Thesis Reference

Thesis

Microbial and geochemical characterization of a contaminated freshwater ecosystem (the case of Vidy Bay, Lake Geneva, Switzerland)

GLASS-HALLER, Laurence

Abstract

Le Léman représente la plus grande réserve d'eau douce d'Europe Occidentale et fournit de l'eau à environ 700 000 personnes. Il s'agit d'un lac tempéré, monomictique, qui était considéré comme eutrophe dans les années 1970 à 1980, mais qui depuis est redevenu mésotrophe grâce à des réductions drastiques des apports en phosphore. Dû à un accroissement de la population, au développement de l'industrie et une agriculture plus intensive durant ces dernières décennies, certaines parties du lac souffrent de pollution d'origine anthropique. La ville de Lausanne, située sur la rive nord, déverse dans le lac les plus gros volumes d'eaux usées et traitées, au niveau de la Baie de Vidy. Les trois principales sources de contamination de cette baie sont les effluents de la station d'épuration de Vidy (STEP) et les embouchures des rivières Chamberonne et Flon. L'objectif principal de cette étude de thèse était d'évaluer et de caractériser la contamination chimique et microbiologique de la baie de Vidy, et d'étudier en particulier les effets de la pollution sur la composition et la diversité des [...]

Reference

GLASS-HALLER, Laurence. Microbial and geochemical characterization of a contaminated freshwater ecosystem (the case of Vidy Bay, Lake Geneva, Switzerland). Thèse de doctorat : Univ. Genève, 2010, no. Sc. 4235

URN : urn:nbn:ch:unige-161678 DOI : 10.13097/archive-ouverte/unige:16167

Available at: http://archive-ouverte.unige.ch/unige:16167

Disclaimer: layout of this document may differ from the published version.

1 / 1 UNIVERSITE DE GENEVE FACULTE DES SCIENCES Section des sciences de la terre et de l’environnement Institut F.-A. Forel Professeur W. Wildi

Département de botanique et biologie végétale Professeur R. Peduzzi

Microbial and geochemical characterization of a

contaminated freshwater ecosystem (the case of

Vidy Bay, Lake Geneva, Switzerland)

THÈSE

présentée à la Faculté des sciences de l’Université de Genève pour obtenir le grade de Docteur ès sciences, mention interdisciplinaire

par

Laurence Glass-Haller

de Baden (AG)

Thèse N°4235

GENÈVE

2010

Table of contents

Abstract ...... v

Résumé ...... viii

Abbreviation list ...... xii

1. Introduction ...... 1 1.1 Chemical contaminants in water ...... 3 1.1.1 Macropollutants in water ...... 3 1.1.2 Micropollutants in water ...... 4 1.2 Pathogenic organisms in water ...... 5 1.3 The role of bacteria in freshwater ecosystems ...... 7 1.3.1 Cellular structure ...... 8 1.3.2 Metabolic diversity of freshwater bacteria ...... 8 1.3.3 Breakdown of organic matter in aerobic and anaerobic environments .. 9 1.3.4 Bacteria in biogeochemical cycles: the case of the carbon, nitrogen, sulphur, phosphorus and iron cycles...... 11 1.4 The role of bacteria in the fate of water contaminants ...... 18 1.4.1 The effect of microorganisms on pollutants ...... 18 1.4.2 The impact of water contaminants on microbial communities ...... 19 1.5 Thesis rationale and objectives ...... 20 1.5.1 Rationale ...... 20 1.5.2 Study site ...... 22 1.5.3 Objectives ...... 27 1.5.4 Structure ...... 28 1.5.5 Institutional framework ...... 28 References ...... 29

2. Effects of a Sewage Treatment Plant Outlet Pipe Extension on the Distribution of Contaminants in the Sediments of the Bay of Vidy, Lake Geneva, Switzerland 35 Abstract 36 2.1 Introduction ...... 37 2.2 Materials and methods ...... 38 2.2.1 Study site and sediments sampling ...... 38 2.2.2 Grain size, organic matter, phosphorous and nitrogen analysis ...... 40 2.2.3 Metal and hydrophobic compound analysis ...... 41 2.2.4 Data analysis ...... 42 2.3 Results ...... 42 2.3.1 Grain size, organic matter, phosphorous and nitrogen ...... 42 2.3.2 Metal concentrations ...... 43 2.3.3 Hydrophobic organic compound concentrations ...... 47 2.3.4 Lake currents and spatial distribution of organic matter and heavy metals in the surface sediments ...... 49 2.4 Discussion ...... 51 2.4.1 Time evolution of contaminants in the surface sediments ...... 51 2.4.2 Heavy metal and hydrophobic organic compound concentration levels: comparison with standard values and potential biological effects...... 56 2.5 Conclusion ...... 58 References ...... 60

3. Composition of bacterial and archaeal communities in freshwater sediments with different contamination levels (Lake Geneva, Switzerland) ...... 65 Abstract ...... 66 3.1 Introduction ...... 67 3.2 Materials and methods ...... 68 3.2.1 Study site description and sampling procedure ...... 68 3.2.2 Chemical analysis ...... 69 3.2.3 DNA extraction ...... 71 3.2.4 PCR amplification ...... 71 3.2.5 Clone library construction and DNA sequencing ...... 72 3.2.6 Phylogenetic analysis ...... 73 3.3 Results ...... 74 3.3.1 Chemical analysis ...... 74 3.3.2 Bacterial 16S rRNA gene clone libraries ...... 79 3.3.3 Archaeal 16S rRNA gene clone libraries ...... 82 3.4. Discussion ...... 83 3.5 Conclusion ...... 95 References 96

4. Origin and spatial-temporal distribution of faecal bacteria contamination in a Bay of Lake Geneva, Switzerland ...... 103 Abstract ...... 104 4.1 Introduction ...... 105 4.2 Materials and methods ...... 107 4.2.1 Study site ...... 107 4.2.2 Water sampling ...... 108 4.2.3 Physicochemical measurements ...... 109 4.2.4 Microbiological analyses ...... 109 4.3 Results and discussion ...... 110 4.3.1 Physicochemical characterization of the lake ...... 110 4.3.2 Bacterial input from the contamination sources ...... 110 4.3.3 Temporal variability of FIB concentrations in the bay ...... 113 4.3.4 Vertical distribution of FIB in the lake water column ...... 116 4.3.5 Horizontal distribution of FIB near the river mouth and in the bay ... 118 4.3.6 Salmonella detection ...... 119 4.4 Conclusions ...... 122 References ...... 124

5. Distribution and survival of faecal indicator bacteria in the sediments of the Bay of Vidy, Lake Geneva, Switzerland ...... 127 Abstract ...... 128 5.1 Introduction ...... 129 5.2 Materials and methods ...... 131 5.2.1 Study site ...... 131 5.2.2 Sediment sampling ...... 132 5.2.3 Sediment characterization ...... 133 5.2.4 Faecal indicator bacteria analysis ...... 133 5.2.5 Persistence study ...... 134 5.3 Results and discussion ...... 134 5.3.1 Sediment characteristics ...... 134 5.3.2 Spatial distribution of FIB in surface sediments ...... 137 5.3.3 Vertical distribution of FIB ...... 141 5.3.4 Survival of FIB in contaminated sediments ...... 144 5.4 Conclusion ...... 145 References ...... 146

iii

6. Influence of freshwater sediment characteristics on persistence of faecal indicator bacteria ...... 151 Abstract ...... 152 6.1 Introduction ...... 153 6.2 Materials and methods ...... 155 6.2.1 Study sites and sampling ...... 155 6.2.2 Microcosms ...... 157 6.2.3 Survival study ...... 158 6.2.4 Sediment and water characterization ...... 158 6.2.5 Bacteria quantification ...... 159 6.2.6 Statistical analysis ...... 160 6.3 Results ...... 160 6.3.1 Water characteristics ...... 160 6.3.2 Sediment characteristics ...... 161 6.3.3 Survival study ...... 163 6.4 Discussion ...... 168 6.5 Conclusion ...... 172 References ...... 173

7. Conclusions and perspectives ...... 177 7.1 Conclusions ...... 178 7.1.1 Distribution of the chemical contamination -due to heavy metals and some organic pollutants- in the sediments of the Bay of Vidy...... 178 7.1.2 Identification and comparison of bacterial communities living in polluted sediments of the Bay of Vidy with non-polluted sediments, using molecular approaches...... 179 7.1.3 ..... Spatial and seasonal distribution of faecal bacteria (FIB) contamination in the water column of the Bay of Vidy, using standard cultivation techniques. .. 180 7.1.4 Spatial distribution of faecal bacteria (FIB) contamination in the sediments of the Bay of Vidy, using standard cultivation techniques...... 181 7.1.5 Importance of faecal bacteria survival and growth in sediments, according to different environmental parameters, such as nutrients and organic matter...... 182 7.2 Perspectives ...... 183

Abstract

Lake Geneva is Western Europe’s largest freshwater reservoir (89 km 3), supplying approximately 700 000 persons with water. It is a monomictic temperate lake which was considered eutrophic in the 70’s and 80’s, but is now mesotrophic after drastic reduction of phosphorus inputs. In the last 30 years, due to a growing population, more intensive agriculture and industrial development, certain areas became affected by contamination. The city of Lausanne, located on the northern shore, discharges the largest volume of treated wastewater into the nearby Bay of

Vidy. There are three main inflows of contaminated water into the bay of Vidy: the wastewater treatment plant (WWTP) effluent, the Chamberonne River and the Flon

River. The WWTP treats currently between 1 and 3 m 3s-1, and exceptionally up to 6 m3s-1, depending on meteorological conditions, corresponding to approximately

220'000 equivalent-inhabitants of wastewater. The Chamberonne River includes water from its natural drainage basin, but also some untreated wastewater from damaged urban collectors. The Flon collects surface and wastewater from the western part of the city, which is usually treated at the WWTP but released into the lake during floods. In 2001, the municipality of Lausanne extended the WWTP outlet pipe as a measure to reduce bacterial water pollution and sediment contamination close to the lake beaches. The sewage pipe now arrives at 700 m from the lake shore at 35 m depth.

The main objective of this interdisciplinary research was to assess the chemical and microbiological contamination of the Bay of Vidy and to look, in particular, at the pollution effects on the composition and the diversity of natural microbial communities living in the sediments.

Results revealed that the sediments sampled around the WWTP outlet were extremely rich in organic matter (between 20 to 30%), nutrients, heavy metals and organic pollutants. The surface area of highly contaminated sediments with heavy metals was reduced by one third after the pipe extension, decreasing from 1.3 to 0.8 km 2. Since 1996, an improvement in the quality of sediments for almost all metals, except for Hg, has been observed. However, concentrations near the outlet pipe were still very high, with for example levels of Hg reaching 8.7 mg kg -1. The highest sediment concentration values in heavy-metals were 3 to 13 times higher than the

PELs (probable effect levels), from the Canadian Sediment Quality Guidelines for the

Protection of Aquatic Life (CCME EPC-98E, 1999), indicating possible toxic effects of sediment contamination on the lake biota.

Indeed, bacterial diversity in Vidy sediments was significantly different from the communities in the uncontaminated sediments. Populations in the uncontaminated area were more diverse than in Vidy. This research suggests that, in addition to environmental variables, pollution could be one of the factors affecting microbial community structure.

Vidy sediments were characterized by intense mineralisation of organic matter under sulphate-reducing and methanogenic conditions, confirmed by the presence of clones related to iron-, sulphate-reducing bacteria and methanogenic archaea, which were more abundant in the contaminated sediments. A large proportion of the

Betaproteobacteria clones, mostly found in Vidy sediments, were related to the reductively dechlorinating Dechloromonas sp . The presence of unique species or specific groups of bacteria in Vidy indicates that certain bacterial and archaeal communities may have adapted to these particular conditions.

The microbiological contamination of the Bay of Vidy was also assessed, using faecal indicator bacteria (FIB). Results of this survey show that FIB levels are highly variable near the water surface. The WWTP outlet pipe at 35 m depth was quantitatively the most important source of faecal bacteria, whereas the FIB input flux rates from the Chamberonne River were typically 1 to 3 orders of magnitude

vi lower. The Chamberonne contributes to the contamination of the surrounding surface waters and beaches. Escherichia coli input flux rates from the WWTP could reach up to 2.5 x 10 10 CFU/s. The highest FIB concentrations in the near-surface water of the bay occurred during floods and during mixed lake conditions in winter. When the lake was stratified, the effluent water was generally trapped below the thermocline. The most favourable situation for bathers happens therefore, when the lake is stratified and the flow rates of the river and WWTP are low, i.e. during the bathing season in summer.

Sediments of the bay of Vidy however, constitute a reservoir of faecal indicator bacteria, which can persist in certain areas of the bay. Escherichia coli and

Enterococcus were particularly abundant near the WWTP outlet discharge and the mouth of the Chamberonne River, where concentrations between 10 5 and 10 7 CFU 100 g-1 were detected. These faecal bacteria could survive up to 50 days in surface sediments. Environmental factors such as organic matter and nutrient content, but also water temperature, influence their survival. Our results revealed an increased growth of FIB in sediments containing higher levels of organic matter and nutrients, and smaller grain size. Extended survival of enteric bacteria in sediments and potential remobilization of pathogens may be responsible for water quality failures and are of considerable significance for the management of risk at specific recreational coastal sites, such as the Bay of Vidy.

Contamination of drinking water resources and recreational waters due to the release of sewage treatment plants and diffuse urban runoffs remains problematic in many parts of the globe. Results from our survey show that chemical and microbiological pollution can exceed, in certain areas of Lake Geneva, the swiss and international recommendations. This pollution may impact the ecosystem functions as well as human health. Measures to carefully monitor wastewater plumes and to reduce pollution at the source are of considerable importance for the sustainable management of freshwater resources in general.

vii

Résumé

Le lac Léman représente la plus grande réserve d’eau douce d’Europe

Occidentale (89 km 3) et fournit de l’eau à environ 700 000 personnes. Il s’agit d’un lac tempéré, monomictique, qui était considéré comme eutrophe dans les années 1970 à

80, mais qui depuis est redevenu mésotrophe grâce à des réductions drastiques des apports en phosphore. Dû à un accroissement de la population, au développement de l’industrie et une agriculture plus intensive durant ces dernières décennies, certaines parties du lac souffrent de pollution d’origine anthropique.

La ville de Lausanne, située sur la rive nord, déverse depuis 1964 les plus gros volumes d’eaux usées et traitées dans le lac, au niveau de la baie de Vidy. Les trois principales sources de contamination de cette baie sont les effluents de la station d’épuration de Vidy (STEP) et les embouchures des rivières Chamberonne et Flon. La

STEP traite, en fonction des conditions météorologiques, 1 à 3 m 3s-1 d’eau usée et exceptionnellement jusqu’à 6 m 3s-1, ce qui correspond à environ 220'000 équivalents- habitants. La Chamberonne draine les eaux de son bassin versant, ainsi que les eaux non traitées de quelques collecteurs non raccordés à la STEP. Le Flon collecte les eaux de surface ainsi que les eaux usées de la partie Ouest de la ville. Ces eaux sont en général traitées à la station d’épuration mais sont rejetées directement dans le lac, en cas de crue. En 2001, la conduite sous-lacustre de l'exutoire de la STEP a été déplacée et prolongée. Elle arrive aujourd’hui à 700 m du bord et à 35 m de profondeur. Le but de cette mesure était de diminuer la contamination microbiologique et chimique au niveau des plages.

L’objectif principal de cette étude interdisciplinaire a été d’évaluer et de caractériser la contamination chimique et microbiologique de la baie de Vidy, et

viii d’étudier en particulier les effets de la pollution sur la composition et la diversité des populations microbiennes dans les sédiments.

Les résultats de cette recherche ont révélé que les sédiments autour de l’exutoire de la STEP sont extrêmement riches en matière organique (entre 20 et 30%), nutriments, métaux-lourds et polluants organiques. La surface fortement contaminée en métaux-lourds a diminué d’un tiers depuis la prolongation de l’exutoire en 2001, passant de 1.3 à 0.8 km 2. Depuis 1996, une amélioration au niveau de la qualité des sédiments a également été observée pour tous les métaux, à l’exception du mercure.

Les teneurs en métaux sont cependant toujours très élevés à proximité de l’exutoire de la STEP, avec par exemple une valeur de mercure atteignant 8.7 mg kg -1. Ces valeurs étaient 3 à 13 fois plus élevées que les PELs («probable effect levels»), mentionnés dans les directives canadiennes concernant la qualité des sédiments et la protection de la vie aquatique (CCME EPC-98E, 1999). Cela nous indique que ces sédiments contaminés ont certainement des effets toxiques sur la faune et la flore du lac.

La diversité microbienne des sédiments contaminés de Vidy est significativement moins importante que celle des sédiments non contaminés du lac.

Cette étude nous suggère donc, que la pollution d’origine anthropique, en plus des paramètres environnementaux, peut affecter la composition et la diversité des populations bactériennes dans les sédiments.

Les sédiments de Vidy ont été caractérisés par une minéralisation intense de la matière organique, grâce aux processus de sulfato-réduction et de méthanogenèse.

Ceci a été confirmé par la présence plus abondante de bactéries sulfato-réductrices et méthanogènes, dans ces sédiments. Une grande proportion des Betaproteobacteria , principalement trouvées à Vidy, sont affiliées au genre Dechloromonas sp . La présence unique de certaines espèces ou familles de bactéries indique que certains groupes de microorganismes ont réussi à s’adapter à un environnement pollué et aux conditions de stress de ce milieu particulier.

La contamination microbiologique de la baie de Vidy a également été évaluée, en utilisant des indicateurs de contaminations fécales (FIB). Les résultats de notre

étude ont montré que les concentrations en FIB dans la colonne d’eau sont extrêmement variables. L’exutoire de la STEP à 35 m de profondeur est définitivement la source la plus importante de bactéries fécales dans la baie de Vidy.

Les flux mesurés d’ Escherichia coli sortant de l’exutoire ont atteint des valeurs allant jusqu’à 2.5 x 10 10 CFU/s. Les apports de la Chamberonne étaient eux plus bas, d’environ 1 à 3 ordres de grandeur, mais jouent un rôle majeur dans la contamination des plages et des eaux de surface. Les concentrations en FIB les plus

élevées, proche de la surface de l’eau, ont été mesurées durant des conditions de crue et de brassage du lac. Lorsque le lac était stratifié, les effluents de la STEP restaient en général piégés sous la thermocline. La situation la plus favorable pour les baigneurs se trouve donc être pendant les jours où le lac est stratifié et les débits des rivières et de la STEP bas, c'est-à-dire pendant la période de baignade en été.

Les sédiments de la baie de Vidy constituent cependant, un réservoir de bactéries fécales, qui peuvent persister dans certaines parties de la baie. Les indicateurs Escherichia coli et Enterococcus étaient particulièrement abondants près de l’exutoire et de l’embouchure de la Chamberonne, où des valeurs allant jusqu’à 10 5 and 10 7 CFU 100 g -1 ont été détectées. Ces populations de bactéries fécales pouvaient survivre jusqu’à 50 jours dans les sédiments de surface. Certains facteurs environnementaux, comme les teneurs en matière organique et nutriments, mais aussi la température de l’eau, influencent la survie de ces microorganismes. Notre

étude a révélé une hausse de la croissance en FIB dans les sédiments fins et plus riches en matière organique et nutriments. L’augmentation de la durée de survie des bactéries entériques dans les sédiments et la remobilisation en particulier des pathogènes dans la colonne d’eau peut avoir des conséquences néfastes sur la qualité de l’eau et doit absolument être pris en compte lors la gestion des risques dans des sites de baignades, comme la baie de Vidy.

La contamination des ressources en eau douce, due aux rejets des stations d’épuration et à la pollution diffuse d’origine anthropique, reste problématique dans de nombreuses régions du globe. Les résultats de notre étude démontrent que la pollution chimique et microbiologique dépasse les normes recommandées par les directives suisses et internationales, dans certaines zones du lac Léman, ce qui

évidemment peut induire des problèmes au niveau du fonctionnement de l’écosystème et au niveau de la santé humaine. Les mesures pour contrôler les rejets de station d’épuration et réduire au maximum la pollution à la source, sont d’une importance primordiale pour pouvoir gérer de manière durable nos ressources en eau douce.

Abbreviation list

AAS: atomic absorption spectrometry

AMA: advanced mercury analyse

AVS: acid volatile sulphur

CFU: colony forming units

CRS: chromium reducible sulphur

DNA: deoxyribose nucleic acid

E. coli: Escherichia coli

ENT: enterococci

FC: faecal coliforms

FIB: faecal indicator bacteria

GC-ECD: gas chromatography/electron capture detector

GC-MS: gas chromatograph mass spectroscopy

HMs: heavy metals

HOCs: hydrophobic organic compounds

HPC: heterotrophic plate count

ICP-MS: quadrupole-based Inductively Coupled Plasma Mass Spectrometry

MFA: multiple factor analysis

OCPs: organochlorine pesticides

OTU: operational taxonomic unit

PAHs: polycyclic aromatic hydrocarbons

PCA: principle component analysis

PCBs: polychlorinated biphenyls

PCR: polymerase chain reaction

xii

PEC: probable effect concentrations

PECQ: PEC quotients

PELs: probable effect levels

RNA: ribonucleic acid

STEP: station d’épuration

TC: total coliforms

TN: total nitrogen

TOC: total organic carbon

TP: total phosphorus

TRS: total reducible sulphur

WWTP: wastewater treatment plant

xiii

xiv

CHAPTER 1

Introduction

Water covers seven tenths of the Earth’s surface. Only 3% of which is fresh water, and of that, less than 1% is available for human use in the form of surface waters such as lakes, rivers, streams and groundwater (Hahn, 2006; U.S.Geological

Survey website, 2010). Although there is ample freshwater on Earth to meet present and future demand, it is often not where it is needed. Many arid and semi-arid parts of the world are without reliable sources of freshwater. In addition to this problem, water is polluted when it is used in industry, agriculture and for domestic purposes, and thus the amount of water of acceptable quality available for human use is reduced still further. Since the Industrial Revolution, the efforts of removing pollutants from the natural environment have not been able to keep pace with the increasing amount of waste materials and a growing population that further aggravates the situation. This has often resulted in the transformation of lakes, rivers and coastal waters into sewage depots where the natural biologic balance is severely upset (Förstner & Wittmann, 1979). Moreover, expected increase in average temperature and changing precipitation patterns linked to climate change may alter the physical, chemical and biological dynamics of the freshwater cycle and could thus affect future water quality in ways that are still unknown.

There are various sources of pollutants in the aquatic ecosystem. Untreated or only partly treated wastewaters including industrial, agricultural and domestic effluents constitute the major contamination sources of freshwater ecosystems.

Sources of surface water pollution are generally grouped into two categories based on their origin: point and non-point sources. Point source pollution refers to contaminants that enter a waterway through a discrete conveyance, such as a pipe or a ditch. Non-point sources are sources that cannot be traced to a single site of discharge. Examples of point sources are: factories, sewage treatment plants, underground mines, oil wells. Examples of non-point sources are: acid deposition from the air, traffic, runoffs, pollutants that are spread through rivers. Nutrient runoff in storm-water over an agricultural field is a typical example of non-point

2 source pollution. This pollution is hard to control because the perpetrators cannot be traced.

The specific contaminants leading to pollution in water include a wide spectrum of chemicals, pathogens, and physical or sensory changes such as elevated temperature and discoloration. Chemicals include macro- and micropollutants but also nano-particles, oils, fuels and plastics, which are harmful to humans and all plants and animals leaving in the aquatic environment. Water-soluble radioactive compounds can be as well considered as dangerous pollutants causing cancer, birth defects and genetic damage. In the following paragraphs, we will succinctly describe some of these contaminants.

1.1 Chemical contaminants in water The first category of contaminants, chemicals, can be subdivided in 2 groups: macropollutants and micropollutants.

1.1.1 Macropollutants in water Macropollutants are chemical compounds, naturally present in the environment, which can become toxic at high concentrations (occurring at μg/liter or mg/liter). These are for example acids, salts, nutrients and natural organic matter.

The source, behaviour and treatment of macropollutants are well understood

(Schwarzenbach et al., 2006). The enrichment of waters with nutrients such as water- soluble nitrates and phosphates is called eutrophication. It can lead to enhanced plant growth (algal blooms) and depleted oxygen levels as this plant material decays.

The problem is often linked to organic matter in municipal waste water and runoff from fields fertilized with chemicals and manure. Organic matter itself is also more discharged into watercourses than any other pollutant. Large amounts of organic matter cause severe oxygen depletion in water, which is then unable to support both

3 the decomposition of organic compounds and many forms of aquatic life. Depending on the degree of eutrophication, severe reductions in water quality may occur, affecting fish and other animal populations. In such cases, the challenges are to optimize treatment technologies and to develop integrated policies at the scale of river basins (Jackson et al, 2001). In developed countries, installing wastewater treatment plants has helped reduce organic matter and nutrient pollution considerably.

1.1.2 Micropollutants in water Micropollutants include inorganic (e.g. heavy metals) and hydrophobic organic compounds.

Micropollutants are different from macropolluants due to:

- Their toxicity at very low concentrations (pg/liter or ng/liter).

- Their very weak biodegradation. Some chemicals are not degraded at all (e.g. heavy metals) or only very slowly (e.g. persistent organic pollutant such as DDT or polychlorinated biphenyls).

- Their negative effect on fundamental life mechanisms (inhibition of biological processes).

- Their potential bioaccumulation in the food chain.

It is far more difficult to assess the effect on the aquatic environment of the thousands of synthetic or natural micropollutants that may be present in the water at very low concentrations (Schwarzenbach et al., 2006). They are released into the environment through urban and industrial wastewater, agricultural runoffs and atmospheric fall-out.

Some heavy metals such as iron, copper, manganese, cadmium, and zinc are required by living organisms. However, excessive levels can be damaging to the organism. Other heavy metals such as mercury and lead are toxic metals that have no known vital or beneficial effect on organisms (Förstner & Wittman, 1979), and their

4 accumulation over time in the bodies of animals can cause serious illness. “Safe” levels of metal concentration are hard to establish and vary according to the type and degree of exposure and the state and toxicity of the metal in question. Toxicity is particularly difficult to measure because effects may not show up for years, or levels may become toxic for humans only through bioaccumulation in aquatic organisms.

For instance, trace levels of lead or mercury may cause damage to the central nervous system through long-term exposure (Förstner & Wittmann, 1979).

Organic compounds are for example polycyclic aromatic hydrocarbons

(PAHs), polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs).

Although the use of DDT was banned in most developed countries in the 1960s, other pesticides, mainly herbicides, are heavily used. More synthetic organic compounds are introduced every year, often without a full understanding of the risk they pose to the environment in general and human health in particular. They can have two effects on human health: immediate short-term toxicity and reactions from long-term exposure both resulting in chronic symptoms and sometimes death.

The increasing contamination of water and sediments by inorganic and organic micropollutants is a big concern in numerous aquatic ecosystems and cause serious environmental risks (Förstner and Wittmann, 1979; Pardos et al., 2004;

Schwarzenbach et al., 2006). Their removal from wastewater or accumulation in sediments should be examined extensively (Wildi et al., 2004).

1.2 Pathogenic organisms in water The second category of water pollutants includes disease-causing microorganisms. These are called pathogens and include bacteria, viruses, protozoa and parasitic worms. Most of the pathogens commonly found in water are enteric in origin. Enteric pathogens are excreted with faecal matter from humans and animals.

The origin of pathogenic microorganisms in surface water comprises municipal

5 wastewater treatment plant discharges, agricultural or storm runoff and other diffuse sources of human and animal wastes. The most important transmission pathway is the faecal-oral route, i.e. insufficiently or untreated wastewaters that contaminate drinking water. Diseases can also be transferred during recreational activities, such as bathing in polluted lake water or seawater (Noble et al. 2003).

Disease-causing bacteria that can be transmitted by water include Vibrio cholerae, Salmonella sp., Campylobacter sp., Shigella sp., and Staphylococcus aureus. There are as well, more than 100 known types of human and animal enteric viruses that may be transmissible through water, which include rotavirus, enterovirus and norovirus. Certain enteric protozoa are also water-borne pathogens, such as

Cryptosporidium parvum and Giardia lamblia. These organisms continue to be problematic even in well-regulated water supplies, because they are found in nearly all surface waters and form thick walled cysts surviving for long periods in the environment, which are highly resistant to chlorine (Madigan et al., 2003).

Hepatitis, cholera, dysentery, gastroenteritis, and typhoid are the more common water-borne diseases that affect large populations. Diarrhoeal diseases, mainly due to microbial water contamination, cause about 1 billion illnesses and 2.2 million deaths per year (Montgomery and Elimelech 2007). Although most of these diseases occur in developing countries, waterborne diseases are a worldwide problem.

Because water-borne pathogens are not easily and rapidly detected in the laboratory, water is instead tested for “indicators”. Indicator bacteria do not necessarily cause illness, but their presence reveals that the water has been contaminated by faecal material; and indicates therefore the potential presence of pathogens (An et al., 2002; Noble et al., 2003). Faecal indicator bacteria (FIB) including Escherichia coli (E. coli ), enteroccoccus sp . (ENT) and total coliform (TC), residing in the gastrointestinal tract of humans and animals, are commonly used to assess the microbiological safety of drinking and recreational waters. According to the WHO, the European and the Swiss guidelines on drinking water quality, no TC,

E. coli or ENT should be detected in 100mL of water (OHyg, 2005; Tallon et al., 2005).

The U.S. Environmental Protection Agency and the European Union recommend the use of E. coli and ENT, to assess the hygienic safety of recreational waters (USEPA

2000; EU 2006). The Swiss regulations are currently following the European Directive

2006/7/CE concerning the management of bathing water quality; and recreational waters are to be classified as poor, if concentrations of E. coli exceed 900 CFU·100 mL -

1 and concentrations of ENT exceed 330 CFU·100 mL -1, based upon a 90 -percentile evaluation.

However, most of the indigenous microorganisms such as bacteria, archaea, protozoa, algae and fungi, numerically dominating water habitats, are not pathogenic. On the contrary, they are key players in biogeochemical processes and also in processes controlling the water quality and the fate of pollution released into the environment. To improve the sustainable management of freshwater resources, detailed knowledge of microbial diversity, functions and ecology in freshwater ecosystems is urgently needed (Hahn, 2006). We will describe in the following paragraphs some of the major roles played by bacteria in the aquatic environment.

1.3 The role of bacteria in freshwater ecosystems Bacteria are a large group of unicellular, prokaryote, microorganisms.

Typically a few micrometres in length, bacteria have a wide range of shapes, ranging from spheres to rods and spirals. Bacteria are ubiquitous in every habitat on Earth, growing in soil, water, air, extreme environments such as acidic hot springs and deep in the Earth's crust, as well as in organic matter and the live bodies of plants and animals. There are typically 40 million bacterial cells in a gram of soil and a million bacterial cells in a millilitre of fresh water. They form much of the world's biomass.

1.3.1 Cellular structure The bacterial cell is surrounded by a lipid membrane, or cell membrane. As they are prokaryotes, bacteria do not tend to have membrane-bound organelles in their cytoplasm and thus contain few large intracellular structures. They consequently lack a nucleus, mitochondria, chloroplasts and the other organelles present in eukaryotic cells. In prokaryotic cells, DNA is present in a large double- stranded molecule called the bacterial chromosome that aggregates to form a visible mass called the nucleoid. DNA is circular in most prokaryotes and most prokaryotes have a single chromosome, containing only a single copy of each gene.

1.3.2 Metabolic diversity of freshwater bacteria Bacteria exhibit an extremely wide variety of metabolic types. Bacterial metabolism is classified into nutritional groups on the basis of three major criteria:

- the kind of energy used for growth,

- the source of carbon required for synthesis of bacterial biomass,

- the source of electron used for growth (electron donors) and the terminal

electron acceptors used for aerobic or anaerobic respiration.

The source of energy used by bacteria comes from either light through photosynthesis (phototrophy ) or from chemical substances ( chemotrophy ).

Chemotrophs use energy obtained from energy-yielding (exergonic) reactions to oxidise organic matter. The source of carbon, required for synthesis of bacterial biomass, is obtained either by reducing CO 2 during photosynthesis ( autotrophs ) or from complex organic compounds ( heterotrophs ). Typical autotrophic bacteria are phototrophic cyanobacteria, green sulfur-bacteria and some purple bacteria. The source of electrons for growth (electron donors), are obtained either from organic

(organotrophs ) or from inorganic ( lithotrophs ) compounds, such as sulphide, hydrogen and water. Respiratory organisms use chemical compounds as a source of energy by taking electrons from the reduced substrate and transferring them to a

8 terminal electron acceptor in a redox reaction. This reaction releases energy that can be used to synthesise ATP and drive metabolism. In aerobic organisms, oxygen is used as the electron acceptor. In anaerobic organisms other inorganic compounds, such as nitrate, sulfate or carbon dioxide are used as electron acceptors. This leads to the ecologically important processes of denitrification, sulfate reduction and acetogenesis, respectively.

The breakdown of biological material by bacteria ultimately involves an oxidation/reduction reaction, with the transfer of electrons from the organic substrate

(oxidation) to an electron acceptor (reduction). Bacteria act as catalysts in mediating this reaction, which can only take place in conditions that are thermodynamically suitable. Only a fraction of the energy that is released by this process becomes directly available to the bacterial cell and promotes bacterial productivity and population increase (Sigee, 2005).

1.3.3 Breakdown of organic matter in aerobic and anaerobic environments One of the main roles of bacteria in the freshwater environment is the breakdown of organic biomass and the recycling of various key elements (nitrogen, phosphorus, sulphur) which are present within the various organic compounds.

Inorganic carbon enters living biomass via carbon fixation by autotrophic organisms.

Decomposing biomass is converted back to inorganic carbon through hydrolysis / fermentation and subsequent mineralization (CO 2) by heterotrophic organisms

(Canfield, 1993; Kristensen and Holmer, 2001). Biomass decomposition, degradation, and decay are three similar terms describing the overall enzymatic process mediating

(1) the dissolution of particulate organic polymers (e.g. polysaccharides, proteins, lipids and nucleic acids) into large macromolecules (hydrolysis), (2) the cleavage of these into smaller moieties (fermentation), and (3) the terminal oxidation

(respiration) of the organic carbon by various electron acceptors such as oxygen,

9 manganese oxides, nitrate, iron oxides, and sulphate (O 2, Mn 4+ , NO 3-, Fe 3+ and SO 42-)

(Canfield et al., 2005).

In a lake, the concentration of dissolved oxygen ranges from supersaturation

(lake surface) to very low levels in the hypolimnion and sediments. This range of oxygen concentration correlates with a range of oxidising ability or oxidation/reduction potential.

Aerobic conditions

In aerobic environments, oxygen will become the main electron acceptor and bacteria that use this to oxidise their substrate will predominate. The surface waters of rivers and lakes are then populated by obligate aerobes which use complex organic compounds as a source of energy, carbon, and electrons.

Anaerobic conditions

Large quantities of detritus are degraded below the oxic zone in stratified water bodies, under anaerobic conditions which occur in the hypolimnion and sediments of eutrophic lakes and in the sediments of ponds, rivers and wastewater treatment plants.

In anaerobic environments, different microbial groups participate in the decomposition pathway and the end product of one metabolism is substrate for another until decomposition is complete (Holmer and Storkholm, 2001). Each process is mediated by a consortium of physiologically different microorganisms (Wellsbury at al., 1996). The different groups of bacteria that carry out that terminal oxidation include the denitrifying, iron-reducing, sulphate-reducing bacteria (which use nitrate, iron oxide, sulphate as the ultimate electron acceptor) and the methane- producing bacteria (which use CO 2 as the electron acceptor). However, many of these different respiration reactions are not mutually exclusive. Thus, Fe reducers can co-

10 exist with sulphate reducer and both of these can co-exist with methanogens

(Canfield, 1993).

1.3.4 Bacteria in biogeochemical cycles: the case of the carbon, nitrogen, sulphur, phosphorus and iron cycles.

1.3.4.1 The carbon cycle

Carbon is an essential part of life on Earth. It plays an important role in the structure, biochemistry, and nutrition of all living cells.

The major ways in which new organic carbon is synthesized on Earth are via photosynthesis and chemosynthesis. Inorganic carbon enters living biomass in aquatic systems, via carbon fixation by autotrophic organisms (such as algae, cyanobacteria). The most important autotrophs for the carbon cycle are trees in forests on land and phytoplankton in aquatic ecosystems.

Photosynthesis follows the reaction 6CO 2 + 6H 2O → C6H12 O6 + 6O 2.

Most carbon leaves the biosphere through respiration. When oxygen is present, aerobic respiration occurs, which releases carbon dioxide into the surrounding air or water, following the reaction C 6H12 O6 + 6O 2 → 6CO 2 + 6H 2O.

Photosynthetically fixed carbon is eventually degraded by heterotrophic microorganisms producing methane (CH 4) and carbon dioxide (CO 2). These 2 products are formed from the activity of methanogens or from various chemoorganotrophs via fermentation, anaerobic respiration, or aerobic respiration.

Methane produced in anoxic habitats is highly insoluble and thus is easily transported to oxic environments where it is oxidised to CO 2 by methanotrophs

(Madigan et al., 2003).

(CH 2O) n Organic compounds Photosynthesis Respiration Algae Plants Green plants Animals Cyanobacteria Microorganisms

Chemolithotrophic bacteria Oxic CO 2 CH4 Anoxic

Anoxygenic Anaerobic respiration and fermentation Methanogens Phototrophic bacteria Anaerobic organisms including phototrophic bacteria

Organic compounds

(CH 2O) n

Fig. 1.1 . The carbon cycle (adapted from Madigan et al., 2003).

1.3.4.2 The nitrogen cycle

Nitrogen is found within freshwater habitats in a wide range of forms (such as nitrate, nitrite, ammonia and dissolved organic nitrogen). Nitrate is the major biologically available form of nitrogen. It enters aquatic systems via rain and soil, passing from rivers to lake.

During the autotrophic growth of freshwater algae and photosynthetic bacteria, the requirement for nitrogen is met in 2 ways:

- assimilation of dissolved nitrate, nitrite or ammonia

- fixation of dissolved molecular nitrogen (N 2).

If nitrate is absorbed, it is first reduced to nitrite ions and then ammonium ions for incorporation into amino acids, intense nucleic acids, and chlorophyll.

Nitrogen is a key constituent of many important biomolecules, such as amino acids, chlorophylls, polymers and is essential to all living organisms.

Complex organic nitrogen in algal and bacterial biomass is subsequently broken down or passes along the food chain, ultimately ending up as organic detritus. Most organic nitrogen is recycled into inorganic form by a process known as ammonification in which nitrogen-containing biomolecules are degraded by microorganisms. The released ammonium can either be re-assimilated by bacteria or algae or it can be oxidised by other specialised bacteria.

The conversion of ammonia to nitrates is performed primarily by nitrifying bacteria. The primary stage of nitrification, the oxidation of NH 4+ is performed by bacteria such as the Nitrosomonas species, which converts ammonia to nitrites (NO 2-).

Other bacterial species, such as the Nitrobacter , are responsible for the oxidation of the nitrites into nitrates (NO 3-). This conversion happens under aerobic conversion.

Removal of dissolved oxygen by this process can lead to anaerobic hypolimnia and sediments in lakes.

Denitrification is the reduction of nitrates back into the largely inert nitrogen gas (N 2), completing the nitrogen cycle. This process is performed by facultative anaerobic bacterial species such as Pseudomonas and Clostridium in anaerobic conditions. They use the nitrate as an electron acceptor in the place of oxygen during respiration.

Nitrification N NO - 2 2

NH groups of 2 Nitrogen protein fixation Assimilation Assimilation Ammonification NO 3- Oxic NH 3 Assimilation Anoxic

NO 2- Ammonification

NH 2 groups of protein Nitrogen fixation N2

Denitrification

Fig. 1.2. The nitrogen cycle (adapted from Madigan et al., 2003).

1.3.4.3 The phosphorus cycle

In freshwater habitats, phosphorus (P) occurs mainly within the lake biota, as insoluble organic phosphorus. Phosphorus is an essential element in all living systems. The primary biological importance of phosphates (PO 43-) is as components of nucleotides, which when linked together, form the nucleic acids DNA and RNA.

Phosphates are also a critical component of ATP and are found in bones (calcium phosphate), and in phospholipids (found in all biological membranes).

The phosphorus cycle differs from the other major biogeochemical cycles in that it does not include a gas phase. The largest reservoir of phosphorus is in sedimentary rock. When it rains, phosphates are removed from the rocks and are distributed throughout both soils and water. Phosphorus is not highly soluble, binding tightly to molecules in soils; therefore it mostly reaches waters by travelling with runoff soil particles. Phosphates also enter waterways through fertilizer runoff, sewage, natural mineral deposits, and wastes from other industrial processes.

Assimilation of phosphorus by freshwater algae is restricted to uptake of phosphate ions, PO 43-. Although the requirement for P by freshwater organisms is considerably less than N, it is normally P which is the growth-limiting factor in freshwater systems. In a lake, the concentration of phosphates in the water directly controls the mass of phytoplankton and other organisms that develop during the major growth phase, and is a key determinant of the trophic status of the water body

(Sigee, 2005). P that is taken up by phytoplankton has three main fates (Sigee, 2005):

- passage to the bottom of the lake within the dead organisms (detritus) that continuously sediment within the water column,

- entering into the food chain, making a long-term contribution to the zooplankton and fish biomass,

- phosphorus recycling: some of it passes back into the water column, mainly due to release from lake biota.

Continuous sedimentation of phosphorus-rich detritus leads to the build-up of a layer of organic material, and continuous release / diffusion of phosphate from sediments back to the water column, due to bacterial degradation.

Particulate phosphorus Fish

Zooplankton Soluble organic phosphorus Phytoplankton Bacteria

PO 4

Organic sediments Inorganic sediments

Streams, rivers, Domestic aerial deposition effluent

Internal loading External loading Fig. 1.3. The phosphorus cycle in lakes (adapted from Sigee, 2005).

1.3.4.4 The sulphur cycle

In freshwater habitats, dissolved organic sulphur occurs primarily as sulphate ions (SO 42-). These ions are taken up by many organisms, including algae, fungi and most prokaryotes, and converted to sulphydryl (-SH) groups in the synthesis of proteins. Sulphate is then reduced to form compounds essential to the life of the cell.

Sulphur is an important nutrient for organisms, being a key constituent of certain amino acids, proteins, and other biochemicals.

The ability to use sulphate as an electron acceptor for energy-generating processes, however, is restricted to the sulphate-reducing bacteria (SRB).

In assimilative sulphate reduction, the H 2S formed is immediately converted into organic sulphur in the form of amino-acids, but in dissimilative sulphate reduction, the H 2S is excreted. The importance of sulphate reduction in the decomposition of organic matter in lakes depends on the amount of organic matter present and the availability of sulphate (Sinke et al., 1992). In these anaerobic conditions, sulphate is used as an electron acceptor by a wide range of SRB, such as Desulfovibrio and

Desulfobacterium .

Death and sedimentation of algae and other biota leads to cell breakdown in sediments, with further reduction of sulphydryl (-SH) groups to hydrogen sulphide

(H 2S), during the process of protein decomposition. This anaerobic process is carried out by a wide range of heterotrophic bacteria, including Pseudomonas liquefaciens

(Sigee, 2005).

H2S generated in the sediments by protein decomposition and reduction of sulphate, diffuses vertically through the hypolimnion and prior to entry or during entry into aerobic conditions, is oxidised to sulphur and then to sulphate. Anaerobic sulphur-oxidising bacteria occur at the top of the hypolimnion and include the

16 phototrophic green and purple sulphur bacteria. The other major group of sulphur- oxidising bacteria are mostly aerobic and oxidise sulphide to sulphate via elemental sulphur, which is deposited either inside ( Beggiatoa ) or outside the cell ( Thiobacillus ) as an intermediate.

Lake biota Organic-SH

Uptake and Degradation by protein heterotrophs synthesis

Reduction SO H S 4 by SRB 2

Oxidation Oxidation e.g. Thiobacillus e.g. Beggiatoa

0 S

Fig.1.4. The sulphur cycle (adapted from Sigee, 2005).

1.3.4.5 The iron cycle

On the Earth’s surface, iron exists naturally in two oxidation states, ferrous

(Fe 2+ ) and ferric (Fe 3+ ).

In temperate climates, the occurrence and availability of iron in natural freshwaters varies with the seasons. In winter, during the period of mixing in lakes, iron is present largely as the oxidised state, as soluble Fe 3+ in the water column and as insoluble precipitates of hydroxyl and phosphate anions on the surface of sediments

(Sigee 2005).

As the lake stratifies and the oxygen level in the hypolimnion declines, Fe 3+ are converted to Fe 2+ by biological and chemical reduction. A large number of organisms

17 can use ferric iron ,Fe 3+ , as an electron acceptor. As part of the oxidation of organic matter, ferric-iron reducing bacteria will use ferric iron as the terminal electron acceptor, once supplies of oxygen, manganese ions (Mn 4+ ), and nitrate (NO 3-) have become exhausted. The internal loading of Fe 2+ from anaerobic sediments can then become very high. Fe 2+ ions that have diffused into the hypolimnion are rapidly removed by sulphide ions to form insoluble ferric sulphide (Sigee, 2005). Fe 2+ ions are also chemically and biologically oxidised when they come into contact with dissolved oxygen. Fe 2+ can be oxidised by iron bacteria such as Gallionella and

Leptothrix for example (Madigan et al., 2003). The seasonal phase of re-oxygenation is accompanied by a dramatic fall in the concentration of soluble iron in the water column.

1.4 The role of bacteria in the fate of water contaminants Pollution of aquatic systems by organic and inorganic pollutants is a major environmental problem in many parts of the world. It is generally agreed that bacteria, as the most abundant sediment organisms, have a major role in the fate of these contaminants (Ford & Ryan, 1995). Two approaches exist when studying the interaction between microorganisms and water contaminants.

- The effect of microorganisms on the transformation and / or mobility of

pollutants,

- The impact of water contaminants on microbial communities and their

functions.

In this study, the impact of pollution and contamination on microbial diversity is of interest and will constitute one of our main objectives.

1.4.1 The effect of microorganisms on pollutants Microorganisms encounter various kinds of contaminants in the environment and it is, therefore, not surprising that they interact with them.

In the case of metals, bacteria possess a variety of properties that can effect changes in metal speciation, toxicity and mobility, as well as mineral formation or mineral dissolution (Gadd, 2004). Bacteria may for example volatize or precipitate metals and transform them into toxic organic derivatives. Mobilization of metals can be achieved by autotrophic and heterotrophic leaching, chelation by microbial metabolites and siderophores, and methylation which result in volatilization (Gadd,

2004). According to the type of physicochemical environment and of microbial metabolism, contaminants may be released from sediments into the water column. In such cases, sediments become a secondary source of pollution, leading to the possible contamination of benthic organisms living in contact with them and the whole benthic food chain (Gillian et al., 2005). On the other hand, immobilization of metals can occur by precipitation of insoluble organic or inorganic compounds or by sorption, uptake and intracellular sequestration (Gadd, 2004).

In the case of pollution with organic contaminants, microorganisms can sometimes be used to clean up the environment (bioremediation). For example a wide variety of bacteria, several molds and yeasts, and certain cyanobacteria and green algae have been shown to be able to oxidise hydrocarbons aerobically

(Madigan et al., 2003). Small-scale oil pollution of aquatic and terrestrial ecosystems from human as well as natural activities is common, and there exists a diverse microbial community that uses hydrocarbons as en electron donor. Microorganisms therefore, participate in oil spills cleanups by oxidising the oil to CO 2 (Madigan et al.,

2003).

1.4.2 The impact of water contaminants on microbial communities Water contaminants on the other hand, can have an impact on bacterial community composition and specific functions. Organic or inorganic pollutants possibly represent a significant source of toxicity for microbial communities and benthic organisms. Diversity of microbial communities may be a sensitive indicator

19 of pollution in aquatic ecosystems. A previous study showed that heavy metal contamination can lead to a reduction of bacterial diversity (Sandaa et al., 1999).

Bacterial key functions in aquatic ecosystems and enzyme activities might also be disturbed due to pollution from industrial and municipal sewage. Several investigations have observed a significant decrease or an inhibition of certain bacterial enzyme activities in polluted rivers, streams, lakes and soils with either heavy metals or organic contaminants (Wei & Morrison, 1992; Andreoni et al., 2004).

However, other studies determined either an increase in microbial diversity along with heavy metal contamination (Sorci et al., 1999) or no significant variation

(Gillan et al., 2005). Many different environmental factors have also an effect on bacterial community composition. Differences in microbial diversity cannot be explained only by pollution of aquatic systems. The different times of exposure to contaminants and the varying geochemical parameters might explain the range of results obtained in previous studies.

1.5 Thesis rationale and objectives

1.5.1 Rationale Urbanization and the consequent increase in population, intensification of agriculture and growth in industries result in increased freshwater pollution, particularly when coupled with inadequate sewage collection and treatment. The main threats to human health are from pathogens in sewage, nutrients, and toxic inorganic and organic chemicals used in industry and agriculture.

In developed countries, most faecal contamination has been contained by separating water cycles, hygienic measures taken at an individual and collective level, wastewater treatment plants and by disinfecting public water supplies.

Although drinking water quality monitoring and treatment reduce the risk, they do

20 not provide absolute safety: some pathogens, such as Cryptosporidium oocysts, may occur when faecal indicator bacteria are absent, and/or they may resist water treatment (Hoxie et al. 1997; Bonadonna et al. 2002). There is also a potential health risk for swimmers going into polluted recreational bathing waters. Many studies report on the microbial contamination of marine or freshwater beaches, due to partially or non-treated wastewater discharges, industrial inputs and non-point source surface runoffs (Noble et al., 2003; An et al., 2002; Alm et al., 2003; Ferguson et al., 2005; Mallin et al., 2007). FIB concentrations very often exceed the legal recommendations regarding bathing water quality, especially after storm and flood events. Faecal pollution originates from a variety of human and non-human sources, but FIB contamination from human faecal material is generally considered to be a greater risk to human health as it is more likely to contain human enteric pathogens

(Scott et al., 2003) Epidemiological studies indicate an increased risk of contracting gastrointestinal and respiratory illnesses, with the number of faecal bacteria in water

(Kay et al. 1994; Prüss, 1998; Haile et al., 1999). From 1999 to 2000, 59 disease outbreaks in the United States were attributed to water exposure, and 61% of these outbreaks were of gastroenteritis (Lee et al., 2002).

Organic and inorganic micropollutants released into the environment, through urban and industrial wastewater, are a concern as well, and represent an unsolved problem in numerous coastal ecosystems, lakes and rivers all over the world.

Contamination of freshwater resources may represent a source of toxicity to the indigenous bacterial communities, which are major players in biogeochemical processes. It might upset the natural biologic balance and the ecosystem as a whole.

Some previous studies have contributed to insights into the diversity and structure of bacterial populations in contaminated and uncontaminated sediments (Gillan et al.,

2005; Powell et al., 2003; Zhang et al., 2008). However, information on the composition of microbial communities in the sediment of freshwater ecosystem is still very fragmentary (Hahn 2006; Briée et al., 2007) and there is a paucity of

21 information on the impact of human pollution and eutrophication on bacterial composition in aquatic ecosystems.

Therefore, wastewater released into drinking water resources and recreational waters are always problematic and should be avoided or limited as much as possible.

Existing wastewater plumes require careful monitoring and research in order to further minimize the risk for the entire ecosystem and consequently human health.

Lake Geneva constitutes an important freshwater reservoir in Western Europe with great economical and societal values, which can be affected by contamination in some areas. Research on Lake Geneva has been carried out for more than 100 years, however, little is known about the spatial variability and fate of pathogens in the water column and the sediments of the lake, and about the microbial composition and activities in sediments with different contamination levels. This study aims to provide information and knowledge regarding these aspects.

1.5.2 Study site Most of this study has been conducted on Lake Geneva. Samples from Lake

Bret, another small lake in the vicinity of Lake Geneva, were needed in the last part of the survey. It will be described later on.

1.5.2.1 Lake Geneva Lake Geneva ( lac Léman ) is Western Europe’s largest freshwater reservoir with a surface area of 582 km 2, a volume of 81 km 3 and a maximum depth of 309 m (Wildi et al., 2004). It is a monomictic temperate lake, with early spring overturn not occurring every year. The lake was considered eutrophic in the 70’s and 80’s, but is now mesotrophic after drastic reduction of phosphorus inputs (Dorioz et al., 1998).

Approximately 700’000 people are supplied with water from Lake Geneva.

1.5.2.2 The Bay of Vidy The city of Lausanne, located on the northern shore, discharges the largest volume of treated wastewater into the nearby Bay of Vidy. A wastewater treatment plant (WWTP) was built in 1964 in Vidy for 220'000 eq.-inhabitants. Initially consisting of a two-stage treatment plant (mechanical and biological treatments), it has been equipped since 1971 with a chemical stage. In this stage, ferric chloride

(FeCl 3) is added and precipitates as Fe(OH) 3, removing phosphates. From 1964 to

2001, the WWTP effluents were discharged into the lake at about 300 m from the lakeshore and 15 m water depth. In 2001, the municipal authorities from Lausanne decided to extend the WWTP outlet pipe and the effluent is now discharged at 700 m from the shore and 35 m depth. The rationale for the extension of the outlet pipe was

1) to significantly reduce the concentration of faecal bacteria in surface waters and 2) to reduce the accumulation of pollutants in sediments close to the shore. The WWTP treats currently between 1 and 3 m 3s-1 and exceptionally up to 6 m 3s-1 of raw water, depending on meteorological conditions, which corresponds to approximately

220'000 equivalent-inhabitants of wastewater. During storms, only part of the wastewaters is treated. As rain water and wastewaters are not fully separated, a large volume of water can arrive at the treatment plant within a few hours. Storm water runoffs are discharged via the Flon River in the eastern part of the bay. On the western side of the bay, the Chamberonne River enters the lake. The Chamberonne

River includes water from its natural drainage basin, but also some untreated wastewater from damaged urban collectors. As a consequence, the Bay of Vidy is the most contaminated area of Lake Geneva.

The city of Lausanne and neighbourhoods receives 58% of its freshwater from the lake and has 2 major drinking-water pumping stations:

- one in Saint Sulpice which is located at 3.8 km west of the WWTP outlet pipe -

pumping 60’000 l/min-

- one in Lutry on the eastern part of Lausanne –pumping 40’000 l/min.

The figures 1.5 and 1.6 show the bay of Vidy and a simplified map of the site, respectively.

Figure 1.5 . The bay of Vidy, Lake Geneva, Switzerland. Topographic map of

Switzerland and information from high resolution sonar survey (unpublished).

Figure 1.6 . Simplified map of the Bay of Vidy.

1.5.2.3 Previous studies on the Bay of Vidy Reports between 2005 and 2008 from the International Commission for the protection of Lake Geneva (CIPEL), mention low concentrations of heavy metals and organic pollutants in the water column and a good water quality in general, in the whole lake. However, regarding the bay of Vidy, which is a popular recreational area near Lausanne, several publications document the accumulation of contaminants in the sediments, and the related ecological impacts and health risks. (Loizeau et al.,

2004; Pardos et al., 2004; Wildi et al., 2004). In Pardos’s study, the spatial distribution of contaminants in Vidy sediments, before the WWTP pipe extension, and their potential biological impacts, are well described. In this study, samples were divided in 3 clusters according to their content levels. Clusters A, B and C represented respectively: samples taken along the shoreline, samples taken from both sides of the outlet pipe and off shore throughout the bay, and finally samples taken near the outlet pipe. They measured in close proximity to the outlet pipe, concentrations of heavy metals and organic compounds above levels believed to evoke toxic biological responses, and concluded that the WWTP effluent is the main source of local contamination (Tables 1.1 and 1.2).

Table 1.1 . Heavy metals contents of surface sediments from Vidy Bay, Lake Geneva

(Pardos et al., 2004).

Cluster A (n = 13) Cluster B (n = 20) Cluster C (n = 5) Median Range Median Range Median Range Depth (m) 4 2-15 39 12-90 19 17-47 Cadmium (mg kg -1) 0.39 0.17–2.8 2 0.8-7.3 13.8 11.2-15.7 Chromium (mg kg -1) 21 12.56 83 38-124 191 172-256 Copper (mg kg -1) 20.3 6.2-153 133 86-404 525 441-765 Iron (mg kg -1) 10.1 7.8-27.8 39.3 24.4-52.6 66.6 59.7-77.7 Mercury (mg kg -1) 0.06 0.03-0.66 1.28 0.43-1.6 3.15 2.79-4.54 Lead (mg kg -1) 16.8 7.9-129 125 75-328 631 529-706 Zinc (mg kg -1) 84 33-476 343 203-1219 2097 1821-2268

Table 1.2. Organic compounds contents of surface sediments from Vidy Bay, Lake

Geneva (Pardos et al., 2004).

Cluster A (n = 3) Cluster B (n = 5) Cluster C (n = 3)

Median Range Median Range Median Range

ΣPAH (mg kg -1) 1.15 0.58-4.73 16.6 9.99-21.03 82.16 62.41-104.62 Fluorene 0.02 0.02-0.52 0.15 0.1-0.31 1.84 1.78-3.69 Phenanthrene 0.09 0.02-0.16 0.78 0.55-1.61 11.39 9.41-17.23 Pyrene 0.15 0.06-0.53 2.02 1.34-2.67 12.53 7.97-15.23

Organotin (μg kg -1) TBT < 7.00 <7-17 <7 <7-10 17 16.6-28.5 DBT <4 <4 <4 <4-8 19 16-27.9

Pesticides (μg kg -1) Heptachlor <0.03 <0.03-0.18 0.34 <0.03-1.29 14.6 1.72-33.12 Heptachlor epoxide <0.06 <0.06-0.39 1.02 <0.06-4.75 45.66 2.75-45.95 p,p' -DDE 0.14 <0.03-0.48 0.86 <0.03-3.38 23.74 1.36-25.69 p,p' -DDD <0.05 <0.05 0.31 <0.05-5.8 0.62 <0.05-4.31 p,p' -DDT <0.03 <0.03-0.15 2.4 0.3-4.19 37.29 3.08-41.18 PAH: polycyclic aromatic hydrocarbon; TBT: tributyltin; DBT: dibutyltin; p,p _-DDE: p,p _-dichlorodiphenyldichloroethylene (DDE); p,p _-DDD: p,p _- dichlorodiphenyldichloroethane; p,p _-DDT: p,p _-1,1,1 trichloro-2,2 bis (4-chlorophenyl) ethane.

1.5.3 Objectives Despite the numerous previous studies undertaken on Lake Geneva and the bay of Vidy in particular, no studies have been performed to look at the microbial composition of sediments in Lake Geneva. Vidy Bay, with these high concentrations in inorganic and organic pollutants around the outlet sewage pipe, represents a polluted environment and the structure of bacterial populations living on the interface sediment–water might be affected by this contamination. Certain functional bacterial groups might have adapted to these specific conditions.

The bacteriological water quality of the Vidy recreational site was also a big concern for the authorities. In 2005, a one year survey was commissioned by the

Lausanne municipality to look at the water and sediment chemical and microbiological quality of the bay, after the outlet pipe extension in 2001. The distribution of contaminants and pathogens and the potential risks for bathers during recreational activities were of major interests for Lausanne authorities. These goals were part of the thesis objectives, which were the following:

- Updated assessment of the chemical contamination (heavy metals, organic pollutants) distribution in the sediments of the Bay of Vidy.

- Phylogenetic diversity of microorganisms in the contaminated Bay of Vidy and non-contaminated sediments of Lake Geneva, using molecular approaches.

- Seasonal and spatial distribution of faecal indicator bacteria in the water column of the Bay of Vidy, using standard cultivation techniques.

- Distribution of faecal indicator bacteria in the sediments of the Bay of Vidy, using standard cultivation techniques.

- Importance of sediment characteristics such as organic matter and nutrients on the survival and growth of faecal indicator bacteria.

1.5.4 Structure This thesis is structured into 7 chapters.

Chapter 1 is the general introduction. In Chapter 2, we describe the effect of the

WWTP outlet pipe extension on the distribution of heavy metals and organic pollutants in the sediments of the bay of Vidy. A comparison of the spatial distribution and the total amounts of contaminants in the bay, before and after the pipe extension, is presented. In chapter 3, we compare the microbial composition and diversity in the contaminated sediments of Vidy with non-contaminated sediments of Lake Geneva, using molecular approaches. The bacteriological contamination in the water column and the sediments of the bay of Vidy is the focus of chapter 4 and

5, respectively. We used faecal indicator bacteria and standard cultivation techniques to assess the impact of the WWTP on the bay. In chapter 6, we look at the growth and survival of these faecal indicator bacteria in sediments, according to several environmental parameters such as organic matter and nutrient content. Finally in chapter 7, we briefly discuss all the results of this study and outline the perspectives and future research work planned to be undertaken in the area. Chapters 2, 4, 5 and 6 were previously published in international journals and chapter 3 is currently under review.

1.5.5 Institutional framework This interdisciplinary research was conducted under the regulations of the

University of Geneva, Faculty of Sciences.

This project was completed through the collaboration of the following institutions:

- Forel Institute, Faculty of Sciences, University of Geneva, route de Suisse 10, 1290

Versoix, Switzerland.

- Cantonal Institute of Microbiology, Via Mirasole 22A, 6500 Bellinzona, Switzerland.

- Laboratory of Microbiology, University of Neuchâtel, Rue Emile-Argand 11, 2009

Neuchâtel, Switzerland.

- Laboratory “Eauservice”, rue de Genève 36, P.O. Box 741, 1002 Lausanne,

Switzerland .

References

Alm EW, Burke J & Spain A (2003) Fecal indicator bacteria are abundant in wet sand and freshwater beaches. Water Research 37 : 3978-3982.

An Y, Kampbell DH & Breidenbach GP (2002) Escherichia coli and total coliforms in water and sediments at lake marinas . Environmental Pollution 120: 771-

778.

Andreoni V, Cavalca L, Rao MA, Nocerino G, Bernasconi S, Dell’Amico E,

Colombo M & Gianfreda L (2004) Bacterial communities and enzyme activities of

PAHs polluted soils. Chemosphere 57 : 401-412.

Bonadonna L, Briancesco R, Coccia AM, Semproni M & Stewardson D (2002)

Occurrence of potential bacterial pathogens in coastal areas of the Adriatic Sea.

Environmental Monitoring and Assessment 77(1) : 31-49.

Briée C, Moreira D & Lòpez-Garcia P (2007) Archaeal and bacterial community composition of sediment and plankton from a suboxic freshwater pond.

Research in Microbiology 158 : 213-227.

Canfield DE, Jørgensen BB, Fossing H, Glud R, Gundersen J, Ramsing NB,

Thamdrup B, Hansen JW, Nielsen LP & Hall POJ (1993) Pathways of organic carbon oxidation in three continental margin sediments. Marine Geology 113 : 27-40.

Canfield DE, Thamdrup B & Kristensen E (2005) Aquatic Geomicrobiology.

Elsevier Academic Press, London, 636 p.

CIPEL (Commission Internationale pour la Protection des Eaux du Léman)

2005, 2006, 2007, 2008. Rapports sur les études et recherches entreprises dans le bassin lémanique. Campagnes 2005, 2006, 2007 et 2008.

Dorioz JM, Pelletier JP & Benoit P (1998) Physico-chemical properties and bioavailability of particulate phosphorus of various origins in a watershed of Lake

Geneva (France) . Water Research 32 : 275-286.

EU, European Directive 2006/7/CE of the European Parliament and of the

Council of 15 February 2006 concerning the management of bathing water quality and repealing Directive 76/160/EEC.

Ferguson DM, Moore DF, Getrich MA & Zhowandai MH (2005) Enumeration and speciation of enterococci found in marine and intertidal sediments and coastal water in southern California. Journal of Applied Microbiology 99 : 598-608.

Ford T & Ryan D (1995) Toxic metals in aquatic ecosystems: a microbiological perspective. Environmental Health Perspect ive 103 : 25-28. (Vu dans Gillian)

Förstner U & Wittmann GTW (1979) Metal pollution in the aquatic environment. Springer Verlag , Berlin, 486 p.

Gadd GM (2004) Microbial influence on metal mobility and application for bioremediation. Geoderma 122 : 109-119.

Gillan CD, Danis B, Pernet P, Joly G & Dubois P (2005) Structure of sediment- associated microbial communities along a heavy-metal contamination gradient in the marine environment. Applied and Environmental Microbiology 71 : 679-690.

Hahn MW (2006) The microbial diversity of inland waters. Current Opinion in

Biotechnology 17 : 256-261.

Haile RW, White JS, Gold M, Cressey R, McGee C, Millikan RC, Glasser A,

Harawa N, Ervin C, Harmon P, Harper J, Dermand J, Alamillo J, Barrett K & Nides M

& Wang GY (1999) The health effects of swimming in ocean water contaminated by storm drain runoff. Epidemiology 10: 355–363.

Holmer M & Storkholm P (2001) Sulphate reduction and sulphur cycling in lake sediments: a review. Freshwater Biology 46 : 431-451.

Hoxie NJ, Davis JP, Vergeront JM, Nashold RD & Blair KA (1997)

Cryptosporidiosis-associated mortality following a massive waterborne outbreak in

Milwaukee, Wisconsin. American Journal of Public Health 87(12) : 2032-2035.

Jackson RB, Carpenter SR, Dahm CN, McKnight DM, Naiman RJ, Postel SL &

Running SW (2001) Water in a changing world. Ecological Applications 11 : 1027-1045

Kay D, Jones F, Wyer MD, Fleisher JM, Salmon RL, Godfree AF, Zelenauch-

Jacquotte A & Shore R (1994) Predicting likelihood of gastroenteritis from sea bathing: results from randomised exposure. The Lancet 344 : 905-909 .

Kristensen E & Holmer M (2001). Decomposition of plant materials in marine sediments exposed to different electron acceptors (O 2, NO 3-, SO 42-), with emphasis on substrate origin, degradation kinetics, and the role of bioturbation. Geochimica and

Cosmochimica Acta 65 : 419-433.

Lee SH, Levy DA, Craun GF, Beach MJ & Calderon RL (2002) Surveillance for waterborne disease outbreaks—United States, 1999-2000. In: CDC Surveillance

Summaries, November 22, 2002 . MVWR 2002: 51 (No. SS-08) p. 1-28.

Loizeau JL, Pardos M, Monna F, Peytremann C, Haller L & Dominik J (2004)

The impact of a sewage treatment plant’s effluent on sediment quality in a small bay in Lake Geneva (Switzerland-France). Part 2 : Temporal evolution of heavy metals.

Lakes & Reservoirs: Research and Management 9 (1 ): 53-63.

Madigan MT, Martinko JM, Parker J (2003) Brock Biology of Microorganisms.

Pearson Education, Inc. Upper Saddle River, NJ, 1019 p.

Mallin MA, Cahoon LB, Toothman BR, Parsons DC, McIver MR, Ortwine ML

& Harrigton RN (2007) Impacts of a raw sewage spill on water and sediment quality in an urbanized estuary. Marine Pollution Bulletin 54 : 81-88.

Montgomery MA & Elimelech M (2007) Water and sanitation in developing countries: Including health in the equation. Environmental Science & Technology 41(1) :

17-24.

Noble RT, Moore DF, Leecaster MK, McGee CD & Welsberg SB (2003)

Comparison of total coliform, fecal coliform, and enterococcus bacterial indicator response for ocean recreational water quality testing. Water Research 37 : 1637-1643.

Ohyg (2005). Ordonnance du département fédéral de l’intérieur (Suisse) sur l’hygiène, du 23 novembre 2005. Ordonnance 817.024.1

Pardos M, Benninghoff C, de Alencastro LP & Wildi W (2004) The impact of a sewage treatment plant's effluent on sediment quality in a small bay in Lake Geneva

(Switzerland-France). Part 1: Spatial distribution of contaminants and the potential for biological impacts. Lakes & Reservoirs: Research and Management 9: 41-52

Powell SM, Bowman JP, Snape I, Stark JS (2003) Microbial community variation in pristine and polluted nearshore Antartic sediments. FEMS Microbiology

Ecology 45 : 135-145.

Prüss A (1998) Review of epidemiological studies on health effects from exposure to recreational water. International Journal of Epidemiology 27 : 1-9.

Sandaa RA, Torsvik V, Enger O, Daae FL, Castberg T & Hahn D (1999)

Analysis of bacterial communities in heavy metal-contaminated soils at different levels of resolution. FEMS Microbiology Ecology 30 : 237-251.

Schwarzenbach RP, Escher BI, Fenner K, Hofstetter TB, Johnson CA, von

Gunten U & Wehrli B (2006) The challenge of micropollutants in aquatic systems.

Science 313 : 1072-1077.

Scott TM, Parveen S, Portier KM, Rose JB, Tamplin ML, Farrah SR, Koo A &

Lukasik J (2003) Geographical variation in ribotype profiles of Escherichia coli isolates from humans, swine, poultry, beef, and dairy cattle in Florida. Applied and

Environmental Microbiology 69 : 1089-1092.

Sigee DC (2005) Freshwater Microbiology. John Wiley & Sons Ltd , The Atrium,

West Sussex, England, 524p.

Sinke AJC, Cornelese AA, Cappenberg TE & Zehnder AJB (1992) Seasonal variation in sulphate reduction and methanogenesis in peaty sediments of eutrophic

Lake Loosdrecht, The Netherlands. Biogeochemistry 16 : 43-61.

Sorci JJ, Paulauskis JD & Ford TE (1999) 16 rRNA Restriction Fragment Length

Polymorphism Analysis of Bacterial Diversity as a Biomarker of Ecological Health in

Polluted Sediments from New Bedford Harbor, Massachusetts, USA . Marine Pollution

Bulletin 38 : 663-675.

Tallon P, Magajna B, Lofranco C & Leung K (2005) Microbial indicators of faecal contamination in water: a current perspective. Water, Air and Soil Pollution 166 :

139-166.

USEPA (2000) Improved enumeration methods for the recreational water quality indicators : enterococci and Escherichia coli EPA-821/R-97/004. U.S.

Environmental Protection Agency, Washington, D.C.

U.S.Geological Survey, http://ga.water.usgs.gov/edu/watercyclehi.html

Wei C & Morrison G (1992) Bacterial enzyme activity and metal speciation in urban river sediments. Hydrobiologia 235 /236 : 597-603.

Wellsbury P, Rodney AH & Parkes RJ (1996) Bacterial activity and production in near-surface estuarine and freshwater sediments. FEMS Microbiology Ecology 19 :

203-214.

Wildi W, Dominik J, Loizeau JL, Thomas RL, Favarger PY, Haller L, Perroud A

& Peytremann, C (2004) River, reservoir and lake sediment contamination by heavy metals downstream from urban areas of Switzerland. Lakes & Reservoirs: Research and

Management. 9: 75-87.

Zhang W, Ki JS & Qian PY (2008) Microbial diversity in polluted harbor sediments I : Bacterial community assessment based on four clone libraries of 16S rDNA. Estuarine, Coastal and Shelf Science 76 : 668-681.

CHAPTER 2

Effects of a Sewage Treatment Plant Outlet Pipe

Extension on the Distribution of Contaminants in

the Sediments of the Bay of Vidy, Lake Geneva,

Switzerland

A similar version of this chapter was published under the following reference:

Poté J., Haller L., Loizeau J- L., Garcia Bravo A., Sastre V., Wildi W. (2008) Effects of a sewage treatment plant outlet pipe extension on the distribution of contaminants in the sediments of the Bay of Vidy, Lake Geneva, Switzerland. Bioresource Technology

99 : 7122-7131.

Abstract

In 2001, the municipality of Lausanne extended the outlet pipe of the sewage treatment plant into the Bay of Vidy (Lake Geneva, Switzerland) as a measure to reduce bacterial water pollution and sediment contamination close to the lake beaches. The aim of the present study was to assess the impact of this measure. Lake bottom sediments were collected and analyzed for grain size, organic matter, organic carbon, nitrogen, phosphorus, heavy metals and hydrophobic organic compounds to evaluate their concentration and spatial distribution. Our results demonstrate that, the surface area of highly contaminated sediments was reduced by one third after the pipe extension, from 1.3 to 0.8 km 2. However, contaminant concentrations are still high near the outlet pipe (the most toxic elements presented the highest concentration values of 17, 8.7 and 18.4 mg kg -1 for As, Hg and Cd, respectively) and the accumulation of pollutants may represent a significant source of toxicity for benthic organisms. One concludes that contaminant reduction at the source will be necessary for further improvement.

2.1 Introduction Sediment contamination is usually due to inorganic and organic compounds including heavy metals (HMs) and hydrophobic organic compounds (HOCs), such as polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs). Anthropogenic activities as well as urban effluent water constitute the main sources of aquatic environment contamination. Sediment contamination with micropollutants might cause irreversible adverse effects to ecosystems and might increase the potential risks to human health.

The main environmental risk is the remobilization of contaminants and their return to the hydrosphere either by sediment re-suspension or by infiltration into the groundwater (Wildi et al., 2004). The modification of environmental conditions such as pH, redox potential, bacterial activities, or ligand concentration can lead for example to the release of heavy metals from the sediment to the water column and increase their bioavailability (Cantwell et al., 2002; Lors et al., 2004). Regarding organic contaminants, PAHs, PCBs and OCPs have been identified as environmental pollutants in all environmental compartments (Wu et al., 1999). Due to their high persistence and low solubility in water, organic contaminants can accumulate in sediments. Therefore, polluted sediments represent an important source of contamination for freshwater organisms (Kang et al., 2000; Verwij et al., 2004).

The city and agglomeration of Lausanne generates large volumes of wastewater. The sewage treatment plant (WWTP) of Vidy, built in 1964, treats today approximately 220'000 equivalent-inhabitants (Fig. 1). From 1964 to 2001, the WWTP effluent was discharged into the lake at about 300 m from the lakeshore and 15 m water depth. In 2001, the Lausanne municipal authorities decided to extend the outlet pipe discharge to a distance of 700 m from the shore, at 35 m water depth. This decision was based on the following assumptions: a) Bacterial contamination of the shore is due to the backflow of contaminated

water from the WWTP outlet to the shore.

37 b) The accumulation of contaminated sediments in the bay is mainly due to local

circular lake currents, retaining water masses and suspensions within the bay,

isolated from the zone concerned by the main lake currents.

The rationale for the prolongation of the WWTP outlet pipe discharge system was therefore

1) to reduce significantly the concentration of faecal bacteria in surface water and

2) to reduce the accumulation of pollutants in sediments close to the shore.

No measure to reduce contaminant concentration at the source, in the WWTP, was considered at that time.

The objective of the present study was to assess and update the quality of bottom sediments of the Bay of Vidy after the WWTP outlet pipe extension. This assessment was based on the determination of the spatial distribution of organic matter and heavy metals (Cu, Cd, As, Fe, Cr, Ni, Zn, Pb, Ag, Hg) in bottom sediment samples, from 25 sites of the bay. The hydrophobic organic compounds (PAHs, PCBs, OCPs) levels in the sediments sampled from 3 sites, close to the present WWTP outlet pipe discharge location, were determined in order to examine the direct effect on the nearby sediments at the WWTP outlet.

2.2 Materials and methods

2.2.1 Study site and sediments sampling There are three potential main inflows of contaminated water into the bay of

Vidy (Fig. 2.1): the WWTP effluent, the Chamberonne River and the Flon River (Flon storm water outlet). The WWTP treats approximately 1 to 3 m 3. s-1 of urban wastewater. The Chamberonne River includes water from its natural drainage basin, but also some untreated wastewater from damaged urban collectors. The Flon

38 collects surface and wastewater from the western part of the city, which is usually treated at the WWTP but released into the lake during floods, via a pipe at 10 m water depth. The pumping station of St. Sulpice, that provides 58% of drinking water for the town of Lausanne, is located at 3.8 km from the outlet of the STP. The drinking water is pumped from a depth of 45 m at an average rate of 385 L .s-1.

During sampling actions, from spring 2005 to spring 2006, the near-surface currents in the bay were measured using drifters that sampled a depth of 2 m and were thus not directly influenced by the wind. They were released at different sampled points and recollected after 5 to 8 h. GPS measurements of their initial and final positions were done in order to determine the displacement vectors. In the area of the WWTP outlet pipe (Swiss coordinates: X 535057, Y 151332), an ADCP flowmeter was installed on the lake bottom to measure continuously the deep-water currents. It measured the water current speeds and directions on a vertical profile between the bottom and surface of the lake with a vertical step of 2 m. Results were published by Goldscheider et al. (2007).

The boat “La Licorne” of the Institute F.-A. Forel was used for sediment sampling in March 2005. The surface sediments (layer of 3 cm thickness) were collected using a “Ponar-type” grab sampler. The layer of 3 cm thickness corresponded to approximately 2 to 5 years, according to sedimentation rates determined by Loizeau et al. (2003) in the most contaminated area. Samples from 25 evenly distributed sites in the whole bay and numbered Vs1-Vs25, were taken (Fig.

2.1). Sampling points are slightly more dense close to the present outlet location than in the rest of the bay. Sediment samples were kept in autoclaved glasses for HOCs analysis and in polyethylene bottles for other analysis. All samples were stored in an icebox at 4 ºC and transported to the laboratory for analysis within 24 h.

Figure 2.1. Location map of the study area, sample sites ( ▲ noticed Vs): * and Vs25 represent the points of outlet pipe of sewage treatment plant discharge in the bay before and after 2001 respectively. The sewage treatment plant is noted as STP.

2.2.2 Grain size, organic matter, phosphorous and nitrogen analysis The particle grain size was measured with a laser Coulter® LS-100 diffractometer (Beckman Coulter, Fullerton, CA, USA), following 5-min ultrasonic dispersal in de-ionized water according to the method described by Loizeau et al.

(1994). The sediment total organic matter content was estimated from loss on ignition at 550°C for 1 hour in a Salvis oven (AG Emmenbrücke, Luzern, Switzerland). The sediment total organic carbon (TOC) content was determined by titrimetry following acid oxidation. Total nitrogen (TN) was determined by the Kjeldahl’s method

(APHA, 1985). The total phosphorus (TP) was measured with a spectrophotometer

(Helios Gamma UV-Vis Thermo Electroporation , Thermo scientific, USA) at 850 nm.

The sediment sample preparations for TP concentration were performed as described below; 50 mg of dry sediments were diluted in 5 mL HCl 1N and introduced in

40 centrifuge tubes. The mixture was ultrasonicated (at ambient temperature) during 16 h and centrifuged (4000 rpm) during 20 min. The supernatant was mineralised during 45 min at 130°C after addition of K 2S2O8 solution (5%). The TP concentration was performed by measuring the absorbance of the blue complex obtained after reduction of molybdophosphoric acid according to the method described by Murphy and Riley (1962), Harwood et al. (1969) and Burrus et al. (1990).

2.2.3 Metal and hydrophobic compound analysis Before analysis, sediment samples were sieved through a 63 µm mesh size sieve and air-dried at ambient room temperature. The total metal concentrations of

Cu, Cd, As, Fe, Cr, Ni, Zn, Pb and Ag were determined by quadrupole-based

Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (HP 4500, Agilent) following the digestion of sediments in Teflon bombs heated to 150°C in analytical grade HNO 3 2 M (Pardos et al., 2004; Loizeau et al., 2004). The total Hg analysis was carried out using the atomic absorption spectrophotometer for mercury determination (Advanced Mercury Analyser; AMA 254, Altec s.r.l., Czech Rep.) following the procedure described by Hall and Pelchat (1997) and Ross-Barraclough et al. (2002). The method is based on sample combustion, gold amalgamation and atomic absorption spectrometry (AAS).

The hydrophobic organic compound analysis (PAHs, PCBs, OCPs) was performed by the Bachema Institute (Schlieren, Switzerland); PAHs were determined by gas chromatograph mass spectroscopy (GC-MS) after extraction by accelerated solvent extraction (ASE) according to the EPA method 8270; PCBs were analyzed by

Gas Chromatography/Electron Capture Detector (GC-ECD) after extraction by ASE and after clean-up according to the EPA method 8082; and OCPs were measured by

GC-ECD after extraction by ASE according to the EPA method 8081. All results are expressed on a sediment dry weight basis.

Triplicate measurements have been performed on selected sediment samples.

The results are expressed in mg kg -1 dry weight sediment for HMs and PAHs, and in

µg kg -1 dry weight sediment for PCBs and OCPs. Total variation coefficients of triplicate sample measurements are smaller than 15% for both ICP-MS and AMA

(Hg) and 2% for HOCs measurements.

2.2.4 Data analysis Statistical treatment of data (Pearson product moment correlation) has been realized using SigmaStat 3.11 (Systat Software, Inc., USA). Mapping of sediment contamination and volume determination were carried out using Surfer 8 (Golden software, USA). Interpolation was based on the kriging method, which better expresses trends in the data than other conventional interpolating methods.

2.3 Results

2.3.1 Grain size, organic matter, phosphorous and nitrogen The grain size and the concentration of TOC, TN and TP in sampled sediments are shown in Table 2.1. The grain size distribution (Fig. 2.2) reflects both the hydrodynamics of the bay, with coarser sediments located close to the shores (high energy environment due to wave action) and the source contributions: the

Chamberonne River delivers coarse sediments close to its mouth.

The sediments sampled around the outlet pipe of the WWTP (Vs12, Vs14,

Vs16, and Vs25) showed high organic matter contents of about 20 to 30%. These values are much higher than the values measured on other sample sites (5 to 8%).

The maximum concentration of TOC, TN and TP was obtained on site Vs25, the closest to the present effluent pipe outlet. Concentrations from 54 004 to 143 971,

42 from 2 181 to 14 395 and from 914 to 1 8871 mg kg -1 were measured for TOC, TN and

TP, respectively.

Figure 2.2 . Median grain size (μm) distribution in the superficial sediments of the

Bay of Vidy obtained using the program surfer 8 (Golden software, USA) with Swiss map program.

STP: sewage treatment plant.

2.3.2 Metal concentrations The concentration values of heavy metals are reported in Table 2.2. These values (in mg kg -1) ranged from 33 to 727 for Cu, 0.45 to 18.4 for Cd, 2.3 to 17.0 for As,

10 470 to 40 555 for Fe, 24 to 337 for Cr, 32 to 87 for Ni, 70 to 3609 for Zn, 16.4 to 620 for Pb, 0.12 to 4.7 for Ag and 0.032 to 8.7 for Hg. With the exception of Ag, the highest metal levels were found in sediments collected from site Vs25. At this location, the most toxic elements presented the highest concentration values of 17, 8.7 and 18.4 mg kg -1 for As, Hg and Cd, respectively.

Table 2.1. Sample sites (GPS location in Swiss coordinates), grain size (median), and concentration of Total Organic Carbon (TOC), Total Nitrogen (TN) and Total

Phosphorous (TP) in sediments of the Bay of Vidy (mg .kg -1 dry sediments). Sample GPS position in Swiss Grain size, median* sites coordinates (mg kg -1 dry sediments) (µm) X (m) Y (m) TOC TN TP Vs1 533209 151446 99074 3943 914 187.2 Vs2 533784 151751 78987 3061 1081 137 Vs3 534287 152092 65063 2310 1196 109.7 Vs4 534910 152081 64938 2886 1184 191 Vs5 535186 151817 59905 2093 1001 68.6 Vs6 535515 151631 58909 2076 1431 138.5 Vs7 536210 151278 70746 3090 1118 41.05 Vs8 536653 151212 88600 4432 1216 219.3 Vs9 534036 151338 58942 2763 1535 24.6 Vs10 534269 151595 60895 3314 1996 27.9 Vs11 535048 151476 56065 2627 1342 46.1 Vs12 534766 151164 62243 3681 2092 23.9 Vs13 535479 150978 56022 2641 1115 23.3 Vs14 534682 151410 87284 7468 7371 40.6 Vs15 534426 151512 65257 3959 4978 29.9 Vs16 534569 151709 54004 2975 3629 80.2 Vs17 534895 151655 56784 2544 1257 84.8 Vs18 534083 151787 67911 2429 1049 136.0 Vs19 534479 151931 70782 2540 968 125.3 Vs20 534958 151871 59538 2181 966 74.4 Vs21 535503 151380 67632 2946 1089 58.3 Vs22 535126 151122 51592 2202 1499 41.3 Vs23 533676 151493 84467 3334 1060 177.2 Vs24 534431 151248 67417 4532 3517 20.8 Vs25** 534676 151543 143972 14395 18871 38.4 * Grain size median gives an unbiased characteristic value of the whole size distribution; ** Closest to the outlet pipe of sewage treatment plant in the lake, at 35 m depth.

Table 2.2. Metal content (mg .kg -1 dry weight sediment)* of surface sediments from

the Bay of Vidy, analyzed by ICP-MS and by AMA for Hg

Concentrations (mg .kg -1 dry weight sediment) Sample sites Cu Cd As Fe Cr Ni Zn Pb Ag Hg DL** 0.63 0.3 0.2 55 0.08 0.17 0.51 0.08 0.3 0.03*** Vs1 43 1.00 4.22 10470 29 35 104 26.9 0.49 0.041 Vs2 40 0.85 4.52 13209 31 40 86 21.7 0.27 0.091 Vs3 33 0.48 4.27 11498 26 32 70 16.4 0.45 0.140 Vs4 66 0.69 4.26 15771 35 40 162 36.3 0.66 0.643 Vs5 42 0.53 3.18 12082 27 32 104 23.8 0.66 0.078 Vs6 48 0.57 4.88 12548 29 32 111 26.2 0.65 0.918 Vs7 61 0.55 2.98 15483 35 43 137 33.8 0.48 0.139 Vs8 49 0.53 3.32 12779 29 37 109 27.5 0.37 0.168 Vs9 108 1.31 5.19 23445 64 61 237 68.2 2.00 0.651 Vs10 106 1.05 3.98 22482 59 61 220 56.5 0.79 0.374 Vs11 196 1.95 4.85 20466 70 60 363 111.9 1.74 1.202 Vs12 143 1.27 6.87 27264 74 70 299 74.3 4.65 0.948 Vs13 144 2.47 6.86 29870 84 75 374 117.5 3.94 1.281 Vs14 266 1.91 7.83 25566 84 55 545 126.0 2.07 1.514 Vs15 181 1.90 7.55 30190 79 63 403 98.6 2.55 1.084 Vs16 130 1.40 5.04 24027 62 52 303 81.2 2.87 1.319 Vs17 71 0.75 3.49 19657 46 51 165 38.9 0.29 0.235 Vs18 37 0.69 4.27 11874 28 34 77 17.3 0.36 0.032 Vs19 37 0.49 3.51 11085 26 32 83 17.5 0.53 0.033 Vs20 38 0.45 2.30 10894 24 29 91 21.2 0.46 0.077 Vs21 63 0.51 3.68 15995 38 44 133 32.2 0.12 0.207 Vs22 138 1.71 5.80 24783 83 70 324 102.8 1.76 0.987 Vs23 39 0.59 4.08 10610 28 34 80 19.8 0.57 0.035 Vs24 186 2.02 8.26 28988 86 71 466 104.6 2.56 1.134 Vs25 727 18.40 17.01 40555 337 87 3609 619.7 0.77 8.650

* Total variation coefficients for triplicate measurements are smaller than 15 % for both ICP-MS and AMA (Hg) measurements. ** DL: detection limit (mg .kg -1) *** the detection limit of Hg is 0.03 g.kg -1 according to manufacturer, with 300 mg of sample.

Table 2.3 shows that almost all metals, together with total nitrogen and total phosphorus, are strongly mutually, and positively correlated, with significant

(p<0.05) correlation coefficient usually higher than 0.8. This means that metals probably originate form a common source, and are transported and deposited by common carriers (particles). An exception to this observation is the Ag behaviour, which shows weaker or not significant correlations with the other metals. It is of note that Ag is correlated with Fe, Ni, and Hg, which show weaker correlations with other metals and nutrients, but are the most negatively correlated to mean grain size.

The correlation of metal concentrations with the organic carbon content is low or absent, despite the fact that the major source of both contaminants is the WWTP effluent. Median grain size is generally negatively correlated with most metals.

Table 2.3. Pearson product moment correlation matrix of selected parameters analyzed in the sediments sampled in 2005, located in the 1.32 km 2 most contaminated zone. Parameters include metals, organic carbon, total nitrogen (TN), total phosphorus (TP) contents, and median grain size. Correlation coefficients have been calculated using the log value of the parameter contents in order to normalize their distribution (n=14, statistically significant coefficients (p< 0.05) are in bold).

median Fe Ni Cu Zn As Ag Cd Hg Pb TN TP Corg size Cr 0.92 0.90 0.97 0.99 0.92 0.43 0.97 0.93 0.98 0.86 0.86 0.63 -0.63 Fe 0.96 0.92 0.89 0.86 0.63 0.83 0.95 0.92 0.80 0.82 0.37 -0.77 Ni 0.88 0.84 0.80 0.61 0.80 0.91 0.88 0.65 0.66 0.27 -0.80 Cu 0.98 0.90 0.53 0.94 0.97 0.99 0.88 0.88 0.57 -0.66 Zn 0.91 0.41 0.98 0.93 0.99 0.90 0.90 0.67 -0.58 As 0.48 0.92 0.82 0.88 0.88 0.88 0.70 -0.56 Ag 0.33 0.64 0.50 0.19 0.41 -0.14 -0.69 Cd 0.87 0.96 0.85 0.85 0.71 -0.50 Pb 0.85 0.85 0.56 -0.64 TN 0.998 0.76 -0.50 TP 0.72 -0.48 Corg -0.11

2.3.3 Hydrophobic organic compound concentrations The hydrophobic organic compounds were analyzed in three locations in the vicinity of the WWTP outlet pipe discharge (sites Vs12, Vs14, and Vs16). The results are presented in Table 2.4. Although three data points are not sufficient to evaluate the area extent of the contamination, one recognizes that the maximum concentrations of dominant PAHs were found at the site Vs12, with values of 2.07,

1.73, 1.51, 1.51 mg ...kg -1 for fluranthene, pyrene, benzo(a)anthracene, and triphenylene, respectively.

Total PCBs concentrations ranged from 10.8 to 156.1 µg...kg -1. Twelve PCB congeners (PCB 31, 70, 74, 87, 95, 101, 110, 132, 138, 149, 180 and 209) were identified in the three sampled sites with concentrations ranging from 0.12 to 17.74 µg...kg -1. The congeners 132 and 138 are found to be dominant with concentration of 17.74 and

16.52 µg...kg -1, respectively. Various OCPs were identified with concentrations ranging from 0.15 to 45.12 µg...kg -1. High levels of heptachlor epoxide, p,p’-DDE, p,p’-DDT and

Mirex were detected in most cases. The concentration values ranged from 0.37 to

45.12, 0.41 to 58.95, 0.84 to 30.40, and 0.14 to 10.32 µg.kg -1 for heptachlor epoxide, p,p’-DDE, p,p’-DDT, and Mirex, respectively. The high concentration levels of these metabolites in the sediments indicate elevated concentration of total DDT in the Bay of Vidy. Important isomers of HCH were detected with the maximum concentration values of 10.98, 5.13 and 56.71 µg.kg -1 respectively for α-HCH, β-HCH and γ-HCH.

HCB was detected in all analyzed sediments, but with a lower concentration range from 0.16 to 3.02 µg...kg -1.

Table 2.4. Concentration of polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyl (PCBs) and organochlorine pesticides (OCPs) in the Bay of

Vidy sediments sampled at the sites Vs12, Vs14 and Vs16 close to the STP outlet pipe discharge

Component name Sampled sites Vs12 Vs14 Vs16 PAHs (mg .kg -1 dry weight sediment; quantification limit 0.05 mg .kg -1)* Naphtalene < 0.05 0.25 < 0.05 Phenanthrene 0.63 0.62 < 0.05 Anthracene 0.31 0.18 < 0.05 Fluoranthene 2.07 1.44 0.17 Pyrene 1.73 1.28 0.14 Benzo(a)Anthracene 1.51 0.61 < 0.05 Triphenylene 1.51 0.79 0.11 Benzo(b)Fluoranthene 0.94 0.57 < 0.05 Benzo(k)Fluoranthene 0.76 0.36 < 0.05 Benzo(a)Pyrene 1.23 0.55 < 0.05 Indene(1,2,3-cd)Pyrene 0.75 0.08 0.06

PCBs ( µµµg.kg -1 dry weight sediment; quantification limit 0.2 µµµg.kg -1) Total PCBs 85.85 156.08 10.08 PCB31 2.52 7.30 0.26 PCB70 1.74 10.16 1.00 PCB74 0.41 7.61 < 0.2 PCB87 4.15 1.63 < 0.2 PCB95 3.83 7.32 0.50 PCB101 5.61 9.99 0.79 PCB110 6.77 11.61 0.76 PCB132 11.39 17.74 1.72 PCB138 11.24 16.52 1.20 PCB149 6.51 9.35 0.83 PCB180 5.20 6.33 0.97

PCB209 13.45 7.05 0.27 OCPs ( µµµg.kg -1 dry weight sediment; quantification limit 0.05 µµµg.kg -1) ) 1,2,4,5 Tetrachlorobenzene 0.32 2.22 <0.05 1,2,3,4-tetrachlorobenzene 0.22 0.95 <0.05

Pentachlorobenzene 0.87 4.03 <0.05 Hexachlorobenzene 0.69 3.02 0.16 α-HCH ( α-hexachlorocyclohexane 5.05 10.98 <0.05 β-HCH ( β-hexachlorocyclohexane) 3.63 5.13 <0.05 γ-HCH ( γ-hexachlorocyclohexane) 56.71 2.08 0.54 Octachlorostyrene 1.86 0.48 <0.05 Heptachlor Epoxide** 4.81 45.12 0.37 Oxychlordane 4.46 39.91 <0.05 p,p'-DDE 1.74 58.95 0.41 p,p'-DDD 1.30 3.40 0.15 p,p'-DDT 2.33 30.40 0.84 Mirex** 3.51 10.32 0.14

* Variation coefficients of triplicate measurements are smaller than 2%; ** Quantification limit 0.01 µg.kg -1.

2.3.4 Lake currents and spatial distribution of organic matter and heavy metals in the surface sediments During the one-year observation period in Vidy Bay, surface lake currents have been shown to be mostly circular, either clockwise, or counter-clockwise.

However, predictions of the surface water flow in function of the wind remains very uncertain. Mean displacement rates ranged in general between 1 and 8 cm s -1, with velocities up to 30 cm s -1 during wind events. Deep currents show a rotation with respect to surface currents; this rotation is known as the "Ekman-spiral ". Velocities are clearly lower than in surface water (Goldscheider et al., 2007).

The spatial distribution of all concentration parameters is almost regularly concentric around the WWTP outlet discharge (Vs25). The strongly impacted surface, with high concentrations, covers approximately 0.8 km 2, as shown by the total organic carbon and Zn areal distribution (Fig. 2.3, Fig. 2.4). As mentioned previously, the distribution of Ag differs significantly from the distribution patterns of the other metals. No explanation can be given concerning this case for the moment.

Figure 2.3. Spatial distribution of total organic carbon (mg.kg -1 dry sediments), in the surface sediments around the actual zone of the WWTP outlet discharge, obtained with the program surfer 8 (Golden software, USA) and Swiss map program.

Figure 2.4. Spatial distribution of Zn (mg.kg -1 dry sediments) in the surface sediments around the actual zone of the WWTP outlet discharge, obtained using the program surfer 8 (Golden software, USA) and Swiss map program.

2.4 Discussion

2.4.1 Time evolution of contaminants in the surface sediments Before 2001, the outlet pipe of the WWTP was discharged at 15 m water depth and 300 m distance from the shore. Previous surveys (Peytremann and Haller, 1997;

Loizeau et al., 2004) have put in evidence the deposition of discharged suspensions from the WWTP effluent, as fine, muddy sediments around the endpoint of the outlet pipe, and throughout the whole bay. These sediments were characterized by high to very high concentrations in heavy metals, polyaromatic hydrocarbons (PAHs), pesticides, phosphorous and organic carbon, and were shown to be toxic for the benthic fauna (Pardos et al., 2004). A strong environmental impact and risk for sediment dwellers and their predators was put in evidence.

In 2001, the WWTP effluent water was extended. The rationale for the WWTP outlet pipe prolongation was to reduce significantly bacterial water pollution and sediment contamination, in this particular part of the lake.

The comparison of our results with former data (Pardos et al., 2004) shows that in addition to the location of the sedimentation centre, the pattern and amplitude of the accumulation of heavy metals has also changed since the outlet pipe extension. Pre-2001 sediments show an oblate distribution pattern in a South-West direction, whereas the present distribution is more circular, with a small contamination “tail” towards the South-East of the outlet, as shown by the Hg and

Cu distribution before and after the pipe extension (Fig. 2.5a, 2.5b, 2.5c and 2.5d).

Moreover, the strongly impacted surface area decreased from 1.3 to 0.8 km 2 after the outlet pipe prolongation. This difference in the distribution of sediments before and after 2001 is explained by the difference in velocity of lake currents at different water depths.

Figure 2.5a . Spatial distribution of Hg (mg .kg -1 dry weight sediment) in 1996, in the surface sediments of the Bay of Vidy. Interpolation was obtained by kriging using the program Surfer 8 .

Figure 2.5b . Spatial distribution of Hg (mg .kg -1 dry weight sediment) in 2005, in the surface sediments of the Bay of Vidy. Interpolation was obtained by kriging using the program Surfer 8 .

Figure 2.5c . Spatial distribution of Cu (mg .kg -1 dry weight sediment) in 1996 in the surface sediments of the Bay of Vidy. Interpolation was obtained by kriging using the program Surfer 8 .

Figure 2.5d . Spatial distribution of Cu (mg .kg -1 dry weight sediment) in 2005, in the surface sediments of the Bay of Vidy. Interpolation was obtained by kriging using the program Surfer 8 .

The last general evaluation of the sediment contamination by metals of Lake

Geneva is 20 years old (Arbouille et al., 1989). Compared to these results, the recent sediments of the Bay of Vidy showed higher concentrations, at least for some metals.

For instance in 2005, contents of Cu, Cd, Hg, Pb, Zn, and Ag, are 4.3 to 6.3 times higher in the effluent outlet region than the mean content in the rest of the lake.

The relationship between pollutant content and grain size is not as strong as one could expect from the general transport and deposition model of hydrophobic pollutants (Salomons and Förstner, 1984). This observation probably results from the fact that pollutants are attached to both large organic and small inorganic particles

(such as clay) which behave in a similar manner, with respect to transport and sedimentation. This hypothesis is supported by a recent study that demonstrates that high, comparable Hg concentrations may be present in all grain size fractions (<

63μm; 63μm – 1mm; > 1mm) in sediments from the Bay of Vidy (Garcia et al., 2007).

Based on the surface distribution of the various metals, it has been possible to estimate their inventory in the upper sediments. For this estimate, we considered the first 3 cm of sediment surface, with a porosity of 90% and a dry sediment density of

2.37 g·cm-3, corresponding to 90% of inorganic material (2.5 g·cm -3) and 10% of organic matter (1.2 g·cm -3) (Monna et al., 1999). This gives a specific mass of sediment of 7.11 kg·m -2. Using this mass of sediment, the directly impacted surface area of 1.32 km 2, and the surface distribution on the contaminants, we estimated

1) the mean concentration and the inventory of metals released by the sewage treatment plant before and after the diversion of the effluent pipe,

2) a reduction factor by comparing the inventory between 1996 and 2005, and

3) a contamination factor for 1996 and 2005, by dividing the levels of the 1996 and 2005 inventory by the “natural background” inventory (Table 2.5).

Table 2.5. Comparison of the inventory of various metals in the upper 3 cm of the sediments. The considered zone has a surface area of 1.32 km 2. Inventory of the 1996 and 2005 surveys include the natural background inventory

Metal Natural background Mean content Inventory Reduction Contamination factor factors Concentration* Inventory 1996 2005 1996 2005 Ratio 1996 2005 survey survey survey survey 1996/2005 survey survey mg .kg -1 kg mg .kg -1 mg .kg -1 kg kg Hg 0.03 0.28 1.5 1.2 14 10.9 1.3 49.6 39.0 Cd 0.2 1.9 4.8 1.7 45 15.6 2.9 24.0 8.4 Zn 50 468 763 66 7130 620 11.5 15.3 1.3 Cu 20 187 243 109 2273 1016 2.2 12.2 5.4 Pb 20 187 222 60 2074 562 3.7 11.1 3.0 Cr 30 281 49 38 462 359 1.3 1.6 1.3 Ag 0.2 1.9 1.4 12.7 6.8 *Natural background concentrations are based on Arbouille et al. (1989).

The comparison of the natural background inventory with 1996 and 2005 inventories shows that in 1996, for all metals except Cr, the anthropogenic sources amounts for more than 90% of the contents, with a maximum of 98% for Hg.

For this period, metals can be grouped into three different classes, according to their contamination factors. A first group comprises Hg and Cd with high contamination factors of 49.6 and 24, respectively. The second group comprises Cu,

Pb and Zn with contamination factors between 11.1 and 15.3, and the last metal is Cr, with a low contamination factor of 1.6.

As seen in the previous section, the diversion of the effluent pipe has reduced the impacted surface area in the bay. Estimation of the inventory of metals between

1996 and 2005, show as well a general reduction of metal loads in the bay. Zn and Cu have decreased by a factor 11.5 and 2.2 respectively. Chromium, which was already low before the modification of the outlet pipe, decreased by a factor of 1.3. The only

55 metal that still presented a high contamination factor is Hg, with a value of 39 for the

2005 survey. Its inventory decreased by a factor of 1.3 only. That means that, although the mean content in Hg was a bit lower in 2005 than 1996, the contaminated surface area decreased from 1996 to 2005, meaning that the concentration of Hg at some points in the remaining contaminated zone was higher than for the 1996 survey. This is shown by the maximum measured value of 8.65 mg .kg -1 in 2005 compared to 4.28 mg .kg -1 in 1996.

Despite the fact that only three samples have been analyzed, total PCB levels have certainly decreased when compared to results of previous studies.

Burgenmeister et al. (1983) measured concentrations between 260 and 430 μg .kg -1 of

PCB in sediments at about 25 m water depth, whereas Thomas et al. (1984) calculated a mean value of 43 ± 28 μg .kg -1 for 80 sediment samples covering the entire lake, with a maximum value of 220 μg .kg -1 in the Bay of Vidy, at about 950 m from the former effluent outlet. In 1986, Corvi et al. measured total PCB levels ranging from 9 to 103

μg .kg -1, in the surface sediments of Lake Geneva. It has been reported that the concentrations of PCBs in mussels and fish decreased from past to present. Recent concentrations were mostly low and compliant with Swiss regulations (Corvi et al.,

2005; Rapin et al., 1995). In this study, the total PCB content was of 156 μg .kg -1 at 130 m (Vs 14), and 86 μg .kg -1 at 340 m (Vs 12) from the current effluent outlet (Table 2.4).

2.4.2 Heavy metal and hydrophobic organic compound concentration levels: comparison with standard values and potential biological effects. The concentration values obtained in this study and the former values of year

1996 (Table 2.6) were compared with the Canadian Sediment Quality Guidelines for the Protection of Aquatic Life (CCME EPC-98E, 1999). The sediment concentration values from the most polluted point in the Bay of Vidy were 3 to 13 times higher than the PELs (probable effect levels) from the Canadian Sediment Quality Guidelines..

This point is strongly contaminated in heavy metals, especially in zinc and mercury,

56 and would be expected to be frequently associated with adverse biological effects.

Although no threshold values for the concentrations of inorganic and organic compounds in the sediments of freshwater reservoirs or lakes are proposed in the

Swiss federal legislation for the protection of freshwater, the document mentions that no accumulation of contaminants in sediments, due to human activities should occur.

According to these standards, it appears clearly that the surface sediments of the Bay of Vidy are heavily contaminated with heavy metals and still present important environmental impacts due to the WWTP effluent discharge into the lake.

Table 2.6. Maximum heavy metals levels in sediments; results of this study and local results obtained in 1996 compared with the Canadian Sediment Quality Guidelines for the Protection of Aquatic Life recommendation (CCME EPC-98E, 1999).

Metal Year 1996* Year 2005 Recommendations** Maximum Maximum Maximum concentration a Sample sites b concentration a Sample sites b concentration a Cd 15.7 534440/151398 18.4 534676/151543 0.60 Cu 765 534440/151398 727 534676/151543 35.70 Cr 256 534440/151398 337 534676/151543 37.30 Pb 706 534516/151794 619.7 534676/151543 35.00 Zn 2268 534655/151780 3609 534676/151543 123.00 Hg 4.54 534655/151780 8.65 534676/151543 0.17 * Sediments recovered year 1996 (Pardos et al., 2004); ** "Canadian Sediment Quality Guidelines for the Protection of Aquatic Life" Recommendations; a Concentration unit; mg .kg -1 dry weight sediment; b X/Y Swiss coordinates (m).

No survey of the recent sediments toxicity has been realized. However an evaluation of the potential deleterious effects of heavy metals towards benthic fauna, based on consensus-based guidelines for the sediment quality (McDonald et al, 2000), gives an estimate of the hazard these sediment may represent for the biota. These guidelines proposed for some metals “probable effect concentrations” (PEC), above which there is a probable toxic effect for the biota. These values can be integrated by calculating the mean PEC Quotients (mPECQ, Table 2.7) which is the mean of the

PEC quotients as defined by Long et al. (2006). In order to compare results,

57 calculation has been performed for the 1.3 km 2 most contaminated area of Vidy Bay.

In 1996, the distribution of the mPECQ shows a median value of 0.93 whereas in 2005 it decreases to 0.68. The distribution of these mPECQ shows that in 1996, the 1.3 km 2 surface impacted by the WWTP presents mPECQ value bigger than 0.5. This surface was reduced by one third in 2005 to 0.81 km 2. However the maximum value observed in 2005, close to the effluent pipe, was 4.3 compared to 3.9 in 1996.

According to Long et al. (2006), sediments with mPECQ larger than 0.5 present an incidence of toxicity of about 40% (amphipod survival tests) whereas an incidence of more than 75% is observed for mPECQ higher than 5.

Table 2.7. Comparison of the Mean Probable Effect Concentration Quotients value

(mPECQ, Long et al., 2006) of the Bay of Vidy sediments for the 1996 and 2005 sampling campaigns mPECQ 1996 2005 Minimum value 0.06 0.22 Maximum value 3.9 4.3 Median value 0.93 0.68 Surface with mPECQ > 0.5 (km 2) 1.21 0.81

2.5 Conclusion The purpose of this study was to measure levels of total heavy metals and hydrophobic organic compound in surface sediments of the Bay of Vidy. The values and spatial distribution of heavy metals were compared with those measured in

1996, before the WWTP outlet pipe extension. The main conclusions are the following:

− High concentrations of HMs and HOCs are still observed in the sediments.

These concentrations vary as a function of the sampled sites. The highest

levels are located close to the current WWTP outlet pipe discharge into the

lake and diminish in a concentric pattern with distance from the point of

discharge.

− In the most impacted area close to the outlet (1.32 km 2), metal inventories

in the surface sediments decreased by a factor between 11 for Zn and 2.2

for Cu in comparison with the pre-2001 outlet location. Chromium, which

was already low before the modification of the outlet pipe, decreased by a

factor of 1.3. The only metal that still presents a very high contamination

factor is Hg, with a value of 39 for the 2005 survey. Its inventory

decreased by a factor 1.3 only.

− The comparison of metal concentrations with “probable effect

concentrations” (PEC), indicates possible toxic effects of sediment

contamination for the biota close to the WWTP outlet.

− The sediment concentration values from the most polluted point in the

Bay of Vidy are 3 to 13 times higher than the PELs ( probable effect levels)

from the Canadian Sediment Quality Guidelines for the Protection of

Aquatic Life (CCME EPC-98E, 1999). However, further ecotoxicological

tests would be necessary to assess the toxicity of these sediments and their

potential biological effects.

− Although the WWTP outlet pipe extension in 2001 improved the situation

with respect to the accumulation of contaminated sediments, the Bay of

Vidy is confirmed to be the most contaminated region of Lake Geneva.

The original goal of extending the WWTP outlet pipe, to significantly

reduce contamination in the coastal area of Vidy Bay, has not been entirely

reached. There are as well, no data available about the evolution of the

water quality at the St. Sulpice water pumping station.

− An improvement of the WWTP’ efficiency will be needed to remediate to

the current situation.

References

APHA (American Public Health Association) (1985) Standard methods for the

examination of water and wastewater. American Public Health Association ,

Washington, 408-410.

Arbouille D, Howa H, Span D & Vernet JP (1989) Etude générale de la pollution par

les métaux et répartition des nutriments dans les sédiments du Léman. Rapp.

Comm. Int. Prot. Eaux Léman Contre Pollut ., campagne 1988, 139-172.

Burgenmeister G, Aswald K, Machao L, Mowrer J & Tarradellas J (1983)

Concentrations en PCB et DDT des sédiments superficiels de la rive Suisse du

lac Léman. Schweiz. Z. Hydrol . 45 : 233-245.

Burrus D, Thomas RL, Dominik B, Vernet JP & Dominik J (1990) Characteristics of

suspended sediment in the Upper Rhone River, Switzerland, including the

particulate forms of phosphorus. Hydrol. Process. 4: 85–98.

Cantwell MG, Burgess RM & Kester DR (2002) Release and phase partitioning of

metals from anoxic estuarine sediments during periods of simulated

resuspension. Environ. Sci. Technol . 36 : 5328-5334.

CCME EPC-98E (Canadian Council of Ministers of the Environment) (1999)

Canadian Sediment Quality Guidelines for the protection of Aquatic Life.

http://www.ccme.ca/

Corvi C, Majeux C & Vogel J (1986) Les polychlorobiphényles et le DDE dans les

sédiments superficiels du Léman et de ses affluents. Rapp. Comm. Int. prot. eaux

Léman contre pollut . Campagne 1985, 205-216.

Corvi C, Zimmerli P, Ortelli D, Khim-Heang S & Becker van Slooten K (2005) Métaux

et micropolluants organiques dans les eaux, les moules et les poisons du Léman.

Rapp. Comm. Int. prot. eaux Léman contre pollut . Campagne 2004, 55-78.

Garcia Bravo A, Willot M & Wildi W (2007) Mercury distribution in the sediment

grain-size fractions of a contaminated site of Lake Geneva. 5 th Swiss Geosciences

Meeting, Geneva, November 16-17, 2007.

Goldscheider N, Haller L, Poté J, Wildi W & Zopfi J (2007) Characterizing Water

Circulation and Contaminant Transport in Lake Geneva Using Bacteriophage

Tracer Experiments and Limnological Methods. Environ. Sci. Technol . 41 : 5252-

5258.

Hall GEM & Pelchat P (1997) Evaluation of a Direct Solid Sampling Atomic

Absorption Spectrometer for the Trace Determination of Mercury in Geological

Samples. The Analyst 122 : 921-924.

Harwood JE, Van Steenderen RA & Kühn AL (1969) A rapid method for

orthophosphate analysis of high concentrations in water. Wat. Res. 3: 417-423.

Kang Y, Sheng G, Fu J, Mai B, Zhang G, Lin Z & Min Y (2000) Polychlorinated

Biphenyls in Surface Sediments from the Pearl River Delta and Macau. Mar.

Pollut. Bull. 40 : 794-797.

Loizeau JL, Arbouillle D, Santiago S & Vernet JP (1994) Evaluation of a wide range

laser diffraction grain size analyser for use with sediments. Sedimentology 41 :

353-361.

Loizeau JL, Pardos M, Monn, F, Peytremann C, Haller L & Dominik J (2004) The

impact of a sewage treatment plant’s effluent on sediment quality in a small bay

in Lake Geneva (Switzerland-France). Part 2 : Temporal evolution of heavy

metals. Lakes Reservoirs: Research and Management 9: 53-63.

Loizeau JL, Rozé S, Peytremann C, Monna F & Dominik J (2003) Mapping sediment

accumulation rate by using volume magnetic susceptibility core correlation in a

contaminated bay (Lake Geneva, Switzerland). Eclogae geol. Helv. 96 : supp. 1, 73-

79.

Long ER, Ingersoll CG & MacDonald DD (2006) Calculation and uses of mean

sediment quality guideline quotients: a critical review. Environ. Sci. Technol . 40 :

1726-1736.

Lors C, Tiffreau C & Laboudigue A (2004) Effects of bacterial activities on the release

of heavy metals from contaminated dredged sediments. Chemosphere 56 : 619-

630.

MacDonald DD, Ingersoll CG & Berger T (2000) Development and evaluation of

consensus-based sediment quality guidelines for freshwater ecosystems. Arch.

Environ. Contam. Toxicol. 39 : 20-31.

Monna F, Dominik J, Loizeau JL, Pardos M & Arpagaus P (1999) Origin and

evolution of Pb in sediments of Lake Geneva (Switzerland - France).

Establishing a stable Pb record. Environ. Sci. Technol. 33 : 2850-2857.

Murphy J & Riley JP (1962) A modified single solution method for the determination

of phosphate in natural waters. Anal. Chim. Acta 27 : 31-36.

Pardos M, Benninghoff C, Alencastro LP, Wildi W (2004) The impact of a sewage

treatment plant's effluent on sediment quality in a small bay in Lake Geneva

(Switzerland-France). Part 1: Spatial distribution of contaminants and the

potential for biological impacts. Lakes Reservoirs: Research and Management 9 : 41-

52.

Peytremann C & Haller L (1997). Etude sédimentologique, géochimique et

écotoxicologique de sédiments de la baie de Vidy. Unpubl. Master Thesis, Univ.

Geneva., 121 p.

Rapin F, Blanc P, Pelletier JP, Balvay G, Gerdeaux D, Corvi C, Perfetta J & Lang C

(1995) Impacts humains sur les systèmes lacustres : exemple du Léman. In :

Pourriot, R., Meybeck, M. (Eds), Limnologie Générale. Masson, Paris, 806-840.

Ross-Barraclough F, Givelet N, Martinez-Cortizas A, Goodsite ME, Biester H &

Shotyk W (2002) An analytical protocol for the determination of total mercury

concentrations in solid peat samples. Sci. Tot. Environ . 292 : 129-139.

Salomons W & Förstner U (1984) Metals in the Hydrocycle. Springer-Verlag, Berlin,

349 p.

Thomas RL, Vernet JP & Frank R (1984) ΣDDT, PCBs, and HCB in the Sediments of

Lake Geneva and the Upper Rhône River. Environ. Geol. 5: 103-113.

Verweij F, Booij K, Satumalay K, Molen N & Oost R (2004) Assessment of

bioavailable PAH, PCB and OCP concentrations in water, using semipermeable

membrane devices (SPMDs), sediments and caged carp. Chemosphere 54 : 1675-

1689.

Wildi W, Dominik J, Loizeau JL, Thomas RL, Favarger PY, Haller L, Perroud A &

Peytremann C (2004) River, reservoir and lake sediment contamination by

heavy metals downstream from urban areas of Switzerland. Lakes Reservoirs:

Research and Management 9: 75-87.

Wu Y, Zhang J & Zhou Q (1999) Persistent organochlorine residues in sediments

from Chinese river/estuary systems. Environ. Pollut. 105 : 143-150.

CHAPTER 3

Composition of bacterial and archaeal

communities in freshwater sediments with

different contamination levels (Lake Geneva,

Switzerland)

A similar version of this chapter is currently under review in Water Research.

Haller L, Tonolla M, Zopfi J, Peduzzi R, Wildi W & Poté J. Composition of bacterial and archaeal communities in freshwater sediments with different contamination levels (Lake Geneva, Switzerland). Submitted.

Abstract Bacterial and archaeal communities in contaminated (Vidy Bay) and uncontaminated (Ouchy area) sediments of Lake Geneva were characterized using

16S rRNA gene diversity. In addition, sediments of both sites were analysed for physicochemical characteristics including porewater composition, organic carbon, and heavy metals. Results show high concentrations of contaminants in sediments from Vidy. Particularly, high contents of fresh organic matter and nutrients led to intense mineralization, which was dominated by sulphate-reduction and methanogenis. Phylogenetic analysis of the bacterial community composition revealed a dominance of Betaproteobacteria , in surface sediments at both sites. A large proportion of Betaproteobacteria clones in Vidy sediments were related to

Dechloromonas sp , a dechlorinating bacteria. Deltaproteobacteria, including clones related to sulphate-reducing bacteria and Fe(III)-reducing bacteria ( Geobacter sp.) were more abundant in the contaminated sediments. The archaeal communities consisted essentially of methanogenic Euryarchaeota , mainly found in the contaminated sediments rich in organic matter. The bacterial diversity in Vidy sediments was significantly different from the communities in the uncontaminated sediments. This research suggests that in addition to environmental variables, pollution could be one of the factors affecting microbial community structure.

3.1 Introduction Diversity of microbial communities may be a sensitive indicator of pollution in aquatic ecosystems, and heavy metal contamination can lead to a reduction of bacterial diversity (Sandaa et al., 1999). However, some studies observed either an increase in microbial diversity along with heavy metal contamination (Sorci et al.,

1999) or no significant variation (Gillan et al., 2005). Other environmental factors as well as time of exposure might explain these differences.

The city of Lausanne discharges the largest volume of treated wastewater into the nearby Bay of Vidy. The last chapter focused on the quality of bottom sediments in the Bay of Vidy and several studies report on the spatial distribution of organic matter, faecal indicator bacteria, heavy metals, and hydrophobic organic compounds

(Poté et al., 2008; Haller et al., 2009). However, there is still a paucity of information regarding sedimentary microbial communities in Lake Geneva and particularly around the WWTP discharge outlet.

Sediments are complex habitats densely colonized by diverse groups of microorganisms, which play key roles in biogeochemical cycling, aquatic food webs and the remobilisation of heavy metals (Nealson, 1997; Lors et al., 2004; Ye et al.,

2009). Many studies have been performed to examine the composition and variability of microbial communities in extreme or complex aquatic ecosystems. However, only a few studies compared microbial community structures in contaminated and uncontaminated sediments (Powell et al., 2003; Zhang et al., 2008). The aims of the present study were to compare the composition of the sediment-associated microbial communities in the contaminated Bay of Vidy with a closeby non-polluted site in the

Ouchy area. The Bay of Vidy is currently used as a model system for several limnological, biogeochemical, and ecotoxicological studies. This research represents the first assessment of Bacteria and Archaea in contaminated and uncontaminated

67 sediments of Lake Geneva and serves as important background information for these studies. To better understand the microbial community structures, molecular analyses were complemented by a detailed physicochemical characterisation of the sediments.

3.2 Materials and methods

3.2.1 Study site description and sampling procedure In August 2005, sediment was collected at two locations in Lake Geneva (Fig.

3.1): (i) within the Bay of Vidy near the outlet pipe of the WWTP of Lausanne (Swiss coordinates X: 534682, Y: 151410) and (ii) near the Ouchy area (Swiss coordinates X:

537985, Y: 150390). Sampling was done from R/V “La Licorne” using a core sampler

(Benthos Inc, USA). Three cores (6.7 cm i.d., 1.5 m length) were retrieved from each site at a depth of 40 m, near the WWTP outlet pipe. For microbiological analyses, the cores were opened longitudinally and sliced into 2 cm thick sections until 10 cm depth. The sediment samples were placed into sterile plastic containers, stored in an icebox and treated in the laboratory within 24 h. For chemical analysis, the intact sediment cores were transported to the laboratory and were stored vertically in a cold-room at 4°C until analysis.

Figure 3.1. Study area with the two sampling locations in the Bay of Vidy and the

Ouchy area.

3.2.2 Chemical analysis Two cores per site were used for the chemical analyses: one was used to measure the organic matter, nutrients and heavy-metals contents and the other one was used to determine the sulphur and iron concentrations and the porewater constituents.

Organic matter and nutrients : the cores were opened longitudinally and sliced every 2 cm down to a depth of 10 cm. Before analysis, sediment samples were air- dried at ambient room temperature. The particle grain size was measured with a laser Coulter® LS-100 diffractometer (Beckman Coulter, Fullerton, CA, USA), after a

5-min ultrasonic dispersal in deionized water according to the method described by

Loizeau et al. (1994). The proportions of three major size classes (clay < 2μm; silt 2-

63μm; and sand > 63μm) were determined from size distributions. Total organic matter content in sediments was estimated by loss on ignition at 550°C for 1 hour in a

Salvis oven (Salvis AG Emmenbrücke, Luzern, Switzerland) on 5 g of dried

69 sediments. Total organic carbon (TOC) was determined by titrimetry following acid oxidation (Méthodes de référence des Stations Fédérales de Recherches

Agronomiques, 1998) on 5 g of dried sediments. Total nitrogen (TN) was determined according to Kjeldahl (APHA, 1985) on 2 g of dried sediments. Total phosphorus (TP) and its different forms were measured on 150 mg of dried sediments with a spectrophotometer (Helios Gamma UV-Vis Thermo Electroporation , Thermo scientific, USA) at 850 nm, following the fractionation scheme of Williams et al.

(1976) as modified by Burrus et al. (1990). The results are expressed in mg kg -1 dry weight sediment (ppm).

Solid phases : sulphur and iron contents were determined every cm down to 10 cm depth. The sediment was sliced at room temperature in a N 2-filled glove bag and fixed in 50mL tubes containing 10mL zinc acetate solution (10%). Sulphide was extracted as acid volatile sulphur [AVS; dissolved sulphides (H 2S) and iron sulphides

(FeS)] and chromium reducible sulphur [CRS; primarily pyrite (FeS 2), elemental sulphur (S 0), and some organic sulphur]. AVS and CRS were measured by a two step distillation process with cold 6 N HCl followed by boiling 1 M acidic CrCl 2 solution

(Fossing and Jørgensen, 1989; Zopfi et al., 2008). Poorly cristalline Fe(III)oxides were extracted with 0.5M HCl according to Thamdrup and colleagues (1994).

Total contents of Cu, Cd, Cr, Zn, and Pb were determined by quadrupole- based Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (HP 4500, Agilent) following the digestion of 1 g of dried sediments in analytical grade 2 M HNO 3

(Pardos et al., 2004). Total Hg was quantified by atomic absorption spectrophotometry (Advanced Mercury Analyser; AMA 254, Altec s.r.l., Czech Rep.) according to Hall and Pelchat (1997) and Ross-Barraclough et al. (2002). The method is based on sample combustion, gold amalgamation and atomic absorption spectrometry (AAS). Results of triplicate measurements are expressed in mg kg -1 dry weight sediment (ppm).

Porewater constituents : Porewater was harvested by centrifugation (4000 rpm) under an N 2 atmosphere to avoid Fe 2+ and H 2S oxidation. Samples for dissolved Fe 2+ were fixed with 0.5M HCl and analyzed by the photometric Ferrozine method

(Thamdrup et al. 1994). Dissolved sulfide was determined on Zn-acetate fixed samples using the colorimetric methylene-blue method (Cline, 1969; Zopfi et al.

2008). The major anions Cl -, SO 42-, NO 3-, and PO 43- were measured by ion chromatography on a DIONEX DX-120 system using an IonPac® AS14A anion exchange column, Na 2CO 3 8mM/NaHCO 3 1mM as eluent, an Anion Self-

Regenerating Suppressor (ASRS® 300, 4mm) module and a conductivity detector.

3.2.3 DNA extraction The samples from 0-2 cm to 8-10 cm depth were used for microbiological analysis. Total DNA was extracted from 250 mg of sediment using the PowerSoil TM

DNA Isolation Kit (Mo Bio Laboratories, Inc.), according to the manufacturer’s instructions. DNA extracts were stored at –20 °C until used for PCR amplification.

3.2.4 PCR amplification Nearly complete bacterial 16S rRNA genes were amplified using the primers

26f (5’-AGAGTTTGATCATGGCTCA-3’) and

1392r (5’-GTGTGACGGGCGGTGTGTA-3’) (Brosius et al., 1981; Lane, 1991).

Archaeal 16S rRNA genes were amplified using the primers

109f (5’-ACKGCTCAGTAACACGT-3’) and

915r (5’-GTGCTCCCCCGCCAATTCCT-3’) (Grosskopf et al., 1998, Stahl and Amann,

1991).

Both archaeal and bacterial PCR amplification were performed separately using the Taq PCR Master Mix Kit (Qiagen, Basel Switzerland). Each PCR reaction was carried out in a volume of 50 μL, containing 1X PCR buffer, 1.5 mM of MgCl 2,

200 μM of each dNTP, 0.3 μM of each primer, 2.5 units of Taq DNA polymerase, 2.5

71 mg/ml of bovine serum albumine (Invitrogen, Basel Switzerland) and 2 μL of DNA

(about 100 ng) of each sample. The following conditions were used for PCR amplification: initial denaturation step at 94°C for 5 min; 35 cycles of denaturation

(94°C for 30 s), annealing (52°C for 30 s), extension (72°C for 1 min) and a final extension step of 10 min at 72°C. PCR products (10 µL) were separated by electrophoresis on 0.8% agarose gels and visualized by ethidium bromide staining and UV illumination.

3.2.5 Clone library construction and DNA sequencing For clone library construction and DNA sequencing only the samples from 0-2 cm and 4-6 cm depth were used. PCR products were purified using the NucleoSpin ®

Extract II Kit (Macherey-Nagel, Oensingen, Switzerland) according to the manufacturer’s instructions. The amplified DNA was then quantified using the

PicoGreen® dsDNA Quantitation Reagent (Molecular Probes Inc.) and a TD-700

Fluorometer (Turner Designs). Approximately 20-30 ng of amplified 16S rDNA were cloned into competent Escherichia coli cells using the TOPO TA cloning kit

(Invitrogen, Basel Switzerland) following the manufacturer’s recommendations. The transformed cells were plated on LB medium containing 50 mg/L ampicillin, 60 mg/L of IPTG (isopropyl-β-D-thiogalactopyranoside), and 100 mg/L of X-gal (5-bromo-4- chloro-3-indolyl-β- D-galactopyranoside), and incubated overnight at 37°C. White recombinants were transferred to LB medium plates for 24h. 80 and 40 recombinants were picked from samples taken at 0-2cm and 4-6cm, respectively, in each location, to constitute the Bacteria clone libraries. For the Archaea , only 2 recombinants (1 per site) were found at 0-2 cm and 34 colonies were picked from 4-6 cm. The insert size, of all picked colonies, was determined by direct PCR using M13 forward and reverse primers included in the cloning kit. The products were subsequently purified with the NucleoSpin ® Extract II Kit (Macherey-Nagel, Oensingen Switzerland) and sequenced using the BigDye® Terminator Cycle Sequencing Ready reaction kit

(Applied Biosystems). Sequences were obtained with an automated sequencing system (ABI PRISM ® 310 Genetic Analyser, Perkin Elmer).

3.2.6 Phylogenetic analysis The obtained sequences were checked and edited using the program EditSeq TM

(DNAStar Inc.).

A Chimera Detection program (http://rdp8.cme.msu.edu/cgis/chimera.cgi?su=SSU ) was used to exclude chimeras. NCBI Blast ( http://www.ncbi.nih.gov ) was used to identify the most closely related 16S rRNA gene sequences.

The partial 16S rRNA gene sequences were then all aligned in the Clustal W implementation of MEGA 3.0 (Kumar et al., 2004). The same program was used to produce neighbour-joining phylogenetic trees (Kimura-2 correction; bootstrap values for 500 replicates). The sequences were identified using the ribosomal database project classifier ( http://rdp.cme.msu.edu/classifier/classifier.jsp All the sequences described in this study have been submitted to the EMBL database under the accession numbers FN679050 to FN679294.

The sequences were assigned to individual OTUs based on the 97% sequence similarity criterion. The number of OTUs per sample and the rarefaction curves were estimated from our sequence data using the program Mothur v.1.12.3 (Schloss et al.,

2009). Comparison between the clone libraries from the two sites was done on the basis of genetic diversity by means of the parsimony test using Mothur v.1.12.3

(Schloss, 2009) and the the FST -test using the program Arlequin, v.2.0 (Schneider et al., 2000).

A Mantel test using the Spearman correlation coefficient (Mantel, 1967) was undertaken to assess the correlations between the microbial dataset from the 2 sites and the environmental variables. A Multiple factor analysis (MFA) was done to obtain an integrative picture of the relationship between the bacterial community structures and the environmental factors at the 2 sites. The environmental parameters

73 were separated into 2 matrices, one including the organic matter and nutrients variables and one with concentrations of heavy metals. Briefly, the MFA allows the simultaneous ordination of a composite table obtained by the juxtaposition of the species and environmental data sets, after weighting the different matrices (Escofier

& Pages, 1994). The final ordination plot shows global points indicating the relative positions of the objects described by the combination of datasets. Each global point is surrounded by partial points indicating the relative positions of the datasets taken separately (Escofier & Pages, 1994; Becue-Bertaut & Pages, 2008). These statistical analyses were carried out on "R", a free software environment for statistical computing and graphics, using the Vegan library (R Development Core Team, 2005).

3.3 Results

3.3.1 Chemical analysis Some general sediment characteristics including particle grain size, organic matter and nutrient contents are presented in Table 3.1. These sediments were mostly composed of silts, approximately 65% and 80% for Vidy and Ouchy, respectively.

Average organic matter contents in Vidy Bay sediments were much higher (21%) than in Ouchy sediments (4.5%). Average nutrient concentrations such as total nitrogen, ammonium and total phosphorus were also found to be higher in Vidy sediments (12.6 ppm, 1.6 ppm and 8217 ppm respectively) while they were considered low in the sediments from Ouchy (2 ppm, 0.5 ppm and 810 ppm respectively). However, at both stations, no significant differences in the nutrient contents between samples collected at 0-2 cm and 4-6 cm were observed.

Table 3.1. Sediment characteristics in different sediment layers of the Ouchy (non- contaminated) and Vidy (contaminated) sites.

Sediment NH4-N Clay/silt/sand OM (%) TOC (%) TN (mg kg -1) TP (mg kg -1) section (mg kg -1) proportion (%) Vidy 0-2cm 18.7 10.8 11.6 1.6 6783.8 0/64/35 Vidy 4-6cm 23.7 13.7 13.5 1.6 9650.9 0/65/34 Ouchy 0-2cm 5 2.9 2.2 0.5 862.1 0.9/78/21 Ouchy 4-6cm 4.3 2.5 1.7 0.4 759.5 0.6/80/19 OM: Organic matter TOC: Total organic carbon TN: Total nitrogen NH4-N: Ammonium TP: Total phosphorus

Porewater concentrations of major ions, including Fe 2+ , H 2S (= H 2S+HS -+S 2-) and SO 42- are shown in Fig. 3.2.

Nitrate was not detected in either one of the two cores. In the Ouchy sediment porewater, sulphate concentrations dropped from about 550 μM at the surface to 0

μM at 7 cm depth, whereby maximum net sulphate reduction was indicated below

3.5 cm. Sulphate was essentially depleted in the Vidy Bay sediment through intense sulphate reduction.

Dissolved Fe 2+ concentration were only high in Vidy sediment where they reached 260 μM. It is unlikely that under such Fe 2+ -rich conditions free sulphide exists in porewater. Sulphide produced during sulphate reduction reacts rapidly with iron to form FeS. Hence, measured porewater sulphide in the Vidy core represents probably soluble FeS complexes or colloidal FeS phases.

High concentrations of porewater PO 43-, reaching about 270 uM were only detected in the uppermost cm of Vidy Bay sediments, indicating intese mineralization and phosporous liberation.

Minor amounts of Fe(III)oxides were only detected in the uppermost section of the Ouchy sediment core. Below 1 cm depth, and in the whole core from the Vidy

Bay, HCl extractable Fe(III) was absent and all of the iron was present in its reduced form (Fig. 3.2).

Differences in the iron-sulphur geochemistry exist between the two sites as indicated as well by the AVS and CRS data (Fig. 3.3) and the S/Fe-ratio (Fig. 3.4).

Unlike in the Vidy sediments, solid phase sulphur species in the Ouchy sediment accumulate gradually with depth and reach a plateau at around 3 cm. Less reduced sulphur species such as S 0, FeS 2, Fe 3S4 may play a role as indicated by higher CRS contents and S/Fe ratios >1 at certain depths (Fig. 3.4). Iron contents in the Vidy sediments were about 50% higher than in the Ouchy site and the constant ration of

S/Fe suggests FeS as dominant phase.

2- µ Concentration SO 4 ( M) 0 100 200 300 400 500 600 0 0

2- 2 SO 4 2

4 4 H2S

Depth (cm) 6 H2S 6

8 8

SO 2- Ouchy 4 Vidy 10 10 0 5 10 15 20 25 0 5 10 15 20 25 µ 2- µ ) Concentration H 2S ( M) Concentration SO 4 ; H 2S ( M

Porewater Fe 2+ ( µM)

0 50 100 150 200 250 300 0 0

2 2 Fe 2+

4 4

Depth (cm) 6 6 Fe 2+

8 8

Fe(III) Fe(III) Ouchy Vidy 10 10 0 10 20 30 40 50 0 10 20 30 40 50 Porewater Fe 2+ (µM) Solid phase Fe(III) ( µmol g -1 ) Solid phase Fe(III) ( µmol g -1 )

Figure 3.2. Porewater sulphate and sulphide concentrations in sediments from Vidy and Ouchy (above). Porewater iron(II) and solid phase iron(III)oxides concentrations in sediments from Vidy and Ouchy (below).

0 0 AVS

CRS 2 2 TRS

4 4

Depth(cm) 6 6

8 8

Ouchy Vidy 10 10 0 25 50 75 100 125 150 0 25 50 75 100 125 150 Concentration S ( µmol g -1 ) Concentration S ( µmol g -1 )

Figure 3.3. Acid volatile sulfur (AVS = H2S + FeS) , chromium reducible sulfur

(CRS = S0 + FeS 2 + Fe 3S4) and total reducible sulfur (TRS = AVS + CRS) contents in sediments from Vidy and Ouchy.

S/Fe Ratio 0.0 0.5 1.0 1.5 2.0 0

Depth(cm) 6

Figure 3.4. Sulfur/iron ratios of the solid phases in sediments from Vidy (open circles) and Ouchy (closed circles).

The contents of heavy metals (in mg kg -1) are reported in Table 2. All measured metal concentrations, except for Cr, were significantly higher in the sediments collected in the Bay of Vidy, where peak concentrations reached 2.8 mg kg -

1 for Cd, 161.4 mg kg -1 for Cu, 164.7 mg kg -1 for Pb, 2.2 mg kg -1 for Hg at 8-10 cm depth.

Table 3.2. Depth variation of heavy metal contents (mg .kg -1 dry weight sediment)* in

Vidy Bay and Ouchy sediments.

Cu Zn Cd Pb Cr Hg

Depth (cm) Ouchy Vidy Ouchy Vidy Ouchy Vidy Ouchy Vidy Ouchy Vidy Ouchy Vidy

0-2 cm 54.7 142.5 126.8 341.3 0.5 1.6 38.8 89.4 60.3 74.2 0.2 1.2

2-4 cm 68.3 135.9 155.1 327.8 0.7 1.5 49.7 89.3 63 68.3 0.6 1

4-6 cm 86.8 133.6 218.1 344.9 1.2 1.7 75 100.7 69.7 67.9 1 2.3

6-8 cm 135.5 181.4 305.5 446.8 1.7 2.3 106.5 135.9 88.4 77.2 1.3 2.3

8-10 cm 136.5 161.4 242.8 518.2 1.2 2.8 91 164.7 81.4 88.9 1.7 2.2

* Total variation coefficients for triplicate measurements are below 15 % for all elements.

3.3.2 Bacterial 16S rRNA gene clone libraries At each site, sediment samples from 0-2 cm and 4-6 cm depth were used for clone library construction of bacterial 16S rRNA genes. Phylogenetic analysis showed that approximately 85% of the retrieved clones fell into known divisions with the rest remaining unclassified. Seven and twelve divisions were identified in the sediments from Vidy and Ouchy, respectively (Fig. 3.5). The dominant groups ( Beta-, Gamma-,

Deltaproteobacteria and Bacteroidetes ) were found at both sites and all depths. Most sequences in the Bay of Vidy were affiliated with Betaproteobacteria 41% (0-2 cm) and

24% (4-6 cm), Deltaproteobacteria 18% (0-2 cm) and 10% (4-6 cm) , Gammaproteobacteria

14% (0-2 cm) and 17% (4-6 cm) , and Bacteroidetes 21% (0-2 cm) and 27% (4-6 cm). In the Ouchy area, most sequences were related as well to Betaproteobacteria 33% (0-2

79 cm) and 5% (4-6 cm), Deltaproteobacteria 2% (0-2 cm) and 13% (4-6 cm),

Gammaproteobacteria 25% (0-2 cm) and 23% (4-6 cm) , and Bacteroidetes 15% (0-2 cm) and 8% (4-6 cm).

45 Ouchy 0-2cm 40 Ouchy 4-6cm 35 Vidy 0-2cm 30 Vidy 4-6cm

% of clones per sample per % of clones 5

a ia a a ia ri r lexi tes ri wn e teria e tes o t te of c ce rob e n c ac r a y d k b itrospira lo b roidetes m a n oba o N h o e o umic U te te C n t rr o a c cidobact ro roteobacteriar y n A e p p p C Bact la V ta a a P e h B mm Deltaproteobacteri a Alp Gemmatimon G

Figure 3.5. Relative abundance of bacterial taxonomic groups in the clone libraries established with sediments from the Bay of Vidy and the Ouchy area.

The sequences were assigned to individual OTUs based on their phylogenetic positions and the 97% sequence similarity criterion. The sequences were assigned to individual OTUs based on their phylogenetic positions and the 97% sequence similarity criterion. The obtained sequences (n = 208) were grouped into 132 OTUs:

40, 26, 53 and 32 for Vidy 0-2cm, Vidy 4-6cm, Ouchy 0-2 cm and Ouchy 4-6 cm respectively (Table 3.3). The number of unique OTUs was higher in Ouchy than in

Vidy, 38% and 55% of OTUs were exclusively found in Vidy and Ouchy respectively, and only 8% of all OTUs were shared between the libraries from both sites under investigation. The OTU richness estimates based on rarefaction curves suggests a higher bacterial diversity for the sediments from Ouchy than from Vidy (Fig. 3.6). By

80 means of the parsimony test (P-Test) and the F ST -test, a significant genetic differentiation between the sediment bacterial communities has been identified between the 2 sites Vidy and Ouchy at all depths (0-2 and 4-6 cm), as well as between the 2 samples from Ouchy taken at different depth (Table 3.4). Significance for both tests signals less genetic diversity within each community than for two communities combined and that the different communities harbour distinct phylogenetic lineages

(Martin 2002).

Table 3.3. Distribution of Bacteria phylotypes in clone libraries from Vidy and Ouchy sediments (Lake Geneva). “A” stands for the 0-2 cm sediment section and “B” for the

4-6 cm sediment section.

Vidy A Vidy B Ouchy A Ouchy B Total no of 68 41 60 40 sequences Total no of 40 26 53 32 OTUs Percentage of unique 17 10 30 18 OTUs

Table 3.4. Summary of F ST and P tests for comparison of microbial communities between Vidy and Ouchy sediments. “A” stands for the 0-2 cm sediment section and

“B” for the 4-6 cm sediment section.

P value a

Group FST test P test Vidy A vs Vidy B NS NS Ouchy A vs Ouchy B < 0.00 0.001 Vidy A vs Ouchy A 0.001 0.001 Vidy B vs Ouchy B 0.005 0.001 a NS, not significant

30 of OTUs observed

o 20 N

Vidy 0-2cm 10 Ouchy 0-2cm Vidy 4-6cm Ouchy 4-6cm

0 0 20 40 60 80 Number of clones sampled Figure 3.6. Rarefaction curves for the Bacteria 16S rRNA gene sequences retrieved fom Vidy and Ouchy sediment samples (depth intervals: 0-2 cm and

4-6 cm). Operational taxonomic units were defined with a 97% sequence similarity cut-off.

3.3.3 Archaeal 16S rRNA gene clone libraries All Archaea found in both sites fell into the Euryarchaeota division and most of them were found in sediments at 4-6cm, except for 2 clones, which were retrieved from surface sediments: one from Vidy and one from Ouchy. The 36 obtained sequences were grouped into 18 OTUs in total: 1 OTU for Vidy 0-2cm, 5 OTUs for

Vidy 4-6cm, 1 OTU for Ouchy 0-2cm, and 12 for Ouchy 4-6cm. A large proportion of these Euryarchaeota phylotypes, mostly originating from Vidy sediment, were assigned to the methanogenic families Methanosaetaceae and Methanomicrobiaceae.

3.4. Discussion Sediments were sampled in the area of Ouchy, a site close to Lausanne city without known release of contaminated water, and in the Bay of Vidy near the

WWTP outlet pipe, a highly polluted area of Lake Geneva. This is a site contaminated with heavy metals, hydrophobic organic compounds and faecal indicator bacteria (Pardos et al., 2004; Poté et al., 2008; Haller et al., 2009).

Organic matter contents in Vidy Bay sediments ranged from 18.7% (0-2 cm) to

23.7% (4-6 cm), and were much higher than in the Ouchy area (max. 5%) and elsewhere in Lake Geneva (max. 5 to 8%) (Poté et al., 2008). As a consequence, the

Vidy sediments are completely reduced by sulphate-reduction and methanogenesis as the dominant degradation processes. In the Ouchy sediments iron-reduction may prevail in the uppermost 1-2 cm of the sediment. Underneath, at 2.5 cm depth, sediments become sulphidic due to the activity of sulphate-reducing bacteria and below 7 cm methanogenic microorganisms take over (Canfield and Thamdrup, 2009).

Heavy metal concentrations in the Bay of Vidy were up to six times higher than in Ouchy. According to the Canadian Sediment Quality Guidelines for the

Protection of Aquatic Life (Conseil canadien des ministres de l’environnement, 1999) the heavy metal concentrations in Vidy Bay are 2 to 8 times higher than reported

PELs (probable effect levels). These results are in agreement with previous data from the same sampling point (see chapter 2). The concentration levels of hydrophobic organic compounds, such PAHs (polycyclic aromatic hydrocarbons), PCBs

(polychlorinated biphenyls) and OCPs (organochlorine pesticides) were also investigated in

Poté’s study and were up to 2, 156 and 45 μg kg -1 respectively. These values are considered high and above average levels for Lake Geneva (Corvi et al., 1986).

Although no threshold values for the concentrations of inorganic and organic compounds in the sediments of freshwater reservoirs or lakes are proposed in the

Swiss federal legislation on freshwater protection, the document mentions that no

83 accumulation of contaminants in sediments, due to human activities, should occur.

According to these standards, it appears clearly that the sediments around the outlet sewage pipe in Vidy are heavily contaminated with many kinds of pollutants, possibly representing a significant source of toxicity for microbial communities and benthic organisms.

Results from this study indicate that the dominant phylogenetic groups at both sites were the Beta-, Gamma- and Delta- subgroups of Proteobacteria and the

Bacteroidetes , which is in agreement with other 16S rRNA analyses of lake bacterioplankton (Zwart et al., 2002; Glöckner at al, 2000; Hiorns et al., 1997).

Nevertheless some differences between the two sites were observed in the relative proportion and abundance of the different Proteobacterial subdivisions . Proteobacteria accounted for 64% of the clones in Vidy and 55% of the clones in Ouchy. Members of the sub-divisions Beta and Deltaproteobacteria were much more abundant in the sediments from Vidy than Ouchy. In contrast, the Gammaproteobacteria were more abundant in Ouchy than Vidy (Fig. 3.5).

Betaproteobacteria appear to be numerically important in freshwater lakes

(Zwart et al., 2002; Hahn, 2006). In our study, that group was well represented in both locations (37% of clones in Vidy, 22% in Ouchy), especially in surface sediments where they were the most abundant (Fig. 3.7). It included different families and species of various metabolic types. A highly prevalent group of bacteria in Vidy sediments, belonged to the Rhodocyclaceae , which are phenotypically, and ecologically very diverse. Twenty clones of this family were affiliated to Dechloromonas , a genus that encloses many heterotrophic and facultative anaerobic bacteria that use chlorate, perchlorate or nitrate as alternative electron acceptors (Wolterink et al. 2005). A few other clones also found only in Vidy sediments were related to Methylophilus

(Methylophilaceae ) a group of methylotrophic organisms. However, while methanol is oxidised as the sole carbon and energy source, some species may grow also on a

84 limited range of other carbon compounds such as methylamines, formate, glucose, and fructose (Jenkins et al., 1987). Similar Methylophilaceae sequences were also found in wastewater treatment pools in China (sequenceAY863077) and in freshwater calcareous mats (sequence EF580978). The rest of the Betaproteobacteria clones were found in both sites and included various genera. Several clones were affiliated to

Propionivibrio (Rhodocyclaceae ), which are aerotolerant or strict anaerobe and chemoorganotrophic bacteria. These bacterial species are typical of anaerobic mud of freshwater sediments (Tanaka et al., 2003). Two other clones were related to

Thiobacillus (Hydrogenophilaceae) , which thrive on reduced sulfur compounds and are ubiquitous in soils and sediments. An additional mixed group of clones was closely related to the genus Rhodoferax (Comamonadaceae). This group of purple non-sulfur bacteria is physiologically very versatile. A recently described isolate from Oyster

Bay (USA) was a facultatively anaerobic bacterium that coupled the oxidation of acetate to the reduction of Fe(III)oxides (Finneran et al., 2003). A few clones, also belonging to the Comamonadaceae , sampled in Vidy sediments, were similar to a sequence that was collected from a group of microorganisms involved in anaerobic digestion of sludge (Riviere et al., 2009. CU918612). Two groups of sequences remained unclassified but were similar to sequences retrieved from Siberian tundra soils (EU644256) and anoxic river sediments (EF667529).

VidyA-51 (FN679057) VidyA-69 (FN679050) VidyA-48 (FN679055) VidyA-66 (FN679053) 85 OuchyB-15 (FN679054) VidyB-50 (FN679052) VidyB-11 (679056) 53 VidyB-13 (679051) OuchyB-12 (FN679058) VidyB-36 (FN679059) Dechloromonas sp. 62 VidyB-44 (FN679060) 65 VidyB-26 (FN679061) 59 VidyB-7 (FN679062) 55 VidyA-1 (FN679063) 65 VidyA-30 (FN679064) Rhodocyclaceae 72 VidyA-50 (FN679065) 76 AF170356 Dechloromonas sp. 99 VidyA-38 (FN679066) VidyA-8 (FN679067) VidyA-64 (FN679069) 99 VidyB-4 (FN679068) 81 86 VidyA-35 (FN679070) EF667730 Dechloromonas sp. VidyA-27 (FN679071) FJ517018 Propionivibrio sp. 96 87 OuchyA-36 (FN679072) 99 VidyA-73 (FN679073) 77 90 VidyA-78 (FN679074) VidyA-21 (FN679075) OuchyA-11 (FN679076) Gallionellaceae 99 EU266836 u. Gallionellaceae 99 VidyA-80 ((FN679077) 57 AY863077 Methylophilus sp. 77 99 VidyB-38 (FN679078) Methylophilaceae VidyA-12 (FN679079) 99 94 DQ857216 Methylophilus sp. EF580978 u. Methylophilaceae 90 99 OuchyA-12 (FN679080) 87 OuchyA-13 (FN679081) VidyA-59 (FN679082) 99 OuchyA-77 (FN679083) Thiobacillus sp. Hydrogenophilaceae 99 EF562571 Thiobacillus sp. VidyA-5 (FN679084) VidyA-24 (FN679087) 89 VidyA-28 (FN679086) 99 VidyA-29 (FN679085) EF667529 u. beta-proteobacterium 88 OuchyA-15 (FN679088) DQ676466 u. beta-proteobacterium 99 OuchyA-44 (FN679089) 99 OuchyA-63 (FN679090) Unclassified

99 OuchyA-76 (FN679092) 98 OuchyA-79 (FN679093) 99 EU644256 u. beta-proteobacterium OuchyA-4 (FN679094) OuchyA-5 (FN679091) 99 EU299971 u. beta-proteobacterium OuchyA-57 (FN679256) 99 EU373128 u. beta-proteobacterium 99 OuchyA-29 (FN679095) 59 AF089858 Aquabacterium sp. Burkholderiales OuchyA-34 (FN679096) OuchyA-25 (FN679097)

99 VidyA-72 (FN679098) 71 X95839 u. beta-proteobacterium 99 VidyA-39 (FN679099) 78 VidyA-76 (FN679100) 69 89 CU918612 u. Comamonadaceae VidyA-60 (FN679101) Comamonadaceae OuchyA-40 (FN679102) 88 66 OuchyA-26 (FN679103) 89 AY788965 Rhodoferax sp.

99 VidyA-36 (FN679104) 51 Rhodoferax sp. VidyA-57 (FN679105) VidyB-42 (FN679106) OuchyA-14 (FN679107)

0.02 Figure 3.7. Neighbour-joining phylogenetic tree of Betaproteobacteria 16S rRNA gene sequences retrieved from Vidy Bay and Ouchy sediments. “A” in clone names stands for the 0-2 cm sediment section and “B” for the 4-6 cm sediment section. Bootstrap values > 50% are shown (500 replicates).

Compared to Betaproteobacteria , the Deltaproteobacteria subdivision (Fig. 3.8) was less abundant (15% and 6% of the clones in Vidy and Ouchy sediments, respectively). Most Deltaproteobacteria are sulphate-, iron- or proton-reducing

(synthrophic) bacteria that play a major role in anoxic settings like meromictic lakes and sediments (Lehours et al, 2007; Karr et al., 2005). The large number of clones related to sulphate-reducing bacteria was not unexpected since a clear consumption of sulphate in the sediment is evident from the porewater data. Furthermore the pool of total reduced sulphur (TRS), including H 2S, was quite high, especially in Vidy sediments. Indeed, several clones related to the Desulfobacteraceae were found in both sites. This family includes various species of sulphate- and sulphur-reducing bacteria that thrive with organic compounds or H 2. Several additional clones, only found in

Vidy sediments, were related to Geobacter sp . ( Geobacteraceae ). These anaerobic organisms have been isolated from freshwater sediments, soils as well as subsurface environments. They are traditionally considered as chemoorganotrophic Fe(III)- reducing bacteria. However, most species utilise a wide range of alternative electron acceptors, including NO 3-, S 0, and other sulfur compounds (Coates et al., 1998).

Similar sequences were retrieved from anthropogenically disturbed urban creek sediments (EU284415) and from aquifers where Fe(III) reduction was associated with aromatic hydrocarbon degradation (AY653549). Another group of clones exclusively found in Vidy sediments belonged to the Synthrophaceae , with one clone related to

Smithella sp . Similar sequences were identified as core microorganisms involved in the anaerobic digestion of sludge (Riviere et al., 2009. CU922073). A few clones remained unclassified and were related to sequences found in low salinity tidal sediments (GQ243164) and agricultural soils (DQ830086).

69 OuchyB-37 (FN679117) 99 GU472640 u. delta-proteobacterium Unclassified 98 EF667759 u. delta-proteobacterium OuchyB-34 (FN679118) 87 100 GQ243164 u. delta-proteobacterium OuchyB-26 (FN679116)

100 AM409935 Desulfobacterium sp. AB239048 Desulfobacula sp.

OuchyB-39 (FN679113) 56 100 EF667823 u. Desulfobacteraceae GQ423356 Desulfococcus sp. VidyA-7 (FN679112) OuchyB-9 (FN679111) Desulfobacteraceae

VidyA-63 (FN679110) 100 51 VidyA-61 (FN679109) 72 VidyA-23 (FN679108) 63 EF520539 u. Desulfobacteraceae DQ415864 u. Desulfobacteraceae 100 VidyB-18 (FN679115) 64 FJ517128 u. Desulfobacteraceae 55 DQ830086 u. delta-proteobacterium Unclassified VidyB-8 (FN679114) 100 99 AJ518376 u. delta-proteobacterium

99 VidyA-40 (FN679119) 59 VidyA-46 (FN679120)

80 GQ183319 Smithella sp. Syntrophobacteraceae VidyA-71 (FN679121) 64 100 CU922073 u. Syntrophaceae VidyA-26 (FN679284) Unclassified 100 GQ354986 u. delta-proteobacterium AY653549 Geobacter psychrophilus

EU284415 Geobacter sp. 100 VidyA-17 (FN679122) 91 VidyA-11 (FN679125) Geobacteraceae 53 VidyB-29 (FN679127) Geobacter sp. VidyA-22 (FN679123) VidyA-52 (FN679126) VidyB-23 (FN679124) OuchyA-56 (FN679128) Bacteriovoraceae 100 DQ984561 Bacteriovorax sp.

0.02

Figure 3.8. Neighbour-joining phylogenetic tree of Deltaproteobacteria 16S rRNA gene sequences from Vidy Bay and Ouchy sediments. “A” in clone names stands for the 0-

2 cm sediment section and “B” for the 4-6 cm sediment section. Bootstrap values >

50% are shown (500 replicates).

The Gammaproteobacteria were more abundant in Ouchy sediments than in

Vidy and were phylogenetically diverse (Fig. 3.9). Clones from both sites were affiliated to Methylobacter sp (Methylococcaceae ), which are methanotroph organisms, characterized by their specialized metabolism restricted to the oxidation of methane or methanol. Similar sequences were retrieved from a permafrost soil in Siberia

(EU124843) and an Arctic wetland soil (AJ414655). Surprisingly, a few clones from

Vidy sediments were related to the phototrophic purple sulfur bacteria

(Chromatiaceae ) despite the fact that there is not much light to be expected at 30 m depth. However, some species can also grow under chemotrophic conditions in the dark, either autotrophically or heterotrophically using oxygen as terminal electron acceptor. Chromatiaceae are known to be indicator organisms for sewage waters and sequences similar to ours were retrieved from a community of microorganisms involved in anaerobic digestion (Riviere et al., 2009) and also from benthic production in Fayetteville Green Lake, USA (FJ437977). Several groups of

Gammaproteobacteria clones remained unclassified but with similar sequences found in bacterioplankton commnunities of Lake Michigan (sequence EU640647), river sediments (sequences EF111188 and EF590053), mangroves (sequence EF125457) and agricultural soils (FJ444695).

High number of sequences affiliated with the division Bacteroidetes (Cytophaga-

Flexibacter-Bacteroidetes ) were found in both sites, particularly in Vidy sediments.

Bacteroidetes constitute the second largest group in Vidy sediments after the

Betaproteobacteria , and the third largest group in Ouchy sediments (Fig. 3.5).

Bacteroidetes phylotypes were diverse and their closest relatives were sequences found in freshwater lakes (Mueller-Spitz et al., 2009), anthropogenically disturbed sediments, tundra soils rich in organic matter (Liebner et al., 2008) and aquifers. Most of the sequences found in the 2 investigated sites remained unclassified. One clone found in Vidy sediments was affiliated to the genus Cytophaga sp . and one clone from

Ouchy was related to the genus Flavobacterium sp.

OuchyA-2 (FN679129) 50 OuchyB-5 (FN679130) 53 EU640647 u. gamma-proteobacterium

100 OuchyB-7 (FN679131) OuchyB-21 (FN679132) 64 67 OuchyB-46 (FN679133) Unclassified FM200957 u. gamma-proteobacterium

99 OuchyB-8 (FN679136) 92 EF111188 u. gamma-proteobacterium

64 OuchyA-31 (FN679134) OuchyB-38 (FN679135) 98 64 EU244006 u. gamma-proteobacterium 54 OuchyA-69 (FN679137) Shewanellaceae 100 EF523608 Shewanella sp.

85 OuchyA-7 (FN679147) EF125457 u. gamma-proteobacterium 55 100 OuchyA-23 (FN679148) Unclassified OuchyB-13 (FN679149) 85 81 99 EU273118 u. gamma-proteobacterium OuchyA-18 (FN679138) OuchyA-78 (FN679139) VidyA-6 (FN679277) 94 OuchyA-19 (FN679140)

74 89 OuchyA-22 (FN679141) 59 OuchyB-30 (FN679142) 80 AJ414655 Methylobacter tundripaludum Methylobacter sp. Methylococcaceae 90 EU124843 Methylobacter sp. OuchyA-16 (FN679143) 64 VidyA-13 (FN679146) 78 VidyA-55 (FN679145) 66 VidyA-19 (FN679144) 94 VidyA-15 (FN679278) VidyB-12 (FN679288)

98 VidyA-58 (FN679279) 100 Unclassified FJ623321 u. gamma-proteobacterium 100 VidyB-31 (FN679289) 100 FJ437975 u. gamma-proteobacterium VidyB-6 (FN679156) 97 VidyB-14 (FN679158) 55 VidyB-24 (FN679157) VidyB-5 (FN679155) Chromatiaceae 100 FJ437977 u. Chromatiaceae VidyA-32 (FN679154) VidyA-33 (FN679151) 76 100 CU918403 u. Chromatiaceae OuchyB-18 (FN679153) OuchyA-70 (FN679152) 100 99 EU546554 u. gamma-proteobacterium 70 100 VidyA-34 (FN679280) GQ183378 u. gamma-proteobacterium

100 OuchyA-53 (FN679259) DQ833482 u. gamma-proteobacterium Unclassified

100 OuchyA-27 (FN679257) 59 EU801093 u. gamma-proteobacterium 100 OuchyA-10 (FN679150) 87 EU546356 u. gamma-proteobacterium 85 OuchyA-43 (FN679258) VidyB-1 (FN679287) 100 100 EU273120 u. gamma-proteobacterium

0.02

Figure 3.9. Neighbour-joining phylogenetic tree of Gammaproteobacteria 16S rRNA gene sequences from Vidy Bay and Ouchy sediments. “A” in clone names stands for the 0-2 cm sediment section and “B” for the 4-6 cm sediment section. Bootstrap values > 50% are shown (500 replicates).

For the Archaeal community (Fig. 3.10), a large proportion of Euryarchaeota phylotypes mostly originating from Vidy sediments, were related to methanogens like Methanosaeta sp. (Methanosaetaceae ) and the Methanomicrobiales group. A few species of Methanosaeta sp . have already been isolated from anaerobic sewage digestors or sewage sludge (Zinder et al., 1984; Huser et al., 1982,). In the presence of

CO 2, these organisms are able to grow and produce significant amount of methane from acetic acid. Similar sequences to the clones found in this study, were retrieved from anaerobic sludge and anoxic freshwater sediments from meromictic lakes

(Lehours et al., 2007). The rest of the archaeal sequences were only distantly related to any cultured species but similar to sequences retrieved from lake sediments

(sequences AY531743 and EU782007), Arctic peat (AM712495) and the anoxic zone from a hydropower plant reservoir in the Brazilian Amazon (GU127420 and

GU127500).

The two investigated sites differ clearly in terms of sediment chemical parameters and degree of pollution. It was therefore expected that the bacterial community composition would be different and reflect the differences in environmental conditions. The Fig. 3.5 and Fig.3.7-3.10 show clearly the diverse bacterial and archaeal lineages detected in the two sites, as confirmed by the results of the genetic diversity tests. Some sequences were common and others different; e.g., the Acidobacteria, Chloroflexi and Cyanobacteria species represented a small fraction of the microbial communities in the two sites whereas Nitropsira,

Planctomycetes, Verrucomicrobia and Gemmatimonadetes were the divisions found exclusively in Ouchy sediments. Some bacteria were also found only in Vidy sediments, like Dechloromonas sp . and are dominating in the contaminated area. The bacterial community composition changed with depth in the uncontaminated sediment. Conversely, no statistically significant variations were observed for the two sediment layers (0-2 and 4-6 cm) in the Bay of Vidy, which is explained by the

91 high sedimentation rates and the non-consolidated nature of the sediment, permitting mobilisation and vertical mixing.

The microbial composition of both sites was correlated with the environmental variables, as shown by the Mantel correlation test (r=0.9429, p=0.044). This result suggests that the diversity of microbial communities may be affected by nutrients, organic matter contents as well as the degree of pollution. Many environmental variables are implicated, which is a situation inherent to all field studies. To focus on only one environmental factor at a time, microcosm studies are required.

An integrative picture of the relationship between bacterial community structures and environmental factors at the 2 sites was obtained through MFA (Fig.

11). This indicated that (i) the sampling sites Ouchy and Vidy were clearly divergent considering both microbial communities and environmental factors; (ii) the two sampling depths in Vidy were statistically similar in terms of microbial communities,

OM and nutrients and heavy metals; (iii) the similarity between all datasets in Ouchy is much lower than in Vidy . Previous results have also demonstrated that bacterial diversities can differ significantly between contaminated and uncontaminated environments. The difference may be explained by several environmental parameters, including the nature of pollution and a wide diversity of organic carbon sources (Sandaa et al., 1999; Sorci et al., 1999; Zhang et al., 2008). The polluted environment of Vidy Bay may have selected among the dispersed microbes in sediments, certain functional bacterial groups which adapted to these conditions and became more dominant in that particular environment.

OuchyB-13 (FN679181) 62 GU257014 u. Methanomicrobiales VidyB-14 (FN679180) 71 OuchyB-12 (FN679179) 87 VidyB-16 (FN679182) 90 FM165672 u. Methanomicrobiales VidyB-6 (FN679177) OuchyB-9 (FN679178) FM165676 u. Methanomicrobiales 97 51 Methanomicrobiales 81 AB479390 isolated Methanomicrobiales VidyB-2 (FN679176) OuchyB-20 (FN679174) 82 70 OuchyB-17 (FN679173) 87 DQ785300 u. Methanomicrobiales 100 VidyB-4 (FN679175) CU917179 u. Methanomicrobiales VidyA-1 (FN679172) 100 85 OuchyB-6 (FN679171) 76 X51423-1 Methanosaeta concilii VidyB-12 (FN679170) VidyB-11 (FN679169) 100 VidyB-15 (FN679168) 80 59 VidyB-1 (FN679159) 88 VidyB-5 (FN679160) Methanosaetaceae AM181940 u. Methanosaetaceae Methanosaeta sp. VidyB-7 (FN679161) VidyB-9 (FN679167) VidyB-13 (FN679166) VidyB-17 (FN679162) 55 VidyB-3 (FN679165) VidyB-8 (FN679164) VidyB-10 (FN679163) 68 CU916809 u. Methanosaetaceae OuchyB-5 (FN679183) 100 EU782007 u. Euryarchaeota OuchyA-1 (FN679189)

67 100 OuchyB-7 (FN679186) 90 OuchyB-11 (FN679187)

83 AJ867619 u. Euryarchaeota 80 OuchyB-10 (FN679188) 100 GU127420 u. Euryarchaeota Unclassified 100 OuchyB-2 (FN679184) 62 GU127500 u. Euryarchaeota Euryarchaeota OuchyB-19 (FN679185)

100 OuchyB-8 (FN679190) AB182794 u. archaeon 100 AY531743 u. Euryarchaeota OuchyB-1 (FN679191) 100 99 OuchyB-3 (FN679192) OuchyB-16 (FN679193)

100 OuchyB-18 (FN679194) 98 AB243808 u. Euryarchaeota

0.05

Figure 3.10. Neighbour-joining phylogenetic tree showing 16S rRNA gene sequences of Archaea retrieved from Vidy Bay and Ouchy sediments. “A” in clone names stands for the 0-2 cm sediment section and “B” for the 4-6 cm sediment section. Bootstrap values > 50% are shown (500 replicates).

3 Microbial composition OM, nutrients Heavy-metals

OuchyA VidyB VidyA 0 1 2

Variation(23.07 %) OuchyB

-4 -3 -2 -1

-4 -2 0 2 Variation (61.4 %)

Fig. 3.11. The Multiple factor analysis (MFA) is a PCA-based technique allowing the simultaneous ordination of a composite table obtained by the juxtaposition of the species and the two environmental data sets, after weighting the different matrices

The superimposed representation shows one global point for each site, Vidy and

Ouchy, at each depth (“A” stands for the 0-2 cm sediment section and “B” for the 4-6 cm sediment section). The three associated partial points correspond to the three data sets (microbial composition, organic matter and nutrients and heavy-metals). The values on the axes indicate the percentage of total variation.

3.5 Conclusion This is the first study reporting on the microbial community structures of

Bacteria and Archaea in contaminated and uncontaminated sediments of Lake

Geneva. Results show that the sediments of the two study sites differed clearly in their organic matter and nutrient contents. Intense mineralisation of organic matter under sulphate-reducing and methanogenic conditions was indicated for the sediments from Vidy Bay. Furthermore, results confirm data of previous studies showing that the area around the WWTP outlet pipe in the Vidy Bay is heavily contaminated with various organic and inorganic pollutants. Phylogenetic analysis of sedimentary procaryotes revealed that (i) archaeal and bacterial communities differed significantly between the contaminated and the non-contaminated sediments. (ii) A correlation was observed between the microbial composition and the environmental variables along the 2 sites, which suggests that diversity of microbial communities may be affected by nutrients, organic matter as well as the degree of pollution. (iii) Betaproteobacteria was the dominant bacterial group, representing more than 30% of analysed clones in surface sediments at both sites. (iv)

A large proportion of Betaproteobacteria clones, mostly from Vidy sediments, were related to the reductively dechlorinating Dechloromonas sp . (iv) Consistent with geochemical data Deltaproteobacteria , including clones related to iron- (Geobacter sp .) and sulphate-reducing bacteria, were more abundant in the contaminated sediments.

(v) The archaeal communities were dominated by methanogenic Euryarchaeota , particularly in the organic matter-rich sediments from Vidy Bay.

95 microbial community composition and specific activities (such as remobilization of contaminants from sediments) of microorganisms inhabiting similar environments should be performed, in order to improve the understanding how pollution and eutrophication may affect microbial communities.

References

Amann RI, Binder BJ, Olsen RJ, Chisholm SW, Devereux R & Stahl DA (1990)

Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry

for analyzing mixed microbial populations. Appl Environ Microbiol 56 :1919-1925.

APHA (1985) Standard methods for the examination of water and wastewater.

American Public Health Association , Washington, 408-410.

Becue-Bertaut M., Pages J., 2008. Multiple factor analysis and clustering of a mixture

of quantitative, categorical and frequency data. Computational Statistics &

Data Analysis 52: 3255-3268.

Brosius J, Dull TJ, Sleeter DD & Noller HF (1981) Gene organization and primary

structure of ribosomal RNA operon from Escherichia coli . J Mol Biol 148 :107-127.

Burrus D, Thomas RL, Dominik B, Vernet JP & Dominik J (1990) Characteristics of

suspended sediment in the Upper Rhone River, Switzerland, including the

particulate forms of phosphorus. Hydrol Process 4: 85-98.

Canfield DE & Thamdrup B (2009) Towards a consistent classification scheme for

geochemical environments, or, why we wish the term “suboxic” would go

away. Geobiol 7: 385-392.

Casamayor EO, Schäfer H, Bañeras L, Pedrós-Alió C & Muyzer G (2000)

Identification of and spatio-temporal differences between microbial

assemblages from two neighborning sulfurous lakes: Comparison by

microscopy and denaturing gradient gel electrophoresis. Appl Environ Microbiol

66 : 499-508

Cline JD (1969) Spectrophotometric determination of hydrogen sulfide in natural

waters. Limnol Oceanogr 14 : 454-458.

Coates JD, Ellis DJ, Blunt-Harris EL, Roden E, Gaw C & Lovley DR (1998) Recovery

and identification of humic-reducing bacteria from a diversity of environments.

Appl Environ Microbiol 64 :1504-1509

Conseil canadien des ministres de l’environnement (1999) Introduction, dans

Recommandations canadiennes pour la qualité de l’environnement. Winnipeg, le

Conseil.

Corvi C, Majeux C & Vogel J (1986) Les polychlorobiphényles et le DDE dans les

sédiments superficiels du Léman et de ses affluents. Rapp. Comm. Int. prot. eaux

Léman contre pollut. Campagne 1985 , 205-216.

Dorioz JM, Pelletier JP & Benoit P (1998) Physico-chemical properties and

bioavailability of particulate phosphorus of various origins in a watershed of

Lake Geneva (France) . Wat Res 32 : 275-286.

Escofier B., Pages J., 1994. Multiple Factor-Analysis (Afmult Package). Computational

Statistics & Data Analysis 18: 121-140.

Finneran KT, Johnsen CV & Lovley DR (2003) Rhodoferax ferrireducens sp . nov., a

psychrotolerant, facultatively anaerobic bacterium that oxidizes acetate with the

reduction of Fe(III). Int J Syst Evol Micr 53 : 669-673.

Fossing H & Jørgensen BB (1989) Measurement of bacterial sulphate reduction in

sediments: evaluation of a single-step chromium reduction method.

Biogeochemistry 8: 205-222.

Gillan CD, Danis B, Pernet P, Joly G & Dubois P (2005) Structure of sediment-

associated microbial communities along a heavy-metal contamination gradient

in the marine environment. Appl Environ Microbiol 71 : 679-690.

Glöckner FO, Zaichikov E, Belkova N, Denissova L, Pernthaler J, Pernthaler A &

Amann R (2000) Comparative 16S rRNA analysis of lake bacterioplankton

reveals globally distributed phylogenetic clusters including an abundant group

of Actinobacteria . Appl Environ Microbiol 66: 5053-5065.

Groβkopf R, Janssen PH & Liesack W (1998) Diversity and structure of the

methanogenic community in anoxic rice paddy soil microcosms as examined by

cultivation and direct 16S rRNA gene sequence retrieval. Appl Environ Microbiol

64 : 960-969.

Hall GEM & Pelchat P (1997) Evaluation of a Direct Solid Sampling Atomic

Absorption Spectrometer for the Trace Determination of Mercury in Geological

Samples. The Analyst 122 : 921-924.

Haller L, Poté J, Loizeau J-L &Wildi W (2009) Distribution and survival of faecal

indicator bacteria in the sediments of the Bay of Vidy, Lake Geneva,

Switzerland . Ecol Indicators 9: 540-547.

Hahn MW (2006) The microbial diversity of inland waters. Curr Opin Biotech 17 : 256-

261.

Hiorns WD, Methé BA, Nierzwicki-Bauer SA & Zehr JP (1997) Bacterial diversity in

Adirondack mountain lakes as revealed by 16S rRNA Gene sequences. Appl

Environ Microbiol 63 : 2957-2960.

Huser BA, Wuhrmann K & Zehnder AJB (1982) Methanothrix soehngenii gen. nov. sp.

nov., a new acetotrophic non-hydrogen-oxidizing methane bacterium. Arch

Microbiol 132 : 1-9.

Jenkins O, Byrom D & Jones D (1987) Methylophilus : a new genus of methanol-

utilizing bacteria. Int J Syst Bacteriol 37 : 446-448.

Karr EA, Sattley WM, Rice MR, Jung DO, Madigan MT & Achenbach LA (2005)

Diversity and distribution of sulphate-reducing bacteria in permanently frozen

Lake Fryxell, McMurdo Dry Valleys, Antarctica. Appl Environ Microbiol 71 : 6353-

6359.

Kumar S, Tamura K & Nei M (2004) MEGA3: integrated software for Molecular

Evolutionary Genetics Analysis and sequence alignment. Brief Bioinform 5: 150-

163.

Lane DJ (1991) 16S/23S rRNA sequencing. In: Stackebrandt E and Goodfellow M

(ed.), Nucleic Acid Techniques in Bacterial Systematics. John Wiley & Sons, Inc.,

New York, pp 115-175.

Lehours AC, Evans P, Bardot C, Joblin K & Gérard F (2007) Phylogenetic diversity of

archaea and bacteria in the anoxic zone of a meromictic lake (Lake Pavin,

France). Appl Environ Microbiol 73 : 2016-2019.

Liebner S, Harder J & Wagner D (2008) Bacterial diversity and community structure

in polygonal tundra soils from Samoylov Island, Lena Delta, Siberia. Int

Microbiol 11 : 195-202.

Loizeau J.-L, Arbouillle D, Santiago S & Vernet J.-P (1994). Evaluation of a wide

range laser diffraction grain size analyser for use with sediments. Sedimentology

41 : 353-361.

Loizeau J.-L, Pardos M, Monna F, Peytremann C, Haller L & Dominik J (2004) The

impact of a sewage treatment plant’s effluent on sediment quality in a small bay

in Lake Geneva (Switzerland-France). Part 2 : Temporal evolution of heavy

metals. Lakes & Reservoirs: Res Manag 9: 53-63.

Lors C, Tiffreau C & Laboudigue A (2004) Effects of bacterial activities on the release

of heavy metals from contaminated dredged sediments. Chemosphere 56 : 619-

630.

Mantel, N. 1967. The detection of disease clustering and a generalized regression

approach. Cancer Research, 27 , 209-220.

Martin, A.P., 2002. Phylogenetic approaches for describing and comparing the

diversity of microbial communities. Applied and Environmental Microbiology

68: 3673-3682.

Méthodes de référence des Stations Fédérales de Recherches Agronomiques

AGROSCOP, 1998. Code C. org. - FAL Zürich-Reckenholz, Switzerland

Mueller-Spitz SR, Goetz GW & McLellan SL (2009) Temporal and spatial variability

in nearshore bacterioplankton communities of Lake Michigan. FEMS Microbial

Ecol 67 : 511-522.

Muyzer G (1993) Profiling of complex microbial populations by denaturing gradient

gel electrophoresis analysis of polymerase chain reaction-amplified genes

coding for 16S rRNA. Appl Environ Microbiol 59 :695-700.

Nealson, KH (1997). Sediment bacteria: who’s there, what are they doing, and what’s

new? Annu Rev Earth Pl Sc 25 : 403-434.

Poté J, Haller L, Loizeau J- L, Garcia Bravo A, Sastre V & Wildi W (2008) Effects of a

sewage treatment plant outlet pipe extension on the distribution of

contaminants in the sediments of the Bay of Vidy, Lake Geneva, Switzerland.

Bioresource Technol 99 : 7122-7131.

Powell SM, Bowman JP, Snape I & Stark JS (2003) Microbial community variation in

pristine and polluted nearshore Antartic sediments. FEMS Microbial Ecol 45 :

135-145.

R Development Core Team, 2005. R: A language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-

900051-07-0, http://www.R-project.org

Riviere D, Desvignes V, Pelletier E, Chaussonnerie S, Guermazi S, Weissenbach J, Li

T, Camacho P & Sghir A (2009) Towards the definition of a core of

microorganisms involved in anaerobic digestion of sludge. ISME J 3: 700-714.

Ross-Barraclough F, Givelet N, Martinez-Cortizas A, Goodsite ME, Biester H &

Shotyk W (2002) An analytical protocol for the determination of total mercury

concentrations in solid peat samples. Sci. Total Environ. 292 : 129-139.

Sandaa R-A, Torsvik V, Enger O, Daae FL, Castberg T & Hahn D (1999) Analysis of

bacterial communities in heavy metal-contaminated soils at different levels of

resolution. FEMS Microbial Ecol 30 : 237-251.

Schloss, P.D., Westcott S.L., Ryabin T., Hall J.R., Hartmann M., Hollister E.B.,

Lesniewski R.A., Oakley B.B., Parks D.H., Robinson C.J., Sahl J.W., Stres B.,

100

Thallinger G.G., Van Horn D.J., Weber C.F., 2009. Introducing mothur: Open-

source, platform-independent, community-supported software for describing

and comparing microbial communities. Applied and Environmental

Microbiology 75: 7537-41.

Schneider S, Roessli D & Excoffier L (2000) Arlequin ver 2.0: A software for

population genetic data analysis. Genetics and Biometry Lab, University of

Geneva, Geneva.

Sorci JJ, Paulauskis JD, & Ford TE (1999) 16 rRNA Restriction Fragment Length

Polymorphism Analysis of Bacterial Diversity as a Biomarker of Ecological

Health in Polluted Sediments from New Bedford Harbor, Massachusetts, USA .

Mar. Poll. Bull . 38 : 663-675.

Stahl DA & Amann RI (1991) Development and application of nucleic acid probes, p.

205-248. In: Stackebrandt E and. Goodfellow M (ed.), Nucleic Acid Techniques

in Bacterial Systematics. John Wiley & Sons, Inc., New York.

Tanaka K, Nakamura K & Mikami E (2003) Fermentation of maleate by a gram-

negative strictly anaerobic non-spore-former, Propionivibrio dicarboxylicus

gen.nov., sp.nov. Arch Microbiol 154: 323–328.

Thamdrup B, Fossing H & Jørgensen BB (1994) Manganese, iron and sulfur cycling in

a coastal marine sediment, Aarthus Bay, Denmark. Geochim Cosmochim Acta 58 :

5115-5129.

Williams JDH, Jaquet JM & Thomas RL (1976) Forms of phosphorus in the superficial

sediments of Lake Erie. J Fish Res Board Canada 33 : 413-429.

Wolterink A, Kim S, Muusse M, Kim IS, Roholl PJM, van Ginkel CG, Stams AJM &

Kengen SWM (2005) Dechloromonas hortensis sp. nov. and strain ASK-1, two

novel (per)chlorate-reducing bacteria, and taxonomic description of strain GR-1.

Int J Syst Evol Microbiol 55 : 2063-2068.

Ye W, Liu X, Lin S, Tan J, Pan J, Li D & Yang H (2009) The vertical distribution of

bacterial and archaeal communities in the water and sediment of Lake Taihu.

FEMS Microbiol Ecol 70 : 263-276.

101

Zhang W, Ki J-S & Qian P-Y (2008) Microbial diversity in polluted harbor sediments

I : Bacterial community assessment based on four clone libraries of 16S rDNA.

Estuar Coast Shelf S 76 : 668-681.

Zinder SH, Cardwell SC, Anguish T, Lee M & Koch M (1984) Methanogenesis in a

thermophilic (58°C) anaerobic digestor: Methanothrix sp. as an important

aceticlastic methanogen. Appl Environ Microbiol 47 : 796-807.

Zopﬁ J, Böttcher ME & Jørgensen BB (2008) Biogeochemistry of sulfur and iron in

Thioploca -colonized surface sediments in the upwelling area off central Chile.

Geochim Cosmochim Acta 72 : 827–843.

Zwart G, Crump BC, Kamst-van Agterveld MP, Hagen F & Han S-K (2002) Typical

freshwater bacteria: an analysis of available 16S rRNA gene sequences from

plankton of lakes and rivers. Aq Microb Ecol 28 : 141-155.

102

CHAPTER 4

Origin and spatial-temporal distribution of faecal

bacteria contamination in a Bay of Lake Geneva,

Switzerland

A similar version of this chapter was published under the following reference:

Poté J, Goldscheider N, Haller L, Zopfi, Khajehnouri F, Wildi W (2009). Origin and spatial-temporal distribution of fecal bacteria in a bay of Lake Geneva, Switzerland.

Environmental Monitoring and Assessment 154 : 337-348.

103

Abstract The origin and distribution of microbial contamination in Lake Geneva’s most polluted bay were assessed using faecal indicator bacteria (FIB) and physicochemical profiling. During one year, water samples were taken at 23 points in the bay and 3 contamination sources: a wastewater treatment plant (WWTP), a river and a storm water outlet. Analyses included Escherichia coli , enterococci (ENT), total coliforms

(TC), and heterotrophic plate counts (HPC). E. coli input flux rates from the WWTP can reach 2.5 x 10 10 CFU/s. The input from the river is lower, 1 to 3 orders of magnitude lower, but still significant. Different pathogenic Salmonella serotypes were identified in selected samples from these sources. FIB levels in the bay are spatially and temporally variable. The results demonstrate that (1) the WWTP outlet at 35 m depth impacts near-surface lake water quality mainly during holomixis in winter; (2) when the lake is stratified, the effluent water is generally trapped below the thermocline; (3) during major floods, upwelling across the thermocline may occur; (4) the river permanently contributes to contamination, mainly near the river mouth and during floods, when the storm water outlet contributes additionally; (5) the lowest

FIB levels, in the near-surface water, occur during low-flow periods in the bathing season.

104

4.1 Introduction Diarrhoeal diseases, mainly due to microbial water contamination, cause about 1 billion illnesses and 2.2 million deaths per year worldwide (Montgomery and

Elimelech 2007). Although most of these infections occur in developing countries, waterborne diseases are a worldwide problem. The most important transmission pathway is the faecal-oral route, i.e. insufficiently or untreated wastewaters that contaminate drinking water. Diseases can also be transmitted during recreational activities, such as bathing in polluted fresh or seawater (Bonadonna et al. 2002a), or by the consumption of contaminated shellfish (Campos and Cachola 2007).

Faecal indicator bacteria (FIB) are commonly used to assess the hygienic quality of drinking and recreational waters. The presence of FIB indicates the possible presence of pathogens of faecal origin, such as S almonella , while their absence suggests that pathogens are also absent (OECD, WHO 2003). Commonly used FIB include total coliforms (TC), faecal coliforms (FC), Escherichia coli , enterococci (ENT). E. coli and ENT have the highest sanitary significance, while TC and even ‘faecal’ coliforms can partly originate from non-faecal sources and thus overestimate pollution (Cabral and Marques 2006; Doyle and Erickson 2006).

Heterotrophic plate count (HPC) is often used as an additional, general water quality indicator. Swiss and most other national legislations and international guidelines for drinking water quality demand that E. coli and ENT should be absent in a 100 mL sample (WHO 2004). Limits for bathing water are higher; according to the European

Directive 2006/7/CE concerning the management of bathing water quality, recreational waters are to be classified as poor if E. coli levels exceed 900 CFU/100mL and concentrations of ENT exceed 330 CFU/100mL, based upon a 90-percentile evaluation.

105

Although water quality monitoring and treatment reduce the risk of waterborne diseases, they do not provide absolute safety: specific pathogens, such as

Cryptosporidium oocysts, may occur when FIB are absent, and/or they may resist water treatment (Hoxie et al. 1997; Bonadonna et al. 2002). Therefore, any release of insufficiently treated wastewaters into drinking and recreational water resources is always problematic and should be avoided or limited as much as possible. Existing microbial contamination requires careful monitoring and research in order to further minimize the risk for human health.

Lake Geneva is Western Europe’s largest freshwater reservoir. Lausanne, with

127,000 inhabitants, receives 58 % of its freshwater from the lake. At the same time, the city and region generate large volumes of wastewater, which are mainly released into the nearby Bay of Vidy. Previous studies focused on heavy metals in the sediments and revealed that the bay is the most contaminated part of part of the lake

(Pardos et al. 2004; Loizeau et al. 2004). Water analyses showed that the bacteriological water quality is often poor. Therefore, the health risk at this recreational site has received attention from the municipality of Lausanne, which recommended the monitoring of the bacterial contamination in the bay during one year.

The main objectives of the study were to quantify the input flux rates of faecal bacteria from the main contamination sources, and to assess the spatial and temporal distribution of bacteria in the bay in order to estimate the human health risk related to recreational activities and drinking water use. Furthermore, we aimed at characterizing the influence of different hydrometeorological and limnological conditions on the bacteriological contamination, including flood events, thermal state of the lake, and other seasonal influences. Sampling and monitoring started in March

2005 and ended in February 2006. The study consisted of FIB analyses ( E. coli , ENT,

106

TC), HPC, the detection of certain pathogenic Salmonella serotypes, and physicochemical measurements.

In parallel to this study, we carried out two multi-tracing experiments with three types of bacteriophages in order to simulate pathogen transport in the bay. The phages were injected at the three sources of contamination. The experiment was carried out twice, first when the lake was stratified and a second time during holomixis. The tracer results demonstrated that contaminated river water spreads rapidly in the bay, and that a well-developed thermocline prevents contamination from the depth to rise up to the surface, while rapid vertical transport occurs under mixed-lake conditions (Goldscheider et al. 2007).

4.2 Materials and methods

4.2.1 Study site The Bay of Vidy is located near the centre of Lausanne, on the northern shore of Lake Geneva (Figure 4.1). There are two major point sources of water contamination: the wastewater treatment plant (WWTP) and the Chamberonne

River. The WWTP treats ~1–3 m 3/s of urban wastewater, which is released into the bay via an underwater pipe at 35 m depth and 700 m distance from the shore. The river includes water from its natural catchment and untreated wastewater. The urban storm water drainage system (Flon) represents a third, transient contamination source. It collects surface and wastewater from the western part of the city, which is usually treated at the WWTP but released into the lake during flood events, via a conduit at 10 m depth. The drinking-water pumping station of St. Sulpice is located at 3.8 km distance from the outlet pipe of the WWTP. The water is pumped from 45 m depth at an average rate of 385 L/s.

107

Figure 4.1. Map of the study area and positions of the three known point sources of water contamination (WWTP, Chamberonne, Flon) and the sampling sites (Swiss coordinates, 1 km grid). The drinking water pumping station is located 3.8 km west of the WWTP outlet (V14).

4.2.2 Water sampling Figure 4.1 shows the location of the sampling sites (V1–V23), which are distributed within the entire bay but particularly focus on zones where contaminated waters enter the lake: V20a–d are situated around the mouth of the Chamberonne

River; V14 is above the outlet pipe of the WWTP, while V14a–d are ~150 m N, E, S and W of this point in order to monitor the direction of the contamination plume;

V17 is located above the storm water outlet pipe. Within the bay, water samples were taken at 2 m depth using a Niskin-type water sampler with a volume of 40 L. V13 and V14a–d were sampled at several depths (2–35 m) in order to assess the vertical distribution of FIB near the WWTP outlet pipe. At the contamination sources (V15,

V16, V17) and near the shore, where the water is less than 2 m deep (V20a, V20d), samples were collected manually. The samples were filled into sterile plastic bottles of 250 mL for FIB analyses and of 2 L for Salmonella determination, placed in a cooling-box at 4 ºC, and analysed within 24 h.

108

4.2.3 Physicochemical measurements Depth profiles of temperature (T), electrical conductivity (EC), pH, turbidity and dissolved oxygen were recorded at V14a–d, using a YSI 600 XL multiprobe. At the outflows of the three contamination sources, and at V17, V20a and V20d, EC and

T were measured with a conductimeter (WTW, LF 325). The flow rates of the three contamination sources were measured continuously, using calibrated weirs.

4.2.4 Microbiological analyses The microbiological analyses were performed in the laboratory “Eauservice” according to international standard methods for water quality determination (APHA,

2005). For TC, E. coli and ENT, 100 mL of water were passed through 0.45 μm membrane filters (47 mm diameter, Millipore), which were subsequently placed onto different culture media (Biolife): TC: Endo Agar, incubated at 35 °C for 24 h; E. coli :

Tryptic Soy Agar, incubated at 20 °C for 24 h and transferred to ECD Agar Mug for

24 h at 44 °C; ENT: Slanetz Bartley Agar, incubated at 37 °C for 48 h and transferred to Bile Aesculin Agar medium at 37 °C for 4 h. Heterotrophic plate counts (HPC) were determined on Plate Count Agar (PCA) after 72 h incubation at 30 °C.

The results are mostly single determinations and are expressed as colony forming units (CFU) per 100 mL. The reproducibility of the analytical procedure was tested by means of triplicates of selected water samples, which revealed a mean standard derivation of 17 %.

For Salmonella spp. detection, 1 L of water was passed trough a 0.45 membrane filter. The filter was placed into 100 mL Rappaport Vassiliadis broth (Biolife 401980) for 18–24 h at 44 ºC in order to enrich for Salmonella spp. A sample of the enrichment was streaked out on selective and differential Hektoen Enteric Agar (Biolife 401541), and incubated for 24h at 37 ºC. A rapid test for Salmonella spp. (Oxoid Biochemical

Identification System) was used to check putative Salmonella spp. colonies. Positive

109 isolates were transferred to Tryptic Soy Agar (ex. Biolife 402150) and sent for serotyping to the Swiss Centre for Salmonellae diagnostics in Lucerne.

4.3 Results and discussion

4.3.1 Physicochemical characterization of the lake The physicochemical depth profiles, particularly temperature, show that the water column was vertically mixed from the beginning of the measurement period in

March 2005 to May 2005 and from December 2005 until the end of the measurement period in February 2006, while it was stratified from June to November 2006. A typical profile for stratified lake conditions is presented further below, together with the FIB results.

The average EC of the surface water is 290 µS/cm when the lake is stratified and 240 µS/cm when it is mixed. Higher values were observed in the WWTP effluent

(average: 863 µS/cm) and the Chamberonne River (415 µS/cm). The pH of the lake water is within a narrow range of 7.5 to 7.8. The lake surface water temperature varied between 5.8 °C (February) and 25 °C (July). The temperatures of the WWTP effluent (10.3–27.4 °C) and Chamberonne River (4.2–23.0 °C) also display important seasonal variability.

4.3.2 Bacterial input from the contamination sources The concentrations of E. coli , ENT, TC and HPC in the WWTP effluent were substantially higher than those found in the Chamberonne River. For example, E. coli varied between 4.0 x 10 4 and 1.2 x 10 6 CFU/100 mL in the WWTP effluent water, and between 1.3 x 10 3 and 5.5 x 10 4 CFU/100 mL in the river. Enterococci levels were also higher in the effluent water (3.3 x 10 3 to 6.3 x 10 5 CFU/100 mL) than in the river (3.0 x

110

10 2 to 1.4 x 10 4 CFU/100 mL). The FIB levels found in the storm water drainage system (Flon) during floods were similar to those in the WWTP effluent water.

During the monitoring period, the discharge of the WWTP outlet varied between 1.0 and 5.5 m 3/s. As mentioned above, the WWTP can only treat up to 3 m 3/s of urban wastewater. The excess of wastewater during flood events is released into the bay without treatment. The measured flow rates of the Chamberonne River varied between 0.4 and 5.2 m 3/s.

For the characterization of contamination sources, mass fluxes are more relevant than concentrations. The FIB flux rates (CFU/s) from the WWTP and

Chamberonne River were calculated by multiplying the measured flow rates (L/s) and bacteria concentrations (CFU/L) (Figure 4.2). The input flux rates from the

WWTP are generally one to three orders of magnitude higher than those from the river but show a similar temporal evolution. A high flood event, for which complete data are available, occurred on the 4 July 2005, when 4 m 3/s were measured in the river and 2.1 m 3/s at the outflow of the WWTP. This flood event corresponds to a major contamination event, with the highest observed flux rates of E. coli (2.5 x 10 10

CFU/s) and ENT (1.3 x 10 10 CFU/s) from the WWTP, as well as the second highest E. coli (8.0 x 10 7 CFU/s) and highest ENT (2.1 x 10 8 CFU/s) flux rates from the river.

111

Figure 4.2. Input flux rates of E. coli and enterococci (ENT) from the WWTP and

Chamberonne River during five sampling campaigns, calculated by multiplying flow rates and concentrations. The flux rates from the WWTP are 1 to 3 orders of magnitude higher than those from the river. Dashed auxiliary lines do not suggest linear interpolation.

The observed average flux rates for E. coli from the WWTP (1.2 x 10 10 CFU/s) and the legal limit for drinking water (< 1 CFU/100 mL) mean that the effluent discharge into the lake has the potential to spoil 1.2 x 10 9 L of drinking water per second, i.e. 3 million times the pumping rate of the pumping station of St. Sulpice, 3.8 km further to the west. Although dispersion, dilution and inactivation will attenuate microbial contamination, this simple calculation illustrates the potential health risk associated to the discharge of effluent water.

Correlation analyses of the four FIB parameters revealed strong positive correlation between ENT and E. coli in the WWTP effluent (linear correlation coefficient r = 0.95), confirming the high sanitary significance of these two parameters

112

(Table 4.1). Total coliforms also correlate well with these two parameters, while HPC show a weak negative correlation with the three other parameters (-0.4 to -0.1). The river water displays a different pattern, with positive correlations between HPC and all other parameters, particularly TC (0.99). Cabral and Marques (2006) also found high correlation between TC, ENT, FC and faecal streptococci in river water.

Table 4.1. Linear correlation coefficients (r) of the four FIB parameters for a) WWTP effluent; b) Chamberonne River; c) lake at 2 m depth, > 300 m from the contamination sources; and d) a vertical profile recorded at V14c when the lake was stratified.

HPC HPC

4.3.3 Temporal variability of FIB concentrations in the bay In order to characterize the variability of FIB in the bay, the results from representative sampling points at 2 m depth were further evaluated. Only sites with complete data series and more than 300m away from the contamination sources were considered. The evaluation thus includes 8 sampling sites: V1–4, V7 and V9–11.

Figure 4.3 presents the results for E. coli during 10 individual sampling campaigns; similar results were found for ENT and TC.

113

Figure 4.3. Box-Whisker Diagrams illustrating the variability of E. coli at 8 representative sampling sites in the bay (> 300 m away from the contamination sources, at a depth of 2 m) during 10 sampling campaigns. The different lines in the diagrams represent the minimum, lower quartile, median, upper quartile and maximum values.

E. coli concentrations during a given sampling campaign are highly variable, e.g. 30–1900 CFU/100 mL on the 19 April, and 0–80 CFU/100 mL on the 4 July 2005.

This finding means that faecal contamination is heterogeneously distributed, and also illustrates that sampling only one site would be insufficient for microbial water quality monitoring.

The highest contaminations, up to 1900 CFU/100 mL, were observed during a flood event on the 19 April 2005, when the lake was mixed and maximum flow rates occurred in the WWTP outflow (4.9 m 3/s) and Chamberonne River (5.2 m 3/s) (input flux rates are not available for this date). This finding suggests upwelling of effluent

114 water from the WWTP outlet pipe and rapid propagation of river water in the bay, as also demonstrated by the tracer test done on 21rst Februrary 2006, during holomixis

(Goldscheider et al., 2007). In this study, phages (type H6/1) were injected near the outlet pipe; only 4h later, extremely high phage concentrations were found in water samples taken in the bay at 2m depth.

Increased FIB levels also occurred during or after several other smaller flood events. Several other studies (e.g. Krometis et al. 2007; Pronk et al. 2006), revealed as well that the highest FIB levels in freshwaters (streams, lakes and springs) often occur during floods, which can be explained by the mobilization of bacteria from soil and by the overflowing of septic tanks and WWTPs.

The lowest E. coli levels, 0–14 CFU/100 mL, were observed on the 27 July, 8

August and 6 September 2005, during low flow and stratified lake conditions, indicating that the effluent from the WWTP was trapped below the thermocline, while the relatively low FIB input from the river was highly diluted in the lake.

Similar conditions (stratified lake, low flow rates) but substantially higher FIB levels

(E. coli : 4–250 CFU/100 mL) occurred on the 7 November 2005. The bacteriophage tracer test done on this very day demonstrated the effectiveness of the thermocline:

6.97 x 10 13 H6/1 were injected near the outlet pipe, but not a single phage was detected in any of the 54 water samples taken at 2 m depth in the bay during a time period of 48 h, confirming that the thermocline prevented vertical transport from the hypolimnion to the epilimnion (Goldscheider et al. 2007). Therefore, the different FIB levels probably reflect the seasonal dependence of FIB survival times in surface lake water. In summer, higher water temperatures and higher biological activity (e.g. protozoan grazing) can cause a rapid decline of faecal bacteria (Brettar and Höfle

1992, Vaque et al. 1994). Brettar and Höfle (1992) also investigated the role of suspended sediment particles on the survival of E. coli and found that, on one hand, particles provide niches for bacterial survival; and on the other hand, the

115 sedimentation of these particles contributes to the removal of E. coli from the water column. More intense UV summary radiation further reduces the FIB levels. Hughes

(2003) assessed the influence of various seasonal environmental parameters on the survival of faecal coliforms around an Antarctic research station and found that solar radiation is the dominant factor. The penetration depth of UV radiation into lake water strongly depends on the concentration of dissolved organic carbon and turbidity, and often ranges between some decimetres and several meters (Laurion et al. 2000).

The near-surface lake water data were also used for a correlation analysis

(Table 4.1). As for the WWTP effluent, the highest correlation was found between E. coli and ENT (r = 0.80). On the other hand, the good correlations between HPC and the other parameters resemble the river water. This finding reflects the combined impact of the two point sources of water contamination on the near-surface lake water quality. Davis et al. (2005) carried out similar statistical analyses in lake water and also found the highest positive correlation between E. coli and ENT.

4.3.4 Vertical distribution of FIB in the lake water column Vertical profiles of temperature, EC, turbidity, FIB and HPC were recorded on the 8 August 2005 during stratified lake and low flow conditions at point V14c, 150 m south of the WWTP outlet pipe (Figure 4.4). The profile displays a thermocline with an inflexion point at 20 m depth. Turbidity and EC generally decrease with depth but show clear local maximums at 20 m, together with HPC, indicating accumulation of suspended and dissolved matter and increased microbiological activity along the thermocline. Turbidity probably originates from the nearby Chamberonne River and from sinking biomass and other particles from the surface lake water. However, it is more likely that most of the turbidity originates from the nearby Chamberonne River and from sinking biomass and other particles from the surface lake water.

116

Figure 4.4. Physicochemical and bacteriological profiles during stratified like conditions (8 Aug 2005, V14c): ENT, E. coli and TC show a maximum below the thermocline, which can be attributed to the WWTP outlet pipe at 35 m, while the

HPC maximum coincides with a local turbidity and EC maximum at the thermocline at 20 m depth. The mean relative standard deviation of the FIB data is 17 %.

The FIB profiles show low values for all parameters at 2 and 10 m water depth, particularly E. coli and ENT, at 2 and 10 m water depth, underscoring the combined influence of higher water temperatures, biological activity and in the upper few meters UV radiation on bacteria degradation. At greater depths, there are notable differences between HPC on one hand and E. coli , ENT and TC on the other hand:

117

HPC display a maximum at 20 m depth, i.e. along the thermocline, while the FIB parameters show low values along the thermocline but maximum values in the hypolimnion at 25 m depth. These finding suggests that the WWTP effluent is trapped below the thermocline, thus preventing water contamination near the surface of the lake. As described above, this interpretation was also confirmed by tracer tests (Goldscheider at al., 2007). The inverse depth distribution of FIB and HPC also demonstrates the low sanitary significance of HPC (Allen et al., 2004).

The correlation analysis of the four FIB parameters from this vertical profile

(Table 4.1) revealed a similar pattern as for the WWTP effluent, i.e. high positive correlation between E. coli , ENT and TC (0.86–0.95), but only weak negative or positive correlation between HPC and the three other parameters. The vertical distribution of FIB in the stratified water column thus mainly reflects the degree of contamination from the WWTP outlet pipe.

4.3.5 Horizontal distribution of FIB near the river mouth and in the bay Figure 4.5 shows the levels of E. coli and ENT at 2 m depth during selected sampling campaigns, during distinct hydrological and limnological situations: stratified lake and flood event (4 July 2005); stratified lake and low flow conditions (8

August 2005); mixed water column and flood event (19 April 2005); and mixed water column and low flow conditions (21 February 2006).

The highest FIB levels often occur near the mouth of the Chamberonne River, mainly during floods, but also, to a lesser degree, during low flow conditions. The observed maximum concentrations reach 26 000 and 14 000 CFU/100 mL for E. coli and ENT, respectively. Lower levels were only observed in September 2005, which can be explained by the fact that the river water is often colder than the lake surface water in autumn and thus sinks into deeper water layers. Further away from the

118 river mouth, the concentrations of E. coli and ENT rarely exceed the limits for bathing water but very often those for drinking water (see introduction). The lowest levels occur during low flow conditions in summer, e.g. on 8 August 2005.

During the flood of 4 July 2005, a contamination plume seems to extend from the river mouth (V20a–d) towards the centre of the bay (V14a–d), with E. coli and

ENT levels up to 17 000 and 7000 CFU/100 mL, respectively, while they were not detectable at more remote sampling sites. However, correlation analysis of the bacteriological data from V14a–d revealed a similar pattern as for the WWTP effluent, i.e. good correlations between TC, E. coli and ENT, with the highest correlation coefficient for E. coli and ENT (r = 0.77), but negative correlation coefficients for HPC and E. coli (r = -0.02) and for HPC and ENT (r = -0.64), confirming once again that HPC are not indicative of faecal contamination.

Therefore, the high FIB concentrations in the centre of the bay can also be attributed to vertical upwelling of effluent water across the thermocline. This seems to contradict the tracer test results, which demonstrated the effectiveness of the thermocline (Goldscheider et al. 2007). The tracer test was done in November during stable low-flow conditions while the bacteriological data reported here suggest that the contaminant plume from the WWTP outlet can possible penetrate the thermocline during summer floods, due to the high flow rates and relatively high temperature of the effluent water.

4.3.6 Salmonella detection While FIB only indicate the probable presence of pathogens in a water sample, their actual occurrence is rarely checked, because such analyses are often quite expensive and time-consuming. However, this type of data is highly relevant for human health risk assessment. In this study, eight different serotypes of pathogenic

Salmonella spp. were identified: 4 in the WWTP effluent and 5 in the Chamberonne

River; only the serotype ‘Give’ was found in both water types (Table 4.2).

119

Figure 4.5. Distribution of E. coli and enterococci (ENT) in the bay during different hydrological and limnological situations: a) July 2005: high flow and stratified lake, b) August 2005: low flow and stratified lake, c) April 2005: high flow and mixed lake, d) February 2006: low flow and mixed lake.

120

Salmonellae can survive for long periods in natural waters and the persistence and virulence of specific epidemic strains is of great concern in public health. Salmonellae can cause gastroenteritis, typhoid and paratyphoid fever (Baudart et al. 2000).

However, there are relatively few studies investigating the occurrence of Salmonellae in the environment, as the method of serotyping is not accessible to many laboratories. Baudart et al. (2000) assessed the diversity of Salmonella strains in different water types and found a higher diversity in river water (35 serotypes) than in wastewater (14 serotypes), and the highest diversity during floods. Although our data are less detailed, they are largely consistent with these findings. Sharma and

Rajput (1996) also assessed Salmonella in freshwater environments and found good correlation with TC, FC and faecal streptococci, underscoring both the health risk stemming from faecal contamination and the usefulness of the faecal indicator approach.

Table 4.2. Salmonella serotypes detected in the WWPT effluent and the Chamberonne

River.

Sampling date Hydrologic conditions Salmonella serotypes detected

WWPT effluent Chamberonne River

14/03/05 low flow Oranienburg ND

24/05/05 after flood Give Give

04/07/05 flood ND Veneziana

27/07/05 low flow Newport Salamae

08/08/05 low flow Infantis ND

22/08/05 flood ― Haifa, Agona

ND: no Salmonella detected ― : not sampled / analysed

121

4.4 Conclusions The input flux rates of faecal indicator bacteria (FIB) from the 2 major contamination sources in the Bay of Vidy, have been determined and their spatial- temporal distribution were assessed. FIB analyses included E. coli , enterococci (ENT), and total coliforms (TC), complemented by heterotrophic plate count (HPC). In addition, several pathogenic Salmonella serotypes were identified.

The WWTP outlet pipe at 30 m depth is quantitatively the most important source of faecal bacteria, whereas the FIB input flux rates from the Chamberonne

River were typically 1 to 3 orders of magnitude lower. However, the effective contribution of these two point sources to near-surface lake water contamination depends greatly on the local hydrological and limnological conditions.

When the water column is mixed, usually in winter and early spring, effluent water from the WWTP rapidly rises up to the surface causing widespread contamination in the bay. When the lake is stratified, the effluent water is usually trapped below the thermocline. Both situations have been confirmed by tracer tests

(Goldscheider et al. 2007). However, the present study suggests that upwelling of effluent water across the thermocline may occur during summer floods, when both the flow rate and temperature of the effluent water are high.

Despite the lower input flux rates, the Chamberonne River significantly contributes to contamination in the bay, mainly during floods, but to a lesser extent also during low flow conditions. The highest FIB levels generally occurred near the river mouth. Depending on the temperature (density) of the river water and the temperature profile of the lake, the river water stays near the surface or enters the water column at a particular depth. When the river water is warmer than the surface lake water, which is often the case in spring, it spreads near the surface. When it is colder than the lake water, which frequently occurs in autumn, it sinks down to

122 deeper zones of the lake, reflected in lower FIB concentrations near the mouth of the river.

The highest FIB concentrations in the near-surface water of the bay consequently occur during floods and during mixed lake conditions. The most favourable situation occurs when the lake is stratified and the flow rates of the river and WWTP are low, i.e. during the bathing season in summer. Increased surface water temperatures, higher biological activity and protozoan grazing, as well as intense UV radiation further contribute to the reduction of FIB levels near the lake surface. The first warm and sunny days after a major summer storm represent the most critical situation for bathers. Although the thermocline protects the epilimnion from contamination during the bathing season in summer, contamination from the

WWTP outlet pipe may spread in the hypolimnion and reach the drinking-water pumping station 3.8 km further to the west. Due to constantly lower temperatures and the absence of UV radiation at greater depths, the attenuation processes described above are expected to be substantially lower in the hypolimnion, resulting in longer survival times of indicator bacteria and pathogens.

123

References

Allen M, Edberg S, & Reasoner D (2004) Heterotrophic plate count bacteria ― what is

their significance in drinking water? International Journal of Food Microbiology

92 (3): 265-274.

APHA, AWWA and WEF (2005). Standard methods for the examination of water and

wastewater, 21 st ed., 1368 pp.

Baudart J, Lemarchand K, Brisabois A. & Lebaron P (2000) Diversity of Salmonella

strains isolated from the aquatic environment as determined by serotyping and

amplification of the ribosomal DNA spacer regions. Applied and Environmental

Microbiology 66 (4): 1544-1552.

Bonadonna L, Briancesco R, Coccia AM, Semproni M & Stewardson D (2002a)

Occurrence of potential bacterial pathogens in coastal areas of the Adriatic Sea.

Environmental Monitoring and Assessment 77 (1): 31-49.

Bonadonna, L., Briancesco, R., Ottaviani, M. & Veschetti, E. (2002b). Occurrence of

Cryptosporidium oocysts in sewage effluents and correlation with microbial,

chemical and physical water variables. Environmental Monitoring and Assessment

75 (3): 241-252.

Brettar I, & Höfle MG (1992) Influence of ecosystematic factors on survival of

Escherichia coli after large-scale release into lake water mesocosms. Applied and

Environmental Microbiology 58 (7): 2201-2210.

Cabral JP & Marques C (2006) Faecal coliform bacteria in Febros river (northwest

Portugal): Temporal variation, correlation with water parameters, and species

identification. Environmental Monitoring and Assessment 118 (1-3): 21-36.

Campos CJA & Cachola RA (2007) Faecal coliforms in bivalve harvesting areas of the

Alvor lagoon (Southern Portugal): Influence of seasonal variability and urban

development. Environmental Monitoring and Assessment 133 (1-3): 31-41.

124

Davis K, Anderson MA & Yates MV (2005) Distribution of indicator bacteria in

Canyon Lake, California. Water Research 39 (7): 1277-1288.

Doyle MP & Erickson MC (2006) Closing the door on fecal coliform assay. Microbe

1(4): 162-163.

Goldscheider N, Haller L, Poté J, Wildi W & Zopfi J (2007) Characterizing water

circulation and contaminant transport in Lake Geneva using bacteriophage

tracer experiments and limnological methods. Environmental Science &

Technology 41 (5): 5252-5258.

Hoxie NJ, Davis JP, Vergeront JM, Nashold RD & Blair KA (1997) Cryptosporidiosis-

associated mortality following a massive waterborne outbreak in Milwaukee,

Wisconsin. American Journal of Public Health 87 (12): 2032-2035.

Hughes KA (2003) Influence of seasonal environmental variables on the distribution

of presumptive fecal coliforms around an Antarctic research station. Applied

and Environmental Microbiology 69 (8): 4884-4891.

Krometis LAH, Characklis GW, Simmons OD, Dilts MJ, Likirdopulos CA & Sobsey

MD (2007) Intra-storm variability in microbial partitioning and microbial

loading rates. Water Research 41 (2): 506-516.

Laurion I, Ventura M, Catalan J, Psenner R & Sommaruga R (2000) Attenuation of

ultraviolet radiation in mountain lakes: Factors controlling the among- and

within-lake variability. Limnology and Oceanography 45 (6): 1274-1288.

Loizeau,JL, Pardos M, Monna F, Peytremann C, Haller L & Dominik J (2004) The

impact of a sewage treatment plant’s effluent on sediment quality in a small bay

in Lake Geneva (Switzerland-France). Part 2: Temporal evolution of heavy

metals. Lakes & Reservoirs: Research and Management 9(1): 53-63.

Montgomery MA & Elimelech M (2007) Water and sanitation in developing

countries: Including health in the equation. Environmental Science & Technology,

41 (1): 17-24.

125

OECD WHO (2003) Assessing microbial safety of drinking water, improving

approaches and methods (296pp.). London: IWA.

Pardos M, Benninghoff C, de Alencastro LP & Wildi W (2004) The impact of a

sewage treatment plant’s effluent on sediment quality in a small bay in Lake

Geneva (Switzerland-France). Part 1: Spatial distribution of contaminants and

the potential for biological impacts. Lakes & Reservoirs: Research and Management

9(1): 41-52.

Poté J, Haller L, Loizeau JL, Garcia Bravo A, Sastre V & Wildi W (2008) Effects of a

sewage treatment plant outlet pipe extension on the distribution of

contaminants in the sediments of the Bay of Vidy, Lake Geneva, Switzerland.

Bioresource Technology 99 : 7122–7131.

Pronk M, Goldscheider N & Zopfi J (2006) Dynamics and interaction of organic

carbon, turbidity and bacteria in a karst aquifer system. Hydrogeology Journal

14 (4): 473-484.

Sharma A & Rajput S (1996) Salmonella as an index of pollution of fresh-water

environments. Environmental Monitoring and Assessment 41 (1): 67-76.

Vaque D, Gasol JM & Marrase C (1994) Grazing rates on bacteria - the significance of

methodology and ecological factors. Marine Ecology-Progress Series 109 (2-3): 263-

274.

WHO (2004) Guidelines for drinking-water quality, 3rd ed., vol. 1, recommendations.

World Health Organisation, Geneva, 515 pp.

Wildi W, Dominik J, Loizeau JL, Thomas RL, Favarger PY, Haller L, Perroud A &

Peytremann C (2004) River, reservoir and lake sediment contamination by

heavy metals downstream from urban areas of Switzerland. Lakes & Reservoirs:

Research and Management 9(1): 75-87.

126

CHAPTER 5

Distribution and survival of faecal indicator bacteria in the sediments of the Bay of Vidy, Lake

Geneva, Switzerland

A similar version of this chapter was published under the following reference:

Haller L, Poté J, Loizeau J-L &Wildi W (2009) Distribution and survival of faecal indicator bacteria in the sediments of the Bay of Vidy, Lake Geneva, Switzerland .

Ecol Indicators 9: 540-547.

127

Abstract The purpose of this study was to determine the concentrations and the horizontal distribution of faecal indicator bacteria (FIB) including Escherichia coli and

Enterococcus sp. in the bottom sediments of the Bay of Vidy, City of Lausanne,

Switzerland. A vertical distribution of FIB in sediments near the municipal wastewater treatment plant (WWTP) outlet was evaluated and their persistence in those sediments was monitored for a period of 90 days. High FIB levels were measured in the sediments sampled near the WWTP outlet pipe and the mouth of the Chamberonne River, at concentrations ranging between 10 5 and 10 7 CFU 100 g -1.

FIB levels at 10 cm depth in the sediments near the WWTP outlet pipe ranged between 10 4 and 10 5 CFU 100 g -1, and were still detected in the top 6 cm after 90 days.

Results of this study indicate that freshwater sediments of the bay of Vidy constitute a reservoir of faecal indicator bacteria, which can persist in certain areas of the bay.

Possible resuspension of FIB and pathogens may affect water quality and may increase health risks to sensitive populations during recreational activities. FIB survival in sediments for long periods is of considerable significance for the understanding of microbial pollution in water and for the management of risk at specific recreational coastal sites.

128

5.1 Introduction Faecal indicator bacteria (FIB), residing in the gastrointestinal tracts of humans and animals, are commonly used to assess the microbiological safety of drinking and recreational waters. Although indicator bacteria do not necessarily cause illness, their presence indicates that the water has been contaminated by faecal material; implying the potential presence of pathogens (An et al., 2002; Noble et al., 2003). The U.S.

Environmental Protection Agency and the European Union recommend the use of

Escherichia coli (E. coli ) and members of the genus Enterococcus , the enterococci (ENT), to assess the hygienic safety of recreational waters (USEPA, 2000; EU, 2006).

Faecal pollution originates from a variety of human and non-human sources, but FIB contamination from human faecal material is generally considered to be a greater risk to human health as it is more likely to contain human enteric pathogens

(Scott et al., 2003). Exposure to bathing waters with high concentrations of E. coli and

ENT are documented in epidemiological studies as being associated with an increased risk of contracting gastrointestinal and respiratory illnesses (Cabelli, 1983;

Kay et al. 1994; Prüss, 1998; Haile et al., 1999). According to the European Directive

2006/7/CE concerning the management of bathing water quality, recreational waters are to be classified as poor, if concentrations of E. coli and ENT exceed 900 CFU·100 mL -1 and 330 CFU·100 mL -1 respectively, based upon a 90 -percentile evaluation.

Lausanne discharges the largest volumes of treated domestic and industrial wastewater into the nearby Bay of Vidy. In 2005, a one year survey was initiated by the Lausanne authorities to assess the water microbiological quality of the bay and to identify potential health risks for bathers. It concluded that major input fluxes of FIB entering into the bay through the wastewater treatment plant pipe and the nearby

Chamberonne River, may under certain hydrological and limnological conditions,

129 seriously affect the surface water quality of the bay (see chapter 4; Goldscheider et al., 2007).

The water quality of beaches along the Bay of Vidy is monitored regularly during the bathing period in summer. However, the microbial quality of sediments is overlooked. According to some studies, sediments may constitute an important reservoir of FIB in freshwater environments (LaLiberte and Grimes, 1982; Burton et al., 1987; Crabill et al., 1999; An et al., 2002; Alm et al., 2003). Accumulation of FIB and pathogenic organisms in sediments has been attributed to the sorption of the microorganisms to particles suspended in water, which then sediment out.

Sediments may contain 100 to 1000 times as many FIB as the overlying water (Davies et al., 1995). FIB can survive longer in sediments than in the water column since sediments provide favourable nutrient conditions (Gerba and McLeod 1976; Laliberte and Grimes, 1982), protection from sunlight inactivation (Sinton et al., 1999) and protozoan grazing (Davies and Bavor, 2000). Resuspension of FIB and pathogens from the sediments to the water column due to recreational activities or natural turbulence may contribute to potential human health risk (An et al, 2002; Craig et al.,

2004). It may cause water failures and also calls into question the choice and specificity of indicators for determining recent faecal contamination (Ferguson et al.,

2005).

Depending on sediment characteristics, coastal sediments can act as a reservoir of FIB; analysis of water quality alone may underestimate the risk of exposure to potentially pathogenic microorganisms in recreational waters (Craig et al., 2002). The main objectives of the present study were (i) to determine FIB concentrations in sediments across the whole bay of Vidy and to assess their horizontal spatial distribution, (ii) to assess FIB contents and their vertical distribution at the most contaminated site of the bay, near the WWTP outlet pipe, and (iii) to examine their persistence and survival in those sediments.

130

The spatial FIB distribution was performed by sampling sediments from 24 sites distributed within the entire bay but particularly focused on zones where potential contaminated water enter into the lake, as well as on the lake shore. The vertical distribution consisted in determining FIB levels in sediments up to 10 cm depth, sampled near the WWTP outlet pipe, in order to examine the direct impact of treated wastewater discharge on the nearby sediments. The influence of sediment characteristics including sediment grain size and organic matter content on FIB accumulation and persistence was also examined.

5.2 Materials and methods

5.2.1 Study site There are three major sources of water contamination of the Bay of Vidy: the wastewater treatment plant (WWTP), the Chamberonne River, and the Flon storm water outlet. The WWTP treats 1 to 3 m 3s-1 and exceptionally up to 6 m 3s-1 of raw water, depending on meteorological conditions. WWTP effluent waters are released into the bay via an underwater pipe at 35 m water depth and at a distance of 700 m from the shore. The Chamberonne River presently drains surface waters from its natural watershed and some untreated wastewaters. The Flon collects surface and wastewater from the western part of the city, which is usually treated at the WWTP but released into the lake during strong rain events, via a conduit at 10 m depth (Fig.

5.1).

131

Figure 5.1 . Location and map of the study area with the positions of the three known point sources of water contamination (WWTP, Chamberonne, Flon) and the sampling sites (Swiss coordinates, 1 km grid).

5.2.2 Sediment sampling Sampling took place in July 2007, during low flow (discharges of the river and

WWTP were low) and stratified lake conditions. The boat “La Licorne” of the Forel

Institute was used to collect surface sediments (layer of 0-3 cm thickness) in the bay using a “Ponar-type” grab sampler. At the site V14, near the WWTP outlet pipe, four sediment cores of 60 to 100 cm length were retrieved, using a gravity corer at a water depth of 40 m. Beach sediment samples were collected manually at 3 m offshore and at less than 1 m depth. The sediment samples were placed into sterile plastic containers, stored in an icebox at 4 ºC and immediately transported to the laboratory for analysis within 24 h.

132

5.2.3 Sediment characterization The particle grain size distribution was measured with a laser Coulter® LS-100 diffractometer (Beckman Coulter, Fullerton, CA, USA), following 5-min ultrasonic dispersal in deionized water according to the method described by Loizeau et al.

(1994). The proportions of three major size classes (clays < 2μm; silts 2-63μm; and sand > 63μm) were determined from size distributions, as well as the median grain size.

The sediments were dried at 60°C for 48 h and the water content was calculated as a weight difference. The dried sediments were then heated at 550°C for

1 h to determine the loss on ignition, used as a surrogate for organic matter content.

5.2.4 Faecal indicator bacteria analysis The faecal indicator bacteria including E. coli and ENT were quantified in the sediment samples. The sediments were resuspended by adding 100 g (wet weight) of sediment to 500 mL of 0.2 % Na 6(PO 3)6 in 1 L sterile plastic bottles and mixed for 30 min using an agitator rotary printing-press Watson-Marlow 601 controller (modified methodology from Balkwill and Ghiorse, 1985). The mixture was centrifuged at 4000 rpm for 15 min at 15°C. FIB in the supernatant were then counted according to the

Swiss standard methods for water quality determination using the membrane filtration method (OHyg, 2005). For each sample, triplicates of 20 mL of supernatant were passed through a 0.45 μm filter (47 mm diameter, Millipore, Bedford, USA), which was placed on different culture media (Biolife Italiana) supplemented with the anti-fungal compound Nystatin (100 µg mL -1 final concentration), using the following incubation conditions: E. coli : Tryptone Soy Agar (TSA) medium, incubated at 30°C for 4 h and transferred to TBX medium at 44°C for 24 h; ENT: Slanetz Bartley Agar medium, incubated at 44°C for 48 h and transferred into Bile Aesculin Agar medium at 44°C for 4 h. The results are expressed as colony forming units (CFU) per 100 g of dry sediments (CFU 100 g -1). The reproducibility of the whole analytical procedure

133 was tested by triplicates of selected sediment samples which revealed a mean variation coefficient of 12 % for E. coli and 22 % for ENT.

5.2.5 Persistence study Four intact sediment cores from the site V14 were stored vertically in a cold room at the in situ sediment temperature of 4°C. After 1, 30 and 90 days a core was opened longitudinally and sliced into 2 cm thick sections down to a depth of 10 cm.

Each layer was sub-sampled for sediment characterisation and FIB analysis. The last core was used to determine sedimentation chronology based on 137 Cs profile (Ritchie and McHenry, 1990). Completeness of the sediment retrieval was verified by measuring the naturally occurring, short-living 7Be (half-life 53 days). The core was sliced into 1 cm thick sections down to 55 cm. Radionuclide activities were measured in all samples by γ-spectrometry in an Ortec HPGe well detector (Ametek, Inc, USA).

5.3 Results and discussion

5.3.1 Sediment characteristics Particle size and organic matter content in sediments are shown in Table 5.1.

The particle size distribution reflects both the hydrodynamics of the bay and the pollution sources. Coarse sediments with 69%-98% sand were located close to the shores, where wave action continuously removes the fine fraction. Around the

Chamberonne River mouth (V20), sediments also contained a quite high proportion of sand (69%), a result of the river deposition. Sediments near the WWTP outlet pipe

(V14) at 40 m depth were generally soft and muddy, with a black colour and a strong smell. These sediments were mostly composed of silts (73%). The percentage of silts and sands varied sensibly with sampling sites but the percentage of clays remained always very low with values not exceeding 0.5%.

134

Table 5.1. Sampling sites, organic matter content and particle grain size. Values correspond to surface sediments (0-3 cm), except at V14.

Sampling site Organic matter (%) Median grain Clay (%) Silt (%) Sand (%) size (µm) V1 5.5 52 0.0 54 46 V2 0.8 192 0.1 6 94 V3 2.7 89 0.0 35 65 V4 6.1 59 0.0 52 48 V5 4.8 54 0.0 56 44 V6 1.6 139 0.0 23 77 V7 7.0 40 0.0 66 34 V8 0.8 383 0.1 12 88 V9 0.4 326 0.0 2 98 V10 0.7 177 0.1 8 92 V11 0.5 271 0.1 3 97 V12 1.6 252 0.1 7 93 V13 2.5 132 0.0 31 69 V14 24.9 29 0.1 73 27 V15 9.4 26 0.0 79 21 V16 7.2 59 0.0 52 48 V17 4.6 53 0.0 54 46 V18 9.6 25 0.1 79 21 V19 1.9 124 0.1 29 71 V20 1.3 117 0.2 31 69 V21 2.9 109 0.1 25 75 V22 5.1 68 0.1 48 52 V23 6.1 49 0.0 58 42 V24 9.1 21 0.2 84 16

V14 0-2 cm 25 39 0.1 65 35 V14 2-4 cm 17.7 38 0.1 64 36 V14 4-6 cm 18.7 39 0.1 65 34 V14 6-8 cm 16.45 37 0.1 66 34 V14 8-10 cm 15.2 36 0.05 69 31

135

This value is certainly underestimated by the analytical technique, due to the presence of large proportion of coarser particles (Loizeau et al., 1994).

The sediments sampled at V14 presented a high organic matter content which decreased with depth, from 25% at 0-2 cm to 15% at 10 cm depth. The high organic matter levels measured in surface sediments at V14 is due to important quantities of organic particles coming out of the WWTP outlet pipe. Sediment accumulation rates in front of the outlet pipe are strongly affected by the large quantities of particles coming out of the WWTP (Loizeau et al., 2003). Samples further away from the

WWTP outlet pipe, taken deeper than 20 m water depth, showed lower values in organic matter, between 6% and 10%. The lowest values were found on the sandy sediments collected on the beaches, with an organic matter content of less than 3%.

Spearman rank order correlation (preferred to Pearson correlation as data are not normally distributed) between organic matter content and sediment median grain size revealed a strong negative correlation with r = -0.941 (p<0.05). The finer the grain size, the higher the content in organic matter.

At V14, a sediment core has been dated to provide a vertical sediment age model. Only one 137 Cs activity peak was present in the sediment core (Fig. 5.2). The absence of the radionuclide below the activity maximum indicates that this peak can be attributed to 1963-64 Nuclear Bomb Test maximum fallout, 137 Cs being an artificial radionuclide released into the environment with measurable fallouts observed only after 1955 (Ritchie and McHenry, 1990). The peak activity of the Chernobyl accident

(1986) was not recorded in the sediment column. Direct visual observation of the sediment shows that a marked sediment texture change (colour, compactness and lamination) is observed at 24 cm depth. This observation, coupled with the water content profile, suggests that a sedimentological event (slump, density current) occurred, which eroded an unknown thickness of sediments, but including the 1986

136 layer. 7Be profile also indicates that the first 4 cm of sediment have been deposited within the last few months or that bioturbation partly mixed the sediments. The latter hypothesis is less probable as very few organisms have been observed within these sediments. The assumption of a continuous sedimentation is therefore not valid and it is impossible to date the 24 cm surface sediments.

Figure 5.2. Porosity and Cesium-137 activity in the sediment core taken at site V14.

5.3.2 Spatial distribution of FIB in surface sediments Figures 5.3a and 5.3b present data on E. coli and ENT levels in surface sediments of the bay, particularly along the beaches and in the vicinity of the known sources of contamination: the WWTP outlet pipe and the Chamberonne River mouth.

Concentrations of FIB in sediments varied significantly with the sampled sites. The maximum levels of FIB were observed at V14 near the WWTP outlet pipe, with

137 concentrations of 7.2 x 10 6 and 2.7 x 10 6 CFU 100 g -1, for E. coli and ENT, respectively.

Levels of FIB near the mouth of the Chamberonne River, at V20, were also high, but lower than V14, with values of 2 x 10 5 and 1 x 10 5 CFU 100 g -1 for E. coli and ENT, respectively. This is explained by the fact that FIB concentrations in the WWTP effluent are substantially higher than those found in the Chamberonne River.

According to a recent water quality survey of the bay of Vidy, E. coli concentrations vary approximately between 10 4 and 10 6 CFU 100 mL -1 in the WWTP effluent water and between 10 3 and 10 4 CFU 100 mL -1 in the Chamberonne River (Poté et al., 2009).

A contamination plume seemed to extend from the river mouth towards the centre of the bay, with FIB levels progressively decreasing with distance from the shore. At

300 m from the river mouth and 20 m depth, E. coli and ENT levels were still between

10 3 and 10 4 CFU 100 g -1. Further away from the WWTP outlet pipe, at V24, which is at

60 m depth, E. coli was not anymore detected and ENT was found at concentrations below 2 x 10 3 CFU 100 g -1.

FIB levels measured on the beach samples (V9-V13) were comprised between

900 and 5 x 10 3 CFU 100 g -1for E .coli and between 190 and 550 CFU 100 g -1 for ENT.

These lower concentrations are accounted for FIB inputs from the WWTP generally trapped under the thermocline when the lake is stratified in summer, as well as a high dilution of FIB coming from the river into the lake (Goldscheider et al., 2007;

Poté et al., 2009). In addition, their accumulation and survival in the sediments close to the shores is not favoured mainly due to lower organic matter content. Moreover, negative effects due to exposure to UV radiation (Sommaruga et al., 1997), and wave action (Lee et al, 2006) can not be excluded. Hughes (2003) assessed the influence of various seasonal environmental parameters on the survival of faecal coliforms around an Antarctic research station and found that solar radiation is the dominant factor. The penetration depth of UV radiation into lake water strongly depends on the concentration of dissolved organic carbon and turbidity, and often ranges between some decimetres and several meters (Laurion et al. 2000).

138

Figure 5.3a. Distribution of Escherichia coli in the sediments of the bay of Vidy.

Figure 5.3b . Distribution of Enterococcus in the sediments of the bay of Vidy.

139

FIB spatial distribution within the bay is mainly controlled by the locations of the two major sources of contamination, the WWTP pipe and the Chamberonne

River, by the water circulation within the bay, and to a lesser extent by the sediment characteristics: organic matter content and particle size. The correlation analyses between the microbiological indicators, E. coli and ENT, and the sediment characteristics revealed significant (p<0.05) but still weak correlation coefficients

(Table 5.2). Except the sites near the two main contamination inputs, FIB concentrations tend to be higher in sediments with finer grain size, a result of a lower energy environment and a higher organic matter content.

Table 5.2. Spearman correlation coefficients of the two FIB parameters and the sediment characteristics (organic matter and particle size). All correlation coefficients are considered as significant (p < 0.05, n=24). median organic particle ENT matter size 0.595 0.461 -0.407 E.coli 0.403 -0.419 ENT organic -0.941 matter

In aquatic environments, sediments may constitute a reservoir of different pollutants including FIB. The average sediment concentrations of FIB throughout all samples taken in the bay of Vidy were 3.1 x 10 5 and 1.2 x 10 5 CFU 100 g -1 for E. coli and ENT, respectively. According to the water microbiological quality survey undertaken in 2005 (Poté et al., 2009), FIB concentrations measured at the same time of the year in the water column at 2 m depth rarely exceeded the limits for bathing waters (European Directive 2006/7/CE) except near the river mouth, where FIB concentrations reached up to 10 4 CFU 100 mL -1. Concentrations of faecal bacteria are therefore significantly higher in sediments than in the water column, which is consistent with previous studies (Crabill et al., 1999; Alm et al., 2003; Craig et al.,

2004; Lee et al., 2006).

140

The risks associated with swimming in microbiologically polluted lakes or rivers during recreational activities are usually not life threatening but could take a substantial toll in children and immune-compromised individuals (Clark et al., 2003).

There are no health standards for beach sediments, hence the difficulty to assess whether the FIB levels measured in sediments along the shores in the Bay of Vidy pose a health risk. In considering a resuspension of 100 mg of sediment in 1L of water, we could determine whether the bacteria originating from that amount of sediment would exceed health standards (Lee et al., 2006). According to the EU directive on bathing waters, the water quality health standards for E. coli and ENT are 900 and 330 CFU 100 mL -1, respectively. Concentrations above 9 x 10 6 and 3 x 10 6

CFU 100 g -1 respectively for E. coli and ENT in sediments would lead to exceedances in the water column, after resuspension of 100 mg L-1. FIB concentrations measured in sediments close to the shore (sites V9 to V13) never rose above those levels.

However, even if FIB concentrations in beach sand were low, their presence suggests that pathogenic bacteria of intestinal origin may also be present. Many parameters such as natural turbulence, currents, floods, and the type of recreational activities can affect FIB and pathogens resuspension which could contribute significantly to higher microbial levels in the water. This may impact on the potential risks of human infections either by ingestion or by infiltration into the groundwater

(Wildi et al., 2004).

5.3.3 Vertical distribution of FIB At V14, the most contaminated site of the bay, a vertical distribution of FIB in the sediments was assessed. The vertical distribution of E. coli and ENT showed a decrease of FIB with depth (Fig. 5.4). In the surface layer, concentration values around 10 6 CFU 100 g -1 were measured for both E. coli and ENT; at 8-10 cm, values ranged

141

Levels of E. coli and ENT (CFU 100g-1) after 1 day

1.E+00 1.E+02 1.E+04 1.E+06 1.E+08

0-2cm

2-4cm

E.coli 4-6cm ENT

Depth (cm) Depth 6-8cm

8-10cm (a)

Levels of E. coli and ENT (CFU 100g-1) after 30 days

1.E+00 1.E+02 1.E+04 1.E+06

0-2cm

2-4cm

E.coli 4-6cm ENT

Depth (cm) Depth 6-8cm

8-10cm (b)

Levels of E. coli and ENT (CFU 100g-1) after 90 days 1.E+00 1.E+02 1.E+04 1.E+06

0-2cm

2-4cm

E.coli 4-6cm ENT

Depth (cm) Depth 6-8cm

8-10cm (c) Figure 5.4. Concentration of Escherichia coli and Enterococcus in function of sediment depth and time at V14 a) after 1 day, b) after 30 days, and c) after 90 days. Error bars represent 1σ (standard deviation of three replicates).

142

between 10 4 and 10 5 CFU 100 g -1, around 2 orders of magnitude lower than at the surface.

The high counts of indicator bacteria at all depths in the sediments at V14 may results from the continuous contamination of the sediments, coming from the WWTP effluent water, and may also indicate an accumulation and persistence of FIB in sediments. The high organic matter content and great proportion of fine particles of the sediments at V14 may lead to a higher rate of FIB adsorption. The disturbed sediment column at this site (see § 5.3.1) precludes a full discussion of the origin of the bacteria found in deep sediments. However, bacteria may also be transported from surface to deeper layers through different mechanisms. Recent studies have shown that bioturbation may influence oxygen, sediment particulates, macrofauna as well as bacteria fluxes from aerobic interface to sediment deeper layers (Nickell et al.,

2003; Michaud et al., 2005). Interactions between bacteria and sediment living organisms are well-documented; but there is still a paucity of information concerning the FIB ingestion and re-excretion by benthic organisms.

Results from this study are similar to those obtained by Desmarais et al.

(2002), which show that E. coli and ENT decrease with depth in Florida river sediment cores. However, in their study E. coli were undetectable at a depth over 5 cm. A study from Alm et al. (2003) detected E. coli and ENT up to a depth of 20 cm in the sand of a lake in Michigan, at levels between 200 and 400 CFU 100 g -1. A survey on seaside beaches in England demonstrates that abundance of faecal indicators in sediments is not due to high deposition rates and suggests that sediments act as a reservoir for faecal indicators bacteria (Obiri-Danso and Jones, 2000).

143

5.3.4 Survival of FIB in contaminated sediments FIB were monitored over a 90 day period at every sampling depth in the cores taken at site V14, to assess their persistence at the most contaminated area of the bay.

This experimental design differed from most of the earlier studies; the present study did not involve the inoculation of laboratory-grown bacteria, but monitored the concentration of indicator bacteria already present in the sediments, during a defined period. Concentrations of E. coli and ENT decreased by around 2 orders of magnitude in the surface layer, between the 1 st and the 30 th day of the experiment, and continued to decrease slightly until the 90 th day (Fig. 5.4). They were still detected after 90 days at concentrations between 104 and 10 5 CFU 100 g -1. At 6 cm of depth, concentrations of E. coli and ENT decreased approximately by one order of magnitude during the first 30 days, to reach concentrations around 10 3 CFU 100 g -1.

They were not detected at the 90 th day of the experiment.

Significant persistence of indicator organisms was observed in the surface sediments near the WWTP outlet pipe in the Bay of Vidy. In spite of no additional input into the cores, the high levels in organic content may have allowed indicator bacteria to survive such a long time. Results of this study also reveal that E. coli and

ENT remained cultivable for at least 90 days and thus were detectable through the culture-based method used here. Many enterobacteria, e.g. Vibrio sp ., E. coli ,

Enterococcus faecalis can activate survival strategies including starvation and the viable but nonculturable (VBNC) state in response to unfavourable growth conditions (Lleò M, 2005). They persist in the environment in conserving their viability despite the loss of their own culturability.

The difference in FIB survival between the surface and bottom layers cannot be explained by a difference in organic matter only. Additional limiting factors, such as predation, temperature, nutrient deficiencies, sediment grain size will also influence FIB persistence (Craig et al., 2002; Hughes, 2003).

144

According to a few studies, the relative amount of organic matter in sediments has a strong effect on persistence and growth of bacterial indicators and may result in different degrees of survival (LaLiberte and Grimes, 1982; Craig et al., 2004).

However, in our study no growth in FIB was recorded. This is consistent with previous studies looking at FIB survival in sediment microcosms, where growth is only observed in inoculated sediment without active predators present, while in the presence of predators, persistence rather than growth is observed (Lee et al., 2006).

The persistence of bacteria at certain levels in the sediments of Vidy may be the result of a balance between their rate of growth and their rate of predation (Marino and

Gannon, 1991). This finding leads to the question of whether pathogenic bacteria are also capable of extended persistence in this environment. Further extension of this work is necessary to demonstrate if this is likely to occur.

5.4 Conclusion The results of this study revealed that faecal indicator bacteria were present in almost all sediment samples collected at 24 locations in the Bay of Vidy. E. coli and

ENT were particularly abundant near the WWTP outlet discharge and the mouth of the Chamberonne River, where concentrations between 10 5 and 10 7 CFU 100 g -1 were detected. Survival of indicator organisms was observed in the surface sediments near the WWTP outlet pipe, for at least 90 days, which can be explained by the presence of high levels of organic matter. High FIB concentration levels in the sediments of the bay suggest that viable human pathogens may also be present. Their potential resuspension in the water column may lead to increased health risks to sensitive populations during recreational activities.

The evaluation of faecal indicator bacteria in sediments may represent a more stable index of overall or long-term water quality than in the overlying water

(Laliberte and Grimes, 1982; Ferguson et al., 2005). Their survival in sediments for

145 long periods is of considerable significance for the understanding of microbial pollution in water and for the management of risk at specific recreational coastal sites.

References

Alm EW, Burke J & Spain A (2003) Fecal indicator bacteria are abundant in wet sand

and freshwater beaches. Wat. Res . 37 : 3978-3982.

An Y, Kampbell DH & Breidenbach GP (2002) Escherichia coli and total coliforms in

water and sediments at lake marinas. Environ. Poll. 120 : 771-778.

Balkwill DL & Ghiorse WC 1985. Characterization of subsurface bacteria associated

with two shallow aquifers in Oklahoma. Appl. Environ. Microbiol. 50, 580-588.

Burton G, Gunnison D & Lanza G (1987) Survival of pathogenic bacteria in various

freshwater sediments. Appl. Environ. Microbiol . 53 : 633-638.

Cabelli VJ (1983) Health effects criteria for recreational waters. Research Triangle

Park, EPA, 1983, EPA-600/1-80-031.

Clark A, Turner T, Dorothy KP, Goutham J, Kalavati C & Rajanna B (2003) Health

hazards due to pollution of water along the coast of Visakhapatnam, east coast

of India. Ecotox. Environ. Safety 56 : 390-397.

Crabill C, Donald R, Snelling J, Foust R & Southam G (1999) The impact of sediment

fecal coliform reservoirs on seasonal water quality in Oak Creek, Arizona. Wat.

Res . 33 : 2163-2171.

Craig DL, Fallowfield HJ & Cromar NJ (2002) Enumeration of faecal coliforms from

recreational coastal sites: evaluation of techniques for the separation of bacteria

from sediments. J. Appl. Microbiol . 93 : 557-565.

146

Craig DL, Fallowfield HJ & Cromar NJ (2004) Use of microcosms to determine

persistence of Escerichia coli in recreational coastal water and sediment and

validation with in situ measurements . J. Appl. Microbiol . 96 : 922-930.

Davies CM, Long JAH, Donald M & Ashbolt N (1995) Survival of fecal

microorganisms in marine and freshwater sediments . Appl. Environ. Microbiol .

61 : 1888-1896.

Davies C & Bavor H (2000) The fate of storm-water associated bacteria in constructed

wetland and water pollution control pond systems. J. Appl. Microbiol . 89 : 349-

360.

Desmarais T, Solo-Gabriele HM & Palmer C (2002) Influence of soil on fecal indicator

organisms in a tidally influenced subtropical environment. Appl. Environ.

Microbiol. 68 : 1165-1172.

EU, European Directive 2006/7/CE of the European Parliament and of the Council of

15 February 2006 concerning the management of bathing water quality and

repealing Directive 76/160/EEC.

Ferguson DM, Moore DF, Getrich MA & Zhowandai MH (2005) Enumeration and

speciation of enterococci found in marine and intertidal sediments and coastal

water in southern California . J. Appl. Microbiol . 99 : 598-608.

Gerba C & McLeod JS (1976) Effect of sediment on the Survival of Escherichia coli in

Marine Water. Appl. Environ. Microbiol . 32 : 114-120.

Goldscheider N, Haller L, Poté J, Wildi W & Zopfi J (2007) Characterizing Water

Circulation and Contaminant Transport in Lake Geneva Using Bacteriophage

Tracer Experiments and Limnological Methods. Environ. Sci. Technol . 41 : 5252-

5258.

Haile RW, White JS, Gold M, Cressey R, McGee C, Millikan RC, Glasser A, Harawa

N, Ervin C, Harmon P, Harper J, Dermand J, Alamillo J, Barrett K, Nides M &

Wang GY (1999) The health effects of swimming in ocean water contaminated

by storm drain runoff. Epidemiology 10: 355–363.

147

Hughes KA (2003) Influence of Seasonal Environmental Variables on the Distribution

of Presumptive Fecal Coliforms around an Antarctic Research Station. Appl.

Environ. Microbiol . 69 : 4884-4891.

Kay D, Jones F, Wyer MD, Fleisher JM, Salmon RL, Godfree AF, Zelenauch-Jacquotte

A & Shore R (1994) Predicting likelihood of gastroenteritis from sea bathing:

results from randomised exposure. The Lancet 344 : 905-909.

LaLiberte P & Grimes DJ (1982) Survival of Escherichia coli in Lake Bottom sediment.

Appl. Environ. Microbiol . 43 : 623-628.

Laurion I, Ventura M, Catalan J, Psenner R & Sommaruga R (2000) Attenuation of

ultraviolet radiation in mountain lakes: Factors controlling the among- and

within-lake variability. Limnol. Oceanogr . 45 : 1274–1288 .

Lee CM, Lin TY, Lin C, Kohbodi GA, Bhatt A, Lee R & Jay JA (2006) Persistence of

fecal indicator bacteria in Santa Monica Bay beach sediments. Wat. Res. 40 : 2593-

2602.

Lleò MM, Bonato B, Benedetti D & Canepari P (2005) Survival of enterococcal species

in aquatic environments. Microbiol. Ecol. 54: 189-196 .

Loizeau J-L, Arbouillle D, Santiago S & Vernet J-P(1994) Evaluation of a wide range

laser diffraction grain size analyser for use with sediments. Sedimentology 41 :

353-361 .

Loizeau J-L, Rozé S, Peytremann C, Monna F & Dominik J (2003) Mapping sediment

accumulation rate by using volume magnetic susceptibility core corelation in a

contaminated bay (Lake Geneva, Switzerland). Eclogae Geol. Helv. 96 : Supplement

1, S73-S79 .

Marino RP & Gannon JJ (1991) Survival of faecal coliforms and faecal streptococci in

storm drain sediment. Wat. Res . 25 : 1089 -1098 .

Michaud E, Desrosiers G, Mermillod-Blondin F, Sundby B & Stora G (2005) The

functional group approach to bioturbation: The effects of biodiffusers and

gallery-diffusers of the Macoma balthica community on sediment oxygen

uptake. J. Exp. Mar. Biol. Ecol. 326 : 77–88 .

148

Nickell LA, Black KD, Hughes DJ, Overnell J, Brand T, Nickell TD, Breuer E &

Harvey SM (2003) Bioturbation, sediment fluxes and benthic community

structure around a salmon cage farm in Loch Creran, Scotland. J. Exp. Mar. Biol.

Ecol. 285–286, 221– 233 .

Noble RT, Moore DF, Leecaster MK, McGee CD & Welsberg SB (2003) Comparison of

total coliform, fecal coliform, and enterococcus bacterial indicator response for

ocean recreational water quality testing. Wat. Res. 37 : 1637-1643 .

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. Wat. Res. 34 : 519-527 .

OHyg. Ordonnance sur l’hygiène (2005) Ordonnance du DFI sur les denrées

alimentaires et les objets usuels (ODIOUS). Berne.

Poté J, Goldscheider N, Haller L, Zopfi J, Khajehnouri F & Wildi W (2009) Origin and

spatial-temporal distribution of fecal bacteria in a bay of Lake Geneva,

Switzerland. Environ. Monit. Assess . 154 : 337-348 .

Poté J, Haller L, Loizeau J-L, Garcia Bravo A, Sastre V & Wildi W (2008) Effects of a

sewage treatment plant outlet pipe extension on the distribution of

contaminants in the sediments of the Bay of Vidy, Lake Geneva, Switzerland.

Biorem. Technol . 9: 7122–7131 .

Prüss A (1998) Review of epidemiological studies on health effects from exposure to

recreational water. Intern. J. Epidemiol . 27 : 1-9.

Ritchie JC & McHenry JR (1990) Application of radioactive fallout cesium-137 for

measuring soil erosion and sediment accumulation rates and patterns – a

review . J. Environ. Qual , 19 : 215-233 .

Scott TM, Parveen S, Portier KM, Rose JB, Tamplin ML, Farrah SR, Koo A & Lukasik

J (2003) Geographical variation in ribotype profiles of Escherichia coli isolates

from humans, swine, poultry, beef, and dairy cattle in Florida. Appl. Environ.

Microbiol. 69 : 1089-1092 .

149

Sinton L, Finlay R & Lynch P (1999) Sunlight inactivation of fecal bacteriophages and

bacteria in sewage-polluted seawater. Appl. Environ. Microbiol . 66 : 230-237.

Sommaruga R, Obernosterer I, Gerhard JH & Psenner R (1997) Inhibitory effect of

solar radiation on thymidine and leucine incorporation by freshwater and

marine bacterioplankton. Appl. Environ. Microbiol . 63 : 4178-4184 .

U.S. Environmental Protection Agency (2000) Improved enumeration methods for

the recreational water quality indicators : enterococci and Escherichia coli EPA-

821/R-97/004. U.S. Environmental Protection Agency, Washington, D.C .

Wildi W, Dominik J, Loizeau JL, Thomas RL, Favarger P-Y, Haller L, Perroud A &

Peytremann C (2004) River, reservoir and lake sediment contamination by

heavy metals downstream from urban areas of Switzerland. Lakes Reserv. Res.

Manag. 9: 75-87 .

150

CHAPTER 6

Influence of freshwater sediment characteristics

on persistence of faecal indicator bacteria

A similar version of this chapter was published under the following reference:

Haller L, Amedegnato E, Poté J, Wildi W (2009) Influence of freshwater sediment characteristics on persistence of faecal indicator bacteria. Water Air & Soil Pollution

203 : 217-227.

151

Abstract Extended persistence of enteric bacteria in coastal sediments and potential remobilization of pathogens during natural turbulence or human activities may induce an increased risk of human infections. In this study, the effect of sediment characteristics such as particle grain size, nutrient and organic matter contents on the survival of faecal indicator bacteria (FIB) including total coliforms, Escherichia Coli and Enterococcus, was investigated. The experimentation was carried out for 50 days in microcosms containing lake water and different contaminated freshwater sediments in continuous-flow and batch conditions. Results of this study revealed:

(1) extended FIB survival in sediments up to 50 days, (2) higher growth and lower decay rates of FIB in sediments with high levels of organic matter and nutrients and small (mainly silt) grain size, and (3) longer survival of Enterococcus sp. compared to

Escherichia coli and total coliforms. FIB survival in sediments and possible resuspension are of considerable significance for the understanding of permanent microbial pollution in water column and therefore human risk during recreational activities.

152

6.1 Introduction The origin of pathogenic bacteria in surface water includes municipal wastewater treatment plant discharges (WWTP), agricultural or storm runoff and other diffuse sources of human and animal wastes. Faecal indicator bacteria (FIB) including Escherichia coli (E. coli ), enteroccoccus sp . (ENT) and total coliform (TC), residing in the gastrointestinal tract of humans and animals, are commonly used to assess the microbiological safety of drinking and recreational waters. Studies have shown that FIB from different sources can be accumulated and distributed in freshwater sediments (Burton et al. 1987; La liberte and Grimes 1982; Davies et al.

1985). The U.S. Environmental Protection Agency and the European Union recommend the use of E. coli and ENT, to assess the hygienic safety of recreational waters (USEPA 2000; EU 2006).

In the aquatic environment, sediments may constitute a reservoir of different pollutants including inorganic and organic compounds, and microorganisms.

Accumulation and survival of FIB and pathogenic organisms in sediments has become a subject of increasing interest due to their negative impact on surface water quality. The survival of faecal bacteria, once released in the aquatic environment, is determined by numerous environmental factors including temperature variations, salinity, oxygen levels, nutrient deficiencies, predation and ultra-violet irradiation

(McFeters and Singh 1991; Davies et al. 1995; Thomas et al. 1999; Hughes 2003; Craig et al. 2004).

Several field and laboratory experiments have documented FIB survival and growth in aquatic systems, especially in the presence of sediments (Gerba and

McLeod 1976; Laliberte and Grimes 1982; Craig et al. 2004; Lee et al. 2006). It has been showed that the level of FIB in aquatic compartments including water column, interstitial water and sediment varied with seasonal period and environmental

153 conditions (Huges 2003; Davies et al. 2005; Goldscheider et al. 2007). There is evidence that FIB can survive and accumulate in sediments at levels 100 to 1000 times higher than in overlying waters (Ashbolt et al. 1993). Higher concentrations of FIB and pathogenic organisms in sediments have been attributed to the sorption of microorganisms to particles suspended in water, which then sediment out (Davies et al. 1995; An et al. 2002; Alm et al. 2003). FIB can also survive longer in sediments than in the water column since sediments provide favourable nutrient conditions (Gerba and McLeod 1976; Laliberte and Grimes 1982), protection from sunlight inactivation

(Sinton et al. 1999) and protozoan grazing (Davies and Bavor 2000). Resuspension of pathogenic organisms from sediments to the water column, due to recreational activities such as swimming in beach waters, natural turbulence such as currents, waves or floods may increase the risk of human infection (An et al. 2002; Evanson and Ambrose 2006), either by direct ingestion or by infiltration into the groundwater

(Wildi et al. 2004).

Numerous studies relied on laboratory microcosm experiments to evaluate

FIB and pathogens survival in the aquatic environment (Gerba and McLeod 1976;

Burton et al. 1987; Thomas et al. 1999, Ghoul et al. 1990; Craig et al. 2004; Anderson et al. 2005). Most of these studies used bacterial strains inoculated in sediment microcosms to monitor the influence of environmental factors such as organic matter, temperature, ultra-violet radiation, on their survival. Microcosms are a very useful tool to understand the complex influence of biotic and abiotics factors on FIB persistence (Fish and Pettibone 1995). Unlike in situ experiments, the use of microcosms facilitates the investigation of faecal bacteria response to specific environmental conditions in isolation (Craig et al. 2004).

Levels of nutrients, organic matter and grain size may vary considerably according to the type of coastal sediments and may influence microorganisms persistence. There is a considerable interest in studying the influence of these

154 parameters on FIB survival in sediments, as these microorganisms and pathogens can be remobilized to the water column, inducing bacteriological pollution and greater risks for human health during recreational activities. The objective of this study was to determine the influence of sediment characteristics including grain size, nutrient and organic matter levels in fresh water sediments, on the survival of FIB.

The experiment was carried out for a period of 50 days using microcosms.

Microcosms were designed to simulate lake conditions and were constituted of lake water and sediments which present different characteristics. The following 3 sites were selected, on the basis of their varying organic matter and nutrient contents: the

Bay of Vidy (Lake Geneva, Switzerland, very rich in organic matter and nutrients),

Lake Bret (eutrophic) and the mouth of the Versoix River in Lake Geneva (poor in nutrients). Before the start of the study, sediments were contaminated with sewage water, containing high concentrations of faecal bacteria. The experimentation was carried out in continuous-flow and batch microcosm conditions. The physicochemical parameters including water temperature, dissolved oxygen, conductivity and pH were monitored during the whole experimentation.

6.2 Materials and methods

6.2.1 Study sites and sampling The sediments were collected from three sites: the Bay of Vidy, Lake Bret and the mouth of the Versoix River (Fig. 6.1). The Global Positioning System (GPS) locations of the sampling sites are presented in Table 6.1.

The Bay of Vidy is located near Lausanne, on the northern shore of Lake

Geneva. Due to the wastewater treatment plant (WWTP) in Vidy, this bay is the most contaminated area of Lake Geneva (Loizeau et al. 2004; Poté et al. 2008, 2009) and sediments are extremely rich in organic matter and nutrients.

155

Lake Bret is a small and shallow reservoir lake in the surrounding area of

Lausanne city. It has a volume of 3 km 3, a length of approximately 1.5 km, and a maximum depth of 20 m. It is a eutrophic lake, poor in oxygen and rich in nutrients.

The Versoix River is a small river flowing from the Jura mountains (France) into

Lake Geneva, with an annual average discharge of 3.4 m 3 s-1. The mouth of the river is poor in nutrients and receives no WWTP discharges.

Figure 6.1. Map of the study area with the positions of the three sampling points

Table 6.1. GPS location of sampling sites in Swiss coordinates Sampling sites X Y Bay of Vidy 534 676 151 543 Lake Bret 548 817 151 659 Versoix River 502 230 125 625

156

Surface sediments (layer of 0-3cm thickness) from the Bay of Vidy were collected in front of the WWTP outlet pipe at 40 m water depth, from a boat, using a “Ponar- type” grab sampler (SDEC, France). Sediment samples from Lake Bret and the

Versoix River were taken manually at 1m from the shore. All sediment samples were placed into sterile plastic containers and stored in a cold room at 4 ºC prior to analysis.

The WWTP effluent water was sampled directly from the WWTP outlet pipe.

Lake water, used for the microcosms, was sampled in the area of Versoix at 5 m depth with a centrifugal pump and filtered at 1.2 µm (CUNO filter).

6.2.2 Microcosms Microcosms consisted of plastic aquaria of a size of 46.5 cm length, 22 cm width, and 26 cm height with an overflow at 21 cm. Prior to the experiment, all microcosm equipment was rinsed with a solution of HCl (1 N) and with deionized water. Microcosms were designed to simulate lake conditions. They contained sediments from three different sites as mentioned above and water from Lake

Geneva. The sampled sediments were first homogenized with a spatula and filled into the microcosms to a height of 2-3 cm. They were then flooded for 24 hours with 5

L of treated wastewater coming from the WWTP. According to a previous study

(Poté et al. 2008a ), concentrations of TC, E. Coli and ENT ranging between 10 6 and 10 9 colony-forming units (CFU) 100 mL -1 are present in the WWTP effluent water (no microbiological treatment of water is done at the WWTP). After 24 h, contaminated water was removed and replaced with 18 L of lake water.

157

6.2.3 Survival study Two series of experiments were performed: a series of three continuous-flow microcosms and a series of three batch microcosms. Each microcosm from each group contained sediments from a selected site and was named as followed:

- MV1: sediments from Vidy + lake water (water renewal)

- MB1: sediments from Bret + lake water (water renewal)

- MVe1: sediments from Versoix + lake water (water renewal)

- MV2: sediments from Vidy + lake water (batch)

- MB2: sediments from Bret + lake water (batch)

- MVe2: sediments from Versoix + lake water (batch)

Continuous-flow microcosms were supplied with lake water, pumped from a reservoir with a peristaltic pump (Tygon R-3607), through silicone tubings. The renewal rate of water was about 10 L day -1 so the mean renewal time of a microcosm was approximately 1.8 days. Water was introduced in microcosms at mid-height close to the wall opposite to the overflow. All microcosms were thermostabilized and kept in a room with artificial light at a temperature of 20 ± 2 °C.

During experimentation, microcosm sediments were sampled on days 1, 5, 10,

20, 30, 40, and 50 for bacterial analysis. For each time of sampling and each set of conditions, all the analyses were conducted in triplicate.

6.2.4 Sediment and water characterization The grain size distribution was measured using a particle size analyzer

Coulter ® LS-100 (Beckman Coulter, Fullerton, CA, USA), following ultrasonic dispersal in de-ionized water (Loizeau et al. 1994). The proportions of three major size classes (clay < 2 μm; silt 2-63 μm; and sand > 63 μm) were determined from size distributions, as well as the median grain size.

158

The sediments were dried at 60 °C during 48 h and the water content was calculated from weight difference. The sediment total organic matter content was measured by loss on ignition at 550°C for 1 hour in a Salvis oven (AG Emmenbrücke,

Luzern, Switzerland). Total nitrogen and ammonium concentrations in the sediments were measured by using the method of Kjeldahl (1883). Total phosphorus and its different forms were determined following the fractionation scheme of Williams et al.

(1976) as modified by Burrus et al. (1990). Sediment characteristics were measured before and at the end of the experimentation.

Water physicochemical parameters including conductivity, temperature, pH and concentration of dissolved oxygen were measured using a Multi 350i (WTW,

Germany) probe, each time sediment microcosms were sampled for microbiological analysis.

6.2.5 Bacteria quantification Faecal indicator bacteria including TC, E. coli and ENT were quantified in the different sediment samples. Bacteria were resuspended by adding 100 g (wet weight) of sediment to 500 mL of 0.2 % sodium hexametaphosphate (Na 6(PO 3)6) in 1 L sterile plastic bottles and mixed for 30 min using the agitator rotary printing-press Watson-

Marlow 60 1 controller (modified methodology from Balkwill and Ghiorse 1985). The mixture was centrifuged at 4000 rpm for 15 min at 15 °C. FIB in the supernatant were then counted according to the Swiss standard methods for water quality determination, using the membrane filtration method (OHyg 2005). For each sample, triplicates of 20 mL of supernatant were passed through a 0.45 μm filter (47 mm diameter, Millipore, Bedford, USA), which was placed on different FIB culture media

(Biolife, Italiana), supplemented with the anti-fungal compound Nystatin (100 µg mL -1 final concentration), using the following incubation conditions: TC: Endo agar medium, incubated at 35 °C for 24 h; E. coli : Tryptone Soy Agar (TSA) medium, incubated at 30 °C for 4 h and transferred to tryptone bile x-glucuronide (TBX)

159 medium at 44 °C for 24 h; ENT: Slanetz Bartley Agar medium, incubated at 44°C for

48 h and transferred into Bile Aesculin Agar medium at 44 °C for 4 h. The results are expressed as colony forming units (CFU) per 100 g of dry sediments (CFU 100 g -1).

6.2.6 Statistical analysis To facilitate comparisons between indicators and types of sediments, the decay rate constant (k) was estimated, by fitting experimental results to the following equation:

Nt = N0 x e-kt , where Nt is the number of bacteria at time t and N0 is the number of bacteria at time t=0 (t=0 corresponds to the start of the decay phase) (Davies and

Evison 1991). The decay rate constants measured for each indicator in every microcosm were compared by the standard statistical t–test. Statistical treatment of data has been realised using SigmaStat 3.11 (Systat software, Inc., USA).

6.3 Results

6.3.1 Water characteristics Table 6.2 shows the physical and chemical characteristics of water monitored throughout the whole experimentation period, in each microcosm. Water temperature was kept at near constant levels at 20 ± 2 °C. pH remained between 7.3 and 8.5 in all microcosms over the sampling period. In the continuous-flow microcosms, conductivity stayed constant with an average of 285 μS cm -1, but in the batch microcosms, it gradually increased with time. The measured values

(beginning-end of experimentation) were 286-404, 276-701 and 279-512 μS cm -1 for

MV2, MB2 and MVe2, respectively. In the lake pre-filtered water, conductivity values ranged from 270 to 280 μS cm -1. Dissolved oxygen dropped significantly in the microcosms with sediments from Vidy (MV1 and MV2), from 8 mg L -1 at the

160 beginning of the experiment to approximately 0.6 mg L -1 at the end. In the microcosms with sediments from Bret and Versoix, dissolved oxygen was around 8 mg L -1 on day 1, decreasing slightly during the sampling period to reach values between 5.3 and 6.3 mg L -1 at the end of the study.

Table 6.2. Water characteristics (range over a 50-day test period)

Dissolved oxygen (mg/liter) Conductivity (µg/cm) pH

MV 1 0.6-8.0 278-293 7.3-7.8 MV 2 0.8-7.9 286-404 7.4-7.8 MB 1 6.3-8.2 275-292 7.9-8.5 MB 2 5.3-8.3 276-701 7.8-8.5 MVe 1 6.3-8.4 278-295 7.9-8.5 MVe 2 5.8-8.5 279-512 7.9-8.5

6.3.2 Sediment characteristics Sediment characteristics including particle grain size, organic matter and nutrient contents are given in Tables 6.3 and 6.4. These parameters were measured at the beginning and the end of the experimentation. Sites were chosen to investigate the effect of sediment type on FIB survival. Sediments from the Bay of Vidy showed a high organic matter content of about 21%. This value is higher than the values measured on other sample sites: 12.6% and 1.8% for the sediments from Lake Bret and Versoix, respectively. Concentrations of total nitrogen, ammonium and total phosphorus measured at the start of the experiment were also higher in the sediments of Vidy, 12.6 ppm, 1.6 ppm and 6784 ppm respectively, while they were considered intermediate in the sediments from Bret (4.7 ppm, 0.6 ppm and 791 ppm respectively) and low in the sediments from Versoix (0.5 ppm, 0.1 ppm and 420 ppm respectively).

161

Sediments varied also in terms of grain size, with a proportion of 72% of silts in Vidy to approximately 48% in the sediments of Bret and Versoix.

Between the beginning and the end of the experiment, organic matter decreased significantly in the microcosms with sediments from Vidy, from 21% at the start, to a final level of 12% in MV1 and 18% in MV2, while it remained quite stable in the other microcosms. Initial concentrations of total nitrogen and ammonium in sediments from Vidy also decreased significantly by approximately 33% and 62%, respectively. In sediments from Bret, total nitrogen and ammonium levels decreased of about 21% and 25%, respectively, and in sediments from Versoix, they stayed quite low during the whole experimentation period. Values showed an average decrease of 13% of organic phophorus in the Vidy sediments while they remained quite constant in the other microcosms.

Table 6.3. Sediment characteristics at day 0 Clay/silt/sand Ntot a NH4-Nb PTot c PO d NAIP e Sediment Organic matter (%) proportion (mg/kg) (mg/kg) (mg/kg) (mg/kg) (mg/kg) (%) Vidy 21.2 12.6 1.6 6'783.8 688.2 3'103.0 0 / 72 / 28 Bret 12.6 4.7 0.6 791.2 364.5 113.5 0.3 / 47.8 / 51.9 Versoix 1.8 0.5 0.1 420.1 33.8 41.5 0.1 / 48.4 / 51.5

Table 6.4. Sediment characteristics at day 50 Organic matter Ntot a NH4-Nb PTot c PO d NAIP e Sediment (%) (mg/kg) (mg/kg) (mg/kg) (mg/kg) (mg/kg) MV1 12.2 8.3 0.7 7'232.6 602.2 5'240.8 MV2 17.9 8.5 0.8 6'512.7 593.5 5'260.5 MB1 11.5 4.1 0.5 748.8 354.4 139.0 MB2 12.2 3.3 0.4 765.0 341.8 138.8 MVe1 2.2 0.7 0.1 483.7 46.2 55.3 MVe2 1.4 0.3 0.1 404.4 21.8 53.6 a Total nitrogen b Ammonium c Total phosphorus d Organic phophorus e Non apatitic inorganic phosphorus

162

6.3.3 Survival study Sediments of all microcosms were contaminated for 24 h with water coming from the WWTP. FIB analysis after contamination gave initial concentrations ranging between 1.9 x 10 5 and 1.2 x 10 6 CFU 100 g -1 for TC, between 1 and 8.5 x 10 5 CFU 100 g -1 for E. coli and between 9.8 x 10 3 and 3.1 x 10 5 CFU 100 g -1 for ENT. FIB initial concentrations varied slightly between the different sediment types and may be explained by a variation in bacterial association with sediment particles.

FIB analysis showed similar patterns between the different types of sediments but differences were observed during growth and decay phases (Fig. 6.2 (a)-(f)).

The survival curves of TC, E. coli and ENT in continuous-flow microcosms with sediments from Vidy, Bret and Versoix are illustrated in Figs 6.2 (a), (b), (c).

During the first 5 days of the experiment, TC and E. coli concentrations generally decreased by around 1 order of magnitude in all types of sediments, from 10 5 to 10 4

CFU 100 g -1, but remained quite constant at 10 4 CFU 100 g -1 for ENT. Between day 5 and 12, microbiological analysis revealed a growth phase for almost all faecal indicators. In sediments from Vidy, TC and E. coli concentrations increased by around 2 orders of magnitude from 10 4 to 10 6 CFU 100 g -1. ENT concentrations increased by 1 order of magnitude from 10 4 to 10 5 CFU 100 g -1. In sediments from Bret and Versoix, FIB growth was not as strong as in the Vidy sediments and increased by not more than 1 order of magnitude. This growth was followed up by a decay phase in all microcosms. After 50 days, TC and E. coli were not detected anymore, but ENT still showed concentrations between 10 2 and 10 3 CFU 100 g -1.

The survival curves of TC, E. coli and ENT in batch microcosms, with sediments from Vidy, Bret and Versoix are illustrated in Figs 6.2 (d), (e), (f). The trend was quite similar to the microcosms with water renewal. A decay phase was observed during the first five days of the experiment. FIB concentrations decreased

163 by 1 to 2 orders of magnitude. This was followed by an increasing phase of 7 to 15 days. E. coli and ENT concentration rates generally reached 10 5 CFU 100 g -1 in sediments from Vidy and Bret, and 10 4 CFU 100 g -1 in sediments from Versoix. There was not much difference between the different types of sediments for TC; they increased to levels around 10 5 CFU 100g -1. As in microcosms with water renewal, this was followed by a decreasing phase in all microcosms.

Figure 6.2: Survival of faecal indicator bacteria in sediment microcosms

1e+7 TC E.coli 1e+6 ENT

1e+5

1e+4

1e+3

1e+2

FIB concentrations FIB (CFU100g-1) 1e+1

1e+0 0 10 20 30 40 50 60 Time (days) (a) Survival in microcosm MV1 (continuous-flow with sediments from Vidy)

164

1e+6 TC E.coli 1e+5 ENT

1e+4

1e+3

1e+2

FIB concentrations FIB (CFU 100g-1) 1e+1

1e+0 0 10 20 30 40 50 60 Time (days) (b) Survival in microcosm MB1 (continuous-flow with sediments from Lake Bret)

TC E.coli 1e+5 ENT

1e+4

1e+3

1e+2

FIB concentrationsFIB 100g-1) (CFU 1e+1

1e+0 0 10 20 30 40 50 60 Time (days) (c) Survival in microcosm MVe1 (continuous-flow with sediments from Versoix)

165

1e+7 TC E.coli 1e+6 ENT

1e+5

1e+4

1e+3

1e+2

FIB concentrations FIB (CFU100g-1) 1e+1

1e+0 0 10 20 30 40 50 60 Time (days) (d) Survival in microcosm MV2 (batch with sediments from Vidy)

1e+6 TC E.coli 1e+5 ENT

1e+4

1e+3

1e+2

FIB concentrationsFIB 100g-1) (CFU 1e+1

1e+0 0 10 20 30 40 50 60 Time (days) (e) Survival in microcosm MB2 (batch with sediments from Lake Bret)

166

TC 1e+5 E.coli ENT

1e+4

1e+3

1e+2

FIB concentrationsFIB 100g-1) (CFU 1e+1

1e+0 0 10 20 30 40 50 60 Time (days) (f) Survival in microcosm MVe2 (batch with sediments from Versoix)

A decline of FIB populations was observed in all microcosms after 12 to 20 days following the start of the experiment. The decay rate constant (k) was estimated for each microcosm, by fitting experimental results to the following equation:

Nt = N0*e -kt , where Nt is the number of bacteria at time t and N0, the number of bacteria at time t=0 (t=0 corresponds to the start of the decay phase which is at day

12 or 20 depending on the microcosm). For all indicators, decay rates were in general lower in sediments from Vidy and Bret (higher nutrients, organic matter and silt content) than in Versoix (Table 6.5).

167

Table 6.5. Decay constants, k (days -1) for TC, E. coli and ENT populations, in microcosms containing different sediment types.

TC E. coli ENT

Decay constant Std. Decay constant Std. Decay constant Std. (k; days -1) Error (k; days -1) Error (k; days -1) Error

MV1 0.28* 0.0183 0.29* 0.0281 0.07 0.0380 MV2 0.22* 0.0249 0.23 0.0819 0.08 0.0259 MB1 0.34* 0.0075 0.45* 0.0095 0.14* 0.0065 MB2 0.61 0.5 0.55* 0.0294 0.23 0.0506 MVe1 0.38* 0.0205 0.80* 0.0446 0.22 0.0569 Mve2 0.36* 0.0148 0.77* 0.0446 0.23 0.1051 *P ≤ 0.05

6.4 Discussion Some studies assessed the survival of FIB in both marine and freshwater sediments (Davies et al. 1995). It has been demonstrated that freshwater sediments can constitute a reservoir of FIB, which can persist according to sediment characteristics.

The results of this experiment showed extended FIB survival, up to 50 days and even more for ENT in freshwater sediments. Between days 5 and 12, some growth occurred in all microcosms, followed by a decay phase. FIB concentrations reached in general higher levels in the Vidy sediments compared to Bret and Versoix.

The lowest rate of decay also occurred in the microcosms containing sediments from

Vidy. For TC and E. coli , the decay rates were significantly lower in Vidy than Bret or

Versoix (t-test, p < 0.05). A significant difference in E. coli decay rates was also observed between sediments from Bret and Versoix (t-test, p < 0.05). ENT followed the same trend than TC and E. coli but statistical analysis could not show any significant difference in ENT survival, between sediments types. ENT decay rate constants (k) were surrounded by quite important uncertainties, therefore the

168 comparison between two values was quite difficult to assess. In general, higher growth rates and lower decay rate constants were measured in sediments from Vidy containing higher nutrient and organic matter levels and finer grain size.

Results of this study confirm, in agreement with previous studies that nutrient availability may have a significant impact on microbial survival, especially E. coli .

Gerba and McLeod (1976) attributed the longer survival of E. coli in estuarine sediments to the greater content of organic matter present in the sediment than in seawater. In addition to longer survival times, they also observed some growth where E. coli was added to samples taken from most polluted sites. In a study of pathogenic and indicator organisms survival in freshwater sediments at 20 ºC,

Burton et al. (1987) identified prolonged survival and lower decay rates of E. coli and

S. Newport in sediments containing higher proportions of clay, organic matter and nutrients compared to sandy sediment with low nutrient levels. Results from Craig et al. (2004) showed that in general the decay rate of E. coli was more important in water than in sediments. Small particle size and high organic carbon content were also enhancing E. coli survival in coastal sediments in the microcosms. To test the hypothesis that sediment organic content is an important determinant of FIB survival, Lee et al. (2006) monitored E. coli growth in microcosms with sediments in absence of their natural organic matter. E. coli levels measured in these microcosms were below the detection limit.

However, no relationship was observed between the loss of nutrients and FIB decay rates. In the microcosms filled with sediments from Vidy, organic matter and nutrient levels decreased significantly in comparison to microcosms with sediments from Bret and Versoix, in which organic matter and nutrient contents did not show any substantial variation. Despite this difference in nutrient consumption, indicator bacteria evolution followed more or less the same trend in all microcosms. This indicates that sufficient nutrients were present, even in the sediments from Versoix,

169 to support FIB limited growth and significant persistence. All types of sediments used in that experiment may not only act to extend the survival time of indicator bacteria, but may also support a certain growth. These results demonstrate that indicator bacteria are capable of utilizing nutrients adsorbed to sediments or in the interstitial water from areas polluted by sewage discharges as well as from areas free of pollution. The significant decrease in organic matter, nutrients and dissolved oxygen observed in the microcosms with Vidy sediments, may be explained by consumption from native microflora naturally present in the sediments.

Our results support previous findings where in general only limited growth or no growth at all was observed in experiments with non-autoclaved sediments. Of the studies undertaken, many have used sterile sediment (Gerba et al. 1976; Laliberte and

Grimes 1982; Thomas et al. 1999; Lee et al. 2006). The use of sterile sediment and water removes the pressure on survival induced by the competition with and predation by, naturally occurring organisms (Craig et al. 2004). In this study, FIB persistence was determined using intact non-sterile sediments, therefore retaining the effect of natural flora. This shows that predation such as protozoan and viruses may contribute greatly to FIB die-off (Laliberte and Grimes 1982; Davies et al. 1995).

FIB may also not effectively compete with native microflora for available nutrients.

Long term persistence in sediments could indicate that some growth took place but

FIB were eliminated at a faster rate than growth occurred.

The survival curves of TC, E. coli and ENT were not significantly different between continuous-flow and batch microcosms. The main difference, among the observed environmental factors, between continuous-flow and batch microcosms was conductivity measured during the whole experimentation period, in the water column. In the continuous-flow microcosms, conductivity stayed constant with an average of 285 μS cm -1, similar to the values measured in lake water, but in the batch microcosms, it gradually increased with time certainly due to the remobilization of

170 dissolved salts from sediments to the water column, from organic matter degradation, and also due to water evaporation. The variation of conductivity in the batch microcosms seems to have no impact on FIB survival.

The survival of TC, E. coli and ENT has been discussed by Noble et al. (2003).

In our experiments, after 50 days TC and E. coli were not detected anymore. ENT die- off was also quite significant but they were still measured at day 50 at concentrations between 10 2 and 10 3 CFU 100 g -1. ENT survived longer than E. coli and TC in all types of sediments. Hanes and Fragala (1967) also found that E. coli degraded more rapidly with increased sunlight intensity than did ENT, a result that was recently confirmed for bacterial samples from Southern California (Noble et al. 2003).

Results of this study revealed that TC, E. coli and ENT remained cultivable for at least 40 to 50 days and thus were detectable through the culture-based method used here. Many enterobacteria e.g. Vibrio sp ., E. coli , Enterococcus faecalis, can activate survival strategies including the viable but nonculturable state (bacteria still viable but cannot be shown as colony forming units by the conventional plate counts), in response to unfavourable growth conditions and starvation (Lleò et al. 2005). They persist in the environment in conserving their viability despite the loss of their own culturability.

All these results suggest that indicator organisms released into the coastal environment can accumulate in sediment, leading to increased persistence. These findings lead to the question of whether pathogenic bacteria are also capable of extended survival in this environment. More work is needed to correlate the presence and survival of E. coli and other faecal indicator bacteria with that of relevant pathogens in order to assess the need for the use of additional indicators

(Tallon et al. 2005).

171

6.5 Conclusion Results of this experiment confirm extended FIB survival in freshwater sediments, up to 50 days. ENT survived longer than TC and E. coli ; they were still present at concentrations between 10 2 and 10 3 CFU 100 g -1 after 50 days. This study also reveals that FIB persistence is influenced by nutrients and organic matter content in sediments. An increased growth and significantly lower decay rates were observed in sediments containing higher levels of organic matter and nutrients, and smaller grain size. However, FIB followed a quite similar trend in all microcosms. All types of sediments were able to support faecal bacteria limited growth and significant persistence. These results demonstrate that indicator bacteria prove to be capable of utilizing nutrients present in sediments from areas polluted by sewage discharges as well as from areas free of pollution. Their presence and survival in coastal sediments may induce an increased risk of human infection due to the possible resuspension of other pathogenic microorganisms during natural turbulence or human activities. Extended survival of enteric bacteria in sediments and potential remobilization of pathogens may be responsible for water quality failures and are of considerable significance for the management of risk at specific recreational coastal sites.

172

References

Alm EW, Burke J & Spain A (2003) Fecal indicator bacteria are abundant in wet sand

and freshwater beaches. Water Research , 37 : 3978-3982 .

An Y, Kampbell DH & Breidenbach GP (2002) Escherichia coli and total coliforms in

water and sediments at lake marinas. Environmental Pollution , 120 : 771-778 .

Anderson LK, Whitlock JE & Harwood VJ (2005) Persistence and differential survival

of fecal indicator bacteria in subtropical waters and sediments. Applied &

Environmental Microbiology , 71 : 3041-3048 .

Ashbolt NJ, Grohmann GS & Kueh C (1993) Significance of specific bacterial

pathogens in the assessment of polluted receiving waters to Sydney. Water

Science & Technology , 27 : 449-452.

Balkwill DL & Ghiorse WC (1985) Characterization of subsurface bacteria associated

with two shallow aquifers in Oklahoma. Applied & Environmental Microbiology ,

50 : 580-588 .

Burton G, Gunnison D & Lanza G (1987) Survival of pathogenic bacteria in various

freshwater sediments. Applied & Environmental Microbiology , 53 : 633-638 .

Burrus D, Thomas RL, Dominik B, Vernet JP & Dominik J (1990) Characteristics of

suspended sediment in the Upper Rhone River, Switzerland, including the

particulate forms of phosphorus. Hydrological Processes , 4: 85-98 .

Craig DL, Fallowfield HJ & Cromar NJ (2004) Use of microcosms to determine

persistence of Escerichia coli in recreational coastal water and sediment and

validation with in situ measurements. Journal & Applied Microbiology , 96 : 922-

930.

Davies CM & Evison LM (1991) Sunlight and the survival of enteric bacteria in

natural waters. Journal of Applied Bacteriology , 70 : 265-274.

173

Davies CM, Long JAH, Donald M & Ashbolt N (1995) Survival of fecal

microorganisms in marine and freshwater sediments. Applied & Environmental

Microbiology , 61 : 1888-1896.

Davies C & Bavor H (2000) The fate of storm-water associated bacteria in constructed

wetland and water pollution control pond systems. Journal of Applied

Microbiology , 89 : 349-360.

EU (2006) European Directive 2006/7/CE of the European Parliament and of the

Council of 15 February 2006 concerning the management of bathing water

quality and repealing Directive 76/160/EEC.

Evanson M & Ambrose RF (2006) Sources and growth dynamics of fecal indicator

bacteria in a coastal wetland system and potential impacts to adjacent waters.

Water Research , 40 : 475-486.

Fish JT & Pettibone GW (1995) Influence of freshwater sediment on the survival of

Escherichia coli and Salmonella sp . as measured by three methods of enumeration.

Letters in Applied Microbiology , 20 : 277-281.

Gerba C & McLeod JS (1976) Effect of sediment on the Survival of Escherichia coli in

Marine Water. Applied & Environmental Microbiology , 32 : 114-120.

Ghoul M, Bernard T & Cormier M (1990) Evidence that Escherichia coli Accumulates

Glycine Betaine from Marine Sediments. Applied & Environmental Microbiology ,

56 : 551-554.

Goldscheider N, Haller L, Poté J, Wildi W & Zopfi J, (2007) Characterizing water

circulation and contaminant transport in Lake Geneva using bacteriophage

tracer experiments and limnological methods. Environmental Science &

Technology , 41 : 5252--5258.

Hanes NB & Fragala C (1967) Effect of seawater concentration on the survival of

indicator bacteria. Journal Water Pollution Control Federation , 39-97.

Hughes KA (2003) Influence of Seasonal Environmental Variables on the Distribution

of Presumptive Fecal Coliforms around an Antarctic Research Station. Applied

& Environmental Microbiology , 69 : 4884-4891.

174

Kjeldhal J (1883) A new method for the determination of nitrogen in organic matter.

Zeitschrift fur Analytische Chemie , 22 : 366.

LaLiberte P & Grimes DJ (1982) Survival of Escherichia coli in Lake Bottom sediment.

Applied and Environmental Microbiology , 43 : 623-628.

Lee CM, Lin TY, Lin C, Kohbodi GA, Bhatt A, Lee R & Jay JA (2006) Persistence of

fecal indicator bacteria in Santa Monica Bay beach sediments. Water Research ,

40 : 2593-2602.

Lleò MM, Bonato B, Benedetti D & Canepari P (2005) Survival of enterococcal species

in aquatic environments. FEMS Microbiology Ecology , 54 : 189-196.

Loizeau J-L, Arbouillle D, Santiago S & Vernet J-P (1994) Evaluation of a wide range

laser diffraction grain size analyser for use with sediments. Sedimentology , 41 :

353-361.

Loizeau J-L, Pardos M, Monna F, Peytremann C, Haller L & Dominik J (2004) The

impact of a sewage treatment plant’s effluent on sediment quality in a small bay

in Lake Geneva (Switzerland-France). Part 2 : Temporal evolution of heavy

metals. Lakes & Reservoirs: Research & Management 9, 1: 53-63.

McFeters GA & Singh A (1991) Effects of aquatic environmental stress on enteric

bacteria. Journal of Applied Bacteriology , 66 : 559-569.

Noble RT, Moore DF, Leecaster MK, McGee CD & Welsberg SB (2003) Comparison of

total coliform, fecal coliform, and enterococcus bacterial indicator response for

ocean recreational water quality testing. Water Research , 37 : 1637-1643.

OHyg (Ordonnance du DFI sur l’hygiène) (2005) Ordonnance sur les denrées

alimentaires et les objets usuels (ODIOUS). Le Département fédéral de l’intérieur

(DFI), Berne, Switzerland .

Thomas C, Hill DJ & 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 : 1024-1032.

175

Poté J, Haller L, Loizeau J- L, Garcia Bravo A, Sastre V & Wildi W (2008) Effects of a

sewage treatment plant outlet pipe extension on the distribution of

contaminants in the sediments of the Bay of Vidy, Lake Geneva, Switzerland.

Bioresource Technology , 99 : 7122–7131.

Poté J, Goldscheider N, Haller L, Zopfi J, Khajehnouri F & Wildi W (2009) Origin and

spatial-temporal distribution of fecal bacteria in a bay of Lake Geneva,

Switzerland. Environmental Monitoring Assessment , 154 : 337-348

Sinton L, Finlay R & Lynch P (1999) Sunlight inactivation of fecal bacteriophages and

bacteria in sewage-polluted seawater. Applied & Environmental Microbiology , 66 :

230-237.

Tallon P, Magajna B, Lofranco C & Leung KT (2005) Microbial indicators of faecal

contamination in water: a current perspective. Water, Air & Soil pollution , 166 :

139-166.

U.S. Environmental Protection Agency (2000) Improved enumeration methods for

the recreational water quality indicators : enterococci and Escherichia coli EPA-

821/R-97/004. U.S. Environmental Protection Agency, Washington, D.C.

Wildi W, Dominik J, Loizeau JL, Thomas RL, Favarger P-Y, Haller L, Perroud A &

Peytremann C (2004) River, reservoir and lake sediment contamination by

heavy metals downstream from urban areas of Switzerland. Lakes & Reservoirs:

Research & Management , 9: 75-87.

Williams JDH, Jaquet JM & Thomas RL (1976) Forms of phosphorus in the superficial

sediments of Lake Erie. Journal of the Fisheries Research Board of Canada , 33 : 413-

429.

176

CHAPTER 7

Conclusions and perspectives

177

7.1 Conclusions The main objective of this interdisciplinary research was to assess and to characterize the chemical and microbiological contamination of a freshwater site, the

Bay of Vidy (Lake Geneva, Switzerland) due mainly to the release of a sewage treatment plant (WWTP), the discharge of a little river and diffuse agricultural and domestic runoffs.

Wastewaters released into drinking water resources and recreational waters are always problematic and should be carefully monitored in order to minimize the risks for human health and the ecosystem itself. The main conclusions of this survey are summarized below.

7.1.1 Distribution of the chemical contamination -due to heavy metals and some organic pollutants- in the sediments of the Bay of Vidy.

In 2001, the municipality of Lausanne extended the Vidy WWTP outlet pipe as a measure to reduce bacterial water pollution and sediment contamination close to the lake beaches. Our results demonstrate that, the surface area of highly contaminated sediments was reduced by one third after the pipe extension. An improvement in the quality of sediments in 2005, compared to 1996, has been observed. Indeed, metal inventories in the surface sediments decreased by a factor between 11 for Zn and 2.2 for Cu in comparison with the pre-2001 outlet location.

Chromium, which was already low before the modification of the outlet pipe, decreased by a factor of 1.3. The only metal that still presented a very high contamination factor was Hg. Concentrations of organic pollutants such as HMs and

HOCs were also still quite high.

The comparison of metal concentrations with “probable effect concentrations”

(PEC), indicates possible toxic effects of sediment contamination for the biota close to the WWTP outlet. The sediment concentration values from the most polluted point in the Bay of Vidy were 3 to 13 times higher than the PELs (probable effect levels) from

178 the Canadian Sediment Quality Guidelines for the Protection of Aquatic Life (CCME

EPC-98E, 1999).

Although the WWTP outlet pipe extension in 2001 improved the situation with respect to the accumulation of contaminated sediments, the Bay of Vidy is confirmed to be the most contaminated region of Lake Geneva. The original goal of extending the WWTP outlet pipe, to significantly reduce contamination in the coastal area of Vidy Bay, has not been entirely reached. Pollution reduction at the source will be necessary in order to obtain a further improvement.

7.1.2 Identification and comparison of bacterial communities living in polluted sediments of the Bay of Vidy with non-polluted sediments, using molecular approaches.

The aim of that study was to characterize and compare bacterial and archaeal communities in contaminated (Vidy Bay) with uncontaminated (Ouchy area) sediments of Lake Geneva, using 16S rRNA gene diversity.

Results of this survey show that the sediments of the two study sites differed clearly in their organic matter and nutrient contents. High contents of fresh organic matter and nutrients in the Bay of Vidy led to intense mineralization, which was dominated by sulphate-reduction and methanogenis, as confirmed by the microbial analysis.

Phylogenetic analysis of sedimentary prokaryotes revealed that:

(i) Archaeal and bacterial communities differed considerably between the contaminated and the non-contaminated sediments.

(ii) Betaproteobacteria was the dominant bacterial group, representing more than 30% of analysed clones in surface sediments at both sites, which is consistent with previous studies in freshwater environments. A large proportion of

Betaproteobacteria clones, mostly from Vidy sediments, were related to the reductively dechlorinating Dechloromonas sp .

179

(iii) Consistent with geochemical data, Deltaproteobacteria including clones related to iron- (Geobacter sp .) and sulphate-reducing bacteria, were more abundant in the contaminated sediments.

(iv) The archaeal communities were dominated by methanogenic

Euryarchaeota , particularly in the organic matter-rich sediments from Vidy Bay.

This study suggests that each site harbors a specific sediment microbial community. The apparent lower bacterial diversity in Vidy sediments may be explained by the significant concentrations of a variety of contaminants which may have adverse biological effects on benthic metazoa and microbes. A correlation was observed between the microbial composition and the environmental variables at the two sites.This research implies therefore that, in addition to environmental variables, such as nutrients and organic matter contents, pollution could be one of the factors affecting microbial community structure. However, given the long history of pollution in the bay, different bacterial and archaeal communities may have adapted to these particular conditions.

7.1.3 Spatial and seasonal distribution of faecal bacteria (FIB) contamination in the water column of the Bay of Vidy, using standard cultivation techniques.

During one year, the input flux rates of faecal indicator bacteria from the 2 major contamination sources in the Bay of Vidy (the WWTP outlet pipe and the

Chamberonne River), have been determined and their spatial-temporal distribution was assessed

The WWTP outlet pipe at 30 m depth is quantitatively the most important source of faecal bacteria, whereas the FIB input flux rates from the Chamberonne

River were typically 1 to 3 orders of magnitude lower. E. coli input flux rates from the

WWTP could reach up to 2.5 x 10 10 CFU/s. Different pathogenic Salmonella serotypes were identified in selected samples from these sources.

180

FIB levels in the bay were spatially and temporally variable. The results demonstrate that:

(i) The WWTP outlet at 35 m depth impacts near-surface lake water quality mainly during holomixis in winter.

(ii) When the lake is stratified, the effluent water is generally trapped below the thermocline.

(iii) During major floods, upwelling across the thermocline may occur.

(iv) The river permanently contributes to contamination, mainly near the river mouth and during floods, when the storm water outlet contributes additionally.

(v) The lowest FIB levels, in the near-surface water, occur during low-flow periods in the bathing season.

The highest FIB concentrations in the near-surface water of the bay consequently occur during floods and during mixed lake conditions. The most favourable situation happens when the lake is stratified and the flow rates of the river and WWTP are low, i.e. during the bathing season in summer . The first warm and sunny days after a major summer storm represent the most critical situation for bathers.

Improving the capacity and the microbiological treatment of the WWTP is necessary in order to obtain a better water quality of this very popular recreational site.

7.1.4 Spatial distribution of faecal bacteria (FIB) contamination in the sediments of the Bay of Vidy, using standard cultivation techniques.

The results of this study revealed that faecal indicator bacteria were present almost everywhere in the sediments of the Bay of Vidy.

E. coli and ENT were particularly abundant near the WWTP outlet discharge and the mouth of the Chamberonne River, where concentrations between 10 5 and 10 7

CFU 100 g -1 were detected. FIB levels at 10 cm depth in the sediments near the

181

WWTP outlet pipe ranged between 10 4 and 10 5 CFU 100 g -1, and were still detected in the top 6 cm after 90 days.

Results of this study indicate that freshwater sediments of the bay of Vidy constitute a reservoir of faecal indicator bacteria, which can persist in certain areas of the bay. Possible resuspension of FIB and pathogens may affect water quality and may increase health risks to sensitive populations during recreational activities.

7.1.5 Importance of faecal bacteria survival and growth in sediments, according to different environmental parameters, such as nutrients and organic matter .

Extended persistence of enteric bacteria in coastal sediments and potential remobilization of pathogens during natural turbulence or human activities may induce an increased risk of human infections. Results of this study revealed the following:

(i) Extended FIB survival in sediments, up to 50 days.

(ii) FIB persistence is influenced by nutrients and organic matter content in sediments. An increased growth and significantly lower decay rates were observed in sediments containing higher levels of organic matter and nutrients, and smaller grain size.

(iii) Some limited growth and significant persistence was observed even in sediments poor in nutrients and organic matter. Indicator bacteria, therefore, prove to be capable of utilizing nutrients present in sediments from areas polluted by sewage discharges as well as from areas free of pollution.

(iv) The longer survival of Enterococcus sp. compared to Escherichia coli and total coliforms.

Extended survival of enteric bacteria in sediments and potential remobilization of pathogens may be responsible for water quality failures and are of

182 considerable significance for the management of risk at specific recreational coastal sites, such as the Bay of Vidy.

7.2 Perspectives Our understanding of freshwater ecosystems in general has increased tremendously in the last decades as well as our knowledge of the microbial diversity in freshwater habitats. However, it is still difficult to assess the effects of pollution on water resources, due to pathogens, nutrients, toxic inorganic and organic chemicals, antibiotics, endocrine disruptors and the thousands of different natural or synthetic substances. These pollutants constitute major threats to human health and aquatic environments, and remain active research fields in every parts of the globe.

Our survey reports on the spatial distribution and the high contamination in organic matter, faecal indicator bacteria, heavy metals, and hydrophobic organic compounds of the sediments of the Bay of Vidy, in Lake Geneva. It suggests as well that the diversity and composition of indigenous bacteria living in the sediments of

Vidy have been affected by pollution, mainly released from the WWTP. Results from this survey will serve as background information for future research in the area.

Indeed, many questions still remain open, in particular:

(i) Some of the major bacteria present in Vidy and Ouchy sediments have been identified, but nothing is known about their ecological functions. The lack of genomic insight into the metabolic capacities of many bacteria is one of the reasons for the very limited functional insight into bacterial communities of freshwater habitats in general (Hahn, 2006). Detailed knowledge on the specific functions and ecology of microorganisms inhabiting freshwater ecosystems is urgently needed.

183

The change in environmental factors such as organic matter, nutrients, toxic micropollutants, can influence the microbial community structure and their functions.

(ii) Therefore, better understanding the interactions between microorganisms and heavy metals has also to be addressed by future investigations. Do bacteria develop a certain resistance to heavy metals and do they play a role into the possible remobilization of these contaminants from the sediments back into the water column?

(iii) The effects of antibiotics on indigenous bacteria are also a big concern all over the world, especially in areas contaminated by WWTP effluents. Many studies report on the contamination of natural aquatic systems with antibiotics. A previous survey on resistance to antibiotics on several water courses and lakes in Switzerland has shown interesting results («Rôle des résidus d’antibiotiques dans l’environnement hydrique sur la selection et la diffusion de bactéries résistantes des genres Aeromonas,

Acinetobacter et Legionella», A.R. Corvaglia, Thèse no 3796, Université de Genève,

2006). Do indigenous bacteria become resistant to these substances? What are the mechanisms involved?

One of the modules of a current project funded by the Swiss science foundation (“ProDoc 21”) focuses on the Bay of Vidy and has the following objectives: (1) assessing the major functions of the bacterial biomass living on the interface sediment-water; (2) determining the heavy metal and antibiotic resistance patterns of bacteria; (3) determining the effects of bacterial activities on the release of heavy metals from contaminated sediments to the water column; (4) assessing the effects of sediment toxicity on the microbial community.

Future results will be of considerable significance for the sustainable management of specific recreational coastal sites and of freshwater resources in general.

184

Remerciements / Acknowledgements

Je tiens tout d’abord à remercier mon directeur et mon co-directeur de thèse,

Professeur Walter Wildi et Professeur Raffaele Peduzzi, pour m’avoir offert l’opportunité d’effectuer cette recherche dans d’excellentes conditions et pour m’avoir accueillie dans leurs instituts respectifs. Ils ont su me guider avec gentillesse et enthousiasme pour mener à bien ce travail et je leur en suis très reconnaissante.

Un merci tout particulier à John Poté qui m’a suivie tout au long de ce travail et qui a su me conseiller et m’encourager durant ces quatre années de thèse. Ses connaissances scientifiques, ses conseils pertinents et sa bonne humeur ont été d’un précieux soutien. Un grand merci à Mauro Tonolla pour avoir accepté de codiriger ce travail de thèse et pour avoir suivi tous mes travaux. Je remercie également Nadia

Ruggeri-Bernardi, Anna-Paola Caminada, Michele Bottinelli, Antonella Demarta et toute l’équipe de l’institut cantonal de microbiologie à Bellinzona, pour m’avoir chaleureusement accueillie dans leur laboratoire et pour m’avoir initiée au monde fascinant de la biologie moléculaire.

J’adresse mes sincères remerciements à Jakob Zopfi pour m’avoir aidée à effectuer les analyses géochimiques à l’Université de Neuchâtel et pour tout le temps consacré à la lecture des articles publiés dans le cadre de cette thèse. Merci infiniment

Jakob pour tes précieux conseils.

Je remercie également le Professeur Pascal Simonet d’avoir accepté de faire parti de mon jury de thèse et d’avoir bien voulu contribuer à la critique du manuscrit.

185

Un grand merci à Jean-Luc Loizeau pour m’avoir aidée à l’échantillonnage, l’ouverture de carottes, la datation au césium et pour ses conseils en statistiques.

Merci Jean-Luc pour avoir gentiment pris le temps de relire certains articles.

Je remercie Pierre Rossi et Noam Shani de l’EPFL pour leur aide concernant certaines analyses statistiques.

Merci également à tous les amis et collègues de l’institut Forel qui m’ont aidé de près ou de loin aux analyses de laboratoire et à l’échantillonnage, en particulier

Philippe Arpagaus et Vincent Sastre pour avoir piloter le bateau par tous les temps afin de récolter nos échantillons, Françoise, Pierre-Yves, Prof. Janusz Dominik,

Alexandra, Essofli, Stéphanie, Andrea, Claudia, Florian, Stéphane, Frédéric, Bian,

Lucie, Serge, Anh-Dao, Niel, Davide, Benoît, Marion, Régis, Vincent C., Thomas,

Raphaël, ainsi que tous les diplômants avec lesquels j’ai collaboré.

Je remercie enfin du fond du cœur toute ma famille, Andrew Glass, mon p’tit bout de chou Liam, ma sœur et sa famille, Cécile, Dominique, Zoe et Killian Chariatte et bien sûr mes parents, Louis et Isabelle Haller, pour leur irremplaçable et indispensable soutien et grâce à qui j’ai pu mener à terme ce travail.

186