasons, a Mann-Whitney test was used. To determine whether there were statistically significant differences in the presence of pathogens on different surface types, a Pearson Chi-square test was used. Only trends were reported to compare the frequencies of pathogens in the wet and the dry season, since the final 2x2 contingency table for each pathogen violated both Fisher Exact test assumptions and Chi-square test assumptions.  38  2.3 Results 2.3.1 Comparing indicator organism numbers and pathogen frequencies recovered from water column, sediment and biofilms samples at each site Figure 2.3 shows the mean number of indicator organisms and frequency of pathogen presence in all water samples collected for each site (five visits to every site). Figures 2.4 shows the mean number of indicator organisms and frequency of pathogen presence for all sediment samples (also five visits per site), while Figure 2.5 shows the mean number of indicator organisms and frequency of pathogen presence from biofilm samples of all substrata collected at each site (five substrata types retrieved from every site every each visit). The indicator organism numbers presented in Figures 2.4 and 2.5 were not adjusted by ash-free dry weight to facilitate comparison with the water column grab sample indicator concentrations. The pathogen presence frequency presented in the three figures is the total number of positive samples collected at each site per substratum. By comparing the three figures, it becomes clear that using water column grab samples often will underestimate the number of indicator organisms at some sites, and definitely underestimates pathogen presence. Heterotrophic plate counts were much lower in water samples than in sediment and biofilm samples at all sites. At site 1, biofilms had the highest numbers of HPCs and EC, while sediment had the highest numbers of FCs and ENT. Also, the highest number of pathogenic E. coli O157 was recovered from biofilms, while only sediment samples at site 1 contained Salmonella. At site 2, sediment samples exhibited higher indicator organism numbers than water and biofilms, but the frequency of pathogens recovered was highest in biofilms. Interestingly, at site 3, water carried the highest number of FCs, but the lowest number of EC. Finally, the lowest number of HPCs, FCs, ENT and no EC were isolated from water at site 4. With the exception of site 4, FC numbers in biofilms were lower than those in water or sediment. However, analyses to compare these differences were limited due to the use of different units to detect the numbers of indicators in water, sediment and biofilms.  39  Nukber of indicator organisms in 100 mL of water Frequency of pathogen presence in water samples  10000000 1000000 300000  100000 40000  Y  10000  Mean(HPC (CFU/100 mL))  4000  1000  Mean(FC (MPN/100 mL))  400  Mean(EC (MPN/100 mL))  100  Mean(ENT (MPN/100 mL))  30  10  Sum(Campylobacter +)  3  1  Sum(Salmonella +)  0.3  Sum(E.coli O157 +)  0.1 0.03  0.01  1  2  3  4  Site  Nukber of indicator organisms in 100 mL of sediment-PBS suspension (10 g sediment) Frequency of pathogen presence in sediment samples  Figure 2.3 Mean number of indicator organism per 100 mL of water and total number of samples positive for pathogens recovered from through water column grab samples at sites 1 to 4 (n=5, total number of processed samples for all sites= 40). The order of the legend from top to bottom corresponds with the bars from left to right.  10000000 3000000  1000000 500000 300000  Y  100000 50000  Mean(HPC CFU/100 mL)  20000  10000 5000 3000  Mean(FC MPN/100 mL)  1000  Mean(EC MPN/100 mL)  500  Mean(ENT MPN/100 mL)  200  100 50 30  Sum(Campylobacter +)  10  Sum(Salmonella +)  5 2  Sum(E.coli O157 +)  1 0.5 0.2  0.1 0.05 0.02  0.01  1  2  3  4  Site  Figure 2.4 Mean number of indicator organism per 100 mL of sediment-PBS suspension and total number of samples positive for pathogens recovered from sediment samples at sites 1 to 4 (n=5, total number of processed samples for all sites= 40). The order of the legend from top to bottom corresponds with the bars from left to right.  40  Nukber of indicator organisms in 100 mL of biofilm-PBS suspension Frequency of pathogen presence in artificial biofilm samplers  10000000 3000000  1000000 400000  100000  Y  40000  Mean(HPC (CFU/100 mL))  10000 4000  Mean(FC (MPN/100 mL))  1000  Mean(EC (MPN/100 mL))  400  Mean(ENT (MPN/100 mL))  100 40  Sum(Campylobacter +)  10  Sum(Salmonella +)  4  Sum(E.coli O157 +)  1 0.4  0.1 0.04  0.01  1  2  3  4  Site  Figure 2.5 Mean number of indicator organism per 100 mL of biofilm-PBS suspension and total number of samples positive for pathogens recovered from all tested artificial substrata (river rock, slate rock, wood, Lexan and sandpaper) at sites 1 to 4 (n=25, total number of processed samples for all sites= 200). The order of the legend from top to bottom corresponds with the bars from left to right.  Consistently, the highest number of pathogens was recovered from biofilms, as opposed to sediment or water. The use of grab samples here also masked the actual trends that indicator organisms and pathogens were exhibiting. For example, in Figure 2.3, the number of FCs actually decreases by the time the water reaches site 4 and no EC was found in the same water samples, while Figure 2.5 showed that the numbers of FCs and EC increased.  2.3.2 Indicator concentrations Tables 2.1 and 2.2 display the mean indicator organism numbers and mean ash-free dry weights for each substratum type analyzed, as well as water and sediment. Dry season heterotrophic plate counts per mg organic matter (HPC CFU/ mg AFDW) were higher than HPC numbers in the wet season for all substrata types expect for wood and sandpaper. HPCs remained fairly similar in both seasons for sediment, but AFDW was higher in the wet season. Faecal coliforms (FC MPN/ mg AFDW) decreased for all substrata types in the summer. The number of E. coli (EC MPN/mg AFDW) also decreased in the dry season for all substrata. Water appeared to have increased in number of ECs in Table 2.2, but this increase may be the result of the decrease in AFDW in the  41  summer. For enterococci (ENT MPN/mg AFDW), the values increased in the dry season for water, river rock and slate rock, but decreased for other substrata. ENT concentrations were higher than FC concentrations in water in both seasons, but only higher than FC concentrations in the wet season for sediment. In the dry season, ENT concentrations in all biofilms were higher than FC concentrations (equal in one case). During the wet season, FC concentrations were higher than ENT concentrations in all biofilms except wood. Tables 2.1 and 2.2 also show the results of the Kruskal-Wallis test in terms of χ2 statistics and p-values. Only the numbers of HPCs, ENTs and the AFDWs were determined to be statistically different amongst different substratum types in both seasons. No statistically significant differences were found in the numbers of indicator organisms found on a substratum between different seasons.  2.3.3 Pathogen presence When comparing the presence of pathogens on the different substratum types, the comparison becomes much easier than looking at indicators since the numbers shown in Table 2.3 are just percentages of substrata examined that were positive for the pathogen of interest. As can be seen from Table 2.3, the percentage of substrata positive for Salmonella sp. in the wet season did not decrease below 15%, with frequency in sediment reaching 41%. Even the 500 mL water samples were positive for Salmonella 21% of the time. Also, the frequency of Campylobacter recovery in the wet season ranged from 0% to 27%, being the highest in sediment. No Campylobacter was recovered from sandpaper biofilms in the wet, nor the dry season. Pathogenic E. coli O157, on the other hand, exhibited some interesting patterns in the wet season. It was found in water 4% of the time. The highest presence was on slate rock, followed by sediment and Lexan. There was no E. coli O157 at all found on wood or sandpaper in the wet season. The frequencies of Campylobacter, Salmonella and E. coli O157 presence on different substratum surface types in the wet season were statistically significant, resulting in p-values of less than 0.001.  42  Table 2.1 Descriptive statistics and Kruskal-Wallis test results for indicator organism numbers (heterotrophic plate counts (HPC CFU/mg), faecal coliforms (FC MPN/mg), E. coli (EC MPN/mg), and enterococci (ENT MPN/mg)) in biofilms on different substrata in the wet season (n=11, total number of analyzed samples for all substrata=132) Descriptive statistics are shown as means, standard error in means (SE), medians, and variances (σ). Kruskal-Wallis analysis of variance test results are shown in terms of χ2 statistics, and p-values, and statistically significant results are bolded. Wet Season  Substratum Type Water River Rock Slate Rock Wood Lexan Sand Paper Sediment 2  χ  p-value  43  HPC CFU/mg Mean (SE) Median 9.9 x104 (5.0 x104) 31 x103 13 x103 (4.5 x103) 9.2 x103 8.8 x103 (2.3 x103) 7.0 x103 91 x103 (77 x103) 9.0 x103 8.7 x103 (2.6 x103) 6.3 x103 21 x103 (12 x103) 7.2 x103 6.9 x102 (3.4 x102) 2.4x101  FC MPN/mg σ  Mean (SE) Median 3.2 x102 4 17 x10 (1.6 x102) 24 15 x103  71 (68)  1.0  7.8 x103  7.0 (5.0)  1.0  27 x104  5.0 (2.0)  3.0  9.1 x103  29 (24)  2.0  37 x103  52 (45)  2.0  1.1 x103  0.0 (0.0)  0.0  24.385 <0.001  EC MPN/mg ENT MPN/mg AFDW (mg/L) Mean Mean σ (SE) Median σ (SE) Median σ Mean (SE) Median 5.6 8.0 4.3 x102 x102 62 918 0.034 (0.020) 0.015 (6.0) 0.0 20 (2.7 x102) 2.3 x102 69 (68) 0.0 226 6.0(2.0) 3.0 6.0 0.61 (0.19) 0.34 1.0 16 (1.0) 0.0 3.0 11 (5.0) 7.0 19 1.6 (0.58) 0.69 3.0 7.0 (2.0) 1.0 7.0 13 (5.0) 5.0 16 1.6 (2.4) 0.77 2.0 85 (2.0) 0.0 6.0 22 (12) 9.0 41 0.46 (0.13) 0.30 1.4 x102 22 (15) 0.0 47 30 (13) 14 40 0.74 (0.27) 0.44 0.0 1.0 (0.0) 0.0 0.0 2.0 (1.0) 0.0 4.0 40 (5.2) 35 6.885 7.430 20.179 0.332 0.283 0.003  σ 0.069 0.66 2.0 2.4 0.46 0.85 17 51.159 <0.001  43  Table 2.2 Descriptive statistics and Kruskal-Wallis test results for indicator organism numbers (heterotrophic plate counts (HPC CFU/mg), faecal coliforms (FC MPN/mg), E. coli (EC MPN/mg), and enterococci (ENT MPN/mg)) in biofilms on different substrata in the dry season (n=6, total number of analyzed samples=72). Descriptive statistics are shown as means, standard error in means (SE), medians, and variances (σ). Kruskal-Wallis analysis of variance test results are shown in terms of χ2 statistics and p-values, and statistically significant results are bolded.  Dry Season  Substratum Type Water River Rock Slate Rock Wood Lexan Sand Paper Sediment  χ2 p-value  44  HPC CFU/mg Mean (SE) 2.9x105 (2.3 x105) 2.2 x104 (1.5 x104) 1.8 x104 (1.0 x104) 3.6 x104 (1.1 x104) 1.1 x104 (3.7 x103) 1.2 x104 (5.0 x103) 6.8 x102 (4.5 x102)  FC MPN/mg EC MPN/mg ENT MPN/mg AFDW (mg/L) Mean Mean Mean Median σ (SE) Median σ (SE) Median σ (SE) Median σ Mean (SE) Median σ 1.2 x103 2.3 x102 0.012 57 x103 5.7 x105 (98) 1.4 x102 239 13 (8.0) 0.0 20 (6.2x102) 8.2 x102 1.5 x103 (0.0021) 0.01 5.2 x103 1.0 7.4 x103 3.6 x104 3.0 (2.0) 1.0 4.0 (1.0) 0.0 1.0 12 (4.0) 10 10 0.43 (0.12) 0.43 0.29 1.0 0.0 3.0 10 (6.0) 5.0 15 0.47 (0.11) 0.39 0.27 10 x103 2.5 x104 10 (9.0) 1.0 21 (1.0) 2.0 31 x103 2.7 x104 1.0 (0.0) 1.0 1.0 (1.0) 0.0 3.0 9 (4.0) 5.0 9.0 1.1 (0.35) 1.1 0.85 0.0 14 x103 8.9 x103 5.0 (3.0) 1.0 8.0 (0.0) 0.0 0.0 12 (4.0) 12 10 0.27 (0.13) 0.18 0.31 2.0 8.4 x103 1.2 x104 3.0 (2.0) 2.0 4.0 (1.0) 2.0 2.0 14 (7.0) 5.0 16 0.68 (0.15) 0.86 0.38 0.0 2.5 x102 1.1 x103 0.0 0.0 0.0 (0.0) 0.0 0.0 0.0(0.0) 0.0 1.0 24 (2.3) 25 5.5 10.337 6.213 18.913 21.725 29.163 0.111 0.400 0.004 0.001 <0.001  44  In the dry season, Salmonella presence was highest on slate rock at 33%, followed by sandpaper at 27%. River rock, wood and sediment were similar at about 8%. Water and Lexan showed no signs of Salmonella in the dry season. Water also did not result in any positive E. coli O157 plates, nor did slate rock, Lexan, sandpaper, or sediment (see Table 2.3 and Figure 2.6). No Campylobacter was recovered from Elk Creek in the dry season. These differences in distribution of Salmonella and E. coli O157 on different surface types were statistically significant in the dry season. With the exception of slate rock and sandpaper, the percentage of positive substrata decreased in the dry season. For E. coli O157, the percentage of positive substrata decreased for all substrata except for river rock and wood.  Table 2.3 Percentage of samples positive for Campylobacter sp., Salmonella sp. and pathogenic E. coli O157 in biofilms, water and sediment on different surface types in the wet and the dry seasons. Table also presents number of samples tested in both seasons (nwet and nDry) and results of χ2 test. Substratum Water nwet=24, nDry=12 River Rock nwet=22, nDry=12 Slate Rock nwet=24, nDry=12 Wood nwet=24, nDry=12 Lexan nwet=24, nDry=12 Sandpaper nwet=20, nDry=11 Sediment nwet=22, nDry=12 χ2 value p-value  Wet Season Prevalence Campylobacter Salmonella E. coli 0157 8.3% 21% 4.2%  Dry Season Prevalence Campylobacter Salmonella E. coli O157 0% 0% 0%  9.1%  18%  4.6%  0%  8.3%  8.3%  22%  26%  17%  0%  33%  0%  13%  25%  0%  0%  8.3%  17%  13%  25%  13%  0%  0%  0%  0%  15%  0%  0%  27%  0%  27%  41%  14%  0%  8.3%  0%  36.3 <0.001  16.8 <0.01  49.0 <0.001  -  26.3 <0.001  36.3 <0.001  45  10 9 8 7 6  Wet  5 4 3 2 1 Season  Total number of samples that tested positive for every substratum  11  10 9 8 7 6 Dry  Total number of samples that tested positive for every substratum  0 11  5 4 3 2 1  Y Sediment  Sand Paper  Lexan  Wood  Slate Rock  River Rock  Water  0  Sum(Campylobacter +) Sum(Salmonella +) Sum(E.coli O157 +)  Substratum  Figure 2.6 Total number of samples that tested positive for Campylobacter (Campylobacter +), Salmonella (Salmonella +) and pathogenic E. coli O157 (E. coli O157 +) in biofilms on different substrata surface types, water and sediment in the wet season and the dry season. The order of the legend from top to bottom corresponds with the bars from left to right.  2.4  Discussion  Studies investigating biofilms occurring in natural aquatic environments have often found that substratum choice influences bacterial biomass (Hunt and Parry, 1998), species richness (Barbiero, 2000), species succession and colonization patterns (Tank and Dodds, 2003), and pollutant concentrations in the EPS (Kroepfl et al., 2006), which all ultimately could affect pathogen and indicator survival potential in a biofilm. The differences we observed in indicator  46  and pathogen distribution amongst different surface type biofilms could be related to a variety of factors. Different substrata tend to have different levels of hydrophobicity (Cerca et al., 2005), surface tension (Characklis and Cooksey, 1983), and surface roughness (Hunt and Parry, 1998), which all seem to affect biofilm microbial populations in different ways and to different extents. Surface characteristics affect the strength of EPS attachment to a surface and thus how fast colonization occurs, how thick a biofilm can become, and how fast detachment occurs. Different surface types result in different microbial succession patterns as well. For example, Tank and Dodds (2003) showed that rock biofilms (periphyton) tend to encourage higher rates of algal colonization, while wood results in more fungal colonization. This would explain the high ashfree dry weight found in wood biofilms, since ash-free dry weights do not distinguish algal organic material from other organic material sources, such as bacteria and fungi. The substrata used in this study were selected due to the differences in their hydrophobicity, surface areas and porosity. These characteristics were thought to affect biofilm colonization, and thus the numbers of indicator organisms and pathogens in those biofilms. While the characteristics of these different substrata were not measured, some speculation can be made about their nature. Lexan had the highest hydrophobicity, the lowest surface area and was not porous. Therefore, the accumulation of FCs and ENT on this substratum in the dry and wet season may be related to its high hydrophobicity. Sandpaper has high surface area due to its roughness, but is neither hydrophobic nor porous. Sandpaper performed better in the wet season than in the dry season, accumulating high numbers of FCs, EC and ENT in that season. This may be the result of the formation of more protected biofilm communities on rough surfaces. This may also explain why slate rock, a very smooth, hydrophobic surface, did not perform very well in the wet season but accumulated FCs and ENT in the dry season. Loss of biofilm due to sloughing in high flow conditions would be higher on a smooth surface. Wood and river rock have a lot of surface characteristics in common. Both have high porosity, surface area and hydrophobicity, and thus rank highest compared to other substrata when it comes to characteristics that allow for biofilm colonization. Both substrata performed well in accumulating HPCs and indicators. While river rock performs better under high flow conditions (it is more porous than wood), wood accumulated more HPCs in both seasons and more EC in the dry season, which could be due to the high organic content of the substratum. This becomes even more relevant in the dry season when nutrient influx into the stream is sparse. The analysis shows that while some substratum characteristics have been associated with enhanced biofilm colonization, fluctuations in flow and nutrients have to be taken into account to choose the most  47  suitable substratum for monitoring purposes in a natural system. To enhance the process of substratum selection, substratum characteristics need to be measured in a reliable manner. Standardizing indicator concentrations across different substrata using ash-free dry weight of the biofilm suspension was necessary considering that different substrata have different surface areas and surface characteristics and thus different abilities to accumulate material per cm surface area. For example, porosity can affect how much biofilm accumulates on a surface, as do cracks in rocks and the grit size of a sandpaper coupon. It is important to note that the method used in this study (using the same volumes of PBS suspensions) does not completely normalize the biofilm related measurements. This is mainly due to the difference in amounts of biofilm that can accumulate on different surfaces, which results in different biomass in each suspension despite the equal volumes. Heterotrophic plate counts were higher in the dry season than in the wet season for all substrata expect for wood and sandpaper. The increase in HPCs is very likely related to the increase in water temperature in the summer (Reynolds, 1999). The increase in HPCs may also be related to a dilution effect. When the creek flow and volume decrease in the dry season, the ratio of microorganism concentration to creek water volume increases, making the retrieval of microorganisms more likely. HPCs on sediment, wood and sandpaper did not decrease in the dry season, but that may be related to the amount of organic material available to the heterotrophic population through the substratum. Also, organic substrata, such as wood, are often colonized by fungi, which can contribute carbon to the biofilms matrix (Tank and Dodds, 2003). Studies have also shown that organic materials in sediment can be used as resources by attached bacteria (Davies et al., 1995). Sediment has been frequently found to carry higher indicator organism concentrations and faecal coliforms than water (Davis et al., 2005; Davies et al., 1995; Anderson et al., 2005; Crabill et al., 1999). This was explained as a result of bacterial association with settling particles in the stream, as well as increased availability of nutrients, and protection from sunlight and predation (Davis et al., 2005). Romani and Sabater (2000) state that sediment bacterial community efficiency is higher than that of rock biofilms due to its higher surface area and greater amount of organic matter associated with the particles. However, in the present study, one problem with using sediment as a monitoring substratum is the fact that sediment exhibited very high variability among its replicates. This was observed by other researchers as well (Anderson et al., 2005), and was probably the result of uneven microbial distribution on sediment, as well as difficulty of detaching attached microbial communities for analysis. Also, it  48  is important to note that the sediment particle colonization period was much longer than the colonization period for the other substrata tested, which may explain the high numbers of HPCs and indicators in sediment throughout the study period. Finally, it must be emphasized that sediment transport increases the difficulty of using sediment samples to assess the microbial quality of a particular site. This present study corresponds with Muirhead et al. (2004) in that the concentrations of indicator bacteria in sediment were higher than those in biofilms throughout the year. However, in the study, actual pathogen numbers associated with sediment were only higher than those in biofilms in the wet season. During the dry season, higher pathogen concentrations and more species were isolated from biofilms than from sediment. In the case of the across site comparison, the observed difference in numbers of organisms may have been the effect of comparing different units. One of the major challenges of this present study was to identify units that can be compared across the different substrata. Different authors have used different techniques to compare the numbers of faecal indicator organisms across different surfaces and media (water, sediment and biofilms). Muirhead et al. (2005) analyzed biofilms on surfaces with a known surface area and compared those to a blanket of sediment with the same surface area. This method obviously does not take into account the complete surface area of sediment particles, or the actual volume of biofilm analyzed. Other studies comparing the numbers of faecal indicator organisms used equal volumes of water and sediment. Sediments samples collected for those studies were collected by coring (collection of a cylinder with known dimensions) (Whitman and Nevers, 2003), or by collecting wet sediment in containers with known volumes (Crabill et al., 1999). However, many studies still compare the differences in organism numbers utilizing different units (An et al., 2002; Davies and Bavor, 2000). The decline in both FC and EC could be related to the decrease in manure application to the agricultural land (Vingarzan et al., 2002), as well as a decrease in rainfall during the dry season. This decrease in rainfall results in reduced agricultural runoff input into the watershed (Shanks et al., 2006). The decrease in runoff could also result in a decrease in nutrient availability, such as nitrate, phosphate, ammonia and DOC, during the dry season. This may have resulted in increased competition for limited resources, which could eliminate more vulnerable coliforms, such as FC and EC, and allow the flourishing of more versatile HPC (Reynolds, 1999). Another factor influencing HPC, FC, EC and ENT input into the stream is the ability of these bacteria to survive outside the host intestinal tract. While Kress and Gifford (1984) showed that faecal  49  coliforms are able to survive in cow manure for up to seven weeks under hot and dry summer conditions, other studies have shown that lower temperatures, manure crusting, moist weather conditions and the amount of organic matter in manure, allow regrowth of some coliform populations (Stoddard et al., 1998). Other studies have shown that age of faecal pat and amount of pat exposure to solar radiation affect the final concentration of E. coli in pats on grazing lands (Meays et al., 2005). All of these factors could reduce the amount of coliform bacteria, as well as pathogens reaching the water in runoff in the dry season. The effects of environmental factors, such as flow, temperature, and nutrient levels, are not accounted for in this paper but are discussed in another publication (Maal-Bared, 2008). The high levels of manure application in the wet season present a management challenge in this watershed. The high frequency of positive samples for pathogens may be related to the fact that Elk Creek has high wildlife densities at the headwaters (temperate rainforest) that could contaminate the water at the source (Meays et al., 2006). Also, Elk Creek is an intensive agricultural watershed with high levels of cattle, chicken, and dairy farming, all of which have been associated with Campylobacter sp., Salmonella sp., and E. coli, contamination of waterways (Jafari et al., 2006). Pathogen results showed that rainfall had some effects on the frequency of pathogens detected in Elk Creek. The frequency of pathogen detection from water column grab samples during the wet season suggests that water column testing is an acceptable method of microbial risk evaluation when rainfall is high. Many studies have found a relationship between waterborne disease outbreaks and extreme rainfall events (Thomas et al., 2006; Auld et al., 2004). Dorner et al. (2004) observed that E. coli O157 concentrations in a heavily impacted drinking watershed were sporadic and correlated with manure spreading followed by runoff events. This could have been the case in this watershed as well. Unfortunately, we only monitored in-stream contamination and have no access to data describing the effect of other inputs, such as runoff or septic tank leakage. Other studies, such as Jenkins et al. (2006), found that runoff from poultry farms carried high concentrations of faecal indicators, but no Campylobacter or Salmonella. However, this present study did not find that water column grab samples were representative of pathogen presence in the creek in the dry season. In the dry season, water carried low concentrations faecal contamination indicators and no pathogens (see Figure 2.6), while biofilms became the main sinks of microbial contaminants and pathogens in particular. Pathogen distribution during the low rainfall, dry season is of particular interest from a public health perspective, especially if we consider that waterborne outbreaks in Canada tend to be  50  highest during the spring/summer season (Schuster et al., 2005). Spring conditions involve spring melt runoff, while summer conditions involve intense rainfall separated by dry periods. During the dry periods in the summer, biofilms are able to grow until a disturbance causes detachment and sloughing of biofilm, which might contain pathogens. Our results are also significant in the context of climate change, which is expected to cause more extreme weather events separated by prolonged dry periods (Auld et al., 2004). These dry periods would allow pathogens and indicators to accumulate in the biofilm prior to resuspension into the water column and contribution to pathogen concentrations in the creek. The detachment of pathogens from biofilms into the water column to contribute to pathogen loading has not been shown in situ in natural aquatic environments, but it has previously been shown in water distribution systems (Stoodley et al., 2001). The present study suggests that biofilms on some surfaces act as reservoirs for opportunistic pathogens in the dry season. This is particularly relevant when considering microbial risk assessment as a watershed management and public health protection strategy. Perhaps looking further into the effect of nutrient conditions and other physical and chemical characteristics on pathogens detachment from the biofilm might lead to some promising new findings, although we have to be aware of the fact that most microbial biofilm detachment studies are confounded by flow conditions (Rubin and Leff, 2007) and protozoan and metazoan predation (Hunt and Parry, 1998). Despite these difficulties, the differences in indicator and pathogen distribution on different substratum types in our study highlights the need to examine other microbial populations beyond the suspended populations to get a better understanding of what water consumers may be exposed to if the water source experienced any sudden changes. From these results, further examination of river rock and sediment may be of interest to researchers, since it accumulated indicators and pathogens throughout all seasons. Replicating the present study is definitely needed across other watersheds. Also, despite the difficulty of attributing an outbreak to a particular source, coupling this kind of study with outbreak data in a community would help to evaluate epidemiological relevance and the effectiveness of biofilms as water quality prediction tools. Future studies should consider evaluating the effect of external inputs into the stream, such as runoff and septic tank leakage to gain a more complete understanding of this complex and open system. Finally, we would like to re-emphasize that we are not suggesting that testing of the water column is of no relevance. As a matter of fact, our study indicates that monitoring of the water column might be all that is needed  51  in high flow, high rainfall seasons. The question is: should we be looking at other surface types besides water in low flow, low rainfall scenarios? This may be particularly important when trying to: a) put together a microbial risk assessment for a watershed; b) put together watershedspecific management plans; c) alert water supply professionals, small system managers, and health professionals to what the community might be exposed to in the near future after rainfall events; and d) protect the public in watersheds that have high inputs of pathogens with low infective doses, like Campylobacter jejuni. Campylobacter jejuni has been known to cause outbreaks in communities dependent upon small water distribution systems in Canada and the US and has been known to survive and proliferate in biofilms rather than the water column. Designing appropriate water quality monitoring systems will become even more important in the context of constant agricultural expansion and climate change.  2.5  Conclusion  The present study raises many questions about the potential use of biofilms in water quality monitoring. There was a difference between the distribution of indicator bacteria and pathogens in water and the distribution of indicators and pathogens in biofilms in Elk Creek. When monitoring certain sites, the number of indicator organisms was almost one order of magnitude higher in sediment and in biofilms on some surfaces, than it was in the water column grab samples. In some cases, specific pathogens were not detected at all in the water samples, while they were present in sediment or biofilm samples. There was also a significant difference amongst the concentrations of faecal coliforms, E. coli and enterococci on river rock, slate rock, wood, Lexan, sandpaper, and sediment in Elk Creek in both the wet and the dry season. There was also a clear difference in pathogen detection frequency (Campylobacter, E. coli O157 and Salmonella) on the different surface types and water. What was even more interesting was the absence of pathogens in the water column grab samples in the dry season, thus suggesting that biofilms and sediment could serve as reservoirs of pathogens. When the aquatic environment experiences change, the result could be a release of microbes from biofilm and sediment into the water column. This indicates the need for different water source monitoring strategies in different seasons.  2.6 Acknowledgements This research has been funded by the CIHR Strategic Training Program in Public Health and the Agricultural Rural Ecosystem (PHARE) and Partner Institutes including the Institute of Cancer Research, Institute of Circulatory and Respiratory Health, Institute of Infection and Immunology,  52  and the Institute of Population and Public Health. We would also like to acknowledge the School of Environmental Health where all the lab work was completed.  53  2.7 References APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater, 20th Edition. American Public Health Association (APHA), American Waterworks Association (AWWA), and Water Environment Association (WEF). Washington, DC. An, Y.J., Kampbell, D.H., Breidenbach, G.P., 2002. Escherichia coli and total coliforms in water and sediment at lake marinas. Environmental Pollution 120, 771-778. Anderson, K.L., Whitlock, J.E., Harwood, V.J., 2005. Persistence and differential survival of faecal indicator bacteria in subtropical waters and sediment. Applied and Environmental Microbiology 71(6), 3041-3048. Auld, H., MacIver, D., Klaassen, J., 2004. Heavy rainfall and waterborne disease outbreaks: the Walkerton example. Journal of Toxicology and Environmental Health: Part A 67 (2022), 1879-1887. Barbiero, R.P., 2000. A multi-lake comparison of epilithic diatom communities on natural and artificial substrates. Hydrobiologia 438, 157-170. Bartram, J., Howard, G., 2003. Drinking water standards for the developing world: Emerging waterborne pathogens. In: Mara, D., Horan, N. (Eds.) Handbook for water and wastewater microbiology. Elsevier Press, London, 221-241. Cerca, N., Pier, G.B., Vilanova, M., Oliveira, R., Azeredo, J., 2005. Quantitative analysis of adhesion and biofilm formation on hydrophilic and hydrophobic surfaces of clinical isolates of Staphylococcus epidermis. Research in Microbiology 15, 506-514. Characklis, W.G., Cooksey, K.E., 1983. Biofilms and microbial fouling. Advances in Applied Microbiology 29, 93-138. Crabill, C., Donald, R., Snelling, J., Foust, R., Southam, G., 1999. The impact of sediment faecal coliform reservoirs on seasonal water quality in Oak Creek, Arizona. Water Research 33 (9), 2163-2171. Davies, C.M.,. Long, J.A.H., Donald, M., Ashbolt, N.J., 1995. Survival of fecal microorganisms in marine and freshwater sediments. Applied and Environmental Microbiology 61(5), 1888-1896. Davies, C.M., Bavor, H.J., 2000. The fate of stormwater-associated bacteria in constructed wetland and water pollution control pond systems. Journal of Applied Microbiology 89 (2), 349-360.  54  Davis, K., Anderson, M.A., Yates, M.V., 2005. Distribution of indicator bacteria in Canyon Lake, California. Water Research 39, 1277-1288. Decho, A.W., 2000. Microbial biofilms in intertidal systems: An overview. Continental Shelf Research 20, 1257-1273. Derksen, G., Schreier, H., Bestbier, R., Brown, S., 2004. Community watershed information system: Elk Creek watershed case study. Environment Canada & IRE-UBC (MultiMedia CD-ROM). Dorner, S.M., Huck, P.M., Slawson, R.M., Gaulin, T., Anderson, W.B., 2004. Assessing levels of pathogenic contamination in a heavily impacted river used as a drinking water source. Journal of Toxicology and Environmental Health: Part A 67, 1813-1823. Glasner, A., McKee, L., 2002. Pathogen occurrence and analysis in relation to water quality attainment in San Francisco Bay area watersheds. San Francisco Estuary Institute Report, San Francisco, USA. Document was retrieved on January 29, 2008 and is available at available at: http://www.sfei.org/watersheds/reports/Pathogens/Pathogenlitrevv6.pdf Holme, R., 2003. Drinking water contamination in Walkerton, Ontario: positive resolutions from a tragic event. Water Science & Technology 47 (3), 1–6. Hrudey, S.E., Hrudey, E.J., 2007. Published case studies of waterborne disease outbreaksevidence of a recurrent threat. Water Environment Research 79, 233-245. Hunt, A.P., Parry, J.D., 1998. The effect of substratum roughness and river flow rate on the development of a freshwater biofilm community. Biofouling 12(4), 287-303. Jafari, R.A., Fazlara, A., Gohavi, M., 2006. An investigation into Salmonella and fecal coliform contamination of drinking water in broiler farms in Iran. International Journal of Poultry Science 5 (5), 491-493. Jenkins, M.B., Endale, D.M., Schomberg, H. H., Sharpe, R. R., 2006. Faecal indicators and sex hormones in soil and runoff from cropped watersheds amended with poultry litter. Science of the Total Environment 358 (1-3), 164-177. Kress, M., Gifford, G.F. 1984. Fecal coliform release from cattle fecal deposits. Water Resources Bulletin 20(1), 61-66. Kroepfl, K., Vladar, P., Szabo, K.A.E., Borsodi, A.K., Szikora, S., Caroli, S., Zaray, G., 2006. Chemical and biological characterization of biofilms formed on different substrata in Tisza River (Hungary). Environmental Pollution 144 (2), 626-31. Maal-Bared, R., 2008. Comparing the distribution of pathogenic bacteria and common indicator microorganisms in biofilms on different surface types in an agricultural watershed in  55  British Columbia (Canada). Unpublished PhD Thesis, University of British Columbia, British Columbia, Canada. Meays, C. L., Broersma, K., Nordin, R., Mazumder, A., 2005. Survival of Escherichia coli in beef cattle fecal pats under different levels of solar exposure. Rangeland Ecology and Management 58 (3), 279-283. Meays, C.L., K. Broersma, R. Nordin, Mazumder, A., Samadpour, M., 2006. Spatial and annual variability in concentrations and sources of Escherichia coli in multiple watersheds. Environmental Science & Technology 40, 5289-5296. Muirhead, R.W., Davies-Colley, R.J., Donnison, A.M., Nagels, J.W., 2004. Faecal bacteria yields in artificial flood events: quantifying in-stream stores. Water Research 38, 1215-1224. Newman, R.F., Hooper, T.D., Powell, G.W., Njenga, F.M., 2003. The influence of range practices on waterborne disease organisms in surface water of British Columbia: a problem analysis. Resource Branch, B.C. Ministry of Forestry, Victoria, B.C. Tech. Rep. 008. Document was retrieved on January 29, 2008. Available at: http://www.for.gov.bc.ca/hfd/pubs/Docs/Tr/Tr008.htm Rahim, Z., Aziz, K.M.S., Islam, S., Huq, M.I., 1985. A preliminary survey of the relative abundance of fecal coliform bacteria in water and sediment and in the fresh water bivalve Lamellidens marginalis, of the Buriganga River, Bangladesh. Microbial Ecology 11 (3), 277-280. Reynolds, K.A., 1999. Heterotrophic plate count bacteria and drinking water. Water Conditioning and Purification 41, 1-2. Romani, A.M., Sabater, S., 2000. Variability of heterotrophic activity in Mediterranean stream biofilms: A multivariate analysis of physical-chemical and biological factors. Aquatic Sciences 62 (3), 205-215. Rood, K.M., Hamilton, R.E., 1995. Hydrology and water use for salmon streams in the Chilliwack/Lower Fraser habitat management area, BC. Prepared for the Fraser River Action Plan by Northwest Hydraulic Consultants Ltd., North Vancouver, BC. Rubin, M.A., Leff, L.G., 2007. Nutrients and other abiotic factors affecting bacterial communities in an Ohio River. Microbial Ecology 54 (2), 374-383. Sabater, S., Guasch, H., Ricart, M., Romaní, A., Vidal, G., Klünder, C., Schmitt-Jansen, M., 2007. Monitoring the effect of chemicals on biological communities: The biofilm as an interface. Analytical and Bioanalytical Chemistry 387(4), 1425-34.  56  Schendel, E.K., Schreier, H., Lavkulich, L.M., 2004. Linkages between phosphorus index estimates and environmental quality variables. Soil and Water Conservation Journal 59, 243-251. Schuster, C.J., Ellis, A.G., Robertson, W.J., Charron, D.F., Aramini, J.J., Marshall, B.J, Medeiros, D.T., 2005. Infectious disease outbreaks related to drinking water in Canada, 1974-2001. Canadian Journal of Public Health 96 (4), 254-258. Shanks, O.C., Nietch, C., Simonich, M., Younger, M., Reynolds, D., Field, K.G., 2006. Basinwide analysis of the dynamics of fecal contamination and fecal source identification in Tillamook Bay, Oregon. Applied and Environmental Microbiology 72 (8), 5537-5546. Stirling, R., Aramini, J., Ellis, A., Lim, G., Meyers. R., Fleury, M., Werker. D., 2001. Waterborne Cryptosporidiosis outbreak, North Battleford, Saskatchewan, Spring 2001. Canadian Communicable Disease Report 2722. Document was retrieved on January 29, 2008. Available at URL: http://www.phac-aspc.gc.ca/publicat/ccdrrmtc/01vol27/dr2722ea.html Stoddard, C.S., Coyne, M.S., Grove, J.H., 1998. Fecal bacteria survival and infiltration through a shallow agricultural soil: timing and tillage effects. Journal of Environmental Quality 27, 1516-1523. Stoodley, P.S., Wilson, L., Hall-Stoodley, J. D., Boyle, H. M., Lappin-Scott, Costerton, J. W., 2001. Growth and detachment of cell clusters from mature mixed-species biofilms. Applied and Environmental Microbiology 67, 5608-5613. Swain, L.G., Holms, G.B., 1985. Ambient water quality objectives for the Fraser River Subbasin from Hope to Kanaka Creek. BC Ministry of Environment: Waste Management Branch. Document was retrieved on January 29, 2008. Available at: http://www.env.gov.bc.ca/wat/wq/objectives/fraserhope/fraserhope.html Tank, J.L., Dodds, W.K., 2003. Nutrient limitation of epilithic and epixylic biofilms in ten North American streams. Freshwater biology 48, 1031-49. Thomas, K.M., Charron, D.F., Waltner-Toews, D., Schuster, C., Maaroof, A., Holt, J.D., 2006. A role of high impact weather events on waterborne disease outbreaks in Canada 1975-2001. Environmental Health Research 16 (3), 167-180. Vingarzan, R., Belzer, W., Thompson, B., 2002. Nutrient levels in the atmosphere of the Elk Creek Watershed, Chilliwack, BC (1999-2000). Environment Canada Technical Report. Aquatic and Atmospheric Sciences Division, Vancouver, Canada.  57  Whitman, R.L., Nevers, M.B., 2003. Foreshore sand as a source of Escherichia coli in nearshore water of a Lake Michigan beach. Applied and Environmental Microbiology 69 (9), 5555–5562. Yates, M.V., 2007. Classical indicators in the 21st century-far and beyond the coliform. Water Environment and Research 79(3), 279-86.  58  3  3.1  Effects of surface types, water quality indicators and colonization period on the distribution of indicator bacteria and pathogenic organisms in biofilms in an agricultural watershed 2  Introduction  Water monitoring is an essential component in the delivery of safe drinking water to the public. Currently, there are no specific protocols for estimating raw water quality at the source (Dorner et al., 2004), particularly ones that take into account attached bacterial communities or actual pathogen concentrations at the source. This may be problematic since future regulation will require drinking water treatment levels to be based on raw water quality (US EPA, 2003). The most commonly used methods in the estimation of pathogen loading into water treatment plants are based on estimation of indicator organism numbers, such as thermotolerant coliforms (also called faecal coliforms), enterococci and E. coli. However, many studies have shown that faecal contamination indicator organism concentrations in aquatic environments do not always correlate with pathogen presence (Glasner and McKee, 2002; Bartram and Howard, 2003; Yates, 2007). More studies are needed to address the conditions under which a correlation between indicators and pathogens does exist (Ouyang et al., 2006). Also, in aquatic environments, large numbers of bacteria and other microbes attach to surfaces instead of remaining suspended in the water column (Decho, 2000; Maal-Bared et al., 2007). These communities remain attached and protected until changes in the aquatic environment encourage their detachment and resuspension into the water column. The detachment of pathogens from biofilms into the water column to contribute to pathogen loading has not been shown in situ in natural aquatic environments, but it has previously been shown in water distribution systems (Stoodley et al., 2001). The majority of water pollutant monitoring programs rely on the collection of discrete water column samples at infrequent points in time (Vrana et al, 2005). Thus, most monitoring systems may only partially reveal the aquatic 2  A version of this chapter will be submitted for publication. Maal-Bared, R., Bartlett, K.H., Bowie, W.R., Hall, E.R., Hall, K.J. 2008. Effects of surface types, water quality indicators and colonization period on the distribution of indicator bacteria and pathogenic organisms in biofilms in an agricultural watershed. Applied and Environmental Microbiology.  59  microbial population that water consumers may be exposed to by revealing its planktonic components (Byrd, et al., 1991, Emtiazi et al., 2004). An important concept in understanding biofilm community changes is that of ecological succession (coined by Odum in 1969), which was explained in a review by Jackson (2003). Succession in an ecosystem is a process through which one community is replaced by another until one stable community is reached. In a biofilm, a heterotrophic community would first colonize a surface and rely on dissolved organic carbon (DOC) from the water column as a nutritional source. As the biofilm becomes thicker, DOC diffusion through the biofilm matrix becomes more difficult. Thus, autotrophic microorganisms, such as algae and cyanobacteria, become more welcome members of the biofilm community. Also, heterotrophs can either be replaced by other communities, lost due to sloughing, or decay as a result of inability to adapt to the new environment. Jackson (2003) suggested that competition in biofilms becomes less relevant in thicker and more mature biofilms, due to an increase in microniches over time. He stated that within a biofilm, an increase in species diversity would occur at the beginning of the colonization process, as the substratum is colonized. This would be followed by a decrease in numbers as members of the community start competing for space and resources (fine and broad scale competition). Finally, another increase in numbers would occur as the biofilm becomes thicker and develops more ecological niches that can be colonized by different organisms, which would still compete amongst themselves (fine scale competition). At later stages, Jackson (2003) suggests that protozoan predation and viral lysis become important players in structuring the biofilm community. In a previous study (Maal-Bared et al., 2007), the potential role of biofilms on different substratum types (river rock, slate rock, wood, Lexan, sandpaper and sediment) as sinks for indicator organisms and pathogens in an agricultural watershed was examined. It was concluded that substratum surface type and seasonality (particularly rainfall) in an agriculturally impacted stream affected the concentrations of indicator bacteria (heterotrophic plate counts (HPC), faecal coliforms (FC), E. coli (EC) and enterococci (ENT)) and frequencies of pathogen detection (Campylobacter sp., Salmonella sp., and E. coli O157) in biofilms. In the present study, the environmental factors associated with the distribution of indicator organisms and pathogens on different surface types in the aquatic environment were examined. Our research questions were the following.  60  1. What are the relationships between water quality characteristics and the concentrations of indicator bacteria and pathogens in biofilms on each substratum type? 2. How does colonization period (short-term and long–term) affect the distribution of the indicator bacteria and pathogens in biofilms on different surface types? 3. Do water quality characteristics exhibit different relationships with biofilms at different stages of maturation (or colonization)? Our hypothesis was that water quality variables, substratum surface type and surface colonization time may affect biofilm thickness, the availability of nutrients and oxygen, competition amongst biota, and predation rates within the biofilm, all of which may alter population densities of bacteria and ecological niches available to the bacterial community at different points in time.  3.2  Methods  3.2.1 Sampling location- Elk Creek The Elk Creek Watershed is located in the Fraser Valley in British Columbia (Canada). It has an area of about 28 km2 and the watercourse is about 12 km long (Rood and Hamilton, 1995). It is mainly agricultural in nature, although it also has forested headwaters located in the mountains, and limited urban residential influences. Four sites were located on the stream for this study (see Figure 3.1). A detailed description of the activities and impacts in the watershed is provided in Maal-Bared et al. (2007).  61  Hope Slough  Site 4 Site 3  Nevin Creek  Site 2 Dunville Big Ditch  Creek  Ford Creek  Elk Creek  Site 1  North  Figure 3.1 Map of Elk Creek in Chilliwack, British Columbia, and location of sites 1 to 4 (marked by the triangles) (Scale about 1: 90,000, 16.7 x 11.7 km). Arrow indicates north.  3.2.2 Biofilm samplers Biofilm samplers were built using the method described in Maal-Bared et al. (2007). The only difference in the method was that for long and short-term samplers, sandpaper was not used as a substratum due to its deterioration and loss when left in the water for longer than one month.  3.2.3 Sample collection Samples were collected between December 2005 and April 2007 on a rotating basis. One-month samplers were collected from December 2005 to December 2006 using the method described in Maal-Bared et al. (2007). In the wet season, 11 samplers were retrieved. Only six samplers were left intact and retrieved in the dry season. All samplers collected between December 2005 and December 2006 were left at a site for four weeks before collection and analysis, since previous investigations indicated that 2 to 4 weeks were required for mature biofilm community formation (Hunt and Parry, 1998). Short- and long-term samplers were collected at site 2, 3, and 4 between January 2007 and April 2007 on nine consecutive dates. The samplers were left at different sites for one week (one sampler with duplicate substrata at site 3), two weeks (one sampler with duplicate substrata at site 3), three weeks (one sampler with duplicate substrata at site 3), four weeks (one sampler with duplicate substrata at site 3), twelve weeks (three samplers with duplicate substrata each at  62  site 2, 3, and 4; n=6) and twenty-four weeks (two samplers with duplicate substrata at sites 2 and 3). At the time of sampler collection, duplicate water column grab samples of river water were collected into sterilized 500 mL bottles, as well as duplicate 100 g surface sediment samples from the same site. The surface sediment grab samples were taken from the top 10-15 cm of the sediment and were collected using a Teflon-coated pot that was attached to an extendable pole. All samples were kept on ice packs in a cooler for transport to the laboratory. In the laboratory, to remove biofilms from substrata, substrata were placed in 500 ml of sterile PBS (PBS, pH=7.2, 9 g NaCl/L, 0.0067M PO4) in a sterile beaker, which was placed into a sonication bath for 30 minutes. The bath was continuously cooled using ice packs. Sonication conditions were tested and determined during previous work to minimize sample loss. The PBS-biofilm suspension was used for analysis. Fifty grams of wet sediment samples were blended in 450 mL of sterile PBS for 3 minutes at low power in an autoclaved stainless steel blender, and then transferred to a sterile bottle. Prior to analysis, the sediment suspension was shaken for 20 seconds to allow for resuspension and homogenization. Water was analyzed directly.  3.2.4 Microbial Analysis Standard Methods procedures were used for the enumeration and isolation of indicator organisms (APHA et al., 1998). For a detailed description of the indicator organism and pathogen analysis methods, see Maal-Bared et al. (2007).  3.2.5 Standardizing across different substrata For a detailed description of the loss on ignition method used to approximate ash-free dry weight (AFDW) of biofilms on each substratum, as well as AFDW in water and sediment, see MaalBared et al. (2007).  3.2.6 Other water quality parameters During every sampling trip, physical and chemical water quality parameters were measured. Water depth was manually measured at the locations at which the biofilm sampler was placed on the sediment. Stream bank width was also measured. A stopwatch was used to estimate water velocity by measuring how long it took a water bubble to travel a specific distance. Three variables (water depth x width x velocity) were the basis of the volumetric water flow calculation (m3/s). A YSI 85 instrument (YSI Inc., Yellow Springs, Ohio) was used for water temperature  63  (oC), conductivity (µS/cm), specific conductivity (µS/cm), and dissolved oxygen (mg/l) in situ measurements. The instrument was calibrated on a regular basis using the instructions and standard solutions provided by the manufacturer. Chilliwack ambient rainfall measurements (collected at the Abbotsford Airport weather station about 45 km from the watershed) were retrieved from the Weather Network website (http://www.theweathernetwork.com/weather/CABC0308) and were used to calculate the cumulative rainfall over 7 days before the sampling date (mm). To measure chemical water quality variables (nitrate+nitrite, ortho-phosphate, ammonia and dissolved organic carbon), water samples were collected and returned to the laboratory for analysis. The same water samples were used for nitrate+nitrite and ortho-phosphate analysis. The samples were dispensed in duplicate to 10 mL test tubes using a 10 mL syringe with a 0.45 µm pore size filter attached. The test tubes contained a drop of 0.1% phenyl mercuric acetate to preserve the samples for analysis. Ammonia samples were filtered in the same manner as above, but preserved using a drop of 0.5% HCl solution. Finally, dissolved organic carbon (DOC) samples were filtered and dispensed into 30 mL glass vials and preserved by lowering the pH of the samples below 2 using phosphoric acid. The nutrient analyses for nitrate+nitrite, orthophosphate and ammonia were performed using a Lachat QuikChem 8000 Flow Injection Analyzer. Ortho-phosphate concentrations were determined using Standard Method 4500-P (Flow injection analysis for orthophosphate). Ammonia concentrations were determined using Proposed Method 4500-NH3 H (Flow injection analysis). And nitrate+nitrite concentrations were found using proposed Standard Method 4500-NO3-I (Cadmium flow injection method). DOC concentrations were obtained using the persulfate-ultraviolet oxidation method (Standard Method 5310 C) and using a Dohrmann Phoenix 8000 TOC analyzer. For quality control purposes, all nutrient samples were collected and analyzed in duplicate. Field and method blanks were collected and run with sample batches. Standards of each analyte were included in each sample run. For total protozoa counts, 500 mL of river water were collected at the time of biofilm collection and fixed with 10% Formalin. The samples were kept at 4oC until examination. For examination, six subsamples from the fixed water sample were analysed in the 1-mm2 counting chamber in the middle of a hemocytometer using a light microscope under several different magnifications (40x, 100x, 250x, and 1000x). The identification of the organisms (to genera) was completed using Patterson’s Freshwater Protozoa Identification Guide (Patterson, 2003). This method was chosen  64  to identify the most common protozoan co-isolates, rather than determining protozoan concentration or diversity. For the long- and short-term samplers, protozoa within the biofilm were identified, as well as water column grab samples.  3.2.7 Statistical Analyses All statistical analyses were completed using JMP (Version 7.0, SAS Institute, NC, USA). Nonparametric tests were used because the distributions of some variables could not be transformed to fit a normal Gaussian distribution. Correlations among water quality variables and microbial variables (one-month, short-term and long-term) were examined using Spearman Correlations and reported in terms of Spearman ρ statistics and p-values. Alpha was set at 0.05. To evaluate the length of colonization required to reach a stable biofilm community short-term, short-term data will include results from weeks 1 to 4, as well as long-term data. To investigate differences in correlations between short-term, long-term and one month biofilms and water quality parameters, short-term biofilm results only included data from weeks 1 to 3.  3.3 Results 3.3.1 Water quality in Elk Creek Different measures of location and dispersion of physical, chemical and biological water quality parameters are given in Tables 3.1, 3.2, and 3.3, respectively. During the dry period (April to September), the mean ambient temperature was 18.1OC, and mean rainfall was 11 mm over a 7day period (approximately 38 mm per month). On average the water velocity was highest at site 1 (mean = 0.7 m/s), while the streambed at site 4 was the widest and had the highest flow (mean width = 5.4 m and mean flow = 1.6 m3/s). Water temperature was highest at site 3 (mean = 13.3OC). Dissolved oxygen levels were highest at site 4 during the dry season (means were 10.7 mg/L and 109.0%). Mean conductivity and specific conductivity were highest at site 2 (185 µS/cm and 237 µS/cm respectively). Salinity did not change during the dry season (mean = 0.1 ppt). Nitrate levels were highest at site 2 (mean = 0.35 mg N/L), and DOC levels were highest at site 1 (mean = 3.6 mg C/L). Ammonia levels were comparable between sites 2 to 4 (mean = 0.04 mg N/L), but PO4 concentrations were highest at site 4. Protozoa counts were highest at site 2 and 4. During the wet season (October to March), the mean ambient temperature was 6.6OC, and mean rainfall was 44 mm over a 7-day period (approximately 167 mm per month). With the exception of site 4, stream velocity, depth, width and flow were higher in the wet season than in the dry  65  season. The depth measurement refers to both the depth of the water at each site, as well as the depth at which the biofilm samplers were placed onto the sediment. Dissolved oxygen and conductivity did not exhibit any clear patterns between seasons. Salinity ranged between 0.1 ppt and 0.2 ppt. Nitrate concentrations increased at all sites during the wet season. Similar to the dry season results, the water velocity was highest at site 1 (mean = 0.94 m/s), while the streambed at site 4 was the widest and had the highest flow (mean width = 5.8 m and mean flow = 3.1 m3/s). Water temperature was highest at site 4 (mean = 7.8OC) and lowest at site 1 (mean = 5.8OC). Dissolved oxygen levels were lowest at site 4 during the wet season (means were 10.0 mg/L and 84.3%). Mean conductivity and specific conductivity were highest at site 2 (176 µS/cm and 258 µS/cm respectively). Nitrate levels were highest at site 3 (mean = 0.79 mg N/L), while PO4 concentrations were highest at site 2 (mean = 0.03 mg P/L). DOC levels were highest at site 2 and 4 (mean = 3.0 mg C/L). Ammonia levels gradually increased between sites 1 and 4 reaching a mean value of 0.11 mg N/L. There were two extreme rainfall events in the watershed. The first occurred in January 2006 (90 mm of rainfall during a 48 hour period on January 9 and 10), while the second occurred in March 2007 (78 mm of rainfall on March 11, 2007). The results show that Elk Creek is an agriculturally impacted stream and that nutrient levels (with the exception of DOC) at sites 2 to 4 are higher than those at the reference site (site 1). However, there does not seem to be a constant increase in contaminant gradient between sites 1 and 4 as Hope Slough is approached. Protozoa counts were also highest at site 2 and 4 in the wet season.  66  Table 3.1 Physical water quality variables measured at four sites in Elk Creek (BC) between December 2005 and April 2007 presented as means, medians, standard deviations, and sample sizes (n). Season Dry Season Site  1  2  3  4  Water Quality Variable Velocity (m/s) Depth (cm) Width (cm) Water Temperature (oC) Ambient Temp (oC) Rainfall (mm) Flow (m3/s) Velocity (m/s) Depth (cm) Width (cm) Water Temperature (oC) Ambient Temp (oC) Rainfall (mm) Flow (m3/s) Velocity (m/s) Depth (cm) Width (cm) Water Temperature (oC) Ambient Temp (oC) Rainfall (mm) Flow (m3/s) Velocity (m/s) Depth (cm) Width (cm) Water Temperature (oC) Ambient Temp (oC) Rainfall (mm) Flow (m3/s)  Wet Season  Mean  Median  Std Dev  n  Mean  0.7 25 517 10.8 18.1 11 1.2 0.3 40 374 12.9 18.1 11 0.4 0.3 67 408 13.3 18.1 11 0.8 0.4 56 542 12.9 18.1 11 1.6  0.6 19 500 11.2 16.8 2.0x10-1 0.5 0.3 40 400 12.6 16.8 2.0x10-1 0.3 0.3 65 500 12.7 16.8 2.0x10-1 0.7 0.3 50 550 12.1 16.8 2.0x10-1 0.8  0.4 12 52 2.1 4.2 18 1.3 0.2 17 44 3.6 4.2 18 0.4 0.2 7 187 4.0 4.2 18 0.7 0.3 19 38 4.1 4.2 18 1.6  6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 7 7 7 6 6 6 6 6 6 6  0.9 21 528 5.8 6.5 44 1.7 0.3 66 450 7.6 6.5 44 1.1 0.3 86 522 7.6 6.5 44 1.6 0.4 121 577 7.8 6.5 44 3.1  Median Std Dev 0.7 20 500 5.6 6.5 31 0.5 0.3 58 450 7.2 6.5 31 0.6 0.3 56 525 7.1 6.5 31 1.0 0.4 110 550 7.2 6.5 31 2.1  1.0 16 70 1.4 3.5 47 4.9 0.2 44 66 1.8 3.5 47 1.4 0.1 74 64 2.1 3.5 47 1.6 0.2 88 93 1.9 3.5 47 3.3  n 23 23 23 23 23 23 23 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22  67  Table 3.2 Chemical water quality variables measured at four sites in Elk Creek (BC) between December 2005 and April 2007 presented as means, medians, standard deviations, and sample sizes (n). Season Dry Season Site  1  2  3  4  Wet Season  Water Quality Variable  Mean  Median  Std Dev  n  Mean  DO (mg/L) DO (% air saturation) Conductivity (µS/cm) Sp. Conductivity (µS/cm) Salinity (ppt) NOx (mg N/L) PO4 (mg P/L) NH3 (mg N/L) DOC (mg C/L) DO (mg/L) DO (% air saturation) Conductivity (µS/cm) Sp. Conductivity (µS/cm) Salinity (ppt) NOx (mg N/L) PO4 (mg P/L) NH3 (mg N/L) DOC (mg C/L) DO (mg/L) DO (% air saturation) Conductivity (µS/cm) Sp. Conductivity (µS/cm) Salinity (ppt) NOx (mg N/L) PO4 (mg P/L) NH3 (mg N/L) DOC (mg C/L) DO (mg/L) DO (% air saturation) Conductivity (µS/cm) Sp. Conductivity (µS/cm) Salinity (ppt) NOx (mg N/L) PO4 (mg P/L) NH3 (mg N/L) DOC (mg C/L)  10.6 95.2 149 201 0.1 0.34 0.01 0.02 3.6 9.8 92.2 185 237 0.1 0.35 0.01 0.04 2.6 10.4 98.7 182 230 0.1 0.34 0.01 0.04 2.8 10.7 100.9 179 228 0.1 0.32 0.20 0.04 3.1  10.3 94.5 166 223 0.1 0.36 0.01 0.02 2.4 9.0 89.3 204 267 0.1 0.35 0.01 0.03 2.4 10.2 100.8 195 254 0.1 0.35 0.01 0.03 2.4 10.5 102.1 179 246 0.1 0.32 0.02 0.02 2.9  1.2 7.0 69 87 0.0 0.15 3.0x10-3 0.01 3.3 1.5 7.7 65 68 0 0.09 3.0x10-3 0.01 1.0 1.5 7.8 68 70 0 0.08 2.0x10-3 0.03 0.9 1.3 6.2 72 74 0.0 0.08 0.29 0.03 0.8  6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 6 6 6 6 6 6 6 6 6  13.2 105 116 182 0.1 0.40 0.01 0.04 2.3 10.0 86.0 176 258 0.1 0.73 0.03 0.09 3.0 10.6 88.6 166 251 0.1 0.79 0.02 0.10 2.7 10.0 84.3 172 255 0.1 0.76 0.02 0.11 3.0  Median Std Dev 13.3 105 128 204 0.1 0.39 0.01 0.01 1.9 10.4 86.7 173 259 0.1 0.76 0.01 0.06 2.8 10.8 87.9 169 258 0.1 0.81 0.01 0.07 2.9 10.0 83.8 180 268 0.1 0.80 0.01 0.09 2.9  0.6 2.8 46 70 3x10-2 0.10 0.01 0.08 1.1 1.0 10.0 33 49 2x10-2 0.20 0.04 0.09 1.2 0.9 9.6 34 43 2x10-2 0.27 0.01 0.10 1.0 0.7 6.9 35 48 2x10-2 0.29 0.02 0.10 1.5  68  n 23 23 23 23 23 23 23 23 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22  Table 3.3 Biological water quality variables measured at four sites in Elk Creek (BC) between December 2005 and April 2007 presented as means, medians, standard deviations, and sample sizes (n). Season Dry Season Water Quality Variable Site (measured in cell numbers) Amoebae Flagellates 1 Ciliates  2  3  4  Total Protozoa Amoebae Flagellates Ciliates Total Protozoa Amoebae Flagellates Ciliates Total Protozoa Amoebae Flagellates Ciliates Total Protozoa  Wet Season  Mean  Median  Std Dev  n  Mean  Median  Std Dev  n  5 3 2  5 3 2  1 0 2  2 2 2  5 2 5  4 1 4  3 2 3  5 5 5  9 9 6 1 15 10 1 2 13 10 2 7 19  9 9 6 1 15 10 1 2 13 10 2 7 19  3 0 0 0 0 5 2 6 3 0 0 0 0  2 1 1 1 1 3 3 3 3 1 1 1 1  10 12 5 2 19 11 3 2 16 13 4 3 20  10 12 4 2 15 6 2 2 10 12 4 3 20  5 8 4 2 7 8 3 2 3 5 3 3 8  5 5 5 5 5 5 5 5 5 5 5 5 5  3.3.2 Variations in one-month biofilm samplers analyzed by substratum type and site Relationships between water quality variables and concentrations of indicator organisms and pathogen detection frequencies in one-month biofilm samplers between December 2005 and December 2006 among sites 1 to 4 in Elk Creek can be seen in Table 3.4. Table 3.4 only contains statistically significant relationships. The table collapses velocity, width and flow in to one variable (flow). It also collapses specific conductivity and conductivity into one variable (conductivity), dissolved oxygen measures (concentration and air % saturation) into dissolved oxygen (DO), and flagellate, ciliate, amoeba and total protozoa counts into one variable (protozoa). Also, it is important to note that the correlations with protozoa numbers presented in Table 3.4 are based on counts from the water column rather than the biofilm itself. In the case of water, many of the indicator organisms correlated with each other (ENT with FC and EC), as  69  well as with specific conductivity. Also, stream velocity and width affected numbers of heterotrophic organisms at site 1. At site 2, HPCs in water showed a negative relationship with temperature (both water and ambient), but no indicator organisms correlated with each other. Also, flagellates and ciliates showed a positive relationship with FC and Salmonella respectively. Salmonella also correlated with PO4 concentrations, which may be a reflection of runoff at that site. The numbers of significant correlations decreased closer to Hope Slough, with only FC and HPC showing some relationships with water quality variables at sites 3 and 4. This decrease in number of correlations seemed to be common for all substrata except for sediment. Indicator organism numbers in sediment correlated positively with water conductivity measurements. Also, EC numbers in sediment correlated with both HPC and FC numbers. At sites 2 to 4, the concentrations of NH3, PO4, and DOC correlated with several of the indicator organisms and pathogens. Phosphate in particular correlated well with EC numbers and pathogen frequencies at site 3, which may be another site highly affected by runoff.  Biofilm indicator organisms on different substrata at site 1 often correlated positively with ambient temperature, velocity, and conductivity measurements. At sites 2 and 4, many of the correlations between indicator organism numbers and variables such as water temperatures, rainfall, flow, and nutrient concentrations may be related to runoff events. In general, microbial variables in biofilms frequently correlated more with nutrient concentrations and conductivity measures than with water temperature. From these correlations, several conclusions can be made. Faecal coliforms and EC correlated with each other to different extents, depending on the monitoring site’s water quality and the substratum examined. Enterococci numbers in biofilms were negatively correlated with rainfall runoff measures and HPCs at sites 2 and 3, highlighting conditions under which ENT survival and competitive abilities may be enhanced, perhaps making it a better indicator under those conditions. The numbers of significant correlations between sediment pathogens and nutrients in the agricultural stretch of Elk Creek indicated the possibility of using nutrients as pathogen presence surrogates for sediment. Finally, it is important to note the large number of significant relationships between indicators and pathogens, and protozoa. Accounting for protozoan relationships with indicators and pathogens in drinking water quality studies may be far more important than was previously assumed.  70  Table 3.4 Results of Spearman correlations between water quality variables and indicator organism numbers (Heterotrophic plate counts (HPC), faecal coliforms (FC), E. coli (EC) and enterococci (ENT)), pathogen frequencies (Campylobacter sp. (Cam), Salmonella sp. (Sal), and pathogenic E. coli O157 (ECO)), ash-free dry weights (AFDW) in water, sediment and biofilms at different sites in Elk Creek (BC).  Water  HPC FC EC ENT Sal  +  Protozoa  Water Temperature  DOC  NOx  NH3  DO  Flow  -  +  +  Conductivity  Depth  Microbial  Protozoa  Water Temperature  DO  Site 4  NOx  Rainfall  Conductivity  Flow  Depth  Microbial  Protozoa  DOC  + +  FC + FC+ EC +  FC EC ENT FC + AFDW  Water Temperature  PO4  Site 3  NH3  DO  Conductivity  Depth  Rainfall  Flow  Microbial  DOC  Protozoa  NOx  NH3  Rainfall  Conductivity  Flow  Site 2  +  HPC River Rock FC EC ENT AFDW EC Cam Slate Rock HPC  Cam  Depth  Microbial  Variable correlates with  Substratum  Site 1  +  + -  + +  -  -  -  -  Cam + +  + +  -  -  -  -  +  -  -  + + + + -  -  -  ECO +  +  HPC +  +  +  +  +  -  -  +  +  ECO +  Sal  -  * A positive sign (+) indicates a positive relationship, while a negative sign (-) indicates a negative relationship.  71  71  Table 3.4 Results of Spearman correlations between water quality variables, indicator organisms, pathogen frequencies, and ash-free dry weight continued.  Wood HPC ENT AFDW Sal Lexan  FC EC  FC +  NH3  DO  NOx  Conductivity  Rainfall  Flow  Depth  Microbial  Protozoa  Water Temperature  DOC  PO4  NH3  NOx  DO  Rainfall  Conductivity  Depth  Flow  EC+ ENT-  -  + ECO +  ECO +  Sal +  + +  +  +  -  +  -  +  -  ENT - -  +  -  + +  +  +  EC+ ENT+  HPC FC EC ENT + ENT AFDW Sal Sediment HPC FC HPCEC FC+ ENT AFDW ENT +  ENT+ EC-  +  +  + ECEC-  +  + +  + + +  +  +  ENT -  -  -  +  + -  -  -  +  +  + + -  +  +  + +  FC +  +  -  + +  + + ENT + +  + +  Cam+ ECO +  +  +  Cam Sal ECO  Site 4  +  +  ENT AFDW Sal Sandpaper  Microbial  Protozoa  DOC  Water Temperature  NH3  DO  Site 3  PO4  Conductivity  +  Rainfall  Flow  ENT -  Depth  Microbial  DOC  Protozoa  Site 2  NH3  NOx  Conductivity  Flow  Microbial  Substratum  Variable correlates with  Site 1  + + +  EC -  +  + +  +  +  +  +  Cam+ Cam+  + +  Cam + Sal+  * A positive sign (+) indicates a positive relationship, while a negative sign (-) indicates a negative relationship.  72 69 72  3.3.3  Long- and short-term biofilms  Figure 3.2 shows the variations in indicator bacteria, pathogen numbers and ash-free-dry weight over time for the long- and short-term samplers. For this component of the study, data collected from weeks 1 to 4 and weeks 12 and 24 were included in the analysis. The indicator concentrations were not standardized by ash-free dry weight. The numbers of pathogens were averaged. Table 3.5 briefly describes the changes in indicator and pathogen numbers and AFDW in biofilms on different substrata with different colonization periods. In general, different substrata showed different trends. Enterococci were frequently higher than faecal coliforms. Campylobacter (Campy) and Salmonella (Sal) often exhibited reversed peaks (when Campylobacter was present in high numbers Salmonella was not detected and vice versa). Wood and sediment were the best performing substrata when it came to numbers of indicators and pathogens retrieved. Also, any study of indicators and pathogens in biofilms in natural environments needs to take into account the optimal colonization period for the substratum under investigation. In most cases, a two to four week colonization period is optimal to retrieve the highest number of indicators and pathogens. Long- and short-term biofilm protozoa counts are shown in Appendix C. Table 3.5 Trends and variations in the concentrations of indicator organisms and presence of pathogens in biofilms collected from short- and long-term samplers across different sites. Short- and long-term samplers Substratum  Water River Rock  Slate Rock  Wood  Lexan Sediment  AFDW  HPC  FC  EC  ENT  Slight decrease over time Slight increase over time  Peaked at week 2 and 4 Peaked at week 2 and 24  Higher than EC until week 24 None at week 4  Not present at week 3,4 and 12 None present after week 3  Higher than FC  Highest at week 3 and 4  Trend similar to river rock  Higher than ENT at week 2 and 3  Only present in long term samples  Highest at week 3 and 4  Highest at week 12  FC higher than ENT except in long term  Highest at week 3 and 4 Decreased until week 4  Slight increase over time Decreased until week 12  FC higher than ENT except at week 4 Lowest at week 24  Not present at week 1 and 24, highest at 4 Present week 1 to 3 Present week 1 to 24  None at week 4 and lower than FC Not present at week 1 (like FC) and peaked at week 4 Lowest at week 2 and 12 Highest at week 4 Higher than FC week 1, 12 and 24  E.coli O157  Campy  Sal  Peaked week 2 and 12 Peaked week 2,3 and 24  Peaked week 4 and 24 Only present week 1 and 24 Present at week 2 and 12  Not present  Present from week 2 to week 12 Present week 12 and 24 Present week 1,3,4, 12  Present at week 3 and 4  Present at week 4 and 24 Present from week 2 to 4 and week 24 Not present Present at week 4 and 12  73  Not present Not present  Present at week 2 Present at week 4  a) Water  CFU/100ml or MPN/100ml  Organic matter weight in water column  10000000 1.5  1000000 100000 10000  1  1000 100 0.5  10 1 0.1  0  24w  4w  12w  b) River Rock  3w  2w  1w  0.01  Average number of samples positive for pathogen  100000000  A ccumulation period  1  CFU/100ml or MPN/100ml Organic matter weight in biofilm  1000000  0.8 1000  0.6 0.4  1  0.2 0  0.001  Average number of samples positive for pathogen  1.2  24w  12w  4w  3w  2w  1w  -0.2  Accumulation period  c) Slate Rock 1000000  1  CFU/100ml or MPN/100ml Organic matter weight in biofilm  400000  100000  0.8  40000  10000 0.6  4000  1000 400  0.4  100 40  0.2  10 4  0  1 0.4  0.1  Average number of samples positive for pathogen  10000000  24w  12w  4w  3w  2w  1w  -0.2  Accumulation period  Left Scale: HPC/100ml ENT/100ml Right Scale:  FC/100ml OM Weight Campylobacter  EC/100ml  E.coli O157  Salmonella  Figure 3.2 Effect of different substratum colonization periods (1 week (n=2), 2 weeks (n=2), 3 weeks (n=2), 4 weeks (n=2), 12 weeks (n=6) and 24 weeks (n=4)) on indicator bacteria concentration, pathogen presence and ash-free dry weight (OM Weight) in water and river and slate rock.  74  d) Wood 3000000  2  1000000 CFU/100ml or MPN/100ml Organic matter weight in biofilm  400000  100000  1.5  40000  10000 4000  1  1000 400  0.5  100 40  10  0  4  1 0.4 0.2  Average number of samples positive for pathogen  10000000  24w  12w  4w  3w  2w  1w  -0.5  Accumulation period  e) Lexan 1.2  100000  1  CFU/100ml or MPN/100ml Organic matter weight in biofilm  40000  10000  0.8  4000  1000  0.6  400  100 0.4  40  10 4  0.2  1 0.4  0  0.1 0.04  Average number of samples positive for pathogen  1000000  24w  12w  f) Sediment  4w  3w  2w  1w  -0.2  2  CFU/100ml or MPN/100ml Organic matter weight in sediment  10000000 3000000  1000000 1.5  300000  100000 30000  10000  1  3000  1000 300  0.5  100 30  10  0  3  Right Scale:  24w  12w  4w  3w  Accumulation period  Left Scale: HPC/100ml OM Weight  2w  1w  1  Average number of samples positive for pathogen  Accumulation period  100000000  FC/100ml  Campylobacter  EC/100ml  E.coli O157  ENT/100ml  Salmonella  Figure 3.2 Effect of different substratum colonization periods (1 week (n=2), 2 weeks (n=2), 3 weeks (n=2), 4 weeks (n=2), 12 weeks (n=6) and 24 weeks (n=4)) on indicator bacteria concentration, pathogen presence and ash free dry weight in wood and Lexan, and in sediment.  75  3.3.4 Relationships between water quality variables, indicator organisms and pathogens in biofilms in short and long-term biofilms Table 3.6 shows the results of the Spearman correlation assessment between water, biofilms and sediment microbial variables and the measured water quality variables in Elk Creek. While data was collected for four weeks for short-term samplers, only results from weeks 1 to 3 were included in this analysis. Also, all short-term samplers were allowed to colonize and collected at site 3. Long-term samplers were the samplers that allowed substrata to be colonized for either 12 or 24 weeks, and were collected from sites 2, 3 and 4. Since the calculated p-values were suspect due to the very small sample size, p-values are not presented and only relationships are shown in Table 3.6. Flow, dissolved oxygen, conductivity and protozoa count variables were collapsed in a manner similar to that in Table 3.4. Both long and short-term samplers were colonized and collected during the wet season and therefore only reflect wet season relationships. Short-term samplers had very different relationships with water quality data than those from long-term samplers. The numbers of microbial indicators, pathogens and ash-free dry weight in biofilms were generally positively correlated with nitrate, ammonia, velocity and flow, but were negatively correlated with conductivity measures and water depth. This may be a reflection of the formation of an initial colonizing microbial biofilm. This will be addressed in detail in the discussion. The positive correlation with nutrients, flow and velocity may be an indication that organisms and nutrients were being carried to the substratum from within-stream sites or through runoff through mass transport events, which are the limiting factors in the initial stages of biofilm formation. The relationship with nutrients may be a reflection of the dependence on “outside” sources of nutrient acquisition. Finally, the negative correlation between microbial variables, conductivity, and water depth may be a result of the decreased chance of bacterial attachment to a biofilm if the volume of water in the creek was high. In the case of long-term biofilm samplers, fewer relationships were observed between the numbers of indicators, pathogens and ash-free dry weights, and water quality variables. The relationships between microbial variables and dissolved oxygen levels and water temperature remained positive in many cases. However, the correlations between microbial variables and nutrients, as well as rainfall, were negative. The results indicate that long- and short-term biofilms are at different stages of succession and therefore have different characteristics and requirements for continuation of growth within the biofilm.  76  Table 3.6 Comparing water quality variables relationships with indicator bacteria concentrations and pathogen presence in biofilms on different surfaces, which were colonized for short- and longterm periods.  Water  HPC FC  + EC +  EC ENT  River Rock  +  -  -  -  +  -  + ENT +  FC  +  - + +  +  ENT +  EC  FC + +  - + +  +  ENT  + -  -  +  - +  HPC  +  - +  +  +  FC  +  - +  +  +  +  +  AFDW  +  +  +  - +  Sal +  +  EC  FC + Sal +  +  ENT Campy  +  +  +  +  +  Sal  +  +  +  -  -  +  -  +  + -  + +  + +  FC + +  -  +  +  -  +  FC  +  +  Sal +  ENT  +  AFDW  +  +  +  HPC  + EC +  EC  Sal  +  +  + - +  + +  +  EC O157  AFDW  -  +  + - +  AFDW  +  +  Campy + +  FC  ENT  +  Campy -  ENT  FC  -  Sal +  -  +  HPC  Protozoa  Microbial variables Flow Depth Rainfall Conductivity DO NOx PO4 NH3 DOC AFDW Water Temperature  +  HPC  Lexan  Sal +  - + +  Sal  +  + -  AFDW  Slate Rock  ENT +  -  FC + EC + +  HPC  Wood  - + -  Campy +  AFDW  Sediment  Long-term samplers (12 and 24 week colonization)  Microbial variables Flow Depth Rainfall DO Conductivity NOx PO4 NH3 DOC AFDW Water Temperature Protozoa  Variable  Substratum  Short-term samplers (1-3 weeks colonization)  FC + EC +  -  +  +  +  -  +  -  +  -  +  -  +  -  +  -  +  +  -  +  + -  77  3.4  Discussion  Agriculture and rainfall amounts clearly affected water quality in the Elk Creek watershed. The highest concentrations of nitrate, phosphate, ammonia and dissolved organic compounds were found at the impacted sites in the agricultural watershed and concentrations were higher during the wet season than the dry season. However, the fluctuations in conductivity, nutrient concentrations and dissolved oxygen between sites 1 and 4 indicate that the pollution levels did not just constantly increase as Hope Slough was approached. Ford Creek (another impacted stream which merges with Elk Creek between sites 3 and 4) contributed to within-stream nutrient loading and pathogen concentrations, however the observed differences in the present study between these sites were minimal. Table 3.2 highlights some expected natural variations in water quality, as well as some fluctuations that are less explicable. For example, dissolved oxygen levels were highest in the wet season at all sites. This is due to decrease in Brownian motion of oxygen at lower temperatures, which results in its decreased release from the water into the atmosphere (Schwedt, 2001). However, the reason behind the decrease in percentages of air saturation at sites 2 and 3 in the wet season, as the concentration of oxygen in the water increased is not clear. To be able to ascertain what caused that decrease, other inputs into the system would have needed to be monitored. The increase in dissolved organic carbon at site 1 in the forested headwaters in the dry season may be due to an increase in litter degradation processes (Cleveland et al., 2004), which tend to be high in forested areas. Phosphate levels at sites 1 and 4 were relatively stable in both seasons, while phosphate levels at sites 2 and 3 increased in the wet season. An increase in nitrate and ammonia levels in the wet season was also observed at all sites. While this increase could have been due to natural transformation processes in the aquatic ecosystem, it seems unlikely since the decrease in temperature levels in the wet season should slow down metabolic processes (Q 10 rule) (Montagnes et al., 2003). Thus, the observed increases in nutrient concentrations may be due to external inputs into the system (particularly runoff or septic tank leakage). The only way to ascertain the source of nutrient fluctuations in this system would be through the monitoring of external inputs. Since survival and inactivation of indicator organisms and pathogens is dependent on physical, chemical, and biological water quality characteristics (Anderson et al., 2005), the differences in the values of water quality variables at sites 1 to 4 were useful for the examination of  78  relationships between these variables and microbial indicators and pathogen numbers in biofilms. Positive correlations between indicator concentrations and flow measurements were observed in water samples. Stream flow correlated with rainfall in many cases at several sites. Shehane et al. (2005) also reported a positive relationship between rainfall, stream flow and indicator organism concentrations in water samples in their study. In the present study, sediment results, as well as one-month biofilm results, showed similar positive relationships between rainfall and indicator and pathogen numbers (particularly on slate rock, wood and sand paper). It was interesting to note that the majority of the one-month biofilm indicators and pathogens data did not exhibit relationships with nutrient concentrations in the water column (compared to short-term samplers). Often biofilm communities, particularly mature ones, are less affected by external environment nutrient changes (such as N or P enrichment) since nutrient demand is being met from within the biofilm (Tank and Dodds, 2003). The relationships between dissolved oxygen levels and microbial variables varied substantially based on substratum type and indicator organism. Positive relationships may be due to the need for oxygen to diffuse through the biofilm matrix to supply residing organisms and to maintain extracellular polymeric substance strength (EPS). The negative relationships may be due to the fact that oxygen levels decrease as pollution increases through runoff. Runoff also tends to carry microorganisms that could potentially be recruited into the biofilm. Water depth exhibited varying effects on pathogens and indicators. For example, depth affected ENTs on slate rock, Lexan and sandpaper positively. In these cases, depth is probably an inverse surrogate for solar radiation, which is known to inactivate microorganisms. This corresponds with other inactivation research (Noble et al., 2004; Davis et al., 2005; Whitman et al., 2004). The commonly observed negative relationship between FC and salinity (salinity is the concentration of salts in the water column measured in mg/L or ppm) (Shehane et al., 2005) was not found in the present study due to the very small variation in salinity in this watershed. There was, however, one relationship involving salinity, which was the positive relationship between Salmonella and salinity in wood biofilms. Salinity may have been another surrogate for agricultural intensity in this case, since crop agriculture often results in soil salinity (Flowers and Yeo, 1995). It is, however, most likely that the relationship between Salmonella and salinity is just a statistical error (significant by chance), since salinity was mostly constant in Elk Creek. Interestingly, Salmonella in biofilms also exhibited negative relationships with water depth, but positive relationships with PO4 levels (a relationship also observed by Campylobacter and E. coli O157 in sediment). While phosphate levels may reflect the amount of runoff that reached the  79  stream, the negative relationship with depth is not easy to explain. Studies have found that Salmonella survives better in cool water temperatures (Guan and Holley, 2003) with little solar radiation. These conditions would be met better in deeper waters. Perhaps, Salmonella is not a good competitor in the presence of other microbial organisms and thus flourishes better in conditions that do not support the growth of others. This may explain the reverse peaks observed between Salmonella and Campylobacter on some substrata. Both positive and negative relationships between water temperature and microbial concentrations were found. Water temperature is hypothesized to have major effects on microbial decay and inactivation (Noble et al., 2004; An et al., 2002). Temperature is also known to have major effects on the growth rates of both bacteria and protozoa (Q10 Microbial Rule) (Montagnes et al., 2003). While the present study found some positive correlations between indicator numbers and water temperature, the results corresponded with Romani and Sabater (2000), who found that the effects of water temperature on biofilms in streams were far less relevant than nutrient conditions, particularly when flow stops. Romani and Sabater (2000) also suggested that conductivity could be used as an indirect measure of discharge changes in streams. This may be the case in the present study as well. In this study, conductivity mostly had positive relationships with microbial indicators and pathogens in water and biofilms, especially at site 1. At site 4, the relationship frequently became negative, especially in the case of ENTs. Finally, stream flow seemed to exhibit negative relationships with indicator organisms (such as FC and ENT), but a positive correlation with pathogens (such as Campylobacter) in one-month samplers. Stream width increased as the stream flowed towards Hope Slough and was more exposed to agricultural activities. While we cannot be certain what width was a covariate for, due to its constantly negative effect on microbial variables it may have been a surrogate for toxic chemicals, antibiotics, pesticides, and/or pollutant with detrimental effects on the microbial communities in the stream. It may also be worth mentioning that some microbial variables on particular substrata were positively correlated with conductivity, nutrients or water velocity at site 1, but correlated negatively with the same variable at site 4. This may imply that biofilms growing in more nutrient-rich environments, such as site 4, grow thicker faster and become more vulnerable and likely to slough off. Our results also showed that the optimal colonization period for biofilm sampling would be between two and four weeks. Hunt and Parry (1998) suggested that biofilm studies examining surface colonization should allow for exposure periods of 14 days to one month to insure that a  80  mature biofilm community was reached. The authors also suggested that after a 14-day period initial colonization often reached a plateau. This was not observed in the present study, particularly not on the hydrophobic surface Lexan. The increase in indicator numbers until week 4 may be related to the enhanced ability of bacteria to attach to hydrophobic surfaces (Cerca et al., 2005). The results also showed that substrata with high organic matter content, such as wood or sediment would accumulate the highest numbers of indicators and pathogens. On all substrata (with the exception of sediment and wood), concentrations of HPCs initially increased, then decreased, to finally increase again. This may be the result of broad and fine spectrum competition that Jackson (2003) suggested occurs at the beginning of colonization. Wood and sediment may have been particularly different from other substrata (surface area, amount of nutrients available) allowing for different growth to occur. It is also interesting to note that with the exception of sediment, other examined biofilms exhibited higher concentrations of FC than ENT at the beginning of the study. ENT concentrations often exceeded FC concentrations by week 24 for all substrata except Lexan. Studies have shown that FC have lower decay rates in freshwater environments compared to ENT and are therefore better faecal contamination indicators in freshwater (Anderson et al., 2005). Considering that ENT are far better at resisting environmental stressors than FC (Yates, 2007), perhaps the increase in concentration after week 4 is an indication of better ability to utilize resources in the biofilm. However, the observation may be based on the fact that water (the indicator and pathogen providing medium) also exhibited higher levels of enterococci starting at week 4. Another interesting observation is that water (and river rock) carried no FC at week 4, while all other substrata did. Wood and sediment contained the highest concentrations of E. coli throughout the study period, with different substrata peaking at different points in the colonization period. Finally, the phenomenon of reverse pathogen peaks in all substrata except for wood and sediment is interesting and may be related to fine scale competition amongst pathogens. It is likely that short- and long-term sampler indicator and pathogen number correlations with water quality variables were a reflection of the different stages of succession the biofilms had reached. If so, short–term biofilms may have exhibited positive relationships with water quality variables that supplied the biofilm with more organisms, such as velocity and flow. The depth of water was, however, negatively correlated with microbial variables and AFDW perhaps because the volume of water in the creek would have influenced the probability of bacterial attachment. The negative relationship between microbial variables and conductivity may be related to that finding, since conductivity decreases with increased water volume. AFDW often correlated  81  positively with indicator numbers in the first three weeks of colonization. This makes sense since bacterial mass would contribute to the calculated AFDW. In most cases, dissolved oxygen exhibited positive correlations with microbial variables. This could be due to the dependence of the biofilm community on oxygen diffusion through the extracellular polymeric substance (EPS) to maintain biofilm integrity and EPS strength. This may also be the reason nutrients had a positive correlation with microbial variables in biofilms (with the exception of phosphate and ENT in sediment and river rock biofilms). Wood short-term biofilms exhibited the highest correlations between pathogen frequencies and water quality variables. Long-term biofilm relationships with water quality variables reflected a very different microbial community. The negative relationships between microbial variables and rainfall and stream width may be related to the fact that older biofilms tend to be thicker, weaker and less oxygenated and thus flow changes may result in sloughing events. The negative relationship microbial variables exhibited with nutrients, such as ammonia and nitrates, may be related to two issues. Studies have shown that sloughing commonly occurs in nutrient rich environments and thick biofilms (Ohashi and Harada, 1996). Also, nutrient presence in the water column may be a reflection of runoff events, which are usually accompanied by flow changes. Enterococci exhibited some relationships that were more challenging to explain and require further investigation. It would have been interesting to compare wet season long and short-term variations in relationships between indicator and pathogen numbers and water quality variables with those in the dry season, when biofilms become the main sinks for pathogens. Positive relationships observed between amoebae, flagellates, ciliates and total protozoa counts, and measured microbial variables are interesting. Caution must be taken when interpreting these results since the method used here is not standardized and does not give a good representation of the range of protozoa in the environment. Relationships between water microbial variables and protozoa were negative in the short-term samplers. That was not the case in biofilms. The relationships in the short-term biofilm community were mostly positive (with the exception of ciliates in wood biofilms). This remained the case in long-term biofilm samplers as well. Negative correlations between protozoa and indicators and pathogens are usually explained through the most obvious relationships between bacteria and protozoa: predation (Huws et al, 2005). However, in this study more positive than negative relationships between protozoa numbers, and indicator organisms and pathogens were observed. This could be related to several phenomena reported in the literature. In some cases, protozoan predators suppress competitors,  82  and thus allow less competitive members of the biofilm community to flourish. Simek (1992) observed that protozoa preferred to feed on larger, more successful bacteria in pelagic communities. Also, different bacteria evolve mechanisms to avoid predation. For example, Salmonella enterica serovars have evolved different O-antigen variability on their cell surfaces, which affect whether or not intestinal amoebae within the host will graze upon them (Matz and Kjelleberg, 2005). Also, existing in a biofilm would allow bacteria with beneficial genotypes in the community to exchange their genes more easily (Bryers, 2000). There have been some studies that documented the phenomenon of microcolonies in studies of Pseudomonas spp. grown with flagellates. At early stages of biofilm development, prey bacteria stuck together and formed structures that were outside the size range of predator flagellates (Hahn et al., 2000). At later stages of biofilm maturation, Pseudomonas aeruginosa cells up-regulated a lethal factor, which produced a chemical compound toxic to flagellates in the same culture (Matz et al., 2004). Quorum sensing would also facilitate cell-to-cell communication, which would result in better defense responses in a biofilm community (Matz and Kelleberg, 2005). Finally, a phenomenon that may be resulting in positive relationships among protozoa and pathogens is the ability of protozoa to internalize other unicellular organisms. Internalization refers to the ability of many intracellular pathogens to survive within a protozoan host when environmental conditions are harsh. Results by Snelling et al. (2005) showed that Campylobacter jejuni was internalized by waterborne Tetrahymena and Acanthamoebae (both protozoa genera were found in our water samples). Similar results have been found by TeczanMerdol et al. (2004) when investigating ability of three Salmonella enterica serovars to be internalized by five species of Acanthamoebae in gentamycin protection assays. Also, Salmonella internalization by Acanthamoebae was reported by King et al. (1988) during chlorination processes. However, it is also possible that protozoa counts were covariates for unmeasured independent variables in the stream; since we know that protozoan predation increases in eutrophic environments (Noble et al., 2004). The study has a variety of limitations worth discussing. Primarily, the sample size of this study was fairly small, particularly in the case of short and long-term biofilm observations. This is why presented statistical correlations need to be interpreted with caution. Secondly, explaining some of the correlations between water quality variables and microbial variables in biofilms may be difficult due to three factors. The water quality variables measured in the present study were collected at discrete points in time and are only snapshots of the actual chemical, physical and  83  biological changes colonizing biofilms were exposed to. This is particularly problematic in the case of long-term biofilms correlations with water quality variables. Also, biofilms exhibit wide ranges of metabolic, physical and chemical heterogeneities within the EPS itself, thereby making the environment within the biofilm different from that in the external water column. For example, the dense aggregation of microorganisms leads to physical and environmental changes within the same biofilm that can be quite dramatic over a couple of micrometers. These broad ecological microniches allow for the formation of very complex microbial communities (Flemming et al., 1999; Jackson, 2003). The third issue that makes interpretation of our results difficult is that some of the measured variables may be covariates for other variables that were not measured. We can only attempt to explain the relationship among these covariables and the microbial concentrations measured based on our knowledge of the system and previous findings of the decay and inactivation literature. For example, we suggested that depth might be a covariate for solar radiation. Some of the variables we measured were confounded by each other, such as stream flow, velocity, depth and width. Future research could try to take into account inputs from the surrounding areas in the watersheds as well as dominant land-uses around sites to be able to interpret the water quality data better. Also, when evaluating sediment results, the reader should keep in mind that sediment colonization period, particle size, and particle within-stream transport was not taken into account in the present study. This study requires replication in other watersheds to see if any of these relationships hold. The lack of information about other variables related to indicator and pathogen survival and inactivation in biofilms was another limitation of this study. There was no measure of algal biomass. Also, more precise estimates of protozoa concentrations in the water column and the biofilm itself could have been useful. Some physical factors, such as solar radiation and turbidity, could have had strong effects or correlations with faecal indicator and pathogen concentrations in the watershed but were not measured. Also, this study does not include estimates of viral lysis and concentrations of toxic chemicals, organic contaminants, and heavy metals, which could affect biofilm community structure and dynamics.  3.5  Conclusions  In this study, the environmental factors associated with the distribution of indicator organisms and pathogens on different surface types in an agricultural stream were examined. The results showed that the relationships between water quality and microbial variables in one-month  84  biofilms were different on different substrata. Also, some relationships between water quality and microbial variables changed among sites on the same substratum. For example, some microbial variables were positively correlated with conductivity, nutrient concentrations (such as nitrate and dissolved organic carbon) or water velocity at site 1, but correlated negatively with the same variable at site 4. This may imply that biofilms growing in eutrophic environments, like site 4, grow thicker faster and become more likely to slough off. Colonization period (short-term and long–term) also affected the distribution of the indicator bacteria and pathogens in biofilms on different surface types. The study indicated that a two to four week period of colonization might be optimal to study indicator and pathogen distribution in biofilms. The study also showed that wood and sediment performed best when it came to the accumulation of pathogens and indicators. Finally, relationships amongst water quality variables and indicator and pathogens numbers changed depending on the stage of maturation (or colonization) the biofilm was undergoing. In young biofilms (colonized substrata for 1 to 3 weeks), indicator and pathogen numbers correlated directly with flow and nutrients, but had negative relationships with depth and conductivity. In older biofilms (colonized 12 or 24 weeks), indicator and pathogen numbers started showing negative correlations with nutrients and rainfall, but positive correlations with protozoa and dissolved oxygen. The study highlights the importance of considering substratum type, water quality variables and colonization period when investigating the distribution of indicators and pathogens in biofilm in natural aquatic environments.  3.6 Acknowledgements This research has been funded by the CIHR Strategic Training Program in Public Health and the Agricultural Rural Ecosystem (PHARE) and Partner Institutes including the Institute of Cancer Research, Institute of Circulatory and Respiratory Health, Institute of Infection and Immunology, and the Institute of Population and Public Health. We would also like to acknowledge the School of Environmental Health where all the microbiological lab analyses were completed. Finally, we would like to thank the University of British Columbia Environmental Engineering Laboratory, particularly Paula Parkinson, for conducting the water nutrient analyses.  85  3.7 References An, Y.J., Kampbell, D.H., Breidenbach, G.P., 2002. Escherichia coli and total coliforms in water and sediment at lake marinas. Environmental Pollution 120, 771-778. Anderson, K.L., Whitlock, J.E., Harwood, V.J., 2005. Persistence and differential survival of faecal indicator bacteria in subtropical waters and sediment. Applied and Environmental Microbiology 71(6), 3041-3048. APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater, 20th Edition. American Public Health Association (APHA), American Waterworks Association (AWWA), and Water Environment Association (WEF). Washington, DC. Bartram, J., Howard, G., 2003. Drinking water standards for the developing world. Emerging waterborne pathogens. In: Mara, D., Horan, N. (Eds.) Handbook for Water and Wastewater Microbiology. Elsevier Press, London, 221-241 Bryers, J.D. 2000. Biofilms: An introduction. In: Bryers, J.D. (Ed.) Biofilms II: Process Analysis and Applications. New York Wiley-Liss, 3-12. Byrd, J., Xu, H., Colwell, R., 1991. Viable but nonculturable bacteria in drinking water. Applied and Environmental Microbiology 57 (3), 875–878. Cerca, N., Pier, G.B., Vilanova, M., Oliveira, R., Azeredo, J., 2005. Quantitative analysis of adhesion and biofilm formation on hydrophilic and hydrophobic surfaces of clinical isolates of Staphylococcus epidermis. Research in Microbiology 156, 506-514. Cleveland, C.C., Neff, J.C., Townsend, A.R., Hood, E., 2004. Composition, dynamics, and fate of leached dissolved organic matter in terrestrial ecosystems: results from a decomposition experiment. Ecosystems 7, 275–285. Davis, K., Anderson, M.A., Yates, M.V., 2005. Distribution of indicator bacteria in Canyon Lake, California. Water Research 39, 1277-1288. Decho, A.W., 2000. Microbial biofilms in intertidal systems: an overview. Continental Shelf Research 20, 1257-1273. Dorner, S.M., Huck, P.M., Slawson, R.M., Gaulin, T., Anderson, W.B., 2004. Assessing levels of pathogenic contamination in a heavily impacted river used as a drinking water source. Journal of Toxicology and Environmental Health: Part A, 67, 1813-1823. Emtiazi, F., Schwartz, T., Marten, S.M., Krolla-Sidenstein, P., Obst, U., 2004. Investigation of natural biofilms formed during the production of drinking water from surface water embankment filtration. Water Research 38 (5), 1197-1206.  86  Flemming, H.C., Wingender, J., Moritz, R., Borchard, W., Mayer, C., 1999. Physico-chemical structure of biofilms – A short review. In: Keevil, C.W. Godfree, A., Holt, D., Dow, C. (Eds.) Biofilms in the Aquatic Environment. Royal Society of Chemistry, Cambridge, 1-12. Flowers, T.J., Yeo, A.R., 1995. Breeding for salinity resistance in crop plants: where next? Australian Journal of Plant Physiology 22 (6), 875 - 884 Glasner, A., McKee, L., 2002. Pathogen occurrence and analysis in relation to water quality attainment in San Francisco Bay Area watersheds. San Francisco Estuary Institute Report, San Francisco, USA. Retrieved on January 29, 2008. Document available at: http://www.sfei.org/watersheds/reports/Pathogens/Pathogenlitrevv6.pdf Guan, T. Y., Holley, R. A., 2003. Pathogen survival in swine manure environments and transmission of human enteric illness: a review. Journal of Environmental Quality 32, 383-392. Hahn, M.W., Moore, E.R.B., Hofle, M.G., 2000. Role of microcolony formation in the protistan grazing defense of the aquatic bacterium Pseudomonas sp. MWH1. Microbial Ecology 39 (3), 175-185. Hunt, A.P., Parry, J.D., 1998. The effect of substratum roughness and river flow rate on the development of a freshwater biofilm community. Biofouling 12 (4), 287-303. Huws, S.A., McBain, A.J., Gilbert, P., 2005. Protozoan grazing and its impact upon population dynamics in biofilm communities. Journal of Applied Microbiology 98 (1), 238–244. Jackson, C.R., 2003. Changes in community properties during microbial succession. Oikos 101 (2), 444-448. King, C. H., Shotts, E. B. Jr., Wooley, R. E., Porter, K. G., 1988. Survival of coliforms and bacterial pathogens within protozoa during chlorination. Applied and Environmental Microbiology 54, 3023-3033 Maal-Bared, R., Bartlett, K.H., Bowie, W.R., Hall, E.R., Hall, K.J., 2007. Seasonal and spatial distribution of indicators and pathogens in biofilms in an agricultural watershed. Annual Conference Proceedings of the American Waterworks Association, Toronto 2007. Matz, C., Bergfeld, T., Rice, S.A., Kjelleberg, S., 2004. Microcolonies, quorum sensing and cytotoxicity determine the survival of Pseudomonas aeruginosa biofilms exposed to protozoan grazing. Environmental Microbiology 6, 218–226. Matz, C., Kjelleberg, S., 2005. Off the hook-how bacteria survive protozoan grazing. Trends in Microbiology 13 (7), 302-307.  87  Montagnes, D.J.S., Kimmance, S.A., Atkinson, D., 2003. Using Q10: can growth rates increase linearly with temperature? Aquatic Microbial Ecology 32, 307-313. Noble, R.T., Lee, I.M., Schiff, K.C., 2004. Inactivation of microorganisms from various sources of faecal contamination in seawater and freshwater. Journal of Applied Microbiology 96, 464-472. Ohashi, A., Harada, H., 1996. A novel concept for evaluation of biofilm adhesion strength by applying tensile force and shear force. Water, Science and Technology 34 (5-6), 201211. Ouyang, Y., Nkedi-Kizza, P., Wu, Q.T., Shinde, D., Huang, C.H., 2006. Assessment of seasonal variation in surface water quality. Water Research 40, 3800-3810. Patterson, D.J., 2003. Free-Living Freshwater Protozoa: A colour guide. ASM Press, Washington, DC. Romani, A.M., Sabater, S., 2000. Variability of heterotrophic activity in Mediterranean stream biofilms: A multivariate analysis of physical-chemical and biological factors. Aquatic Sciences 62 (3), 205-215. Rood, K.M., Hamilton, R.E., 1995. Hydrology and water use for salmon streams in the Chilliwack/Lower Fraser habitat management area, BC. Prepared for the Fraser River Action Plan by Northwest Hydraulic Consultants Ltd., North Vancouver, BC. Schwedt, G., 2001. The essential guide to environmental chemistry. Wiley Press, West Sussex, UK. Shehane, S.D., Harwood, V.J., Whitlock, J.E., Rose, J.B., 2005. The influence of rainfall on the incidence of microbial faecal indicators and the dominant sources of faecal pollution in a Florida river. Journal of Applied Microbiology 98, 1127-1136. Simek, K., 1992. Direct and indirect evidence of size selective grazing on pelagic bacteria by freshwater nanoflagellates. Applied and Environmental Microbiology 58, 3715-3720. Snelling, W. J., McKenna, J. P., Lecky, D. M., Dooley, J. S. G., 2005. Survival of Campylobacter jejuni in waterborne protozoa. Applied and Environmental Microbiology 71, 5560-5571. Stoodley, P., Wilson, S., Hall-Stoodley, L., Boyle, J. D., Lappin-Scott, H.M., Costerton, J.W., 2001. Growth and detachment of cell clusters from mature mixed-species biofilms. Applied and Environmental Microbiology 67, 5608-5613. Tank, J.L., Dodds, W.K., 2003. Nutrient limitation of epilithic and epixylic biofilms in ten North American streams. Freshwater Biology 48, 1031-49.  88  Tezcan-Merdol, D., Ljungstrom, M., Winiecka-Krusnell, J., Linder, E., Engstrand, L., Rhen, M., 2004. Uptake and replication of Salmonella enterica in Acanthamoeba rhysodes. Applied and Environmental Microbiology 70, 3706-3714 U.S. Environmental Protection Agency, 2003. National primary drinking water regulations: Long term 2 enhanced surface water treatment rule; Proposed rule. Federal Regulation 68 (154), 47640–47795. Vrana, B., Allan, I.J., Greenwood, R., Mills, G.A., Dominiak, E., Svensson, K., Knutsson, J., Morrison, G., 2005. Passive sampling techniques for monitoring pollutants in water TrAC - Trends in Analytical Chemistry 24(10), 845-868. Yates, M.V., 2007. Classical indicators in the 21st century-far and beyond the coliform. Water Environment and Research 79(3), 279-86.  89  4  4.1  Campylobacter spp. distribution in Elk Creek British Columbia: improving sampling techniques3  Introduction  Campylobacter is a gram negative, microaerophilic bacterium that causes acute gastro- intestinal illness worldwide (Moore et al., 2005). Twelve of the 15 identified Campylobacter species cause disease in humans, and C. jejuni and C. coli cause the majority of those outbreaks (Moore et al., 2005). Campylobacter is of particular interest to public health practitioners for the following reasons. In Canada and the US, Campylobacter is the leading recognized cause of bacterial enteritis. The bacterium caused 24 outbreaks in Canada between 1974 and 2001 (Schuster et al., 2005), and 26 outbreaks in the United States between 1971 and 1996 (Leclerc et al., 2002). Campylobacter is estimated to cause 5%-14% of diarrhoea worldwide (Coker et al., 2002). Campylobacter has a low infective dose (approximately 800 bacteria). This is comparable to the infective dose of pathogens like Shigella sp., Salmonella typhi and E. coli O157:H7, and much lower than the infective doses of non-typhoidal Salmonella (105-1010 bacteria) and Vibrio cholerae (108 bacteria) (Ekdahl et al., 2005). Campylobacter outbreaks display a clear seasonal pattern with highest incidence occurring in the summer months (Nylen et al., 2002; Patrick et al., 2005). The highest levels of environmental Campylobacter prevalence, however, occur in fall, winter and early spring (Flanders and Yildiz, 2004; Clark et al., 2003). This discrepancy raises many questions about our understanding of the bacterium’s routes of transmission. The majority of Campylobacter cases are related to consumption of contaminated poultry, consumption of contaminated raw milk or water, and direct contact with pets and farm animals (Ekdahl et al., 2005). Nylen et al. (2002) suggest that these routes of transmission only explain 50% of cases. Clark et al. (2003) suggest that about 20% of Campylobacter jejuni infections can be attributed to transmission routes other than food, which include water exposure. Continuing agricultural expansion, increasing contact with 3  A version of this chapter will be submitted for publication. Maal-Bared, R., Bartlett, K.H., Bowie, W.R., Hall, E.R., Hall, K.J. 2007. Campylobacter sp. Distribution in Elk Creek British Columbia: improving sampling techniques. Journal of Water and Health.  90  farm animals (known reservoirs for Campylobacter) (Moore et al., 2005), and our limited understanding of transmission routes are likely to result in more outbreaks in the future. Prevalence of Campylobacter in environmental samples, and particularly freshwater, has been linked to warmer temperatures and extreme rainfall (Thomas et al., 2006). Climate change is expected to increase extreme rainfall events, thereby potentially increasing the incidence of campylobacteriosis. Animals (human and avian populations, pigs, cattle, dogs, cats, waterfowl, and small mammals on farms) play an important role as Campylobacter sp. reservoirs. Thus, the bacterium reaches most surface water through faecal contamination (Leclerc et al., 2002). There have been many studies investigating the prevalence, as well as the spatial and temporal distribution of Campylobacter in natural aquatic systems (Kuusi et al., 2005; French et al., 2005; Kemp et al. 2005; Brown et al., 2004). The presence of Campylobacter sp. in biofilms in the environment has, however, rarely been investigated (Flanders and Yildiz, 2004). Biofilms are attached multi-species microbial communities that are surrounded by an extracellular polymeric substance (EPS). EPS benefits the attached community through concentration of nutrients, genes and extracellular enzymes, and protection from external stressors. Buswell et al. (1998) investigated Campylobacter survival in biofilms and found that survival times in biofilms for some species of Campylobacter almost doubled compared to survival times in water microcosms. However, Campylobacter was still strongly affected by oxygen levels and temperature. Also, many studies point out the potential biofilms could play as reservoirs of disease (Flanders and Yildiz, 2004). Studies in water distribution systems have shown that pathogens such as Vibrio cholerae and Salmonella enterica serotype Typhi are not likely to survive in biofilms (Rittmann, 2004). Campylobacter jejuni, however, can survive and proliferate in biofilms in water distribution systems (Rittmann, 2004; Lehtola et al., 2006). The purpose of the present study was to investigate the presence of Campylobacter sp. in biofilms on a variety of surfaces and in water in an agricultural watershed to: a) explore if current water sampling protocols are sufficient to give a good representation of Campylobacter distribution in water sources, and b) investigate if any commonly monitored chemical, physical, or biological water quality parameters could be used as Campylobacter sp. or faecal contamination surrogates in water sources. River rock, slate rock, wood, Lexan™, sandpaper and sediment were chosen for investigation to represent a variety of substrate surfaces, whose microbial colonization and succession patterns may be differently affected by microbial, chemical or physical water quality parameters. These interactions between biofilm attachment surface and water characteristics  91  would affect Campylobacter recovery from a natural watercourse. The hypothesis was that due to biofilm community and ecology benefits, Campylobacter sp. might be more likely to be found in biofilms than in water column grab samples. Also, different support surface types exhibit varying amounts of surface tension, hydrophobicity, porosity, and surface area, which all affect biofilm thickness, EPS chemical characteristics, strength of biofilm attachment and sloughing rates. Thus, different surfaces might have different Campylobacter prevalences and correlate differently with other water quality parameters.  4.2  Methods  4.2.1 Evaluating substrata for Campylobacter spp. monitoring purposes Substrata were evaluated using the following criteria. 1. Prevalence of Campylobacter sp. in biofilms on the surface. 2. Correlation of Campylobacter sp. presence in the biofilm with concentration of faecal bacterial indicators in the biofilm. Faecal contamination indicator bacteria of choice were: heterotrophic plate counts (HPCs), faecal coliforms (FC), E. coli (EC), and enterococci (ENT). These indicators were chosen due to their common use in water quality testing. 3. Correlation of Campylobacter sp. presence with physical and chemical water quality characteristics. Physical water quality characteristics used were water temperature, flow, velocity, specific conductivity, and rainfall. Chemical water quality indicators were dissolved oxygen, nitrate + nitrite, ortho-phosphate, ammonia and dissolved organic carbon. These are parameters that are frequently included in water quality assessments. Protozoal counts were included as another variable of interest. 4. Cost and availability. This parameter was chosen due to its relevance in smaller and remote communities with limited resources. 5. Potential for standardizing a substratum colonization surface area. It was important to use substrata of similar (if not equal) sizes to be able to compare the data collected from different surfaces, sites or dates. Standardization would also facilitate cross-laboratory and cross-watershed comparisons. 6. Ease of biofilm removal using only sonication. This was essential when detaching the bacterial population in biofilms for analysis. 7. Probability of substratum loss in situ. Substrata were left in situ for two to four weeks to allow colonization. Some of these substrata, like river rocks, were very difficult to anchor  92  and retrieve. Thus, this category would also affect substratum choice for monitoring purposes.  4.2.2 Sampling location- Elk Creek The Elk Creek Watershed is located in the Fraser Valley in British Columbia (Canada). It has an area of about 28 km2 and the watercourse is about 12 km long (Rood and Hamilton, 1995). It is mainly agricultural in nature, although it also has forested headwaters located in the mountains, and limited urban residential influences. The watercourse experiences an increasing pollution gradient as it flows north and drains into the Fraser River at Hope Slough. Four sites were located on the stream for this study. The first site was located in the headwaters as a reference site. Sites 2, 3 and 4 were located in the agricultural area (Figure 4.1). For a more detailed description of the watershed see Maal-Bared et al. (2007).  Hope Slough  Site 4 Site 3  Nevin Creek  Site 2 Dunville Big Ditch  Creek  Ford Creek  Elk Creek  Site 1  North  Figure 4.1 Map of Elk Creek in Chilliwack, British Columbia, and location of sites 1 to 4 (marked by the triangles) (Scale about 1: 90,000, 16.7 x 11.7 km). Arrow indicates north.  4.2.3 Biofilm samplers Artificial substrata used for biofilm colonization were river rock, slate rock, wood, Lexan™, and fine grit sandpaper. Biofilm samplers were built using the method described in Maal-Bared et al. (2007).  93  4.2.4 Sample collection Samples were collected between December 2005 and December 2006, and only one site at a time was visited. In the wet season (October to April), 11 samplers were retrieved, while only six samplers were left in place and retrieved in the dry season (May- September). Every sampler was left at a site for four weeks before collection and analysis, since other investigations have shown that a mature biofilm community requires somewhere between 2 weeks and one month for formation (Hunt and Parry, 1998). At the time of sampler collection, duplicate water column grab samples of river water and wet sediment samples were collected. In the laboratory, 500 mL of sterile Phosphate Buffer Saline (PBS, pH=7.2, 9 g NaCl/L, 0.0067M PO4) suspension was used for substrata and sediment analysis. Water was analyzed directly. For a detailed description of the sample collection and processing used for substrata, water and sediment, see Maal-Bared et al. (2007).  4.2.5 Campylobacter spp. analysis To determine presence of Campylobacter sp., 40 ml of PBS suspension was tempered in peptone water for 4-5 hours at 37oC (1:9 ratio). Then a 10 ml sample of the tempered medium was added to an enrichment broth for 8 hours or overnight at 37oC (1:1 enrichment ratio). The enrichment broth was made up of 475 ml of Brucella broth (BBL™ BD Microbiological Systems, Sparks, MD) with one vial of Preston Campylobacter Supplement (Oxoid™, Basingstoke, Hampshire, UK, SR0117) and 25 ml of defibrinated horse blood (PML Microbiologicals, Wilsonville, OR). Finally, the enriched suspension was plated onto Campylobacter isolation plates (475 ml Campylobacter base agar (Oxoid™ Basingstoke, Hampshire, UK) with 25 ml defibrinated horse blood and one vial of Preston Supplement). The plates were incubated at 42.5oC for 48 hours in microaerophilic conditions using a BBL™ Campy Pak without the palladium catalyst (BBL™ BD Microbiology Systems, Sparks, MD). Colonies conforming to Campylobacter morphology were tested using the Oxoid™ DrySpot test (Dryspot Campylobacter Test Kit DR0150), which identifies five species of the Campylobacter genus (C. jejuni, C. lari, C. coli, C. fetus, and C. upsaliensis). It does not separate the species from each other. Throughout all tests, sterile distilled water was used as a negative control, while an ATCC C. jejuni control culture (ATCC 29428) was used as a positive control.  94  4.2.6 Sample analysis Standard Methods procedures were used for the enumeration and isolation of indicator organisms (APHA et al., 1998). Heterotrophic plate counts (HPC) were used as indicators of heterotrophic activity. Faecal coliforms (FC), E. coli (EC) and enterococci (ENT) were used as faecal contamination indicators. For a detailed description of the indicator organism and pathogen analysis methods, see Maal-Bared et al. (2007).  4.2.7 Standardizing across different substrata For a detailed description of the loss on ignition method used to approximate ash-free dry weight (AFDW) of biofilms on each substratum, as well as AFDW in water and sediment, see MaalBared et al. (2007).  4.2.8 Water quality parameters During every sampling trip, physical and chemical water quality parameters were measured. The measured parameters were: water depth (cm), water width (m), water velocity (m/s), stream flow (m3/s), water temperature (oC), conductivity (µS/cm), specific conductivity (µS/cm), and dissolved oxygen (mg/l), nitrate+nitrite (NOx) (mg N/L), ortho-phosphate (PO4) (mg N/L), ammonia (NH3) (mg N/L), dissolved organic carbon (DOC) (mg N/L) and protozoa counts. For a detailed description of the water quality parameter measurement methods, see Maal-Bared (2008, p. 61).  4.2.9 Statistical Analyses All statistical analyses were completed using JMP 7.0 (SAS Institute, NC, USA). To determine whether there were statistically significant differences between different substrata types, a Chisquare test was used. To find the relationships between Campylobacter presence and different water quality parameters, Spearman Correlation tests were used. Alpha was set at 0.05.  4.3  Results and Discussion  4.3.1 Campylobacter prevalence on different substrata The prevalence of Campylobacter sp. retrieved from all surfaces during the wet season (October to beginning of April) and the dry season (May-September) in biofilms on different substrata is  95  shown in Table 4.1. Amongst the substrata that performed well in accumulating Campylobacter sp. were sediment, slate rock, wood and Lexan. Water column samples performed poorly in comparison, only being positive 8% of the time during the wet season. Sandpaper did not accumulate Campylobacter at all. The differences amongst substrata were statistically significant in the wet season χ2 (p= 4.38 x 10-8, df=11). No Campylobacter sp. was recovered in the dry season.  Campylobacter sp. average prevalence between December 2005 and December 2006 during both the wet and dry season was 9%. The prevalence observed here is lower than that of other studies using isolation and culturing techniques to examine the prevalence of Campylobacter sp. in water samples. For example, Nam et al. (2005) found that 30% of water samples tested by PCR were positive for Campylobacter, but only 68% of those samples (20% of all samples) were positive using culture techniques. Another study by Brown et al. (2004) found that only 15% of water samples collected were positive for C. jejuni using PCR, and only 17% of water samples were positive for Campylobacter coli. In general, authors suggest that using culturing techniques to detect environmental Campylobacter sp. is not sufficient and underestimates the Campylobacter burden in a water source (Kemp et al., 2005; Lehtola et al., 2006; Waage et al., 1999). It is proposed that Campylobacter sp. reverts to a viable, non-culturable state when under environmental stress. Organisms in this state cannot be detected using traditional culturing techniques and their role in infection and disease remains debated (Thomas et al., 1999). Other studies have critiqued the steps involved in culturing Campylobacter sp. Work by Abatay and Corey (1997) suggests that pre-enrichment steps used in the isolation of thermotolerant Campylobacters selectively enrich Arcobacter sp. instead and do not enhance Campylobacter recovery. Other work by Abulreesh et al. (2005) indicates that filtering water samples can reduce Campylobacter recovery rates. While using PCR might have reduced the limit of detection for this present study and increased Campylobacter prevalence, the objective of the present study was to obtain an estimate of viable, culturable, and infective Campylobacter sp. in the watershed.  96  Table 4.1 Prevalence of Campylobacter sp. during the wet and the dry season in water and in biofilms on different surface types in Elk Creek, British Columbia.  Wet season  Dry season Total number of samples of substratum  Substratum  Prevalence  n  Prevalence  n  Water  8%  24  0%  12  36  River rock Slate rock Wood Lexan Sandpaper Sediment Total number of substrata tested per season  9% 22% 13% 13% 0% 27%  22 23 24 24 20 22  0% 0% 0% 0% 0% 0%  12 12 12 12 11 12  34 35 36 36 31 34  13%  159  0%  83  242  The complete absence of Campylobacter during the dry season may be related to several other issues besides the sensitivity of the culturing techniques used. As rainfall decreases in the dry season, less agricultural runoff reaches the stream, resulting in reduced pathogen loading. It also decreases nutrients in the stream, which may affect bacterial survival rates (see Figure 4.2). Also, farmers in the watershed use large amounts of organic fertilizer, such as dairy cattle slurry (mixture of cattle waste and water) and poultry manure, particularly in late October, before their spreading is limited by the Manure Management guidelines. Studies measuring Campylobacter sp. concentrations directly in liquid (slurry) or solid (farmyard waste) animal waste fertilizers have often found that livestock manure can carry high concentrations of Campylobacter (Brown et al., 2004; Hutchinson et al., 2005). Some authors suggest that the highest levels of manure application in the watershed occur between February and October in compliance with the Manure Management Guidelines of the British Columbia Ministry of Water, Land and Air Protection (Vingarzan et al., 2002). An assessment administered by the Ministry of Environment in early spring 2005, however, reported 40 manure spreadings, which were not in compliance with the manure spreading advisories in Chilliwack and Aggasiz in the lower Fraser Valley (Rushworth and Younie, 2006). There were also indications that manure had been spread on another 20 farms (mostly dairy farms) before the aerial investigation began during the advisory. About eighty-eight percent of farms were in compliance (Rushworth and Younie, 2006). This would partially explain why Campylobacter levels were high in the watershed between December 2005 and March 2006.  97  4.3.2 Campylobacter correlation with other microbial water quality indicator bacteria Figure 4.2 shows Campylobacter sp. prevalence in biofilms on different surfaces in the wet season and compares it to the numbers of HPCs, FCs, EC, and ENT. A Spearman correlation was used to study relationships between the prevalence of Campylobacter sp. within any particular biofilm and water column bacterial water quality indicators. No statistically significant relationships were found (all p>0.05). A relationship between those variables would have been very useful, since most water managers and public health inspectors already test the water column for indicator organisms. Another correlation anlaysis was performed to investigate relationships between numbers of indicator organisms (CFU/mL or MPN/mL) in biofilms and Campylobacter presence within the same biofilm during the complete sampling period. Presence of Campylobacter on slate rock and wood significantly correlated with ENT numbers, and Campylobacter presence in wood biofilms also significantly correlated with FC numbers (see Table 4.2). This may be of particular interest to public health officials, engineers and professionals that already monitor indicator bacteria in the water column, but want to increase their knowledge of Campylobacter presence in the watershed indirectly using faecal contamination surrogates. It is also interesting to note there was a trend of Campylobacter presence in slate rock biofilms with FC numbers (p=0.0753). Increasing the sample size in future research may be of interest to see if this relationship changes and/or becomes statistically significant. Table 4.2 Spearman coefficients and p-values for associations between Campylobacter sp. (Camp) presence in water, slate rock, wood and sediment and the average numbers of heterotrophic plate counts (HPC), faecal coliforms (FC), E. coli (EC) and enterococci (ENT) found in the same media between December 2005 and December 2006. Statistically significant relationships (p<0.05) are bolded. Water Slate Rock Wood Sediment by Spearman ρ p-value Spearman ρ p-value Spearman ρ p-value Spearman ρ p-value Variable variable NS 0.102 NS* -0.218 NS -0.418 0.084 -0.053 Camp HPC NS NS -0.087 0.430 0.075 0.196 0.627 0.005 Camp FC NS -0.187 NS -0.064 NS -0.065 NS 0.399 Camp EC NS 0.017 NS 0.377 Camp ENT 0.522 0.026 0.633 0.005 * NS stands for “not significant” and designates p-values that were much larger than 0.05.  98  30%  CFU/mg OM or MPN/mg OM  1000000  25%  100000 20%  10000 1000  15%  100 10%  10 1  5%  % Samples positive for Campylobacter  10000000  0.1 0%  Substratum  Sediment  Sand Paper  Lexan  Wood  Slate Rock  River Rock  Water  0.01  Left Scale: HPC CFU/mg EC MPN/mg Right Scale:  FC MPN/mg ENT MPN/mg Campy %  Figure 4.2 Comparing mean indicator numbers per mg of ash-free dry weight extracted from biofilm (n=11) and percentage of substrata positive for Campylobacter (n=22) across different surface types in the wet season.  There was a negative but statistically insignificant correlation between HPCs and Campylobacter prevalence in all biofilms. Previous biofilm work suggested that microaerophiles, such as Campylobacter jejuni, Helicobacter pylori, and Legionella pneumophila migrate to low redox zones in heterogeneous biofilms (Azevedo et al., 2003) that were created by the respiration of heterotrophic species (Keevil, 2003). We did not observe that relationship in our data. This may be related to several issues. High HPC concentrations lead to increased anoxic zones and waste products in the extracellular polymeric substance (EPS). This may decrease EPS strength and increase biofilm sloughing rates (Rittman, 1989). Also, mixed cultures are more susceptible to sloughing. This could either be due to the fact that bacteria depolymerise the EPS of other species, making them less stable, or because EPS and microbial surfaces are incompatible causing a weakness in binding (Brading et al., 1995). Our water column results correspond with studies that found no statistically significant correlation between Campylobacter presence in water and faecal coliform concentrations (Carter et al., 1987; Waage et al., 1999) or E. coli concentrations (Kemp et al., 2005). Vereen et al.  99  (2007) found a correlation between faecal coliforms and Campylobacter presence in the water column, but this study also found that Campylobacter levels peaked in the summer when rainfall was highest. This points out that different watersheds will experience different environmental characteristics, which will affect the Campylobacter population in the environment. This eliminates the possibility of creating “one-size-fits-all” monitoring programs, which we can apply under all environmental conditions. In our study, the relationship between E. coli and Campylobacter was negative, but not statistically significant, in all samples except for those from sediment. E. coli and Campylobacter have different cell surface characteristics (Bolster et al., 2006), which may explain differences in their attachment rates to different surfaces and survival strategies. This highlights that using E. coli as an indicator for Campylobacter presence in water sources may not be an option for public health officials and water managers.  4.3.3 Correlating Campylobacter prevalence in different biofilms with physical and chemical water quality characteristics Table 4.3 shows the Spearman correlation coefficients and p-values for Campylobacter prevalence in biofilms and some frequently measured water quality characteristics. Figure 4.3 shows the frequency of positive Campylobacter sp. samples and water quality parameter measurements over time during the study period. Only substrata that showed statistically significant relationships are shown in Table 4.3. Lexan exhibited a statistically significant relationship with protozoa counts in the water column, but with no chemical or physical water quality characteristics. The positive relationship between Campylobacter sp. presence in biofilms (except those on rocks) and water column protozoa is interesting. This may be related to the fact that some protozoa can internalize Campylobacter and offer the bacterium some protection (Snelling et al., 2005). The positive relationship may also be related to protozoan predation. Predation may reduce competition and allow Campylobacter sp. to survive in the biofilm longer before the next sloughing event occurs. The only negative, but insignificant, relationship between protozoa counts and Campylobacter presence was seen in slate rock biofilms. These results bring up more questions regarding the nature of the symbioses between protozoa and waterborne pathogens, such as Campylobacters, and what environmental factors trigger different symbioses, since the relationships cannot just be explained through predation (Parry, 2004).  100  Campylobacter presence in water was not correlated with any nutrients in the water column. This result was also observed by Thomas et al. (1999). However, Campylobacter presence in sediment and on wood was statistically correlated with nitrate concentrations in the water column (also see Figure 4.3). Both substrata also exhibited high Campylobacter prevalence. Thus, using nitrate concentrations in the water column as an early, indirect Campylobacter detection system may be a possibility, since many monitoring stations already measure nitrates. Table 4.3 Spearman coefficients and p-values for associations between Campylobacter sp. presence in water, slate rock, wood and sediment and water quality characteristics (velocity, temperature (Temp), dissolved oxygen (DO), nitrate + nitrite (NOx), ortho-phosphate (PO4), ammonia (NH3), dissolved organic carbon (DOC), rainfall, and protozoa counts (Prtz counts)). Statistically significant relationships (p<0.05) are bolded. Water Slate Rock Wood Sediment Variable Spearman ρ p Spearman ρ p Spearman ρ p Spearman ρ p * NS NS -0.324 NS -0.279 -0.285 -0.460 0.063 Velocity (m/s) NS NS NS Temperature -0.222 -0.138 -0.153 -0.278 NS (oC) NS NS NS -0.375 -0.212 -0.221 -0.463 0.061 DO (mg/L) NS 0.375 0.455 0.058 0.484 0.042 0.668 0.003 NOx (mg/L) NS NS NS 0.465 0.052 0.193 0.176 0.232 PO4 (mg/L) NS NS NS NS 0.069 0.386 0.386 0.379 NH3 (mg/L) NS NS NS NS -0.136 -0.145 -0.122 -0.285 DOC (mg/L) NS NS NS NS 0.068 0.004 0.007 -0.063 Rainfall (mm) NS Prtz counts (cell -0.033 numbers) 0.495 0.037 0.503 0.034 0.576 0.016 * NS stands for “not significant” and designates p-values that were much larger than 0.05.  Another interesting relationship is the negative association between Campylobacter prevalence and water velocity in all substrata in Table 4.3. Bolton et al. (1987) also reported that fast flowing sites in their study river had lower Campylobacter prevalences than slow flowing sites. This negative relationship may also explain the reverse peaks in Figure 4.3 when looking at rainfall and Campylobacter counts on January 10, 2006 (an extreme rainfall event resulting in more than 90 mm of precipitation over 48 hours). It is possible that high flow events decrease Campylobacter survival rates or encourage the organism to switch to a viable, non-culturable state. The negative relationship may also be the result of biofilm sloughing events. Biofilms were more likely to contain Campylobacter than water samples in this study. Thus, the loss of Campylobacter from biofilms during high flow events would affect Campylobacter recovery rates. As the number of extreme weather events increase with climate change, the risk of drinking water source contamination with Campylobacters will increase as well. If we are unable to detect Campylobacter during these events, public health is at risk if the water is used. Also, when  101  looking at Figure 4.3, we see that nutrients were still reaching the watershed during the extreme flow event. Since both nutrients and pathogen loading mainly results from the same source (agricultural runoff), the question becomes why Campylobacter was not detected and what are the effects of high velocity on Campylobacter populations in aquatic environments. Another factor affecting Campylobacter prevalence in biofilms and the water column is water temperature. Water temperatures in our study ranged from 9.8oC to 19.1oC. Table 4.3 shows the negative, but statistically insignificant, associations between temperature and Campylobacter numbers. That same effect can be seen in Figure 4.3. As temperature increased, Campylobacter survival declined. This relationship may explain the absence of Campylobacter from the watershed in the dry season. If the increase in water temperature and exposure to sunlight did not directly kill the bacterium, it may have converted it to a viable non-culturable state. Many studies show Campylobacter’s affinity for lower temperatures (Waage et al., 1999; Carter et al., 1987; Buswell et al., 1998; Rollins and Colwell, 1986). Some studies, like Vereen et al. (2007) did not observe the same temperature effects.  102  Wet season  Dry season  Wet season  200  Rainfall (mm)  150  100  03-Dec-06  05-Oct-06  23-Jul-06  05-Jul-06  25-May-06  16-May-06  08-Apr-06  30-Mar-06  22-Mar-06  W a ter T em p er a tu r e  D O (m g /L )  10  3  7 .5  2  5  1  2 .5  0  0 08-Apr-06  21-Feb-06  10-Jan-06  13-Dec-05  03-Dec-06  4  05-Oct-06  1 2 .5  23-Jul-06  5  05-Jul-06  15  25-May-06  6  16-May-06  1 7 .5  30-Mar-06  7  22-Mar-06  20  07-Feb-06  8  Water Temperature (oC)  Right Sc ale: Frequency of samples positive for Campylobacter sp.  D a te  Ca m py +  Dissolved Oxygen Concentrations (mg/L)  Lef t Sc ale:  21-Feb-06  10-Jan-06  13-Dec-05  0  07-Feb-06  50  D a te  0.25  0.15  5 4  0.1  3 0.05  2 1  Ammonia concetrations (mg/L)  0.2 6  Phosphate concentrations (mg/L)  7 Nitrate concentrations (mg/L)  Dissolved organic carbon concetrations (mg/L)  Number of samples positive for Campylobacter  8  0  03-Dec-06  05-Oct-06  23-Jul-06  05-Jul-06  25-May-06  16-May-06  08-Apr-06  30-Mar-06  22-Mar-06  21-Feb-06  07-Feb-06  10-Jan-06  13-Dec-05  0  Date  Lef t Sc ale: Right Sc ale:  Ca m p y + P O 4 (m g /L )  DO C (m g /L )  NO x (m g /L )  NH3 (m g /L )  Figure 4.3 Cumulative frequencies of positive Campylobacter sp. samples (Campy +) and water quality variables (rainfall (mm), temperature in oC, dissolved oxygen (DO), nitrate + nitrite (NOX), ortho-phosphate (PO4), ammonia (NH3), and dissolved organic carbon (DOC)) in Elk Creek (BC) between December 2005 and December 2006. Frequencies include Campylobacters recovered from water, biofilms and sediment. Dotted lines distinguish the dry from the wet season.  103  4.3.4 Substrata evaluation for Campylobacter spp. monitoring The final assessment of all substrata used in this study is presented in Table 4.4. Most value is placed on the first criterion. Sandpaper is of no interest to this study, because it exhibited a Campylobacter sp. prevalence of 0%. River rock and water exhibited low prevalences. Sediment, slate rock, wood and Lexan™ exhibited different prevalences ranging from medium to high. The final four criteria for assessment are related to financial and technical issues, and may be of more interest to small communities with limited resources and access to equipment. Cost and availability would mainly be an issue with Lexan™. Lexan is also likely to float away during the two to four week colonization period when biofilms grow on surfaces. Comparing all criteria presented below, we believe that sediment, slate rock or wood might be of high interest to researchers and public health inspectors interested in Campylobacter sp. monitoring in natural systems. The final choice of substratum should be made based on what indicators are frequently measured in the water source. While testing for Campylobacter sp. would still be required when indicators in the water column are present at high concentrations, using these surrogates may increase the probability of detecting Campylobacter in the watershed if biofilms are tested. This may be worthwhile if the water body has just experienced a heavy rainfall event and is designated for drinking water use. In situations where Campylobacter testing is not possible, surrogates (such as nitrates or ENT) may be used to decide when to administer boil water advisories.  4.4  Conclusion  The distribution and prevalence of Campylobacter sp. in biofilms in the aquatic environment are rarely studied phenomena that may be of interest to public health practitioners and water managers. This study has shown that most biofilms on a variety of surfaces can serve as reservoirs of Campylobacter to varying degrees and that Campylobacter frequencies in water column grab samples were low. However, Campylobacter recovery rates were low in this study due to the use of culture-based identification methods. Based on the analyses performed, we recommend the use of sediment, slate rock or wood as substrata for Campylobacter monitoring, depending on what water quality indicators are measured routinely. Of these suggested substrata, sediment definitely is the easiest to acquire and does not require early placement in the water body. We also recommend the use of nitrates and enterococci as potential faecal contamination surrogates that could indicate when Campylobacter may be present at high concentrations in a drinking water source.  104  Table 4.4 Evaluating substrata for use as Campylobacter sp. monitoring tools in natural aquatic systems and agricultural watersheds.  Correlations with Correlation with Substratum or Campylobacter microbial water water quality Cost and Medium presence quality parameters parameters frequently availability in biofilm measured Water  Low  Sediment  High  River Rock  Low  Slate Rock  High  Wood  Medium  Lexan  Medium  Sandpaper  None  No cost and high availability No cost and None Nitrate and Protozoa high availability No cost and None None high availability Medium cost ENT, and trend with Trend with nitrate* and medium FC* availability Low cost and ENT, FC and trend high Nitrate and Protozoa with HPC* availability High cost and None Protozoa low availability Medium cost None None and high availability None  Protozoa, and trend with phosphate*  Standardization Ease of biofilm Possibility of potential a removal loss in situ  High  None  None  High  Easy  None  Low  Difficult  High  Medium  Medium  Low  High  Medium  Low  High  Difficult  Medium  High  Medium  High  * The asterix designates trends between the variables, where the relationship was not statically significant (p>0.05). Highlighting these relationships allows other researchers to investigate further. a The criterion “Standardization potential” refers to the likelihood of acquiring the same volumetric or gravimetric amount of medium for analysis in the case of sediment and water, or the same amount of surface area in the case of artificial substrata. The criterion does not take into account the difficulty of assuring sediment has the same particle size, colonization period, or microbial population.  105 105  4.5 Acknowledgements This research has been funded by the CIHR Strategic Training Program in Public Health and the Agricultural Rural Ecosystem (PHARE) and Partner Institutes including the Institute of Cancer Research, Institute of Circulatory and Respiratory Health, Institute of Infection and Immunology, and the Institute of Population and Public Health. We would also like to thank the Environmental Engineering laboratory and Paula Parkinson (Civil Engineering Department, University of British Columbia) for working on the water nutrients, and the School of Environmental Health for providing some equipment, as well as transportation.  106  4.6 References Atabay, H.I., Corry, J.E.L., 1997. The prevalence of Campylobacters and arcobacters in broiler chickens. Journal of Applied Microbiology 83,619-626. Abulreesh, H.H., Paget, T.A., Goulder, R., 2005. Recovery of thermophilic Campylobacters from pond water and sediment and the problem of interference by background bacteria in enrichment culture. Water Research 39 (13), 2877-2882. APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater, 20th Edition. American Public Health Association (APHA), American Waterworks Association (AWWA), and Water Environment Association (WEF). Washington, DC. Azevedo, N.F., Vieira, M.J., Keevil, C.W., 2003. Establishment of a continuous model to study H. pylori survival in potable water biofilms. Water, Science and Technology 47 (5), 155-160. Bolster, C.H., Walker, S.L., Cook, K.L., 2006. Comparison of the transport behaviour of Escherichia coli and Campylobacter jejuni in saturated porous media. Journal of Environmental Quality 35, 1018-1025. Bolton, F.J., Coates, D., Hutchinson, D.N., Godfree, A.F., 1987. A study of thermophilic Campylobacters in a river system. Journal of Applied Bacteriology 62, 167-176. Brading, M.G., Jass, J., Lappin-Scott, H.M., 1995. Dynamics of bacterial biofilm formation. In: Microbial Biofilms. Cambridge, University Press, 46-63. Brown, P.E., Christensen, O.F., Clough, H.E., Diggle, P.J., Hart, C.A., Hazel, S., Kemp, R., Leatherbarrow, A.J.H., Moore, A., Sutherst, J., Turner, J., Williams, N.J., Wright, E.J., French, N.P., 2004. Frequency and spatial distribution of environmental Campylobacter spp. Applied and Environmental Microbiology 70, 6501-6511. Buswell, C.M., Herlihy, Y.M., Keevil, C.W., Leach, S.A., 1999. The effect of aquatic biofilm structure on the survival of enteric pathogen Campylobacter jejuni in water. In: Keevil, C.W. Godfree, A., Holt, D., Dow, C. (Eds.) Biofilms in the Aquatic Environment. Royal Society of Chemistry, Cambridge, 153-159. Carter, A.M., Pacha, R.E., Clark, G.W., Williams, E.A., 1987. Seasonal occurrence of Campylobacter spp. in surface waters and their correlation with standard indicator bacteria. Applied and Environmental Microbiology 53, 523-526 Clark, C.G., Price, L., Ahmed, R., Woodward, D.L., Melito, D.L., Rodgers, F.G., Jamieson, F., Ciebin, B., Li, A., Ellis, A., 2003. Characterization of waterborne outbreak–associated  107  Campylobacter jejuni, Walkerton, Ontario. Emerging Infectious Diseases 9 (10), 12321241. Coker, A.O., Isokpehi, R.D., Thomas, B.N., Amisu, K.O., Obi, C.L., 2002. Human Campylobacteriosis in developing countries. Emerging Infectious Diseases 8, 237–44. Ekdahl, K., Normann, B., Andersson, Y., 2005. Could flies explain the elusive epidemiology of Campylobacteriosis? BMC Infectious Diseases 5 (1), 11. Flanders, J.R., Yildiz, F.H., 2004. Biofilms as reservoirs for disease. In: Ghannoum, M., O’Toole, G.A. (Eds.) Microbial Biofilms. ASM Press, Washington, DC, 314-31. French, N., Barrigas, M., Brown, P., Ribiero, P., Williams, N., Leatherbarrow, H., Birtles, R., Bolton, E., Fearnhead, P., Fox, A., 2005. Spatial epidemiology and natural population structure of Campylobacter jejuni colonizing a farmland ecosystem. Environmental Microbiology 7 (8), 1116-26. Hunt, A.P., Parry, J.D., 1998. The effect of substratum roughness and river flow rate on the development of a freshwater biofilm community. Biofouling 12 (4), 287-303. Hutchison, M., Walters, L.D., Moore, T., Crookes, K.M., Avery, S.M., 2004. Effect of length of time before incorporation on survival of pathogenic bacteria present in livestock wastes applied to agricultural soil. Applied and Environmental Microbiology 70, 5111–5118. Keevil, C.W., 2003. Rapid detection of biofilms and adherent pathogens using scanning confocal laser microscopy and episcopic differential interference contrast microscopy. Water, Science and Technology 47 (5), 105-116. Kemp , R., Leatherbarrow, A.J., Williams, N.J., Hart, C.A., Clough, H.E., Turner, J., Wright, E.J., French, N.P., 2005 Prevalence and genetic diversity of Campylobacter spp. in environmental water samples from a 100-square-kilometer predominantly dairy farming area. Applied and Environmental Microbiology 71 (4), 1876-82. Kuusi, M., Nuorti, J.P., Hanninen, M.L., Koskela, M., Jussila, V., Kela, E., Miettinen, I., Ruutu, P., 2005. A large outbreak of Campylobacteriosis associated with a municipal water supply in Finland. Epidemiology and Infection 133 (4): 593-601. Leclerc, H., Schwartzbrod, L., Dei-Cas, E. 2002. Microbial agents associated with waterborne diseases. Critical Reviews in Microbiology 28, 371-409. Lehtola, M.J., Pitkanen, T., Miebach, L., Miettinen, I.T., 2006. Survival of Campylobacter jejuni in potable water biofilms: A comparative study with different detection methods. Water, Science and Technology 54 (3), 57-61. Maal-Bared, R., Bartlett, K.H., Bowie, W.R., Hall, E.R., Hall, K.J., 2007. Seasonal and spatial distribution of indicators and pathogens in biofilms in an agricultural watershed.  108  Annual Conference Proceedings of the American Waterworks Association, Toronto 2007. Maal-Bared, R., 2008. Comparing the distribution of pathogenic bacteria and common indicator microorganisms in biofilms on different surface types in an agricultural watershed in British Columbia (Canada). Unpublished PhD Thesis, University of British Columbia, British Columbia, Canada. Moore, J.E., Corcoran, D., Dooley, J.S.G., Fanning, S., Lucey, B., Matsuda, M., McDowell, D.A., Mégraud, F., Millar, B.C., O’Mahony, R., O’Riordan, L., O’Rourke, M., Rao, J.R., Rooney, P.J., Sails, A., Whyte, P. 2005. Campylobacter. Veterinary Research 36 (3), 351-82. Nam, H.M., Srinivasan, V., Murinda, S.E., Oliver, S.P., 2005. Detection of Campylobacter jejuni in dairy farm environmental samples using SYBR Green real-time polymerase chain reaction. Foodborne Pathogens and Disease 2 (2), 160-168. Nylen, G., Dunstan, F., Palmer, S.R., Andersson, Y., Bager, F., Cowden, J., Feierl, G, Galloway, Y., Kapperud, G., Megraud, F., Molbak, K., Petersen, L.R., Ruutu, P., 2002. The seasonal distribution of Campylobacter infection in nine European countries and New Zealand. Epidemiology and Infection 128, 383–90. Parry, J.D., 2004. Protozoan grazing of freshwater biofilms. Advances in Applied Microbiology 54, 167-196. Patrick, M.E., Christiansen, L.E., Waino, M., Ethelberg, S., Madsen, H., Wegener, H.C., 2005. Effects of climate on incidence of Campylobacter spp. in humans and prevalence in broiler flocks in Denmark. Applied and Environmental Microbiology 70, 7474–80. Rittmann, B.E., 1989. Detachment from biofilms. In: Characklis, W.G., Wilderer, P.A. (Eds.) Structure and Function in Biofilms. New York, Wiley, 49-58. Rittmann, B.E., 2004. Biofilms in the water industry. In: Ghannoum, M., O’Toole, G.A. (Eds.) Microbial Biofilms. ASM Press, Washington, DC, 359-78. Rollins, D.M., Colwell, R.R., 1986. Viable but nonculturable stage of Campylobacter jejuni and its role in survival in the natural aquatic environment. Applied and Environmental Microbiology 52, 531-538. Rood, K.M., Hamilton, R.E., 1995. Hydrology and water use for salmon streams in the Chilliwack/Lower Fraser habitat management area, BC. Prepared for the Fraser River Action Plan by Northwest Hydraulic Consultants Ltd., North Vancouver, BC. Rushworth, G., Younie, M., 2006. Compliance Assessment of Manure Application Practices in the Chilliwack and Agassiz Areas of the Lower Fraser Valley, British Columbia  109  February 7 – March 4, 2005. British Columbia Ministry of Environment. Report retrieved September 20, 2007. Available at: http://wlapwww.gov.bc.ca/sry/p2/compliance_assess/manure_chilliwack/compl_manur e.pdf Schuster, C.J., Ellis, A.G., Robertson, W.J., Charron, D.F., Aramini, J.J., Marshall, B.J., Medeiros, D.T., 2005. Infectious disease outbreaks related to drinking water in Canada, 1974-2001. Canadian Journal of Public Health 96 (4), 254-258. Snelling, W.J., McKenna, J.P., Lecky, D.M., Dooley, J.S.G., 2005. Survival of Campylobacter jejuni in waterborne protozoa. Applied and Environmental Microbiology 71, 55605571. Thomas, C., Hill, D.J., Mabey, M., 1999. Evaluation of the effect of temperature and nutrients on the survival of Campylobacter spp. in water microcosms. Journal of Applied Microbiology 86 (6), 1024–1032. Thomas, K.M., Charron, D.F., Waltner-Toews, D., Schuster, C., Maarouf, A.R., Holt, J.D., 2006. A role of high impact weather events in waterborne disease outbreaks in Canada, 1975 - 2001. International Journal of Environmental Health Research 16 (3), 167-80. Vereen, E., Lowrance, R.R., Cole, D.J., Lipp, E.K., 2007. Distribution and ecology of campylobacters in coastal plain streams (Georgia, United States of America). Applied and Environmental Microbiology 73, 1395-1403 Vingarzan, R., Belzer, W., Thompson, B., 2002. Nutrient levels in the atmosphere of the Elk Creek Watershed, Chilliwack, BC (1999-2000). Environment Canada Technical Report. Aquatic and Atmospheric Sciences Division, Vancouver, Canada. Waage, A.S., Varund, T., Lund, V., Kapperud, G., 1999. Detection of small numbers of Campylobacter jejuni and Campylobacter coli cells in environmental water, sewage, and food samples by a semi-nested PCR assay. Applied and Environmental Microbiology 65, 1636-1643.  110  5  5.1  Distribution and patterns of phenotypic antibiotic resistant Escherichia coli isolates from an agricultural watershed (Elk Creek, British Columbia)4  Introduction  The increase in antibiotic resistance of microbial populations is an issue of global concern (Rooklidge, 2004). Antibiotics have been extensively used in humans, agriculture, aquaculture, fruit growing, beekeeping, and the food and glue industries (Kümmerer, 2004). Antibiotics are commonly used in agriculture to prevent and treat disease, as well as to promote growth (Rooklidge, 2004). Consequently, animal waste that is produced can potentially contain both large numbers of bacteria that may be resistant to one or multiple antibiotics due to exposure in the gut (Diarra et al., 2007; Witte, 2000), or animal waste can contain some concentration of antimicrobial agents, which once in the environment, can affect selective pressures acting on bacteria thereby promoting resistance (Richardson and Bowron, 1985; Hirsch et al., 1999; Hamscher et al., 2002; Martinez-Carballo et al., 2007). Animal waste produced in agriculture is often used to fertilize agricultural lands. When manure is applied to land, antimicrobial compounds concentrate and mobilize in soil. Antimicrobials often end up leaching into ground water (Hirsch et al., 1999; Chee-Sanford et al., 2001), or are carried to surface water through agricultural runoff (Pederson et al., 2003). Agricultural runoff also carries large concentrations of faecal bacteria (resistant and susceptible) from manure-applied lands to water sources (Duriez and Topp, 2007). Thus, it is not surprising that many studies have found antibiotic resistant bacteria in aquatic environments (Schwartz et al, 2003; Sayah et al., 2005; McKeon et al., 1995, Ash et al., 2002; Watkinson et al., 2007). Understanding the distribution of antibiotic resistant bacteria in water sources in an agricultural watershed is important since antimicrobial resistance is a threat to public health, animal health, and to the occupational health of farmers (Rooklidge, 2004). This study was conducted to increase our knowledge of the distribution of antibiotic resistant E. coli in an agricultural watershed. E. coli was chosen because it is a commonly used faecal 4  A version of this chapter will be submitted for publication. Maal-Bared, R., Bartlett, K.H., Bowie, W.R., Hall, E.R., Hall, K.J. 2007. Distribution and patterns of phenotypic antibiotic resistant Escherichia coli isolates from an agricultural watershed (Elk Creek, British Columbia). Applied and Environmental Microbiology.  111  contamination indicator that is transient in aquatic environments, but still survives well in water. It also is a gut commensal whose antibiotic resistance has been well documented (Edge and Hill, 2005). The study was also designed to investigate the effects that a variety of physical and chemical factors had on antimicrobial resistance to several commonly used antibiotics in the watershed. It is useful to observe the effects of chemical and physical variables on bacterial antimicrobial resistance in natural aquatic environments, as opposed to laboratory microcosms, due to the large differences in the results of these different types of studies (Ford, 1994; Williams et al., 1996). The working hypothesis was that due to the high levels of antibiotic use and manure spreading in the Elk Creek agricultural watershed, E. coli isolates may experience varying levels of antibiotic resistance, as measured by broth microdilution Minimum Inhibitory Concentrations (MIC). The MIC for a microorganism is the lowest concentration of antimicrobial agent that will inhibit the growth of the microorganism after overnight incubation (Andrews, 2001). Thus, MICs at the agriculturally impacted sites would be higher than those at the pristine reference site. The second hypothesis was that MICs may be directly or indirectly influenced by physical and chemical factors (such as water quality parameters and biofilm development). Methods used to evaluate antibiotic resistance range from molecular techniques to growth-based techniques. For the purposes of this study, a growth-based technique was chosen.  5.2 Methods 5.2.1 Sample collection All isolated E. coli and E. coli O157 samples were collected, processed, and identified as described in Maal-Bared et al. (2007). E. coli isolates were collected from the Elk Creek watershed in British Columbia (Canada) from December 2005 to April 2007. Four sites were chosen along the length of the creek. Site one was the control site (in the headwaters) and sites two, three, and four were located in an area of agricultural land-use. Samples were collected from these sites on a rotating basis, and from water, sediment and biofilms of river rock, slate rock, wood, Lexan, and sandpaper. Biofilm samples collected between December 2005 and 2006 biofilms were allowed to colonize the substrata for one-month. Biofilms collected between January 2007 and April 2007 were allowed to colonize the substrata for different periods of time (1, 2, 3, 4, 12 and 24 weeks). All samples were returned to the lab for processing and analysis. Identification procedures followed Standard Methods (APHA et al., 1998). After isolation and identification, samples were kept in 5 mL glass vials in a sterile 50% Glycerol-TSA solution  112  (Difco™, Becton, Dickinson & Co., Sparks, MD) in a –80oC freezer until antibiotic resistance testing was performed.  5.2.2 Sample preparation Solutions of each isolate were prepared for the broth microdilution MICs. Each frozen culture isolate was plated onto TSB (Difco™, Becton, Dickinson & Co., Sparks, MD) and incubated overnight to ascertain culture purity. The pure cultures were suspended in about 4 mL of 0.85% sterile saline solution in 5 mL test tubes. Sufficient pure colonies were added until a 0.5 McFarland turbidity level was reached. The assessment was made visually using McFarland turbidity standards (bioMérieux Vitek, Durham, NC).  5.2.3 Broth microdilution MIC The National Committee for Clinical Laboratory Standards (NCCLS) standard methods were used to perform broth microdilution testing (NCCLS, 2000). The following antibiotics (all from Sigma-Aldrich Inc, St Louis, USA) and concentrations were used: Ampicillin (Potency: 925 µg/mg, final concentration range 1 µg/mL-512 µg/mL), Cefotaxime (Potency: 940 µg/mg, final concentration range 2 µg/mL-1024 µg/mL), Ciprofloxacin (Potency: 850 µg/mg, final concentration range 0.125 µg/mL-64 µg/mL), Nalidixic Acid (Potency: 990 µg/mg, final concentration range 1 µg/mL-512 µg/mL), Streptomycin (Potency: 750 µg/mg, final concentration range 2 µg/mL-1024 µg/mL), and Tetracycline (Potency: 900 µg/mg, final concentration range 0.5 µg/mL-256 µg/mL). Antibiotic stock solutions were prepared using the highest desired concentration needed for testing. Ninety-six well plates were filled with 0.05 mL Mueller-Hinton Broth (BBL™, Bencton, Dickinson & Co., Sparks, MD), 0.05 mL of antibiotic solution, and 0.01 mL of the prepared E. coli culture suspension. The first well in the 12-well row was always a negative control (MH broth), while the second well always was the positive control (MH broth with bacterial inoculum). The third well in a row was always inoculated with 0.05 mL of the antibiotic stock, which was then used to transfer the antibiotic to the remaining wells with a two-fold dilution. Once wells were inoculated, they were incubated at 35oC overnight and read spectrophotometrically at wavelength 405 nm the next day. MIC50% (from hereon referred to as MIC unless otherwise stated) was determined to be the lowest concentration of the antibiotic that reduces the growth of the organism by 50% compared to the positive control. The streptomycin MIC was not listed in the NCCLS standard method, thus it was chosen from the literature (Sunde and Norstrom, 2005).  113  5.2.4 Water quality data Water samples were collected to obtain physical and chemical water quality information as described in Maal-Bared (2008, p.61). Physical water quality data included: water depth (manually measured), stream width (manually measured), water velocity (m/s), and water flow (calculated in m3/s and result of three variables water depth x width x velocity). Other physical and chemical water quality variables measured in situ using a YSI-85 instrument (Yellow Springs, OH) were water temperature (oC), conductivity (µS/cm), specific conductivity (µS/cm), and dissolved oxygen (% saturation and mg/L). Chilliwack ambient rainfall measurements (collected at the Abbotsford Airport weather station about 45 km from the watershed) were retrieved from the Weather Network website (http://www.theweathernetwork.com/weather/CABC0308) and were calculated as the cumulative rainfall over 7 days before the sampling date (mm). Preserved water samples were analyzed for nitrate+nitrite (NOx), ortho-phosphate (PO4), ammonia (NH3) and dissolved organic carbon (DOC) upon return to the laboratory.  5.2.5 Statistical analyses All statistical analyses were completed using JMP 7.0 (SAS Institute, NC, USA). Chi-square tests were used to determine the statistical significance of differences in susceptibility/resistance patterns and MICs of E. coli isolated from different sites and from different substrata. To describe relationships among MICs and water quality variables, logistic regressions were used for each antibiotic. Logistic regressions were used instead of linear regressions, since the goal was to predict an ordinal, non-continuous response outcome (antimicrobial resistance measured by MICs) through a set of variables that were continuous (water quality variables). The results were displayed in terms of chi-square values and p-values (Quinn and Keogh, 2002). To determine whether the effect of substratum colonization period (biofilm age) on resistance was significant, another chi-square test was used. Alpha was set at 0.05.  5.3 Results 5.3.1 General phenotypic antibiotic resistance patterns in the watershed based on broth microdilution MIC results Many of the E. coli isolates tested showed some levels of antibiotic resistance, as can be seen in Figure 5.1. Resistance to all antibiotics (with the exception of tetracycline) exhibited positive skew. Tetracycline exhibited a more bimodal distribution, with MICs peaking at very low and  114  very high concentrations. Table 5.1 shows the number of isolates that were susceptible, moderately susceptible or resistant to each antibiotic. The highest levels of resistance were to tetracycline, followed by ampicillin, streptomycin and nalidixic acid. While no isolates were resistant to cefotaxime, 16 isolates were moderately susceptible to the antibiotic.  Table 5.1 Total number of E. coli isolates from Elk Creek (British Columbia), which were susceptible, moderately susceptible or resistant to different antibiotics tested based on broth microdilution MICs (n=214)  Antibiotic  Susceptible  Moderately susceptible  Resistant  Ampicillin  165  14  35  Cefotaxime  198  16  0  Ciprofloxacin  213  0  1  Nalidixic Acid  203  0  8  Streptomycin  197  0  17  Tetracycline  85  2  127  115  214 200  150  100 76 50  39  35  15 0  1  2  4  8  9  16  12  3  32  4  2  64  128  256  512  250  250 214 200 178 150  100  50 15 0  14  C) Streptomycin (susceptible ≤ 8 µg/mL; resistant ≥ 16 µg/mL) Frequency of E.coli isolates with inhibited growth  Frequency of E.coli isolates with inhibited growth  250  B) Cefotaxime (susceptible ≤ 8 µg/mL; moderately susceptible = 16-32 µg/mL; resistant ≥ 64µg/mL) Frequency of E.coli isolates with inhibited growth  A) Ampicillin (susceptible ≤ 8 µg/mL; moderately susceptible = 16 µg/mL; resistant ≥ 32µg/mL)  2  4  13  5 8  3  16  32  214 200  158 150  100  50  0  Total  5  2  20  19  4  8  4  1  5  4  3  16  64  128  256  512  MIC 50%  >512 Total  Total  MIC 50%  MIC 50%  E) Ciprofloxacin (susceptible ≤ 1 µg/mL; moderately susceptible = 2 µg/mL; resistant ≥ 4  µg/mL) 214  200  150  100  101  49  50  36 14  0  1  2  4  8  3  1  1  1  16  32  64  128  MIC 50%  8 >512  Total  250  250 Frequency of E.coli isolates with inhibited growth  Frequency of E.coli isolates with inhibited growth  250  214  207 200  150  100  50  0  F) Tetracycline (susceptible ≤ 4 µg/mL; moderately susceptible = 8 µg/mL; resistant ≥ 16µg/mL)  0.125  4  1  1  1  0.25  0.5  1  >64  MIC 50%  Total  Frequency of E.coli isolates with inhibited growth  D) Nalidixic Acid (susceptible ≤ 16 µg/mL; resistant ≥ 32µg/mL)  214 200  150 112 100 81 50  0  0.5  2  2  2  2  5  2  3  3  2  4  8  16  32  64  128  256  >256  MIC 50%  Figure 5.1 Minimum inhibitory concentration distribution of E. coli isolates from Elk Creek (British Columbia) to Ampicillin, Cefotaxime, Ciprofloxacin, Nalidixic acid, Streptomycin, and Tetracycline (n=214).  116  116  Total  5.3.2 Differences in phenotypic antibiotic resistance by site Table 5.2 shows the resistance levels of E. coli isolated from the four different sites at Elk Creek to the six tested antibiotics. From the table, we can see that the highest frequency of resistance was associated with site 4, while the lowest frequency was associated with site 1. Also, site 1 was most likely to be associated with moderate resistance. The results show that while site 1 is more pristine than the other three sites located in the agricultural reach, it is still contaminated with resistant E. coli. Table 5.2 Distribution of susceptible, moderately susceptible, and resistant E. coli isolated from four different sites in Elk Creek (British Columbia) for all tested antibiotics. Site 1 is the control site and is located in the headwaters, and sites 2, 3, and 4 are situated in the agricultural reach.  Site 1 2 3 4  Susceptible 67 % 58 % 61 % 45 %  Resistance Level (% Isolates) Moderately Susceptible Resistant 13 % 21 % 6% 35 % 3% 35 % 10 % 45 %  Two chi-square tests were performed to evaluate whether the differences in E. coli resistance levels among different sites were statistically significant. The first test evaluating general E. coli resistance levels (susceptible, moderately susceptible and resistant) among different sites was not statistically significant (p-value=0.449). The second chi-square test evaluated whether the differences in MIC 50% of E. coli isolated from different sites were statistically significant for each antibiotic. The differences in Ampicillin, Cefotaxime, Nalidixic Acid and Tetracycline resistance at different sites were statistically significant, with p-values of 0.001, 0.002, 0.002, and 0.041 respectively. Differences in E. coli resistance to Ciprofloxacin and Streptomycin from different sites were not statistically significant. Figure 5.2 shows the differences in E. coli antibiotic resistance MIC levels to all the antibiotics tested at different sites. In Figure 5.2, the yaxis depicts the total number of E. coli isolates that were inhibited at the stated MIC level on the x-axis throughout the whole study period. The x-axis was grouped by site. For example, ten E. coli isolates were inhibited by Ampicillin at site 1 at 8 µg/mL out of a total of 69 isolates tested for that site. Figure 5.2 shows that E. coli isolate resistance to particular antibiotics differed depending on the site. At site 1, E. coli isolates were mostly resistant to ampicillin and tetracycline. E. coli isolates at sites 2 and 4 exhibited similar resistance patterns. Site 3 E. coli isolates were most resistant to tetracycline. The results indicate that sites 2 and 4 are probably the  117  most affected by runoff. Also, the tetracycline results indicate that the tetracycline resistance observed may be influenced by factors other than runoff and manure spreading.  118  A) Ampicillin 80 70  Sum(Frequency)  60 50 40 30 20 10 1 2 4 8 16 32 64 128 256 512 >512 Total 1 2 4 8 16 32 64 128 256 512 >512 Total 1 2 4 8 16 32 64 128 256 512 >512 Total 1 2 4 8 16 32 64 128 256 512 >512 Total  0  1  2  3  4  MIC 50% w ithin Site  B) Cefotaxime 80  Sum(Frequency)  70 60 50 40 30 20  32  Total  >64  Total  3  16  8  4  2  Total  32  16  8  4  2  Total  32  2  1  1  16  8  4  2  Total  32  16  8  4  0  2  10  4  MIC 50% w ithin S ite  C) Ciprofloxacin 80 70  Sum(Frequency)  60 50 40 30 20  1  2  0.5  0.25  0.125  Total  1 3  >64  0.5  0.25  0.125  Total  >64  1  0.5  0.25  0.125  Total  >64  1  0.5  0.25  0  0.125  10  4  MIC 5 0% w ith in Site  Figure 5.2 Total number of E. coli isolates resistant to Ampicillin, Cefotaxime, and Ciprofloxacin at the stated minimum inhibitory concentration (MIC 50%) at sites 1 to 4 in Elk Creek (British Columbia).  119  D) Nalidixic Acid 80  Sum(Frequency)  70 60 50 40 30 20  1 2 4 8 16 32 64 128 >512 Total  1 2 4 8 16 32 64 128 >512 Total  1 2 4 8 16 32 64 128 >512 Total  0  1 2 4 8 16 32 64 128 >512 Total  10  1  2  3  4  MIC 50% w ithin Site  E) Streptomycin 80 70 Sum(Frequency)  60 50 40 30 20  1  3  512  Total  256  64  128  8  16  4  2  512  Total  256  64  128  8  2  16  4  2  512  Total  256  64  128  8  16  4  2  512  Total  256  64  128  8  16  2  0  4  10  4  MIC 5 0 % w ith in S ite  F) Tetracycline  Sum(Frequency)  70 60 50 40 30 20  0.5 2 4 8 16 32 64 128 256 >256 Total  0.5 2 4 8 16 32 64 128 256 >256 Total  0.5 2 4 8 16 32 64 128 256 >256 Total  0  0.5 2 4 8 16 32 64 128 256 >256 Total  10  1  2  3  4  MIC 50% w ith in Site  Figure 5.2 Total number of E. coli isolates resistant to Nalidixic Acid, Streptomycin and Tetracycline at the stated minimum inhibitory concentration (MIC 50%) at sites 1 to 4 in Elk Creek (British Columbia).  120  5.3.3 Differences in phenotypic antibiotic resistance by substratum Table 5.3 shows the distribution of susceptible, moderately susceptible and resistant E. coli on different surface types at all sites in Elk Creek. The frequencies of antibiotic resistance level were reported in percentages to take into account the fact that isolation frequencies of E. coli differed based on substratum type. Table 5.3 shows that resistance was most common in sediment and river rock isolates. Lexan and sandpaper biofilms, as well as water, were the substrata most likely to be associated with antibiotic susceptible E. coli. Two chi-square tests were performed to evaluate whether the differences in E. coli resistance levels among different substrata types were statistically significant. The first tests evaluating general E. coli resistance levels (susceptible, moderately susceptible and resistant) on different substrata were not statistically significant (pvalue=0.432). Table 5.3 Distribution of susceptible, moderately susceptible, and resistant E. coli in water, sediment and in biofilms on river rock, slate rock, wood, Lexan, and sandpaper isolated from all sites at Elk Creek British Columbia Resistance Level (% Isolates) Substratum Susceptible  Moderately Susceptible Resistant  Water River Rock Slate Rock Wood Lexan Sandpaper Sediment  6% 8% 5% 0% 12 % 4% 3%  71 % 53 % 68 % 72 % 76 % 74 % 53 %  24 % 39 % 27 % 28 % 12 % 22 % 44 %  Total number of isolates 17 36 22 25 17 23 34  The second chi-square test evaluated the differences in E. coli antibiotics resistance levels for each antibiotic and each substratum separately using the actual MICs. The differences in ampicillin, cefotaxime and tetracycline were statistically significant, with p-values of <0.0001, 0.008 and 0.0427 respectively. Differences in E. coli resistance to Ciprofloxacin, Nalidixic Acid, and Streptomycin on different substrata were not statistically significant.  5.3.4 Pathogenic E. coli O157 antibiotic resistance Twenty-seven of the 204 E. coli isolates were determined to be pathogenic E. coli O157. Table 5.4 summarizes susceptibility and resistance patterns of the pathogen. Resistance levels to tetracycline were highest, while there was no resistance to Cefotaxime and Ciprofloxacin. Only one pathogenic E. coli O157 isolated from site 1 was Ampicillin resistant, while four were tetracycline resistant. At site 2, one E. coli O157 was resistant to each ampicillin, nalidixic acid,  121  and streptomycin, while five isolates were resistant to tetracycline. Seven isolates were resistant to tetracycline at site 3. Finally, one isolate of E. coli O157 was resistant to each ampicillin, nalidixic acid, and streptomycin, and two isolates were resistant to tetracycline at site 4.  Table 5.4 Total number of pathogenic E. coli O157 isolates from Elk Creek (British Columbia), which were susceptible, moderately susceptible or resistant to different antibiotics tested based on broth microdilution MICs (n=27)  Antibiotic Ampicillin Cefotaxime Ciprofloxacin Nalidixic Acid Streptomycin Tetracycline  Susceptible 23 24 27 25 25 9  Moderately susceptible 1 3 0 0 0 0  Resistant 3 0 0 2 2 18  5.3.5 Relationships between antibiotic resistance and other water quality parameters in the watershed Logistic regressions were used to determine whether values of water quality variables affected antibiotic resistance levels in the Elk Creek watershed. Logistic regression results can be seen in Table 5.5 and are presented in terms of chi square statistic and p-values. Some variables showed interesting relationships with antibiotic resistance in the Elk Creek watershed. Streptomycin seemed to be least affected by changes in water quality. Dissolved oxygen concentration, as well as dissolved oxygen measured in percent air saturation, appeared as a predictor of resistance for several antibiotics, as did nutrient concentrations, rainfall and water depth. Two more chi-square tests were performed; the first to determine whether substratum colonization period had any effect on MICs, and the second to determine whether the season (dry versus wet) had any effect on MICs. The first test showed that the colonization period of the biofilm substratum only had an effect on nalidixic acid (p=0.006) and tetracycline resistance (p=0.010). For nalidixic acid, the highest E. coli MIC levels were found when the substratum had been colonized for 4 weeks. For tetracycline, the highest E. coli MIC levels were found when the substratum had been colonized for 12 weeks, while the lowest were found when the substratum was colonized for 4 weeks. However, seasonality had no statistically significant effect on antibiotic resistance.  122  Also, Figure 5.3 shows the distribution of MICs between December 2005 and April 2007, along with the rainfall data (in mm). Ampicillin MIC peaks did seem to follow some rainfall patterns peaking several times during the year (February, March, July and December 2006 and in January 2007). Cefotaxime MICs experienced three peaks: two in February and one in May 2006. Highest levels of Ciprofloxacin resistance in Elk Creek were observed between December 2005 and February 2006, as well as one peak in May 2006, when rainfall peaked as well (the same relationship was seen in the logistic regression in Table 5.5). Nalidixic acid MICs peaked in December 2005, and in January February, and March of 2006. Streptomycin resistance peaked in February and March 2006 and between October and December 2006. Tetracycline resistance did not show any particular peaks and MICs were high through the sampling period.  123  Table 5.5 Results of logistic regression models highlighting relationships between different antibiotic resistant E. coli isolates (measured by MICs) and water quality parameters in Elk Creek. The positive and negative signs next to the antibiotic name indicate whether relationships were positive or negative. Statistically significant relationships are bolded, while strong trends that are not statistically significant are not bolded.  Type of variable Variable Velocity (m/s)  Physical  Depth (cm) Water Temp (oC)  Ambient Temp (oC) Rainfall (mm) DO (mg/L)  DO (% air saturation)  Chemical  Sp. Conductivity (µS/cm) NOx (mg/L)  PO4 (mg/L) NH3 (mg/L) DOC (mg/L)  Antibiotic Resistance  χ2  p-value  Ciprofloxacin (+)  3.117  0.078  Tetracycline (-)  3.960  0.047  Cefotaxime (-)  4.140  0.042  14.348  0.001  Tetracycline (+)  3.687  0.055  Ampicillin (-)  8.338  0.004  Nalidixic Acid (+)  10.373  0.001  Ampicillin (-)  14.216  <0.001  Nalidixic Acid (+)  9.683  0.002  Ciprofloxacin (+)  4.946  0.026  Tetracycline (-)  4.174  0.041  Ampicillin (+)  4.806  0.028  Tetracycline (-)  6.225  0.013  Cefotaxime (+)  7.241  0.007  Nalidixic Acid (+)  4.880  0.027  Tetracycline (-)  3.084  0.079  Ampicillin (-)  5.814  0.016  Ampicillin (+)  4.178  0.041  Cefotaxime (-)  3.632  0.057  Nalidixic Acid (-)  5.489  0.019  Ampicillin (-)  4.741  0.029  Nalidixic Acid (-)  15.862  <0.001  Nalidixic Acid (-)  5.200  0.023  Nalidixic Acid (-)  7.839  0.005  Streptomycin (+)  3.641  0.056  Tetracycline (-)  4.867  0.027  Nalidixic Acid (-)  124  05/02/2007 14/02/2007 19/02/2007 09/03/2007 26/03/2007  05/02/2007 04/04/2007 14/02/2007 19/02/2007 09/03/2007  04/04/2007  25/01/2007  25/01/2007  26/03/2007  18/01/2007  18/01/2007  12/07/2006  05/07/2006  25/05/2006  16/05/2006  09/05/2006  30/03/2006  22/03/2006  01/03/2006  21/02/2006  08/02/2006  18/01/2006  10/01/2006  03/12/2006  0.1 03/12/2006  1  28/10/2006  10  06/10/2006  10 0  28/10/2006  10 00  23/07/2006  10 00 0  06/10/2006  Date  23/07/2006  12/07/2006  05/07/2006  25/05/2006  16/05/2006  09/05/2006  30/03/2006  22/03/2006  01/03/2006  21/02/2006  08/02/2006  18/01/2006  12/12/2005  0  10/01/2006  12/12/2005  MICs for E. coli Isolates  Mean(Rainfall) 200  175  150  125  100  75  50  25  Date  Y Ampi ci l l i  Cefotaxi m  Ci profl ox  Nal i di xi c  Streptom  Tetracycl  Figure 5.3 Minimum Inhibitory Concentrations of E. coli isolates (in µg/mL) and rainfall data (in mm) from Elk Creek (British Columbia) from December 2005 to April 2007.  125  5.4 Discussion The development of antibiotic resistance is either driven by selective pressures or by the presence of genes within the organism that allows it to be resistant (Witte, 2000). The prevalence of antibiotic resistance, however, is based on how resistance is spread. For example, the fertilization of land with manure from animals treated with veterinary antibiotics would contribute higher concentrations of resistant bacteria to the environment than natural pressures would (Kümmerer, 2004). All of these processes would affect the E. coli isolate antibiotic resistance in the Elk Creek watershed. The main selective pressure contributing to resistance within the gut of poultry, livestock and other animals, and resistance in the external environment outside the host would be exposure to antibiotics. The Elk Creek watershed is an intensive agricultural watershed with an animal density of 2.46 animal units per hectare. This means that Elk Creek is mainly used for dairy and poultry production, as well as animal rearing, all of which utilize a large amount of antibiotics. Also, manure is applied to about 58% of the agricultural area to fertilize the land and dispose of the waste. It has been estimated that an approximate 140,000 tonnes of poultry manure are applied to the land in the Lower Fraser Valley annually (Vingarzan et al., 2002). Another factor contributing to antibiotic resistance in the agricultural watershed would be exposure to heavy metals and pesticides, since those types of resistance are associated with the same plasmids (Ford, 1994). Calomiris et al. (1984) suggested that presence of metals (Cu, Pb, and Zn) in water distribution systems in a community in Oregon was positively correlated with multiple antibiotic resistances in treated water. Agriculture often results in the addition of heavy metals to soil through waste disposal, pesticides, and inorganic fertilizers use, but atmospheric deposition can have a large influence on heavy metal concentrations in soil as well (Nicholson et al., 2003). The acquisition of antibiotic resistance genes is also a complicated process. The bacterium may be able to mutate through the processes of conjugation (transfer of DNA from one bacterial cell to another), transformation (picking up DNA from the environment) or transduction (exchange of DNA among bacteria using viruses). Once resistance genes reach the bacterium, resistance is spread further by clonal spread of resistant strains or dissemination of antibiotic resistant genes to other individuals (Witte, 2000). For example, the presence of low concentrations of antibiotics, such as oxytetracycline, in the environment has been shown to result in plasmid conjugation in anaerobes (Salyers and Shoemaker, 1996).  126  Antimicrobial resistance of E. coli isolated from water to a variety of antibiotics has been shown in many studies (Sayah et al., 2005; Edge and Hill, 2005; Schwartz et al., 2003; Watkinson et al., 2007). Also, some studies were able to isolate resistant E. coli O157 from multiple use watersheds (Hamelin et al., 2006; Hamelin et al., 2007). The results of the present study indicate that E.coli and pathogenic E. coli O157 isolates exhibited the highest levels of resistance to tetracycline, followed by ampicillin and streptomycin. High levels of E. coli antibiotic resistance to tetracycline have been observed in several studies. Wilkerson et al. (2004) observed the occurrence of antibiotic resistant E. coli O157, particularly to tetracyclines, in bovine populations. Watkinson et al. (2007) also observed the highest levels of resistance in E. coli isolates to tetracycline. Edge and Hill (2005) found high levels of ampicillin and tetracycline resistance in E. coli isolates from surface water in Hamilton (Ontario). Edge and Hill (2005) also found low levels of resistance to ciprofloxacin, which corresponds with our results as well. High levels of resistance to ampicillin were expected due to fact that ampicillins are older antibiotics that have been extensively used over the years (Rooklidge, 2004). It is difficult to understand what the driving force behind tetracycline resistance is. Work by Khachatryan et al. (2006) suggested that tetracycline-streptomycin-sulfonamide-resistant E. coli presence in dairy cows maybe related to other unknown environmental factors (such as vitamin D supplementation), besides antibiotic selective pressures. This result was also found by Bartolino et al. (2006), who investigated tetracycline resistance in a remote population of naïve unexposed individuals in Bolivia and found that 64% of them were tetracycline resistant. Walk et al. (2007) also agreed with those findings and suggested that tetracycline resistance is probably the result of genetic hitchhiking of resistant loci (Tetr) and not the direct result of tetracycline selective pressures. Walk et al. (2007) also suggested that ampicillin resistance is probably the result of clonal resistance. As seen in the results, site 4 was most closely associated with resistance. Sites 2 and 3 were more likely to be resistant than site one. However, site 1 was more likely to be moderately resistant. This may be an indication of different antibiotic resistance acquisition processes or lower selective pressures being applied at site 1, compared to those at the other sites. The presence of E. coli in the headwaters is not a surprise due to the high wildlife density in the forested component of the watershed, which could contaminate the stream (Meays et al., 2006). Antibiotic resistance at the headwaters in Elk Creek has not previously been reported, however both resistant and  127  pathogenic E. coli have previously been observed in unpolluted sites by other studies (Hamelin et al., 2007). The resistance may be related to antibiotic resistance crossover between domesticated animals, livestock and wildlife. Many studies have observed the spread of antibiotic resistance outside the farm ecosystem to wildlife populations (Sayah et al., 2005). Work by Hagedorn et al. (1999) showed high levels of similarity between antibiotic resistant isolates of beef cows and wildlife and suggested some cross contamination. The prevalence of pathogenic E. coli O157 in the headwaters itself may imply cross-contamination between cattle and wildlife, since ruminants are considered to be the main reservoirs of E. coli O157:H7 (Hancock et al., 1998; Seurinck et al., 2003). Contamination of wildlife feces, such as those from deer, possums, racoons and birds, with pathogenic E. coli O157 has been repeatedly reported in the literature (Renter et al., 2001; Renter et al., 2003; Shere et al., 1998; Hancock et al., 1998). Although unlikely, the resistance at site 1 in the headwaters could have been affected by naturally produced antimicrobial agents in the environment by competing microorganisms. In the natural environment, bacteria and fungi in soil both produce a variety of antibiotics that affect growth rates and population dynamics of other microbial populations (Kümmerer, 2004). The claim has been made, however, that the concentration of antibiotic produced by bacteria is too small to cause antibiotic resistance and only affects the producer’s immediate surroundings in soil, except in the tropics where bacterial concentrations in soil are higher (Kümmerer, 2004). Bacteria also produce allelopathic compounds (also called bacteriocins) that should allow the producing bacterium to compete better. E. coli produces two of these bacteriocins (colicins and microcins) under various environmental conditions such as iron deficiency and nutrient stress (Gordon and O’Brien, 2006), and their role will change based on its biotic and abiotic environment (Riley and Gordon, 2002). Edwards et al. (2001) found that bacteriocins of heterotrophic bacteria in the water column affected symbioses among closely related bacteria. The authors also pointed out that the effects of bacteriocins on biofilms and flocs would be much higher than the effects on planktonic bacteria. Another factor worth considering in examining bacterial population dynamics are the interactions among bacteria and algae. Algae, as well as cyanobacteria, are also known for their ability to secrete two types of chemical agents that affect the microbial population surrounding them (LeFlaive and Ten-Hage, 2007): a) allelopathic chemicals that restrict the activity of competing symbionts or predators through a variety of effects ranging from inhibition of photosynthesis to cellular paralysis; and b) toxins that are toxic and inhibit enzyme production or interfere with cell membrane receptors. The amount of allelopathic chemical produced is highly affected by nutrient stress, such as phosphate limitation (von Elert and Jüttner,  128  1997), culture temperature (Issa, 1999), and pH (Ray and Bagchi, 2001). A study by Matsui et al (2003) suggested that algae, such as Euglena grasilis, Microcystis aeruginosa, and Chlamydomonas neglecta, were involved in bacterial horizontal gene transfer (transformation in particular) by stimulating the release of DNA into the aquatic environment for unknown reasons. In biofilms, algae can enhance bacterial growth through the release of extracellular dissolved organic compounds, but inhibit bacterial growth through the release of algal exudate toxins (Olapade and Leff, 2006). More information about the effects of all these chemicals on microbial communities in natural aquatic environments and particularly in biofilms would improve our understanding of the role naturally produced antimicrobials play in antibiotic resistance. The results also indicated that sediment was a major reservoir of antibiotic resistant E. coli. These findings have been supported in the literature by other studies investigating the presence of antibiotic resistant bacteria in aquaculture sediments and marine sediments (Samuelson et al., 1992; Andersen and Sandaa, 1994). Also, the high levels of resistance observed in E. coli isolates from sediment may be related to the fact that antibiotics, such as quinolones and tetracyclines, are sorbed by organic matter and can therefore accumulate in some sediments (Kümmerer, 2004). In the present study, E. coli isolated from Lexan, sandpaper and water were most likely to be susceptible to antibiotics. E. coli isolated from water were likely to be susceptible for several reasons. Bacteria in water tend to be very sparsely distributed, thus making any genetic material exchange very difficult (Murray, 1997). Antibiotics in water also tend to be attached to particles or organic matter, thus concentrating in compartments that planktonic bacteria do not have exposure to, which may decrease selective pressures on planktonic bacteria. It is not clear why sandpaper biofilms did not accumulate many resistant bacteria. E. coli isolated from river rock biofilms were likely to be resistant. This is not surprising since biofilms are known to be able to avoid the effects of antibiotics better than their planktonic counterparts (Lejeune, 2003) using several hypothesized mechanisms (Stewart and Costerton, 2001), such as: a) failed penetration of the antibiotic through the extracellular polymeric substance created by the biofilm community, b) differentiation of biofilm members into protected phenotypes, and c) antagonizing the antibiotic effects by altering the biofilm microenvironment. Biofilms in general present a great opportunity for horizontal gene transfer. Conjugation is often assumed to have an essential role in antibiotic resistance transfer via plasmids, particularly in biofilms on sediment and rocks (Witte, 2000). Also, natural transformation has been demonstrated in river epilithon of Acinetobacter calcoaceticus by Williams et al. (1996).  129  The water quality variables could have affected observed antibiotic resistance patterns in several ways. The variables could have affected antibiotic concentrations by reducing selective pressures on E. coli isolates through decreasing antibiotic concentrations in the aquatic environment. Depth would be an example of such a case, since depth is a determinant of light penetration through the water column. Oxytetracyclines have been shown to degrade faster in water at high temperatures, alkaline conditions and high light exposure (Doi and Stoskopf, 2000). This could explain why tetracycline MICs had a positive relationship with depth, but not why cefotaxime and nalidixic acid MICs exhibited a negative relationship with this variable. Photodegradation of cefotaxime, along with 18 other commonly used antibiotics, has been shown by Alexy et al. (2004) in the laboratory. Water temperature (also reflected in ambient temperature) exhibited a negative association with MICs for ampicillin, but a positive one with nalidixic acid. The negative relationship can be explained by antibiotic degradation again, but the positive relationship is harder to explain. Williams et al. (1996) showed an increase in natural transformation rates in river biofilms on rocks with increased temperature. Thus, in some cases, water quality variables may have affected antibiotic resistance in E. coli isolates by affecting the transfer and/or stability of the genetic material that needs to be transferred from one individual to another (Kümmerer, 2004). However, that would be very difficult to show in an in situ experiment such as this one. The positive relationship between rainfall and ciprofloxacin MICs was also observed through velocity, since water velocity correlated with rainfall during the sampling period at most sites in Elk Creek (Maal-Bared, 2008). This is interesting because it indicates that the observed ciprofloxacin resistance may be a result of agricultural runoff. Tetracycline resistance also exhibited a negative relationship with both rainfall and velocity, indicating that runoff was not what affected resistance levels. Most of the nutrients examined (nitrate, phosphate, ammonia and dissolved organic carbon) exhibited negative associations with MICs. A similar increase in susceptibility to ciprofloxacin and tobramycin in Pseudomonas aeruginosa mature biofilms has been reported in an experiment performed by Borriello et al. (2006) when nitrate and arginine were added to a biofilm reactor under anaerobic conditions. A previous study by Borriello et al. (2004) suggested that nitrate addition had opposite effects on antimicrobial susceptibility. The authors explain the difference by emphasizing the difference in biofilm age (or colonization period) in both experiments. This corresponded with work by Anderl et al. (2003), which showed that nutrient limitation, and growth phase of the biofilm decreased the ability of antibiotics to kill cells. Borriello et al. (2004) also suggested that oxygen limitation and anaerobic regions within the biofilms contribute to resistance in young Pseudomonas aeruginosa biofilms. The lower  130  levels of oxygen in the water may explain the increased level of resistance to tetracycline. However, caution must be taken when interpreting these results since the dissolved oxygen and the nutrient concentrations gradients within the biofilms would be quite different from the concentrations in the water column. Also, different E. coli isolates tend to have very different responses to environmental condition changes (Reisner et al., 2006). Finally, biofilm age (colonization period) did have an effect on nalidixic acid and tetracycline resistance patterns. The effects of biofilm age on bactericidal effects and susceptibility have also been shown by some studies (Monzon et al, 2001; Borriello et al., 2003; Borriello et al, 2006). The present study had a variety of limitations. Since we have no estimate of natural resistance in the E. coli population in Elk Creek (even at the headwaters), we assumed that external pressures or addition of already resistant bacteria affected MICs. That may not be the case, especially for tetracycline. No studies have previously investigated the concentrations of antibiotics in the watershed, so we assumed that some concentrations of antibiotics were reaching the creek. Also, this study can only comment on the levels of resistance in the watershed, but it cannot attribute the resistance to any particular source (selective pressures or runoff, for example). The study attempted to associate resistant individuals with particular sites or substrata; however, transport of E. coli was not taken into account. The water quality variables used in this study may only have affected resistance indirectly and many of the variables examined are confounded in the natural environment. Most importantly, the correlations between antibiotic resistant E. coli in biofilms and water quality variables should be interpreted with caution, since the water quality variables measured were snapshots, and do not represent the actual variation the variable would exhibit over a prolonged period of time. In conclusion, both resistant E. coli and resistant pathogenic E. coli were isolated from Elk Creek. The largest number of resistant E. coli isolates was found in the agricultural reach of Elk Creek. Highest levels of resistance were found in sediment and river rock, while Lexan, sandpaper and water carried the lowest numbers of resistant E. coli isolates. Colonization period of the substratum only affected E. coli MICs for nalidixic acid and tetracycline. Also, a variety of water quality variables showed strong relationships with antimicrobial resistance that may be worth examining further, particularly relationships that may be driving factors of antimicrobial resistance of bacteria in drinking water sources. Finally, the contamination of the headwaters with antibiotic resistant E. coli and E. coli O157 provides more evidence that cross-contamination between wildlife and livestock populations may be occurring and may need to be controlled in the  131  future. One method to reduce the high levels of resistance in the agricultural watershed may be more stringent regulation on growth promoting antibiotics and manure management practices in the agricultural industry in the Fraser Valley.  5.5 Acknowledgements This research has been funded by the CIHR Strategic Training Program in Public Health and the Agricultural Rural Ecosystem (PHARE) and Partner Institutes including the Institute of Cancer Research, Institute of Circulatory and Respiratory Health, Institute of Infection and Immunology, and the Institute of Population and Public Health. We would also like to acknowledge the School of Environmental Health where all the microbiological lab analyses were completed. Finally, we would like to thank the University of British Columbia Environmental Engineering Laboratory, particularly Paula Parkinson, for conducting the water nutrient analyses.  132  5.6 References Alexy, R., Kumpel T., Kummerer, K., 2004. Assessment of degradation of 18 antibiotics in the Closed Bottle Test, Chemosphere 57 (6), 505–512. Anderl, J.N., Zahller, J., Roe, F., Stewart, P.S., 2003. Role of nutrient limitation and stationaryphase existence in Klebsiella pneumoniae biofilm resistance to ampicillin and ciprofloxacin. Antimicrobial Agents and Chemotherapy 47,1251-1256. Andrews, J., 2001. Determination of minimum inhibitory concentrations. Journal of Antimicrobial Chemotherapy 48, 5-16. Andersen, S.R., Sandaa, R.A., 1994. Distribution of tetracycline resistance determinants among Gram-negative bacteria isolated from polluted and unpolluted marine sediments. Applied and Environmental Microbiology 60, 908–912 APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater, 20th Edition. American Public Health Association (APHA), American Waterworks Association (AWWA), and Water Environment Association (WEF). Washington, DC. Ash, R., Mauck, B., Morgan, M., 2002. Antibiotic resistance of gram-negative bacteria in rivers, United States. Emerging and Infectious Diseases 8 7, 713–716. Bartoloni, A., Pallecchi, L., Benedetti, M., Fernandez, C., Vallejos, Y., Guzman, E., Villagran, A. L., Mantella, A., Lucchetti, C., Bartalesi, F., Strohmeyer, M., Bechini, A., Gamboa, H, Rodriguez, H., Falkenberg, T., Kronvall, G., Gotuzzo, E., Paradisi, F., Rossolini, G. M., 2006. Multidrug-resistant commensal Escherichia coli in children, Peru and Bolivia. Emerging and Infectious Diseases 12, 907-913. Borriello, G., Werner, E., Roe, F., Kim, A.M., Ehrlich, G.D., Stewart, P.S., 2004. Oxygen limitation contributes to antibiotic tolerance of Pseudomonas aeruginosa in biofilms. Antimicrobial Agents and Chemotherapy 48, 2659-2664. Borriello, G., Richards, L., Ehrlich, G.D., Stewart, P.S., 2006. Arginine or nitrate enhances antibiotic susceptibility of Pseudomonas aeruginosa in biofilms. Antimicrobial Agents and Chemotherapy 50, 382-384. Calomiris, J.J., Armstrong J.L., Seidler R.J., 1984. Association of metal tolerance with multiple antibiotic resistance of bacteria isolated from drinking water. Applied and Environmental Microbiology 47, 1238-1242.  133  Chee-Sanford, J.C., Aminov, R.I., Krapac, I.J., Garrigues-Jeanjean, N., Mackie, R.I., 2001. Occurrence and diversity of tetracycline resistance genes in lagoons and groundwater underlying two swine production facilities. Applied and Environmental Microbiology 67, 1494-1502. Diarra, M.S., Silversides, F.G., Diarrassouba, F., Pritchard, J., Masson, L., Brousseau, R., Bonnet C., Delaquiz, P., Bach, S., Skura, B.J., Topp, E., 2007. Impact of feed supplementation with antimicrobial agents on growth performance of broiler chickens, Clostridium perfringens and Enterococcus number, antibiotic resistant phenotype, and distribution of antimicrobial resistance determinants in Escherichia coli. Applied and Environmental Microbiology 73 (20), 6566-6576 Doi, A.M., Stoskopf, M.K., 2000. The kinetics of oxytetracycline degradation in deionized water under varying temperature, pH, light, substrate, and organic matter. Journal of Aquatic Animal Health 12, 246-253. Duriez, P., Topp, E., 2007. Temporal dynamics and impact of manure storage on antibiotic resistance patterns and population structure of Escherichia coli isolates from a commercial swine farm. Applied and Environmental Microbiology 73, 5486-5493 Edge , T.A., Hill, S., 2005. Occurrence of antibiotic resistance in Escherichia coli from surface waters and faecal pollution sources near Hamilton Ontario. Canadian Journal of Microbiology 51, 501-505. Edwards, M.L., Lilley, A.K., Timms-Wilson, T.H., Thompson, I.P., Cooper, I., 2001 Characterization of the culturable heterotrophic bacterial community in a small eutrophic lake (Priest Pot). FEMS Microbiology Ecology 35, 295–304. Ford, T., 1994. Pollutant effects on the microbial ecosystem. Environmental Health Perspectives 102 (12), 45-48. Gordon, D. M., O'Brien, C. L., 2006. Bacteriocin diversity and the frequency of multiple bacteriocin production in Escherichia coli. Microbiology 152, 3239-3244. Hagedorn, C., Robinson, S.L., Filtz, J.R., Grubbs, S.M., Angier, T.A., Reneau, R.B. Jr., 1999. Using antibiotic resistance patterns in the fecal streptococci to determine sources of fecal pollution in a rural Virginia watershed. Applied and Environmental Microbiology 65, 5522–5531. Hamelin, K., Bruant, G., El-Shaarawi, A., Hill, S., Edge, T.A., Bekal, S., Fairbrother, J.M., Harel, J., Maynard, C., Masson, L., Brousseau, R., 2006. A virulence and antimicrobial resistance DNA microarray detects a high frequency of virulence genes in Escherichia  134  coli isolates from Great Lakes recreational waters. Applied and Environmental Microbiology 72, 4200-4206. Hamelin, K., Bruant, G., El-Shaarawi, A., Hill, S., Edge, T. A., Fairbrother, J., Harel, J., Maynard, C., Masson, L., Brousseau, R., 2007. Occurrence of virulence and antimicrobial resistance genes in Escherichia coli isolates from different aquatic ecosystems within the St. Clair River and Detroit River areas. Applied and Environmental Microbiology 73, 477-484 Hamscher, G., Sczesny, S., Höper, H., Nau, H., 2002. Determination of persistent tetracycline residues in soil fertilized with liquid manure by high-performance liquid chromatography with electrospray ionization tandem mass spectrometry. Analytical Chemistry 74, 1509– 18. Hancock, D.D., Besser, T.E., Rice, D.H., Ebel, E.D., Herriott D.E., Carpenter L.V., 1998. Multiple sources of Escherichia coli O157 in feedlots in the Northwestern USA. Preventative Veterinary Medicine 35, 11-19. Hirsch, R., Ternes, T., Haberer, K., Kratz, K.L., 1999. Occurrence of antibiotics in the aquatic environment. Science of the Total Environment 225, 109-118. Issa, A.A., 1999. Antibiotic production by the cyanobacteria Oscillatoria angustissima and Calothrix parietina. Environmental Toxicology and Pharmacology 8, 33–37. Khachatryan, A.R., Besser, T.E., Hancock, D.D., Call, D.R., 2006. Use of a non-medicated dietary supplement correlates with increased prevalence of streptomycin-sulfatetracycline-resistant Escherichia coli on a dairy farm. Applied and Environmental Microbiology 72, 4583-4588. Kümmerer, K., 2003. The significance of antibiotics in the environment. Journal of Antimicrobial Chemotherapy 52, 5–7 Kümmerer, K., 2004. Resistance in the environment. Journal of Antimicrobial Chemotherapy 54, 311-320. Lejeune, P., 2003. Contamination of abiotic surfaces: what a colonizing bacterium sees and how to blur it. Trends in Microbiology 11, 179–184. LeFlaive, J., Ten-Hage, L., 2007. Algal and cyanobacterial secondary metabolites in freshwaters: a comparison of allelopathic compounds and toxins. Freshwater Biology 52 (2), 199–214. Maal-Bared, R., Bartlett, K.H., Bowie, W.R., Hall, E.R., Hall, K.J., 2007. Seasonal and spatial distribution of indicators and pathogens in biofilms in an agricultural watershed. Annual Conference Proceedings of the American Waterworks Association, Toronto 2007.  135  Maal-Bared, R., 2008. Comparing the distribution of pathogenic bacteria and common indicator microorganisms in biofilms on different surface types in an agricultural watershed in British Columbia (Canada). Unpublished PhD Thesis, University of British Columbia, British Columbia, Canada. Martínez-Carballo, E., González-Barreiro, C., Scharf, S., Gans., O., 2007. Environmental monitoring study of selected veterinary antibiotics in animal manure and soils in Austria. Environmental Pollution 148 (2), 570-579 Matsui, K., Ishii, N., Kawabata Z., 2003. Release of extracellular transformable plasmid DNA from Escherichia coli cocultivated with algae. Applied and Environmental Microbiology 69, 2399-2404. McKeon, D.M., Calabrese, J.P., Bissonnette, G.K., 1995. Antibiotic resistant gram-negative bacteria in rural groundwater supplies. Water Research 29, 1902–1908. Meays, C.L., Broersma, K., Nordin, R., Mazumder, A., Samadpour, M., 2006. Spatial and annual variability in concentrations and sources of Escherichia coli in multiple watersheds. Environmental Science & Technology 40, 5289-5296. Monzon, M., Oteiza, C., Leiva, J., Amorena, B., 2001. Synergy of different antibiotic combinations in biofilms of Staphylococcus epidermidis. Journal of Antimicrobial Chemotherapy 48, 793-801. Murray, B. E., 1997. Antibiotic resistance. Advances in Internal Medicine 42, 339–367. National Committee for Clinical Laboratory Standards, 2000. Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically—Sixth Edition: Approved Standard M7-A5. NCCLS, Wayne, PA, USA. Nicholson, F.A., Smith, S.R., Alloway, B.J., Carlton-Smith, C., Chambers, B.J., 2003. An inventory of heavy metals inputs to agricultural soils in England and Wales, Science of the Total Environment 311, 205–219. Olapade, O.A., Leff, L.G., 2006. Influence of dissolved organic matter and inorganic nutrients on biofilm bacterial community on artificial substrates in a north eastern Ohio, USA, stream. Canadian Journal of Microbiology 52, 540-9. Pedersen, J., Yeager, M., Suffet, I., 2003. Xenobiotic organic compounds in runoff from fields irrigated with treated wastewater. Journal of Agricultural and Food Chemistry 51, 1360– 1372. Quinn, G.P. Keogh, M.J., 2002. Experimental Design and Data Analysis for Biologists. Cambridge University Press, Cambridge, England.  136  Ray S., Bagchi S.N., 2001. Nutrients and pH regulate algicide accumulation in culture of the cyanobacterium Oscillatoria laetevirens. New Phytologist 149, 455–460. Reisner, A., Krogfelt, K.A., Klein, B.M., Zechner E.L., Molin, S., 2006. In vitro biofilm formation of commensal and pathogenic Escherichia coli strains: impact of environmental and genetic factors. Journal of Bacteriology 188, 3572-3581. Renter, D.G., Sargeant, J.M., Hyngstorm, S.E., Hoffman, J.D., Gillespie, J.R., 2001. Escherichia coli in free-ranging deer in Nebraska. Journal of Wildlife Diseases 37, 755-760. Renter, D.G., Sargeant, J.M., Oberst, R.D., Samadpour, M., 2003. Diversity, frequency, and persistence of Escherichia coli O157 strains from range cattle environments. Applied and Environmental Microbiology 69, 542-547 Richardson, M.L, Bowron, J.M., 1985. The fate of pharmaceutical chemicals in the aquatic environment. Journal of Pharmacy and Pharmacology 37, 1–12. Riley, M.A., Gordon, D.M., 1996. The ecology and evolution of bacteriocins. Journal of Industrial Microbiology and Biotechnology 17, 151–158. Rooklidge, S.J., 2004. Environmental antimicrobial contamination from terraccumulation and diffuse pollution pathways. Science of the Total Environment 325, 1-13. Salyers, A.A., Shoemaker, N., 1996. Resistance gene transfer in anaerobes: new insights, new problems. Clinical Infection and Disease 23 (1), 36-43. Samuelson, O.B., Torsvik, V., Ervik, A., 1992. Long-range changes in oxytetracycline concentration and bacterial resistance towards oxytetracycline in fish farm sediment after medication. Science of the Total Environment 114, 25–36 Sayah, R.S., Kaneene, J.B., Johnson, Y., Miller, R.A., 2005. Patterns of antimicrobial resistance observed in Escherichia coli isolates obtained from domestic and wild-animal fecal samples, human septage, and surface water. Applied and Environmental Microbiology 71, 1394–1404. Schwartz, T., Kohnen, T., Jansen, B., Obst, U., 2003. Detection of antibiotic-resistant bacteria and their resistance genes in wastewater, surface water, and drinking water biofilms. FEMS Microbiology Ecology 43, 325–35. Seurinck, S., Verstraete, W., Siciliano S.D., 2003. Use of 16S-23S rRNA intergenic spacer region PCR and repetitive extragenic palindromic PCR analyses of Escherichia coli isolates to identify nonpoint fecal sources. Applied and Environmental Microbiology 69, 4942-4950. Shere, J.A., Bartlett, K.J., Kaspar, C.W., 1998. Longitudinal study of Escherichia coli O157:H7 dissemination on four dairy farms in Wisconsin. Applied and Environmental Microbiology 64, 1390-1399.  137  Stewart, P.S., Costerton, J.W., 2001. Antibiotic resistance of bacteria in biofilms. Lancet 258, 135-138. Sunde, M., Norstrom, M., 2005. The genetic background for streptomycin resistance in Escherichia coli influences the distribution of MICs. Journal of Antimicrobial Chemotherapy 56, 87–90. Vingarzan, R., Belzer, W., Thompson, B., 2002. Nutrient levels in the atmosphere of the Elk Creek Watershed, Chilliwack, BC (1999-2000). Environment Canada Technical Report. Aquatic and Atmospheric Sciences Division, Vancouver, Canada. von Elert, E., Jüttner, F., 1997. Phosphorus limitation and not light controls the extracellular release of allelopathic compounds by Trichormus doliolum. Limnology and Oceanography 42, 1796–1802. Walk, S.T., Mladonicky, J.M., Middleton, J.A., Heidt, A.J., Cunningham, J.R., Bartlett, P., Sato, K., Whittam, T.S., 2007. Influence of antibiotic selection on genetic composition of Escherichia coli populations from conventional and organic dairy farms. Applied and Environmental Microbiology 73, 5982-5989. Watkinson, A.J., Micalizzi, G.B., Graham, G.M., Bates, J.B., Costanzo, S.D., 2007. Antibioticresistant Escherichia coli in wastewaters, surface waters, and oysters from an urban riverine system. Applied and Environmental Microbiology 73 (17), 5667-5670. Wilkerson, C., Samadpour, M., van Kirk, N., Roberts, M.C., 2004. Antibiotic resistance and distribution of tetracycline resistance genes in Escherichia coli O157:H7 isolates from humans and bovines. Antimicrobial Agents and Chemotherapy. 48, 1066-1067. Williams H.G., Day, M.J., Fry, J.C., Stewart, G.J., 1996. Natural transformation in river epilithon. Applied and Environmental Microbiology 62 (8), 2994-2998. Witte, W. 2000. Ecological impact of antibiotic use in animals on different complex microflora: environment. International Journal of Antimicrobial Agents 14, 321-325.  138  6  Discussion  6.1 Summary of results and status of working hypotheses This exploratory study was designed to examine the distribution of indicator bacteria (HPC, FC, EC, and ENT) and common waterborne pathogens (pathogenic E. coli 0157, Campylobacter sp. and Salmonella sp.) in biofilms on different surface types (river and slate rock, wood, sandpaper, and LexanTM), as well as in water and sediment. Samples were collected from the Elk Creek intensive agricultural watershed (British Columbia, Canada) between December 2005 and April 2007 from four sites along the stream. The ultimate goal of the study was to improve our understanding of the concentration of waterborne disease-related pathogens and commonly used bacterial indicator organisms in different aquatic environment compartments (see Figure 6.1). This information is relevant for several reasons. Currently, microbial water quality monitoring protocols utilize water column grab samples to ensure public health protection. If this investigation revealed that pathogens and indicators are found in higher numbers in biofilm than in the water column and that these pathogens could resuspend into the water column following water quality and quantity changes, then the results of this study could be used to improve our monitoring and sampling techniques. This information would be particularly useful to water managers if raw water testing becomes mandatory (Dorner et al., 2004). While this watershed is not used for drinking water purposes anymore, site one was used for that purpose in the past and the results from that site are of interest to individuals investigating contamination in drinking water streams. Also, the presence of pathogens in Elk Creek could affect livestock, wildlife, fish, ecosystem, and human health. There were several driving hypotheses and research questions in this study, which are all summarized along with the main findings in Figure 6.1. The first hypothesis was that indicator bacteria and pathogens would prefer the more protected and beneficial environment of biofilms, as opposed to that of the water column, unless environmental changes result in detachment. To test this hypothesis the distribution of indicator organisms and pathogens in water column grab samples was compared to the distribution of indicator organisms and pathogens in biofilms on different substratum surface types and sediments at different sites in Elk Creek and during different seasons (see Figure 6.1).  139  Do biofilms act as sinks for indicator organisms and pathogens in a natural aquatic environment? If yes, should biofilms be included in water source monitoring protocols to protect animal, wildlife, ecosystem and public health? Was there a difference in the distribution of indicator organisms and pathogens in biofilms on different substrata, and in water and sediment? Q. Was there a difference in the distribution of indicators and pathogens in biofilms on different substratum surface types, and in water and sediment during different seasons? YES. The distribution of indicators and pathogens on different surfaces was different. No pathogens were found in water in the dry season. Biofilms on different surfaces became sinks for pathogens.  Q. Did Campylobacter exhibit any patterns different from other pathogens?  Q. Was there a difference in the distribution of indicators and pathogens in biofilms on different substratum surface types, and in water and sediment at different sites?  Q. Was there a difference among the concentrations of faecal indicator bacteria and prevalences of pathogens in biofilms on different substratum surface types based on biofilm age (or substratum colonization period)?  YES. Both biofilms and sediment contained indicator numbers and pathogens frequencies that were different from those of water. Yes. Campylobacter was not retrieved at all during the dry season and it was not found in sandpaper biofilms.  Yes. Different numbers of indicators and pathogens were recovered from different substrata at different weeks. Sediment and wood accumulated the highest numbers. Two to four weeks of accumulation usually resulted in the highest retrieval numbers.  Q. Was there a difference in the prevalences of antibiotic resistant E. coli in biofilms on different substratum surface types, sediment and in the water column?  Yes. Sediment and river rock biofilms contained the highest number of resistant E. coli.  Q. Did changes in water quality parameters (physical, chemical and biological) correlate with any of these findings?  Campylobacter frequencies in slate rock and wood biofilms correlated with enterococci numbers. Sediment and wood Campylobacter frequencies correlated with nitrates in water. Sediment, slate rock and wood are the best substrata for monitoring.  Water quality variables showed different correlations with indicator and pathogen numbers depending on the type of colonized substratum type and site.  Young biofilms correlated positively with flow measures and nutrients, but negatively with conductivity. Old biofilms correlated negatively with rainfall and nutrients.  Yes. Water quality variables did correlate with E. coli MICs for different antibiotics. Among these variables were rainfall, substratum colonization period, dissolved oxygen levels and nutrient concentrations.  Figure 6.1 Summary of research questions and results found in this study. Double lined boxes indicate major research questions asked during the study.  140  140  Results in Chapter 2 showed that using water column grab samples alone underestimated microbial burdens at each site and led to misleading conclusions about faecal contamination at some sites. For example, no indicator E. coli was retrieved from site 4 using water column grab samples. The use of biofilms and sediment, however, revealed the presence of E. coli in high concentrations. Thus, the hypothesis that attached bacterial survival is the more preferred form of life in situ as opposed to planktonic survival stands. Similar findings have been reported in the literature on bacterial concentrations in sediment (Muirhead et al., 2004; Crabill et al., 1999; Obiri-Danso and Jones, 2000) and in biofilms in drinking distribution systems (Rittman et al., 2000). To my knowledge, similar results have not been found in biofilms in natural aquatic systems.  Also in Chapter 2, the results showed that the heterotrophic plate counts and enterococci numbers on different substratum surface types were significantly different during both the wet and dry season. Another interesting finding in Chapter 2 was that there was little difference in the prevalence of pathogens between water, sediment and biofilms in the wet season. In the dry season, however, more pathogens were isolated from biofilms than from sediment and no pathogens were isolated from water. This indicated that biofilms might be the main reservoir of Salmonella sp. and pathogenic E. coli O157 during the summer when rainfall is limited. This could be due to the role agricultural runoff (after rainfall events) plays in freshwater contamination with faecal bacteria and pathogens in the vicinity of agricultural lands, especially if rain falls after manure application to the land (Gilpin et al., 2002; Duriez and Topp, 2007; Shanks et al, 2006; Dorner et al., 2004). Muirhead et al. (2004) who investigated how in-stream flooding affected release of E. coli from sediment and river rock biofilms into the water column found that the contribution of sediment was much higher than that of biofilms from river rocks. However, the investigators only examined the concentration of indicator bacteria in that study. It was also hypothesized that the agricultural reach of Elk Creek would have higher inputs of faecal indicators and pathogens as a result of agricultural practices, while site one would remain relatively pristine. The hypothesis that the headwaters would be pristine due to the absence of agricultural activity was rejected. As was seen in Chapters 2, 3 and 5, site one was contaminated with high concentrations of faecal indicator bacteria and both E. coli O157 and Salmonella, as well as with ampicillin and tetracycline resistant E. coli. The faecal contamination of headwaters could be caused by wildlife (Meays et al., 2006). The presence of pathogenic E. coli O157 in the headwaters, in particular, could have been the result of contamination crossover between livestock in the agricultural ecosystem and the wildlife population (Sayah et al., 2005; Hagedorn et al., 1999), since ruminants are considered the main reservoirs of E. coli O157:H7.  141  Contamination of wildlife feces (deer, possums, racoons and birds) with pathogenic E. coli O157 has been repeatedly reported in the literature (Renter et al., 2001; Renter et al., 2003; Shere et al., 1998; Hancock et al., 1998). The present study supports the hypothesis that cross-contamination between livestock and wildlife is occurring (Hancock et al., 1998; Seurinck et al., 2003). In Chapter 3, the relationships between water quality variables and distribution of indicators and pathogens on different surface types were examined. The results showed that the relationships between water quality and microbial variables in one-month biofilms were different on different substrata. Also, some relationships between water quality and microbial variables changed among sites on the same substratum. For example, some microbial variables were positively correlated with conductivity, nutrient concentrations (such as nitrate and dissolve organic carbon) or water velocity at site 1, but correlated negatively with the same variable at site 4. This may imply that biofilms growing in eutrophic environments, like site 4, grow faster and become more vulnerable to environmental changes. Numerous positive relationships between protozoa and biofilm microbial variables were found as well. Chapter 3 also investigated the influence of biofilm age (colonization period) on indicator numbers and pathogen prevalence on different surfaces. Different substrata peaked in concentrations of indicators and pathogens at different points in time. Wood and sediment accumulated the highest numbers of pathogens and indicators. The study also indicated that a two to four week period of colonization would be optimal to study indicator and pathogen distributions in biofilms, which is comparable to results by Hunt and Parry (1998). In young biofilms (substrata colonized for 1 to 3 weeks), indicators and pathogen numbers correlated directly with flow and nutrients, but exhibited negative relationships with depth and conductivity. In older biofilms (substrata colonized for 12 or 24 weeks), indicators and pathogen numbers started showing negative correlations with nutrients and rainfall, but positive correlations with protozoa and dissolved oxygen. These relationships may again be the result of biofilm needs at a particular point in succession (Jackson, 2003) and development. The positive relationship between young biofilms and flow variables may be a reflection of rainfall runoff adding more bacteria to the water column that can be sequestered to the growing biofilm community. However, high flow and rainfall in older biofilms results in sloughing (Brading et al., 1995). Also, the negative relationship between older biofilms and nutrients may be due to the fact that sloughing occurs most commonly in nutrient rich environments and thick biofilms (Ohashi and Harada, 1996). It is also worth mentioning that positive relationships between indicators, pathogens and protozoa numbers were observed in this study as well. Chapter 3 highlights that biofilms colonizing different surfaces in natural aquatic environments are complex ecosystems  142  within themselves that are highly affected by their external environment and can be substantially different from each other. These types of differences need to be taken into account if biofilms were used as monitoring tools. Chapter 4 showed that Campylobacter was not retrieved from the stream in the dry season at any of the sites and it was not retrieved from sandpaper biofilms throughout the study period. Considering that Campylobacter is the main cause of bacterial waterborne gastroenteritis in Canada (Schuster et al., 2005), an analysis of variables that may have affected the prevalence of Campylobacter sp. in different biofilms, sediment and water was conducted in Chapter 4. The use of growth-based isolation methods, as opposed to molecular techniques, may have resulted in the low Campylobacter recovery rates in the wet season, as well as the pathogen’s complete absence during the dry season (Kemp et al., 2005; Lehtola et al., 2006; Waage et al., 1999). Also, the decrease in rainfall and the change in environmental conditions may have contributed to die off rates of this pathogen. Substratum type did affect the prevalence of Campylobacter sp. in the stream, being highest in sediment, slate rock and wood. From this analysis, we suggest that sediment, slate rock or wood be used as a substratum for Campylobacter sp. monitoring, since Campylobacter presence on those surfaces also correlated with nitrate concentrations in the water column and enterococci in the same biofilms. The final substratum of choice should depend on what other water quality parameters are tested for regularly. Also, the use of nitrates and enterococci as potential faecal contamination markers in freshwater may be of interest to water managers and public health inspectors, since these two variables correlated very well with Campylobacter frequencies in the stream in the wet season.  Finally, antibiotic resistance was hypothesized to be highest in the agricultural reach of the watershed due to the high levels of resistant bacterial inputs, as well as high selective pressures. This hypothesis still stands since the highest prevalences of antibiotic resistant E. coli were retrieved from sites 2,3 and 4 in the agricultural watershed. The effects of agriculture on antimicrobial resistance have often been documented in the literature (Witte, 2000; Kümmerer, 2004). E. coli isolates showed highest levels of resistance to tetracyclines, followed by ampicillin, streptomycin and nalidixic acid. These findings correspond with Edge and Hill (2005), who found high levels of ampicillin and tetracycline resistance in E. coli isolates from surface water in Hamilton (Ontario) and low levels of resistance to ciprofloxacin. High levels of resistance to ampicillin were expected due to fact that ampicillins are older antibiotics that have been extensively used over the years (Rooklidge, 2004). Antibiotic resistant E. coli O157 were isolated from all sites. Sediment and river rocks were the most likely substrata to be associated with  143  resistant E. coli, while water was the least likely examined medium to be associated with resistant E. coli. Also, nutrient concentrations (nitrate, phosphate, ammonia and dissolved organic carbon) mostly had negative effects on minimum inhibitory concentrations of E. coli isolates, while the effects of other variables depended on the antibiotic tested. Biofilm colonization period only affected resistance levels for tetracycline and nalidixic acid. The effect of biofilms age on organism susceptibility has previously been found in lab-based studies (Borriello et al., 2004; Borriello et al., 2006; Anderl et al., 2003). In general, this study suggests that different surfaces in natural aquatic systems can accumulate different concentrations of indicator bacteria and pathogens, and act as microbial sinks under the right conditions. In the wet season, runoff and sediment transport may have been the main contributors to indicator and pathogen loading in the watershed. In the dry season, biofilms played an increasingly important role in pathogen retention, which was not reflected in measuring indicator concentrations. The monitoring of the water column remains the most direct measure of the risk to the consumers, but perhaps monitoring biofilms during the dry season may be beneficial because it includes protecting public health from potential risk.  6.2 Significance of the research This is the first study exploring the distribution of indicators and pathogens in biofilms on a variety of surfaces in a natural aquatic system and it makes several original contributions to the literature. Most importantly, the study showed that during the dry season, biofilms and sediment are important sinks for indicator bacteria and pathogens that could potentially be used to predict water quality. Previous studies investigating pathogens presence in natural aquatic environments have usually focused on sediment concentrations relative to water column grab samples, but not on biofilms. This study suggests that biofilms have higher pathogen concentrations and more genera of pathogens than sediments did in Elk Creek during the dry season. Also, only one study has investigated the distribution of organisms in sediment relative to biofilms (Muirhead et al., 2004). That study only used indicator E. coli numbers for comparison. The findings from the present study, like those of Muirhead et al. (2004), showed that sediment contributes higher numbers of E. coli to a stream during an in-stream flood event compared to river rock biofilms, but the same was not true to actual pathogens. In Chapter 3, 4 and 5, this work revealed a multitude of relationships between physical and chemical variables and indicator numbers and pathogen frequencies, such as E. coli O157,  144  Campylobacter and Salmonella in biofilms in situ. Conducting these types of experiments in situ is quite rare. The field observations provide a variety relationships between indicators, pathogens and water quality variables that laboratory-based studies (mesocosm and microcosm studies) could examine under more controlled conditions. For example, Salmonella in biofilms in Chapter 3 correlated positively with phosphate, but negatively with depth. In Chapter 4, Campylobacter seemed to exhibit more positive relationships with protozoa counts than negative ones. The negative relationships, discussed in Chapter 5, between MICs for nalidixic acid and tetracyclines with dissolved organic carbon have not previously been reported and may be of interest in watersheds where the water is used for drinking purposes. In Chapter 4, a platform was presented that can be used to compare the effectiveness of different substratum surface types as monitoring surfaces for particular pathogens. The method utilized several criteria such as: Campylobacter sp. correlation with indicator bacteria presence in the biofilm, correlations with water quality parameters, cost and availability, potential for standardizing substratum, ease of biofilm removal from substratum for analysis and probability of substratum loss in situ. This method can easily be reproduced if water managers or public health officials want to determine what substratum to use to monitor for a pathogen of interest at a water source. Finally, this study suggested that wildlife might have negative effects on water sources, which should be investigated further. Like Elk Creek, other watersheds with forested headwaters, which may be used for drinking water purposes, may have high wildlife densities, and may thus be exposed to high levels of E. coli (resistant and susceptible), E. coli O157 (resistant and susceptible), and Salmonella. If we start also considering the potential for cross-contamination of antibiotic resistance in the watershed between wildlife and livestock, we quickly realize this phenomenon is a threat to public health and needs to be addressed.  6.3 Strengths and limitations of the research The study has a variety of limitations, many of which are related to the fact that it was a fieldbased study. Due to sampler and substrata loss in situ, the sample size of this study was fairly small (particularly in the case of short- and long-term biofilm observations). Many of our tests were statistically insignificant, which could be related to the fact that showing statistical significance in field studies is usually more difficult than in laboratory-based systems. Considering the alpha value selected for statistical testing and the actual number of tests performed, it is likely that some of our statistically significant results were significant by chance (Type I statistical error-false positive).  145  Since this was a field-based study, many variables that were not tested may still have had some effects on indicators and pathogens. Among those variable were algal biomass and more precise protozoa concentrations in the water column and the biofilm itself. Some physical factors, such as solar radiation and turbidity, would also have had strong effects on pathogen inactivation and resuspension but were not measured. Turbidity has been correlated with outbreak data (Payment, 2003). Also, the present study does not include estimates of viral lysis and concentrations of toxic chemicals and heavy metals, which might affect biofilm community structure and dynamics. Some of the variables analyzed in Chapters 3, 4 and 5 are clearly confounded and must be interpreted with care. Also, the water quality variables, which were correlated with indicator numbers, pathogen detection frequencies and MICs, do not reflect the different fluctuations of the variables the biofilm microbial community would have been exposed to over the colonization periods. The present study attempted to minimize major variations in water quality variables by allowing the biofilms to colonize within the same season (particularly for long-term samplers). This study requires replication in other watersheds to see if observed relationships hold. If repeated, it may be beneficial to also collect data about disease outbreaks in human and/or animal populations utilizing the watershed, although researchers need to remember that associating an outbreak with a particular source is very difficult. Since the Elk Creek watershed is no longer used for drinking water purposes, it was not possible to make connections between biofilm indicator concentrations and pathogen prevalence, and disease outbreaks in human and/or animal populations. The study does have several strengths. The study was conducted in situ, which makes the findings more applicable to real environments. Also, the study investigated the frequencies of some of the most common pathogens associated with waterborne disease outbreaks in rural and agricultural watersheds in Canada. The study investigated the survival of these pathogens in biofilms. Biofilms are rarely tested ecosystems in which pathogen survival has often been hypothesized to increase. The use of growth-based methods to isolate these pathogens may have resulted in low prevalences, but it also makes the results relevant from a public health perspective. The use of molecular techniques also includes viable, non-culturable organisms in recovery rates, the infective role of which is still unknown. However, the use of molecular methods would have allowed us to identify and quantify the pathogens to the species level with high levels of accuracy and in a shorter period of time.  146  6.4 Future research As previously mentioned, this study was exploratory and thus replication is required to test our hypotheses in other watersheds. If these future study watersheds are used for drinking water sourcing, collection of data related to outbreaks in human populations may be of interest. It may be beneficial to test some of the observed relationships between indicators, bacteria and water quality variables in laboratory settings to see if the observations hold under controlled conditions. Future research should also consider performing some continuous water quality variable monitoring while the biofilms are being grown to examine relationships between indicator and pathogen concentrations in biofilms and water quality variables more closely. However, it is still important to remember that while water quality parameters were useful in the interpretation of relationships among variables, chemical and physical characteristics within the biofilm itself are usually very different from the external aquatic environment, and even vary within the biofilm itself (Rittmann, 2004). It would be beneficial if future research also took into account the contribution of external inputs into the creek such as runoff and septic tank leakage, to be able to assess the role biofilms play in the retention of bacterial pathogens during the high flow season. This could be accomplished by directly examining sources, such as runoff, for microbial contaminants and nutrient concentrations. The effects could also be estimated through the determination of dominant land uses in the watersheds in the proximity of different sites. Future research should focus on substrata that successfully accumulated pathogens of interest. This may eliminate the need for standardization among different substrata types and facilitate the investigation. Future studies should include measurement of some of the variables that may affect pathogen and indicator survival within the biofilm, such as algal concentrations, fungal concentrations and solar radiation. Also, while the present study examined the presence of bacterial pathogens and indicators in biofilms, the presence of Cryptosporidium oocysts survival has been shown in biofilms (Searcy et al., 2006; Momba et al., 2000). Thus, investigating the presence of protozoan parasites may be interesting since they cause even higher outbreak rates in Canada than bacterial pathogens (Schuster et al., 2005), are harder to eliminate from treated drinking water due to small size and resistance to oxidation (Health Canada, 2004) and are carried by wildlife populations in headwaters (Cacció et al., 2005). Well water in rural and agricultural communities in Canada is also used as a drinking water source and is often contaminated with runoff, thus a similar study could be conducted in well water systems. Finally, there is a wide  147  array of emerging waterborne pathogens of concern, such as Helicobacter pylori, Mycobacterium avium complex, Cyclospora, Toxoplasma, enteric viruses, and Legionella (Percival et al., 2000; Momba et al., 2000; Huffman et al., 2003) whose survival in freshwater biofilms could be investigated as well. The question becomes what method of growth and detection is chosen to examine pathogen presence and prevalence in the water source. Molecular methods can detect organisms in viable non-culturable states, while standard culturing and isolation techniques do not. However, sophisticated, molecular-based techniques are less likely to be available to small water system operators who may be interested in finding out whether their raw water source is contaminated with pathogens. Also, the role of viable, non-culturable bacteria in infection remains debated and unclear (Thomas et al., 1999; Reissbrodt et al., 2002). So, the choice of technique must be based on the research question and organism of interest. It is clear, however, that there is definitely more research that could be conducted in the investigation of biofilms as sinks for pathogens in natural aquatic ecosystems.  148  6.5 References Anderl, J.N., Zahller, J., Roe, F., Stewart, P.S., 2003. Role of nutrient limitation and stationaryphase existence in Klebsiella pneumoniae biofilm resistance to ampicillin and ciprofloxacin. Antimicrobial Agents and Chemotherapy 47,1251-1256. Borriello, G., Werner, E., Roe, F., Kim, A.M., Ehrlich, G.D., Stewart, P.S., 2004. Oxygen limitation contributes to antibiotic tolerance of Pseudomonas aeruginosa in biofilms. Antimicrobial Agents and Chemotherapy 48, 2659-2664. Borriello, G., Richards, L., Ehrlich, G.D., Stewart, P.S., 2006. Arginine or nitrate enhances antibiotic susceptibility of Pseudomonas aeruginosa in biofilms. Antimicrobial Agents  and Chemotherapy 50, 382-384. Brading, M.G., Jass, J., Lappin-Scott, H.M., 1995. Dynamics of bacterial biofilm formation. In: Microbial Biofilms. Cambridge, University Press, 46-63. Cacció, S.M., Tompson, R.C.A., McLauchlin, J., Smith, H.V., 2005. Unravelling Cryptosporidium and Giardia epidemiology, Trends in Parasitology 21, 430–437. Crabill, C., Donald, R., Snelling, J., Foust, R., Southam, G., 1999. The impact of sediment faecal coliform reservoirs on seasonal water quality in Oak Creek, Arizona. Water Research 33 (9), 2163-2171. Dorner, S.M., Huck, P.M., Slawson, R.M., Gaulin, T., Anderson, W.B., 2004. Assessing levels of pathogenic contamination in a heavily impacted river used as a drinking water source. Journal of Toxicology and Environmental Health: Part A 67, 1813-1823. Duriez, P., Topp, E., 2007. Temporal dynamics and impact of manure storage on antibiotic resistance patterns and population structure of Escherichia coli isolates from a commercial swine farm. Applied and Environmental Microbiology 73, 5486-5493 Edge , T.A., Hill, S., 2005. Occurrence of antibiotic resistance in Escherichia coli from surface waters and faecal pollution sources near Hamilton Ontario. Canadian Journal of Microbiology 51, 501-505. Gilpin, B.J., Gregor, J.E., Savill, M.G., 2002. Identification of the source of faecal pollution in contaminated rivers. Water, Science and Technology 46 (3), 9-15. Hagedorn, C., Robinson, S.L., Filtz, J.R., Grubbs, S.M., Angier, T.A., Reneau, R.B. Jr., 1999. Using antibiotic resistance patterns in the fecal streptococci to determine sources of fecal pollution in a rural Virginia watershed. Applied and Environmental Microbiology 65, 5522–5531.  149  Hancock, D.D., Besser, T.E., Rice, D.H., Ebel, E.D., Herriott, D.E., Carpenter, L.V., 1998. Multiple sources of Escherichia coli O157 in feedlots in the Northwestern USA. Preventative Veterinary Medicine 35, 11-19. Health Canada, 2004. Protozoa. In: Canadian Guidelines for Drinking Water Quality: supporting documents. Health Canada, Ottawa. Retrieved on January 6, 2007. Report available at: http://www.hc-sc.gc.ca/ewh-semt/pubs/water-eau/doc_supappui/protozoa/index_e.html Huffman, D.E., Quinter-Betancourt, W., Rose, J., 2003. Emerging waterborne pathogens. In: Mara, D., Horan, N. (Eds.) The Handbook of Water and Wastewater Microbiology. London, Elsevier, 193-207. Hunt, A.P., Parry, J.D., 1998. The effect of substratum roughness and river flow rate on the development of a freshwater biofilm community. Biofouling 12 (4), 287-303. Jackson, C.R., 2003. Changes in community properties during microbial succession. Oikos 101 (2), 444-448. Kemp , R., Leatherbarrow, A.J., Williams, N.J., Hart, C.A., Clough, H.E., Turner, J., Wright, E.J., French, N.P., 2005 Prevalence and genetic diversity of Campylobacter spp. in environmental water samples from a 100-square-kilometer predominantly dairy farming area. Applied and Environmental Microbiology 71 (4), 1876-82. Kümmerer, K., 2004. Resistance in the environment. Journal of Antimicrobial Chemotherapy 54, 311-320. Lehtola, M.J., Pitkanen, T., Miebach, L., Miettinen, I.T., 2006. Survival of Campylobacter jejuni in potable water biofilms: A comparative study with different detection methods. Water, Science and Technology 54 (3), 57-61. Meays, C.L., Broersma, K., Nordin, R., Mazumder, A., Samadpour, M., 2006. Spatial and annual variability in concentrations and sources of Escherichia coli in multiple watersheds. Environmental Science & Technology 40, 5289-5296. Momba, M.N.B., Kfir, R., Venter, S.N., Cloete, T.E., 2000. An overview of biofilm formation in distribution systems and its impact on the deterioration of water quality. Water SA 26 (1): 59-66. Muirhead, R.W., Davies-Colley, R.J., Donnison, A.M., Nagels, J.W., 2004. Faecal bacteria yields in artificial flood events: quantifying in-stream stores. Water Research 38, 1215-1224. Obiri-Danso, K., Jones. K., 2000. Intertidal sediments as reservoirs for hippurate negative Campylobacters, Salmonellae and faecal indicators in three EU recognised bathing waters in North West England. Water Research 34(2), 519- 527.  150  Ohashi, A., Harada, H., 1996. A novel concept for evaluation of biofilm adhesion strength by applying tensile force and shear force. Water, Science and Technology 34 (5-6), 201211. Payment, P., 2003. Health effects of water consumption and water quality. In: Mara, D., Horan, N. (Eds.) The Handbook of Water and Wastewater Microbiology. London, Elsevier, 209-219. Percival, S.L., Walker, J.T., Hunter, P.R., 2000. Microbiological Aspects of Biofilms and Drinking Water. CRC Press, Boca Raton. Reissbrodt, R., Rienaecker, I., Romanova, J.M., Freestone, P.P.E., Haigh, R.D., Lyte, M., Tschäpe, H., Williams, P.H., 2002. Resuscitation of Salmonella enterica serovar Typhimurium and enterohemorrhagic Escherichia coli from the viable but nonculturable state by heat-stable enterobacterial autoinducer. Applied and  Environmental Microbiology 68, 4788-4794 Renter, D.G., Sargeant, J.M., Hyngstorm, S.E., Hoffman, J.D., Gillespie, J.R., 2001. Escherichia coli in free-ranging deer in Nebraska. Journal of Wildlife Diseases 37, 755-760. Renter, D.G., Sargeant, J.M., Oberst, R.D., Samadpour, M., 2003. Diversity, frequency, and persistence of Escherichia coli O157 strains from range cattle environments. Applied and Environmental Microbiology 69, 542-547 Rittmann, B.E., 2004. Biofilms in the water industry. In: Ghannoum, M., O’Toole, G.A. (Eds.) Microbial Biofilms. ASM Press, Washington, DC, 359-78. Rooklidge, S.J., 2004. Environmental antimicrobial contamination from terraccumulation and diffuse pollution pathways. Science of the Total Environment 325, 1-13. Sayah, R.S., Kaneene, J.B., Johnson, Y., Miller, R.A., 2005. Patterns of antimicrobial resistance observed in Escherichia coli isolates obtained from domestic- and wild-animal fecal samples, human septage, and surface water. Applied and Environmental Microbiology 71, 1394–1404. Searcy, K.E., Packman, A.I., Atwill, E.R., Harter, T., 2006. Capture and retention of Cryptosporidium parvum oocysts by Pseudomonas aeruginosa biofilms. Applied and Environmental Microbiology 72, 6242-6247. Seurinck, S., Verstraete, W., Siciliano S.D., 2003. Use of 16S-23S rRNA intergenic spacer region PCR and repetitive extragenic palindromic PCR analyses of Escherichia coli isolates to identify nonpoint fecal sources. Applied and Environmental Microbiology 69, 4942-4950. Schuster, C.J., Ellis, A.G., Robertson, W.J., Charron, D.F., Aramini, J.J., Marshall, B.J., Medeiros, D.T., 2005. Infectious disease outbreaks related to drinking water in Canada, 1974-2001. Canadian Journal of Public Health 96 (4), 254-258.  151  Shanks, O.C., Nietch, C., Simonich, M., Younger, M., Reynolds, D., Field, K.G., 2006. Basinwide analysis of the dynamics of fecal contamination and fecal source identification in Tillamook Bay, Oregon. Applied and Environmental Microbiology 72 (8), 5537-5546. Shere, J.A., Bartlett, K.J., Kaspar, C.W., 1998. Longitudinal study of Escherichia coli O157:H7 dissemination on four dairy farms in Wisconsin. Applied and Environmental Microbiology 64, 1390-1399. Thomas, C., Hill, D.J., Mabey, M., 1999. Evaluation of the effect of temperature and nutrients on the survival of Campylobacter spp. in water microcosms. Journal of Applied Microbiology 86 (6), 1024–1032. Waage, A.S., Varund, T., Lund, V., Kapperud, G., 1999. Detection of small numbers of Campylobacter jejuni and Campylobacter coli cells in environmental water, sewage, and food samples by a semi-nested PCR assay. Applied and Environmental Microbiology 65, 1636-1643. Witte, W., 2000. Ecological impact of antibiotic use in animals on different complex microflora: environment. International Journal of Antimicrobial Agents 14, 321-325.  152  Appendix A Elk Creek Pictures  Figure A. Image of the Elk Creek agricultural watershed.  153  Figure B. Elk Creek forested headwaters, where the control site (site 1) was located.  Figure C. Warning sign outside old drinking water treatment station at site 1 in the forested headwaters of Elk Creek.  154  Figure D. Area between sites 1 (control site in forested area) and 2 (first impacted site studied). The area holds several livestock production operations and pasture land.  Figure E. First site located in the agriculturally impacted area (site 2) and in the middle of Elk Creek.  155  Figure F. Site 3 located in the agricultural reach of Elk Creek and after the merger of Elk Creek and Ford Creek.  Figure G. Site 4 located at the end of Elk Creek at the Big Ditch right before Elk Creek pours into Hope Slough.  156  Appendix B Biofilm Sampler Pictures  Figure A. Prepared biofilm samplers (five samplers appear in this image). These were the short and long-term biofilm samplers that did not contain sandpaper as a substratum.  157  Figure B. Image showing how substrata (wood is shown here) were attached to the galvanized steel bars with plastic ties (zap straps) while making sure that minimal contact occurred between the substratum and the surrounding surfaces.  Figure C. Image showing how the river rocks were attached. These substrata were the hardest to fasten and required more contact with the plastic ties and the galvanized steel bar.  158  Appendix C Protozoa numbers within biofilms colonizing different substrata in Elk Creek Date  Site  Acc Period  Jan-18-07  4  4 months  Jan-25-07  2  4 months  Feb-05-07  3  4 months  Feb-14-07  3  1 week  Feb-19-07  3  2 weeks  Mar-02-07  3  3 weeks  Mar-09-07  3  4 weeks  Mar-29-07  2  6 months  Apr-04-07  3  6 months  Substratum Water River Rock Slate Rock Wood Lexan Water River Rock Slate Rock Wood  Total number of organisms per mm3 Amoeba Flagellates Ciliates Total Protozoa 23 0 28 51 70 5 30 105 160 28 55 243 190 25 73 288 43 5 23 71 3 0 8 11 258 45 123 426 150 60 268 478 125 25 8 158  Lexan Water River Rock Slate Rock Wood  33 40 368 268 428  10 0 35 38 43  38 8 195 188 700  81 48 598 494 1171  Lexan Water River Rock Slate Rock Wood Lexan Water River Rock Slate Rock Wood Lexan Water River Rock Slate Rock Wood Lexan Water River Rock Slate Rock Wood  200 4 56 20 89 11 6 331 250 210 139 7 5130 240 310 190 8 44 258 85  10 3 14 15 8 4 3 26 34 13 9 4 30 30 23 9 9 18 15 35  28 10 30 61 26 14 19 50 210 68 58 3 12 52 12 43 5 19 141 44  238 17 100 96 123 29 28 407 494 291 206 14 5172 322 345 242 22 81 414 164  Lexan Water River Rock Slate Rock Wood Lexan Water River Rock Slate Rock Wood Lexan  21 35 105 203 395 60 8 383 315 448 113  11 10 10 68 60 10 3 45 23 48 15  45 13 60 360 203 63 25 185 130 140 85  77 58 175 631 658 133 36 613 468 636 213  159  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0066404/manifest

Comment

Related Items