Effect-Directed Analysis for Identification of River Basin Specific Pollutants

Effect-directed analysis for identification of river basin specific pollutants.

Ph.D. Dissertation

Zuzana Toušová

Masaryk University, Faculty of Science,

Centre RECETOX

Brno, Czech Republic 2020

Supervisor: Dr. Ing. Jaroslav Slobodník

BIBLIOGRAPHIC ENTRY

Author: Mgr. Zuzana Toušová Faculty of Science, Masaryk University Centre RECETOX

Title of Dissertation: Effect-directed analysis for identification of river basin specific pollutants

Degree Programme: Environmental Health Sciences

Specialization: Environmental Chemistry and Toxicology

Supervisor: Dr. Ing. Jaroslav Slobodník

Supervisor specialist: prof. RNDr. Luděk Bláha, Ph.D.

Year: 2020

Keywords: contaminants of emerging concern, effect directed analysis, in-vitro bioassays, prioritization, non- target screening.

BIBLIOGRAFICKÝ ZÁZNAM

Autor: Mgr. Zuzana Toušová Přírodovědecká fakulta, Masarykova univerzita Centrum RECETOX

Název práce: Užití efektem řízené analýzy pro identifikaci specifických znečišťujících látek v povodích

Studijní program: Životní prostředí a zdraví

Specializace: Environmentální chemie a toxikologie

Školitel: Dr. Ing. Jaroslav Slobodník

Školitel specialista: prof. RNDr. Luděk Bláha, Ph.D.

Rok: 2020

Klíčová slova: nově se objevující znečišťující látky, efektem řízená analýza, biotesty in-vitro, prioritizace, necílená analýza.

ACKNOWLEDGEMENTS

I would like to thank my supervisor Dr.. Jaroslav Slobodník for his great guidance, encouragement, support and patience during my postgraduate study. My deep gratitude also goes to prof. RNDr. Luděk Bláha, Ph.D., who carefully co- supervised my work, inspired me, provided me with great scientific training and contributed to the direction and richness of this research. Prof. Bláha was always ready to give advice and discuss immediate problems, which helped me greatly to progress smoothly with both laboratory work and writing papers. I am also very grateful to doc. Mgr. Klára Hilscherová, PhD., who largely contributed to the study design and results’ interpretation in publications II and III. Doc. Hilscherová kindly shared her rich practical experience with me and devoted a lot of her time to help me finalize the publications. My thanks also go to all co-authors of our common publications, for their great work on the research and writing papers. Namely, I would like to thank Dr. Tobias Schulze and doc. Ing. Branislav Vrana, PhD. for their great work on the sampling part of this research. I gratefully acknowledge Dr. Peter Oswald and Dr. Natalia Glowacka., colleagues from the EI, who contributed on the analytical part and helped me with lots of practical issues. I extend my thanks to all members of the EDA EMERGE Project for all those inspirational discussions, training courses, hospitality and fun. My thanks go to all RECETOX colleagues for making such a friendly environment at the workplace and being helpful at all times. I express my deep gratitude to all my family. I would like to thank my parents and grandparents for their huge support throughout my studies. Great thanks go to my husband for his encouragement, support, patience and time with the kids. I am very grateful to my mother and especially to my late mother- in-law for being such good grandmas and all the baby-sitting, without which this thesis would never had taken shape.

The PhD training programme was funded by the European Union under the Marie Curie Actions—Initial Training Networks, FP7-PEOPLE-2011-ITN, EDA-EMERGE project, Grant Agreement No. 290100.

ABSTRACT

Pollution of surface waters with contaminants of emerging concern (CECs) has become a pressing environmental issue of global importance. CECs is a group of organic compounds of diverse structure, physicochemical properties and usage patterns comprising pesticides, pharmaceuticals, personal care products, flame retardants and others. CECs can be found in surface and waste waters, they are not monitored or regulated and raise concerns of the research community and public due to their extensive occurrence and possible adverse effects on biota, which are mostly unknown. Innovative approaches to monitor surface waters, with regard to CECs, have been proposed. These include novel sampling techniques, effect assessment with sensitive bioassays, non-target chemical screening and use of advanced software tools for data interpretation. Further, a powerful technique to identify bioactive compounds in complex environmental mixtures, known as effect-directed analysis (EDA), is based on the combination of bioassays, fractionation and chemical analysis. The objective of this thesis was to contribute to the identification of relevant toxicity drivers in surface waters by advancing the EDA utilization. The specific objectives, addressed by particular studies, were to apply a newly developed simplified EDA protocol (study I and IV), identify compounds responsible for algal growth and acetylcholine esterase inhibition in a WWTP effluent extract (study II and V) and characterize contamination of the Bosna River with CECs (study III). Novel sampling device, based on solid phase extraction of large water volumes (LVSPE), was tested and optimized in a sampling campaign spanning across four European river basins (Studies I and IV). In the effect assessment of extracts, the most pronounced effects were estrogenicity, toxicity to algae and fish embryo toxicity, whereas the major portions of the observed effects could not be explained by analyzed compounds. Simplified risk assessment procedure with analyzed compounds enabled identification and prioritization of 21 candidate compounds for future monitoring efforts. Effect directed analysis was applied to identify compounds contained in a wastewater treatment plant (WWTP) effluent extract that were responsible for growth inhibition of green algae (Study II) and acetylcholinesterase inhibition in vitro (Study V). Our results suggest that pesticides and their transformation products, pharmaceuticals (barbiturate derivatives and macrolide antibiotics e.g. azithromycin), industrial compounds or caffeine and its metabolites were the most likely toxicity drivers for green algae in study II. 7

A combined chemical and effect screening of water quality in the River Bosna was carried out, to characterize, in some detail, its pollution with CECs (Study III). The assessment of cumulative pollutant concentrations and hazard profiles enabled to identify the major source of contamination with CECs and hotspots of biological potency. Simplified risk assessment procedure of detected target compounds suggested that 7 compounds, namely diazinon, diclofenac, 17-β estradiol, estrone, benzo[k]fluoranthene, fluoranthene and benzo[k]fluoranthene, might pose serious risks to aquatic biota in the Bosna River. The presented dissertation demonstrates that combination of novel sampling techniques with effect-based methods and chemical screening is a functional strategy to address CECs in surface waters. This thesis shows that EDA is a useful tool to reduce the complexity of environmental mixtures and to identify relevant toxicity drivers. Risk assessment and prioritization of detected chemicals is a crucial step in the process of identifying river basin specific pollutants.

ABSTRAKT

Znečištění povrchových vod nově se objevujícími znečišťujícími látkami (Contaminants of emerging concern - CECs) se stalo vážným environmentálním problémem celosvětového rozsahu. CECs jsou velmi rozmanitou skupinou organických látek, co se týče struktury, fyzikálně chemických vlastností a oblastí jejich využití. Mezi CECs řadíme například pesticidy, farmaka, výrobky pro osobní hygienu, zpomalovače hoření a další. CECs se vyskytují v odpadních a povrchových vodách, přičemž nejsou pravidelně monitorovány ani nijak regulovány. Díky častému výskytu a možným nežádoucím účinkům na živé organismy vyvolává skupina CECs znepokojení veřejnosti a velkou pozornost vědecké komunity. Nově navrhované přístupy k monitoringu povrchových vod s ohledem na CECs proto zahrnují speciální vzorkovací techniky, sledování biologických účinků citlivými biotesty, necílený chemický screening a užití pokročilých softwarových nástrojů k interpretaci získaných dat. Dále také efektem řízená analýza (EDA – effect directed analysis), založená na kombinaci biotestů, frakcionace, a chemické analýzy, je účinnou metodou k identifikaci látek zodpovědných za pozorované biologické účinky v komplexních environmentálních směsích. Cílem této disertace bylo přispět k identifikaci látek zodpovědných za toxické účinky sledované v povrchových vodách užitím a dalším vývojem metody EDA. Dílčími cíli pak byly i) využití zjednodušeného protokolu pro efektem řízenou analýzu (studie I a IV); ii) identifikace inhibitorů růstu zelených řas a acetylcholinesterázy v extraktu přečištěné odpadní vody (studie II a V) a iii) charakterizace znečištění řeky Bosny nově se objevujícími znečišťujícími látkami V rámci studie I a IV proběhla vzorkovací kampaň ve čtyřech povodích Evropských řek, přičemž bylo testováno a optimalizováno nové vzorkovací zařízení založené na extrakci velkého objemu vody na pevné fázi. Nejčastěji sledovaným biologickým efektem byla estrogenita, dále toxicita pro zelené řasy a rybí embrya. Převážnou část těchto efektů nebylo možné vysvětlit výskytem cílových analytů. Na základě zjednodušené analýzy rizik cílových analytů bylo identifikováno a prioritizováno 21 látek zajímavých pro budoucí monitoring povrchových vod. K identifikací látek způsobujících toxicitu pro zelené řasy (Studie II) a inhibici acetylcholinesterázy in vitro (studie V) byla využita efektem řízená analýza extraktu přečištěné odpadní vody. Výsledky studie II naznačují, že pro inhibici růstu zelených řas jsou určující pesticidy a produkty jejich přeměny, 9 některá farmaka (deriváty barbiturátů, makrolidy), průmyslové látky a kofein včetně jeho metabolitů. Znečištění řeky Bosny CECs bylo zkoumáno využitím kombinace screeningu biologických účinků a chemických analýz provedené na výluzích pasivních vzorkovačů (Studie III). Hlavní zdroj kontaminace řeky Bosny CECs a ohniska sledované biologické aktivity se podařilo identifikovat porovnáním kumulativních koncentrací cílových analytů a profilů nebezpečnosti. Zjednodušená analýza rizik cílových analytů poukázala na 7 látek (diazinon, diclofenac, 17-β estradiol, estrone, benzo[k]fluoranthene, fluoranthene and benzo[k]fluoranthene), které mohou představovat riziko pro vodní biotu. Tato disertční práce demonstruje, že kombinace nových vzorkovacích technik s metodami k posouzení biologických účinků a chemickými analýzami je dobrou strategií, jak přistupovat ke znečištění povrchových vod CECs. Tato práce také prokazuje užitečnost metody EDA pro snížení komplexity environmentálních směsí a identifikaci látek způsobujících toxicitu. Analýza rizik a prioritizace detekovaných látek je důležitým krokem v procesu hledání specifických znečišťujících látek povodí.

LIST OF ORIGINAL PUBLICATIONS AND THE AUTHOR’S CONTRIBUTION

Publication I (see ANNEX I) Tousova, Z., Oswald, P., Slobodnik, J., Blaha, L., Muz, M., Hu, M., Brack, W., Krauss, M., Di Paolo, C., Tarcai, Z., Seiler, T.-B., Hollert, H., Koprivica, S., Ahel, M., Schollée, J.E., Hollender, J., Suter, M.J.-F., Hidasi, A.O., Schirmer, K., Sonavane, M., Ait-Aissa, S., Creusot, N., Brion, F., Froment, J., Almeida, A.C., Thomas, K., Tollefsen, K.E., Tufi, S., Ouyang, X., Leonards, P., Lamoree, M., Torrens, V.O., Kolkman, A., Schriks, M., Spirhanzlova, P., Tindall, A., Schulze, T., 2017. European demonstration program on the effect-based and chemical identification and monitoring of organic pollutants in European surface waters. Sci. Total Environ. 601–602. https://doi.org/10.1016/j.scitotenv.2017.06.032

Zuzana Toušová participated in the planning of the study, conducted designated parts of the study (sampling in Danube RB, sample processing, algal bioassays), evaluated and interpreted relevant results (linking of biological effects and detected compounds, risk assessment and prioritization) and prepared the manuscript.

Publication II (see ANNEX II) Tousova, Z., Froment, J., Oswald, P., Slobodník, J., Hilscherova, K., Thomas, K.V., Tollefsen, K.E., Reid, M., Langford, K., Blaha, L., 2018. Identification of algal growth inhibitors in treated waste water using effect-directed analysis based on non-target screening techniques. J. Hazard. Mater. 358. https://doi.org/10.1016/j.jhazmat.2018.05.031

Zuzana Toušová participated in the planning of the study, conducted designated parts of the study (sampling, sample processing, algal bioassays, fractionation, toxicity data search in literature and databases, ECOSAR modelling), evaluated and interpreted relevant results and prepared the manuscript.

Publication III (see ANNEX III) Toušová, Z., Vrana, B., Smutná, M., Novák, J., Klučárová, V., Grabic, R., Slobodník, J., Giesy, J.P., Hilscherová, K., 2019. Analytical and bioanalytical assessments of organic micropollutants in the Bosna River using a combination of passive sampling, bioassays and multi-residue analysis. Sci. Total Environ. 650. https://doi.org/10.1016/j.scitotenv.2018.08.336 11

Zuzana Toušová participated in the planning of the study, conducted designated parts of the study (sample processing, in vitro bioassays, literature and databases, ECOSAR modelling), evaluated and interpreted relevant results (linking of biological effects and detected compounds, contamination profiling and hazard assessment), and prepared the manuscript.

Publication IV (see ANNEX IV) Schulze, T., Ahel, M., Ahlheim, J., Aït-Aïssa, S., Brion, F., Di Paolo, C., Froment, J., Hidasi, A.O., Hollender, J., Hollert, H., Hu, M., Kloß, A., Koprivica, S., Krauss, M., Muz, M., Oswald, P., Petre, M., Schollée, J.E., Seiler, T.-B., Shao, Y., Slobodnik, J., Sonavane, M., Suter, M.J.-F., Tollefsen, K.E., Tousova, Z., Walz, K.-H., Brack, W., 2017. Assessment of a novel device for onsite integrative large-volume solid phase extraction of water samples to enable a comprehensive chemical and effect-based analysis. Sci. Total Environ. 581–582. https://doi.org/10.1016/j.scitotenv.2016.12.140

Zuzana Toušová participated in the planning of the study, conducted designated parts of the study (sampling in Danube RB, sample processing, algal bioassay) and contributed to writing of the manuscript.

Publication V (see ANNEX V) Ouyang, X., Leonards, P.E.G., Tousova, Z., Slobodnik, J., De Boer, J., Lamoree, M.H., 2016. Rapid Screening of Acetylcholinesterase Inhibitors by Effect- Directed Analysis Using LC × LC Fractionation, a High Throughput in Vitro Assay, and Parallel Identification by Time of Flight Mass Spectrometry. Anal. Chem. 88. https://doi.org/10.1021/acs.analchem.5b04311

Zuzana Toušová carried out the sampling and sample processing, and contributed to writing of the manuscript. 12

LIST OF ABBREVIATIONS

AR androgen receptor AChE acetylcholinesterase BEQ bioanalytical equivalent concentration CA concentration addition CECs contaminants of emerging concern CI contamination index CUPs currently used pesticides [1-(3,4-dichlorophenyl)-3-methylurea], diuron DCPMU transformation product Dex-EQ dexamethasone equivalent DHT dihydrotestosterone DHT-EQ dihydrotestosterone equivalent DMSO dimethylsulfoxid E1 estrone E2 17ß-estradiol E2-EQ 17-ß-estradiol equivalent E3 estriol effect concentration 50 - concentration at which the effect EC50 reaches 50% of the effect in untreated control EDA effect directed analysis EDP European demonstration program EE2 17a-ethinylestradiol EQS environmental quality standard ER estrogen receptor EU European union Flu-EQ flutamide equivalent GC gas chromatography (two dimensional) gas chromatography coupled to mass GC(xGC)-MS spectrometry HI hazard index HQ hazard quotient HR-ToF MS high resolution time of flight mass spectrometry IC inhibitory concentration IA independent action LC liquid chromatography liquid chromatography coupled to high resolution mass LC-HRMS spectrometry LOD limit of detection; 13

LOEC lowest observed effect concentration LOQ limit of quantification LVSPE large volume solid phase extraction MEC95 95th percentile of maximum environmental concentration MeOH methanol MTPexp maximum toxicity potential - experimental

MTPpred maximum toxicity potential - predicted with ECOSAR OCPs organochlorine pesticides OH-Tam -EQ hydroxytamoxifen equivalent OPE organophosphate esters PAH polycyclic aromatic hydrocarbons PBDE polybrominated diphenylethers PCBs polychlorinated biphenyls PCPs personal care products PEC predicted environmental concentration PFAS per- and polyfluoroalkyl substances PFOA perfluorooctanoic acid PFOS perfluorooctanesulfonate PNEC predicted no effect concentration POCIS polar organic chemical integrative sampler PPP plant protection product PRC performance reference compound RB river basin RBSP river basin specific pollutants REF relative enrichment factor REP relative effect potency RP-HPLC reverse phase - high performance liquid chromatography RP-SPE reverse phase-solid phase extraction SPMD Semi permeable membrane device 2,3,7,8-Tetrachlorodibenzo-p-dioxin; TCDD-EQ – TCDD TCDD equivalent TU toxic unit UHPLC ultra-high-performance liquid chromatography US United states WFD Water framework directive WW wastewater WWTP wastewater treatment plant

TABLE OF CONTENTS

1 INTRODUCTION ...... 16 1.1 Monitoring of water quality ...... 16 1.2 Contaminants of emerging concern (CECs) ...... 17 Main classes of CECs: ...... 18 1.2.1 Artificial sweeteners ...... 18 1.2.2 Per- and polyfluoroalkyl compounds (PFASs) ...... 19 1.2.3 Pharmaceuticals and steroids ...... 19 1.2.4 Personal care products (PCPs) ...... 20 1.2.5 Brominated and emerging flame retardants ...... 20 1.2.6 Benzotriazoles and benzothiazoles ...... 21 1.2.7 Pesticides ...... 21 1.2.8 Others ...... 21 1.3 Sampling ...... 22 1.3.1 Large volume solid phase extraction (LVSPE) ...... 22 1.3.2 Passive sampling ...... 23 1.4 Chemical analyses and non-target screening ...... 25 1.5 Bioassays ...... 25 1.6 Effect directed analysis (EDA) ...... 26 1.7 Risk assessment (RA) ...... 27 2 Aims of the study ...... 29 3 SUMMARY OF MATERIALS AND METHODS ...... 30 3.1 European demonstration program (EDP) ...... 30 3.1.1 LVSPE sampling ...... 30 3.1.2 Effect assessment ...... 30 3.1.3 Chemical analyses ...... 31 3.1.4 Linking effects and detected compounds ...... 32 3.1.5 Risk assessment and prioritization ...... 32 3.2 Effect directed analysis of WWTP effluent extract ...... 33 15

3.2.1 Fractionation ...... 33 3.2.2 Algal growth inhibition assay ...... 33 3.2.3 Non-target screening ...... 33 3.2.4 Linking effects and detected compounds ...... 33 3.2.5 LCxLC system for identification of AChE inhibitors ...... 34 3.3 Analytical and bioanalytical assessment of the Bosna River ...... 34 3.3.1 Passive sampling ...... 34 3.3.2 Chemical analyses ...... 34 3.3.3 In vitro bioassays ...... 35 3.3.4 Contamination profiling and hazard assessment ...... 35 4 SUMMARY OF RESULTS AND DISCUSSION ...... 36 4.1 European demonstration program (EDP) ...... 36 4.1.1 Sampling ...... 36 4.1.2 Effect assessment ...... 37 4.1.3 Chemical analyses ...... 37 4.1.4 Risk assessment and prioritization ...... 39 4.2 Effect directed analysis of the WWTP effluent extract ...... 41 4.2.1 Toxicity to algae ...... 41 4.2.2 AChE inhibition ...... 43 4.3 Analytical and bioanalytical assessment of the Bosna River ...... 45 5 CONCLUSIONS AND FUTURE PROSPECTS...... 48 6 REFERENCES ...... 50 7 ANNEXES...... 60

1 INTRODUCTION

1.1 Monitoring of water quality Environmental quality monitoring of surface waters is fundamental to the sustainable management of water resources and to reducing risks posed by multiple anthropogenic stressors (Geissen et al., 2015). In spite of the marked improvement of water quality in European surface water bodies, the ultimate goal of the European Union (EU) water policy, i.e. reaching good ecological and chemical status, still remains a challenge in most cases (European Environment Agency, 2018). The current monitoring scheme delineated by the Water Framework Directive (WFD), based on a few priority substances and biological indicators, fails to address a wide range of chemicals occurring in complex environmental mixtures, which may elicit adverse biological effects (Directive 2000/60/EC, Directive 2008/1005/EC, Directive 2013/39/EC). Novel holistic approaches to water quality monitoring with focus on the mixture risks and identification of the main toxicity drivers have been proposed (Fig. 1). These empathize the use of tailored sampling techniques, effect-based methods, chemical suspect and non-target screening followed by linking of the chemical and biological data (Altenburger et al., 2019). Integrative monitoring strategies reflect the complex character of water pollution and aim to bring relevant information to the water management bodies about the deleterious biological effects of chemical contaminants at reasonable cost (Brack et al., 2019).

Figure 1: Scheme of the current and proposed future approaches to surface water quality monitoring under the WFD (adapted from European Environment Agency, 2018). EQS – Environmental quality standard, RBSP – River basin specific pollutant INTRODUCTION 17

1.2 Contaminants of emerging concern (CECs) Chemicals known as contaminants of emerging concern (CECs) comprise a diverse group of compounds including personal care products, human and veterinary pharmaceuticals, surfactants and surfactant-derived compounds, X-ray contrast media, pesticides, disinfection by-products, algal toxins, flame retardants, plasticizers, UV-filters, industrial compounds and transformation products (Sima et al., 2014). These compounds typically occur at rather small (down to sub-ng L-1) concentrations. However, some of these are sufficiently potent or have the potential to be accumulated to concentrations such that they can elicit biological effects. Due to the broad range of physicochemical properties and usage patterns, CECs constitute a major challenge for analytical chemists and environmental regulators (Ginebreda et al., 2014). The main sources of CECs entering surface waters are wastewater treatment plant (WWTP) effluents, agricultural and urban runoff (Fairbairn et al., 2018). The occurrence and fate of CECs in surface waters varies widely, depending on their physicochemical properties, regional usage and efficiency of wastewater treatment. In surface waters, CECs attenuate naturally by biotransformation, photolysis, sorption, volatilization, and dispersion, or a combination there of (Pal et al., 2010). While some CECs have short half-lives e.g. certain pharmaceuticals, others are susceptible to poor removal during conventional wastewater treatment processes e.g. artificial sweeteners (Loos et al., 2013) Eventually, some compounds e.g. bisphenol A, became virtually ubiquitous as they occur in remote areas with no direct anthropogenic impact (Weissinger et al., 2018). Exposure of CECs to aquatic biota is essentially constant, even for those with short half- lives, because the feed from wastewater treatment plants is continuous (Vera-Candioti et al., 2008). CECs are an issue of major concern as they may have negative impact on human health and aqueous ecosystems through various modes of toxic action e.g. endocrine disruption, immunotoxicity, oxidative stress or genotoxicity and embryotoxicity (Connon et al., 2012). These specific biological effects were repeatedly observed for surface waters, drinking waters, raw and treated wastewaters and technical systems (Dizer et al., 2002, Loos et al., 2013; Jálová et al., 2013; Osman et al., 2015; Zhao et al., 2011) and some can be associated with particular compounds as the main toxicity drivers, e.g., estrogenicity due to E2 and EE2 (Aerni et al., 2004; Koenig et al., 2017; Miège et al., 2009) or phytotoxicity to herbicides (Tang and Escher, 2014). However, the cause of many effects observed in bioassays remains unexplained by the chemicals detected in the same samples (Burgess et al., 2013; Neale et al., 2015). 18

CECs occur in highly complex and variable environmental mixtures. Combined effects of individual constituents, phenomenon known as ‘cocktail effect’, present a major challenge, since considering one substance at a time may lead to underestimation of the overall environmental toxicity (Posthuma et al., 2019). Research is moving towards new strategies to evaluate the effects of chemical mixtures. Two main approaches are available to study the toxicity of mixtures: “whole mixture” and “component-based”. The first directly assesses the toxicity of a chemical mixture, considers all the chemicals in the mixture and evaluates the realistic exposure scenario. However, the results cannot be extrapolated to different samples and are only applicable to the investigated mixture. The component-based approach determines the toxicity of the mixture on the basis of the effects of its individual components, following two concepts of potential interaction between chemicals: Concentration Addition (CA) and Independent Action (IA). CA assumes similar activity and a common mode of action for all the compounds in the mixture, while IA implies that the compounds act on different molecular targets that lead to a common toxicological effect (Riva et al., 2019). CECs are largely unregulated, although several CECs were added to the EU WFD list of priority substances for chemical monitoring and three compounds, i.e. diclofenac, 17-beta-estradiol and 17-alpha-ethinylestradiol, have been monitored as a part of the WFD Watch list (Directive 2008/1005/EC, Directive 2013/39/EC). However, the focus on a few routinely monitored compounds encourages reduction in their use, while replacement of these chemicals by alternatives that pose similar hazards is an unresolved problem e.g. atrazine vs. terbuthylazine in the EU (Brack et al., 2019).

Main classes of CECs:

1.2.1 Artificial sweeteners Artificial sweeteners such as sucralose, acesulfame, saccharin, aspartame, etc. are widely used in high concentrations as sugar substitutes in food, drinks and pharmaceuticals (Scheurer et al., 2009). Artificial sweeteners are rarely metabolized and primarily enter the environment through treated wastewater. Due to their stability and high levels found in the environment (units of µg/L in river waters), sucralose and acesulfame are recognized as potential tracers of anthropogenic inputs into environmental waters (Richardson and Ternes, 2018). Artificial sweeteners are not expected to bioaccumulate, but their impacts on the behavior of aquatic organisms and toxicity of sweetener transformation products were shown in several studies. INTRODUCTION 19

Environmentally relevant levels (0.05 and 155 μg/L) of sucralose caused changes in oxidative biomarkers on the the gills, muscle, brain, and liver of carp and also oxidative damage in lipids and proteins (Saucedo-Vence et al., 2017). Increased oxidative stress in the liver of Carassius auratus after the exposure to phototransformation products of acesulfame was reported by Ren et al., (2016).

1.2.2 Per- and polyfluoroalkyl compounds (PFASs) PFASs such as perfluorooctanoic acid (PFOA) and perfluorooctanesulfonate (PFOS) are substances containing a fully fluorinated hydrophobic alkylchain with variable number of carbon atoms. Due to their unusual chemical properties, repelling both grease and water, PFASs are used in food packaging, fabrics (e.g., waterproof jackets), carpets, nonstick cooking pans, paints, adhesives, electronics, personal care products, and firefighting foams. PFASs are highly stable in the environment, can be transported over long distances and accumulate in red blood cells (Richardson and Ternes, 2018). PFASs may enter the aquatic environment directly from production facilities, further due to the use of products with their content, from wastewater treatment plants and furthermore from degradation of long-range- transported volatile precursors (Bečanová et al., 2016). The most common PFASs (PFOA and PFOS) have been regulated in the EU and US as they were identified as persistent, bioaccumulative and toxic (OECD, 2002), however, the replacement PFASs remain a concern. While the acute toxicity of PFASs is moderate, toxicological studies reported adverse effects of PFASs spanning from hepatotoxicity, neurotoxicity, immunotoxicity, developmental and reproduction toxicity to endocrine disruption (Rainieri et al., 2017).

1.2.3 Pharmaceuticals and steroids Pharmaceuticals bear diverse classes of organic compounds such as analgesics, antiepileptics, antihyperlipidemics, non-steroidal anti-inflammatory drugs, synthetic hormones, antimicrobials and many more. Pharmaceutical administrations are not always entirely metabolized within the body and can be excreted as unaltered or an active metabolite of the parent compound, which may undergo further transformation or remain intact when they pass through WWTPs (Salimi et al., 2017). Globally, the major source of pharmaceuticals in surface waters is the discharge of urban wastewater, although emissions from the pharmaceutical industry, agriculture and aquaculture can be important on a local scale (Richardson and Ternes, 2018). Pharmaceuticals, their metabolites and transformation products may pose both acute and long-term chronic 20

effects on aquatic biota at low, environmentally relevant concentrations (Patel et al., 2019). Adverse biological effects of pharmaceuticals on biota comprise e.g. embryotoxicity, mutagenicity, gill alterations, growth inhibition and changes in foraging behavior. Among all, the great advance of antibiotic resistance in environmental microbial communities raises major concerns (Patel et al., 2019). Natural endogenous (e.g. 17β-estradiol, estrone, estriol, progesterone, cortisone, testosterone and thyroid hormone) and synthetic steroids (17α- ethinyloestradiol, mestranol) excreted by humans and livestock enter surface waters by direct discharges or WWTP effluents (Thomaidis et al., 2012). Steroids together with non-steroidal chemicals from various classes known as endocrine disruptors (e.g. bisphenol A, phthalates, phenothrin) cause adverse endocrine effects such as fish feminization.

1.2.4 Personal care products (PCPs) Personal care products (PCPs) are a diverse group of chemicals including disinfectants (e.g. triclosan), fragrances (e.g. musks), insect repellants (e.g.DEET), preservatives (e.g. parabens) and UV filters (Benzophenone-3), which are used in consumer products like soaps, lotions, toothpaste, fragrances, sunscreens etc., and they are applied directly on the human body to change appearance, taste or odour (Brausch and Rand, 2011). PCPs enter the aquatic environments through recreational activities such as swimming and also via showering and bathing as well as other technological process (Tijani et al., 2016). PCPs are ubiquitous and even though some may have restricted lifetime in the aquatic environment, their levels in surface waters are very stable as they are released constantly (Ohoro et al., 2019). PCPs were reported to cause a wide range of adverse effects in aquatic organisms such as endocrine disruption (parabens), changes in developmental rates of invertebrates and fish (polycyclic musks), or acute toxicity to algae (disinfectant triclosan), (Brausch and Rand, 2011).

1.2.5 Brominated and emerging flame retardants Flame retardants are chemical additives used in plastics, textiles, electronic circuitry, and other materials, from the class of polybrominated diphenyl ethers (PBDEs), organophosphate esters (OPEs) and chloro- organophosphates (Thomaidis et al., 2012). PBDEs were banned in some countries, due to their widespread presence in the environment and adverse effects on biota (developmental neurotoxicity and others), but they are still being released to the environment from older products (Linares et al., 2015). INTRODUCTION 21

New replacement flame retardants, OPEs and chloro-organophosphates, mainly released from WWTPs, have emerged in surface waters (Salimi et al., 2017). However, these compounds were also reported to cause neurotoxicity, embryotoxicity or disturbance of thyroid signaling at higher than current environmental levels, and therefore need to be carefully monitored (Greaves and Letcher, 2017).

1.2.6 Benzotriazoles and benzothiazoles Benzotriazoles and benzothiazoles are widely used as corrosion inhibitors, UV stabilizers in plastics and polymers or in rubber production (Richardson and Ternes, 2018) . They are highly soluble in water, resistant to biological degradation, and not well removed in wastewater treatment. In surface waters, they are present at low ng/L levels but can reach up to tens of µg/L (Shi et al., 2019). Benzotriazoles have high bio-accumulation potential and they were shown to elicit multiple toxic effects including changes in mitochondrial bioenergetics, embryonic development, and locomotor activity of zebrafish, in vitro anti-androgenic and in vivo estrogenic effects, neurotoxicity and hepatotoxicity in fish or alteration of molting frequency in Daphnia (Liang et al., 2019; Shi et al., 2019; Tangtian et al., 2012).

1.2.7 Pesticides Pesticides are a wide class of chemicals used to limit, inhibit, prevent the growth or repel harmful animals, insects, weeds, invasive plants, and fungi. Agricultural activity is the main source of pesticides in the environment, however, industrial emissions during their production and urban run-off also significantly contribute to the levels of pesticides in surface waters (Pietrzak et al., 2019). The levels of pesticides in surface waters fluctuate in time, depending on the application pattern and weather. Pesticides in surface waters may undergo degradation by hydrolysis, oxidation, biodegradation, or photolysis, while for some compounds the transformation products can occur at higher levels and show higher toxicity than the parent compounds (Houtman, 2010). Pesticides elicit a multitude of toxic effects to aquatic life, threaten human health and water pollution due to pesticides is a priority issue of global concern (Pietrzak et al., 2019).

1.2.8 Others Besides the main, above mentioned groups, CECs comprise other compounds or compound classes which have been increasingly detected in surface waters and therefore should not stay out of the scientific and regulatory 22

focus, e.g. halogenated methanesulfonic acids, ionic liquids, algal toxins, dioxane, disinfection byproducts, siloxanes and bisphenol A.

1.3 Sampling Water is an extremely heterogeneous matrix where the distribution and mixing of waterborne chemicals are affected by the hydrodynamics of the water, the sorption partition coefficients of the chemicals, and the amount of organic matter. The concentrations of contaminants vary spatially and temporally, because episodic events from surface runoff, spills, and other point source contamination can result in isolated or short-lived chemical pulses in the water (Alvarez and Jones-Lepp, 2010). Traditional sampling methods i.e. grab sampling have several limitations for monitoring of CECs in surface waters. Modern chemical analytical instrumentation allows for the analysis of small water volumes with no or only low sample enrichment for most of the typical water micropollutants pollutants, which occur at trace levels, while the analysis of some priority substances with very low EQS values as well as in vivo and in vitro tests may require greater enrichment and larger water volumes (Neale et al., 2015). Grab samples represent the concentration of CECs only at the instant of sampling and even repetitive grab sampling schemes usually fail to capture the concentration pulses during episodic events (Vrana et al., 2005). To overcome these limitations, integrative sampling techniques like passive sampling or solid phase extraction have been developed (Dimpe and Nomngongo, 2016; Jones et al., 2015).

1.3.1 Large volume solid phase extraction (LVSPE) The combination of targeted and nontargeted chemical screening analysis in combination with bioassays has been recommended for the identification of (eco-)toxicologically active compounds and mixtures by several studies (Altenburger et al., 2015; Brack et al., 2015). Large sample volumes are required for such a combined approach, which may present logistic, technical, economic and scientific issues related to the storage and transport. To prevent these problems, a novel automated sampling device based on solid phase extraction (SPE) was developed by Schulze et al. (2017). Solid phase extraction (SPE) is the most powerful sampling and enrichment technique for complex mixtures of known and unknown contaminants, which are captured and stabilized on the sorbent during the sampling. Several well- tested and widely used solid phases, based on C18 or polystyrene- divinylbenzene (co-)polymers, that trap organic compounds with a broad range INTRODUCTION 23

of properties (nonpolar to polar, neutral to charged) are commercially available (Fontanals et al., 2010). The newly introduced on-site large volume solid phase extraction (LVSPE) device combines a pre-filtration cartridge (separation of suspended particulate matter (SPM) from the water phase) with a set of tailor- made columns that allow customizable selection of sorbents to cover a broad range of compounds with different properties (Fig. 2). To force the water through the extraction columns, the LVSPE device uses a pressurized system which is powered by a car battery and controlled by 12 V electronic components with low energy consumption (Schulze et al., 2017).

Figure 2: (a) Picture of the LVSPE50 device; (1): Dosing system (500 mL), (2): pre-filter (3): ball valve, (4): pressure chamber (550 mL), (5):extraction cartridge, (6): controller (Photo by MAXX GmbH); (b) LVSPE device during field sampling at the WWTP effluent in Brno – Modřice, Czech. Rep. in August 2014.

1.3.2 Passive sampling Passive sampling is a sampling technique based on free flow of analyte molecules from the sampled medium to a collecting medium, as a result of a difference in chemical potentials of the analyte between the two media. (Górecki and Namienik, 2002). Passive samplers are non-mechanical devices, that require minimal resources of personnel and equipment for sampling and avoid almost every disadvantage of active sampling and/or of the methods of preparing the samples mentioned above. Materials used in passive sampler construction have constant composition and well-defined diffusion and partition properties a constant uptake capacity (Taylor et al., 2019). Passive sampling provides a sensitive measurement of dissolved concentrations that is integrated over time (Cfree). Due to its proportionality to the chemical activity and chemical potential, Cfree is considered a key parameter in understanding 24

chemical’s exposure of aquatic organisms (Reichenberg and Mayer, 2006). Passive sampling enables integrative collection of contaminants over an extended period of time and captures residues from episodic events, which typically remain undetected when using grab sampling (Alvarez et al., 2004; Vrana et al., 2005). Nowadays, various passive samplers’ designs, multiple receiving phases and diffusion membranes are available. Passive samplers can be used for sampling of a wide variety of compounds, e.g. semipermeable membrane device (SPMD) for hydrophobic substances such as PAHs or PCBs (Huckins et al., 1993) and polar organic chemical integrative sampler (POCIS) for hydrophilic substances such as polar pesticides and pharmaceuticals (Alvarez et al., 2004), (Figure 3). Because of the integrative character of sampling, passive samplers accumulate a sufficient amount of sampled chemicals for detection of small concentrations in water and samples for multiple analyses including bioassays (Jones et al., 2015; Moschet et al., 2014; Vrana et al., 2014). Passive sampling has been successfully used for combined chemical and effect analyses of CECs in surface waters in many studies (Emelogu et al., 2013; Jalova et al., 2013; Jarosova et al., 2012; Novák et al., 2018). Nevertheless, the recognition of the added value of passive sampling by regulators is still rather limited. In Europe, the strict monitoring requirements of the EU WFD hinder the implementation of passive sampling for regulatory purposes, whereas environmental managers in the United States have more freedom to apply approaches in line with the current scientific insights (Booij et al., 2016). The main weakness of passive samplers from the regulators’ perspective is insufficient quality control related to variability in the reported results due to inaccuracies of the partition coefficients of target analytes (Booij et al., 2016). This issue is more pronounced in case of polar compounds (Křesinová et al., 2016).

Figure 3: Stainless steel protective cage with POCIS discs in triplicate (left); POCIS disc (middle) and SPMD membrane stretched in the stainless steel protective cage (right) with visible biofouling after deployment in surface water (photos by Dr. Vrana) INTRODUCTION 25

1.4 Chemical analyses and non-target screening Chemical analyses of CECs in surface waters require highly sensitive analytical methods, which have become available with the advance of gas chromatography (GC) and (ultra)-high performance liquid chromatography [(U)HPLC] coupled either to tandem mass spectrometry (MS/MS), or a wide range of high resolution mass spectrometers (García-Córcoles et al., 2019; Richardson and Ternes, 2018). The selection for GC or LC separation usually depends on the physicochemical properties of the analytes. Thus, the less volatile and polar compounds are normally separated with LC, whereas GC is often used for the volatile compounds (Martín-Pozo et al., 2019). Conventional multi-residue target analyses of complex environmental samples like WWTP effluents provide only a limited picture of a mixture’s complete chemical composition. Many compounds are co-extracted in the extraction process, they remain unnoticed during the analysis, but they may contribute significantly to the observed biological effects. Non-target screening techniques using GC and LC-MS together with high-resolution mass spectrometers (HR-MS) have therefore been increasingly applied in environmental analysis (Gómez et al., 2009; Ibáñez et al., 2008; Schymanski et al., 2015; Zedda and Zwienerbu, 2012). A study by Schymanski et al. (2015) showed that while the non-target screening analytical techniques were substantially harmonized among laboratories, the data-processing and evaluation part still presented a challenge. Currently, powerful chromatographic deconvolution and structure elucidation software, in silico prediction tools for fragmentation, retention time and ionization, as well as MS and MS/MS libraries of environmental contaminants are being developed rapidly and enable tentative identification of unknowns in a reasonable time frame with a reasonable confidence (Krauss et al., 2010; Schymanski et al., 2015). Non-target screening has a great potential for characterization of complex environmental mixtures, identification of unknown toxicity drivers, prioritization of substances for monitoring programs and assessment of environmental quality (Du et al., 2017; Hollender et al., 2019).

1.5 Bioassays Adverse biological effects of CECs occurring in surface waters have been studied extensively in the recent decade (Richardson and Ternes, 2018; Rykowska and Wasiak, 2015). To assess these effects, a large number of bioassays indicative of different endpoints at various levels of biological organization have been developed (Escher et al., 2018). In vivo-bioassays using 26

whole organisms like green alga Raphidocelis subcapitata, crustacean Daphnia magna or fish Danio rerio investigate the effects of CECs on apical endpoints such as survival, growth and reproduction. In vitro assays generally measure effects at the cellular level and are mode-of action specific. They allow rapid and sensitive detection and are designed for high throughput applications in the laboratory. Further, several in vitro approaches allow to assess specific endpoints, relevant to CECs, including neurotoxicity, immunotoxicity, oxidative stress response, xenobiotic metabolism regulation, genotoxicity, dioxin-like and hormonal activities like (anti-)estrogenicity, (anti-)androgenicity, glucocorticoidand thyroid activity, etc. (Connon et al., 2012; Neale et al., 2015). These specific biological effects were repeatedly observed in the samples from surface waters, drinking waters, raw and treated wastewaters and technical systems (Dizer et al., 2002, Loos et al., 2013; Jálová et al., 2013; Osman et al., 2015; Zhao et al., 2011). Some can be associated with particular compounds as the main toxicity drivers, e.g., estrogenicity due to E2 and EE2 (Aerni et al., 2004; Koenig et al., 2017; Miège et al., 2009) or phytotoxicity to herbicides (Tang and Escher, 2014). However, the cause of many effects observed in bioassays remains unexplained by the chemicals detected in the same samples (Burgess et al., 2013; Neale et al., 2015). By applying a panel of cellular and small-scale whole- organism assays it is possible to obtain a more holistic profile of the effects of all chemicals present in a water sample without identifying the causative compounds individually. (Escher et al., 2018)

1.6 Effect directed analysis (EDA) Effect-directed analysis (EDA) is a method, which combines bioassays, fractionation and chemical analysis with the aim to identify compounds responsible for the observed biological effects (Brack, 2003). The procedure is typically applied to organic extracts of the sample and starts with biotesting. In case significant effects are detected, the complexity of the sample is sequentially reduced by fractionation, repeated biotesting and elimination of fractions with low or no biological activity (Fig. 4). Several fractionation steps can be carried out until the isolated toxic fractions are subjected to the identification of toxicity drivers by target chemical analysis and non-target screening. Finally, the identified toxicity drivers need to be confirmed as the cause of the measured effect (Brack et al., 2016). EDA is applied in drug discovery, toxicology, forensics, and environmental sciences, where it proved to INTRODUCTION 27

be a powerful tool to facilitate identification of unknown toxicants (Brack et al., 2016; Burgess et al., 2013). EDA studies are very costly, tedious and time consuming. However, novel sampling, bioanalytical, chemical and software tools make EDA studies increasingly feasible through miniaturized and automated high-throughput formats, hyphenated tools, lowered detection limits, and optimized multi-target and non-target screening (Guijarro et al., 2015; Brack et al., 2013).

Figure 4: General scheme of effect directed analysis. Redrawn from Brack (2003)

1.7 Risk assessment (RA) The assessment of potential risk of CECs to aquatic ecosystems has been addressed by several studies (Ginebreda et al., 2014; Kuzmanovic et al., 2014; Smital et al., 2013). These studies combine the established methods of risk assessment of chemicals with novel approaches to overcome the greatest challenge which is data scarcity (Von der Ohe et al., 2011). The risk assessment process considers the exposure level, often referred to as – for example – Predicted Environmental Concentration (PEC), and potential effect of a given substance, referred to as Predicted No-effect Concentration (PNEC). The PECs can be derived from available measured data (concentrations in the environment) and/or model calculations. The PNEC values are ecological safety thresholds usually determined on the basis of results from single species laboratory tests or, in a few cases, established effect and/or no-effect concentrations from model ecosystem tests (e.g. pesticides). PNEC is regarded as a concentration below which an unacceptable effect will most likely not 28

occur (European Commission, 2003). PEC/PNEC risk ratios above 1 indicate that the substance poses risk to aquatic life in the investigated area. Due to the great number of CECs potentially occurring in the environment, there is a need to prioritize these compounds for the management optimization purposes (Kuzmanovic et al., 2014; Von der Ohe et al., 2011). Existing prioritization schemes, mainly based on monitoring and ecotoxicity data as well as predictive modelling, are varied and often highlight different compounds of concern (Letsinger and Kay, 2019). A methodology to prioritize CECs based on the risk of individual compounds was developed within the NORMAN Network (Dulio and Von der Ohe, 2013). This scheme seems most suitable for CECs as it reflects the ever growing list of the compounds of interest along with the evolving availability of monitoring and toxicity data (Dulio et al., 2018). The prioritization is based on the lowest NORMAN PNECs, which were determined with the use of either experimental data, existing environmental quality standards (EQSs), or in silico predictions. Within this prioritization approach, maximum environmental concentrations 95 (MEC95), which are the 95th percentiles of the measured concentrations at the investigated area, are compared to the lowest NORMAN PNEC values to determine frequency of PNEC exceedance (indicator 1), and the extent of exceedance (indicator 2). Indicators 1 and 2 are given risk scores which are used to derive the final NORMAN risk score. Measured compounds on the priority list are then ranked in order of the resulting NORMAN risk score. This ranking than allows to efficiently prioritize various CECs based on their potential risks to aquatic biota.

AIMS OF THE STUDY 29

2 AIMS OF THE STUDY

The overall aims of the present study were to apply and assess different combinations of effect directed analysis (EDA) tools to identify and prioritize river basin specific pollutants (RBSPs) from the class of contaminants of emerging concern (CECs), particularly endocrine disrupters, compounds toxic to microalgae and neurotoxic compounds. Specific objectives of the study were then as follows:

1. to evaluate applicability of a newly developed simplified EDA protocol for effect-based monitoring of surface waters including the use a novel sampling device, set of bioassays, extensive multi-residue analysis and non-target screening (publication I and IV) 2. to identify compounds responsible for algal growth inhibition and acetylcholinesterase inhibition in a WWTP effluent extract using complex EDA approaches based on non-target screening (publication II) and 2- dimensional liquid chromatography system (publication V) 3. to asses water quality in the Bosna River by use of a combination of passive sampling, a set of in vitro bioassays and multi-residue analysis of several compound classes (publication III)

3 SUMMARY OF MATERIALS AND METHODS

3.1 European demonstration program (EDP) – a case study on effect-based water monitoring (publications I and IV) Surface water samples were collected at 18 sampling sites in four European river basins in six countries (Figure 5) from July 2013 to August 2014 by means of grab sampling and large volume solid phase extraction (LVSPE). The grab samples (2L) were designated for in-vivo thyroid assay with Xenopus tadpoles. The LVSPE extracts were subject to a set of bioassays addressing 9 endpoints, multi-residue target analysis addressing 151 target compounds and GC-MS non-target screening. Resulting effect data were linked to the measured concentrations of target compounds, which were also used for basic risk assessment and prioritization.

3.1.1 LVSPE sampling An onsite LVSPE device (UFZ, Leipzig, Germany; Maxx GmbH, Rangendingen, Germany) was used to extract organic micropollutants from 50L of river water. The extraction device comprised three different sorbents in sequence, designed to capture neutral, weakly acidic and weakly basic organic compounds. Laboratory blanks were prepared by percolation of 2L of mineralized LC-grade distilled water through the LVSPE device for 100 cycles (equivalent of 50L).

3.1.2 Effect assessment A set of 4 in vitro and 3 in vivo bioassays was applied to screen for 10 endpoints covering both non-specific and specific toxicity (Tab. 1). The effective concentrations of surface water samples were expressed in relative enrichment factors (REFs) as proposed by Escher et al. (2006). The LVSPE extracts were reconstituted in a solvent and mixed with the test media to reach final REFs from 1 to 100, meaning that the tested range covered the original river concentrations (REF=1) as well as concentrations up to 100-times higher (REF=100). The in vivo thyroid activity assay with Xenopus was performed with whole raw water samples. For active samples in the receptor-mediated assays, effect equivalents of standard agonists or antagonists were calculated.

SUMMARY OF MATERIALS AND METHODS 31

Table 1: Overview of bioanalytical methods

Bioassay Biological model Endpoint Method reference

in vitro U2-OS (GR- Glucocorticoid GR – CALUX® Sonneveld et al. (2005) CALUX®) activity Estrogenic MELN (MCF7- Balaguer et al. (1999) ER-mediated activity ERE-Luciferase- activity Antiestrogenic Neomycin) Creusot et al. (2016) activity Androgenic Wilson et al. (2002) AR-mediated activity MDA-kb2 activity Antiandrogenic Creusot et al. (2014) activity Purified enzyme Enzyme Ellman et al. (1961) AChE inhibition AChE inhibition Galgani and Bocquene (1991) in vivo Algal growth Raphidocelis Growth rate OECD guideline 201 inhibition subcapitata inhibition Rojickova and Dvorakova (1998) Survival OECD guideline 236 Zebrafish embryo Danio rerio - Sum of acute toxicity embryo sublethal OECD guideline 236 endpoints Thyroid activity Xenopus laevis - Thyroid Fini et al. (2007) embryo activity

3.1.3 Chemical analyses Neutral LVSPE extracts were subject to target analysis of 151 compounds and to GC-MS non-target screening analysis using state-of-the-art LC-(HR)MS/MS and GC-MS tools (Tab. 2). The list of target compounds was set up to cover several classes of environmentally and toxicologically relevant CECs and their transformation products.

Table 2: Overview of chemical analytical methods

Analytical method Target compound Nr. of Method reference /Instrumentation group targets LC-ESI-QFT-HRMS/MS (Thermo Q-Exactive Polar micropollutants 44 Schymanski et al. (2014) Orbitrap) LC-QgQ-MS/MS Industrial solvents, 4 developed for EDP (Thermo TSQ-Vantage) pesticide LC-ESI-HRMS/MS (Thermo LTQ-FT Glucocorticoids 20 Schriks et al. (2010) Orbitrap) LC-ESI-MS/MS Phenolic (Thermo TSQ Quantum 6 Petrovic et al. (2003) xenoestrogens Ultra AM) LC-ESI-ITFT-HRMS (Thermo LTQ Orbitrap Polar micropollutants 47 Hug et al. (2014) XL) LC-MS/MS Steroids 27 Griffith et al. (2014) (AB Sciex Qtrap 6500) GC-EI-MS Legacy pesticides, WFD (Agilent 5795C) priority compounds, 22 modified ISO 6468:1996 industrial compounds GC-EI-MS Non-target screening NA Slobodnik et al. (2012) (Agilent 5795C)

3.1.4 Linking effects and detected compounds For receptor mediated in vitro assays, mass balance calculations were conducted using relative effect potencies (REPs) of known agonists according to Kinani et al. (2010) to quantify the contribution of the detected target compounds to the observed biological activity. For in vivo assays, the zebrafish embryo acute toxicity assay and algal growth inhibition assay, the link between measured compounds and observed effects was calculated using toxic units (TU) and their sum, similarly to Booij et al. (2014) and Kuzmanovic et al. (2014). Toxicity of detected compounds for fish and

green algae was estimated by the program ECOSAR (v1.11) and the 96h-EC50 value

of the most toxic ECOSAR chemical class (minECOSAR 96h-EC50i) was selected (US-EPA, 2012).

3.1.5 Risk assessment and prioritization Risk assessment and prioritization of detected target compounds was based on the NORMAN lowest predicted no effect concentrations (PNECs) developed within the NORMAN Network (Dulio and Von der Ohe, 2013). Target compounds were prioritized in order of NORMAN risk scores, which take into account the frequency and extent of the NORMAN lowest PNEC exceedance by SUMMARY OF MATERIALS AND METHODS 33

each compound measured at individual sampling sites and their 95th percentile (MEC95), respectively.

3.2 Effect directed analysis of WWTP effluent extract (publication II and V) The LVSPE extract of WWTP effluent in Brno Modřice (Czech Rep.) from publication I elicited toxic effects in the algal growth and AChE inhibition assays. The sample was subject to a full higher-tier EDA study to identify AChE inhibitors and a new LVSPE extract was collected at the site to perform a higher-tier EDA study focused on toxicity to algae.

3.2.1 Fractionation (publication II) The LVSPE extract was fractionated in three fractionation steps. The first fractionation was achieved on site by the LVSPE50 sampling device followed by 5 elution steps in the laboratory (5 fractions; F1-F5). The second fractionation of the bioactive fraction F1, was performed with RP-SPE, followed by stepwise elution with water and MeOH mixtures with increasing eluotropic strength (9 fractions; F1.1 - F1.9). The third fractionation of pooled bioactive fractions F1.2, F1.3 and F1.4, was conducted with RP-HPLC resulting in 31 fractions (F1.4.1-F1.4.31).

3.2.2 Algal growth inhibition assay (publication II) Miniaturized algal growth inhibition assay with unicellular green alga Pseudokirchneriella subcapitata (syn. Raphidocelis subcapitata) was carried out according to a modified OECD method 201 (2011) as described by Rojickova and Dvorakova (1998). Sample extracts in DMSO (0.5% v/v) were tested at a relative enrichment factor (REF) up to 200.

3.2.3 Non-target screening (publication II) Both GC- and LC-MS non target screening techniques were applied for analysis of the bioactive fractions. Specific workflows were followed to tentatively identify proposed chemical structures. The workflows are described in detail in the supplementary material of publication II.

3.2.4 Linking effects and detected compounds (publication II) To link the observed bioactivity and the tentatively identified compounds, several experimental datasets and ECOSAR modelling were used. Firstly, effect concentrations (ECx, ICx and LOEC) from laboratory studies with algae and cyanobacteria were extracted from the US-EPA Ecotox database (US-EPA, 2015). 34

Secondly, experimental toxicity data for photosynthesizing organisms, i.e. algae and cyanobacteria, were collected from open literature and thereafter an internal database with more than 1700 experimental data points was built to derive the lowest effective experimental concentration value for each compound (maximum

toxicity potential - MTPexp). Thirdly, green algae 96h EC50s were predicted by the ECOSAR program, v1.11(US-EPA, 2012) to fill the gaps in experimental data. The

toxicity of the ECOSAR class with the lowest EC50 value was selected to represent

the predicted toxicity of each compound (maximum toxicity potential - MTPpred). The detected compounds were ranked in order of their maximum toxicity potential (MTP), whereas experimental data were always given priority over the predicted ones

3.2.5 LCxLC system for microfractionation, high throughput AChE bioassay and parallel HR-TOF detection (publication V) The primary fractionation was achieved on site by the LVSPE50 sampling device followed by 3 elution steps in the laboratory (3 fractions; F1-F3). Neutral eluate (F1), which elicited AChE inhibition, was subject to microfractionation (384 fractions) using a 2-dimensional LC system. The flow of LCxLC effluent was split into the in vitro AChE bioassay (four 96-well plates, 80%) and into the parallel high-resolution time of flight (HR-TOF) mass spectrometric detection (20%).

3.3 Analytical and bioanalytical assessment of the Bosna River (publication III)

3.3.1 Passive sampling Passive samplers (SPMD and POCIS) were deployed at 10 locations (S1- S10) along the Bosna River, BiH, in 2012. At each site 3 POCIS discs and 3 SPMD replicates were co-deployed in a protective cage. Concentrations of compounds from SPMDs were estimated according to an approach based on the sampler-

water partition coefficients (KSW) of the compounds and their sampling rate (RS) calculated on the basis of dissipation of performance reference compounds (PRCs) as described in Vrana et al. (2014), Rusina et al. (2010) and Booij and Smedes (2010). Concentrations of compounds from POCIS samplers were estimated using

median RS value of 0.2 L d-1 according to Harman et al. (2012).

3.3.2 Chemical analyses Chemical analyses of 168 target compounds (134 in POCIS and 34 in SPMD extracts) in 5 compound classes, namely hydrophobic compounds (PAHs, PCBs, SUMMARY OF MATERIALS AND METHODS 35

OCPs), current use pesticides, estrogens and pharmaceuticals, were conducted by use of state-of-the-art GC-MS(-MS) and HPLC-MS(-MS) techniques.

3.3.3 In vitro bioassays Three in vitro cell-based reporter gene bioassays were used to assess ER-, (anti)AR- and AhR-mediated potencies and cytotoxicity of the SPMD and POCIS extracts (Tab. 3). Cytotoxicity of the sample extracts was tested with combination of three dyes (Alamar Blue, CFDA-AM, neutral red) according to Schirmer et al. (1998).

Table 3: Overview of bioanalytical methods Effect Cell line Standard reference Method reference Wilson et al. Androgenicity MDA-kb2 cells DHT: 3.3 pM–100 nM (2002) Antiandrogenicity MDA-kb2 cells FLU: 110 nM – 100 µM Jálová et al. (2013) Demirpence et al. Estrogenicity MVLN cells E2: 1–500 pM (1993) H4G1.1c2 cells Dioxin-like activity (AhR) TCDD: 1–500 pM Nagy et al. (2002) (CAFLUX assay)

3.3.4 Contamination profiling and hazard assessment Toxicity profiles based on the results of applied bioassays were translated into site-specific contamination profiles according to an approach outlined by Hamers et al. (2010). The ratio between the biological response of downstream sites (S2-S10) and a reference site (S1), contamination index (CI), was regarded as a measure of contamination by each of the tested endocrine effects. Assessments of hazards of detected target compounds were conducted by use of the lowest predicted no effect concentrations (PNEC), values derived by the NORMAN Network (Dulio and Von der Ohe, 2013). Hazard quotients (HQs) were calculated as a ratio of the estimated dissolved concentration of an individual compound at a particular sampling site and the NORMAN lowest PNEC value, whereas compounds with HQs exceeding 1 might pose risk to aquatic life. Overall hazard index (HI) was calculated by summation of all HQs of compounds detected at each sampling site.

4 SUMMARY OF RESULTS AND DISCUSSION

4.1 European demonstration program (EDP) – a case study on effect-based water monitoring (publications I and IV)

4.1.1 Sampling The novel LVSPE device itself (Fig. 2) and sample processing steps were optimized within the study and its repeated application in field during the EDP sampling campaign (18 sites shown in Fig. 5) demonstrated the suitability of LVSPE for combined chemical and effect-based screening of water quality. Within 2-4 hours, the LVSPE device enabled extraction of 50L, which proved to be sufficient volume for a set of 6 bioassays (with required enrichment up to REF 100) and chemical analyses. The LVSPE device proved to be useful for sampling of complex environmental mixtures of known water contaminants, as the recoveries of 251 spiked compounds (log D range -3,6 – 9,4 at pH 7, concentration range 1-2400 ng.L-1) were acceptable (60-123%) for most compounds (Fig. 6).

Figure 5: Overview map of the EDP sampling locations (clockwise from top left to bottom left) in the Saale RB (Germany), Danube RB (Czech Rep., Slovakia and Hungary), Sava RB (Croatia) and Emme RB (Switzerland). SUMMARY OF RESULTS AND DISCUSSION 37

The effect assessment showed the LVSPE extract enriched at REF 100 allowed a discrimination of active from non-active samples in case of 8 out of 10 toxic endpoints. The neutral (HR-X) sorbents performed better than the weak acidic (HR-XAW) or weak basic (HR-XCW) ones in terms of recovery and capture of toxic effects.

Fig. 6: Scatterplot of the total recoveries (in %) of compounds (N = 251) spiked in a pristine water sample of Wormsgraben (Harz Mountains, Germany) and extracted with the LVSPE50 device versus the water-octanol partition coefficient at pH 7.0 corrected for the speciation (log D).

4.1.2 Effect assessment The most frequently observed effects were estrogenicity, zebrafish acute embryo toxicity and algal growth inhibition (Tab. 4). The biological effects assessed in the EDP program were environmentally relevant, except for the AChE inhibition, which was measurable only in highly enriched extracts

(EC50s>REF 100). With the receptor mediated assays, effects could be detected even for diluted (estrogenicity) or moderately enriched samples (REF<20, androgenicity, anti-estrogenicity and glucocorticoid activity). In the algal growth inhibition and the FET assay, the EC50s were never lower than REF 10 and at some cases exceeded REF 100. In vivo thyroid activity in transgenic Xenopus, using whole water samples, was detected at one sampling site (Saale RB).

4.1.3 Chemical analyses Target analysis of 151 compounds in neutral LVSPE extracts (HR-X) provided extensive chemical characterization of the sampling sites. Out of the total 151 compounds, 107 compounds were detected at least at one sampling 38

site, while 44 compounds, mainly glucocorticoids, steroids and WFD priority compounds, were below detection limits. Twelve compounds occurred at all sampling sites with the exception of the reference site (Saale RB). The concentrations of detected target compounds ranged from a few ng L-1 to a few µg L-1, which is in line with a similar study conducted by Kuzmanovic et al. (2014) and Smital et al., (2013). The highest median concentrations were detected for 1H-benzotriazole, sucralose, phenylbenzimidazolesulfonic acid, 4- toluenesulfonamide and 5-methyl-1H-benzotriazole.

Table 4: Overview of the results of effect assessment linked to the results of target analyses (results of mass balance calculations and identification of the main effect drivers) Range of measured bioactivity Main effect drivers (percentage of Bioactivity (Nr. of active/examined sites) effect explained) Glucocorticoid activity cortisol (29%) 0.3 - 30.5 (4/18) Dex-EQ [ng L−1] 6α-methylprednisolone (2%) estrone (E1) nonylphenoxyacetic Estrogenicity 0.06-1.85 (15/18) acid, bisphenol A and octylphenol E2-EQ [ng L−1] (0 - 77 %) Antiestrogenic activity 35 - 185 ng/L (5/13) unexplained OH-Tam EQ [ng L−1] Androgenic activity 0.93 - 2.7 ng L-1 (4/18) unexplained DHT-EQs [ng L−1] Antiandrogenic activity n.d. n.a. FLU-EQ [pg L−1] AChE inhibition IC50 > 100 (5/18) n.a. [REF] Algal growth inhibition diuron, atrazine and IC50 = 17 – 259 (14/18) [REF] terbuthylazine (30-166%) nonylphenol, nonylphenoxyacetic Zebrafish embryo LC50 = 12-83 (15/18) acid, octylphenol, diazinon, acute toxicity [REF] fipronil (0-20.6%) Thyroid activity 1/18 unexplained n.d. – never detected, n.a. – not available

SUMMARY OF RESULTS AND DISCUSSION 39

4.1.4 Risk assessment and prioritization NORMAN PNEC values were available for 94 out of 151 target compounds. The MEC95 of 15 compounds exceeded the PNEC value and measured concentrations of another 6 compounds exceeded their PNEC values at least once. Target compounds were ranked according to their NORMAN risk score and listed in Table 5. The EDP priority list contained 21 compounds including 4 pesticides and 2 pesticide transformation products, 7 pharmaceuticals (3 of them macrolide antibiotics), 3 surfactant-derived compounds and 1 PAH, biocide, repellent and plasticizer each. 40

Table 5: The list of priority compounds from the EDP case study based on NORMAN prioritization methodology Frequency of Frequency of MEC95 NORMAN PNEC MEC95 Median occurrence PNEC Rank Compound name Compound group / Usage pattern exceedance risk [ng/L] [ng/L] [ng/L] N=18 exceedance of PNEC score [%] [%] Perfluorooctanesulfonic 1 surfactant, WFD priority 0.13 31

4.2 Effect directed analysis of the WWTP effluent extract (publication II and V)

4.2.1 Toxicity to algae (Publication II) Algal growth inhibition was observed in two primary fractions of the

WWTP extract, namely F1 (EC50=18.6 REF) and F3 (EC50=56.2 REF). Consequently, F1 was subject to further fractionation by RP-SPE creating 9 fractions, out of which F1.3 was the most active. F1.3 and its neighbouring fractions F1.2, and F1.4 were pooled (EC50=32.9) and further fractionated on

RP-HPLC. Out of the total 31 fractions, F1.4.7 (EC50=82.3 REF), F1.4.8

(EC50=84.3) and F1.4.31 (EC50=105.5) were toxic to algae. Toxicity of treated waste waters to algae was described in earlier studies (Köhler et al., 2006; Magdaleno et al., 2014; Maselli et al., 2013), and it remains a concern regarding the chemical and ecological status of the recipient surface waters. In the GC(xGC)-MS non-target screening, 41 and 22 compounds were detected and tentatively identified in the phytotoxic primary fraction F1 and combined tertiary fractions F.1.4.7 and F1.4.8, respectively. These compounds comprised of organophosphates, pharmaceuticals, personal care products, pesticides and industrial pollutants. Two-dimensional GC separation enabled identification of several compounds with co-eluting peaks in the one- dimensional setup, e.g. ticlopidine and fluconazole, promethryn and clorophene, tramadol and desmethyltramadol, venlafaxine and norvenlafaxine. In the LC-HRMS non-target screening of three tertiary fractions active in the bioassay (F1.4.7, F1.4.8, F1.4.31), the total of 345 compounds, including mostly pharmaceuticals, industrial chemicals and by-products, biocides, herbicides and their transformation products, were detected and tentatively identified. The collection of experimental and predicted ecotoxicity data for the detected compounds resulted in the total of 377 Maximum toxicity potential

(MTP) values. Experimental values (MTPexp) were available for 68 compounds

(17.8%). ECOSAR predicted maximum toxicity potential (MTVpred) values were used for the remaining 309 compounds, while no MTP could be derived for 6 chemicals. Subsequently, detected compounds were ranked according to their MTPs to identify those with the greatest potential to contribute to the observed toxicity. The list of top 10 compounds (Tab. 6), identified as the most likely toxicity drivers, was dominated by herbicides and their transformation products, which is in line with earlier studies focused on phytotoxicity of 42

surface and waste waters (Bengtson Nash et al., 2005; Beate I Escher et al., 2006; Tang and Escher, 2014; Vermeirssen et al., 2010).

Table 6: List of the top 10 compounds detected and tentatively identified by GC- and LC-MS non-target screening of phytotoxic fractions of WWTP effluent extract (F1, F1.4.7, F1.4.8 and F1.4.31) ranked according to their estimated maximum toxicity potential.

Status under EU Reference to Name Use pesticide MTP [µg/L] the MTP regulation* selective triazine herbicide, US EPA- 1 Terbutryn not approved 0.02 photosystem II inhibitor ECOTOX chlorotriazine herbicide, Booij et al., 2 Terbuthylazine approved 0.40 photosystem II inhibitor 2014 selective triazine herbicide, US EPA- 3 Prometryn not approved 0.41 photosystem II inhibitor ECOTOX industrial formulation of 2-(2-heptadec-8- lubricant additives, 4 enyl-2-imidazolin-1- n.a. 0.84 ECOSAR lubricants, functional fluids yl)ethanol and greases chloroacetanilide herbicide, Nagai and 5 Acetochlor not approved 1.40 elongase inhibitor Taya, 2015 6- chlorinated degradation parent compound US EPA- 6 2 deisopropylatrazine product herbicide atrazine not approved ECOTOX selective herbicide, 7 Flurochloridone biosynthesis of carotenoids approved 3 ECOSAR inhibitor broad- US EPA- 8 Carbendazim spectrum benzimidazole not approved 3.3 ECOTOX fungicide central nervous system (CNS) stimulant of the US EPA- 9 Caffeine n.a. 5 methylxanthine class of ECOTOX psychoactive drugs parent compound metabolite of herbicide not approved in US EPA- 10 DCPMU 5 diuron most EU member ECOTOX states Information on the EU regulatory status retrieved from the Pesticide Properties Database (PPDB) by Lewis et al., 2015

Several pesticides, detected in the active fractions, were not approved in the Czech Republic for use in either plant protection products (PPPs) or biocides i.e. – prometryn and acetochlor. The triazine herbicide, prometryn, detected by GC(xGC)-MS, was ranked as number 3 in the list of likely toxicity drivers in the studied sample. This is an interesting finding because the use of prometryn in PPPs has been banned in the EU since 2003 (or since 2007 in some countries) and the compound is not registered for use in biocides either. Prometryn was reported to be present in effluents and the concentrations in the post ban years did not decrease (Quednow and Püttmann, 2009) and similar trend was reported also for atrazine (Vonberg et al., 2014). Possible SUMMARY OF RESULTS AND DISCUSSION 43

explanations for the continuous input of banned compounds to the environment could be the use of the old stocks, leaching from reservoirs in contaminated soils or the run-off from roof paints, plasters or other building materials enhanced by these compounds (Quednow and Püttmann, 2009; Venzmer, 2008). The results of this study indicated that active substances entering the aquatic environment from biocidal products should be carefully considered. Among the other top-ranking compounds following the group of herbicides, were industrial chemicals, other pesticides, caffeine and its metabolites, barbiturate drugs, macrolide antibiotics and other pharmaceuticals.

4.2.2 AChE inhibition (publication V) The optimized LCxLC system enhanced peak capacity and near perfect orthogonality to provide the basis for high resolution microfractionation in a relatively short time to support fast and accurate identification of active compounds after bioassay screening. Out of total 384 microfractions, 7 were active in the AChE bioassay (threshold 10% effect) and high resolution microfractionation greatly reduced the number of candidate chemicals in the active fractions (Fig. 7). Several pharmaceuticals and their metabolites were tentatively identified in the bioactive microfractions and among those, three pharmaceuticals with applications in neurology/psychiatry were selected for further investigation. Tiapride, lamotrigine and amisulpride were identified as the drivers of AChE inhibition based on the bioanalytical confirmation and earlier findings (Ferrer and Thurman, 2010; Fontaine and Reuse, 1980; Waldmeier et al., 1995; Wode et al., 2015).

Figure 7: Heatmap of AChE inhibition (%, n = 3) of the 384 microfractions of the Brno WWTP effluent extract after LC × LC fractionation. The numbers on the top are the plate column numbers (1−12) and the letter and number combinations to the left of the graph show the plate number together with the row number (A-H) of the 4 plates.

SUMMARY OF RESULTS AND DISCUSSION 45

4.3 Analytical and bioanalytical assessment of the Bosna River Of the 168 target compounds, 103 compounds were detected in extracts of samplers from at least one sampling site. Specifically, 71 out of 134 compounds (52.9%) were found in POCIS and 32 out of 34 (94.1%) in SPMD extracts. 65 (38.7%) compounds never exceeded their LOQ. There is a clear trend of decreasing cumulative concentration from S3 downstream to S10 in the POCIS samples, while no such pattern can be seen in case of hydrophobic compounds determined in the SPMD. Most analysed compounds were undetectable at the reference site S1 (spring of Bosna) with only a few compounds detected in concentrations near their LOQs. Within the class of PAHs, compounds with 3 and 4 condensed aromatic rings were detected at the greatest concentrations and occurred at all sampling sites. In case of CUPs (20 out of 52 compounds detected), the greatest concentrations were detected for diazinon and prometryn (53 ng L−1 and 17 ng L−1 at sites S2 and S5, respectively) and these two compounds occurred at all sampling sites except for the reference site (S1). Natural estrogens E1, E2 and E3 were detected at all sites in a range of 2.0×10-2-5.8 ng L−1 except for sites S1 and S5. EE2, a synthetic estrogen contained in hormonal contraceptives, was less than the LOQ of 1.0×10-2 – 3.0×10-2 ng L−1 at all sampling sites. This might be related to the 8- to 9-fold lesser prevalence of hormonal contraceptives in BiH compared to western European countries like Belgium or the UK due to cultural and economic reasons as well as limited availability of hormonal contraceptives (Boussen, 2012). Analyses of 77 pharmaceuticals in 11 subclasses resulted in detection of 47 compounds (61%) at least once. The most frequently detected compounds, disopyramide, carbamazepine and caffeine, occurred at all sampling sites. Bioanalytical assessment of the samples resulted in detection of anti- /androgenic, estrogenic and dioxin-like activities (Tab. 7) and the results were in line with earlier studies (Bain et al., 2014; Creusot et al., 2013; Jálová et al., 2013; König et al., 2017; Miège et al., 2009). A profile of integrated effects of mixtures at various locations in the Bosna RB, based on potencies observed in the three bioassays for extracts of the two types of sampler at each location were developed based on comparison to the reference site (S1). Contamination indices (CI), the ratio between the response of downstream sites (S2-S10) and a reference site (S1), in combination with the overall cumulative concentration and number of detected target compounds for 46

each sampling site are shown in Fig. 8. None of the sampling sites downstream of S1 can be considered as uncontaminated as the reference site because all sites exceeded a CI of 1.0 for at least two endpoints. The CI profiles, as well as the cumulative concentration and the number of detected hydrophobic compounds, differed less between individual sites in extracts of SPMD than in extracts of POCIS. Contamination indices indicate that the most contaminated sites were S2 and S3. The greatest cumulative concentrations and numbers of detected compounds were observed for the extract of the POCIS at S3 (a complete analysis for S2 was not available). This result implies that the major source of pollution to the Bosna River was Sarajevo (S2), the capital with a population of about 300,000 (Milinovic, 2013). The trend of decreasing cumulative concentrations in extracts of POCIS samplers downstream of S3 cannot be clearly seen in CIs and no patterns between CIs and cumulative concentrations in extracts of SPMD were observed, despite extensive, multi- residue analyses. Hazard quotients based on the NORMAN lowest PNEC values exceeded the threshold value 1.0 in case of 7 compounds, namely diazinon, diclofenac, E1, E2, benzo[b]fluoranthene, fluoranthene and benzo[k]fluoranthene. The overall hazard index (HI), resulting from the summation of all HQs at each sampling site, indicates that all sites downstream of the reference site S1 might cause adverse effects to aquatic biota as their HIs exceed 1.0.

Table 7. Results of the bioanalytical assessment.

Sampler Range of measured Nr. of Effect drivers (percentage Effect type bioactivity (Nr. of compounds with of effect explained) active sites) REP Androgenicity SPMD 4.2 – 10 (2) n.a. n.a. DHT-EQs [pg L−1] POCIS 2.1×102 -1.7×103(2) n.a. n.a. SPMD 3.4×104 -5.1×104 2 benzo[a]pyrene (0.03- Antiandrogenicity (5) 0.05%) FLU-EQ [pg L−1] POCIS 2.8×106 - 10 diazinon (0.04-0.06%) 3.2×106(3) SPMD n.d. Estrogenicity POCIS 2.3 ×102 - 2.5×103 8 natural estrogens - E1, E2 E2-EQ [pg L−1] (8) and E3 (0.84-305%) SPMD 9 benzo[b]fluoranthene, 2.9 - 7.3 (9) benzo[k]fluoranthene and Dioxin-like activity chrysene (6.1-24%) TCDD-EQ [pg L−1] POCIS 7 propiconazole (0.02- 31-2.2×102 (6) 0.04%) n.a. – not available, n.d. – not detected

SUMMARY OF RESULTS AND DISCUSSION 47

SPMD

400 35 ]

pM 30 [ 300 25 20 200 15 100 10 5

Sumconcentration 0 0

S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 Numbercompoundsofdetected Androgenicity 1 74.1 30.0 1 1 1 1 1 1 1 Anti-androgenicity 1 1 1 25.4 22.8 20.0 1 17.1 1 21.6 Estrogenicity 1 1 1 1 1 1 1 1 1 1 Dioxin-like activity 1 17.3 14.8 15.5 13.2 8.43 7.33 7.93 18.3 11.8

] POCIS

pM 5000 70 [ 60 4000 50 3000 40 2000 30 20 1000

10 Sum Sum concentration 0 0

S1 S2* S3 S4 S5 S6 S7 S8 S9 S10 Number of detected compounds Androgenicity 1 5628 707 1 1 1 1 1 1 1 Anti-androgenicity 1 1 1 1 1 1 31.8 30.6 1 27.2 Estrogenicity 1 42.6 194 83.6 86.4 71.9 122 25.5 17.7 1 Dioxin-like activity 1 2.52 1.11 0.87 0.36 0.95 1 1 1 1

0.1-1 Sum concentration 1 Nr of detected compounds 1-10 10-100 100-1000 Reference >1000 Fig. 8: Contamination profiles of sampling sites S2-S10 combined with the total number and cumulative concentration (pM) of detected compounds in SPMD (top) and POCIS (bottom) extracts. Colours green to red indicate to what extent the bioassay responses exceed the bioassay response of the reference site S1 (on a logarithmic scale). *77 pharmaceuticals and 16 target CUPs were not analysed in the POCIS extract from site S2

5 CONCLUSION AND FUTURE PROSPECTS

The presented dissertation demonstrated a successful application of a newly developed simplified effect-directed analysis (EDA) protocol. The novel on-site LVSPE sampling device proved to be a valuable instrument for integrative effect-based and chemical monitoring purposes. The set of selected bioassays enabled detection of effects relevant for surface waters on an EU- wide scale whereas the most frequently observed were estrogenicity, fish embryo toxicity and toxicity to algae. The findings of this dissertation confirmed that the non-target screening techniques may bring significant added value to the classical chemical target monitoring anchored in the current EU legislation (WFD). In the case of LC-MS non-target screening, however, the data have only qualitative character, and further research is needed to quantify the real concentrations of the identified compounds. This may be challenging especially in case of transformation products, for which analytical standards are hardly available. In this thesis, several compounds were identified as novel candidates for future surface water monitoring campaigns across Europe, e.g. ibuprofen, terbuthylazine, triphenylphosphate and nonylphenoxyacetic acid, as well as macrolide antibiotics such as azithromycin. The results also showed that many WFD priority compounds are of lesser concern, while several CECs, particularly from the class of pesticides and pharmaceuticals may pose risk to aquatic life. On the other hand, our studies indicated that several pesticides, highly prioritized by our research, were no longer registered for use in plant protection products or biocides, which indicated that their non-agricultural input into aquatic environment via WWTPs is not negligible and the sources of these compounds should be explored further. Most of the biological effects observed and described in the particular studies could not be explained by an extensive list of target compounds. which clearly demonstrated the need for implementation of the effect-based tools in surface water monitoring programs. Further investigation using higher tier EDA studies could possibly identify the yet unknown toxicity drivers of the observed effects. Another important aspect, which was only partially addressed by the research in this dissertation, is mixture effects. Understanding the interactions between compounds in complex environmental mixtures is crucial to proper evaluation of the water quality assessment based on individual components of the mixture. CONCLUSIONS AND FUTURE PROSPECTS 49

In conclusion, the objectives of the present dissertation were successfully addressed by conducted research. The studies demonstrated the advantages and needs to combine novel sampling tools, effects-based methods and chemical analyses including non—target screening, in order to address the myriad of contaminants of emerging concern and to achieve a comprehensive assessment of surface water quality. In future, the combined approaches should be further enhanced by application of modelling of mixtures responses and effect-directed analysis together with non-target identification of chemicals to prioritize the most relevant toxicants and effect drivers.

6 REFERENCES

Aerni, H.R., Kobler, B., Rutishauser, B. V., Wettstein, F.E., Fischer, R., Giger, W., Hungerbuhler, A., Marazuela, M.D., Peter, A., Schonenberger, R., Vogeli, A.C., Suter, M.J.F., Eggen, R.I.L., 2004. Combined biological and chemical assessment of estrogenic activities in wastewater treatment plant effluents. Anal. Bioanal. Chem. 378, 688–696. https://doi.org/10.1007/s00216-003-2276-4 Altenburger, R., Ait-Aissa, S., Antczak, P., Backhaus, T., Barceló, D., Seiler, T.-B., Brion, F., Busch, W., Chipman, K., de Alda, M.L., de Aragão Umbuzeiro, G., Escher, B.I., Falciani, F., Faust, M., Focks, A., Hilscherova, K., Hollender, J., Hollert, H., Jäger, F., Jahnke, A., Kortenkamp, A., Krauss, M., Lemkine, G.F., Munthe, J., Neumann, S., Schymanski, E.L., Scrimshaw, M., Segner, H., Slobodnik, J., Smedes, F., Kughathas, S., Teodorovic, I., Tindall, A.J., Tollefsen, K.E., Walz, K.-H., Williams, T.D., Van den Brink, P.J., van Gils, J., Vrana, B., Zhang, X., Brack, W., 2015. Future water quality monitoring — Adapting tools to deal with mixtures of pollutants in water resource management. Sci. Total Environ. 512–513, 540–551. https://doi.org/10.1016/j.scitotenv.2014.12.057 Altenburger, R., Brack, W., Burgess, R.M., Busch, W., Escher, B.I., Focks, A., Hewitt, L.M., Jacobsen, B.N., Alda, M.L. De, Aissa, S.A., Backhaus, T., Ginebreda, A., Hilscherová, K., Hollender, J., Hollert, H., Neale, P.A., Schulze, T., 2019. Future water quality monitoring : improving the balance between exposure and toxicity assessments of real ‑ world pollutant mixtures. Environ. Sci. Eur. 1–17. https://doi.org/10.1186/s12302-019-0193-1 Alvarez, D., Jones-Lepp, T., 2010. Sampling and Analysis of Emerging Pollutants, Water Quality Concepts, Sampling, and Analyses. CRC Press. https://doi.org/10.1201/b10157-12 Alvarez, D.A., Petty, J.D., Huckins, J.N., Jones-Lepp, T.L., Getting, D.T., Goddard, J.P., Manahan, S.E., 2004. Development of a passive, in situ, integrative sampler for hydrophilic organic contaminants in aquatic environments. Environ. Toxicol. Chem. 23, 1640–1648. https://doi.org/10.1897/03-603 Bain, P., Williams, M., Kumar, A., 2014. Assessment of multiple hormonal activities in wastewater at different stages of treatment. Environ. Toxicol. Chem. 33, 2297–2307. https://doi.org/10.1002/etc.2676 Balaguer, P., François, F., Comunale, F., Fenet, H., Boussioux, A.M., Pons, M., Nicolas, J.C., Casellas, C., 1999. Reporter cell lines to study the estrogenic effects of xenoestrogens. Sci. Total Environ. 233, 47–56. https://doi.org/10.1016/S0048-9697(99)00178-3 Bečanová, J., Komprdová, K., Vrana, B., Klánová, J., 2016. Annual dynamics of perfluorinated compounds in sediment: A case study in the Morava River in Zlín district, Czech Republic. Chemosphere 151, 225–233. https://doi.org/10.1016/j.chemosphere.2016.02.081 Bengtson Nash, S.M., Schreiber, U., Ralph, P.J., Müller, J.F., 2005. The combined SPE:ToxY-PAM phytotoxicity assay; application and appraisal of a novel biomonitoring tool for the aquatic environment. Biosens. Bioelectron. 20, 1443–1451. https://doi.org/10.1016/j.bios.2004.09.019 Booij, K., Robinson, C.D., Burgess, R.M., Mayer, P., Roberts, C.A., Ahrens, L., Allan, I.J., Brant, J., Jones, L., Kraus, U.R., Larsen, M.M., Lepom, P., Petersen, J., Pröfrock, D., Roose, P., Schäfer, S., Smedes, F., Tixier, C., Vorkamp, K., Whitehouse, P., 2016. Passive Sampling in Regulatory Chemical Monitoring of Nonpolar Organic Compounds in the Aquatic Environment. Environ. Sci. Technol. 50, 3–17. https://doi.org/10.1021/acs.est.5b04050 Booij, K., Smedes, F., 2010. An improved method for estimating in situ sampling rates of nonpolar passive samplers. Environ. Sci. Technol. 44, 6789–6794. Booij, P., Vethaak, D., Leonards, P.E.G., Sjollema, S.B., Kool, J., De Voogt, P., Lamoree, M.H., 2014. Identification of photosynthesis inhibitors of pelagic marine algae using 96-well plate microfractionation for enhanced throughput in effect-directed analysis. Environ. Sci. Technol. 48, 8003–8011. https://doi.org/10.1021/es405428t Brack, W., 2003. Effect-directed analysis: a promising tool for the identification of organic REFERENCES 51

toxicants in complex mixtures? Anal. Bioanal. Chem. 377, 397–407. Brack, W., Aissa, S.A., Backhaus, T., Dulio, V., Escher, B.I., Faust, M., Hilscherova, K., Hollender, J., Hollert, H., Müller, C., Munthe, J., 2019. Effect ‑ based methods are key . The European Collaborative Project SOLUTIONS recommends integrating effect ‑ based methods for diagnosis and monitoring of water quality. Environ. Sci. Eur. 4–9. https://doi.org/10.1186/s12302-019-0192-2 Brack, W., Ait-Aissa, S., Burgess, R.M., Busch, W., Creusot, N., Di Paolo, C., Escher, B.I., Mark Hewitt, L., Hilscherova, K., Hollender, J., Hollert, H., Jonker, W., Kool, J., Lamoree, M., Muschket, M., Neumann, S., Rostkowski, P., Ruttkies, C., Schollee, J., Schymanski, E.L., Schulze, T., Seiler, T.B., Tindall, A.J., De Aragão Umbuzeiro, G., Vrana, B., Krauss, M., 2016. Effect-directed analysis supporting monitoring of aquatic environments - An in-depth overview. Sci. Total Environ. 544, 1073–1118. https://doi.org/10.1016/j.scitotenv.2015.11.102 Brack, W., Altenburger, R., Schüürmann, G., Krauss, M., López Herráez, D., van Gils, J., Slobodnik, J., Munthe, J., Gawlik, B.M., van Wezel, A., Schriks, M., Hollender, J., Tollefsen, K.E., Mekenyan, O., Dimitrov, S., Bunke, D., Cousins, I., Posthuma, L., van den Brink, P.J., López de Alda, M., Barceló, D., Faust, M., Kortenkamp, A., Scrimshaw, M., Ignatova, S., Engelen, G., Massmann, G., Lemkine, G., Teodorovic, I., Walz, K.H., Dulio, V., Jonker, M.T.O., Jäger, F., Chipman, K., Falciani, F., Liska, I., Rooke, D., Zhang, X., Hollert, H., Vrana, B., Hilscherova, K., Kramer, K., Neumann, S., Hammerbacher, R., Backhaus, T., Mack, J., Segner, H., Escher, B., de Aragão Umbuzeiro, G., 2015. The SOLUTIONS project: Challenges and responses for present and future emerging pollutants in land and water resources management. Sci. Total Environ. 503–504, 22–31. https://doi.org/10.1016/j.scitotenv.2014.05.143 Brack, W., Govender, S., Schulze, T., Krauss, M., Hu, M., Muz, M., Hollender, J., Schirmer, K., Schollee, J., Hidasi, A., Slobodnik, J., Rabova, Z., Ait-Aissa, S., Sonavane, M., Carere, M., Lamoree, M., Leonards, P., Tufi, S., Ouyang, X., Schriks, M., Thomas, K., Almeida, A.C. De, Froment, J., Hammers-Wirtz, M., Ahel, M., Koprivica, S., Hollert, H., Seiler, T.-B., Paolo, C. Di, Tindall, A., Spirhanzlova, P., 2013. EDA-EMERGE: an FP7 initial training network to equip the next generation of young scientists with the skills to address the complexity of environmental contamination with emerging pollutants. Environ. Sci. Eur. 25, 1–7. https://doi.org/10.1186/2190-4715-25-18 Brausch, J.M., Rand, G.M., 2011. A review of personal care products in the aquatic environment: Environmental concentrations and toxicity. Chemosphere 82, 1518–1532. https://doi.org/10.1016/j.chemosphere.2010.11.018 Burgess, R.M., Ho, K.T., Brack, W., Lamoree, M., 2013. Effects-directed analysis (EDA) and toxicity identification evaluation (TIE): Complementary but different approaches for diagnosing causes of environmental toxicity. Environ. Toxicol. Chem. 32, 1935–1945. https://doi.org/10.1002/etc.2299 Connon, R.E., Geist, J., Werner, I., 2012. Effect-based tools for monitoring and predicting the ecotoxicological effects of chemicals in the aquatic environment. Sensors (Switzerland) 12, 12741–12771. https://doi.org/10.3390/s120912741 Creusot, N., Aït-Aïssa, S., Tapie, N., Pardon, P., Brion, F., Sanchez, W., Thybaud, E., Porcher, J.M., Budzinski, H., 2014. Identification of synthetic steroids in river water downstream from pharmaceutical manufacture discharges based on a bioanalytical approach and passive sampling. Environ. Sci. Technol. 48, 3649–3657. https://doi.org/10.1021/es405313r Creusot, N., Dévier, M.H., Budzinski, H., Aït-Aïssa, S., 2016. Evaluation of an extraction method for a mixture of endocrine disrupters in sediment using chemical and in vitro biological analyses. Environ. Sci. Pollut. Res. 23, 10349–10360. https://doi.org/10.1007/s11356- 016-6062-1 Creusot, N., Tapie, N., Piccini, B., Balaguer, P., Porcher, J.M., Budzinski, H., Aït-Aïssa, S., 2013. Distribution of steroid- and dioxin-like activities between sediments, POCIS and SPMD in a French river subject to mixed pressures. Environ. Sci. Pollut. Res. 20, 2784–2794. https://doi.org/10.1007/s11356-012-1452-5 Demirpence, E., Duchesne, M.-J., Badia, E., Gagne, D., Pons, M., 1993. MVLN Cells: A bioluminescent MCF-7-derived cell line to study the modulation of estrogenic activity. J. Steroid Biochem. Mol. Biol. 46, 355–364. https://doi.org/10.1016/0960-0760(93)90225-L 52

Dimpe, K.M., Nomngongo, P.N., 2016. Current sample preparation methodologies for analysis of emerging pollutants in different environmental matrices. TrAC - Trends Anal. Chem. 82, 199–207. https://doi.org/10.1016/j.trac.2016.05.023 DIRECTIVE 2000/60/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 23 October 2000 establishing a framework for Community action in the field of water policy., 2000. . Off. J. Eur. Union L327, 1e77. DIRECTIVE 2008/105/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 16 December 2008 on environmental quality standards in the field of water policy, amending and subsequently repealing Council Directives 82/176/EEC, 83/513/EEC, 84/156/EEC, 84/491/EEC, 2008. . Off. J. Eur. Union L 348/84. DIRECTIVE 2013/39/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 12 August 2013 amending Directives 2000/60/EC and 2008/105/EC as regards priority substances in the field of water policy, 2013. . Off. J. Eur. Union L 226/15. Dizer, H., Wittekindt, E., Fischer, B., Hansen, P.D., 2002. The cytotoxic and genotoxic potential of surface water and wastewater effluents as determined by bioluminescence, umu-assays and selected biomarkers. Chemosphere 46, 225–233. https://doi.org/10.1016/S0045- 6535(01)00062-5 Du, B., Lofton, J.M., Peter, K.T., Gipe, A.D., James, C.A., McIntyre, J.K., Scholz, N.L., Baker, J.E., Kolodziej, E.P., 2017. Development of suspect and non-target screening methods for detection of organic contaminants in highway runoff and fish tissue with high-resolution time-of-flight mass spectrometry. Environ. Sci. Process. Impacts 19, 1185–1196. https://doi.org/10.1039/c7em00243b Dulio, V., van Bavel, B., Brorström-Lundén, E., Harmsen, J., Hollender, J., Schlabach, M., Slobodnik, J., Thomas, K., Koschorreck, J., 2018. Emerging pollutants in the EU: 10 years of NORMAN in support of environmental policies and regulations. Environ. Sci. Eur. 30, 5. https://doi.org/10.1186/s12302-018-0135-3 Dulio, V., Von der Ohe, P.C. (Eds.), 2013. NORMAN prioritisation framework for emerging substances. [WWW Document]. URL http://www.norman- network.net/sites/default/files/files/Publications/NORMAN_prioritisation_Manual_15 April2013_final for website-f.pdf (accessed 10.16.15). Ellman, G.L., Courtney, K.D., Andres, V., Featherstone, R.M., 1961. A new and rapid colorimetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 7, 88–95. Emelogu, E.S., Pollard, P., Dymond, P., Robinson, C.D., Webster, L., McKenzie, C., Dobson, J., Bresnan, E., Moffat, C.F., 2013. Occurrence and potential combined toxicity of dissolved organic contaminants in the Forth estuary and Firth of Forth, Scotland assessed using passive samplers and an algal toxicity test. Sci. Total Environ. 461–462, 230–239. https://doi.org/10.1016/j.scitotenv.2013.05.011 Escher, B.I., Aїt-Aїssa, S., Behnisch, P.A., Brack, W., Brion, F., Brouwer, A., Buchinger, S., Crawford, S.E., Du Pasquier, D., Hamers, T., Hettwer, K., Hilscherová, K., Hollert, H., Kase, R., Kienle, C., Tindall, A.J., Tuerk, J., van der Oost, R., Vermeirssen, E., Neale, P.A., 2018. Effect-based trigger values for in vitro and in vivo bioassays performed on surface water extracts supporting the environmental quality standards (EQS) of the European Water Framework Directive. Sci. Total Environ. 628–629, 748–765. https://doi.org/10.1016/j.scitotenv.2018.01.340 Escher, B.I., Pronk, W., Suter, M.J.F., Maurer, M., 2006. Monitoring the removal efficiency of pharmaceuticals and hormones in different treatment processes of source-separated urine with bioassays. Environ. Sci. Technol. 40, 5095–5101. Escher, Beate I, Quayle, P., Muller, R., Schreiber, U., Mueller, J.F., 2006. Passive sampling of herbicides combined with effect analysis in algae using a novel high-throughput phytotoxicity assay (Maxi-Imaging-PAM). J. Environ. Monit. 8, 456–464. https://doi.org/10.1039/b517512g European Commission, 2003. Document on Risk Assessment. Tech. Guid. Doc. Ris Assess. Part II 337. https://doi.org/10.1002/mp.12308 European Environment Agency, 2018. European waters — Assessment of status and pressures 2018. https://doi.org/0.2800/303664 REFERENCES 53

Fairbairn, D.J., Elliott, S.M., Kiesling, R.L., Schoenfuss, H.L., Ferrey, M.L., Westerhoff, B.M., 2018. Contaminants of emerging concern in urban stormwater: Spatiotemporal patterns and removal by iron-enhanced sand filters (IESFs). Water Res. 145, 332–345. https://doi.org/10.1016/j.watres.2018.08.020 Ferrer, I., Thurman, E.M., 2010. Identification of a New Antidepressant and its Glucuronide Metabolite in Water Samples Using Liquid Chromatography/Quadrupole Time-of-Flight Mass Spectrometry. Anal. Chem. 82, 8161–8168. https://doi.org/10.1021/ac1014645 Fini, J.B., Le Mével, S., Turque, N., Palmier, K., Zalko, D., Cravedi, J.P., Demeneix, B.A., 2007. An in vivo multiwell-based fluorescent screen for monitoring vertebrate thyroid hormone disruption. Environ. Sci. Technol. 41, 5908–5914. Fontaine, J., Reuse, J., 1980. The effects of substituted benzamides on frog rectus abdominis. Eur. J. Pharmacol. 68, 55–60. https://doi.org/10.1016/0014-2999(80)90060-6 Fontanals, N., Marcé, R.M., Borrull, F., 2010. Núria Fontanals, Rosa Maria Marcé, Francesc Borrull_2010 6, 199–213. https://doi.org/10.2436/20.7010.01.97 Galgani, F., Bocquene, G., 1991. Semi-automated colorimetric and enzymaic assays for aquatic organisms using microplate readers. Water Res. 25, 147–150. García-Córcoles, M.T., Rodríguez-Gómez, R., de Alarcón-Gómez, B., Çipa, M., Martín-Pozo, L., Kauffmann, J.M., Zafra-Gómez, A., 2019. Chromatographic Methods for the Determination of Emerging Contaminants in Natural Water and Wastewater Samples: A Review. Crit. Rev. Anal. Chem. 49, 160–186. https://doi.org/10.1080/10408347.2018.1496010 Geissen, V., Mol, H., Klumpp, E., Umlauf, G., Nadal, M., van der Ploeg, M., van de Zee, S.E. a. T.M., Ritsema, C.J., 2015. Emerging pollutants in the environment: A challenge for water resource management. Int. Soil Water Conserv. Res. 3, 57–65. https://doi.org/10.1016/j.iswcr.2015.03.002 Ginebreda, A., Kuzmanovic, M., Guasch, H., de Alda, M.L., López-Doval, J.C., Muñoz, I., Ricart, M., Romaní, A.M., Sabater, S., Barceló, D., 2014. Assessment of multi-chemical pollution in aquatic ecosystems using toxic units: Compound prioritization, mixture characterization and relationships with biological descriptors. Sci. Total Environ. 468–469, 715–723. https://doi.org/10.1016/j.scitotenv.2013.08.086 Gómez, M.J., Gómez-Ramos, M.M., Agüera, a., Mezcua, M., Herrera, S., Fernández-Alba, a. R., 2009. A new gas chromatography/mass spectrometry method for the simultaneous analysis of target and non-target organic contaminants in waters. J. Chromatogr. A 1216, 4071–4082. https://doi.org/10.1016/j.chroma.2009.02.085 Górecki, T., Namienik, J., 2002. Passive sampling. TrAC - Trends Anal. Chem. 21, 276–291. https://doi.org/10.1016/S0165-9936(02)00407-7 Greaves, A.K., Letcher, R.J., 2017. A Review of Organophosphate Esters in the Environment from Biological Effects to Distribution and Fate. Bull. Environ. Contam. Toxicol. 98, 2–7. https://doi.org/10.1007/s00128-016-1898-0 Griffith, D.R., Kido Soule, M.C., Matsufuji, H., Eglinton, T.I., Kujawinski, E.B., Gschwend, P.M., 2014. Measuring free, conjugated, and halogenated estrogens in secondary treated wastewater effluent. Environ. Sci. Technol. 48, 2569–2578. https://doi.org/10.1021/es402809u Guijarro, C., Fuchs, K., Bohrn, U., Stütz, E., Wölfl, S., 2015. Simultaneous detection of multiple bioactive pollutants using a multiparametric biochip for water quality monitoring. Biosens. Bioelectron. 72, 71–79. https://doi.org/10.1016/j.bios.2015.04.092 Hamers, T., Leonards, P.E.G., Legler, J., Dick Vethaak, A., Schipper, C.A., 2010. Toxicity profiling: An integrated effect-based tool for site-specific sediment quality assessment. Integr. Environ. Assess. Manag. 6, 761–773. https://doi.org/10.1002/ieam.75 Harman, C., Allan, I.J., Vermeirssen, E.L.M., 2012. Calibration and use of the polar organic chemical integrative sampler-a critical review. Environ. Toxicol. Chem. 31, 2724–2738. https://doi.org/10.1002/etc.2011 Hollender, J., van Bavel, B., Dulio, V., Farmen, E., Furtmann, K., Koschorreck, J., Kunkel, U., Krauss, M., Munthe, J., Schlabach, M., Slobodnik, J., Stroomberg, G., Ternes, T., Thomaidis, N.S., Togola, A., Tornero, V., 2019. High resolution mass spectrometry-based non-target screening can support regulatory environmental monitoring and chemicals management. Environ. Sci. Eur. 31. https://doi.org/10.1186/s12302-019-0225-x 54

Houtman, C.J., 2010. Emerging contaminants in surface waters and their relevance for the production of drinking water in Europe. J. Integr. Environ. Sci. 7, 271–295. https://doi.org/10.1080/1943815X.2010.511648 Huckins, J.N., Manuweera, G.K., Petty, J.D., Mackay, D., Lebo, J.A., 1993. Lipid-containing semipermeable membrane devices for monitoring organic contaminants in water. Environ. Sci. Technol. 27, 2489–2496. Hug, C., Ulrich, N., Schulze, T., Brack, W., Krauss, M., 2014. Identification of novel micropollutants in wastewater by a combination of suspect and nontarget screening. Environ. Pollut. 184, 25–32. https://doi.org/10.1016/j.envpol.2013.07.048 Ibáñez, M., Sancho, J. V., Hernández, F., McMillan, D., Rao, R., 2008. Rapid non-target screening of organic pollutants in water by ultraperformance liquid chromatography coupled to time- of-light mass spectrometry. TrAC - Trends Anal. Chem. 27, 481–489. https://doi.org/10.1016/j.trac.2008.03.007 Jálová, V., Jarošová, B., Bláha, L., Giesy, J.P., Ocelka, T., Grabic, R., Jurčíková, J., Vrana, B., Hilscherová, K., 2013. Estrogen-, androgen- and aryl hydrocarbon receptor mediated activities in passive and composite samples from municipal waste and surface waters. Environ. Int. 59, 372–383. https://doi.org/10.1016/j.envint.2013.06.024 Jalova, V., Jarosova, B., Blaha, L., Giesy, J.P.P., Ocelka, T., Grabic, R., Jurcikova, J., Vrana, B., Hilscherova, K., Jálová, V., Jarošová, B., Bláha, L., Giesy, J.P.P., Ocelka, T., Grabic, R., Jurčíková, J., Vrana, B., Hilscherová, K., 2013. Estrogen-, androgen- and aryl hydrocarbon receptor mediated activities in passive and composite samples from municipal waste and surface waters. Environ. Int. 59, 372–83. https://doi.org/10.1016/j.envint.2013.06.024 Jarosova, B., Blaha, L., Vrana, B., Randak, T., Grabic, R., Giesy, J.P., Hilscherova, K., 2012. Changes in concentrations of hydrophilic organic contaminants and of endocrine-disrupting potential downstream of small communities located adjacent to headwaters. Environ. Int. 45, 22–31. https://doi.org/10.1016/j.envint.2012.04.001 Jones, L., Ronan, J., McHugh, B., McGovern, E., Regan, F., 2015. Emerging priority substances in the aquatic environment: a role for passive sampling in supporting WFD monitoring and compliance. Anal. Methods 7, 7976–7984. https://doi.org/10.1039/C5AY01059D Kinani, S., Bouchonnet, S., Creusot, N., Bourcier, S., Balaguer, P., Porcher, J.-M., Aït-Aïssa, S., 2010. Bioanalytical characterisation of multiple endocrine- and dioxin-like activities in sediments from reference and impacted small rivers. Environ. Pollut. 158, 74–83. https://doi.org/10.1016/j.envpol.2009.07.041 Koenig, M., Escher, B.I., Neale, P.A., Krauss, M., Hilscherova, K., Novak, J., Teodorovic, I., Schulze, T., Seidensticker, S., Kamal Hashmi, M.A., Ahlheim, J., Brack, W., 2017. Impact of untreated wastewater on a major European river evaluated with a combination of in??vitro bioassays and chemical analysis. Environ. Pollut. 220, 1220–1230. https://doi.org/10.1016/j.envpol.2016.11.011 Köhler, A., Hellweg, S., Escher, B.I., Hungerbühler, K., 2006. Organic pollutant removal versus toxicity reduction in industrial wastewater treatment: The example of wastewater from fluorescent whitening agent production. Environ. Sci. Technol. 40, 3395–3401. https://doi.org/10.1021/es060555f König, M., Escher, B.I., Neale, P.A., Krauss, M., Hilscherová, K., Novák, J., Teodorović, I., Schulze, T., Seidensticker, S., Kamal Hashmi, M.A., Ahlheim, J., Brack, W., 2017. Impact of untreated wastewater on a major European river evaluated with a combination of in vitro bioassays and chemical analysis. Environ. Pollut. 220, 1220–1230. https://doi.org/10.1016/j.envpol.2016.11.011 Krauss, M., Singer, H., Hollender, J., 2010. LC-high resolution MS in environmental analysis: From target screening to the identification of unknowns. Anal. Bioanal. Chem. 397, 943–951. https://doi.org/10.1007/s00216-010-3608-9 Křesinová, Z., Petrů, K., Lhotský, O., Rodsand, T., Cajthaml, T., 2016. Passive sampling of pharmaceuticals and personal care products in aquatic environments. Eur. J. Environ. Sci. 6, 43–56. https://doi.org/10.14712/23361964.2016.8 Kuzmanovic, M., Ginebreda, A., Petrovic, M., Barcelo, D., 2014. Risk assessment based prioritization of 200 organic micropollutants in 4 Iberian rivers. Sci. Total Environ. 504, REFERENCES 55

289–299. https://doi.org/10.1016/j.scitotenv.2014.06.056 Letsinger, S., Kay, P., 2019. Comparison of Prioritisation Schemes for Human Pharmaceuticals in the Aquatic Environment. Environ. Sci. Pollut. Res. 26, 3479–3491. https://doi.org/10.1007/s11356-018-3834-9 Liang, X., Adamovsky, O., Souders, C.L., Martyniuk, C.J., 2019. Biological effects of the benzotriazole ultraviolet stabilizers UV-234 and UV-320 in early-staged zebrafish (Danio rerio). Environ. Pollut. 245, 272–281. https://doi.org/10.1016/j.envpol.2018.10.130 Linares, V., Bellés, M., Domingo, J.L., 2015. Human exposure to PBDE and critical evaluation of health hazards. Arch. Toxicol. 89, 335–356. https://doi.org/10.1007/s00204-015-1457-1 Loos, R., Carvalho, R., António, D.C., Comero, S., Locoro, G., Tavazzi, S., Paracchini, B., Ghiani, M., Lettieri, T., Blaha, L., Jarosova, B., Voorspoels, S., Servaes, K., Haglund, P., Fick, J., Lindberg, R.H., Schwesig, D., Gawlik, B.M., 2013. EU-wide monitoring survey on emerging polar organic contaminants in wastewater treatment plant effluents. Water Res. 47, 6475–6487. https://doi.org/10.1016/j.watres.2013.08.024 Magdaleno, A., Juárez, Á.B., Dragani, V., Saenz, M.E., Paz, M., Moretton, J., 2014. Ecotoxicological and Genotoxic Evaluation of Buenos Aires City ( Argentina ) Hospital Wastewater 2014. Martín-Pozo, L., de Alarcón-Gómez, B., Rodríguez-Gómez, R., García-Córcoles, M.T., Çipa, M., Zafra- Gómez, A., 2019. Analytical methods for the determination of emerging contaminants in sewage sludge samples. A review. Talanta 192, 508–533. https://doi.org/10.1016/j.talanta.2018.09.056 Maselli, B.D.S., Luna, L.A.V. De, Palmeira, J.D.O., Babosa, S., Beijo, L.A., Umbuzeiro, G.D.A., Kummrow, F., 2013. Ecotoxicity of raw and treated effluents generated by a veterinary medicine industry. Ambient. e Agua - An Interdiscip. J. Appl. Sci. 8, 795–804. https://doi.org/10.4136/ambi-agua.1121 Miège, C., Gabet, V., Coquery, M., Karolak, S., Jugan, M.L., Oziol, L., Levi, Y., Chevreuil, M., 2009. Evaluation of estrogenic disrupting potency in aquatic environments and urban wastewaters by combining chemical and biological analysis. TrAC - Trends Anal. Chem. 28, 186–195. https://doi.org/10.1016/j.trac.2008.11.007 Milinovic, Z., 2013. PRELIMINARY RESULTS Of the 2013 Census of Population, Households and Dwellings in Bosnia and Herzegovina [WWW Document]. Agency Stat. Bosnia Herzegovina. URL http://www.bhas.ba/obavjestenja/Preliminarni_rezultati_bos.pdf (accessed 10.10.17). Moschet, C., Vermeirssen, E.L.M., Seiz, R., Pfefferli, H., Hollender, J., 2014. Picogram per liter detections of pyrethroids and organophosphates in surface waters using passive sampling. Water Res. 66, 411–422. https://doi.org/10.1016/j.watres.2014.08.032 Nagy, S.R., Sanborn, J.R., Hammock, B.D., Denison, M.S., 2002. Development of a green fluorescent protein-based cell bioassay for the rapid and inexpensive detection and characterization of Ah receptor agonists. Toxicol. Sci. 65, 200–210. https://doi.org/10.1093/toxsci/65.2.200 Neale, P.A., Ait-Aissa, S., Brack, W., Creusot, N., Denison, M.S., Deutschmann, B., Hilscherova, K., Hollert, H., Krauss, M., Novak, J., Schulze, T., Seiler, T.B., Serra, H., Shao, Y., Escher, B.I., 2015. Linking in Vitro Effects and Detected Organic Micropollutants in Surface Water Using Mixture-Toxicity Modeling. Environ. Sci. Technol. 49, 14614–14624. https://doi.org/10.1021/acs.est.5b04083 Novák, J., Vrana, B., Rusina, T., Okonski, K., Grabic, R., Neale, P.A., Escher, B.I., Macová, M., Ait- Aissa, S., Creusot, N., Allan, I., Hilscherová, K., 2018. Effect-based monitoring of the Danube River using mobile passive sampling. Sci. Total Environ. 636, 1608–1619. https://doi.org/10.1016/j.scitotenv.2018.02.201 OECD, 2013. Test No. 236: Fish Embryo Acute Toxicity (FET) Test [WWW Document]. https://doi.org/http://dx.doi.org/10.1787/9789264203709-en OECD, 2011. Test No. 201: Freshwater Alga and Cyanobacteria, Growth Inhibition Test [WWW Document]. https://doi.org/http://dx.doi.org/10.1787/9789264069923-en OECD, 2002. Hazard assessment of perfluorooctane sulfonate (PFOS) and its salts.Co-operation on existing chemicals. Dev. 362. Ohoro, Adeniji, Okoh, Okoh, 2019. Distribution and Chemical Analysis of Pharmaceuticals and Personal Care Products (PPCPs) in the Environmental Systems: A Review. Int. J. Environ. 56

Res. Public Health 16, 3026. https://doi.org/10.3390/ijerph16173026 Osman, A.G.M., AbouelFadl, K.Y., Krüger, A., Kloas, W., 2015. Screening of multiple hormonal activities in water and sediment from the river Nile, Egypt, using in vitro bioassay and gonadal histology. Environ. Monit. Assess. 187, 317. https://doi.org/10.1007/s10661-015- 4553-z Pal, A., Gin, K.Y.-H., Lin, A.Y.-C., Reinhard, M., 2010. Impacts of emerging organic contaminants on freshwater resources: review of recent occurrences, sources, fate and effects. Sci. Total Environ. 408, 6062–9. https://doi.org/10.1016/j.scitotenv.2010.09.026 Patel, M., Kumar, R., Kishor, K., Mlsna, T., Pittman, C.U., Mohan, D., 2019. Pharmaceuticals of emerging concern in aquatic systems: Chemistry, occurrence, effects, and removal methods. Chem. Rev. 119, 3510–3673. https://doi.org/10.1021/acs.chemrev.8b00299 Petrovic, M., Barceló, D., Diaz, A., Ventura, F., 2003. Low nanogram per liter determination of halogenated nonylphenols, nonylphenol carboxylates, and their non-halogenated precursors in water and sludge by liquid chromatography electrospray tandem mass spectrometry. J. Am. Soc. Mass Spectrom. 14, 516–527. https://doi.org/10.1016/S1044- 0305(03)00139-9 Pietrzak, D., Kania, J., Malina, G., Kmiecik, E., Wątor, K., 2019. Pesticides from the EU First and Second Watch Lists in the Water Environment. Clean - Soil, Air, Water 47. https://doi.org/10.1002/clen.201800376 Posthuma, L., Altenburger, R., Backhaus, T., Kortenkamp, A., Müller, C., Focks, A., de Zwart, D., Brack, W., 2019. Improved component-based methods for mixture risk assessment are key to characterize complex chemical pollution in surface waters. Environ. Sci. Eur. 31. https://doi.org/10.1186/s12302-019-0246-5 Quednow, K., Püttmann, W., 2009. Temporal concentration changes of DEET, TCEP, terbutryn, and nonylphenols in freshwater streams of Hesse, Germany: Possible influence of mandatory regulations and voluntary environmental agreements. Environ. Sci. Pollut. Res. 16, 630–640. https://doi.org/10.1007/s11356-009-0169-6 Rainieri, S., Conlledo, N., Langerholc, T., Madorran, E., Sala, M., Barranco, A., 2017. Toxic effects of perfluorinated compounds at human cellular level and on a model vertebrate. Food Chem. Toxicol. 104, 14–25. https://doi.org/10.1016/j.fct.2017.02.041 Reichenberg, F., Mayer, P., 2006. Two complementary sides of bioavailability: accessibility and chemical activity of organic contaminants in sediments and soils. Environ. Toxicol. Chem. 25, 1239–45. Ren, Y., Geng, J., Li, F., Ren, H., Ding, L., Xu, K., 2016. The oxidative stress in the liver of Carassius auratus exposed to acesulfame and its UV irradiance products. Sci. Total Environ. 571, 755–762. https://doi.org/10.1016/j.scitotenv.2016.07.047 Richardson, S.D., Ternes, T.A., 2018. Water Analysis: Emerging Contaminants and Current Issues. Anal. Chem. 90, 398–428. https://doi.org/10.1021/acs.analchem.7b04577 Riva, F., Zuccato, E., Davoli, E., Fattore, E., Castiglioni, S., 2019. Risk assessment of a mixture of emerging contaminants in surface water in a highly urbanized area in Italy. J. Hazard. Mater. 361, 103–110. https://doi.org/10.1016/j.jhazmat.2018.07.099 Rojickova, R., Dvorakova, D., 1998. The Use of Miniaturized Algal Bioassays in Comparison to the Standard Flask Assay. Env. Toxicol Water Qual 13, 235–241. Rusina, T.P., Smedes, F., Koblizkova, M., Klanova, J., 2010. Calibration of silicone rubber passive samplers: Experimental and modeled relations between sampling rate and compound properties. Environ. Sci. Technol. 44, 362–367. Rykowska, I., Wasiak, W., 2015. Research trends on emerging environment pollutants - A review. Open Chem. 13, 1353–1370. https://doi.org/10.1515/chem-2015-0151 Salimi, M., Esrafili, A., Gholami, M., Jonidi Jafari, A., Rezaei Kalantary, R., Farzadkia, M., Kermani, M., Sobhi, H.R., 2017. Contaminants of emerging concern: a review of new approach in AOP technologies. Environ. Monit. Assess. 189. https://doi.org/10.1007/s10661-017-6097-x Saucedo-Vence, K., Elizalde-Velázquez, A., Dublán-García, O., Galar-Martínez, M., Islas-Flores, H., SanJuan-Reyes, N., García-Medina, S., Hernández-Navarro, M.D., Gómez-Oliván, L.M., 2017. Toxicological hazard induced by sucralose to environmentally relevant concentrations in common carp (Cyprinus carpio). Sci. Total Environ. 575, 347–357. REFERENCES 57

https://doi.org/10.1016/j.scitotenv.2016.09.230 Scheurer, M., Brauch, H.J., Lange, F.T., 2009. Analysis and occurrence of seven artificial sweeteners in German waste water and surface water and in soil aquifer treatment (SAT). Anal. Bioanal. Chem. 394, 1585–1594. https://doi.org/10.1007/s00216-009-2881-y Schirmer, K., Chan, A.G.J., Greenberg, B.M., Dixon, D.G., Bols, N.C., 1998. Ability of 16 priority PAHs to be photocytotoxic to a cell line from the rainbow trout gill. Toxicology 127, 143–155. https://doi.org/10.1016/S0300-483X(98)00031-6 Schriks, M., Van Leerdam, J.A., Van Der Linden, S.C., Van Der Burg, B., Van Wezel, A.P., De Voogt, P., 2010. High-resolution mass spectrometric identification and quantification of glucocorticoid compounds in various wastewaters in the Netherlands. Environ. Sci. Technol. 44, 4766–4774. https://doi.org/10.1021/es100013x Schulze, Tobias, Ahel, M., Ahlheim, J., Aït-Aïssa, S., Brion, F., Di Paolo, C., Froment, J., Hidasi, A.O., Hollender, J., Hollert, H., Hu, M., Kloß, A., Koprivica, S., Krauss, M., Muz, M., Oswald, P., Petre, M., Schollée, J.E., Seiler, T.-B., Shao, Y., Slobodnik, J., Sonavane, M., Suter, M.J.-F., Tollefsen, K.E., Tousova, Z., Walz, K.-H., Brack, W., 2017. Assessment of a novel device for onsite integrative large-volume solid phase extraction of water samples to enable a comprehensive chemical and effect-based analysis. Sci. Total Environ. 581–582, 350–358. https://doi.org/10.1016/j.scitotenv.2016.12.140 Schulze, T., Ahel, M., Ahlheim, J., Aït-Aïssa, S., Brion, F., Di Paolo, C., Froment, J., Hidasi, A.O., Hollender, J., Hollert, H., Hu, M., Kloß, A., Koprivica, S., Krauss, M., Muz, M., Oswald, P., Petre, M., Schollée, J.E., Seiler, T.-B., Shao, Y., Slobodnik, J., Sonavane, M., Suter, M.J.-F., Tollefsen, K.E., Tousova, Z., Walz, K.-H., Brack, W., 2017. Assessment of a novel device for onsite integrative large-volume solid phase extraction of water samples to enable a comprehensive chemical and effect-based analysis. Sci. Total Environ. 581–582. https://doi.org/10.1016/j.scitotenv.2016.12.140 Schymanski, E.L., Singer, H.P., Longrée, P., Loos, M., Ruff, M., Stravs, M. a., Ripollés Vidal, C., Hollender, J., 2014. Strategies to characterize polar organic contamination in wastewater: Exploring the capability of high resolution mass spectrometry. Environ. Sci. Technol. 48, 1811–1818. https://doi.org/10.1021/es4044374 Schymanski, E.L., Singer, H.P., Slobodnik, J., Ipolyi, I.M., Oswald, P., Krauss, M., Schulze, T., Haglund, P., Letzel, T., Grosse, S., Thomaidis, N.S., Bletsou, A., Zwiener, C., Ibáñez, M., Portolés, T., de Boer, R., Reid, M.J., Onghena, M., Kunkel, U., Schulz, W., Guillon, A., Noyon, N., Leroy, G., Bados, P., Bogialli, S., Stipaničev, D., Rostkowski, P., Hollender, J., 2015. Non-target screening with high-resolution mass spectrometry: critical review using a collaborative trial on water analysis. Anal. Bioanal. Chem. 6237–6255. https://doi.org/10.1007/s00216- 015-8681-7 Shi, Z.Q., Liu, Y.S., Xiong, Q., Cai, W.W., Ying, G.G., 2019. Occurrence, toxicity and transformation of six typical benzotriazoles in the environment: A review. Sci. Total Environ. 661, 407–421. https://doi.org/10.1016/j.scitotenv.2019.01.138 Sima, L., Amador, J., Silva, A.D.K., Miller, S.M., Morse, A.N., Pellegrin, M.-L., Rock, C., Wells, M.J.M., 2014. Emerging Pollutants – Part I: Occurrence, Fate and Transport. Water Environ. Res. 86, 1994–2035. https://doi.org/10.2175/106143014X14031280668731 Slobodnik, J., Mrafkova, L., Carere, M., Ferrara, F., Pennelli, B., Schüürmann, G., von der Ohe, P.C., 2012. Identification of river basin specific pollutants and derivation of environmental quality standards: A case study in the Slovak Republic. TrAC Trends Anal. Chem. 41, 133– 145. https://doi.org/10.1016/j.trac.2012.08.008 Smital, T., Terzić, S., Lončar, J., Senta, I., Žaja, R., Popović, M., Mikac, I., Tollefsen, K.-E., Thomas, K. V., Ahel, M., 2013. Prioritisation of organic contaminants in a river basin using chemical analyses and bioassays. Environ. Sci. Pollut. Res. 20, 1384–1395. https://doi.org/10.1007/s11356-012-1059-x Sonneveld, E., Jansen, H.J., Riteco, J.A.C., Brouwer, A., van der Burg, B., 2005. Development of androgen- and estrogen-responsive bioassays members of a panel of human cell line-based highly selective steroid-responsive bioassays. Toxicol. Sci. 83, 136–148. https://doi.org/10.1093/toxsci/kfi005 Tang, J.Y.M., Escher, B.I., 2014. Realistic environmental mixtures of micropollutants in surface, 58

drinking, and recycled water: Herbicides dominate the mixture toxicity toward algae. Environ. Toxicol. Chem. 33, 1427–1436. https://doi.org/10.1002/etc.2580 Tangtian, H., Bo, L., Wenhua, L., Shin, P.K.S., Wu, R.S.S., 2012. Estrogenic potential of benzotriazole on marine medaka (Oryzias melastigma). Ecotoxicol. Environ. Saf. 80, 327–332. https://doi.org/10.1016/j.ecoenv.2012.03.020 Taylor, A.C., Fones, G.R., Vrana, B., Mills, G.A., 2019. Applications for Passive Sampling of Hydrophobic Organic Contaminants in Water—A Review. Crit. Rev. Anal. Chem. 0, 1–35. https://doi.org/10.1080/10408347.2019.1675043 Thomaidis, N.S., Asimakopoulos, A.G., Bletsou, A.A., 2012. Emerging contaminants: A tutorial mini-review. Glob. Nest J. 14, 72–79. https://doi.org/10.30955/gnj.000823 Tijani, J.O., Fatoba, O.O., Babajide, O.O., Petrik, L.F., 2016. Pharmaceuticals, endocrine disruptors, personal care products, nanomaterials and perfluorinated pollutants: a review. Environ. Chem. Lett. 14, 27–49. https://doi.org/10.1007/s10311-015-0537-z Tousova, Z., Oswald, P., Slobodnik, J., Blaha, L., Muz, M., Hu, M., Brack, W., Krauss, M., Di Paolo, C., Tarcai, Z., Seiler, T.-B., Hollert, H., Koprivica, S., Ahel, M., Schollée, J.E., Hollender, J., Suter, M.J.-F., Hidasi, A.O., Schirmer, K., Sonavane, M., Ait-Aissa, S., Creusot, N., Brion, F., Froment, J., Almeida, A.C., Thomas, K., Tollefsen, K.E., Tufi, S., Ouyang, X., Leonards, P., Lamoree, M., Torrens, V.O., Kolkman, A., Schriks, M., Spirhanzlova, P., Tindall, A., Schulze, T., 2017. European demonstration program on the effect-based and chemical identification and monitoring of organic pollutants in European surface waters. Sci. Total Environ. 601–602, 1849–1868. https://doi.org/10.1016/j.scitotenv.2017.06.032 US-EPA, 2015. ECOTOX Database [WWW Document]. URL http://cfpub.epa.gov/ecotox/ (accessed 4.20.15). US-EPA, 2012. ECOSAR 1.11 [WWW Document]. URL http://www.epa.gov/oppt/newchems/tools/21ecosar.htm Venzmer, H.V., 2008. Forum Altbausanierung 2. Biofilme und funktionale Baustoffoberflächen.: 8. Dahlberg-Kolloquium vom 25. bis 26. September 2008 im Zeughaus Wismar. Fraunhofer IRB Verlag. Vera-Candioti, L., Gil García, M.D., Martínez Galera, M., Goicoechea, H.C., 2008. Chemometric assisted solid-phase microextraction for the determination of anti-inflammatory and antiepileptic drugs in river water by liquid chromatography-diode array detection. J. Chromatogr. A 1211, 22–32. https://doi.org/10.1016/j.chroma.2008.09.093 Vermeirssen, E.L.M., Hollender, J., Bramaz, N., Van Der Voet, J., Escher, B.I., 2010. Linking toxicity in algal and bacterial assays with chemical analysis in passive samplers deployed in 21 treated sewage effluents. Environ. Toxicol. Chem. 29, 2575–2582. https://doi.org/10.1002/etc.311 Von der Ohe, P.C., Dulio, V., Slobodnik, J., De Deckere, E., Kühne, R., Ebert, R.U., Ginebreda, A., De Cooman, W., Schüürmann, G., Brack, W., 2011. A new risk assessment approach for the prioritization of 500 classical and emerging organic microcontaminants as potential river basin specific pollutants under the European Water Framework Directive. Sci. Total Environ. 409, 2064–2077. https://doi.org/10.1016/j.scitotenv.2011.01.054 Vonberg, D., Vanderborght, J., Cremer, N., Pütz, T., Herbst, M., Vereecken, H., 2014. 20 years of long-term atrazine monitoring in a shallow aquifer in western Germany. Water Res. 50. https://doi.org/10.1016/j.watres.2013.10.032 Vrana, B., Allan, I.J., Greenwood, R., Mills, G.A., Dominiak, E., Svensson, K., Knutsson, J., Morrison, G., 2005. Passive sampling techniques for monitoring pollutants in water. TrAC - Trends Anal. Chem. 24, 845–868. Vrana, Branislav, Klučárová, V., Benická, E., Abou-Mrad, N., Amdany, R., Horáková, S., Draxler, A., Humer, F., Gans, O., 2014. Passive sampling: An effective method for monitoring seasonal and spatial variability of dissolved hydrophobic organic contaminants and metals in the Danube river. Environ. Pollut. 184, 101–112. https://doi.org/10.1016/j.envpol.2013.08.018 Vrana, B, Klučárová, V., Benická, E., Abou-Mrad, N., Amdany, R., Horáková, S., Draxler, A., Humer, F., Gans, O., 2014. Passive sampling: An effective method for monitoring seasonal and spatial variability of dissolved hydrophobic organic contaminants and metals in the REFERENCES 59

Danube river. Environ. Pollut. 184, 101–112. Waldmeier, P.C., Baumann, P.A., Wicki, P., Feldtrauer, J.-J., Stierlin, C., Schmutz, M., 1995. Similar potency of carbamazepine, oxcarbazepine, and lamotrigine in inhibiting the release of glutamate and other neurotransmitters. Neurology 45, 1907 LP – 1913. https://doi.org/10.1212/WNL.45.10.1907 Weissinger, R.H., Blackwell, B.R., Keteles, K., Battaglin, W.A., Bradley, P.M., 2018. Bioactive contaminants of emerging concern in National Park waters of the northern Colorado Plateau, USA. Sci. Total Environ. 636, 910–918. https://doi.org/10.1016/j.scitotenv.2018.04.332 Wilson, V.S., Bobseine, K., Lambright, C.R., Gray, L.E., 2002. A novel cell line, MDA-kb2, that stably expresses an androgen- and glucocorticoid-responsive reporter for the detection of hormone receptor agonists and antagonists. Toxicol. Sci. 66, 69–81. Wode, F., van Baar, P., Dünnbier, U., Hecht, F., Taute, T., Jekel, M., Reemtsma, T., 2015. Search for over 2000 current and legacy micropollutants on a wastewater infiltration site with a UPLC-high resolution MS target screening method. Water Res. 69, 274–283. https://doi.org/10.1016/J.WATRES.2014.11.034 Zedda, M., Zwiener, C., 2012. Is nontarget screening of emerging contaminants by LC-HRMS successful? A plea for compound libraries and computer tools. Anal. Bioanal. Chem. 403, 2493–2502. https://doi.org/10.1007/s00216-012-5893-y Zhao, J.-L., Ying, G.-G., Yang, B., Liu, S., Zhou, L.-J., Chen, Z.-F., Lai, H.-J., 2011. Screening of multiple hormonal activities in surface water and sediment from the Pearl River system, South China, using effect-directed in vitro bioassays. Environ. Toxicol. Chem. 30, 2208–15. https://doi.org/10.1002/etc.625 60

7 ANNEXES

List of annexes:

ANNEX I Tousova, Z., Oswald, P., Slobodnik, J., Blaha, L., Muz, M., Hu, M., Brack, W., Krauss, M., Di Paolo, C., Tarcai, Z., Seiler, T.-B., Hollert, H., Koprivica, S., Ahel, M., Schollée, J.E., Hollender, J., Suter, M.J.-F., Hidasi, A.O., Schirmer, K., Sonavane, M., Ait-Aissa, S., Creusot, N., Brion, F., Froment, J., Almeida, A.C., Thomas, K., Tollefsen, K.E., Tufi, S., Ouyang, X., Leonards, P., Lamoree, M., Torrens, V.O., Kolkman, A., Schriks, M., Spirhanzlova, P., Tindall, A., Schulze, T., 2017. European demonstration program on the effect-based and chemical identification and monitoring of organic pollutants in European surface waters. Sci. Total Environ. 601–602. https://doi.org/10.1016/j.scitotenv.2017.06.032

ANNEX II Tousova, Z., Froment, J., Oswald, P., Slobodník, J., Hilscherova, K., Thomas, K.V., Tollefsen, K.E., Reid, M., Langford, K., Blaha, L., 2018. Identification of algal growth inhibitors in treated waste water using effect-directed analysis based on non-target screening techniques. J. Hazard. Mater. 358. https://doi.org/10.1016/j.jhazmat.2018.05.031

ANNEX III Toušová, Z., Vrana, B., Smutná, M., Novák, J., Klučárová, V., Grabic, R., Slobodník, J., Giesy, J.P., Hilscherová, K., 2019. Analytical and bioanalytical assessments of organic micropollutants in the Bosna River using a combination of passive sampling, bioassays and multi-residue analysis. Sci. Total Environ. 650. https://doi.org/10.1016/j.scitotenv.2018.08.336

ANNEX IV Schulze, T., Ahel, M., Ahlheim, J., Aït-Aïssa, S., Brion, F., Di Paolo, C., Froment, J., Hidasi, A.O., Hollender, J., Hollert, H., Hu, M., Kloß, A., Koprivica, S., Krauss, M., Muz, M., Oswald, P., Petre, M., Schollée, J.E., Seiler, T.-B., Shao, Y., Slobodnik, J., Sonavane, M., Suter, M.J.-F., Tollefsen, K.E., Tousova, Z., Walz, K.-H., Brack, W., 2017. Assessment of a novel device for onsite integrative large-volume solid phase extraction of water samples to enable a comprehensive chemical and effect-based analysis. Sci. Total Environ. 581–582. https://doi.org/10.1016/j.scitotenv.2016.12.140

LIST OF ANNEXES 61

ANNEX V Ouyang, X., Leonards, P.E.G., Tousova, Z., Slobodnik, J., De Boer, J., Lamoree, M.H., 2016. Rapid Screening of Acetylcholinesterase Inhibitors by Effect- Directed Analysis Using LC × LC Fractionation, a High Throughput in Vitro Assay, and Parallel Identification by Time of Flight Mass Spectrometry. Anal. Chem. 88. https://doi.org/10.1021/acs.analchem.5b04311

ANNEX I Science of the Total Environment 601–602 (2017) 1849–1868

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

European demonstration program on the effect-based and chemical identiﬁcation and monitoring of organic pollutants in European surface waters

Zuzana Tousova a,b,PeterOswalda, Jaroslav Slobodnik a, Ludek Blaha b, Melis Muz c,d,MengHuc,d, Werner Brack c,d, Martin Krauss c, Carolina Di Paolo d, Zsolt Tarcai d, Thomas-Benjamin Seiler d, Henner Hollert d, Sanja Koprivica e,MarijanAhele,JenniferE.Scholléef,g, Juliane Hollender f,g, Marc J.-F. Suter f,g, Anita O. Hidasi f,h, Kristin Schirmer f,g,h, Manoj Sonavane i, Selim Ait-Aissa i, Nicolas Creusot i, Francois Brion i, Jean Froment c,j, Ana Catarina Almeida j, Kevin Thomas j,k, Knut Erik Tollefsen j,l,SaraTuﬁ m,XiyuOuyangm, Pim Leonards m,MarjaLamoreem, Victoria Osorio Torrens n, Annemieke Kolkman n, Merijn Schriks n,o, Petra Spirhanzlova p,AndrewTindallp, Tobias Schulze c,⁎ a Environmental Institute (EI), Okruzna 784/42, 972 41 Kos, Slovak Republic b Masaryk University, Faculty of Science, RECETOX, Kamenice 753/5, 625 00 Brno, Czech Republic c UFZ Helmholtz Centre for Environmental Research GmbH, Permoserstrasse 15, 04318 Leipzig, Germany d RWTH Aachen University, Institute for Environmental Research (Biology V), Department of Ecosystem Analysis, Worringerweg 1, 52074 Aachen, Germany e Rudjer Boskovic Institute, Bijenicka cesta 54, 10000 Zagreb, Croatia f Eawag, Überlandstrasse 133, 8600 Dübendorf, Switzerland g Institute of Biogeochemistry and Pollutant Dynamics, ETH Zürich, 8092 Zürich, Switzerland h EPF Lausanne, School of Architecture, Civil and Environmental Engineering, 1015 Lausanne, Switzerland i Institut National de l'Environnement Industriel et des Risques (INERIS), Unité ECOT, Parc ALATA - BP2, 60550 Verneuil-en-Halatte, France j Norwegian Institute for Water Research (NIVA), Ecotoxicology and Risk Assessment, Gaustadallèen 21, NO-0349 Oslo, Norway k Queensland Alliance for Environmental Health Sciences (QAEHS), The University of Queensland, 39 Keesels Road, Coopers Plains 4108, Australia l Norwegian University of Life Sciences (NMBU), Faculty of Environmental Science & Technology, Dept. for Environmental Sciences, Post Box 5003, N-1432 Ås, Norway m Vrije Universiteit Amsterdam, Department Environment & Health, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands n KWR, Watercycle Research Institute, Department of Chemical Water, Quality and Health, P.O. Box 1072, 3430 BB Nieuwegein, The Netherlands o Vitens drinking water company, P.O Box 1205, 8001 BE Zwolle, The Netherlands p WatchFrog S. A., 1 rue Pierre Fontaine, 91000 Evry, France

HIGHLIGHTS GRAPHICAL ABSTRACT

• Simpliﬁed protocol for effect-based monitoring of micropollutants was applied. • Large volume solid phase extraction device was applied to European rivers. • Target compounds did not explain major portions of observed biological effects • 21 micropollutants were prioritized based on risk assessment.

http://dx.doi.org/10.1016/j.scitotenv.2017.06.032 0048-9697/© 2017 Elsevier B.V. All rights reserved. 1850 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 article info abstract

Article history: Growing concern about the adverse environmental and human health effects of a wide range of micropollutants Received 12 April 2017 requires the development of novel tools and approaches to enable holistic monitoring of their occurrence, fate Received in revised form 4 June 2017 and effects in the aquatic environment. A European-wide demonstration program (EDP) for effect-based moni- Accepted 4 June 2017 toring of micropollutants in surface waters was carried out within the Marie Curie Initial Training Network EDA- Available online xxxx EMERGE. The main objectives of the EDP were to apply a simplified protocol for effect-directed analysis, to Editor: D. Barcelo link biological effects to target compounds and to estimate their risk to aquatic biota. Onsite large volume solid phase extraction of 50 L of surface water was performed at 18 sampling sites in four European river Keywords: basins. Extracts were subjected to effect-based analysis (toxicity to algae, fish embryo toxicity, neurotoxicity, Adverse effects (anti-)estrogenicity, (anti-)androgenicity, glucocorticoid activity and thyroid activity), to target analysis (151 or- Large volume solid phase extraction ganic micropollutants) and to nontarget screening. The most pronounced effects were estrogenicity, toxicity to EDA-EMERGE algae and fish embryo toxicity. In most bioassays, major portions of the observed effects could not be explained fi Simpli ed effect-directed analysis protocol by target compounds, especially in case of androgenicity, glucocorticoid activity and fish embryo toxicity. Estrone Environmental health and nonylphenoxyacetic acid were identified as the strongest contributors to estrogenicity, while herbicides, Human health with a minor contribution from other micropollutants, were linked to the observed toxicity to algae. Fipronil and nonylphenol were partially responsible for the fish embryo toxicity. Within the EDP, 21 target compounds were prioritized on the basis of their frequency and extent of exceedance of predicted no effect concentrations. The EDP priority list included 6 compounds, which are already addressed by European legislation, and 15 micropollutants that may be important for future monitoring of surface waters. The study presents a novel simplified protocol for effect-based monitoring and draws a comprehensive picture of the surface water status across Europe. © 2017 Elsevier B.V. All rights reserved.

1. Introduction sampling techniques like passive sampling or solid phase extraction have been developed (Dimpe and Nomngongo, 2016; Jones et al., Environmental quality monitoring of surface waters is fundamental 2015). Onsite large volume solid phase extraction (LVSPE), introduced to the sustainable management of water resources and to reducing risks by Schulze et al. (2017), enables collection and pre-concentration of posed by multiple anthropogenic stressors (Geissen et al., 2015). Cur- large samples for effect assessment and chemical analyses and allows rent EU legislation requires that member states regularly monitor a direct quantification of water concentrations of the identified com- wide range of biological and chemical parameters with the objective pounds. Chemical analyses of CECs are based on highly sensitive analyt- of achieving good chemical and ecological status of all European ical methods, which have become available with the advance of gas waterbodies (Directive 2000/60/EC). Forty-five priority substances chromatography (GC) and (ultra)-high performance liquid chromatog- have been established by the European Commission in the Water raphy [(U)HPLC] coupled either to tandem mass spectrometry (MS/ Framework Directive (WFD) for chemical monitoring and results of tar- MS), or a wide range of high resolution mass spectrometers (HR-MS), get analyses are compared against the environmental quality standards the latter allowing also for identification of unknowns in nontarget (EQSs), while biological monitoring evaluates flora, benthic inverte- screening workflows (Schymanski et al., 2015). brates and fish communities (Directive 2000/60/EC; Directive The biological effects of CECs have been assessed by means of in vitro 2008/1005/EC; Directive 2013/39/EC). However, adverse effects on and in vivo methods at various levels of biological organization, among aquatic biota cannot be easily associated with the analysis of priority which the use of cell-based methods, bacterial algal and fish models substances and the assessment of chemical and ecological status often such as zebrafish (Danio rerio), medaka (Oryzias latipes) or fathead min- results in contradictory outcomes (Altenburger et al., 2015). Therefore, now (Pimephales promelas) have been well established. For assessing ef- novel holistic monitoring approaches combining sensitive effect-based fects of CECs in surface waters and in addition to the traditional tools with sophisticated chemical analysis are needed to address a ecotoxicological endpoints like survival, growth and reproduction, spe- wide range of contaminants of emerging concerns (CEC) occurring in cific assays on neurotoxicity, immunotoxicity, oxidative stress or complex environmental mixtures. The growing list of CECs currently in- genotoxicity and hormonal activity such as (anti-)estrogenicity, (anti- cludes personal care products, human and veterinary pharmaceuticals, )androgenicity, glucocorticoid activity and thyroid activity are available surfactants and surfactant-derived compounds, X-ray contrast media, (Connon et al., 2012). These specific biological effects were repeatedly pesticides, disinfection by-products, algal toxins, flame retardants, plas- observed for surface waters, drinking waters, raw and treated wastewa- ticizers, UV-filters, industrial compounds and transformation products ters and technical systems (Dizer et al., 2002; Loos et al., 2013; Jálová et (Sima et al., 2014). Due to the broad range of physicochemical proper- al., 2013; Osman et al., 2015; Zhao et al., 2011) and some can be associ- ties, toxic modes of action and usage patterns, CECs constitute a major ated with particular compounds as the main toxicity drivers, e.g., challenge for analytical chemists, ecotoxicologists and environmental estrogenicity due to E2 and EE2 (Aerni et al., 2004; Koenig et al., 2017; regulators (Ginebreda et al., 2014). Miège et al., 2009) or phytotoxicity to herbicides (Tang and Escher, CECs in surface waters tend to occur at rather low (down to sub- 2014). However, the cause of many effects observed in bioassays re- ng L−1)andalsofluctuating concentrations, and hence integrative mains unexplained by the chemicals detected in the same samples

Abbreviations: AR, androgen receptor; Dex-EQ, dexamethasone equivalent; CEC, contaminants of emerging concern; DHT, dihydrotestosterone; DHT-EQ, dihydrotestosterone equivalent;

E1, estrone; E2, 17β-estradiol; E2-EQ, 17-β-estradiol equivalent; E3, estriol; EC50, concentration at which the effect reaches 50% of the effect in untreated control; EDA, effect directed analysis; EDP, European demonstration program; EE2, 17α-ethinylestradiol; EEQ-SSE, concentration of E2-EQ which is safe regarding major steroid estrogens E1, E2, E3 or EE2; EQS, environmental quality standard; ER, estrogen receptor; Flu-EQ, ﬂutamide equivalent; GC, gas chromatography; GR, glucocorticoid receptor; HPLC, high performance liquid chromatography; (HR)MS,(highresolution)massspectrometry;LOEC, lowest observed effect concentration; LOD, limit of detection; LOQ, limit of quantiﬁcation; LVSPE, large volume solid phase extraction;

MEC95, 95th percentile of maximum environmental concentration; MeOH, methanol; OH-Tam-EQ, hydroxytamoxifen equivalent; PAH, polycyclic aromatic hydrocarbons,; (P-)PNEC, (provisional) predicted no effect concentration; RB, river basin; REF, relative enrichment factor; SPE, solid phase extraction; T, testosterone; TOF, time of ﬂight; TU, toxic unit. ⁎ Corresponding author at: UFZ Helmholtz Centre for Environmental Research GmbH, Department of Effect-Directed Analysis, Permoserstrasse 15, 04318 Leipzig, Germany. E-mail address: [email protected] (T. Schulze). Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1851

(Burgess et al., 2013; Neale et al., 2015). Effect-directed analysis (EDA), anthropogenic pressures is given in Table 1. The main parameters of se- a method of combining biotesting, fractionation and chemical analysis lected RBs differed; however, in general all were affected by wastewater of bioactive fractions, is a powerful tool to facilitate identification of un- discharges and to a lesser extent by industrial and/or agricultural pollu- known toxicants (Brack, 2003; Brack et al., 2016). EDA studies are very tion. The impact of WWTP effluents were assessed for the Saale and costly, tedious and time consuming. However, novel sampling, Sava basin as well as for Swiss sites in SD (Table S12). The overall bioanalytical, chemical and software tools make EDA studies increasing- scheme of the EDP is given in Fig. 2. ly feasible through miniaturized and automated high-throughput formats, lowered detection limits, and optimized multi-target and non target screening (Guijarro et al., 2015; Brack et al., 2013). 2.2. Sampling In the presented study, a newly developed simplified EDA protocol was applied within a European demonstration program (EDP), which Single sampling was carried out in the different RBs between July included effect-based monitoring of CECs at 18 sampling sites in 4 Euro- 2013 and August 2014 using both grab sampling and LVSPE sampling. pean river basins, as a part of the EU-funded EDA-EMERGE project. The The 2 L-grab samples were kept at 4 °C and shipped immediately to simplified EDA protocol was based on existing EDA methods and includ- WatchFrog laboratory (Evry, France) for an in vivo thyroid activity ed a broad extraction procedure, simple clean-up and fractionation assay. The LVSPE sampling was performed according to a method de- techniques, followed by a set of sensitive, rapid and low volume bioas- scribed (Schulze et al., 2017). In brief, 50 L of river water were extracted says and extensive chemical screening analyses designed to add signif- with an onsite LVSPE device (UFZ, Leipzig, Germany; Maxx Mess- und icant value to classical chemical target monitoring. The specific goals of Probenahmetechnik GmbH, Rangendingen, Germany) containing this study were to i) combine the LVSPE sampling with a set of bioas- three different sorbents in sequence, designed to capture neutral says, chemical target analysis and nontarget screening in a novel simpli- (Chromabond® HR-X, Macherey-Nagel, Düren, Germany), weakly acid- fied EDA protocol; ii) identify target compounds responsible for ic (Chromabond® HR-XAW) and weakly basic (Chromabond® HR- biological effects and quantify their contribution to the observed effects; XCW) organic compounds. Each sorbent was eluted separately and iii) prioritize target compounds according to their estimated ecological the extracts were then divided into 13 aliquots according to the volume risk. This study is a result of the joint research activities of the EDA- requirements of each bioassay and chemical analysis. The aliquots were EMERGE fellows as a training event. dried and shipped to the respective laboratories for bioassays or chemical analyses (supplementary data (SD) - Table S3). Two blank samples 2. Materials and methods were collected with the LVSPE sampler. Extraction cartridges filled with clean sorbents were conditioned and extracted accordingly to obtain 2.1. River basins and sampling sites fabrication blanks. Laboratory blanks were prepared by percolation of 2 L of mineralized LC-grade distilled water through the LVSPE device Four case studies in four river basins (RBs) in six European countries for 100 cycles, which is equivalent to sampling of 50 L. Both blanks (i.e., Germany, Czech Republic, Slovakia, Hungary, Croatia and Switzer- were otherwise treated identically to the field samples. The detailed de- land) were selected for the EDP (Fig. 1) based on results of earlier mon- scription of the LVSPE sampling method and individual samples is pro- itoring campaigns. Overview of the sampling sites and major vided in SD (Table S1 and S2).

Fig. 1. Overview map of the EDP sampling locations (from top right to bottom left) in the Saale RB (Germany), Danube RB (Czech Rep., Slovakia and Hungary), Sava RB (Croatia) and Emme RB (Switzerland). 1852 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868

Table 1 Description of the EDP sampling sites. For more details see supplementary data (Table S1).

Code Site Coordinates Site characterization and major pressures

Danube RB 1-1 WWTP effluent - Svratka, 49.12447N 16.62697E Site at the effluent of large scale municipal WWTP (400,000 citizens); estrogenicity, Brno Modrice, Czech Republic dioxin-like activity, cytotoxicity, pharmaceuticals, polar pesticides, bisphenol A, PFOA and steroid hormones were previously observed (Vrana et al., 2011, Jedličková et al., 2011) 1-2 Danube - Morava Tributary, 48.17864N 16.97597E Site 300 m upstream of Morava confluence to Danube; site affected by automobile Devinska Nová Ves, Slovakia industy and agriculture; polluted sediments and SPM - DDT,DDE, DDD, PAHs, petroleum hydrocarbons and DEHP were previously observed (“Danube River Basin Water Quality Database, https://www.icpdr.org/wq-db” 2012) 1-3 Danube - Bratislava, Slovakia 48.11834N 17.14435E Site in the city centre of Bratislava, site affected by urban pressures and navigation; nonylphenols, triazine pesticides, 2,4-D, pharmaceuticals, polluted sediments and SPM with PAHs, nonylphenols, DDT, dioxins, PBDEs, PCBs and DEHP were previously observed (“Danube River Basin Water Quality Database,” 2012) 1-4 Danube - Szob, Hungary 47.59890N 19.08361E Site downstream of towns - Estergom and Sturovo and tributary of Hron; river bank filtration facility for drinking water production in the vicinity; PFCs, triazine pesticides, DEHP and pharmaceuticals were observed (“Danube River Basin Water Quality Database,” 2012); high dilution by the Danube River expected, reference to the upstream sites Sava RB 2-1 Sava - Otok Samoborski, 45.843083N 15.729167E 10 km upstream of the city of Zagreb, upstream reference location to the downstream Samobor, Croatia sites affected by various pollution sources in the area of the city of Zagreb; a checkpoint for transboundary pollution from Slovenia (Smital et al., 2013) 2-2 Sava - Podsused, Zagreb, Croatia 45.793583N 15.852783E Situated in the western part of the city of Zagreb; about 5 km downstream from the discharge point of the wastewaters from the WWTP of the city of Zapresic, including significant contribution of effluents from pharmaceutical industry; marked input of macrolide antimicrobials (Terzic and Ahel, 2011) 2-3 Sava - Oborovo, 45.686450N 16.246900E Several km downstream of the main wastewater outlets of WWTPs of the cities of Velika Gorica, Croatia Zagreb (800,000 inh.) and Velika Gorica (60,000 inh.); the wastewaters of both cities are of the mixed type, including significant contribution of industrial WWs; both WWs receive full mechanical and biological treatment before the discharge 2-4 Sava - Crnac, Sisak, Croatia 45.445267N 16.419267E 2 km downstream from the city of Sisak (50,000 inhabitants), affected by urban WW, iron works and oil refinery; notable petroleum hydrocarbon pollution in water and sediments (Smital et al., 2013) Emme RB 3-1 Limpach – Messen, Switzerland 47.10458N 7.44811E Upstream of WWTP Messen effluent to the Limpach Creek, site unaffected by WW 3-2 Limpach - Messen, Switzerland 47.10989N 7.46659E Downstream of WWTP Messen effluent to the Limpach Creek, no other pressures 3-3 Urtenen - Kernenried, Switzerland 47.05544N 7.53750E Upstream of WWTP effluent Kernenried to the Urtenen Creek, site unaffected by WW 3-4 Urtenen - Kernenried, Switzerland 47.05864N 7.54021E Downstream of WWTP effluent Kernenried to the Urtenen Creek, no other pressures 3-5 Emme - Wiler bei Utzenstorf, 47.16023N 7.54703E Confluence of Limpach Creek to Emme, downstream of both Messen and Kernenried WWTPs Switzerland Saale RB 4-1 Holtemme - Steinerne Renne, Germany 51.81794 N 10.72889E Pristine water in Harz mountains; reference site to 4-2, 4-5 4-2 Holtemme - Derenburg, Germany 51.86645N 10.879617E Site downstream of WWTP in Silstedt, impact of hospitals and wellness clinics 4-3 Saale - Rudolstadt, Germany 50.71865N 11.3984E Site with industrial impact from paper and chemical industry 4-4 Saale - Klein Rosenburg, Germany 51.931944N 11.888889E Integrative site on the Saale before entering into the Elbe river in Groß Rosenburg, sampling station of the Watch List Study 2012 4-5 Holtemme - Nienhagen, Germany 51.94153N 11.15856E Integrative site of Holtemme near the confluence with Bode; downstream of WWTP Halberstadt; urban and agricultural impact

Abbreviations: DDT - 1,1,1-trichlor-2,2-bis(4-chlorfenyl)ethan; DDD - dichlorodiphenyldichloroethane; DDE - dichlorodiphenyldichloroethylene; DEHP - Bis(2-ethylhexyl) phthalate; PAHs - polycyclic aromatic hydrocarbons; PBDEs - polybrominated diphenyl ethers or PBDEs- polybrominated diphenyl ethers; PCBs - polychlorinate bifenyls; PFCs - perﬂuorocarbons; SPM - suspended particulate matter; WW - waste water; WWTP - water water treatment plant.

2.3. Effect assessment available RegTox Microsoft Excel™ Macro (http://www.normalesup. org/~vindimian/fr_index.html). A basic description of the assays is pro- A set of seven bioassays performed in six different laboratories was vided below, while more details on the bioassays methods and test con- applied to screen for both nonspecificandspecifictoxicityofthe ditions are provided in the SD (Table S4). water samples, including both in vitro and in vivo methods (SD - Table S3). The effective concentrations of surface water samples were 2.3.1. ER-mediated activity (MELN cells) assay expressed in relative enrichment factors (REFs) as proposed by Escher The estrogen receptor mediated bioassay based on human et al. (2006). For active samples in the receptor-mediated assays, effect breast cancer cell line (MELN cells transfected with a promoter contain- equivalents of standard agonists or antagonists were calculated. Prior to ing estrogen responsive elements driving expression of luciferase testing, the extracts of all sample fractions were reconstituted in a sol- (Balaguer et al., 1999)) was used for assessment of estrogenicity and vent and mixed with the test media to reach safe solvent concentration antiestrogenicity of the sample. The receptor mediated activity of a sam- related to the bioassay used and final REFs from 1 to 100, meaning that ple was assessed in 96-well plates by measurement of luminescence the tested range covered the original river concentrations (REF = 1) as after 16 h of exposure at 37 °C. Effective concentrations and equivalents well as concentrations up to 100-times higher (REF = 100). In excep- of estrogenic potency (using the reference estrogen estradiol, E2) and tional cases, REF 500 were used for the upper limit (e.g., in the AChE in- anti-estrogenic potency (with OH-tamoxifen as the reference) were de- hibition assay, where no effects were observed up to REF 100). One of termined as previously detailed (Creusot et al., 2016). the bioassays (in vivo thyroid activity assay with Xenopus)wasper- formed with whole raw water samples. Laboratory blank, fabrication 2.3.2. AR-mediated activity (MDA-kb2 cells) assay blank, solvent control and positive control with standard reference The androgen receptor mediated bioassay based on human breast compounds were tested in parallel with the samples in each bioassay. cancer cell line (MDA-kb2 cells transfected with a promoter containing The results of bioassays were evaluated using nonlinear regression androgen responsive elements driving expression of luciferase models to derive the effective concentrations (EC50 or EC20)in (Wilson et al., 2002)) was used for assessment of androgenicity and GraphPad Prism 6 (GraphPad Software, Inc., La Jolla, USA) or freely antiandrogenicity of the sample. The receptor mediated activity of the Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1853

Fig. 2. Scheme of the European demonstration programme. For more details see supplementary data (Table S1–S5).

sample was assessed in 96-well plates by measurement of lumines- Potencies were expressed as equivalents (EQ) of standard reference com- cence after 16 h of exposure at 37 °C. Potencies were expressed as pound dexamethasone Dex-EQ as previously described in Macikova et al. equivalents (EQ) of standard reference compounds, i.e., DHT-EQ for (2014). androgenicity and flutamide-EQ for anti-androgenicity as described previously (Creusot et al., 2014). The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide 2.3.4. Zebrafish embryo acute toxicity assay (FET) (MTT) test, based on the color change of tetrazolium dye induced by ox- Fish embryo acute toxicity was assessed in zebrafish (Danio rerio) idoreductase enzymes of viable cells as previously described by Creusot according to the OECD guideline 236 (OECD, 2013). Freshly fertilized et al. (2015) was used to assess cytotoxicity of the samples in the cell zebrafish eggs were exposed to six sample concentrations in 24-well based bioassays (MELN and MDA-kb2 cells). plates for 96 h, at 26 °C and 14 h light : 10 h dark cycle. Four apical endpoints were observed daily as indicators of lethality: embryo coagula- tion, lack of somite formation, nondetachment of the tail and absence 2.3.3. GR-CALUX® assay of heartbeat. Additionally, the occurrence of sublethal morphological ef- GR-CALUX is a patented glucocorticoid receptor mediated bioassay fects was also recorded. Effects are described either as the occurrence of based on human osteoblastic cells (U2-OS) transfected with a promoter lethality, or as the cumulative occurrence of any endpoint for lethality or containing glucocorticoid responsive elements driving expression of lucif- sublethal toxicity (i.e., any lethal or sublethal effect). In each test, ten erase. Glucocorticoid activity of the samples was assessed in 96-well embryos were exposed per exposure concentration. 3,4-Dichloraniline plates by measurement of luminescence after 24 h of exposure at 37 °C. was used as the standard reference compound. 1854 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868

Table 2 Summary results of target analysis. For more details see supplementary data (Table S5 and S7).

Compound name CAS Compound LOQ PNEC PNEC MEC95 Median Frequency of PNEC Frequency of group/usage pattern [ng/L] [ng/L] type [ng/L] [ng/L] occurrence [%] exceedance PNEC exceedance N=18 [%] N = 18

17-alpha-Ethinylestradiol 57-63-6 Pharmaceutical, 0.40 0.007 PNEC ND ND 0 no NA estrogen, WFD watch list 17-beta-Estradiol 50-28-2 Hormone, estrogen, 0.30 0.08 PNEC ND ND 0 no NA WFD watch list 1H-Benzotriazole 95-14-7 Corrosion inhibitor, 1.0 4840 P-PNEC 1026 135 94 no NA industrial compound 2,4-Dichlorophenoxyacetic acid 94-75-7 Herbicide 1.0 200 PNEC 4.2 bLOQ 44 no NA 2,4-Dinitrophenol 51-28-5 Herbicide, 1.0 1 PNEC 7.7 1.9 72 7.70 56 transformation product 2,6-Dichlorobenzamide 2008-58-4 Herbicide, 1.0 NA NA 20.1 bLOQ 50 NA NA transformation product 4-Androstene-3,17-dione 63-05-8 Hormone, androgen 1.50 NA NA 2.99 bLOQ 11 NA NA 4-Hydroxytamoxifen 68392-35-8 Pharmaceutical, 0.60 NA NA ND ND 0 NA NA anti-estrogen 4-Nitrophenol 100-02-7 Industrial compound 5.0 NA NA bLOQ bLOQ 22 NA NA 4-Toluenesulfonamide 70-55-3 Industrial compound 10 NA NA 308 56 94 NA NA 5-Methyl-1H-benzotriazole 136-85-6 Corrosion inhibitor 1.0 5285 P-PNEC 955 53.7 94 no NA 6-alpha-Methylprednisolone 83-43-2 Pharmaceutical, 0.63 NA NA 1.44 bLOQ 17 NA NA glucocorticoid Acesulfame 55589-62-3 Artificial sweetener, 1.0 1731 P-PNEC 19.8 2.3 72 no NA Marker compound Acetyl-Sulfamethoxazole 21312-10-7 Pharmaceutical, 1.0 178177 PNEC 5.9 1.7 61 no NA transformation product Aclonifen 74070-46-5 Herbicide, WFD 0.10 12.0 PNEC ND ND 0 no NA priority Alachlor 15972-60-8 Herbicide, WFD 1.0 300 PNEC ND ND 0 no NA priority Aldosterone 52-39-1 Hormone, 14 NA NA ND ND 0 NA NA glucocorticoid Amcinonide 51022-69-6 Pharmaceutical, 0.88 NA NA ND ND 0 NA NA glucocorticoid Anastrozole 120511-73-1 Pharmaceutical, 0.15 NA NA 0.24 bLOQ 6 NA NA anti-estrogen Androsterone 53-41-8 Hormone, androgen 2.0 NA NA ND ND 0 NA NA Anthracene 120-12-7 PAH, WFD priority 1.0 100 PNEC 5.9 bLOQ 67 no NA Atrazine 1912-24-9 Herbicide, WFD 1.0 600 PNEC 18.0 5.0 94 no NA priority Atrazine-desethyl 6190-85-4 Herbicide, 1.0 300 PNEC 23.3 9.0 72 no NA transformation product Azithromycin 83905-01-5 Pharmaceutical, 1.0 90 PNEC 1022 bLOQ 17 11.36 11 antibiotic Azoxystrobin 131860-33-8 Fungicide 1.0 56 PNEC 5.0 b LOQ 67 no NA Bentazone 25057-89-0 Herbicide 1.0 60 PNEC 52.2 4.5 89 no 6 Benzo(a)pyrene 50-32-8 PAH, WFD priority 1.0 0.17 PNEC ND ND 0 no NA Benzo(b)fluoranthene 205-99-2 PAH, WFD priority 1.0 0.17 PNEC ND ND 0 no NA Benzo(g,h,i)perylene 191-24-2 PAH, WFD priority 1.0 0.17 PNEC ND ND 0 no NA Benzo(k)fluoranthene 207-08-9 PAH, WFD priority 1.0 0.17 PNEC ND ND 0 no NA Benzophenone-3 131-57-7 Sunscreen agent 1.0 345 P-PNEC 26.8 bLOQ 56 no NA Benzophenone-4 4065-45-6 UV filter 1.0 NA NA 205 40.5 89 NA NA Betamethasone 378-44-9 Pharmaceutical, 0.52 658425 P-PNEC bLOQ bLOQ 6 no NA glucocorticoid Bezafibrate 41859-67-0 pharmaceutical, 1.0 460 PNEC 5.1 bLOQ 28 no NA lipid regulator Bifenox 42576-02-3 Herbicide, WFD 20 1.2 PNEC ND ND 0 no NA priority Bisphenol A 80-05-7 Plasticizer, 4.0 100 PNEC 43.3 5.3 89 no NA xenoestrogen Caffeine 58-08-2 Stimulant in 1.0 100 PNEC 237 12.8 72 2.37 17 beverages, marker compound Canrenone 976-71-6 Pharmaceutical, 0.70 NA NA 2.64 bLOQ 11 NA NA diuretic, anti-androgen Carbamazepine 298-46-4 Pharmaceutical, 1.0 500 PNEC 393 25.2 94 no 6 antiepileptic, marker compound Carbendazim 10605-21-7 Fungicide 1.0 150 PNEC 2.8 bLOQ 44 no NA Ciprofloxacin 85721-33-1 Pharmaceutical, 10 89 PNEC bLOQ bLOQ 6 no NA Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1855

Table 2 (continued)

Compound name CAS Compound LOQ PNEC PNEC MEC95 Median Frequency of PNEC Frequency of group/usage pattern [ng/L] [ng/L] type [ng/L] [ng/L] occurrence [%] exceedance PNEC exceedance N=18 [%] N = 18

antibiotic Chlorfenvinphos 470-90-6 Insecticide, WFD 1.0 100 PNEC ND ND 0 no NA priority Chlorotoluron 15545-48-9 Herbicide 1.0 100 PNEC 2.4 bLOQ 56 no NA Chlorpyrifos 2921-88-2 Insecticide, WFD 1.0 30 PNEC 1.7 bLOQ 11 no NA priority Clarithromycin 81103-11-9 Pharmaceutical, 1.0 60 PNEC 57.6 3.0 94 no 6 antibiotic Clobetasol propionate 25122-46-7 Pharmaceutical, 0.93 NA NA ND ND 0 NA NA glucocorticoid Clobetasone butyrate 25122-57-0 Pharmaceutical, 0.39 NA NA ND ND 0 NA NA glucocorticoid Clopidogrel 113665-84-2 Pharmaceutical, 0.20 NA NA ND ND 0 NA NA anti-coagulant Clozapine 5786-21-0 Pharmaceutical, 0.25 NA NA 11.98 0.35 61 NA NA anti-psychotic Cortisone 53-06-5 Hormone, 1.0 NA NA bLOQ bLOQ 28 NA NA glucocorticoid Cyproterone 2098-66-0 Pharmaceutical, 0.70 NA NA ND ND 0 NA NA Anti-androgen Desoximethasone 382-67-2 Pharmaceutical, 0.42 NA NA ND ND 0 NA NA glucocorticoid Dexamethasone 50-02-2 Pharmaceutical, 0.54 663141 P-PNEC bLOQ bLOQ 6 no NA glucocorticoid Diazinon 333-41-5 Insecticide 1.0 1.0 PNEC 10.8 bLOQ 50 10.75 50 Diclofenac 15307-86-5 Pharmaceutical, 1.0 10 PNEC 443 2.9 83 44.34 39 analgesic, WFD watch list Dicofol 115-32-2 Insecticide, WFD 10 0.03 PNEC ND ND 0 no NA priority Didecyldimethylammonium 2390-68-3 Surfactant 1.0 NA NA bLOQ bLOQ 6 NA NA Diethyltoluamid 134-62-3 Repellent 1.0 360 P-PNEC 191 16.7 94 no 6 Diﬂuprednate 23674-86-4 Pharmaceutical, 0.73 NA NA bLOQ bLOQ 6 NA NA glucocorticoid Diglyme 111-96-6 Industrial solvent 1.0 NA NA 5.3 bLOQ 22 NA NA Dihydrotestosterone 521-18-6 Hormone, androgen 1.5 NA NA ND ND 0 NA NA Dichlorvos 62-73-7 Insecticide, WFD 1.0 0.6 PNEC 3.8 bLOQ 11 6.40 11 priority Dimethenamid-p 87674-68-8 Herbicide 1.0 200 PNEC 1.9 bLOQ 22 no NA Dimethoate 60-51-5 insecticide 1.0 70 PNEC bLOQ bLOQ 6 no NA Diuron 330-54-1 Herbicide, WFD 1.0 200 PNEC 29.6 1.8 72 no NA priority Drospirenone 67392-87-4 Pharmaceutical, 1.0 NA NA ND ND 0 NA NA progesterone Endosulfan aplha and beta 115-29-7 Insecticide, WFD 1.0 5 PNEC bLOQ bLOQ 6 no NA priority Epi-Androsterone 481-29-8 Hormone, androgen 2.0 NA NA 4.6 bLOQ 33 NA NA Erythromycin 114-07-8 Pharmaceutical, 10 40 PNEC 54 bLOQ 39 1.34 11 antibiotic Estriol 50-27-1 Hormone, estrogen 1.0 232530 P-PNEC ND ND 0 no NA Estrone 53-16-7 Hormone, estrogen 0.10 3.6 PNEC 1.04 0.27 78 no NA Fipronil 120068-37-3 Insecticide 0.10 12.0 PNEC 1.51 0.22 72 no NA Flunisolide 3385-03-3 Pharmaceutical, 0.65 NA NA ND ND 0 NA NA glucocorticoid Fluoranthene 206-44-0 PAH, WFD priority 1.0 6.3 PNEC 13.8 1.6 61 2.19 17 Fluorometholone 426-13-1 Pharmaceutical, 0.41 NA NA ND ND 0 NA NA glucocorticoid Gestoden 60282-87-3 Pharmaceutical, 1.0 NA NA ND ND 0 NA NA progesterone Heptachlor 76-44-8 Insecticide, WFD 1.0 0.00001 PNEC ND ND 0 no NA priority Heptachlor epoxide 1024-57-3 Insecticide, 1.0 0.00001 PNEC ND ND 0 no NA transformation product Hexachlorobenzene 118-74-1 Fungicide, WFD 1.0 10 PNEC bLOQ bLOQ 11 no NA priority Hexachlorocyclohexane 608-73-1 Insecticide, WFD 1.0 NA NA bLOQ bLOQ 6 NA NA priority Hexamethoxymethylmelamine 3089-11-0 Industrial 5.0 1350 PNEC 152 27.8 78 no NA compound, marker compound Hydrocortisone 50-23-7 Hormone, 0.86 NA NA 2.47 bLOQ 11 NA NA glucocorticoid

(continued on next page) 1856 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868

Table 2 (continued)

Compound name CAS Compound LOQ PNEC PNEC MEC95 Median Frequency of PNEC Frequency of group/usage pattern [ng/L] [ng/L] type [ng/L] [ng/L] occurrence [%] exceedance PNEC exceedance N=18 [%] N = 18

Ibuprofen 15687-27-1 Pharmaceutical, 10 10 PNEC 20 bLOQ 28 2.01 22 analgesic Indeno(1,2,3-cd)pyrene 193-39-5 PAH, WFD priority 1.0 0.17 PNEC ND ND 0 no NA Irgarol (cybutryn) 28159-98-0 Biocide, anti-fouling 1.0 2.5 PNEC bLOQ bLOQ 11 no NA agent Isoproturon 34123-59-6 Herbicide, WFD 1.0 300 PNEC 21.8 bLOQ 89 no NA priority Ketoprofen 22071-15-4 Pharmaceutical, 1.0 3452 P-PNEC 5.6 bLOQ 28 no NA analgesic Lauryl diethanolamide 120-40-1 Surfactant 9.4 NA NA bLOQ bLOQ 6 NA NA Levo-norgestrel 797-63-7 Pharmaceutical, 1.0 NA NA ND ND 0 NA NA progesterone Mecoprop 93-65-2 Herbicide 1.0 18000 PNEC 13.9 1.5 94 no NA Medroxyprogesterone 520-85-4 Pharmaceutical, 0.40 NA NA bLOQ bLOQ 6 NA NA progesterone Metoprolol 37350-58-6 Pharmaceutical, beta 1.0 64000 PNEC 507 4.4 83 no NA blocker N,N-Dimethyldodecylamine-N-oxide 1643-20-5 Surfactant 1.0 NA NA 10.3 bLOQ 11 NA NA Naphthalene 91-20-3 PAH, WFD priority 1.0 2000 PNEC 2.8 bLOQ 50 no NA Naproxen 22204-53-1 Pharmaceutical, 1.0 1700 PNEC 38.6 1.4 50 no NA analgesic Nonylphenol 25154-52-3 Surfactant, WFD 10 21 PNEC 148 17 78 7.17 39 priority, xenoestrogen Nonylphenoxyacetic acid 3115-49-9 Surfactant, 1.0 176 P-PNEC 276 12.0 94 1.57 6 xenoestrogen Nonylphenoxydiethoxyacetic acid 106807–78–7 Surfactant, 1.0 NA NA 69.6 2.3 72 NA NA xenoestrogen Norethindrone 68-22-4 Pharmaceutical, 1.0 NA NA ND ND 0 NA NA progesterone Norfloxacin 70458-96-7 Pharmaceutical, 10 16600 PNEC ND ND 0 no NA antibiotic Norgestimate 35189-28-7 Pharmaceutical, 0.30 NA NA ND ND 0 NA NA progesterone Octylphenol 140-66-9 Surfactant, 3.0 100 PNEC 16.1 bLOQ 22 no NA xenoestrogen Octylphenoxyacetic acid 15234-85-2 Surfactant, 1.0 319 P-PNEC ND ND 0 no NA xenoestrogen Pentabromodiphenylether 32534-81-9 Flame retadant, WFD 1.0 NA NA ND ND 0 NA NA (congener numbers 28, 47, 99, priority 100, 153 and 154) Pentachlorobenzene 608-93-5 Industrial 1.0 7 PNEC 1.8 bLOQ 28 no NA compound, WFD priority Pentachlorophenol 87-86-5 Fungicide, WFD 100 400 PNEC ND ND 0 no NA priority Perfluorooctanesulfonic acid 1763-23-1 Surfactant, WFD 10 0.13 PNEC 31 bLOQ 67 239.23 67 priority Perfluorooctanoic acid 335-67-1 Surfactant 1.0 10000 PNEC 71.5 17.5 83 no NA Phenazone 60-80-0 Pharmaceutical, 1.0 2969 P-PNEC 20.8 bLOQ 67 no NA analgesic Phenylbenzimidazolesulfonic acid 27503-81-7 Sunscreen agent 1.0 NA NA 1026 58.9 94 NA NA Pirimicarb 23103-98-2 Insecticide 1.0 NA NA bLOQ bLOQ 17 NA NA Prednicarbate 73771-04-7 Pharmaceutical, 0.41 NA NA ND ND 0 NA NA glucocorticoid Prednisolone 50-24-8 Pharmaceutical, 0.88 2450000 PNEC ND ND 0 no NA glucocorticoid Prednisone 53-03-2 pharmaceutical, 0.76 NA NA bLOQ bLOQ 6 NA NA glucocorticoid Progesterone 57-83-0 Hormone, 1.0 NA NA bLOQ bLOQ 6 NA NA progesterone Propiconazole 60207-90-1 Fungicide 1.0 230 PNEC 4.3 bLOQ 72 no NA Prothioconazole-desthio 120983-64-4 Fungicide, 1.0 NA NA 1.2 bLOQ 39 NA NA transformation product Quinoxyfen 124495-18-7 Fungicide, WFD 10 15 PNEC ND ND 0 no NA priority Raloxifene 84449-90-1 Pharmaceutical, 0.40 NA NA bLOQ bLOQ 6 NA NA anti-/estrogen Rimexolone 49697-38-3 Pharmaceutical, 0.39 NA NA ND ND 0 NA NA glucocorticoid Roxithromycin 80214-83-1 Pharmaceutical, 1.0 1000 PNEC 31.9 bLOQ 22 no NA antibiotic Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1857

Table 2 (continued)

Compound name CAS Compound LOQ PNEC PNEC MEC95 Median Frequency of PNEC Frequency of group/usage pattern [ng/L] [ng/L] type [ng/L] [ng/L] occurrence [%] exceedance PNEC exceedance N=18 [%] N = 18

Simazine 122-34-9 Herbicide, WFD 1.0 1000 PNEC 6.2 1.7 72 no NA priority Sotalol 3930-20-9 Pharmaceutical, beta 1.0 1958 P-PNEC 82.4 2.9 78 no NA blocker Spiroxamine 118134-30-8 Fungicide 1.0 NA NA bLOQ bLOQ 11 NA NA Sucralose 56038-13-2 Artificial sweetener, 10 346561 P-PNEC 1579 100 89 no NA marker compound Sulfadimidin/sulfamethazine 57-68-1 Pharmaceutical, 1.0 1735 P-PNEC 4.6 bLOQ 39 no NA sulfonamide antimicrobial Sulfamethoxazole 723-46-6 Pharmaceutical, 1.0 600 PNEC 92.2 8.0 83 no NA sulfonamide antimicrobial Sulfapyridine 144-83-2 Pharmaceutical, 1.0 10500 P-PNEC 88.5 1.8 61 no NA sulfonamide antimicrobial Sulfathiazole 72-14-0 Pharmaceutical, 1.0 NA NA 1.5 bLOQ 6 NA NA sulfonamide antimicrobial Tamoxifen 10540-29-1 Pharmaceutical, 0.70 NA NA ND ND 0 NA NA anti-estrogen Tebuconazole 107534-96-3 Fungicide 1.0 100 PNEC 7.4 bLOQ 56 no NA Terbuthylazine 5915-41-3 Herbicide 1.0 190 PNEC 17.7 4.7 83 no NA Terbuthylazine-desethyl 30125-63-4 Herbicide, 1.0 2.4 PNEC 8.1 2.8 94 3.38 61 transformation product Terbutryn 886-50-0 Herbicide, WFD 1.0 6.5 PNEC 4.9 1.5 94 no 6 priority Testosterone 58-22-0 Hormone, androgen 0.40 NA NA ND ND 0 NA NA Tetraglyme 143-24-8 Industrial solvent 0.10 NA NA 66.20 3.90 67 NA NA Thiacloprid 111988-49-9 Insecticide 1.0 NA NA bLOQ bLOQ 17 NA NA Trenbolone 10161-33-8 Pharmaceutical, 2.0 NA NA bLOQ bLOQ 6 NA NA growth promoter Triamcihexacetonide 5611-51-8 Pharmaceutical, 18 NA NA ND ND 0 NA NA glucocorticoid Triamcinolone 83474-03-7 Pharmaceutical, 1.2 NA NA bLOQ bLOQ 6 NA NA glucocorticoid Triamcinolone acetonide 8054-16-8 Pharmaceutical, 0.59 NA NA 1.73 bLOQ 33 NA NA glucocorticoid Triclosan 3380-34-5 Biocide 1.0 0.7 PNEC 2.4 bLOQ 39 3.43 28 Triethylcitrate 77-93-0 Plasticizer 1.0 240700 P-PNEC 126 bLOQ 39 no NA Trifluralin 1582-09-8 Herbicide, WFD 1.0 30 PNEC ND ND 0 no NA priority Triglyme 112-49-2 Industrial solvent 0.10 88221 P-PNEC 4.6 0.2 61 no NA Trimethoprim 738-70-5 Pharmaceutical, 1.0 60000 PNEC 9.4 1.2 83 no NA antibiotic Trimethyloctylammonium 2083-68-3 Surfactant 1.0 NA NA bLOQ bLOQ 6 NA NA Triphenylphosphate 115-86-6 Plasticizer, flame 5.0 30 PNEC 51.7 bLOQ 11 1.72 6 retadant Triphenylphosphine oxide 791-28-6 Industrial 5.0 17092 P-PNEC 294 bLOQ 72 no NA compound, marker compound Tris(2-butoxyethyl)phosphate 78-51-3 Flame retardant 5.0 6800 PNEC 140 bLOQ 50 no NA Verapamil 152-11-4 Pharmaceutical, 1.0 30 P-PNEC 14.5 bLOQ 56 no 6 antihypertensive

MEC95 - maximum environmental concentration 95 or 95th percentile of measured concentration; ND - non-detected; NA - not available; LOQ - limit of quantiﬁcation; PNEC - predicted no effect concentration based on experimental value or existing environmental quality standard (EQS); P-PNEC - predicted-PNEC based on in silico prediction.

2.3.5. Algal growth inhibition assay coding sequence under the control of thyroid hormone responsive re- Growth of a population of unicellular green alga (Raphidocelis gion of the TH/bZIP-eGFP promoter) were exposed to the sample for subcapitata) was assessed after 72 h of exposure to the sample by mea- 72 h to assess thyroid axis activity by fluorescence microscopy (Fini et suring absorbance at 680 nm. The assay was performed in transparent al., 2007). The assay was performed with unfiltered whole water 96-well plates, at 24 °C and under permanent illumination (1600 lx). samples. Potassium dichromate was used as the standard reference compound. The method was based on OECD guideline 201 (OECD, 2011)andmod- ified according to Rojickova and Dvorakova (1998). 2.3.7. Acetylcholine esterase (AChE) inhibition assay Changes in the activity of the AChE were determined after 1 h of exposure at room temperature using the color change of 2.3.6. In vivo thyroid activity assay dithiobisnitrobenzoate to 5-thio-2-nitrobenzoic acid, which is a direct Tadpoles of a stable line of transgenic Xenopus (Xenopus laevis), har- measure of hydrolysis catalyzed by AChE (Ellman et al., 1961; Galgani boring the TH/bZIP-eGFP genetic construct (green fluorescent protein and Bocquene, 1991). The assay was performed in 96-well plates, 1858 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868

Table 3 Frequency of effects in bioassays (expressed as percentage) for all sampling sites and different LVSPE sorbents. For more details see supplementary data (Table S6).

Bioassay Endpoint Sampling sites Neutral sorbent (HR-X) Weak anionic exchanger (HR-XAW) Weak cationic exchanger (HR-XCW) N=18 N=18 N=18 N=17

ER-mediated activity Estrogenicity 83 83 28 12 MELN cells Anti-estrogenicity 46 46 0 (N = 13) 0 (N = 12) AR-mediated activity Androgenicity 22 22 0 0 MDA-kb2 cells Anti-androgenicity 17 17 0 0 GR CALUX Glucocorticoid activity 28 28 0 (N = 13) 0 (N = 12) Zebraﬁsh embryo acute toxicity Survival 83 83 28 13 (N = 16) Sublethal endpoints 73 73 28 13 (N = 16) Algal growth inhibition Growth rate inhibition 83 78 6 18 In vivo thyroid activity Thyroid activity 6 –– – AChE inhibition AChE inhibition 61 17 22 41

using an automated liquid handling system. Dichlorvos was used as a detected compounds in a sample was used to quantify the overall con- standard reference compound. tribution of target compounds to the observed toxicity.

ci REFðÞ EC50 2.4. Chemical analyses TU ¼ ð1Þ minECOSAR 96h−EC50i Neutral extracts were subject to target analysis of 151 compounds and to GC-MS nontarget screening, performed in 5 different laborato- 2.6. Risk assessment and prioritization ries. Each laboratory required an amount of extract which was equivalent of 1 L of river water. The target analysis was carried out using LC- Basic risk assessment and prioritization of detected target com- (HR)MS/MS and GC-MS tools, respective methods for each laboratory pounds was carried out according to a methodology developed within are detailed in SD (Table S3). The list of target compounds was set up the NORMAN Network (Dulio and Von der Ohe, 2013; von der Ohe et to cover several classes of environmentally and toxicologically relevant al., 2011). In brief, risk assessment was based on the NORMAN predicted CECs and their transformation products including 55 pharmaceuticals, no effect concentrations (PNECs), which are ecotoxicological threshold 43 pesticides, 11 surfactant-derived compounds and several com- values determined on the basis of experimental data, existing environ- pounds from other classes like industrial compounds, corrosion inhibi- mental quality standards (EQSs), or in silico predictions in order to pro- tors, sunscreen agents, sweeteners and flame retardants (Table 2). The tect aquatic biota. Concentrations of a compound lower than PNEC are target compound list included 31 Water Framework Directive (WFD) considered as safe, while concentrations exceeding PNEC might pose a priority substances and three compounds from the WFD watch risk to aquatic life. Measured concentrations of target compounds at in- list (DIRECTIVE 2013/39/EU). Nineteen compounds were measured dividual sampling sites were compared to NORMAN PNEC values to de- by two laboratories in parallel and the average value is presented in termine frequency of PNEC exceedance (indicator 1). Maximum the results. Concentrations of compounds detected below limit of quan- environmental concentrations 95 (MEC95) of target compounds, tification (LOQ), were replaced by half of the LOQ for statistics and which are the 95th percentiles of the measured concentrations at all further calculations. More details are provided in SD (Table S3, S5 EDP sites, were also compared to NORMAN PNECs to calculate extent and S7). of exceedance (indicator 2). Indicators 1 and 2 were given risk scores and combined to derive a final NORMAN risk score. Target compounds 2.5. Linking effects and detected compounds were then prioritized according to the resulting NORMAN risk score.

For receptor mediated in vitro assays, mass balance calculations 3. Results and discussion were conducted using relative effect potencies (REPs) of known agonists according to Kinani et al. (2010) to quantify the contribution 3.1. Effect assessment of the detected target compounds to the observed biological activity. In brief, chemical equivalents (chem-EQ) were calculated The results of the bioassays are summarized in Table 3 (Table S6 in by multiplying the measured concentration (cl) of a known agonist SD). The most frequently observed effects were estrogenicity, zebrafish with its REP, reported in earlier studies (Table 6). The sum of acute embryo toxicity and algal growth inhibition. Androgenic and glu- chem-EQs of all agonists at each sampling site was then divided by cocorticoid activities were detected for 2 and 4 sites, respectively. Most the measured biological equivalent and the resulting ratio indicated biological activity was observed for neutral extracts (HR-X), whereas how much of the biological effect could be explained by detected acidic (HR-XAW) and basic extracts (HR-XCW) only rarely elicited ef- chemicals. fects, with the exception of the AChE inhibition assay (full results for For in vivo assays, the zebrafish embryo acute toxicity assay and algal acidic and basic extracts can be found in SD - Table S6). In the AChE in- growth inhibition assay, the link between measured compounds and hibition assay, several basic extracts were active (1-1: EC50 = REF 290; observed effects was calculated using toxic units (TU) similar to Booij 1–3: EC50 = REF 332; 3-3: EC50 = REF 106; 1-4; 2-2 and 2-3: LOEC = et al. (2014) and Kuzmanovic and Ginebreda (2014).Toxicityofdetect- REF 500), while most neutral extracts gave no effect. Biological effects ed compounds for fish and green algae was estimated by the program at individual sampling sites are shown and quantified in Tables 4 and

ECOSAR (v1.11) and the 96h-EC50 value of the most toxic ECOSAR 5. Our results indicate that all biological effects assessed in the EDP pro- chemical class (minECOSAR 96h-EC50i) was selected (US-EPA, 2012). gram are environmentally relevant, except for the AChE inhibition, TUs were calculated according to equation (Eq. 1), where ci is the mea- which was measurable only in highly enriched extracts (EC50s N REF sured concentration of individual target compounds, REF(EC50)ofa 100). With the receptor mediated assays, effects could be detected sample is REF at which EC50 was reached. Sum of TUs (∑TUs) of all even for diluted (estrogenicity) or moderately enriched samples (REF Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1859

b 20, androgenicity, anti-estrogenicity and glucocorticoid activity). In LOD LOD LOD b b b the algal growth inhibition and the FET assay, the EC50s were never lower than REF 10 and at some cases exceeded REF 100. Similar effect LOD LOQ LOD concentrations were seen in a study evaluating treated wastewaters b b b and river waters by Escher et al. (2014). The values for FET assay were approximately one order of magnitude lower than those reported by LOD LOD LOD b b b Neale et al. (2015) for samples collected from the highly diluted Danube. LOD 0,93 b In vivo thyroid activity in transgenic Xenopus, using whole water samples, was detected at one sampling site (4-2) in the Saale RB. Thy- LOD 5.2 LOD LOD roid activity of river environments is relatively rare and therefore, the Saale RB b b b detection of thyroid effects directly in a whole water sample, which had not been pre-concentrated, is an important ﬁnding (Chinathamby LOD LOQ 0.06 1.26 0.14 0.64 0.80 LOD NA NA NA NA NA LOD LOQ b b b b b et al., 2013; Inoue et al., 2009; Jugan et al., 2009). The thyroid activity, observed at site 4-2, may be related to strong embryo toxicity also de- LOD LOQ b b tected in the sample (Jomaa et al., 2014). plementary data (Table S4 and Table S6). LOD 6.4 LOQ LOD 3.2. Chemical analyses b b b

3.2.1. Target analysis LOD LOD LOD b b b Based on the detected bioactivities, only the neutral LVSPE extracts were subject to extensive target analysis of 151 compounds, which pro- LOD LOD LOD Emme RB b b b vided expansive chemical characterization of the sampling sites summarized in Table 3 and detailed in SD (Table S7). Out of the total 151 LOD LOD 0.07 0.07 0.04 0.15 LOD LOD 128 35 47 79 LOD compounds, 107 compounds were detected at least at one sampling b b b b b site, while 44 compounds, mainly glucocorticoids, steroids and WFD priority compounds, were below detection limits. Twelve compounds LOD LOD LOD b b b occurred at all sampling sites with the exception of the pristine mountain site in the Saale RB (4-1). The concentration of detected target com- − − LOD 0.32 LOD LOD LOD 1 μ 1 b b b b pounds ranged from a few ng L to a few gL , which is in line with a similar study conducted by Kuzmanovic and Ginebreda (2014) and LOD LOD LOD LOD Smital et al. (2013). As expected, the highest cumulative concentration Sava RB b b b b and the highest number of compounds (74) were detected at site 1-1, which is a WWTP efﬂuent, and the concentrations of the marker com- LOD LOQ LOQ LOD b b b b pounds carbamazepine and sucralose corresponded to the results of Loos et al. (2013), who evaluated 90 WWTP efﬂuents in an EU-wide LOD LOD LOD LOD 0.08 0.16 0.24 0.42 LOD b b b b b study. The highest median concentrations were detected for 1H- benzotriazole, sucralose, phenylbenzimidazolesulfonic acid, 4- LOD LOQ LOD LOD toluenesulfonamide and 5-methyl-1H-benzotriazole. b b b b

3.2.2. GC-MS nontarget screening LOD LOD 185 1-1 1-2 1-3 1-4 2-1 2-2 2-3 2-4 3-1 3-2 3-3 3-4 3-5 4-1 4-2 4-3 4-4 4-5 b b GC-MS nontarget screening had been previously applied in several studies dealing with surface waters (Gómez et al., 2009; Ma et al., 2014; Schymanski et al., 2011; Slobodnik et al., 2012). In the presented work, it provided valuable complementary information to the target analysis and proved to be a useful analytical tool for the assessment of 0.4 1.2 30.5 120 618 0.009 0.024 1.85 2.4 6.6 0.021 0.085 2.7 complex environmental samples. In the nine analyzed samples (Danube and Emme RBs), 55 compounds were tentatively identified and quantified, while nineteen remained unknown. The identified compounds often appeared at multiple sampling sites and one compound (1,3- diacetylbenzene) was detected at all sites. 1,3-Diacetylbenzene, a transformation product of the aromatic solvent and minor jet fuel component diethylbenzene, is not specifically toxic unlike its ortho isomer Glucocorticoid activity Dex-EQ [ng/L]: Endpoint, BEQ, unit LOD LOQ Danube RB Anti-androgenicity Flu-EQ [ng/L]: E2-EQ [ng/L]: Anti-estrogenicity OH-Tam-EQ [ng/L]: Androgenicity DHT-EQ [ng/L]: 1,2-diacetylbenzene, which has been shown to cause neurotoxicity and neuropathy (Kim et al., 2001; Sabri et al., 2007). The estimated con- −1 bioassays expressed by biological effect equiva lents (BEQs). Shown are the results for the neutra l sorbent (HR-X) extracts. For more details see Sup centrations of several compounds ranged from high ng L to low μgL−1, e.g., gabapentin, a common anticonvulsant and analgesic, or

in vitro phthalimide, a transformation product of the fungicide folpet (Viviani- −1 MELN cells (100) Estrogenicity Nauer et al., 1997). Concentrations above 7 and 1 μgL at sampling – – sites 1-2 and 1-1, respectively, were estimated for TMDD (2,4,7,9- tetramethyl-5-decindiol), a high production volume chemical used as a wetting and anti-foaming agent in paper, ink, pesticide and adhesives industries. This compound has been linked to embryo toxicity in

CALUX® (100) zebraﬁsh by Vincze et al. (2014), which may partially explain the effects – seen in the zFET assay (see Table 5). In the WWTP efﬂuent sample (1-1), (maximal tested REF) GR Bioassay ER-mediated activity MDA-kb2 cells (100) AR-mediated activity −1

Table 4 Results of receptor-mediated μ NA - not assessed. concentrations above 1 gL were also estimated for 1860 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868

Table 5 Mass balance calculations for estrogenic, androgenic and glucocorticoid activities in the EDP samples. Chemical equivalents (chem-EQ), calculated by multiplying measured concentrations of known agonists with their relative potencies in the particular bioassays, were compared to biological equivalents (bio-EQ) and the resulting ratio (expressed as a percentage) shows the extent to which the observed biological activity can be explained by target compounds.

Compound name CAS LOQ REP Reference Danube RB Saale RB [ng/L] 1-1 1-4 4-1 4-2 4-3 4-4 4-5

Estrogenicity (MELN) 4-Androstene-3,17-dione 63-05-8 1.5 9.70E-07 Creusot et al. 10.1 ND ND ND ND ND ND (2014) Benzophenone-3 131-57-7 1.0 7.04E-07 Molina-Molina et 141.6 bLOQ ND 1.8 ND ND ND al. (2008) Bisphenol A 80-05-7 4.0 4.49E-05 Neale et al. (2015) 62 bLOQ ND 15.0 16 6.0 ND (0.2%) (0.5%) Endosulfan aplha and beta 115-29-7 1.0 2.00E-06 Lemaire et al. ND ND ND ND ND ND 90.0 (2006) Estrone 53-16-7 0.1 0.02 Kinani et al. ND 0.38 ND ND 0.3 ND 0.23 (2010) (53%) (23%) (3.2%) Nonylphenol 25154-52-3 10 3.30E-06 Creusot et al. 74.0 18.0 ND ND ND ND 119.0 (2013) Nonylphenoxyacetic acid 3115-49-9 1.0 4.10E-04 Creusot et al. 1418 5.6 ND 71 25 55 75 (2013) (31%) (3%) (2.3%) (7.2%) (3.5%) (3.9%) Octylphenol 140-66-9 3.0 1.10E-04 Creusot et al. 14 ND ND ND ND ND 7.5 (2013) (0.1%) (0.1%) Progesterone 57-83-0 1.0 7.50E-07 Creusot et al., 2014 ND ND ND 1.2 ND ND ND Triphenylphosphate 115-86-6 5.0 6.90E-06 Creusot ESPR in ND ND ND ND ND ND ND revision bio-E2-EQ 1850 0.08 0.06 1.26 0.14 0.64 0.80 chem-E2-EQ 0.59 0.04 0.00 0.03 0.04 0.02 0.06 chem/bio ratio (%) 31.7 56 0 2.3 30.7 3.6 7.2 Androgenicity 4-Androstene-3,17-dione 63-05-8 1.50 3.90E-03 Ait-Aissa et al. 10.1 ND ND ND ND ND ND (MDAkb2) (2010) (1.5%) Epi-Androsterone 481-29-8 2.0 8.10E-06 Bellet et al. (2012) 4.1 ND ND ND 7.47 ND 4.0 Trenbolone 10161-33-8 2.0 1.8 Blake et al. (2010) ND ND ND ND ND ND ND bio-DHT-EQ 2.67 0.9 chem-DHT-EQ 0.04 0 0 0 6.05E-05 0 3.20E-05 chem/cio ratio (%) 1.5 Glucocorticoid activity 6-alpha-Methylprednisolone 83-43-2 0.63 0.54 Macikova et al. 1.24 bLOQ ND ND ND ND ND (GR CALUX) (2014) (2.2%) Cortisone 53-06-5 1.0 8.00E-04 Schriks et al. ND ND ND ND bLOQ bLOQ ND (2010) Hydrocortisone 50-23-7 0.86 3.60E-02 Macikova et al. ND ND ND ND ND ND 2.5 (2014) bio-Dex-EQ 30.50 5.20 chem-Dex-EQ 0.67 0 0 0 0 0 0.09 chem/bio ratio (%) 2.2 0

Compound name CAS LOQ REP Reference Sava RB Emme RB [ng/L] 2-1 2-2 2-3 3-1 3-2 3-3 3-4

Estrogenicity (MELN) 4-Androstene-3,17-dione 63-05-8 1.5 9.70E-07 Creusot et al. ND ND ND ND ND ND 1.7 (2014) Benzophenone-3 131-57-7 1.0 7.04E-07 Molina-Molina et ND 1.7 bLOQ ND ND 1.7 ND al. (2008) Bisphenol A 80-05-7 4.0 4.49E-05 Neale et al. 5.8 17 12 11 bLOQ bLOQ 40 (1%) (2015) (0.2%) (0.3%) (0.1%) (0.8%) Endosulfan aplha and beta 115-29-7 1.0 2.00E-06 Lemaire et al. ND ND bLOQ ND ND ND ND (2006) Estrone 53-16-7 0.1 0.02 Kinani et al. 0.43 0.45 1.28 0.23 0.19 0.28 0.57 (2010) (30%) (21%) (34%) (38%) (32%) (75%) (40%) Nonylphenol 25154-52-3 10 3.30E-06 Creusot et al. 22.0 22.0 17.0 315 bLOQ 10 33.0 (2013) (1.6%) (0.1%) Nonylphenoxyacetic acid 3115-49-9 1.0 4.10E-04 Creuszt et al. 12 6.2 12 3.9 12 2 (2%) 65 (17%) (2013) (3%) (1%) (1.2%) (2.4%) (7.6%) Octylphenol 140-66-9 3.0 1.10E-04 Creusot et al. ND ND ND 14 ND ND ND (2013) (2.3%) Progesterone 57-83-0 1.0 7.50E-07 Creusot et al. ND ND ND ND ND ND ND (2014) Triphenylphosphate 115-86-6 5.0 6.90E-06 Creusot ESPR in ND bLOQ ND ND ND ND ND revision bio-E2-EQ 0.16 0.24 0.42 0.07 0.07 0.04 0.16 chem-E2-EQ 0.05 0.05 0.15 0.03 0.03 0.03 0.09 chem/bio ratio 33 23 35 45 40 77 59 (%) Androgenicity 4-Androstene-3,17-dione 63-05-8 1.50 3.90E-03 Ait-Aissa et al. ND ND ND ND ND ND 1.7 (MDAkb2) (2010) Epi-Androsterone 481-29-8 2.0 8.10E-06 Bellet et al. ND ND ND ND ND 3.6 2.6 (2012) Trenbolone 10161-33-8 2.0 1.8 Blake et al. 2.7 ND ND ND ND ND ND Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1861

Table 5 (continued)

Compound name CAS LOQ REP Reference Sava RB Emme RB [ng/L] 2-1 2-2 2-3 3-1 3-2 3-3 3-4

(2010) bio-DHT-EQ chem-DHT-EQ 4.86 0 0 0 0 2.92E-05 6.81E-03 chem/bio ratio (%) Glucocorticoid activity 6-alpha-Methylprednisolone 83-43-2 0.63 0.54 Macikova et al. ND ND ND ND 2.6 ND ND (GR CALUX) (2014) Cortisone 53-06-5 1.0 8.00E-04 Schriks et al. 1.8 ND bLOQ ND ND bLOQ ND (2010) Hydrocortisone 50-23-7 0.86 3.60E-02 Macikova et al. ND ND 2.56 ND ND ND ND (2014) (29%) bio-Dex-EQ 0.32 chem-Dex-EQ 0.00 0 0.09 0 1.39 0 0 chem/bio ratio 28.9 (%)

ND non-deteceted. bLOQ - below limit of quantiﬁcation.

benzenesulfonamide (an intermediate for organic synthesis), 1H- detection of estrogenicity, which is a highly relevant endpoint for mon- isoindole-1,3(2H)-dione (a synthetic androgen receptor antagonist, ac- itoring of European surface waters (Loos et al., 2013; Jarošová et al., cording to Salvati et al. (2005)) and 1,1-dichloro-2,2-diphenylethane. 2014b), is necessary because large portions of the observed effects still Complete results of GC–MS nontarget screening including analysis de- cannot be explained by target chemical analysis mainly due to problems tails can be found in SD (Table S8). of analytical detection limits. The risk of the observed estrogenicity can be expressed by comparison of the measured bio-E2-EQs with safe es- 3.3. Linking effects and detected compounds trogenic equivalents regarding steroid estrogens (EEQ-SSE) as proposed by Jarošová et al. (2014a, 2014b). Based on the MELN cells, the author 3.3.1. ER-mediated activity (MELN cells) assay derived EEQ-SSEs, threshold values for detected bioactivity expressed Estrogenic activity was observed for 15 out of 18 sampling sites and as bio-E2-EQ, below which no adverse effects are expected in aquatic the effects could be quantified for 14 of them. The biological estradiol biota due to estrogens. The EEQ-SSE for short term exposure equivalents (bio-E2-EQs) ranged from 0.06 to 1.85 ng L−1 and these (0.8 ng L−1 bio-E2-EQ) was exceeded at three sites (1-1, 4-2 and 4-4) levels are similar to those reported by Jálová et al. (2013) for WWTP ef- and the EEQ-SSE for chronic exposure (0.2 ng L−1 bio-E2-EQ) was fluent and river water (Table 4). The mass balance calculations resulted exceeded at another three sites (2-2, 2-3 and 4-5), which means that in ratios of observed estrogenic activity that could be explained by the one third of the sampling locations within our study were characterized detected ER-agonists, ranging from 0 to 77% (Table 5). For the sampling by estrogenic activity above the threshold for concern. site 1-1 with the highest estrogenic activity (1.85 ng L−1 E2-EQ), the ef- Anti-estrogenic activity was detected at five out of 13 sampling sites fect could be partially explained by the three chemicals with the OH-Tam EQ reaching from 35 to 185 ng/L−1. Anti-estrogenicity nonylphenoxyacetic acid, bisphenol A and octylphenol, which contrib- was observed in WWTP effluent and river water samples in many earli- uted 31% (0.581 ng L−1 E2-EQ), 0.2% (0.003 ng L−1 E2-EQ) and 0.1% er studies, however, the use of different biological models makes any di- (0,002 ng L−1 E2-EQ), respectively. Interestingly, none of the target es- rect comparisons of the measured effect levels very difficult (Gehrmann trogens (E1, E2, EE2 or E3 with LODs of 0.1, 0.3, 0.4 and 1 ng L−1,respec- et al., 2016; Jálová et al., 2013; Oh et al., 2006). The measured anti-estro- tively) was detected at this site, which is a direct WWTP effluent. genic effects could not be explained by the content of target anti-estro- Estrone (E1) was a dominant driver of the observed estrogenicity for gens (4-hydroxytamoxifen, anastrozole, raloxifene and tamoxifen) as sites (1-4; 2-1; 2-2; 2-3; 3-1; 3-2; 3-3; 3-4) where its contribution their concentrations were below the LOQ at all active sites except for ranged from 21 to 75%. Nonylphenoxyacetic acid, bisphenol A and site 1-1. However, the concentrations of anastrozole and raloxifene octylphenol also accounted for certain portions of the observed (1.6 and 1.3 ng L−1, respectively) detected at site 1-1 are too low to elicit estrogenicity at several sites (0.1–17%). It is important to note that de- the observed anti-estrogenicity (187 ng/L−1 OH Tam-EQ), since the spite ever-decreasing LODs of chemical analytical methods, a mixture contribution of raloxifene with REP = 4.4 accounts for only 3.1% of estrogens, where each component occurs at a concentration below (Witters et al., 2010). The contribution of anastrozole could not be cal- LOD, may still elicit measurable biological effects (Conley et al., 2015). culated because it's REP for MELN or similar model is unknown to the Instrumental analytical methods for determination of estrogens in envi- authors. The detected anti-estrogenity likely results from combined ac- ronmental samples remains a challenge due to high sample complexity, tion of nontarget pharmaceuticals, PAHs (anthracene and fluoranthene) need of appropriate clean-up, special analytical procedures, e.g., deriva- or other compounds (triclosan, PFOA and PFOS) directly or indirectly af- tization, and their very low physiologically active concentrations (Fang fecting the ER-mediated activity (Henry and Fair, 2013; Hilscherova et et al., 2016). The limits of quantification for E2 and EE2 are mostly above al., 2002). the proposed annual average Environmental Quality Standards under the WFD (AA-EQS) of 0.4 and 0.035 ng L−1, respectively, and therefore 3.3.2. AR-mediated activity (MDA-kb2 cells) assay bioanalytical tools are very valuable for environmental monitoring Androgenic activity was detected at four out of 18 sampling sites and (Kunz et al., 2017). the effects above LOQ of the assay were observed at sites 1-1 and 4-2 Low estrogenic activity (0.06 ng L−1E2-EQ) was observed also at the with 2.7 and 0.93 ng L−1 DHT-EQ, respectively (Table 4). Similar levels pristine mountain site (4-1), where none of the target ER-agonists was of DHT-EQ were reported by Bain et al. (2014). At both active sites, the detected. This effect could be caused by natural compounds such as contribution of target AR-agonists was negligible and most of the ob- phytoestrogens, phytosterols, lignans or mycotoxins (Ribeiro et al., served androgenicity remained unexplained (Table 5). At site 1-1, the 2016). Our results demonstrate that the use of bioanalytical tools for contribution of 4-androstene-3,17-dione reached only 1.5% and at site 1862

Table 6 Results of bioassays expressed as LC50,EC50 or LOEC values, where relative enrichment factor (REF) is used as the unit. Shown are the results for the neutral sorbent (HR-X) extracts (algal growth inhibition, AChE inhibition and zebraﬁsh embryo acute toxicity)andwholewatersamples(in vivo thyroid activity). Active samples with resulting REF b 20 are marked in red, REF between 20 and 100 in orange, equaling 100 in yellow and N100 in green. For more details see supplementary data (Table S4 and S6). 601 Environment Total the of Science / al. et Tousova Z.

Bioassay Sample type (maximal Endpoint, unit Danube RB Sava RB Emme RB Saale RB tested REF) 1-1 1-2 1-3 1-4 2-1 2-2 2-3 2-4 3-1 3-2 3-3 3-4 3-5 4-1 4-2 4-3 4-4 4-5 Growth rate inhibition Algal growth 17.4 NA NA NA 259 94.6 132 NA 158NANANANANA61.6 NA 38.9 69.4 EC50 [REF]: inhibition Growth rate inhibition (100) 11 100 100 100 100 33 33 100 100 >100 >100 100 >100 >100 33 100 33 11 LOEC [REF]: AChE inhibition NA NA NA NA NA NA NA NA NA NA NA NA NA NA 106 NA NA 231 AChE EC50 [REF]: inhibition AChE inhibition (500) >250 >250 >250 >250 >250 >250 >250 >250 >250 >250 >250 >250 >250 >250 NR >250 >250 NR Neutral HR-X LOEC [REF]: sorbent Survival <100 <100 <100 <100 19.1 72 23.3 NA 22.3 NA NA 38.6 NA 83.1 13.3 38.2 36.9 12.3 LC [REF] 50 – 0 21)1849 (2017) 602 Survival NR NR NR NR 12.5 25 25 >100 NR >100 >100 NR 100 25 12.5 50 25 12.5 Zebrafish LOEC [REF] embryo acute Sublethal endpoints toxicity (100) NA NA NA NA NA NA NA NA 1.92 NA NA 4.95 NA NA NA NA NA NA EC50 [REF] Sublethal endpoints –

NA NA NA NA NA NA NA >100 NR >100 >100 NR 100 25 6,25 50 50 12,5 1868 LOEC [REF] Whole water In vivo thyroid Thyroid activity >1 >1 >1 >1 >1 >1 >1 >1 >1 >1 >1 >1 >1 >1 <1 >1 >1 >1 sample activity (1) LOEC [REF] NA – not assessed NR – not reported Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1863

4-2, none of the target AR-agonists was detected. The most potent AR- Escher, 2014), may possibly have contributed to the observed embryo agonists T and DHT were not detected at any sampling site within the toxicity at site 4-1, which was unaffected by anthropogenic pressures. EDP (LODs 0.4 and 1.5 ng L−1, respectively). Interestingly, the chem- Therefore, the role of natural compounds in acute toxicity to ﬁsh embry- DHT-EQ at site 2-1 reached 4.9 ng L−1, due to the presence of os observed in surface water should be addressed by further research. trenbolone; however, no androgenicity could be detected in the bioassay. Similarly to our results, only partly explained androgenicity in 3.3.5. Algal growth inhibition assay WW was reported by Bellet et al. (2012). But in contrast, Thomas et al. Toxicity of surface waters to microalgae has been described in many (2002) found that up to 99% of observed androgenicity in estuarine wa- earlier studies (Emelogu et al., 2013; Escher et al., 2006) and it is a cause ters from the UK could be explained by natural steroids or a steroid of major environmental concern because microalgae constitute the base transformation product of DHT. of aquatic food webs (Booij et al., 2015). In our study, algal growth inhibition was detected at 14 out of 18 sampling sites and the effect concen-

3.3.3. GR-CALUX® assay trations were relatively high (EC50 N REF 100) at seven of them (Table GR activity was detected at four out of the 18 sampling sites distrib- 6). Extracts from four sites (1-1; 4-2; 4-4; 4-5) elicited higher toxicity uted over all investigated RBs and the Bio-Dex-EQ values were slightly (EC50 = REF 17 to 69) to algae and the effects were linked to detected higher than those reported for treated WW and river water by target chemicals (Table 7, in detail SD - Table S11). The proportion of ef- Macikova et al. (2014). Eleven out of 20 target glucocorticoids were fects which could be explained by detected compounds was higher than not detected at any sampling site and the concentrations of nine detect- in the FET assay and always exceeded 30%. At site 1-1, the ∑TUs ed glucocorticoids were similar to those reported by Isobe et al. (2015). reached higher than one, meaning that the detected compounds were The highest GR activity, 30.5 ± 6.4 ng L−1 bio-Dex-EQ was detected at expected to cause even stronger effects, according to the approach ap- the site 1-1, where chemical analysis of glucocorticoids identified 6α- plied. The herbicides diuron, atrazine and terbuthylazine were found methylprednisolone, but its concentration explained only 2% of the ob- to be the dominant components causing the observed toxicity. The con- served biological activity (Table 5). A bio-Dex-EQ of 6.4 ± 1.2 ng L−1 centrations of diuron (150 ng L−1), atrazine (27 ng L−1)and was detected at the site 3-4; however, no glucocorticoids were detected terbuthylazine (5 ng L−1)weresufficient to explain the observed effect in the sample. Similarly, no glucocorticoids were detected at site 4-2 independently of the other compounds detected in the extract. This with observed glucocorticoid activity (5.2 ± 0.4 ng L−1 bio-Dex-EQ). finding is in line with the results of earlier studies, which indicated her- The site 2-3 showed activity of 0.3 ± 0.1 ng L−1 bio-Dex-EQ and cortisol bicides were responsible for most of the toxicity to algae detected in was identified at this site at concentrations that could explain 29% of the treated wastewaters and surface waters (Tang and Escher, 2014; biological activity. The chemical results also showed a small amount of Cleuvers, 2003). ECOSAR analysis additionally revealed other contribu- cortisone and hydrocortisone at sites 2-1 and 4-5 (1.8 ng L−1 and tors including sucralose, carbamazepine and azithromycin that collec- 2.5 ng L−1, respectively). While no GR activity was detected, at either tively contributed to about another 44% of the toxicity. Our result of them. GR activity was mostly detected in samples taken from the ef- (∑TUs N 1) may suggest that the highly complex extract originating fluent or downstream of WWTPs in the four RBs, while sites upstream of from WWTPs contained some components that reduced the combined the WWTPs showed no GR activity. It has already been shown that effect of the compounds mentioned above. However, it may also be as- WWTPs cannot completely remove glucocorticoids (Kugathas et al., sociated with the fact that the toxic components of the complex envi- 2012). Inspite of extensive target analysis of glucocorticoids, most of ronmental mixture may act by different mode of action used in our the GR activity could not be explained and thus a higher-tier EDA ap- study, in which case the use of concentration addition model is not approach (HT-EDA) should be used at the relevant sites in order to further propriate (Ginebreda et al., 2014). Algal growth inhibition detected at investigate which compounds are responsible for the detected activi- sites 4-2, 4-4, and 4-5 could be caused by the combined effect of several ties. Indeed, an HT-EDA study investigated the 3-4 site in Switzerland, compounds. The toxicity of sucralose, and carbamazepine was predicted including upstream and effluent of the WWTP (unpublished data). as relatively high by the ECOSAR model and these compounds were therefore identified as major contributors to the observed toxicity. 3.3.4. Zebrafish embryo acute toxicity test assay (FET) However, sucralose and carbamazempine are unlikely to cause effects −1 In our study, acute toxicity on the zebrafish embryos was observed on green algae as experimental EC50s higher than 0.1 g L were report- at 15 out of 18 sampling sites and the toxic effects were linked to target ed for both of the compounds (Ferrari et al., 2003; Tollefsen et al., 2012). chemicals at 10 sites, where the embryo toxicity could be quantified This shows that the use of the ECOSAR model has its limitations and re- with LC50 values (Table 6). Fish embryo toxicity of surface waters was sults should therefore always be treated cautiously. Further investiga- reported in several studies (Maier et al., 2014; Keiter et al., 2006) and tion of the role of nonherbicidal components of WWTP effluents it is believed that micropollutants originating from wastewater, espe- contributing to the toxicity on algae with focus on the use of available cially pharmaceuticals, play an important role for this endpoint (van experimental data was carried out within a HT-EDA study (unpublished Woudenberg et al., 2014; Petrie et al., 2014). Our approach of linking ef- data). fects in the FET assay to the detected target chemicals was based on the assumption that all chemicals act through the same mode of action and 3.4. Risk assessment and prioritization that summation of toxic units' results in the overall portion of effect that can be explained by target chemicals (Table 7, in detail SD - Table S10). NORMAN PNEC values were available for 94 out of 151 target com- Similar simplification was used before in a study by Neale et al. (2015). pounds, while 73 PNECs were derived from experimental data or The percentage of effects explained by target compounds ranged from 0 existing EQS values and 21 PNECs were predicted (P-PNECs) using to 21% and several main contributors to embryo toxicity appeared at QSAR tools (Table 2). Almost all sampling sites exceeded the PNEC multiple sampling sites, e.g., surfactants-derived compounds like values of at least two compounds except for 4-1. At site 1-1, a direct nonylphenol, nonylphenoxyacetic acid, octylphenol and WWTP effluent, PNEC exceedances were most abundant (13) as op- nonylphenoxyacetic acid or pesticides such as diazinon and fipronil. posed to site 4-3, affected by industry, with the least PNEC exceedances Surprisingly, embryo toxic effects were also detected at the pristine (2). The highest cumulative extent of PNEC exceedance (373) was ob- mountain site (4-1) with an approximately LC50 of REF 83. This effect served at sampling site 3-4 downstream of a WWTP effluent. The could not be explained by target analysis because only triethylcitrate MEC95 of 15 compounds exceeded the PNEC value and measured con- (which is nontoxic to fish) was detected in the sample. Co-extracted centrations of another 6 compounds exceeded their PNEC values at natural dissolved organic material, which showed biological activity in least at one sampling site. Perfluorooctanesulfonic acid exceeded the other bioassays (Bittner et al., 2009; Janošek et al., 2007; Neale and PNEC most frequently (12 sites) and its MEC95 exceedance of the 1864

Table 7 Link between effects observed in the in vivo bioassays (zebraﬁsh embryo acute toxicity assay and algal growth inhibition assay) and measured concentrations of target compounds. Listed are the main contributors to the observed toxicity and their relative contribution (expressed as percentage) based on ECOSAR toxicity predictions for ﬁsh and algae. The sum of toxic units (∑TUs) shows the total extent to what the observed biological activity can be explained by target compounds. For more details see supplementary data (Table S10 and S11).

Zebraﬁsh embryo acute toxicity assay 601 Environment Total the of Science / al. et Tousova Z.

Sampling site 2-1 2-2 2-3 3-1 3-4 4-1 4-2 EC50 [REF] 19.2 72 23.3 22.3 38.6 83.1 13.3 Sum of toxic 1.82 6.32 5.67 20.60 9.69 0 0.84 units [%] Main Nonylphenol 1.17 Nonylphenol 4.40 Chlorpyrifos 2.94 Nonylphenol 19.51 Fipronil 4.71 Didecyldimethyl- 0.42 contributors to ammonium observed Fipronil 0.38 Fipronil 1.42 Fipronil 1.18 Fipronil 0.69 Nonylphenol 3.54 Nonylphenoxy- 0.11 toxicity [%] acetic acid Diazinon 0.21 Azithromycin 0.14 Nonylphenol 1.10 Octylphenol 0.27 Diazinon 0.53 5-Methyl-1H- 0.07 benzotriazole Diazinon 0.26 Nonylphenoxy-acetic acid 0.29

Algal growth inhibition assay Sampling site 4-3 4-4 4-5 1-1 4-2 4-4 4-5 EC50 [REF] 38.2 36.9 12.3 17.4 61.6 38.9 69.4 ∑TUs[%] 1.30 0.72 4.78 166 30.1 42.9 32.1

Main Fipronil 0.99 Nonylphenoxy- 0.32 Nonylphenol 4.07 Diuron 109 Sucralose 14.22 Caffeine 18.58 Sucralose 13.75 – contributors to diethoxyacetic acid 1849 (2017) 602 observed Nonylphenoxy- 0.11 Nonylphenoxy-acetic 0.23 Fipronil 0.35 Sucralose 38.3 Carbamazepine 8.75 Terbuthylazine 12.86 Carbamazepine 7.87 toxicity [%] acetic acid acid Nonylphenoxy- 0.09 Pentachloro-benzene 0.04 Nonylphenoxy-acetic 0.11 Carbamazepine 3.80 Hexamethoxy- 1.45 Diuron 4.33 Nonylphenol 6.21 diethoxyacetic acid methylmelamine

acid – 1868 Bisphenol A 0.05 5-Methyl-1H- 0.03 Octylphenol 0.08 Terbuthylazine 2.83 5-Methyl-1H-benzotriazole 1.35 Triﬂuralin 2.67 benzotriazole Nonylphenoxy- 0.07 Azithromycin 1.86 1-H Benzotriazole 0.92 Sucralose 2.36 diethoxyacetic acid Atrazine 1.19 Didecyldimethyl-ammonium 0.52 Carbamazepine 1.25 Fipronil 1.11 Hexamethoxy-methylmelamine 0.86 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1865

Table 8 Prioritization of target compounds based on NORMAN risk score. For more details see supplementary data (Table S9).

Rank Compound name CAS Compound LOQ PNEC PNEC MEC95 Frequency of NORMAN hazard NORMAN hazard NORMAN group/usage [ng/L] [ng/L] type exceedance PNEC ratio for frequency ratio for extent of risk score pattern of PNEC exceedance of exceedance exceedance [ng/L] [%] N = 18 (indicator 1) (indicator 2)

1 Perﬂuorooctanesulfonic 1763-23-1 Surfactant, WFD 10 0.13 PNEC 239.23 67 0.67 0.50 0.583 acid priority 2 Terbuthylazine-desethyl 30125-63-4 Herbicide, 1.0 2.4 PNEC 3.38 61 0.61 0.10 0.356 transformation Product 3 Diazinon 333-41-5 Insecticide 1.0 1.0 PNEC 10.75 50 0.50 0.20 0.350 4 2,4-Dinitrophenol 51-28-5 Herbicide, 1.0 1 PNEC 7.70 56 0.56 0.10 0.328 transformation product 5 Diclofenac 15307-86-5 Pharmaceutical, 1.0 10 PNEC 44.34 39 0.39 0.20 0.294 analgesic, WFD watch list 6 Nonylphenol 25154-52-3 Surfactant, WFD 10 21 PNEC 7.17 39 0.39 0.10 0.244 priority, xenoestrogen 7 Triclosan 3380-34-5 Biocide 1.0 0.7 PNEC 3.43 28 0.28 0.10 0.189 8 Ibuprofen 15687-27-1 Pharmaceutical, 10 10 PNEC 2.01 22 0.22 0.10 0.161 analgesic 9 Azithromycin 83905-01-5 Pharmaceutical, 1.0 90 PNEC 11.36 11 0.11 0.20 0.156 antibiotic 10 Fluoranthene 206-44-0 PAH, WFD 1.0 6.3 PNEC 2.19 17 0.17 0.10 0.135 priority 11 Caffeine 58-08-2 Stimulant in 1.0 100 PNEC 2.37 17 0.17 0.10 0.133 beverages, marker compound 12 Dichlorvos 62-73-7 Insecticide, WFD 1.0 0.6 PNEC 6.40 11 0.11 0.10 0.106 priority 12 Erythromycin 114-07-8 Pharmaceutical, 10 40 PNEC 1.34 11 0.11 0.10 0.106 antibiotic 13 Nonylphenoxyacetic 3115-49-9 Surfactant, 1.0 176 P-PNEC 1.57 6 0.06 0.10 0.078 acid xenoestrogen 14 Triphenylphosphate 115-86-6 Plasticizer, ﬂame 5.0 30 PNEC 1.72 6 0.06 0.01 0.033 retadant 15 Bentazone 25057-89-0 Herbicide 1.0 60 PNEC No 6 0.06 0.00 0.028 15 Carbamazepine 298-46-4 Pharmaceutical, 1.0 500 PNEC No 6 0.06 0.00 0.028 antiepileptic, marker compound 15 Clarithromycin 81103-11-9 Pharmaceutical, 1.0 60 PNEC No 6 0.06 0.00 0.028 antibiotic 15 Diethyltoluamid 134-62-3 Repellent 1.0 360 P-PNEC No 6 0.06 0.00 0.028 15 Terbutryn 886-50-0 Herbicide, WFD 1.0 6.5 PNEC No 6 0.06 0.00 0.028 priority 15 Verapamil 152-11-4 Pharmaceutical, 1.0 30 P-PNEC No 6 0.06 0.00 0.028 antihypertensive

PNEC was also the highest (239). Therefore this compound reached the analgesic), transformation products of terbuthylazine (a widely used highest NORMAN risk score and was ranked number one in the EDP pri- herbicide), triphenylphosphate (plasticizer and flame retardant, ority list. The other compounds were ranked accordingly and the results whose release into the environment is expected to increase due to in- are shown in Table 8 (detailed results including PNEC exceedances at increased production and use after the phase out of some brominated dividual sites can be found in SD - Table S9). The EDP priority list con- flame retardants), nonylphenoxyacetic acid (surfactant biotransforma- tains 21 compounds included four pesticides and two pesticide tion product (Ahel et al., 1994) and xenoestrogen), and a group of transformation products, seven pharmaceuticals (three of them macrolide antibiotics such as azithromycin, which can reach relatively macrolide antibiotics), three surfactant-derived compounds and one high concentrations in treated wastewaters, due to low elimination PAH, biocide, repellent and plasticizer each. Five compounds on the rates in the WWTP process (McArdell et al., 2003). The EDP priority EDP priority list are also WFD priority compounds and one compound list partially overlaps with compounds prioritized in earlier studies. is included in the WFD watch list. Our results indicate that these WFD Von der Ohe et al. (2011), who worked with a much larger dataset priority compounds are still relevant for monitoring in the EU, even (MODELKEY database), identified 73 out of 500 compounds that though several of them were phased out for major applications e.g., exceeded the PNEC values and four compounds, i.e. diazinon, nonylphenol, perflouorooctanesulfonic acid, dichlorvos and terbutryn. terbutryn, terbuthylazine-desethyl, and fluoranthene, were also Studies by Von der Ohe et al. (2011) and Kuzmanovic and Ginebreda identified as priority compounds in our study. In a case study per- (2014) also reported several WFD priority compounds in their priority formed in Greece by Thomaidi et al. (2015), 30 compounds were pri- lists. The EDP priority list includes several CECs which might be interest- oritized and 6 of them can also be found in the EDP priority list, i.e. ing for future surface water monitoring, e.g., ibuprofen (a widely used azithromycin, clarithromycin, caffeine, diclofenac, triclosan and 1866 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 nonylphenol. The EDP priority compounds diazinon and (ESR11), Jean Froment (ESR12), Xiyu Ouyang (ESR13) and Victoria nonylphenol were also found among ten most important contami- Osorio Torrens (ER). nants within a prioritization exercise from Spain carried out by Kuzmanovic and Ginebreda (2014). On the contrary, there was no Appendix A Supplementary data overlap with compounds prioritized in the Sava RB in a study by Smital et al. (2013). Supplementary data associated with this article can be found in the online version, at http://dx.10.1016/j.scitotenv.2017.06.032. These 4. Conclusion data include the details on sampling, sampling sites, analytical methods, chemicals, and the full raw and assessment dataset. These data include The European demonstration program showed a successful applica- the Google maps of the most important areas described in this article. tion of a newly developed simplified effect-directed analysis (EDA) protocol. The novel onsite LVSPE sampling device proved to be a valuable instrument for integrative effect-based and chemical monitoring pur- References poses. The set of selected bioassays enabled detection of effects relevant Aerni, H.R., Kobler, B., Rutishauser, B.V., Wettstein, F.E., Fischer, R., Giger, W., for surface waters on an EU-wide scale whereas the most frequently ob- Hungerbuhler, A., Marazuela, M.D., Peter, A., Schonenberger, R., Vogeli, A.C., Suter, served were estrogenicity, fish embryo toxicity and toxicity to algae. M.J.F., Eggen, R.I.L., 2004. Combined biological and chemical assessment of estrogenic activities in wastewater treatment plant effluents. Anal. Bioanal. Chem. 378:688–696. Nontarget screening with GC-MS helped to complement the target http://dx.doi.org/10.1007/s00216-003-2276-4. analysis by revealing several frequently occurring compounds outside Ahel, M., Giger, W., Koch, M., 1994. Behaviour of alkylphenol polyethoxylate surfactants in the target list, such as 1,3-diacetylbenzene, gabapentin, phthalimide the aquatic environment—I. Occurrence and transformation in sewage treatment. – fi fi Water Res. 28:1131 1142. http://dx.doi.org/10.1016/0043-1354(94)90200-3. and TMDD. These ndings con rmed that the nontarget screening Ait-Aissa, S., Laskowski, S., Laville, N., Porcher, J.M., Brion, F., 2010. Anti-androgenic activ- techniques may bring significant added value to the classical chemi- ities of environmental pesticides in the MDA-kb2 reporter cell line. Toxicol. Vitr. 24: cal target monitoring anchored in the current EU legislation (WFD). 1979–1985. http://dx.doi.org/10.1016/j.tiv.2010.08.014. Altenburger, R., Ait-Aissa, S., Antczak, P., Backhaus, T., Barceló, D., Seiler, T.-B., Brion, F., Our study clearly demonstrated the need for effect-based tools in Busch, W., Chipman, K., de Alda, M.L., de Aragão Umbuzeiro, G., Escher, B.I., Falciani, surface water monitoring programs because most biological effects F., Faust, M., Focks, A., Hilscherova, K., Hollender, J., Hollert, H., Jäger, F., Jahnke, A., could not be explained by an extensive list of target compounds. Kortenkamp, A., Krauss, M., Lemkine, G.F., Munthe, J., Neumann, S., Schymanski, Thus, higher-tier EDA studies would be necessary to possibly close E.L., Scrimshaw, M., Segner, H., Slobodnik, J., Smedes, F., Kughathas, S., Teodorovic, I., Tindall, A.J., Tollefsen, K.E., Walz, K.-H., Williams, T.D., Van den Brink, P.J., van this gap for the bioactive samples. Risk assessment revealed that Gils, J., Vrana, B., Zhang, X., Brack, W., 2015. Future water quality monitoring - most EDP sites were at risk because a minimum of two target adapting tools to deal with mixtures of pollutants in water resource management. – – compounds exceeded the predicted no effect concentration Sci. Total Environ. 512 513:540 551. http://dx.doi.org/10.1016/j.scitotenv.2014.12. 057. values. Based on the estimated risk, 21 compounds were priori- Bain, P., Williams, M., Kumar, A., 2014. Assessment of multiple hormonal activities in tized and 15 identified as novel candidates for future surface wastewater at different stages of treatment. Environ. Toxicol. Chem. 33:2297–2307. water monitoring campaigns, e.g. ibuprofen, terbuthylazine, http://dx.doi.org/10.1002/etc.2676. Balaguer, P., François, F., Comunale, F., Fenet, H., Boussioux, A.M., Pons, M., Nicolas, J.C., triphenylphosphate and nonylphenoxyacetic acid, as well as Casellas, C., 1999. Reporter cell lines to study the estrogenic effects of xenoestrogens. macrolide antibiotics such as azithromycin. The EDP priority list Sci. Total Environ. 233:47–56. http://dx.doi.org/10.1016/S0048-9697(99)00178-3. showed that many WFD priority compounds are of lesser con- Bellet, V., Hernandez-Raquet, G., Dagnino, S., Seree, L., Pardon, P., Bancon-Montiny, C., Fenet, H., Creusot, N., Aït-Aïssa, S., Cavailles, V., Budzinski, H., Antignac, J.-P., cern, while several CECs, particularly from the class of pesticides Balaguer, P., 2012. Occurrence of androgens in sewage treatment plants influents is and pharmaceuticals may pose risk to aquatic life. The study in- associated with antagonist activities on other steroid receptors. Water Res. 46: troduced a novel simplified protocol for effect-based monitoring 1912–1922. http://dx.doi.org/10.1016/j.watres.2012.01.013. and provided a comprehensive insight into surface water status Bittner, M., Hilscherova, K., Giesy, J.P., 2009. In vitro assessment of AhR-mediated activities of TCDD in mixture with humic substances. Chemosphere 76:1505–1508. across Europe. http://dx.doi.org/10.1016/j.chemosphere.2009.06.042. Blake, L.S., Martinovic, D., Gray, L.E., Wilson, V.S., Regal, R.R., Villeneuve, D.L., Ankley, G.T., Conflict of interest 2010. Characterization of the androgen-sensitive mda-kb2 cell line for assessing complex environmental mixtures. Environ. Toxicol. Chem. 29:1367–1376. http://dx.doi. org/10.1002/etc.166. The authors declare no conflict of interest. All in vivo experiments Booij, P., Vethaak, D., Leonards, P.E.G., Sjollema, S.B., Kool, J., De Voogt, P., Lamoree, M.H., fi were performed according to ethical and legal regulations. 2014. Identi cation of photosynthesis inhibitors of pelagic marine algae using 96- well plate microfractionation for enhanced throughput in effect-directed analysis. Environ. Sci. Technol. 48:8003–8011. http://dx.doi.org/10.1021/es405428t. Acknowledgements Booij, P., Sjollema, S.B., van der Geest, H.G., Leonards, P.E.G., Lamoree, M.H., de Voogt, W.P., Admiraal, W., Laane, R.W.P.M., Vethaak, D., 2015. Toxic pressure of herbicides on microalgae in Dutch estuarine and coastal waters. J. Sea Res. 102:48–56. http://dx. The study was supported by the EDA-EMERGE project (FP7-PEOPLE- doi.org/10.1016/j.seares.2015.05.001. 2011-ITN, grant agreement number 290100). We thank Ondrej Sanka Brack, W., 2003. Effect-directed analysis: a promising tool for the identification of organic for creating maps, and Robert Hrich for his kind cooperation and for en- toxicants in complex mixtures? Anal. Bioanal. Chem. 377:397–407. http://dx.doi.org/ ř 10.1007/s00216-003-2139-z. abling us to access and collect samples at the WWTP Brno-Mod ice. We Brack, W., Govender, S., Schulze, T., Krauss, M., Hu, M., Muz, M., Hollender, J., Schirmer, K., are particularly grateful to Margit Petre, Emma Schymanski and Jörg Schollee, J., Hidasi, A., Slobodnik, J., Rabova, Z., Ait-Aissa, S., Sonavane, M., Carere, M., Ahlheim for logistical and technical help. The Eawag team gratefully Lamoree, M., Leonards, P., Tufi, S., Ouyang, X., Schriks, M., Thomas, K., Almeida, A.C. De, Froment, J., Hammers-Wirtz, M., Ahel, M., Koprivica, S., Hollert, H., Seiler, T.-B., acknowledges Birgit Beck for participation in sampling and analysis Paolo, C. Di, Tindall, A., Spirhanzlova, P., 2013. EDA-EMERGE: an FP7 initial training as well as the EcoImpact team for information about the sampling network to equip the next generation of young scientists with the skills to address locations. Jean Froment and Ana Caterina Almeida were supported the complexity of environmental contamination with emerging pollutants. Environ. – by additional funding from the Research Council of Norway Sci. Eur. 25:1 7. http://dx.doi.org/10.1186/2190-4715-25-18. Brack, W., Ait-Aissa, S., Burgess, R.M., Busch, W., Creusot, N., Di Paolo, C., Escher, B.I., Mark (221324). Zsolt Tarzai received a scholarship from German Federal Hewitt, L., Hilscherova, K., Hollender, J., Hollert, H., Jonker, W., Kool, J., Lamoree, M., Environmental Foundation. The study was carried out as a training Muschket, M., Neumann, S., Rostkowski, P., Ruttkies, C., Schollee, J., Schymanski, area of the EDA EMERGE project based on close collaboration of 13 E.L., Schulze, T., Seiler, T.B., Tindall, A.J., De Aragão Umbuzeiro, G., Vrana, B., Krauss, M., 2016. Effect-directed analysis supporting monitoring of aquatic environments - early research scientists (ESRs) and an experienced scientist (ER), an in-depth overview. Sci. Total Environ. 544:1073–1118. http://dx.doi.org/10. i.e. Carolina di Paolo (ESR1), Manoj Sonavane (ESR2), Petra 1016/j.scitotenv.2015.11.102. Spirhanzlova (ESR3), Ana Catarina Almeida (ESR4), Anita Hidasi Burgess, R.M., Ho, K.T., Brack, W., Lamoree, M., 2013. Effects-directed analysis (EDA) and fi fi toxicity identi cation evaluation (TIE): complementary but different approaches for (ESR5), Sara Tu (ESR6), Sanja Koprivica (ESR7), Zuzana Tousova diagnosing causes of environmental toxicity. Environ. Toxicol. Chem. 32: (ESR8), Meng Hu (ESR9), Melis Muz (ESR10), Jennifer Schollee 1935–1945. http://dx.doi.org/10.1002/etc.2299. Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868 1867

Chinathamby, K., Allinson, M., Shiraishi, F., Lopata, A.L., Nugegoda, D., Pettigrove, V., Galgani, F., Bocquene, G., 1991. Semi-automated colorimetric and enzymatic assays for Allinson, G., 2013. Screening for potential effects of endocrine-disrupting chemicals aquatic organisms using microplate readers. Water Res. 25:147–150. http://dx.doi. in peri-urban creeks and rivers in Melbourne, Australia using mosquitofish and re- org/10.1016/0043-1354(91)90023-J. combinant receptor-reporter gene assays. Environ. Sci. Pollut. Res. 20:1831–1841. Gehrmann, L., Bielak, H., Behr, M., Itzel, F., Lyko, S., Simon, A., Kunze, G., Dopp, E., Wagner, http://dx.doi.org/10.1007/s11356-012-1188-2. M., Tuerk, J., 2016. (Anti-)estrogenic and (anti-)androgenic effects in wastewater Cleuvers, M., 2003. Aquatic ecotoxicity of pharmaceuticals including the assessment of during advanced treatment: comparison of three in vitro bioassays. Environ. Sci. combination effects. Toxicol. Lett. 142:185–194. http://dx.doi.org/10.1016/S0378- Pollut. Res.:1–11 http://dx.doi.org/10.1007/s11356-016-7165-4. 4274(03)00068-7. Geissen, V., Mol, H., Klumpp, E., Umlauf, G., Nadal, M., van der Ploeg, M., van de Zee, Conley, J.M., Evans, N., Mash, H., Rosenblum, L., Schenck, K., Glassmeyer, S., Furlong, E.T., S.E.A.T.M., Ritsema, C.J., 2015. Emerging pollutants in the environment: a challenge Kolpin, D.W., Wilson, V.S., 2015. Comparison of in vitro estrogenic activity and estro- for water resource management. Int. Soil Water Conserv. Res. 3:57–65. http://dx. gen concentrations in source and treated waters from 25 U.S. drinking water treat- doi.org/10.1016/j.iswcr.2015.03.002. ment plants. Sci. Total Environ. 579:1610–1617. http://dx.doi.org/10.1016/j. Ginebreda, A., Kuzmanovic, M., Guasch, H., de Alda, M.L., López-Doval, J.C., Muñoz, I., scitotenv.2016.02.093. Ricart, M., Romaní, A.M., Sabater, S., Barceló, D., 2014. Assessment of multi-chemical Connon, R.E., Geist, J., Werner, I., 2012. Effect-based tools for monitoring and predicting pollution in aquatic ecosystems using toxic units: compound prioritization, mixture the ecotoxicological effects of chemicals in the aquatic environment. Sensors 12: characterization and relationships with biological descriptors. Sci. Total Environ. 12741–12771. http://dx.doi.org/10.3390/s120912741. 468–469:715–723. http://dx.doi.org/10.1016/j.scitotenv.2013.08.086. Creusot, N., Budzinski, H., Balaguer, P., Kinani, S., Porcher, J.-M., Aït-Aïssa, S., 2013. Effect- Gómez, M.J., Gómez-Ramos, M.M., Agüera, A., Mezcua, M., Herrera, S., Fernández-Alba, directed analysis of endocrine-disrupting compounds in multi-contaminated sedi- A.R., 2009. A new gas chromatography/mass spectrometry method for the simulta- ment: identification of novel ligands of estrogen and pregnane X receptors. Analytical neous analysis of target and non-target organic contaminants in waters. and Bioanalytical Chemistry 405, 2553–2566. J. Chromatogr. A 1216:4071–4082. http://dx.doi.org/10.1016/j.chroma.2009.02.085. Creusot, N., Aït-Aïssa, S., Tapie, N., Pardon, P., Brion, F., Sanchez, W., Thybaud, E., Porcher, Guijarro, C., Fuchs, K., Bohrn, U., Stütz, E., Wölfl, S., 2015. Simultaneous detection of mul- J.M., Budzinski, H., 2014. Identification of synthetic steroids in river water down- tiple bioactive pollutants using a multiparametric biochip for water quality monitor- stream from pharmaceutical manufacture discharges based on a bioanalytical ap- ing. Biosens. Bioelectron. 72:71–79. http://dx.doi.org/10.1016/j.bios.2015.04.092. proach and passive sampling. Environ. Sci. Technol. 48:3649–3657. http://dx.doi. Henry, N.D., Fair, P.A., 2013. Comparison of in vitro cytotoxicity, estrogenicity and anti- org/10.1021/es405313r. estrogenicity of triclosan, perfluorooctane sulfonate and perfluorooctanoic acid. Creusot, N., Brion, F., Piccini, B., Budzinski, H., Porcher, J.M., Aït-Aïssa, S., 2015. BFCOD ac- J. Appl. Toxicol. 33:265–272. http://dx.doi.org/10.1002/jat.1736. tivity in fish cell lines and zebrafish embryos and its modulation by chemical ligands Hilscherova, K., Kannan, K., Holoubek, I., Giesy, J.P., 2002. Characterization of estrogenic of human aryl hydrocarbon and nuclear receptors. Environ. Sci. Pollut. Res. 22: activity of riverine sediments from the Czech Republic. Arch. Environ. Contam. 16393–16404. http://dx.doi.org/10.1007/s11356-014-3882-8. Toxicol. 43:175–185. http://dx.doi.org/10.1007/s00244-002-1128-0. Creusot, N., Dévier, M.H., Budzinski, H., Aït-Aïssa, S., 2016. Evaluation of an extraction Inoue, D., Nakama, K., Matsui, H., Sei, K., Ike, M., 2009. Detection of agonistic activities method for a mixture of endocrine disrupters in sediment using chemical and in against five human nuclear receptors in river environments of Japan using a yeast vitro biological analyses. Environ. Sci. Pollut. Res. 23:10349–10360. http://dx.doi. two-hybrid assay. Bull. Environ. Contam. Toxicol. 82:399–404. http://dx.doi.org/10. org/10.1007/s11356-016-6062-1. 1007/s00128-008-9616-1. Dimpe, K.M., Nomngongo, P.N., 2016. Current sample preparation methodologies for Isobe, T., Sato, K., Joon-Woo, K., Tanabe, S., Suzuki, G., Nakayama, K., 2015. Determination analysis of emerging pollutants in different environmental matrices. TrAC Trends of natural and synthetic glucocorticoids in effluent of sewage treatment plants using Anal. Chem. 82:199–207. http://dx.doi.org/10.1016/j.trac.2016.05.023. ultrahigh performance liquid chromatography-tandem mass spectrometry. Environ. DIRECTIVE 2000/60/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 23 Sci. Pollut. Res. 22:14127–14135. http://dx.doi.org/10.1007/s11356-015-4626-0. October 2000 establishing a framework for Community action in the field of water Jálová, V., Jarošová, B., Bláha, L., Giesy, J.P., Ocelka, T., Grabic, R., Jurčíková, J., Vrana, B., policy, 2000. Off. J. Eur. Union L327, 1e77. Hilscherová, K., 2013. Estrogen-, androgen- and aryl hydrocarbon receptor medi- DIRECTIVE 2008/105/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 16 ated activities in passive and composite samples from municipal waste and sur- December 2008 on environmental quality standards in the field of water policy, facewaters.Environ.Int.59:372–383. http://dx.doi.org/10.1016/j.envint.2013. amending and subsequently repealing Council Directives 82/176/EEC, 83/513/EEC, 06.024. 84/156/EEC, 84/491/EEC, 2008. Off. J. Eur. Union (L 348/84). Janošek, J., Bittner, M., Hilscherová, K., Bláha, L., Giesy, J.P., Holoubek, I., 2007. AhR-medi- DIRECTIVE 2013/39/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 12 ated and antiestrogenic activity of humic substances. Chemosphere 67:1096–1101. August 2013 amending Directives 2000/60/EC and 2008/105/EC as regards priority http://dx.doi.org/10.1016/j.chemosphere.2006.11.045. substances in the field of water policy, 2013. Off. J. Eur. Union (L 226/15). Jarošová, B., Bláha, L., Giesy, J.P., Hilscherová, K., 2014a. What level of estrogenic activity Dizer, H., Wittekindt, E., Fischer, B., Hansen, P.D., 2002. The cytotoxic and genotoxic po- determined by in vitro assays in municipal waste waters can be considered as safe? tential of surface water and wastewater effluents as determined by bioluminescence, Environ. Int. 64:98–109. http://dx.doi.org/10.1016/j.envint.2013.12.009. umu-assays and selected biomarkers. Chemosphere 46:225–233. http://dx.doi.org/ Jarošová, B., Erseková, A., Hilscherová, K., Loos, R., Gawlik, B.M., Giesy, J.P., Bláha, L., 2014b. 10.1016/S0045-6535(01)00062-5. Europe-wide survey of estrogenicity in wastewater treatment plant effluents: the Dulio, V., Von der Ohe, P.C. (Eds.), 2013. NORMAN prioritisation framework for need for the effect-based monitoring. Environ. Sci. Pollut. Res. 10970–10982. http:// emerging substances, http://www.norman-network.net/sites/default/files/norman_ dx.doi.org/10.1007/s11356-014-3056-8. prioritisation_manual_15%20April2013_final_for_website.pdf, (accessed 10.16.15). Jedličková,B.,Hilscherová,K.,Bláha,L.,2011.Monitoring of Endocrine Disruptors in Ellman, G.L., Courtney, K.D., Andres, V., Featherstone, R.M., 1961. A new and rapid color- Czech Waste Waters and Their Recipients. ACTA Environ. Univ. COMENIANAE 19, imetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 7: 122–125. 88–95. http://dx.doi.org/10.1016/0006-2952(61)90145-9. Jomaa, B., Hermsen, S.A.B., Kessels, M.Y., Van Den Berg, J.H.J., Peijnenburg, A.A.C.M., Aarts, Emelogu, E.S., Pollard, P., Dymond, P., Robinson, C.D., Webster, L., McKenzie, C., Dobson, J., J.M.M.J.G., Piersma, A.H., Rietjens, I.M.C.M., 2014. Developmental toxicity of thyroid- Bresnan, E., Moffat, C.F., 2013. Occurrence and potential combined toxicity of dis- active compounds in a zebrafish embryotoxicity test. ALTEX 31:303–317. http://dx. solved organic contaminants in the forth estuary and firth of forth, Scotland assessed doi.org/10.14573/altex.1402011. using passive samplers and an algal toxicity test. Sci. Total Environ. 461–462: Jones, L., Ronan, J., McHugh, B., McGovern, E., Regan, F., 2015. Emerging priority sub- 230–239. http://dx.doi.org/10.1016/j.scitotenv.2013.05.011. stances in the aquatic environment: a role for passive sampling in supporting WFD Escher, B.I., Quayle, P., Muller, R., Schreiber, U., Mueller, J.F., 2006. Passive sampling of her- monitoring and compliance. Anal. Methods 7:7976–7984. http://dx.doi.org/10. bicides combined with effect analysis in algae using a novel high-throughput phyto- 1039/C5AY01059D. toxicity assay (maxi-imaging-PAM). J. Environ. Monit. 8:456–464. http://dx.doi.org/ Jugan, M.L., Oziol, L., Bimbot, M., Huteau, V., Tamisier-Karolak, S., Blondeau, J.P., Lévi, Y., 10.1039/b517512g. 2009. In vitro assessment of thyroid and estrogenic endocrine disruptors in wastewa- Escher, B.I., Allinson, M., Altenburger, R., Bain, P.A., Balaguer, P., Busch, W., Crago, J., ter treatment plants, rivers and drinking water supplies in the greater Paris area Denslow, N.D., Dopp, E., Hilscherova, K., Humpage, A.R., Kumar, A., Grimaldi, M., (France). Sci. Total Environ. 407:3579–3587. http://dx.doi.org/10.1016/j.scitotenv. Jayasinghe,B.S.,Jarosova,B.,Jia,A.,Makarov,S.,Maruya,K.A.,Medvedev,A., 2009.01.027. Mehinto, A.C., Mendez, J.E., Poulsen, P., Prochazka, E., Richard, J., Schifferli, A., Keiter, S., Rastall, A., Kosmehl, T., Wurm, K., Erdinger, L., Braunbeck, T., Hollert, H., 2006. Schlenk, D., Scholz, S., Shiraishi, F., Snyder, S., Su, G., Tang, J.Y.M., van der Burg, Ecotoxicological assessment of sediment, suspended matter and water samples in B., van der Linden, S.C., Werner, I., Westerheide, S.D., Wong, C.K.C., Yang, M., the upper Danube River. A pilot study in search for the causes for the decline of Yeung, B.H.Y., Zhang, X., Leusch, F.D.L., 2014. Benchmarking organic fish catches. Environ. Sci. Pollut. Res. Int. 13:308–319. http://dx.doi.org/10.1065/ micropollutants in wastewater, recycled water and drinking water with in vitro espr2006.04.300. bioassays. Environ. Sci. Technol. 48:1940–1956. http://dx.doi.org/10.1021/ Kim, M.S., Sabri, M.I., Miller, V.H., Kayton, R.J., Dixon, D.A., Spencer, P.S., 2001. 1,2- es403899t. Diacetylbenzene, the neurotoxic metabolite of a chromogenic aromatic solvent, in- Fang, T.Y., Praveena, S.M., deBurbure, C., Aris, A.Z., Ismail, S.N.S., Rasdi, I., 2016. Analytical duces proximal axonopathy. Toxicol. Appl. Pharmacol. 177:121–131. http://dx.doi. techniques for steroid estrogens in water samples - a review. Chemosphere 165: org/10.1006/taap.2001.9301\rS0041-008X(01)99301-9. 358–368. http://dx.doi.org/10.1016/j.chemosphere.2016.09.051. Kinani, S., Bouchonnet, S., Creusot, N., Bourcier, S., Balaguer, P., Porcher, J.-M., Aït-Aïssa, S., Ferrari, B., Paxéus, N., Lo Giudice, R., Pollio, A., Garric, J., 2003. Ecotoxicological impact of 2010. Bioanalytical characterisation of multiple endocrine- and dioxin-like activities pharmaceuticals found in treated wastewaters: study of carbamazepine, clofibric in sediments from reference and impacted small rivers. Environ. Pollut. 158:74–83. acid, and diclofenac. Ecotoxicol. Environ. Saf. 55:359–370. http://dx.doi.org/10. http://dx.doi.org/10.1016/j.envpol.2009.07.041. 1016/S0147-6513(02)00082-9. Koenig, M., Escher, B.I., Neale, P.A., Krauss, M., Hilscherova, K., Novak, J., Teodorovic, I., Fini, J.B., Le Mével, S., Turque, N., Palmier, K., Zalko, D., Cravedi, J.P., Demeneix, B.A., 2007. Schulze, T., Seidensticker, S., Kamal Hashmi, M.A., Ahlheim, J., Brack, W., 2017. Impact An in vivo multiwell-based fluorescent screen for monitoring vertebrate thyroid hor- of untreated wastewater on a major European river evaluated with a combination of mone disruption. Environ. Sci. Technol. 41:5908–5914. http://dx.doi.org/10.1021/ in vitro bioassays and chemical analysis. Environ. Pollut. 220:1220–1230. http://dx. es0704129. doi.org/10.1016/j.envpol.2016.11.011. 1868 Z. Tousova et al. / Science of the Total Environment 601–602 (2017) 1849–1868

Kugathas, S., Williams, R.J., Sumpter, J.P., 2012. Prediction of environmental concentra- Schulze, T., Ahel, M., Ahlheim, J., Aït-Aïssa, S., Brion, F., Di Paolo, C., Froment, J., Hidasi, A.O., tions of glucocorticoids: The River Thames, UK, as an example. Environ. Int. 40: Hollender, J., Hollert, H., Hu, M., Kloß, A., Koprivica, S., Krauss, M., Muz, M., Oswald, P., 15–23. http://dx.doi.org/10.1016/j.envint.2011.11.007. Petre, M., Schollée, J.E., Seiler, T.-B., Shao, Y., Slobodnik, J., Sonavane, M., Suter, M.J.-F., Kunz, P.Y., Simon, E., Creusot, N., Jayasinghe, S.B., Kienle, C., Maletz, S., Schonlau, C., Tollefsen, K.E., Tousova, Z., Walz, K.-H., Brack, W., 2017. Assessment of a novel device Hollert, H., Schifferli, A., Werner, I., Vermeirssen, E.L.M., 2017. Effect-based tools for for onsite integrative large-volume solid phase extraction of water samples to enable monitoring estrogenic mixtures: evaluation of five in vitro bioassays. Water Res. a comprehensive chemical and effect-based analysis. Sci. Total Environ. 581–582: 110:378–388. http://dx.doi.org/10.1016/j.watres.2016.10.062. 350–358. http://dx.doi.org/10.1016/j.scitotenv.2016.12.140. Kuzmanovic, Maja, Ginebreda, A., Petrovič, M., Barcelo, D., 2014. Risk assessment based Schymanski, E.L., Meringer, M., Brack, W., 2011. Automated strategies to identify com- prioritization of 200 organic micropollutants in 4 Iberian rivers. Sci. Total Environ. pounds on the basis of GC/EI-MS and calculated properties. Anal. Chem. 83: 504:289–299. http://dx.doi.org/10.1016/j.scitotenv.2014.06.056. 903–912. http://dx.doi.org/10.1021/ac102574h. Lemaire, G., Mnif, W., Mauvais, P., Balaguer, P., Rahmani, R., 2006. Activation of alpha- and Schymanski, E.L., Singer, H.P., Slobodnik, J., Ipolyi, I.M., Oswald, P., Krauss, M., Schulze, T., beta-estrogen receptors by persistent pesticides in reporter cell lines. Life Sci. 79: Haglund, P., Letzel, T., Grosse, S., Thomaidis, N.S., Bletsou, A., Zwiener, C., Ibáñez, M., 1160–1169. http://dx.doi.org/10.1016/j.lfs.2006.03.023. Portolés, T., de Boer, R., Reid, M.J., Onghena, M., Kunkel, U., Schulz, W., Guillon, A., Loos, R., Carvalho, R., António, D.C., Comero, S., Locoro, G., Tavazzi, S., Paracchini, B., Ghiani, Noyon, N., Leroy, G., Bados, P., Bogialli, S., Stipaničev, D., Rostkowski, P., Hollender, M., Lettieri, T., Blaha, L., Jarosova, B., Voorspoels, S., Servaes, K., Haglund, P., Fick, J., J., 2015. Non-target screening with high-resolution mass spectrometry: critical re- Lindberg, R.H., Schwesig, D., Gawlik, B.M., 2013. EU-wide monitoring survey on view using a collaborative trial on water analysis. Anal. Bioanal. Chem. 6237–6255. emerging polar organic contaminants in wastewater treatment plant effluents. http://dx.doi.org/10.1007/s00216-015-8681-7. Water Res. 47:6475–6487. http://dx.doi.org/10.1016/j.watres.2013.08.024. Sima, L., Amador, J., Silva, A.D.K., Miller, S.M., Morse, A.N., Pellegrin, M.-L., Rock, C., Wells, Ma, H., Zhang, H., Wang, L., Wang, J., Chen, J., 2014. Comprehensive screening and priority M.J.M., 2014. Emerging pollutants - part I: occurrence, fate and transport. Water En- ranking of volatile organic compounds in Daliao River, China. Environ. Monit. Assess. viron. Res. 86:1994–2035. http://dx.doi.org/10.2175/106143014X14031280668731. 186:2813–2821. http://dx.doi.org/10.1007/s10661-013-3582-8. Slobodnik, J., Mrafkova, L., Carere, M., Ferrara, F., Pennelli, B., Schüürmann, G., von der Macikova, P., Groh, K.J., Ammann, A.A., Schirmer, K., Suter, M.J.F., 2014. Endocrine Ohe, P.C., 2012. Identification of river basin specific pollutants and derivation of envi- disrupting compounds affecting corticosteroid signaling pathways in Czech and ronmental quality standards: a case study in the Slovak Republic. TrAC Trends Anal. Swiss waters: potential impact on fish. Environ. Sci. Technol. 48:12902–12911. Chem. 41:133–145. http://dx.doi.org/10.1016/j.trac.2012.08.008. http://dx.doi.org/10.1021/es502711c. Smital, T., Terzić, S., Lončar, J., Senta, I., Žaja, R., Popović, M., Mikac, I., Tollefsen, K.-E., Maier, D., Blaha, L., Giesy, J.P., Henneberg, A., Köhler, H.-R., Kuch, B., Osterauer, R., Peschke, Thomas, K.V., Ahel, M., 2013. Prioritisation of organic contaminants in a river basin K., Richter, D., Scheurer, M., Triebskorn, R., 2014. Biological plausibility as a tool to as- using chemical analyses and bioassays. Environ. Sci. Pollut. Res. 20:1384–1395. sociate analytical data for micropollutants and effect potentials in wastewater, sur- http://dx.doi.org/10.1007/s11356-012-1059-x. face water, and sediments with effects in fishes. Water Res. 2. http://dx.doi.org/10. Tang, J.Y.M., Escher, B.I., 2014. Realistic environmental mixtures of micropollutants in sur- 1016/j.watres.2014.08.050. face, drinking, and recycled water: herbicides dominate the mixture toxicity toward McArdell, C.S., Molnar, E., Suter, M.J.F., Giger, W., 2003. Occurrence and fate of macrolide algae. Environ. Toxicol. Chem. 33:1427–1436. http://dx.doi.org/10.1002/etc.2580. antibiotics in wastewater treatment plants and in the Glatt Valley Watershed, Swit- Terzic, S., Ahel, M., 2011. Nontarget analysis of polar contaminants in freshwater sedi- zerland. Environ. Sci. Technol. 37:5479–5486. http://dx.doi.org/10.1021/es034368i. ments influenced by pharmaceutical industry using ultra-high-pressure liquid chro- Miège, C., Gabet, V., Coquery, M., Karolak, S., Jugan, M.L., Oziol, L., Levi, Y., Chevreuil, M., matography-quadrupole time-of-flight mass spectrometry. Environ. Pollut. 159: 2009. Evaluation of estrogenic disrupting potency in aquatic environments and 557–566. http://dx.doi.org/10.1016/j.envpol.2010.10.009. urban wastewaters by combining chemical and biological analysis. TrAC Trends Thomaidi, V.S., Stasinakis, A.S., Borova, V.L., Thomaidis, N.S., 2015. Is there a risk for the Anal. Chem. 28:186–195. http://dx.doi.org/10.1016/j.trac.2008.11.007. aquatic environment due to the existence of emerging organic contaminants in treat- Molina-Molina, J.M., Escande, A., Pillon, A., Gomez, E., Pakdel, F., Cavailles, V., Olea, N., Ait- ed domestic wastewater? Greece as a case-study. J. Hazard. Mater. 283:740–747. Aissa, S., Balaguer, P., 2008. Profiling of benzophenone derivatives using fish and http://dx.doi.org/10.1016/j.jhazmat.2014.10.023. human estrogen receptor-specific in vitro bioassays. Toxicol. Appl. Pharmacol. 232 Thomas, K.V., Hurst, M.R., Matthiessen, P., McHugh, M., Smith, A., Waldock, M.J., 2002. An (395). http://dx.doi.org/10.1016/j.taap.2008.07.017. assessment of in vitro androgenic activity and the identification of environmental an- Neale, P.A., Escher, B.I., 2014. Does co-extracted dissolved organic carbon cause artefacts drogens in United Kingdom estuaries. Environ. Toxicol. Chem. 21:1456–1461. http:// in cell-based bioassays? Chemosphere 108:281–288. http://dx.doi.org/10.1016/j. dx.doi.org/10.1002/etc.5620210717. chemosphere.2014.01.053. Tollefsen, K.E., Nizzetto, L., Huggett, D.B., 2012. Presence, fate and effects of the intense Neale, P.A., Ait-Aissa, S., Brack, W., Creusot, N., Denison, M.S., Deutschmann, B., sweetener sucralose in the aquatic environment. Sci. Total Environ. 438:510–516. Hilscherova, K., Hollert, H., Krauss, M., Novak, J., Schulze, T., Seiler, T.B., Serra, H., http://dx.doi.org/10.1016/j.scitotenv.2012.08.060. Shao, Y., Escher, B.I., 2015. Linking in vitro effects and detected organic US EPA, Ecological Structure Activity Relationships (ECOSAR) Predictive Model, v.1.11, micropollutants in surface water using mixture-toxicity modeling. Environ. Sci. 2012. United States Environmental Protection Agency. Technol. 49:14614–14624. http://dx.doi.org/10.1021/acs.est.5b04083. Vincze, K., Gehring, M., Braunbeck, T., 2014. (Eco)toxicological effects of 2,4,7,9- OECD, 2011. Test no. 201: freshwater alga and cyanobacteria, growth inhibition test, tetramethyl-5-decyne-4,7-diol (TMDD) in zebrafish (Danio rerio) and permanent http://dx.doi.org/10.1787/9789264069923-en. fish cell cultures. Environ. Sci. Pollut. Res. 21:8233–8241. http://dx.doi.org/10.1007/ OECD, 2013. Test no. 236: fish embryo acute toxicity (FET) test, http://dx.doi.org/10.1787/ s11356-014-2806-y. 9789264203709-en. Viviani-Nauer, A., Hoffmann-Boller, P., Gafner, J., 1997. In vitro detection of Folpet and its Oh, S.M., Park, K., Chung, K.H., 2006. Combination of in vitro bioassays encompassing dif- degradation to Phthalimide in aqueous solutions and in yeast suspensions. Am. ferent mechanisms to determine the endocrine-disrupting effects of river water. Sci. J. Enol. Vitic. 48, 63–66. Total Environ. 354:252–264. http://dx.doi.org/10.1016/j.scitotenv.2005.01.041. von der Ohe, P.C., Dulio, V., Slobodnik, J., De Deckere, E., Kühne, R., Ebert, R.U., Ginebreda, Osman, A.G.M., AbouelFadl, K.Y., Krüger, A., Kloas, W., 2015. Screening of multiple hor- A., De Cooman, W., Schüürmann, G., Brack, W., 2011. A new risk assessment approach monal activities in water and sediment from the river Nile, Egypt, using in vitro bio- for the prioritization of 500 classical and emerging organic microcontaminants as po- assay and gonadal histology. Environ. Monit. Assess. 187:317. http://dx.doi.org/10. tential river basin specific pollutants under the European water framework directive. 1007/s10661-015-4553-z. Sci. Total Environ. 409:2064–2077. http://dx.doi.org/10.1016/j.scitotenv.2011.01.054. Petrie, B., Barden, R., Kasprzyk-Hordern, B., 2014. A review on emerging contaminants in Vrana, B., Foppe, S., Prokes, R., Loos, R., 2011. Norman Interlaboratory Study (ILS) on Pas- wastewaters and the environment: current knowledge, understudied areas and rec- sive Sampling of Emerging Pollutants [WWW Document]. URL. (accessed 12.1.12). ommendations for future monitoring. Water Res. 72:3–27. http://dx.doi.org/10.1016/ http://www.norman-network.net/?q=node/101. j.watres.2014.08.053. Wilson, V.S., Bobseine, K., Lambright, C.R., Gray, L.E., 2002. A novel cell line, MDA-kb2, that Ribeiro, A.R., Maia, A., Santos, M., Tiritan, M.E., Ribeiro, C.M.R., 2016. Occurrence of natural stably expresses an androgen- and glucocorticoid-responsive reporter for the detec- contaminants of emerging concern in the Douro River Estuary, Portugal. Arch. Envi- tion of hormone receptor agonists and antagonists. Toxicol. Sci. 66:69–81. http://dx. ron. Contam. Toxicol. 70:361–371. http://dx.doi.org/10.1007/s00244-015-0212-1. doi.org/10.1093/toxsci/66.1.69. Rojickova, R., Dvorakova, D., 1998. The use of miniaturized algal bioassays in comparison Witters, H., Freyberger, A., Smits, K., Vangenechten, C., Lofink, W., Weimer, M., Bremer, S., to the standard flask assay. Environ. Toxicol. Water Qual. 13:235–241. http://dx.doi. Ahr, P.H.J., Berckmans, P., 2010. The assessment of estrogenic or anti-estrogenic activ- org/10.1007/s11064-007-9392-z. ity of chemicals by the human stably transfected estrogen sensitive MELN cell line: Sabri, M.I., Hashemi, S.B., Lasarev, M.R., Spencer, P.S., 2007. Axonopathy-inducing 1,2- results of test performance and transferability. Reprod. Toxicol. 30:60–72. http://dx. diacetylbenzene forms adducts with motor and cytoskeletal proteins required for ax- doi.org/10.1016/j.reprotox.2010.02.008. onal transport. Neurochem. Res. 32:2152–2159. http://dx.doi.org/10.1007/s11064- van Woudenberg, A.B., Snel, C., Rijkmans, E., de Groot, D., Bouma, M., Hermsen, S., 007-9392-z. Piersma, A., Menke, A., Wolterbeek, A., 2014. Zebrafish embryotoxicity test for devel- Salvati, M.E., Balog, A., Shan, W., Wei, D.D., Pickering, D., Attar, R.M., Geng, J., Rizzo, C.a., opmental (neuro) toxicity: demo case of an integrated screening approach system Gottardis, M.M., Weinmann, R., Krystek, S.R., Sack, J., An, Y., Kish, K., 2005. Structure using anti-epileptic drugs. Reprod. Toxicol. 49:101–116. http://dx.doi.org/10.1016/j. based approach to the design of bicyclic-1H-isoindole-1,3(2H)-dione based androgen reprotox.2014.07.082. receptor antagonists. Bioorg. Med. Chem. Lett. 15:271–276. http://dx.doi.org/10. Zhao, J.-L., Ying, G.-G., Yang, B., Liu, S., Zhou, L.-J., Chen, Z.-F., Lai, H.-J., 2011. Screening of 1016/j.bmcl.2004.10.085. multiple hormonal activities in surface water and sediment from the Pearl River sys- Schriks, M., Van Leerdam, J.A., Van Der Linden, S.C., Van Der Burg, B., Van Wezel, A.P., De tem, South China, using effect-directed in vitro bioassays. Environ. Toxicol. Chem. 30: Voogt, P., 2010. High-resolution mass spectrometric identification and quantification 2208–2215. http://dx.doi.org/10.1002/etc.625. of glucocorticoid compounds in various wastewaters in the Netherlands. Environ. Sci. Technol. 44:4766–4774. http://dx.doi.org/10.1021/es100013x.

ANNEX II Journal of Hazardous Materials 358 (2018) 494–502

Contents lists available at ScienceDirect

Journal of Hazardous Materials

journal homepage: www.elsevier.com/locate/jhazmat

Identiﬁcation of algal growth inhibitors in treated waste water using eﬀect- directed analysis based on non-target screening techniques T

Zuzana Tousovaa,b, Jean Fromentc,d, Peter Oswalda, Jaroslav Slobodníka, Klara Hilscherovab, ⁎ Kevin V. Thomasd,e, Knut Erik Tollefsend, Malcolm Reidd, Katherine Langfordd, Ludek Blahab, a Environmental Institute (EI), Okruzna 784/42, 972 41 Kos, Slovak Republic b Masaryk University, Faculty of Science, RECETOX, Kamenice 753/5, 625 00 Brno, Czech Republic c Helmholtz Centre for Environmental Research (UFZ), Permoserstraße 15, 04318 Leipzig, Germany d Norwegian Institute for Water Research (NIVA), Gaustadallèen 21, NO-0349 OSLO, Norway e Queensland Alliance for Environmental Health Sciences (QAEHS), University of Queensland, 39 Kessels Road, Coopers Plains, Queensland, 4108 Australia

GRAPHICAL ABSTRACT

ARTICLE INFO ABSTRACT

Keywords: Growth inhibition of freshwater microalga Pseudokirchneriella subcapitata caused by a waste water treatment Contaminants of emerging concern plant (WWTP) effluent extract was investigated using an effect directed analysis (EDA) approach. The objective Effect directed analysis was to identify compounds responsible for the toxicity by combining state-of-the-art sampling, bioanalytical, Fractionation fractionation and non-target screening techniques. Three fractionation steps of the whole extract were performed Non-target screening and bioactive fractions were analysed with GC (xGC)-MS and LC-HRMS. In total, 383 compounds were tenta- Toxicity drivers tively identified, and their toxicity was characterized using US EPA Ecotox database, open scientific literature or modelled by ECOSAR. Among the top-ranking drivers of toxicity were pesticides and their transformation products, pharmaceuticals (barbiturate derivatives and macrolide antibiotics e.g. azithromycin), industrial

Abbreviations: CEC, contaminants of emerging concern; DMSO, dimethylsulfoxide; EC50/20, concentration which causes 50% (20%) growth rate inhibition of algae compared to solvent control; EDA, eﬀect directed analysis; EtOAc, ethylacetate; GC(xGC)-MS, (two dimensional) gas chromatography coupled to mass spectrometry; LC-HRMS, liquid chromatography coupled to high resolution mass spectrometry; LVSPE, large volume solid phase extraction; MeOH, methanol; MTV, minimum toxicity value; PCPs, personal care products; PPP, plant protection product; REF, relative enrichment factor; RP-HPLC, reverse phase - high performance liquid chromatography; (RP)-SPE, (reverse phase) solid phase extraction; WWTP, waste water treatment plant ⁎ Corresponding author. E-mail address: [email protected] (L. Blaha). https://doi.org/10.1016/j.jhazmat.2018.05.031 Received 19 January 2018; Received in revised form 11 May 2018; Accepted 15 May 2018 Available online 17 May 2018 0304-3894/ © 2018 Elsevier B.V. All rights reserved. Z. Tousova et al. Journal of Hazardous Materials 358 (2018) 494–502

compounds or caffeine and its metabolites. Several of the top-ranking pesticides are no longer registered for use in plant protection products or biocides in the Czech Republic (e.g. prometryn, atrazine, acetochlor, resmethrin) and some are approved only for use in biocides (e.g. terbutryn, carbendazim, phenothrin), which indicates that their non-agricultural input into aquatic environment via WWTPs should be carefully considered. The study demonstrated a functional strategy of combining biotesting, fractionation and non-target screening techniques in the EDA study focused on the identification of algal growth inhibitors in WWTP effluent.

1. Introduction test toxicity of compounds or their mixtures to microalgae have therefore been developed and standardized [6,7]. Microalgae as primary producers are a key functional group in Despite great progress, WWTP effluents entering surface waters still aquatic food webs and possible adverse effects on algal communities present a major source of toxic pollutants [8,9]. Studies combining may lead to changes at multiple trophic levels and ultimately impair biological and chemical analytical approaches reported that between ecosystem health [1]. Toxic effects of anthropogenic contaminants to one half to two thirds of phytotoxic effects in surface or waste waters phytoplankton have been previously reported [2,3,4,5], and methods to could be explained by herbicides and their metabolites [3,10].

Fig. 1. Overview of the fractionation strategy, biotesting and chemical analyses workflow used for the identification of phytotoxic compounds in the WWTP effluent. The fractions identified as phytotoxic (black boxes) were further fractionated and analyzed.

495 Z. Tousova et al. Journal of Hazardous Materials 358 (2018) 494–502

Nevertheless, contaminants of emerging concern (CEC) such as phar- The first fractionation was achieved on site by the LVSPE50 sampling maceuticals, perfluorinated compounds and UV filters in surface waters device followed by 5 elution steps in the laboratory (5 fractions; F1-F5). also have the potential to cause toxic effects to algae [11,12,13,14,15]. The second fractionation of F1, which showed the most pronounced ® Algal toxicity testing has been well established in ecotoxicology, toxicity to algae, was performed with RP-SPE (C18 – Sep-Pak ,1g, and large amount of toxicity data for microalgae is available in open Waters, Ireland) followed by stepwise elution with water and MeOH scientific literature and ecotoxicological databases like US EPA Ecotox mixtures (9 fractions; F1.1 - F1.9; Fig. 1). The third fractionation of database [16]. The availability of experimental datasets also facilitated pooled F1.2, F1.3 and F1.4, identified as phytotoxic by the bioassay, the development of predictive QSAR tools such as ECOSAR [17], which was conducted by RP-HPLC resulting in 31 fractions (F1.4.1-F1.4.31). enabled modelling of toxicity without the need for experimental testing In each fractionation step, the volume of individual fractions was re- [18]. duced under mild stream of nitrogen, divided into aliquots for chemical To investigate the role of individual compounds in the overall analyses and bioassays and the solvents were exchanged for MeOH or phytotoxicity of highly complex environmental samples, effect directed acetonitrile and DMSO, respectively. More details on the fractionation analysis (EDA), a technique combining biotesting, fractionation and method are available in SM1-Section 2. chemical analysis of bioactive fractions, can be applied [19,20]. EDA is based on the assumption that only a few chemicals drive a biological effect and these chemicals can be isolated in discrete fractionations 2.3. Bioassay with one-dimensional HPLC is the most frequently used in environmental EDA studies [21]. Multidimensional techniques, developed in Algal growth inhibition assay with unicellular green alga other fields, have also found their use in environmental analysis dealing Pseudokirchneriella subcapitata (syn. Raphidocelis subcapitata) was car- fi with highly complex mixtures, and for example two-dimensional gas ried out according to a modi ed OECD method 201 (2011) as described chromatography coupled to mass spectrometry (GCxGC-MS) has been by Rojickova and Dvorakova [32]. Sample extracts in DMSO (0.5% v/v) successfully used [22]; [23,24,25]. were tested at a relative enrichment factor (REF) of 100 or 200, i.e. the fi ffl Conventional multi-residue target analysis of complex samples like nal testing solution contained the e uent (or its fraction) con- fl WWTP effluents provides only a limited picture of a mixture’s complete centrated 100 or 200 times. Absorbance at 680 nm and uorescence chemical composition. Many compounds are co-extracted in the ex- (485/685) were measured as surrogate parameters of biomass every traction process, they remain unnoticed during the analysis, but they 24 h to determine algal growth rate. All fractions were initially tested at fi may contribute significantly to the observed biological effects. Non- a xed concentration (REF = 100 or 200) and phytotoxic fractions were target screening techniques using GC and LC–MS have therefore been thereafter diluted to obtain full concentration-response curves (n = 2) ff increasingly applied in environmental analysis [26,27,28,29]. A recent to derive e ect concentrations (ECs) expressed with REF values as study by Schymanski et al. [28] showed that while the non-target proposed by Escher et al. [2]. More details on the bioassay are listed in screening analytical techniques were substantially harmonized among SM1-Section 3. laboratories, the data-processing and evaluation part still presents a challenge. Currently, powerful chromatographic deconvolution and 2.4. Chemical analyses structure elucidation software, in silico prediction tools for fragmentation, retention time and ionization, as well as MS and MS/MS libraries 2.4.1. GC–MS non-target screening of environmental contaminants are being developed rapidly and enable GC–MS non-target screening was performed according to a method tentative identification of unknowns in a reasonable time frame with a described in Slobodnik et al. [33]. In brief, an Agilent 7890 A GC reasonable confidence [30,28]. (Agilent Technologies, Wilmington, DE, USA) connected with quadru- The objectives of the present study were to identify compounds pole mass spectrometer (qMSD - 5975C) were used to gain mass responsible for algal growth inhibition in a WWTP effluent extract using spectra, followed by identification of candidate structures using the a complex EDA. The study applied an innovative approach combining a NIST database with AMDIS [34] combined in most cases with Mass novel sampling tool, RP-SPE and RP-HPLC fractionation, miniaturized Frontier TM 4.0 [35].GC–MS non-target screening was performed for algal bioassay, GC–MS (complemented with GCxGC-MS) and LC-HRMS fractions F1, F1.4.7 and F1.4.8 (Fig. 1). Detailed description of the non-target screening, exploration of ecotoxicity data and ECOSAR method is available in SM-Section 4. toxicity predictions. The presented work aimed to demonstrate feasi- bility and potential of an EDA case study based on successful use of state-of-the-art sampling, bioanalytical, fractionation and non-target 2.4.2. GCxGC-MS screening tools. GCxGC-MS non-target screening was performed on Agilent 7890 A GC (Agilent Technologies, Wilmington, DE, USA) equipped with dif- 2. Experimental ferential flow modulator (DFM) and quadrupole mass spectrometric detector (qMSD - 5975C). Experimental data were processed with GC 2.1. Sampling Image software (Lincoln, Nebraska, USA) and candidate compounds were identified by use of the same workflow as in case of one-dimen- ffl ř The sample of e uent from the WWTP in Brno-Mod ice, Czech sional GC. More details on the two-dimensional setup can be found in Republic (49.12447 N, 16.62697E) was collected on August 15th, 2014 SM1-Section 5. using large volume solid phase extraction device – LVSPE50 (UFZ, Leipzig, Germany; and Maxx GmbH company, Rangendingen, Germany) as described in Schulze et al. [43,62]. The device enabled 2.4.3. LC-HRMS non-target screening extraction of 50 litres of water sample and primary on-site fractionation LC-HRMS non-target screening was carried out on Acquity UPLC based on the affinity of three sorbents to distinct compound groups (for System coupled to Xevo™ G2-S QTOF mass spectrometer (Waters, more details on the sampling see SM1-Section 1). Milford, MA, USA). Suspect screening was based on comparison of active and non-active fractions by use of principal component analysis 2.2. Fractionation (PCA) and search for matches of retention times and accurate masses of candidate compounds in the STOFF-IDENT database. Detailed descrip- The sample was subjected to an EDA procedure with three fractio- tion of the LC-HRMS non-target screening workflow is given in SM1- nation steps, for which the workflow overview is illustrated in Fig. 1. Section 6.

496 Z. Tousova et al. Journal of Hazardous Materials 358 (2018) 494–502

2.5. Linking phytotoxic effects to compounds detected by chemical analyses Table 1 Effective concentrations of individual fractions of the WWTP effluent extract in To link the observed phytotoxic effects and the compounds detected the algal growth inhibition assay expressed as relative enrichment factors by chemical analyses in the whole HR-X neutral eluate (F1) and active (REF).

HPLC fractions (F1.4.7, F1.4.8 and F1.4.31), i.e. in total 383 chemical Fraction Fraction description EC50 [REF] EC20 [REF] individuals, several experimental datasets and ECOSAR modelling were used (Fig. 2). Firstly, the US-EPA Ecotox database [16] was searched for F1 neutral eluate of HR-X sorbent 18.6 14.3 F2 acidic eluate of HR-X sorbent > 100 > 100 phytotoxicity data by the CAS numbers of the detected compounds. F3 basic eluate of HR-X sorbent 56.2 27.6 ff E ect concentrations (ECx,ICx and LOEC) from laboratory studies with F4 basic eluate of HR-XAW sorbent > 100 > 100 algae and cyanobacteria were extracted from the database. Secondly, F5 acidic eluate of HR-XCW sorbent > 100 > 100 experimental toxicity data for photosynthesizing organisms, i.e. algae F1.2-4 upto 80%MeOH 32.9 19.8 and cyanobacteria, were derived from open literature whereas an in- F1.4.7 upto 62.5%MeOH 82.3 78 F1.4.8 upto 63.6%MeOH 84.3 78.7 ternal database with more than 1700 experimental data points was F1.4.31 backflush of the HPLC column 105.5 54.3 developed as the primary source of effect data (SM2 - Table S3). From these, the lowest effective experimental concentration value for each compound was selected as the conservative estimation of the highest (EC50 = 56.2 REF). Primary fractions F2, F4 and F5 showed no effects toxicity (maximum toxicity potential - MTPexp). Thirdly, green algae up to 100 REF. Consequently, F1 was subject to further fractionation by 96 h EC50s were predicted by the ECOSAR program, v1.11 [17] and RP-SPE creating 9 fractions, out of which F1.3 was the most active used to fill the gaps in experimental data. The toxicity of the ECOSAR (Fig. 3B). F1.3 and its neighbouring fractions F1.2, and F1.4 were fur- class with the lowest EC50 value was selected to represent the predicted ther assessed, pooled (EC50 = 32.9) and further fractionated on RP- toxicity of each compound (maximum toxicity potential - MTPpred). The HPLC. Three out of the total 31 resulting fractions (F1.4.7, F1.4.8 and water solubility of detected compounds was predicted by the US-EPA F1.4.31) were toxic to algae (Fig. 3C). Each fractionation step lead to a ECOSAR program suite [36] using WS Kowwin (v1.43). The detected decrease in toxicity which could be explained by losses of compounds compounds were ranked in order of their maximum toxicity potential. during fractionation and/or possibly by combined effects in complex Experimental data were always given priority over the predicted ones mixtures (Table 1 - compare F1 vs F1.4.7, F1.4.8 and F1.4.31), how- and the compounds with the lowest effective concentrations exceeding ever, the effects of recombined fractions to address the mixture effects their water solubility were excluded from the ranking (SM2-Table S4). were not examined in our study.. In the GC(xGC)-MS non-target screening of primary fraction F1, 41 3. Results compounds were detected, tentatively identified and their concentrations estimated (SM-Table S4). These compounds comprised of orga- Algal growth inhibition was observed in two primary fractions of nophosphates, pharmaceuticals, personal care products, pesticides and the WWTP extract, namely F1 and F3 (Table 1, Fig. 3A). The effect was industrial pollutants. The highest concentrations were estimated for most pronounced in F1 (EC50 = 18.6 REF) and moderate in F3

Fig. 2. Overview of the ecotoxicity data retrieval and processing using US EPA Ecotox database, open peer-reviewed literature and ECOSAR predictions. ("SM-S2″ in parentheses - "Supplementary material – Table S2″).

497 Z. Tousova et al. Journal of Hazardous Materials 358 (2018) 494–502

Fig. 3. Representative results of algal growth rate inhibition bioassay (72 h) testing the WWTP extract after the first fractionation by LVSPE50 (fractions F1-F5 in a five-point dilution series) (A), the second fractionation by RP-SPE (B) and the third fractionation by RP-HPLC (C). Tested concentrations are given as relative enrichment factor (REF). Error bars denote standard deviation between 5 replicates in a single experiment. two plasticizers - benzenesulfonamide (1.7 μg/L) and n-butylbenzene- products, followed by industrial chemicals, other pesticides, caffeine sulfonamide (1.2 μg/L); and a sunscreen agent - octinoxate (1.5 μg/L). and its metabolites, barbiturate drugs and other pharmaceuticals. For Two-dimensional GC separation enabled identification of several com- full list of detected compounds, their MTPs, physico-chemical para- pounds with co-eluting peaks in the one-dimensional setup, e.g. ticlo- meters and ranking see SM2- Table S4. pidine and fluconazole, promethryn and clorophene, tramadol and desmethyltramadol, venlafaxine and norvenlafaxine. In the GC–MS non-target screening of the phytotoxic tertiary fractions (F1.4.7 and 4. Discussion F1.4.8), 22 compounds were detected and tentatively identified. In the LC-HRMS non-target screening of three tertiary fractions Toxicity of treated waste waters to algae was described in earlier active in the bioassay (F1.4.7, F1.4.8, F1.4.31), the total of 345 com- studies [40,41,42], and it remains a concern regarding the chemical pounds, including mostly pharmaceuticals, industrial chemicals and by- and ecological status of the recipient surface waters. Several studies products, biocides, herbicides and their transformation products, were reported that the phytotoxicity of waste and surface waters was detected and tentatively identified (SM2 – Table S2). There was an dominated by herbicides, that large portion of the effects could be ex- overlap between GC–MS and LC-HRMS in case of 4 compounds, namely plained by target analysis (> 50%) and that photosystem II inhibition fluconazole, N-butylbenzensulfonamide, triethylphosphate and tribu- was the driving mode of action [43,2,44,10]. tylphosphate (SM2- Table S4). A novel approach in the present study provided a new insight into The collection of experimental and predicted ecotoxicity data for the complexity of the WWTP effluent sampleand enabled to identify the the detected compounds (summarized in Fig. 2) resulted in the total of likely toxicity drivers. GC(xGC)-MS non-target screening proved to be a 377 Maximum toxicity potential (MTP) values. Experimental values useful tool, despite the fact that GC techniques are considered less

(MTPexp) were available for 68 compounds (17.8%). Among these, the suitable than LC for screening of polar and semi-polar water con- US-EPA Ecotox database covered 39 compounds and the internal da- taminants. As expected, there was only a little overlap between non- tabase based on open literature search (SM- Table S3) covered 53 target GC(xGC)-MS and LC-HRMS analyses (4 out of total 383 com- compounds (overlap of 24 compounds between the two datasets). pounds) and therefore GC–MS non target screening served as an im-

ECOSAR predicted maximum toxicity potential (MTVpred) values were portant complement to the LC-HRMS. Large volume injection used in used for the remaining 309 compounds, while no MTP could be derived GC–MS non-target screening, as earlier described in Gómez et al. [26], for 6 chemicals. Subsequently, detected compounds were ranked ac- was a good strategy to search for a number of compounds in a single cording to their MTPs to identify those with the greatest potential to analytical run. GC(xGC)-MS allowed detection (and also concentration contribute to the observed toxicity. This ranking was based solely on estimation) of personal care products (PCPs)like galaxolides, N,N-Die- the qualitative information from the non-target screening (list of de- thyl-meta-toluamide(DEET) or octinoxate, similarly to the results of tected and tentatively identified compounds), since the quantitative or Gómez et al. [45] and Herrera Lõpez et al. [46]. Within the top 25 semi-quantitative estimations of concentrations could not be retrieved compounds identified as likely toxicity drivers, six were detected ex- in case of the LC-HRMS non-target screening workflow. The total of 77 clusively by GC(xGC)-MS (Table 2). Identification by LC–MS was based compounds were excluded from the ranking because their MTPs ex- on two different parameters (retention time and accurate mass), ceeded the predicted water solubility. As a result of this ranking, the top achieving a level 3 identification according to Schymanski et al. [47]. 25 compounds, identified as the most likely toxicity drivers, are listed Level 2 of identification was obtained by the GC (and GCxGC)-MS in Table 2. This list is dominated by herbicides and their transformation measurements. To eventually achieve level 1 and unequivocal identification, the pure compound has to be injected in the analytical

498 .Tuoae al. et Tousova Z.

Table 2 List of the top 25 compounds detected and tentatively identiﬁed by GC- and LC–MS non-target screening of phytotoxic fractions of WWTP eﬄuent extract (F1, F1.4.7, F1.4.8 and F1.4.31) ranked according to their estimated maximum toxicity potential (for full table with all detected compounds and more details see SM2- Table S4).

Rank Method CAS# Name Use Approval status under EU Maximum toxicity Reference to the maxim Lowest 96 h-EC50 for Chemical class of lowest and Czech pesticide potential [μg/L] toxicity potential used green algae in EC50 in ECOSAR regulationa for ranking ECOSAR [μg/L]

1 GC-MS 886-50-0 Terbutryn selective triazine herbicide, photosystem approved in biocides 0.02 US EPA-ECOTOX 33 Triazines, Aromatic II inhibitor 2 LC-HRMS 5915-41-3 Terbuthylazine chlorotriazine herbicide, photosystem II approved in biocides and 0.40 [9] 5 Triazines, Aromatic inhibitor PPPs 3 GC-MS 7287-19-6 Prometryn selective triazine herbicide, photosystem not approved 0.41 US EPA-ECOTOX 10 Triazines, Aromatic II inhibitor 4 LC-HRMS 95-38-5 2-(2-heptadec-8-enyl-2- industrial formulation of lubricant 0.84 ECOSAR 0.84 Aliphatic Amines imidazolin-1-yl)ethanol additives, lubricants, functional fluids and greases 5 LC-HRMS 34256-82-1 Acetochlor chloroacetanilide herbicide, elongase not approved 1.40 [37] 3 Haloacetamides inhibitor 6 LC-HRMS 1007-28-9 6-deisopropylatrazine chlorinated degradation product parent compuund not 2 US EPA-ECOTOX 281 Triazines, Aromatic herbicide atrazine approved 7 LC-HRMS 61213-25-0 Flurochloridone selective herbicide, biosynthesis of approved in PPPs 3 ECOSAR 3 Haloacetamides carotenoids inhibitor 8 LC-HRMS 10605-21-7 Carbendazim broad-spectrum benzimidazole approved in biocides 3.3 US EPA-ECOTOX 302 Carbamate Esters fungicide 9 GC-MS 58-08-2 Caffeine central nervous system (CNS) stimulant 5 US EPA-ECOTOX 15 Carbonyl Ureas of the methylxanthine class of psychoactive drugs 499 10 LC-HRMS 3567-62-2 DCPMU metabolite of herbicide diuron parent compound 5 US EPA-ECOTOX 329 Substituted Ureas approved in biocides and PPPsb 11 LC-HRMS 26002-80-2 Phenothrin synthetic pyrethroid, component of approved in biocides 6 ECOSAR 6 Esters aerosol insecticides for domestic use 12 GC-MS 101-37-1 Triallyl cyanurate intermediate in the production of other 7 ECOSAR 7 Triazines, Aromatic chemicals such as rubbers, resins and polymers 13 LC-HRMS 6190-65-4 Deethylatrazine degradation product of the herbicide parent compuund not 7.88 US EPA-ECOTOX 198 Triazines, Aromatic atrazine approved 14 LC-HRMS 87392-12-9 S-Metolachlor chloroacetanilide herbicide, elongase approved in PPPs 8 US EPA-ECOTOX 4 Haloacetamides inhibitor 15 LC-HRMS 10453-86-8 Resmethrin synthetic pyrethroid insecticide not approved 11 ECOSAR 11 Esters 16 GC-MS 13401-18-8 1,3-Dimethyluracil metabolite of methylxanthines (caffeine, 11 ECOSAR 11 Carbonyl Ureas theophylline and theobromine) Journal ofHazardousMaterials358(2018)494–502 17 LC-HRMS 82413-20-5 3-hydroxytamoxifen synthetic antiestrogen 15 ECOSAR 15 Aliphatic Amines 18 LC-HRMS 1952-67-6 Crotarbital barbiturate derivative for treatment of 16 ECOSAR 16 Carbonyl Ureas insomnia 19 LC-HRMS 66063-05-6 Pencycuron phenylurea fungicide Approved in PPPs 16 ECOSAR 16 Substituted Ureas 20 GC-MS 2309-49-1 Temurin metabolite of methylxanthines (caffeine, 17 ECOSAR 17 Carbonyl Ureas theophylline and theobromine) 21 LC-HRMS 76-73-3 Secobarbital barbiturate derivative drug 18 ECOSAR 18 Carbonyl Ureas 22 LC-HRMS 7009-43-0 Levometiomeprazin anti-emetic, neuroleptic drug 19 ECOSAR 19 Aliphatic Amines 23 LC-HRMS 77-27-0 Thiamylal barbiturate derivative drug 19 ECOSAR 19 Carbonyl Ureas 24 LC-HRMS 83905-01-5 Azithromycin macrolide antibiotic 19 [38] 1874 Aliphatic Amines 25 LC-HRMS 151-83-7 Methohexital barbiturate derivative drug 20 ECOSAR 20 Carbonyl Ureas

a Information on the EU and Czech regulatory status retrieved from the Pesticide Properties Database (PPDB) by [39] and Information System CHLAP: https://eregpublicsecure.ksrzis.cz/Registr/CHLAP/Biocid. b Diuron is approved in PPPs in Bulgaria and its authorization is in progress in several EU countries. Z. Tousova et al. Journal of Hazardous Materials 358 (2018) 494–502 instrument and data compared with the chemical screening. However, Similar synergism could occur also in the studied WWTP effluent as this was not the aim of the present study, as level 2 of identification multiple macrolides and sulfonamides were detected. provided satisfactory confidence for further combination with toxicity Among the top 25 compounds in the present study were also 4 data and models. barbiturate drugs, nevertheless, their toxicity to algae had to be pre- The ranking of the MTP values, identified herbicides, their trans- dicted by ECOSAR because no experimental data were available. For formation products and other pesticides as the most likely toxicity these compounds, further research is needed to derive the experimental drivers, which is in line with the recent results of Tang and Escher [44]. values and confirm the predicted toxicity. The synthetic antiestrogen, 3- Among the top-ranking pesticides, three compounds were not approved hydroxytamoxifen, was another highly ranking pharmaceutical in the in the Czech Republic for use in either PPPs or biocides i.e. - prometryn, present study with high predicted toxicity to algae. However, experi- acetochlor and resmethrin. The triazine herbicide, prometryn, detected mental IC50 of related compound tamoxifen for P. subcapitata reported by GC(xGC)-MS, was ranked as number 3 in the list of likely toxicity by Orias et al. [57] was 980 μg/L, while its environmental concentra- drivers in the studied sample. This is an interesting finding because the tions usually do no not exceed units of ng/L. A metabolite of another use of prometryn in PPPs has been banned in the EU since 2003 (or pharmaceutical, norfluoxetine, was detected in the active fractions. since 2007 in some countries) and the compound is not registered for Norfluoxetine and its parent compound fluoxetine showed high toxicity use in biocides either. Prometryn was reported to be present in effluents to algae, which even increased by an order of magnitude at pH 10, and the concentrations in the post ban years did not decrease [48]. reaching similar toxicity as the herbicide diuron [58]. This phenom- Similar trend was reported also for atrazine, which was repeatedly enon was explained by the authors by speciation-associated increase in detected in unchanged concentrations in German ground waters over20 toxicokinetic, which could be general to autotrophs. An increase of the years after its ban [49]. In our study, two atrazine transformation pH occurring along with the CO2 fixation may lead to an increase of the products (6-deisopropylatrazine, deethylatrazine), also top-ranked in neutral form of the aliphatic amines increasing thus their bioavail- the list, indicate the occurrence of the parent compound in the en- ability and consequent toxicity. This may have further implications for vironment. Our results brought further evidence that many pesticides, WWTPs as many pharmaceuticals and their transformation products the use of which was either completely banned or restricted for non- can be classified as aliphatic amines and thus may contribute to the agricultural use in biocides, still occur in urban waste water and may observed toxicity to higher extent than expected. pose risk to the aquatic environments [4]; [39]. Non-target screening Caffeine and its 2 metabolites were also ranked as important algal enabled identification of transformation products of several non-ap- growth inhibitors in our study, which complies to results of proved herbicides such as atrazine (mentioned above), diuron (DCPMU Kuzmanovic et al. [50] and Thomaidi et al. [38]. In the mentioned -1-(3,4-dichlorophenyl)-3-methylurea and DCPU - 1-(3,4-di- studies, caffeine was prioritised based on its ECOSAR-predicted value, chlorophenyl)urea) and alachlor (alachlor-ESA - (2-[(2,6-diethyl- while our study used the lowest experimental value for algae phenyl)(methoxymethyl) amino]-2-oxo-ethanesulfonic acid), listed in LOEC = 5 μg/L reported by Lawrence et al. [59].Caffeine toxicity for SM2- Table S4). These results comply with the recent study by Kuz- algae was addressed also by other studies, which in general reported manovic et al. [50], who identified diuron among the 10 most im- orders of magnitude higher effect concentrations [60,61]. The con- portant contaminants in 4 Spanish river basins. However, it should be tribution of caffeine and its metabolites to the observed toxicity noted that diuron was approved for agricultural use in Spain and Bul- therefore remains a matter of research, but as caffeine may occur in garia but not in the rest of the EU [39]. The input of the banned relatively high concentrations, its possible effects on biota should not be compounds and their continuous occurrence in waters years after the ignored. ban remains a principal issue for both researchers and regulators. Possible explanations could be the use of the old stocks, leaching from 5. Conclusions reservoirs in contaminated soils or the run-off from roof paints, plasters or other building materials enhanced by these compounds [48,51]. Our The present EDA study of the WWTP effluent demonstrated a novel result indicated that active substances entering the aquatic environ- approach combining a state-of-the-art sampling tool, biotesting, frac- ment from biocidal products should be carefully considered. tionation and non-target screening techniques. In total 383 compounds Despite of being dominant toxicants for algae, concentrations of were identified in the effluent fractions causing algal toxicity. The in- herbicides estimated by GC(xGC)-MS were at least an order of magni- dicated important contributors to the observed phytotoxicity were tude lower than those of detected PCPs, pharmaceuticals, organopho- herbicides (terbutryn, terbutylazine, prometryn, acetochlor, fluoro- sphates or industrial compounds. This is in line with the results of Neale chloridone, S-metolachlor) and their transformation products, other et al. [31,52] and also our earlier study [53], where 151 CECs were pesticides (carbendazim, phenothrin, resmethrin, pencycuron), phar- determined by multi-residue target analysis in the same WWTP effluent. maceuticals (macrolide azithromycin and several barbiturate drugs), Herbicides occurred within the range of units or tens of ng/L (with an caffeine and its two metabolites and two industrial chemicals (2-(2- exception of diuron - 150 ng/L), while some CECs identified in the heptadec-8-enyl-2-imidazolin-1-yl)ethanol and triallyl cyanurate). As phytotoxic fractions reached up to hundreds of ng/L (benzophenone, the data have qualitative character, further research is needed to diethyltoluamid, sulfamethoxazole) of even units of μg/L (sucralose, quantify the real contribution of individual compounds to the observed azithromycin). These compounds were not ranked among the top toxicity. The sources of likely toxicity drivers, the use of which had toxicity drivers but their higher concentrations may lead to biological been completely banned or restricted to biocidal products, should be effects similar to those of the top-ranking toxicants (cf SM2-Table S4). further explored as well as possible mixture effects of micropollutants, From the class of antibiotics, azithromycin was ranked among the where herbicides, macrolide and sulfonamide antibiotics, and possibly top 25 compounds. Other antibiotics and antimicrobials (oxytetracy- pharmaceuticals with aliphatic amines moiety could play major role. cline, azithromycin metabolite, penicillin V, enoxacin, anhydro-erythromycin, sulfamethoxazole, acediasulfone) were also detected in the Acknowledgements active fractions and could contribute to the toxic effects. It is well recognized that the toxicity of some antibiotics to algae and cyano- The study was supported by EDA-EMERGE project (FP7-PEOPLE- bacteria may be equivalent to herbicides and therefore this class of 2011-ITN, grant agreement number 290100), SOLUTIONS project pharmaceuticals deserves close attention [52]. Several studies also fo- funded by the European Union Seventh Framework Programme (FP7, cused on effects of antimicrobials in multicomponent mixtures, which grant agreement no. 603437), and by the research infrastructure grants often resulted in a strong synergism leading to toxicity increased by an of the Czech Ministry of Education (LM2015051 and CZ.02.1.01/0.0/ order of magnitude for certain combinations of compounds [54,55,56]. 0.0/16_013/0001761) We thank Robert Hrich for enabling the

500 Z. Tousova et al. Journal of Hazardous Materials 358 (2018) 494–502 cooperation and access to the WWTP Brno-Modřice. toxicants in complex mixtures? Anal. Bioanal. Chem. 377 (2003) 397–407. [21] M.G. Weller, A unifying review of bioassay-guided fractionation, effect-directed analysis and related techniques, Sensors 12 (2012) 9181–9209, http://dx.doi.org/ Appendix A. Supplementary data 10.3390/s120709181. [22] P. Lacina, L. Mravcová, M. Vávrová, Application of comprehensive two-dimensional Supplementary material related to this article can be found, in the gas chromatography with mass spectrometric detection for the analysis of selected drug residues in wastewater and surface water, J. Environ. Sci. 25 (2013) 204–212, online version, at doi:https://doi.org/10.1016/j.jhazmat.2018.05.031. http://dx.doi.org/10.1016/S1001-0742(12)60006-0. [23] P.C.F. Lima Gomes, B.B. Barnes, Á.J. Santos-Neto, F.M. Lancas, N.H. Snow, References Determination of steroids, caffeine and methylparaben in water using solid phase microextraction-comprehensive two dimensional gas chromatography-time of flight mass spectrometry, J. Chromatogr. A 1299 (2013) 126–130, http://dx.doi.org/10. [1] J.W. Fleeger, K.R. Carman, R.M. Nisbet, Indirect effects of contaminants in aquatic 1016/j.chroma.2013.05.023. ecosystems, Sci. Total Environ. 317 (2003) 207–233, http://dx.doi.org/10.1016/ [24] V. Matamoros, E. Jover, J.M. Bayona, Part-per-trillion determination of pharma- S0048-9697(03)00141-4. ceuticals, pesticides, and related organic contaminants in river water by solid-phase [2] B.I. Escher, P. Quayle, R. Muller, U. Schreiber, J.F. Mueller, Passive sampling of extraction followed by comprehensive two-dimensional gas chromatography time- herbicides combined with effect analysis in algae using a novel high-throughput of- flight mass spectrometry, Anal. Chem. 82 (2010) 699–706, http://dx.doi.org/10. phytotoxicity assay (Maxi-Imaging-PAM), J. Environ. Monit. 8 (2006) 456–464, 1021/ac902340e. http://dx.doi.org/10.1039/b517512g. [25] R.E. Mohler, K.T. O’Reilly, Da. Zemo, A.K. Tiwary, R.I. Magaw, K.A. Synowiec, Non- [3] S.M. Bengtson Nash, J. Goddard, J.F. Müller, Phytotoxicity of surface waters of the targeted analysis of petroleum metabolites in groundwater using GC×GC-TOFMS, Thames and Brisbane River Estuaries: a combined chemical analysis and bioassay Environ. Sci. Technol. 47 (2013) 10471–10476, http://dx.doi.org/10.1021/ approach for the comparison of two systems, Biosens. Bioelectron. 21 (2006) es401706m. 2086–2093, http://dx.doi.org/10.1016/j.bios.2005.10.016. [26] M.J. Gómez, M.M. Gómez-Ramos, A. Agüera, M. Mezcua, S. Herrera, [4] E.S. Emelogu, P. Pollard, P. Dymond, C.D. Robinson, L. Webster, C. McKenzie, A.R. Fernández-Alba, A new gas chromatography/mass spectrometry method for J. Dobson, E. Bresnan, C.F. Moffat, Occurrence and potential combined toxicity of the simultaneous analysis of target and non-target organic contaminants in waters, dissolved organic contaminants in the forth estuary and firth of forth, Scotland J. Chromatogr. A 1216 (2009) 4071–4082, http://dx.doi.org/10.1016/j.chroma. assessed using passive samplers and an algal toxicity test, Sci. Total Environ. 2009.02.085. 461–462 (2013) 230–239, http://dx.doi.org/10.1016/j.scitotenv.2013.05.011. [27] M. Ibáñez, J.V. Sancho, F. Hernández, D. McMillan, R. Rao, Rapid non-target [5] K.P.J. Viaene, F. De Laender, P.J. Van Den Brink, C.R. Janssen, Using additive screening of organic pollutants in water by ultraperformance liquid chromato- modelling to quantify the effect of chemicals on phytoplankton diversity and bio- graphy coupled to time-of-light mass spectrometry, TrAC - Trends Anal. Chem. 27 mass, Sci. Total Environ. 449 (2013) 71–80, http://dx.doi.org/10.1016/j.scitotenv. (2008) 481–489, http://dx.doi.org/10.1016/j.trac.2008.03.007. 2013.01.046. [28] E.L. Schymanski, H.P. Singer, J. Slobodnik, I.M. Ipolyi, P. Oswald, M. Krauss, [6] C.R. Janssen, D.G. Heijerick, Algal toxicity tests for environmental risk assessments T. Schulze, P. Haglund, T. Letzel, S. Grosse, N.S. Thomaidis, A. Bletsou, C. Zwiener, of metals, Rev. Environ. Contam. Toxicol. 178 (2003) 23–52, http://dx.doi.org/10. M. Ibáñez, T. Portolés, R. de Boer, M.J. Reid, M. Onghena, U. Kunkel, W. Schulz, 1007/0-387-21728-2_2. A. Guillon, N. Noyon, G. Leroy, P. Bados, S. Bogialli, D. Stipaničev, P. Rostkowski, [7] N. Nyholm, T. Kallqvist, Methods for growth inhibition toxicity tests with fresh- J. Hollender, Non-target screening with high-resolution mass spectrometry: critical water algae, Environmnetal Toxicol. Chem. 8 (1989) 689–703. review using a collaborative trial on water analysis, Anal. Bioanal. Chem. (2015) [8] C. Harman, M. Reid, K.V. Thomas, In situ calibration of a passive sampling device 6237–6255, http://dx.doi.org/10.1007/s00216-015-8681-7. for selected illicit drugs and their metabolites in wastewater, and subsequent year- [29] M. Zedda, C. Zwiener, Is nontarget screening of emerging contaminants by LC- long assessment of community drug usage, Environ. Sci. Technol. 45 (2011) HRMS successful? A plea for compound libraries and computer tools, Anal. Bioanal. 5676–5682, http://dx.doi.org/10.1021/es201124j. Chem. 403 (2012) 2493–2502, http://dx.doi.org/10.1007/s00216-012-5893-y. [9] P. Booij, A.D. Vethaak, P.E.G. Leonards, S.B. Sjollema, J. Kool, P. de Voogt, [30] M. Krauss, H. Singer, J. Hollender, LC-high resolution MS in environmental ana- M.H. Lamoree, Identification of photosynthesis inhibitors of pelagic marine algae lysis: from target screening to the identification of unknowns, Anal. Bioanal. Chem. using 96-well plate microfractionation for enhanced throughput in effect-directed 397 (2010) 943–951, http://dx.doi.org/10.1007/s00216-010-3608-9. analysis, Environ. Sci. Technol. 48 (14) (2014) 8003–8011, http://dx.doi.org/10. [31] P.A. Neale, N.A. Munz, S. Aїt-Aїssa, R. Altenburger, F. Brion, W. Busch, B.I. Escher, 1021/es405428t. K. Hilscherová, C. Kienle, J. Novák, T.B. Seiler, Y. Shao, C. Stamm, J. Hollender, [10] E.L.M. Vermeirssen, J. Hollender, N. Bramaz, J. Van Der Voet, B.I. Escher, Linking Integrating chemical analysis and bioanalysis to evaluate the contribution of was- toxicity in algal and bacterial assays with chemical analysis in passive samplers tewater effluent on the micropollutant burden in small streams, Sci. Total Environ. deployed in 21 treated sewage effluents, Environ. Toxicol. Chem. 29 (2010) 576 (2017) 785–795, http://dx.doi.org/10.1016/j.scitotenv.2016.10.141. 2575–2582, http://dx.doi.org/10.1002/etc.311. [32] R. Rojickova, D. Dvorakova, The use of miniaturized algal bioassays in comparison [11] C.H. Johansson, L. Janmar, T. Backhaus, Toxicity of ciprofloxacin and sulfa- to the standard flask assay, Environ. Toxicol. Water Qual. 13 (1998) 235–241. methoxazole to marine periphytic algae and bacteria, Aquat. Toxicol. 156 (2014) [33] J. Slobodnik, L. Mrafkova, M. Carere, F. Ferrara, B. Pennelli, G. Schüürmann, 248–258, http://dx.doi.org/10.7287/peerj.preprints.330v1. P.C. von der Ohe, Identification of river basin specific pollutants and derivation of [12] M.V. Pablos, P. García-Hortigüela, C. Fernández, Acute and chronic toxicity of environmental quality standards: a case study in the Slovak Republic, TrAC Trends emerging contaminants, alone or in combination, in Chlorella vulgaris and Daphnia Anal. Chem. 41 (2012) 133–145, http://dx.doi.org/10.1016/j.trac.2012.08.008. magna, Environ. Sci. Pollut. Res. (2015) 5417–5424, http://dx.doi.org/10.1007/ [34] NIST, Automated Mass Spectral Deconvolution and Identification System (AMDIS), s11356-015-4119-1. (2011). [13] K. Petersen, H.H. Heiaas, K.E. Tollefsen, Combined effects of pharmaceuticals, [35] Thermo Fisher Scientific, U., n.d. Mass FrontierTM; Chemically Intelligent Small personal care products, biocides and organic contaminants on the growth of ske- Molecule Structural Elucidation Software for Mass Spectrometry. letonema pseudocostatum, Aquat. Toxicol. 150 (2014) 450–454, http://dx.doi.org/ [36] US-EPA, 2015. Estimation Programs Interface SuiteTM for Microsoft® Windows, v 10.1016/j.aquatox.2014.02.013. 4.11. [14] R. Rosal, I. Rodea-Palomares, K. Boltes, F. Fernández-Piñas, F. Leganés, S. Gonzalo, [37] T. Nagai, K. Taya, Estimation of herbicide species sensitivity distribution using A. Petre, Ecotoxicity assessment of lipid regulators in water and biologically treated single‐species algal toxicity data and information on the mode of action, Environ. wastewater using three aquatic organisms, Environ. Sci. Pollut. Res. 17 (2010) Toxicol. Chem. 34 (3) (2015) 677–684. 135–144, http://dx.doi.org/10.1007/s11356-009-0137-1. [38] V.S. Thomaidi, A.S. Stasinakis, V.L. Borova, N.S. Thomaidis, Is there a risk for the [15] D. Xu, C. Li, H. Chen, B. Shao, Cellular response of freshwater green algae to per- aquatic environment due to the existence of emerging organic contaminants in fluorooctanoic acid toxicity, Ecotoxicol. Environ. Saf. 88 (2013) 103–107, http:// treated domestic wastewater? Greece as a case-study, J. Hazard. Mater. 283 (2015) dx.doi.org/10.1016/j.ecoenv.2012.10.027. 740–747, http://dx.doi.org/10.1016/j.jhazmat.2014.10.023. [16] US-EPA, 2015. ECOTOX Database [WWW Document]. URL http://cfpub.epa.gov/ [39] Lewis, K.A., Green, A., Tzilivakis, J., Warner, D., 2015. The Pesticide Properties ecotox/ (Accessed 20 April 2015). DataBase (PPDB). [WWW Document]. Dev. by Agric. Environ. Res. Unit (AERU), [17] US-EPA, 2012. ECOSAR 1.11 [WWW Document]. URL http://www.epa.gov/oppt/ Univ. Hertfordshire, 2006-2015. URL http://sitem.herts.ac.uk/aeru/ppdb/en/ newchems/tools/21ecosar.htm. index.htm (Accessed 9 October 2015). [18] P. Reuschenbach, M. Silvani, M. Dammann, D. Warnecke, T. Knacker, ECOSAR [40] A. Köhler, S. Hellweg, B.I. Escher, K. Hungerbühler, Organic pollutant removal model performance with a large test set of industrial chemicals, Chemosphere 71 versus toxicity reduction in industrial wastewater treatment: the example of was- (2008) 1986–1995, http://dx.doi.org/10.1016/j.chemosphere.2007.12.006. tewater from fluorescent whitening agent production, Environ. Sci. Technol. 40 [19] R. Altenburger, S. Ait-Aissa, P. Antczak, T. Backhaus, D. Barceló, T.-B. Seiler, (2006) 3395–3401, http://dx.doi.org/10.1021/es060555f. F. Brion, W. Busch, K. Chipman, M.L. de Alda, G. de Aragão Umbuzeiro, B.I. Escher, [41] A. Magdaleno, Á.B. Juárez, V. Dragani, M.E. Saenz, M. Paz, J. Moretton, F. Falciani, M. Faust, A. Focks, K. Hilscherova, J. Hollender, H. Hollert, F. Jäger, Ecotoxicological and genotoxic evaluation of Buenos Aires city (Argentina) hospital A. Jahnke, A. Kortenkamp, M. Krauss, G.F. Lemkine, J. Munthe, S. Neumann, wastewater, J. Toxicol. 2014 (2014) 10, http://dx.doi.org/10.1155/2014/248461 E.L. Schymanski, M. Scrimshaw, H. Segner, J. Slobodnik, F. Smedes, S. Kughathas, Article ID 248461. I. Teodorovic, A.J. Tindall, K.E. Tollefsen, K.-H. Walz, T.D. Williams, P.J. Van den [42] B.D.S. Maselli, L.A.V. Luna, J.D.O. De Palmeira, S. Babosa, L.A. Beijo, Brink, J. van Gils, B. Vrana, X. Zhang, W. Brack, Future water quality monitoring — G.D.A. Umbuzeiro, F. Kummrow, Ecotoxicity of raw and treated effluents generated adapting tools to deal with mixtures of pollutants in water resource management, by a veterinary medicine industry. Ambient. e Agua - An Interdiscip, J. Appl. Sci. 8 Sci. Total Environ. (2015) 540–551, http://dx.doi.org/10.1016/j.scitotenv.2014. (2013) 795–804, http://dx.doi.org/10.4136/ambi-agua.1121. 12.057. [43] S.M. Bengtson Nash, U. Schreiber, P.J. Ralph, J.F. Müller, The combined SPE:ToxY- [20] W. Brack, Effect-directed analysis: a promising tool for the identification of organic PAM phytotoxicity assay; Application and appraisal of a novel biomonitoring tool

501 Z. Tousova et al. Journal of Hazardous Materials 358 (2018) 494–502

for the aquatic environment, Biosens. Bioelectron. 20 (2005) 1443–1451, http://dx. P. Spirhanzlova, A. Tindall, T. Schulze, European demonstration program on the doi.org/10.1016/j.bios.2004.09.019. effect-based and chemical identification and monitoring of organic pollutants in [44] J.Y.M. Tang, B.I. Escher, Realistic environmental mixtures of micropollutants in European surface waters, Sci. Total Environ. (2017) 1849–1868, http://dx.doi.org/ surface, drinking, and recycled water: Herbicides dominate the mixture toxicity 10.1016/j.scitotenv.2017.06.032. toward algae, Environ. Toxicol. Chem. 33 (2014) 1427–1436, http://dx.doi.org/10. [54] M. Cleuvers, Aquatic ecotoxicity of pharmaceuticals including the assessment of 1002/etc.2580. combination effects, Toxicol. Lett. 142 (2003) 185–194, http://dx.doi.org/10. [45] M.J. Gómez, S. Herrera, D. Solé, E. García-Calvo, A.R. Fernández-Alba, Automatic 1016/S0378-4274(03)00068-7. searching and evaluation of priority and emerging contaminants in wastewater and [55] M. González-Pleiter, S. Gonzalo, I. Rodea-Palomares, F. Leganés, R. Rosal, K. Boltes, River Water by stir Bar sorptive extraction followed by comprehensive Two-di- E. Marco, F. Fernández-Piñas, Toxicity of five antibiotics and their mixtures towards mensional gas chromatography-time-of-flight mass spectrometry, Anal. Chem. 83 photosynthetic aquatic organisms: implications for environmental risk assessment, (2011) 2638–2647. Water Res. 47 (2013) 2050–2064, http://dx.doi.org/10.1016/j.watres.2013.01. [46] S. Herrera Lõpez, M.D. Hernando, M.J. Gõmez, J. Santiago-Morales, R. Rosal, 020. A.R. Fernández-Alba, Investigation of galaxolide degradation products generated [56] L.-H. Yang, G.-G. Ying, H.-C. Su, J.L. Stauber, M.S. Adams, M.T. Binet, Growth- under oxidative and irradiation processes by liquid chromatography/hybrid quad- inhibiting effects of 12 antibacterial agents and their mixtures on the freshwater rupole time-of-flight mass spectrometry and comprehensive two-dimensional gas microalga Pseudokirchneriella subcapitata, Environ. Toxicol. Chem. 27 (2008) chromatography/time-of-flight mass spect, Rapid Commun. Mass. Spectrom. 27 1201–1208, http://dx.doi.org/10.1897/07-471.1. (2013) 1237–1250, http://dx.doi.org/10.1002/rcm.6575. [57] F. Orias, S. Bony, A. Devaux, C. Durrieu, M. Aubrat, T. Hombert, A. Wigh, [47] E.L. Schymanski, J. Jeon, R. Gulde, K. Fenner, M. Ru, H.P. Singer, J. Hollender, Y. Perrodin, Tamoxifen ecotoxicity and resulting risks for aquatic ecosystems, Identifying small molecules via High Resolution mass spectrometry: communicating Chemosphere 128 (2015) 79–84, http://dx.doi.org/10.1016/j.chemosphere.2015. confidence, Environ. Sci. Technol. 48 (2014) 2097–2098, http://dx.doi.org/10. 01.002. 1021/es5002105. [58] J. Neuwoehner, B.I. Escher, The pH-dependent toxicity of basic pharmaceuticals in [48] K. Quednow, W. Püttmann, Temporal concentration changes of DEET, TCEP, ter- the green algae scenedesmus vacuolatus can be explained with a toxicokinetic ion- butryn, and nonylphenols in freshwater streams of Hesse, Germany: possible in- trapping model, Aquat. Toxicol. 101 (2011) 266–275, http://dx.doi.org/10.1016/j. fluence of mandatory regulations and voluntary environmental agreements, aquatox.2010.10.008. Environ. Sci. Pollut. Res. 16 (2009) 630–640, http://dx.doi.org/10.1007/s11356- [59] J.R. Lawrence, B. Zhu, G.D.W. Swerhone, J. Roy, V. Tumber, M.J. Waiser, E. Topp, 009-0169-6. D.R. Korber, Molecular and microscopic assessment of the effects of caffeine, [49] D. Vonberg, J. Vanderborght, N. Cremer, T. Pütz, M. Herbst, H. Vereecken, 20 years acetaminophen, diclofenac, and their mixtures on river biofilm communities, of long-term atrazine monitoring in a shallow aquifer in western Germany, Water Environ. Toxicol. Chem. 31 (2012) 508–517, http://dx.doi.org/10.1002/etc.1723. Res. (2014) 50, http://dx.doi.org/10.1016/j.watres.2013.10.032. [60] K. Pollack, K. Balazs, O. Ogunseitan, Proteomic assessment of caffeine effects on [50] M. Kuzmanovic, A. Ginebreda, M. Petrovic, D. Barcelo, Risk assessment based coral symbionts, Environ. Sci. Technol. 43 (2009) 2085–2091, http://dx.doi.org/ prioritization of 200 organic micropollutants in 4 Iberian rivers, Sci. Total Environ. 10.1021/es802617f. 504 (2014) 289–299, http://dx.doi.org/10.1016/j.scitotenv.2014.06.056. [61] A. Zarrelli, M. DellaGreca, M.R. Iesce, M. Lavorgna, F. Temussi, L. Schiavone, [51] Venzmer, H.V., 2008. Forum Altbausanierung 2. Biofilme und funktionale E. Criscuolo, A. Parrella, L. Previtera, M. Isidori, Ecotoxicological evaluation of Baustoffoberflächen.: 8. Dahlberg-Kolloquium vom 25. bis 26. September 2008 im caffeine and its derivatives from a simulated chlorination step, Sci. Total Environ. Zeughaus Wismar. Fraunhofer IRB Verlag. (2014) 453–458, http://dx.doi.org/10.1016/j.scitotenv.2013.10.005. [52] J. Guo, A. Boxall, K. Selby, Do pharmaceuticals pose a threat to primary producers? [62] T. Schulze, M. Ahel, J. Ahlheim, S. Aït-Aïssa, F. Brion, C. Di Paolo, J. Froment, Crit. Rev. Environ. Sci. Technol. 3389 (2015), http://dx.doi.org/10.1080/ A.O. Hidasi, J. Hollender, H. Hollert, M. Hu, A. Kloß, S. Koprivica, M. Krauss, 10643389.2015.1061873 00–00. M. Muz, P. Oswald, M. Petre, J.E. Schollée, T.-B. Seiler, Y. Shao, J. Slobodnik, [53] Z. Tousova, P. Oswald, J. Slobodnik, L. Blaha, M. Muz, M. Hu, W. Brack, M. Krauss, M. Sonavane, M.J.-F. Suter, K.E. Tollefsen, Z. Tousova, K.-H. Walz, W. Brack, C. Di Paolo, Z. Tarcai, T.-B. Seiler, H. Hollert, S. Koprivica, M. Ahel, J.E. Schollée, Assessment of a novel device for onsite integrative large-volume solid phase ex- J. Hollender, M.J.-F. Suter, A.O. Hidasi, K. Schirmer, M. Sonavane, S. Ait-Aissa, traction of water samples to enable a comprehensive chemical and effect-based N. Creusot, F. Brion, J. Froment, A.C. Almeida, K. Thomas, K.E. Tollefsen, S. Tufi, analysis, Sci. Total Environ. (2017) 350–358, http://dx.doi.org/10.1016/j. X. Ouyang, P. Leonards, M. Lamoree, V.O. Torrens, A. Kolkman, M. Schriks, scitotenv.2016.12.140.

502

ANNEX III Science of the Total Environment 650 (2019) 1599–1612

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Analytical and bioanalytical assessments of organic micropollutants in the Bosna River using a combination of passive sampling, bioassays and multi-residue analysis

Zuzana Toušová a ,b , Branislav Vrana a ,c , Marie Smutná a ,JiříNováka ,VeronikaKlučárová d , Roman Grabic e , Jaroslav Slobodník b ,JohnPaulGiesyf ,g ,h , Klára Hilscherová a ,⁎ a Masaryk University, Faculty of Science, RECETOX, Kamenice 753/5, 625 00 Brno, Czech Republic b Environmental Institute (EI), Okružná 784/42, 972 41 Koš, Slovakia c Water Research Institute, Nabr. Arm. Gen. L. Svobodu 5, 812 49 Bratislava, Slovakia d Slovak University of Technology, Faculty of Chemical and Food Technology, Radlinskeho 9, 812 37 Bratislava, Slovakia e University of South Bohemia in Ceske Budejovice, Faculty of Fisheries and Protection of Waters, South Bohemian Research Center of Aquaculture and Biodiversity of Hydrocenoses, Zatisi 728/II, CZ- 389 25 Vodnany, Czech Republic f Dept. Biomedical Veterinary Sciences and Toxicology Centre, University of Saskatchewan, 52 Campus Drive, Saskatoon, SK S7N 5B4, Saskatchewan, Canada g School of Biological Sciences, University of Hong Kong, Hong Kong, SAR, People's Republic of China h State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, Nanjing University, Nanjing, People's Republic of China

HIGHLIGHTS GRAPHICAL ABSTRACT

• Impact of insufﬁcient water treatment on river water quality was assessed. • Passive sampling, multi-residue analysis and battery of bioassays were combined. • Sarajevo was identiﬁed as a major pollution source in Bosna river. • Diazinon occurred at most sites in concentrations posing risk to aquatic biota. • Most bioactivities (except estrogenicity) were not explained by detected compounds.

article info abstract

Article history: Complex mixtures of contaminants from multiple sources, including agriculture, industry or wastewater enter Received 30 May 2018 aquatic environments and might pose hazards or risks to humans or wildlife. Targeted analyses of a few priority Received in revised form 24 August 2018 substances provide limited information about water quality. In this study, a combined chemical and effect screen- Accepted 24 August 2018 ing of water quality in the River Bosna, in Bosnia and Herzegovina was carried out, with focus on occurrence and Available online 27 August 2018 effects of contaminants of emerging concern. Chemicals in water were sampled at 10 sites along the Bosna River

Abbreviations: AR, androgen receptor; AhR, arylhydrocarbon receptor; BiH, Bosnia and Herzegovina; BEQ, bioanalytical equivalent concentration; CECs, contaminants of emerging concern; CI, contamination index; CUPs, currently used pesticides; DCM, dichloromethane; DDT, 1,1,1-trichloro-2,2-bis(4-chlorphenyl)ethane; DHT, dihydrotestosterone; DHT-EQ, dihydrotestosterone equivalent; DS, downstream; E1, estrone; E2, 17β-estradiol; E2-EQ, 17β-estradiol equivalent; E3, estriol; EC50, concentration at which the effect reaches 50% of the effect in positive control; EE2, 17α-ethinylestradiol; ER, estrogen receptor; Flu-EQ, ﬂutamide equivalent; HCH, hexachlorocyclohexane; HI, hazard index; HQ, hazard quotient; LOD, limit of detection; LOQ, limit of quantiﬁcation; PAHs, polycyclic aromatic hydrocarbons; PCBs, polychlorinated biphenyls; POCIS, polar organic chemical integrative sampler; OCPs, organochlorine pesticides; PRC, performance reference compound; REP, relative effect potency; RB, river basin; SPE, solid phase extraction; SPMD, Semi permeable membrane device; TCDD, 2,3,7,8- Tetrachlorodibenzo-p-dioxin; TCDD-EQ, TCDD equivalent; US, upstream; WWTP, waste water treatment plant. ⁎ Corresponding author at: Masaryk University, Faculty of Science, RECETOX, Kamenice 753/5, 62500 Brno, Czech Republic. E-mail address: [email protected] (K. Hilscherová).

Editor: D. Barcelo by use of passive sampling. The combination of semipermeable membrane devices (SPMDs) and polar organic chemical integrative samplers (POCIS) enabled sampling of a broad range of contaminants from hydrophobic Keywords: (PAHs, PCBs, OCPs) to hydrophilic compounds (pesticides, pharmaceuticals and hormones), which were deter- Contaminants of emerging concern - passive mined by use of GC–MS and LC-MS (MS). In vitro, cell-based bioassays were applied to assess (anti)androgenic, sampling estrogenic and dioxin-like potencies of extracts of the samplers. Of a total of 168 targeted compounds, 107 were In vitro bioassay - endocrine disruption Hazard profiling - water quality monitoring detected at least once. Cumulative pollutant concentrations decreased downstream from the city of Sarajevo, which was identified as the major source of organic pollutants in the area. Responses in all bioassays were observed for samples from all sites. In general, estrogenicity could be well explained by analysis of target estrogens, while the drivers of the other observed effects remained largely unknown. Profiling of hazard quotients identified two sites downstream of Sarajevo as hotspots of biological potency. Risk assessment of detected compounds revealed, that 7 compounds (diazinon, diclofenac, 17β-estradiol, estrone, benzo[k]fluoranthene, fluoranthene and benzo[k]fluoranthene) might pose risks to aquatic biota in the Bosna River. The study brings unique results of a complex water quality assessment in a region with an insufficient water treatment infrastructure. © 2018 Elsevier B.V. All rights reserved.

1. Introduction In vitro bioassays, as sensitive, rapid and cost-effective screening tools, have been applied previously to detect micro-pollutants in Sustainable management of water resources relies on regular moni- water (Escher et al., 2014; Jalova et al., 2013; Neale et al., 2015). Unlike toring of status and trends of qualities of surface waters, which allows target chemical analysis, bioanalytical tools take into account possible identification of hazards and or risks posed by multiple anthropogenic mixture effects of all chemicals present in the sample (Altenburger stressors (Geissen et al., 2015). Chemical pollution of water resources et al., 2015). Bioassays can, therefore, help to provide a more holistic is considered one of the main causes of impairment of aquatic ecosys- picture of possible hazards of complex environmental mixtures of prior- tems and losses of biodiversity (Vörösmarty et al., 2010). The ever- ity pollutants and CECs to aquatic biota (Connon et al., 2012). increasing multitude of chemicals entering aquatic environments con- The Bosna River Basin (RB), with a total surface area of stitutes a challenge for monitoring schemes, because most of these com- 10,809.83 km2 and a population of almost one million is the most pop- pounds typically occur at rather small (sub-ng L−1 ) concentrations. ulated and developed region in Bosnia and Herzegovina (BiH). The However, some of these are sufficiently potent or have the potential to Bosna River is about 275 km long and receives pollution from several be accumulated to concentrations such that they can elicit biological ef- points and diffuse sources. The major point sources comprise various in- fects. Moreover, some chemicals might undergo biotic or abiotic trans- dustries, including among others, leather, pulp and paper, steel making, formation forming very complex environmental mixtures where most oil refining, thermal power, municipalities (Sarajevo, with no WWTP in individual components can only hardly be identified (Ginebreda et al., operation at the time of sampling, Zenica and Doboj) and landfills (Sara- 2014). These compounds, known as contaminants of emerging concern jevo and Zenica). Diffuse pollution originates from agriculture and (CECs), comprise many different chemical and usage pattern groups, i.e. households, because only about 50% of population is connected to sew- personal care products, human and veterinary pharmaceuticals, surfac- erage systems. Releases of untreated effluents from municipalities and tants and surfactant-derived compounds, X-ray contrast media, industrial facilities often dominated by old and generally less effective pesticides, disinfection by-products, algal toxins, flame retardants, plas- technologies are considered a key environmental problem in the region ticizers, UV-filters, industrial compounds and transformation products (Smital et al., 2013). (Sima et al., 2014). CECs together with priority pollutants such as The objective of the study, results of which are presented here, was PAHs, legacy and currently used pesticides (CUPs) might cause adverse to characterize, in some detail, water quality in the Bosna River affected effects in aquatic biota and pose risks to human health. by untreated wastewaters. This was achieved by use of a combination of Passive sampling presents a promising approach for surface water passive sampling, a battery of in vitro bioassays and targeted chemical monitoring of CECs, because it provides a sensitive measurement of dis- analyses for several compound classes. This approach evaluated chemi- solved concentrations that is integrated over time (Cfree). Due to its pro- cal and ecotoxicological status at 10 sampling sites along the river. The portionality to the chemical activity and chemical potential, Cfree is specific goals of this study were to: 1) screen for potencies of response considered a key parameter in understanding chemical's exposure of in bioassays as well as quantification of 168 targeted compounds in ex- aquatic organisms (Reichenberg and Mayer, 2006). Passive sampling tracts of SPMD and POCIS samples; 2) estimate proportions of observed enables integrative collection of contaminants over an extended period responses in bioassay that could be explained by targeted chemicals and of time and captures residues from episodic events, which typically re- 3) identify hotspots by use of contamination profiling and chemicals main undetected when using grab sampling (Alvarez et al., 2004; posing risk to aquatic biota by means of hazard assessment. Vrana et al., 2005). Passive samplers are available for sampling of a wide variety of compounds, e.g. semipermeable membrane device 2. Materials and methods (SPMD) for hydrophobic substances such as PAHs or PCBs (Huckins et al., 1993) and polar organic chemical integrative sampler (POCIS) 2.1. Study area for hydrophilic substances such as polar pesticides and pharmaceuticals (Alvarez et al., 2004). Passive samplers are non-mechanical devices, that Passive samplers (SPMD and POCIS) were deployed at 10 locations require minimal resources of personnel and equipment for sampling, along the Bosna River, BiH, for 28–43 days from mid-October to mid- and they constitute a well-defined sampling medium with a constant November 2012 (Fig. 1, Table S1 in Supplementary Materials SM2). De- uptake capacity. Because of the integrative character of sampling, they tailed information on individual sampling sites, exact dates of sampler's accumulate a sufficient amount of sampled chemicals for detection of deployment, physicochemical parameters of river water during deploy- small concentrations in water and samples for multiple analyses includ- ment and retrieval of the samplers, and estimation of sampled volumes ing bioassays (Jones et al., 2015; Moschetetal.,2014; Vrana et al., 2014). are provided in Table S1 in SM2. The sampling sites were selected to Passive sampling has been successfully combined with in vitro bioassays span the whole waterway from the source, in the south, upstream of Sa- in many earlier studies (Emelogu et al., 2013; Jalova et al., 2013; rajevo to the confluence with the Sava River near Modrica in the north. Jarosova et al., 2012). In order to evaluate absolute and relative sources of target compounds Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612 1601

4.95 mL (triolein + membrane). SPMDs were stored at −20 °C in gas- tight metal containers before use. After exposure, collected SPMD samplers were processed according to Vrana et al. (2014). In brief, they were cleaned of mud and debris, placed in a cooled container and transported to the laboratory. Accumu- lated compounds were extracted by dialysis to hexane (two times for 24 h). The volume of dialysates was reduced and extracts were further cleaned up by gel permeation chromatography (GPC) with dichloromethane as a mobile phase. The SPMD extracts for toxicological analysis were solvent exchanged to 100 μL DMSO. SPMD extracts for chemical analyses were reduced in volume and further fractionated by use of silica gel or sulfuric acid modiﬁed silica gel for PAHs, PCBs and OCPs analyses. The SPMD samplers and their extracts were stored at −20 °C.

2.2.2. POCIS POCIS samplers, consisting of membrane-sorbent-membrane layers compressed between two stainless-steel support rings, were purchased from Exposmeter AB, Sweden (www.exposmeter. com) under the commercial name EWH-Pharm - Exposmeter Water Hydrophilic Pharmaceuticals. The membrane was made of microporous polyethersulphone (PES) with 0.1 μmporesize.The samplers with the surface area of 45.8 cm2 contained 200 mg Oasis HLB powder adsorbent. Following exposure, each POCIS sampler was dismantled and the sorbent was by means of Milli-Q water transferred into an empty 3 mL SPE cartridge fitted with a polypro- pylene frit. The sorbent phase was dried by applying gentle vacuum Fig. 1. Sites (S1-S10) of deployment of passive samplers in autumn 2012 along the Bosna on a vacuum manifold. The mass of recovered sorbent was deter- River, BiH. mined gravimetrically from the mass difference of the SPE cartridge with and without sorbent. For one POCIS, the analytes were eluted and biological effects, sampling sites were situated upstream (US) and from the sorbent with 2 × 3 mL of elution mixture (MeOH:DCM, downstream (DS) of major known point sources of pollution entering 1:1 v/v). The eluate was then evaporated under mild stream of ni- the River Bosna (municipalities, industries and landfills; Fig. 1). trogen and reconstituted into 3 mL of MeOH. Extract of was divided into two aliquots intended for bioassays and chemical analyses of 2.2. Sampling and sample processing CUPs. A solvent of samples for toxicological analysis was exchanged for 0.5 mL of DMSO. Parallelly exposed individual POCIS were used Two passive sampler cages were co-deployed at each sampling loca- for the chemical analyses as described in 2.3.3 and 2.3.4. The tion. One contained samplers intended for bioassay screening and the POCIS samplers and their extracts were stored until analyses at other one samplers for chemical analyses. Cages, made of perforated −20 °C. stainless steel, were commercially available (www.exposmeter.com). At each sampling location, 3 POCIS samplers and 3 replicate SPMD sam- 2.3. Chemical analyses plers were placed into each protective cage. Cages with samplers were installed in the river water, usually from bridge pillars, approximately Chemical analyses of 168 target compounds (134 in POCIS and 34 in 1 m below the surface and fixed in place by use of weights, buoys and SPMD extracts) in 4 compound classes were conducted by use of state- ropes. At the end of exposure, samplers were collected and checked of-the-art GC–MS(-MS) and HPLC-MS(-MS). for formation of biofilms or damage. While samplers were being deployed and collected, an additional field control of each sampler type 2.3.1. Hydrophobic compounds analyzed in SPMD extracts was exposed to air only and processed identically to the deployed in- Hydrophobic compounds were determined according to a method stream samplers. The field control was used to assess potential sampler described elsewhere (Vrana et al. 2014). In brief, identification and contamination during transportation, storage and handling. Potential quantification of PAHs were conducted using 6890 N GC (Agilent Tech- contamination arising from the manufacturing process, sampler compo- nologies, Palo Alto, CA, USA), which was equipped with a 30 m × nents, laboratory storage, processing and analytical procedures, was 0.25 mm × 0.25 μm HP5-MS column (Agilent, USA) coupled to 5972 assessed by analysis of fabrication control passive samplers (3 replicates MS operated in electron impact ionization mode. PCBs and OCPs analy- for each sampler type). Analysis of fabrication controls also served to sis was conducted on GC–MS/MS 6890 N GC (Agilent Technologies, Palo determine the initial concentration of PRCs in the SPMD samplers before Alto, CA, USA) equipped with a 60 m × 0.25 mm × 0.25 μm DB5-MS col- exposure (Booij et al., 2007; Huckins et al., 2002). umn (Agilent J&W, USA) which was coupled to a Quattro Micro GC–MS/ MS (Waters, Micromass, Manchester, UK) and operated in electron ion- 2.2.1. SPMDs ization mode. Details of sample processing and instrumental analysis SPMDs, purchased from Exposmeter AB, Tavelsjö, Sweden (www. are provided in SM1-Section 1. exposmeter.com), consisted of an LDPE (Low-density polyethylene) membrane filled with triolein (1 mL, 95% purity), with nominal dimen- 2.3.2. Currently used pesticides analyzed in POCIS extracts sions 2.54 × 91.4 cm, exposure surface area of 460 cm2 and wall- Agilent 1290 series (Agilent Technologies, Waldbronn, Germany) thickness of 75–90 μm. Samplers designated for chemical analyses HPLC coupled to MS-MS AB Sciex Qtrap 5500 (AB Sciex, Concord, ON, contained 2 μg sampler−1 of individual performance reference com- Canada) with electrospray ionization (ESI) was used for analyses of pounds (PRCs; D10-Acenaphthene, D12-Benzo(e)pyrene, D12-Chrysene, CUPs. A Phenomenex SecureGuard C18 guard column (Phenomenex, D10-Fluorene, D10-Phenanthrene). No PRCs were added to the samplers Torrance, CA, USA) followed by a Phenomenex Synergy Fusion C-18 intended for toxicological analysis. The volume of the sampler was end capped column (100 mm × 2 mm i.d., 4 μmparticles)wasused 1602 Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612 for separation of target compounds. Quantifications were based on iso- 2.4.2. ER-mediated potency topically labelled internal standards. Method details are given in SM1- Estrogenicity of extracts was assessed by use of MVLN cells Section 2 and in Brumovský et al. (2016). (Demirpence et al., 1993). These are human, breast carcinoma cells transfected with a promoter containing estrogen responsive elements driving expression of luciferase, as earlier described in Jálová et al. 2.3.3. Estrogens analyzed in POCIS extracts (2013). In brief, MVLN cells were exposed to sample extracts, calibra- POCIS for analyses of estrogens were processed as described in tion reference and solvent control in DMEM/F12 medium for 24 h at Skodova et al. (2016). Estrogens were analyzed by use of HPLC Agilent 37 °C. Standard calibration was conducted using 17β-estradiol (E2; 1200 Series (Agilent Technologies, Waldbronn, Germany) coupled to six-point dilution series 1–500 pM) and the intensity of luciferase lumi- MS-MS (Agilent 6410 Triple Quad; Agilent Technologies, Waldbronn, nescence was assessed by use of Promega Steady Glo Kit (Promega, Germany) after derivatization with dansyl chloride (Lin et al. 2007). USA). An ACE 3 C18 column (150 mm × 4.6 mm, 3 μm) coupled with a pre- column was used for chromatographic separation. Quantification was 2.4.3. AhR-mediated potency based on internal standards (E2-d4, E3-d2) and a 9-point calibration Dioxin-like activities, mediated through the aryl hydrocarbon recep- curve. Dansyl chloride derivatives exhibited a fragment ion m/z of 171, tor (AhR) were assessed by use of H4G1.1c2 cells (CAFLUX assay), rat present in the MS-MS spectra of all investigated compounds. Detailed hepatoma cells which contain a GFP reporter gene under control of description of POCIS extraction, clean-up, derivatization and LC-MS- dioxin-responsive elements (Nagy et al. 2002). In brief, the H4G1.1c2 MS analysis of estrogens is given in SM1 – Section 2. cells were exposed to extracts, calibration reference and solvent control in DMEM medium for 24 h at 37 °C. 2,3,7,8-Tetrachlorodibenzo-p- 2.3.4. Pharmaceuticals analyzed in POCIS extracts dioxin (TCDD) served as standard reference compound and calibration POCIS for analyses of pharmaceuticals were processed as described curves were established by use of six-point dilution series (1–500 in Fedorova et al. (2014). Pharmaceuticals, illicit drugs and metabolites pM). The intensity of fluorescence as a measure of receptor activation analyses in POCIS samples were carried out by use of an Accela 1250 LC was assessed after medium replacement with phosphate buffer. pump and Accela 600 LC pump (Thermo Fisher Scientific, San Jose, CA, USA) with an HTS XT-CTC autosamplers (CTC Analytics AG, Zwingen, 2.4.4. Cytotoxicity Switzerland) coupled to a Q-Exactive mass spectrometer and a triple Combination of three dyes, according to Schirmer et al. (1998) with stage quadrupole MS/MS TSQ Quantum Ultra mass spectrometer slight modifications, was used to assess cytotoxicity of the sample ex- (both from Thermo Fisher Scientific, San Jose, CA, USA). A HypersilGold tracts. The intensity of fluorescence was measured after 30 min of incu- aQ column (50 mm × 2.1 mm ID × 5 μm particles; Thermo Fisher Scien- bation with CFDA-AM (5-carboxyfluorescein diacetate acetoxy- methyl tific, San Jose, CA, USA) was employed for separation of the target ester, Invitrogen, Basel, Switzerland) reflecting cell membrane integrity analytes. The POCIS extracts were supplemented with isotope-labelled and with AlamarBlue (AB, Invitrogen, Basel, Switzerland) showing cel- internal standards prior to dilution with ultrapure water (1:1) and ana- lular metabolic activity (530/590 nm and 485/520 nm, respectively). Af- lyzed by use of conventional LC injection (10 μL of sample per injection). terwards, lysosomal membrane integrity was assessed by measurement The validated method was described in full detail in previous papers of absorbance (540 nm) after 2 h incubation with neutral red (NR, (Fedorova et al., 2014, 2013; Grabic et al., 2012). Sigma-Aldrich, Buchs, Switzerland). Cell viability was also assessed by microscopic inspection.

2.4. In vitro bioassays 2.4.5. Bioassay data analysis Three cell-based reporter gene bioassays were used to examine ER-, Responses to sample extracts were expressed as percents of the (anti)AR- and AhR-mediated potencies and cytotoxicity of organic ex- maximum response observed in the calibration reference curves (% tracts of passive samplers. DMSO (0.5% v/v, Sigma Aldrich, Czech Rep.) DHTmax/% E2max/% TCDDmax). The response of the solvent control was was used as a solvent for extracts and reference compounds. All assays subtracted from both sample and calibration responses. Nonlinear loga- were conducted in 96-well microplates and included six dilutions of ex- rithmic regression of dose-response curves of calibration reference and tracts in triplicate to characterize a dose-response curve for each sam- samples was used for calculation of effect concentrations equivalent to ple. Samples were always tested in at least two independent 50% of maximal response (EC50; Graph Pad Prism 6, GraphPad® experiments. A brief description of the bioassays is provided below, Software, San Diego, California, USA). Bioanalytical equivalent concen- while more details on the bioassay methods and test conditions are pro- trations (BEQbio) were calculated by relating the EC50 values of calibra- vided in the SM1 – Section 3. tion reference (DHT-EQ, E2-EQ, TCDD-EQ) with the concentration of the tested sample inducing the same response (Escher and Leusch, 2012; Jalova et al., 2013). Percentage of the maximal luminescence inhibition 2.4.1. AR-mediated potency in the calibration curves of reference antiandrogen flutamide co- Androgenicity and antiandrogenicity of extracts were assessed using exposed with competitive concentration of DHT (100 pM) were used MDA-kb2 cells (Wilson et al., 2002). These are human breast cancer to characterize the antiandrogenic effects expressed as FLU-EQ based cells transfected with a promoter containing androgen responsive ele- on EC50 levels. LOQs for individual samples in each bioassay were calcu- ments driving expression of luciferase, as detailed in Jálová et al. lated as 3-fold the standard deviation (SD) of the average response of (2013). In brief, the MDA-kb2 cells were exposed to calibration of refer- solvent control on each assay plate according to Könemann et al. ence compound, solvent control and sample extracts, in L-15 medium (2018). Results of cytotoxicity evaluation were expressed as a fraction for 24 h at 37 °C. Standard reference calibration curves for androgenic of control (FOC) ranging from 0 to 1 and corresponding to a relative de- response were produced with dihydrotestosterone (DHT; eight-point crease of fluorescence/absorbance of samples related to solvent control. dilution series: 3.3 pM–100 nM). To assess antiandrogenicity, the cells Potency balance calculations using the ratio between BEQchem were co-exposed to competing endogenous ligand DHT (100 pM) to- (based on chemical analysis) and BEQbio values (based on bioassays) gether with sample extracts, solvent control or calibration of standard were carried out to quantify the proportion of the response of bioassays antiandrogen flutamide (FLU; five-point dilution series: 110 nM–100 that could be explained by detected chemicals. Total BEQchem was μM). The intensity of luciferase luminescence was measured with pre- calculated as a sum of BEQchem values of individual compounds, for pared luciferase reagent (Pavlíková et al., 2012). which relative effect potency (REP) value was available. REPs for target Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612 1603 compounds based on identical or similar biological models were col- 2.6. Hazard assessment lected in open scientific literature or ToxCast dashboard (US EPA, 2015). Assessments of hazards of detected target compounds were conducted by use of the lowest predicted no effect concentrations (PNEC) 2.4.6. Calculation of dissolved water concentrations from passive sampler values derived by the NORMAN Network to identify compounds with data hazard potential (Dulio et al., 2018; Working Group on Prioritisation of Emerging Substances, 2013). In order to protect aquatic biota, ecotox- 2.4.6.1. SPMD. Concentrations of PAHs, PCBs and OCPs dissolved in water icological threshold values, known as the lowest NORMAN PNECs, were were calculated from amounts accumulated in SPMDs by exactly fol- determined on the basis of experimental data, existing environmental lowing the previously described method (Vrana et al. 2014). Briefly, quality standards (EQSs), or in silico predictions. Hazard quotients the calculations were based on an assumption that the amounts of (HQs) were calculated (Eq. 1), where: ci is the calculated dissolved con- analytes absorbed by samplers follow a first-order approach to equilib- centration of an individual compound in water at a particular sampling rium. Aqueous concentrations of individual compounds were calculated site, and PNEC is the lowest NORMAN PNEC value. Compounds with from the mass absorbed by the SPMD, the in situ sampling rate of the HQs b 1.0 are less than the threshold for a specified level of effect, compounds (RS) and their sampler-water partition coefficients (KSW). while those with HQs exceeding 1 might pose risk to aquatic life. Overall Nonlinear least squares method according to Booij and Smedes (2010) hazard index (HI) was calculated by summation of all HQs of com- was used to estimate the Rs values on the basis of dissipation of PRCs pounds detected at each sampling site and concentrations below LOQ from SPMDs during exposure. Fraction f of individual PRCs (D10- were considered zero. acenaphthene, D10-fluorene, D10-phenanthrene and D10-chrysene and c D12-benzo[e]pyrene), that had remained in the SPMD after exposure, HQ ¼ i ð1Þ was considered as a continuous function of their KSW,withRS as an ad- PNEC justable parameter. A model for water-boundary layer controlled uptake, derived by Rusina et al. (2010) was used to estimate RS of individual target compounds as a function of their molar mass. 3. Results and discussion For the purpose of comparison of toxic potencies of extracts from SPMDs from different sampling sites, the measured bioanalytical equiv- 3.1. Chemical analyses −1 alent concentrations (BEQbio) in extracts [ng.SPMD ] were translated −1 to water concentrations CW-BEQ [ng L ] at the individual sites as All passive samplers were successfully retrieved from sampling sites shown in Jálová et al. (2013). Linear uptake of the compounds that ex- except for the POCIS sampler deployed at site S2, where two out of three hibit bioassay response in the extracts was assumed since their physico- POCIS discs were damaged and therefore not enough extract was avail- chemical properties are not known. SPMDs for toxicological analysis able to carry out the analysis of pharmaceuticals and several CUPs. Sam- were not spiked with PRCs, and thus their sampling rate were for each plers deployed at site S10 were retrieved after 43 instead of the planned sampling site calculated as RS values of co-deployed SPMDs for chemical 30 days due to flooding of the Sava River. Of the 168 target compounds, analysis for a compound with a medium molar mass (MW = 103 compounds were detected in extracts of samplers from at least one 300 g mol−1 ). sampling site. Specifically, 71 out of 134 compounds (52.9%) were found in POCIS and 32 out of 34 (94.1%) in SPMD extracts. 65 (38.7%) compounds never exceeded their LOQ. The concentrations for all individual 2.4.6.2. POCIS. Linear uptake from water during the entire sampling pe- compounds are listed in detail in SM2 in both the format ng POCIS−1 or riod was assumed to assess the concentrations of polar pesticides and ng SPMD−1 as well as recalculated based on the derived sampling rates pharmaceuticals dissolved in water from amounts accumulated in to ng L−1 (SM2 – Tables S2, S3). Summary results of the chemical anal- POCISs. Because the use of PRCs in POCIS is questionable and the varia- yses showing the number of detected compounds and their sum con- −1 tion of published RS data is related not only to compound physicochem- centrations at each site in pM and pmol sampler are reported in ical properties but also to differences in exposure conditions such as Tables 1 and SM2-Table S4, respectively. temperature, water flow rates, salinity, pH, and fouling, it is currently Concentrations of targeted compounds less than the LOQ were con- not possible to select unbiased substance specific Rs values in POCIS in sidered zero in the calculations of summary concentrations. Total sum a specific deployment situation (Harman et al., 2012). molar concentrations of all compounds (in pM) detected at each site Harman et al. (2012) reviewed all published POCIS calibration data in relation to the observed bioactivities expressed as contamination and the median values for all reported Rs values are 0.18 and 0.19 L profiles are presented in Fig. 2. There is a clear trend of decreasing cu- d−1 (stagnant and turbulent exposures, respectively). Therefore, we mulative concentration from S3 downstream to S10 in the POCIS sam- −1 applied a constant value of RS of 0.2 L d for all compounds as well ples, while no such pattern can be seen in case of hydrophobic as for bioassay responses. compounds determined in the SPMD. Most analyzed compounds were undetectable at the reference site S1 (spring of Bosna) with only a few compounds detected in concentrations near their LOQs. The major 2.5. Contamination profiling source of pollution to the Bosna River was Sarajevo, the capital with a population of about 300,000 (Milinovic, 2013). Detailed results of Toxicity profiles based on a set of in vitro bioassays were translated chemical analyses are listed in SM2-Table S2 and S3. into site-specific contamination profiles by use of an approach outlined by Hamers et al. (2010). The measured bioassay response of each ex- 3.1.1. Passive sampler performance characteristics tract was divided by the response measured in the extract from the upstream reference site S1, which was considered unaffected by 3.1.1.1. SPMD. Sampling rates for SPMDs, expressing the equivalent vol- anthropogenic pressures. The ratio between the response of down- ume of water from which a compound is extracted per day, were calcu- stream sites (S2-S10) and a reference site (S1), contamination index lated from in situ dissipation of PRCs. A graphical presentation of fitting

(CI), was regarded as a measure of contamination by each of the tested the PRC data to the model described in 2.4.6., as well as results of the RS endocrine effects. If no potency was detected at the reference site S1, calculation, are shown in Supplementary information, Section 1.3. Since one-half of the LOQ was used to calculate the CI. Contamination index RS weakly depends on molar mass, the Table S1 shows site speciﬁcsam- 1.0 was assigned for the effects less than the LOQ. pling rates for a model compound with a molar mass of 300 g mol−1 , 1604

Table 1 Tou Z. Sum of concentrations of target compounds detected in SPMD and POCIS extracts expressed in pM for different compound classes. Numbers of detected target compounds within each compound class at each sampling site are given in parentheses. š

Concentrations of target compounds less than the LOQ were considered zero in calculations of sums. “n.a.” stands for “not analyzed”. 1599 (2019) 650 Environment Total the of Science / al. et ová

Sampler Compound Total nr. of target S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 class compounds Spring of Bosna Sarajevo, DS Visoko, US Visoko, DS Lasva conﬂuence, US Zepce, US Maglaj, US Doboj, US Modrica, US Modrica, DS

SPMD PAHsb 15 7.2 (9) 2.3 × 102 (14) 2.5 × 102 (14) 1.9 × 102 (14) 1.3 × 102 (14) 2.8 × 102 (14) 1.9 × 102 (14) 3.3 × 102 (14) 2.3 × 102 (14) 2.2 × 102 (14) − − − − − − − − − − PCBs 7 7.3 × 10 2 (7) 4.7 × 10 1 (7) 6.2 × 10 1 (7) 6.3 × 10 1 (7) 4.7 × 10 1 (7) 2.5 × 10 1 (7) 3.4 × 10 1 (7) 3.6 × 10 1 (7) 2.6 × 10 1 (7) 3.8 × 10 1 (7) − − − − − − − − − − OCPs 12 3.8 × 10 2 (5) 2.4 × 10 1 (11) 1.7 × 10 1 (9) 6.5 × 10 1 (11) 3.8 × 10 1 (11) 2.9 × 10 1 (10) 1.9 × 10 1 (10) 2.0 × 10 1 (10) 1.6 × 10 1 (10) 2.2 × 10 1 (10) POCIS CUPs 52 2.9 (7) 2.4 × 102 (11a ) 1.8 × 102 (13) 1.8 × 102 (15) 2.1 × 102 (13) 2.5 × 102 (15) 2.6 × 102 (15) 2.2 × 102 (15) 1.3 × 102 (13) 1.2 × 102 (13) Estrogens 5 0 (0) 2.9 × 101 (4) 1.6 × 101 (4) 1.9 × 101 (4) 0 (0) 7.0 (4) 4.1 (4) 3.4 (4) 2.1 (3) 1.6 (3) Antibiotics 11 0 (0) n.a. 6.9 × 102 (6) 5.3 × 102 (7) 4.6 × 102 (6) 2.9 × 102 (6) 2.3 × 102 (6) 1.8 × 102 (6) 1.7 × 102 (6) 1.2 × 102 (6) − − − − Antidiabetics 2 0 (0) n.a. 1.6 (2) 1.6 (2) 8.1 × 10 1 (2) 6.5 × 10 1 (2) 6.7 × 10 1 (2) 0 (0) 0 (0) 1.6 × 10 1 (1) Antihistamins 6 0 (0) n.a. 1.9 × 101 (2) 2.0 × 101 (2) 7.7 (2) 8.1 (1) 8.3 (1) 6.4 (1) 5.4 (1) 2.3 (1) Cancer treatment 1 0 (0) n.a. 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) − Cardiovascular 10 1.2 × 10 1 (1) n.a. 3.9 × 102 (10) 2.6 × 102 (9) 2.3 × 102 (9) 1.9 × 102 (8) 1.4 × 102 (8) 1.2 × 102 (8) 9.7 × 101 (8) 7.2 × 101 (8) NSAIDS 1 0 (0) n.a. 2.1 × 102 (1) 2.8 × 102 (1) 1.6 × 102 (1) 1.2 × 102 (1) 9.4 × 101 (1) 6.7 × 101 (1) 5.2 × 101 (1) 3.4 × 101 (1) − Psychoactive 20 2.1 × 10 1 (1) n.a. 3.0 × 102 (12) 2.6 × 102 (11) 2.4 × 102 (11) 1.9 × 102 (11) 2.3 × 102 (10) 1.9 × 102 (6) 1.8 × 102 (7) 1.4 × 102 (7) − Statins 4 0 (0) n.a. 1.5 × 101 (4) 1.2 × 101 (3) 4.1 (3) 3.3 (2) 1.9 (1) 2.4 (1) 1.1 (1) 9.4 × 10 1 (1) Illicit drugs 8 0 (0) n.a. 9.9 (3) 8.6 (3) 1.0 × 101 (3) 6.8 (3) 5.2 (3) 3.2 (3) 1.4 (2) 1.3 (3) Metabolites 5 0 (0) n.a. 2.8 × 101 (4) 2.5 × 101 (4) 1.7 × 101 (4) 1.1 × 101 (4) 7.7 (3) 8.8 (4) 6.3 (3) 5.5 (4)

3 3 3 3 2 2 2 2 –

Others 9 2.9 (1) n.a. 2.7 × 10 (2) 1.1 × 10 (2) 1.2 × 10 (2) 1.2 × 10 (2) 4.0 × 10 (2) 3.8 × 10 (2) 4.1 × 10 (2) 1.8 × 10 (2) 1612

a At site S2, only 40 target CUPs were analyzed. b Concentration of naphthalene is not included in the reported sum of PAHs because of poor recoveries and its presence in blanks. Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612 1605

Fig. 2. Proﬁles of contamination at sampling sites S2-S10 combined with the total number and cumulative concentration (pM) of detected compounds in SPMD (top) and POCIS (bottom) extracts. Colors green to red indicate to what extent the bioassay responses exceed the response of the bioassay at reference site S1 (on a logarithmic scale). *77 pharmaceuticals and 16 target CUPs were not analyzed in the POCIS extract from site S2.

−1 Rs,300. Rs,300 values ranged from 5.7 to 10.5 L d . The volume of water lower than the variability of environmental concentrations, data ob- extracted for a compound that remains in the integrative uptake tained by passive sampling represent the contamination in the water phase during the entire sampling period ranged from 154 to 285 L. body equally or better than the low frequency spot sampling that is cur- Compound specific data treatment was not possible for interpreta- rently applied in regulatory monitoring of surface waters (Miège et al., tion of toxic potencies BEQ (ng/SPMD) of SPMD extracts, since the 2015). BEQ comprise many unknown substances. In order to translate toxic potencies to aqueous concentrations, we applied site specific Rs derived 3.1.2. Hydrophobic compounds from PRC data for a compound with MW = 300 (a compound with me- Most hydrophobic compounds were detected at all sampling sites dium molecular size) as a compromise. This is justifiable since Rspres- including site S1. PAHs occurred at the greatest concentrations (sum ent only a very weak function of MW (Rs = B × MW−0.47 ) and thus, of the 16 US EPA PAHs with the exception naphthalene 7.9–3.3 × 102 the uncertainty introduced by accepting an assumed MW of 300 for pM) compared to the other classes of hydrophobic compounds i.e. all compounds active in a bioassay is less than a factor 2, when assuming PCBs and OCPs (sum concentrations 3.8 × 10−2 -6.5 × 10−1 pM). Re- that active ingredients are within the range of MW from 200 to 700 (i.e. sults for naphthalene are not reported because of its poor extraction re- 700–0.47 /200–0.47 = 1.8), which is a typical range for xenobiotics that coveries and high levels found in blank samples. Within the class of are absorbed by SPMDs. PAHs, compounds with 3 and 4 condensed aromatic rings were detected at the greatest concentrations and occurred at all sampling 3.1.1.2. POCIS. Unlike the data from SPMDs, site specificsamplingrates sites. Concentrations of acenaphthene were 10-fold greater at site S6 cannot be derived for POCIS. Therefore, we decided to apply a constant and downstream compared with the sites upstream of S6. This indicates −1 RS value of 0.2 L d for all compounds as well as for bioassay responses, presence of a specific local pollution source for this compound, possibly acknowledging the resulting uncertainty of the reported data, which ul- the thermal power plant in Kakanj. Spatial concentration profiles of the timately renders them semi-quantitative. Despite this introduced un- remaining PAHs, PCBs and OCPs were less variable. Concentrations of certainty, passive sampling with POCIS provides time-integrated detected compounds were comparable with those sampled by SPMDs concentrations of pollutants, in contrast to spot sampling. If the uncer- during a survey in 2008 and 2009 when the river was screened for tainty of water concentrations estimated from passive sampling is Stockholm Convention persistent organic pollutants (Harman et al., 1606 Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612

2013). The comparison for selected compounds is shown in Supplemen- sampling sites. The greatest cumulative concentration of pharmaceuti- tary information, Section 1.4. Similar concentrations of dissolved hydro- cals was observed at site S3 (4.4 × 103 pM) and gradually decreased phobic contaminants collected by passive samplers in central European downstream to site S10. The same trend was observed for numbers of rivers were reported previously (Prokeš et al., 2012; Jálová et al., 2013 pharmaceuticals detected at individual sites. Concentrations of carba- and Vrana et al., 2014). Hydrophobic compounds adsorbed on mazepine and caffeine, indicator compounds of municipal waste suspended particulate matter were not addressed by the present water pressure (Buerge et al., 2003; Clara et al., 2004), were greatest study, but an additional load of these compounds on solid phase can at site S3 (42 and 4.9 × 102 ng L−1 , respectively) and differences be expected to contribute to the overall burden. between concentrations at site S3 and other sites were especially pronounced in case of caffeine. Concentrations of carbamazepine and 3.1.3. CUPs caffeine detected in the Danube River were an order of magnitude lesser Concentrations of most CUPs were less than the LOQ and at sites S6, than in the present study (Neale et al., 2015). Within the subclass of S7, and S8 with the greatest occurrence of CUPs, only 15 out of 52 com- antibiotics, 11 compounds were targeted and 6 were detected at 8 pounds were detected. In total, 20 CUPs (38.5%) were detected at least sampling sites (S3–S10). Sulfamethoxazole, trimethoprim and once. Carbendazim, diuron and isoproturon occurred at all sampling clarithromycin reached the greatest concentrations ranging from a sites including the reference site S1 and their concentrations ranged few to hundreds of ngL−1 . Similar concentrations of sulfamethoxazole from 6.0 × 10−2 –5.6 ng L−1 . The greatest concentrations were detected and clarithromycin in Sava river were reported by Tousova et al. in case of diazinon and prometryn (53 ng L−1 and 17 ng L−1 at sites S2 (2017). However, concentrations of trimethoprim were 10-fold less and S5, respectively) and these two compounds occurred at all sampling than in the present study. Cardiovascular drugs were frequently de- sites except for the reference site (S1). In the EU, both diazinon, a non- tected (8 out of 10 compounds were detected at 8 sites) with greatest systemic organophosphate insecticide, and prometryn, a systemic tri- concentrations observed for valsartan and atenolol (76 ng L−1 ,and azine herbicide, have been banned for use in plant protection products 21 ng L−1 , respectively), which have been reported to be ubiquitous since 2007 and 2002, respectively (European Commission, 2016). How- and persistent in aquatic environments (Ebele et al., 2017). Comparable ever, both these compounds along with other banned pesticides (e.g. at- concentrations of atenolol were detected in the Po and Lambro Rivers in razine, diuron, isoproturon) have been commonly detected in European Northern Italy (Calamari et al., 2003). Diclofenac, a nonsteroidal anti- surface waters despite the ban (Neale et al., 2015; Szekacs et al., 2015). inflammatory drug, was detected at 8 sampling sites (S3–S10) at con- Several studies have prioritized diazinon among the pesticides of major centrations ranging from 10 to 82 ng L−1 . Similar concentrations of concern in surface waters (Guha et al., 2016; Kuzmanovic et al., 2014a). diclofenac in river water and WWTP effluents were previously reported Greatest cumulative concentrations of CUPs were observed at sites S7, (Marsik et al., 2017 and Loos et al., 2013). However, a review paper re- S6 and S2 (2.6 × 102 pM, 2.5 × 102 pM, and 2.4 × 102 pM). Comparable ported a wide range of concentrations reaching μg·L−1 (Tiedeken et al., concentrations of CUPs determined in European surface waters and 2017). Concentrations of psychoactive drugs (n = 20) were rather small WWTP effluents were published previously (Jálová et al., 2013; Loos (hundreds of pg L−1 or single ngL−1 ) with the exception of carbamaz- et al., 2013; Neale et al., 2015 and Tousova et al., 2017). epine. Seven illicit drugs were determined, whereas cocaine, and its metabolite benzoylecgonine, and methadone were detected at 8 sites and 3.1.4. Estrogens 3,4-methylenedioxymethamphetamine (MDMA) was detected at 7 Natural estrogens E1, E2 and E3 were detected at all sites in a range sites. Similar levels of cocaine were reported in Czech and Welsh rivers of 2.0 × 10−2 –5.8 ng L−1 except for sites S1 and S5. Greatest concentra- by Fedorova et al. (2014) and Kasprzyk-Hordern et al. (2008), tions were observed at sites S2-S4. Concentrations of E3 at sites S2-S4 respectively. were 10-fold greater than those measured in the Danube River (Neale et al., 2015) or Sava River (Tousova et al., 2017), where neither E1, E2 3.2. In vitro bioassays nor EE2 were detected. EE2, a synthetic estrogen contained in hormonal contraceptives, was less than the LOQ of 1.0 × 10−2 –3.0 × 10−2 ng L−1 The responses of SPMD blank sample extracts did not significantly at all sampling sites. This might be related to the 8- to 9-fold lesser prev- differ from the response of solvent controls in all in vitro bioassays alence of hormonal contraceptives in BiH compared to western used in the present study. Minor estrogenic and dioxin-like activities European countries like Belgium or the UK due to cultural and economic were observed for POCIS blank sample extracts. Effect equivalents of reasons as well as limited availability of hormonal contraceptives these responses were subtracted from respective effect equivalents de- (Boussen, 2012). Pollution of European rivers with E2 and EE2 is a ubiq- tected in POCIS samples. None of the SPMD or POCIS extracts elicited cy- uitous phenomenon, however, analysis of these compounds still pre- totoxicity up to the greatest tested concentration in any of the used sents a challenge because LODs of most current monitoring techniques mammalian cell lines (0.5% v/v SPMD or POCIS extract mL−1 ). An over- are still greater than the proposed environmental quality standard view of in-vitro bioassay results in pg L−1 is given in Table 2 (results in EQS values (4 × 10−1 and 3.5 × 10−2 ng L−1 for E2 and EE2, respec- pmol L−1 are shown in SM2-Table S5), and more details are reported in tively) under the WFD (Tiedeken et al., 2017). This fact stresses the SM2-Table S6. REP values used for mass balance calculations were avail- need of routine application of bioanalytical tools for monitoring of estro- able for anti-androgenicity (12 compounds), estrogenicity (8 com- genic potency in surface waters, because bioassays can integrate the ef- pounds) and dioxin-like potency (16 compounds). No REPs were fect of multiple compounds contained in complex environmental available for androgenicity. Complete mass balance calculations, REPs mixtures and are more sensitive than most chemical analytical methods with literature references and resulting contributions of each com- (Jarošová et al., 2014; Kunz et al., 2017). LC-MS-MS is widely recognized pound to the observed potencies in bioassays are shown in SM2-Table as the most sensitive technique for the identification and quantification S7. of estrogens in environmental samples and the present study confirms its compliance with the WFD requirements as LODs for E2 was 2.0 3.2.1. AR-mediated potency ×10−3 ng L−1 and for EE2 ranged from 1.0 × 10−2 to 3.0 Androgenic potencies greater than LOQ (DHT-EQ. 1×10−2 – ×10−2 ng L−1 (Tiedeken et al., 2017). 3.5pgL−1 and 0.59–55 pg L−1 for SPMD and POCIS, respectively) were detected at sites S2 and S3. Concentrations of DHT-EQs detected 3.1.5. Pharmaceuticals in extracts of SPMDs were 10 pg L−1 (S2) and 4.2 pg L−1 (S3). Andro- Analyses of 77 pharmaceuticals in 11 subclasses resulted in detec- genic potency detected in POCIS extracts was almost two orders of mag- tion of 47 compounds (61%) at least once. The most frequently detected nitude greater (DHT-EQ. 1.7 × 103 and 2.1 × 102 pg L−1 at sites S2 and compounds, disopyramide, carbamazepine and caffeine, occurred at all S3, respectively). The DHT-EQ concentrations detected in POCIS extracts Table 2 −1 BEQbio and BEQchem values in pg L for SPMD and POCIS extracts determined in in vitro bioassays and chemical analyses. Percentage of effect that can be explained by the detected chemicals is given in parentheses. The uncertainty of BEQbio is expressed with standard deviation (n = 2). Tou Z.

Bioassay Sampler Sampling S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 š v ta./Sineo h oa niomn 5 21)1599 (2019) 650 Environment Total the of Science / al. et ová site Spring of Sarajevo, DS Visoko, US Visoko, DS Lasva conﬂuence, Zepce, US Maglaj, US Doboj, US Modrica, US Modrica, DS Bosna US

−1 −1 −1 −1 −1 −2 AR: DHT-EQ SPMD BEQbio b2.8 × 10 10 ± 2.8 4.2 ± 4.7 b3.5 b2.5 b1×10 b1.5 × 10 b1.6 × 10 b2.8 × 10 b1×10 −1 3 3 2 2 −1 POCIS BEQbio b5.9 × 10 1.7 × 10 ±1×10 2.1 × 10 ±1.8×10 b55 b53 b1.1 b9.5 × 10 b38 b41 b2.7

3 3 3 4 4 4 4 3 4 3 4 Anti-AR: SPMD BEQbio b4×10 b4.2 × 10 b3.6 × 10 5×10 ± 1.2 × 10 4.5 × 10 4.0 × 10 b5.6 × 10 3.4 × 10 b5.4 × 10 4.3 × 10 FLU-EQ BEQchem 3 20 20 20 (0.04%) 10 (0.03%) 20 (0.05%) 20 10 (0.03%) 10 20 (0.05%) 5 5 5 5 5 5 6 6 5 6 POCIS BEQbio b2.0 × 10 b2. 3 × 10 b2.1 × 10 b3.4 × 10 b3.3 × 10 b3.3 × 10 3.2 × 10 ± 3.1 × 10 ± b2.3 × 10 2.8 × 10 ± 9.9 × 105 6.4 × 105 7.2 × 105 2 3 3 3 3 3 3 3 3 3 BEQchem 2×10 4×10 3×10 2×10 1×10 2×10 2×10 (0.06%) 2 × 10 (0.06%) 1 × 10 1×10 (0.04%) −1 −1 −1 −1 −1 −1 −1 −1 −1 ER: E2-EQ SPMD BEQbio b3.7 × 10 b3.9 × 10 b7.8 × 10 b1.0 b2.7 × 10 b2.3 × 10 b1.6 × 10 b1×10 2 b2.7 × 10 b1.7 × 10 BEQchem n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 2 3 3 3 2 3 2 2 POCIS BEQbio b30 5.4 × 10 2.5 × 10 ± 1.1 × 10 ± 1.1 × 10 ± 9.2 × 10 ± 1.6 × 10 3.3 × 10 2.3 × 10 b70 2.3 × 103 1.2 × 102 8.8 × 102 8.7 × 102 3 2 2 2 2 2 2 2 BEQchem 10 1.7 × 10 (305%) 9.88 × 10 (40%) 8.7 × 10 (82%) 10 (0.84%) 3.9 × 10 (43%) 2.1 × 10 (14%) 1.9 × 10 (59%) 1.5 × 10 (64%) 1.0 × 10

−1 −1 −1 −1 −2 AhR: TCDD-EQ SPMD BEQbio b7.9 × 10 6.9 ± 1.4 5.9 ± 6.8 × 10 6.2 ± 5.5 × 10 5.2 ± 8.8 × 10 3.6 ± 1.7 2.9 ± 5 × 10 3.2 ± 1.2 × 7.3 ± 3.1 4.7 ± 2.2 − 10 1

−2 −1 −1 −1 −1 −1 −1 −1 −1 −1 –

BEQchem 7×10 5.6 × 10 (8.2%) 7.2 × 10 (12%) 5.6 × 10 (9.1%) 4.1 × 10 (7.8%) 6.1 × 10 (18%) 7.0 × 10 4.5 × 10 4.4 × 10 5.5 × 10 1612 (24%) (14%) (6.1%) (12%) 2 2 2 POCIS BEQbio b1.2 × 10 2.2 × 10 ±1.8×10 96 ± 53 b76 31 ± 5.2 82 ± 13 b12 b12 b28 b17 −3 −2 −2 −2 −2 −2 −2 −2 −2 −2 BEQchem 1×10 3.0 × 10 (0.02%) 2.0 × 10 (0.02%) 1.0 × 10 1.0 × 10 (0.04%) 1.0 × 10 7.0 × 10 6.0 × 10 4.0 × 10 4.0 × 10 (0.02%) 1607 1608 Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612 were consistent with the data reported in earlier studies addressing 3.2.3. AhR-mediated potency river water (König et al., 2017; Tousova et al., 2017)orWWTPeffluents AhR-mediated (dioxin-like) potency was detected in all SPMD ex- (Bain et al., 2014; Jalova et al., 2013). tracts except for site S1. The TCDD-EQ ranged from 2.9 to 7.3 pg L−1 Anti-androgenic potencies greater than the LOQ (FLU-EQ. 1.2 × 103 – (LOQ 0.4–0.84 pg L−1 ). These concentrations are consistent with previ- 5.6 × 103 pg L−1 ) were detected in extracts of SPMD at 5 sites (S4, S5, ously reported potencies measured in SPMD extracts of WWTP effluents S6, S8, S10) and FLU-EQ ranged between 3.4 × 104 –5.1 × 104 pg L−1 . and river water (Jálová et al., 2013). Potency balance calculations iden- Based on the potency balance, 0.03–0.05% of the response in the bioas- tified benzo[b]fluoranthene, benzo[k]fluoranthene and chrysene as the say could be explained and benzo[a]pyrene was identified as the major main drivers of the observed potencies, with detected compounds contributor (REP was available for only 2 compounds). In the case of ex- explaining 6.1–24% of the potencies measured in the CAFLUX assay. tracts of POCIS, 3 sites (S7, S8, S10) exhibited measurable potencies, REP values were available for 9 compounds. Similarly, 2–10% of the with FLU-EQs of 3.2 × 106 ,3.1×106 and 2.8 × 106 pg L−1 , respectively dioxin-like potency detected in SPMDs deployed in a drinking water (LOQ 1.4 × 105 –3.4 × 105 pg L−1 ). Approximately from 0.04–0.06% of reservoir in China could be explained by analysis of PAHs and PCBs the observed anti-androgenicity could be explained and diazinon was (Wang et al., 2014). According to their results, the contribution of the primary contributor (REP was available for 10 compounds). The PCBs to the overall potency was negligible, which is consistent with anti-androgenic potency of river waters determined by use of passive the findings in the study, results of which are reported here. POCIS ex- sampling was reported in earlier studies (Jálová et al., 2013; Liscio tracts showed AhR-mediated potency at 6 sites (S1, S2, S3, S4, S5 and et al., 2014). Anti-androgenic effects were observed more frequently S6) with TCDD-EQ. 31–2.2 × 102 pg L−1 (LOQ 12–1.2 × 102 pg L−1 ). in SPMDs, which is consistent with previous results (Creusotetal., The greater frequency of AhR mediated potency detected in extracts of 2013). Concentrations of anti-androgenicity measured in river SPMD compared to POCIS samplers was observed previously in an ear- water affected by untreated waste water, by use of large volume lier study (Creusot et al., 2013). Concentrations detected in POCIS ex- SPE and MDA-kb2 cells (König et al., 2017), were 10-fold less than tracts were as much as 10-fold greater than AhR-mediated potency the concentrations in extracts of POCIS observed during the present observed in extracts of POCIS, collected in headwaters in the Czech study. Similar to our results, only a minor portion of the observed Republic (Jarosova et al., 2012) or in large volume SPE samplers placed anti-androgenicity (up to 3%) could be explained by targeted in the Danube River (König et al., 2017; Neale et al., 2015). Only chemicals (König et al., 2017). When the yeast androgen screen 0.02–0.04% of the effect detected in POCIS could be explained and assay (YAS), was used to measure anti-androgenicity of river water propiconazole contributed most to potency measured in the bioassay, potency measured in extracts of SPMDs were almost 100-fold REPs were available for 7 compounds. In the Danube River, 3–71% of de- greater than those observed in extracts of POCIS (Liscio et al., 2014 tected AhR potency could be explained by three chemicals, daidzein, and Chen and Chou 2016). More than 31 compounds were identified terbuthylazine and carbaryl (Neale et al., 2015). Concentrations of that could have contributed to the observed anti-androgenicity terbuthylazine and carbaryl in extracts of the POCIS were less than (Liscio et al., 2014). These chemicals accounted for N50% of potency the LOQ or small. Daidzein, a natural isoflavone, was not analyzed observed in the bioassay. This list of compounds included several in our study. Significant amounts of AhR-mediated potency in river pharmaceuticals with confirmed anti-androgenic potency in the waters has been reported to be of anthropogenic origin, particularly YAS (i.e. carbamazepine, citalopram, codeine, diclofenac, diltiazem, from treated and untreated wastewater (Long and Bonefeld- irtesartan, trimethoprim and venlafaxine), which were also detected Jørgensen, 2012). Potential contributors could include also polar de- in our study and could therefore possibly contribute to the anti- rivatives of PAHs, some pharmaceuticals, CUPs or other weakly or androgenic potency (REPs for MDA-kb2 not available). The main moderately polar natural compounds. A larger list of target com- drivers of anti-androgenicity were not identified since the major pounds associated with REP values of particular compounds is portion of the effect could not be explained by target chemicals. needed for successful identification of the main AhR-mediated potency drivers.

3.2.2. ER-mediated potency 3.3. Contamination profiling No estrogenic potency was detected at concentrations greater than the LOQ (E2-EQ. 0.12–1pgL−1 ) in any of the SPMD extracts. Aprofile of integrated effects of mixtures at various locations in the POCIS extracts elicited estrogenic potency above the LOQ (E2-EQ. Bosna, based on potencies observed in the three bioassays (Table 2)for 30–1.1 × 102 pg L−1 ) at 8 sampling sites (S2–S9) and the E2-EQ extracts of the two types of sampler at each location were developed ranged from 2.3 × 102 to 2.5 × 103 pg L−1 .S3wasthesitewith based on comparison to the reference site (S1), which was defined as greatest estrogenic potency. These results are consistent with those the location least affected by direct and indirect inputs from human ac- reported for surface waters during previous studies (Jalova et al., tivities, including urbanization and industrialization. Contamination in- 2013; Jarosova et al., 2012; Jugan et al., 2009; Tousova et al., 2017). dices (CI), the ratio between the response of downstream sites (S2–S10) Potency balance calculations revealed that 0.84–305% of the and a reference site (S1), in combination with the overall cumulative estrogenicity could be explained by target compounds (REP was concentration and number of detected target compounds for each sam- available for 8 compounds), dominated by natural estrogens (E1, pling site are shown in Fig. 2. None of the sampling sites downstream of E2 and E3). This is in line with earlier findings of Miège et al. S1 can be considered as uncontaminated as the reference site because (2009), who identified estrogens as the main drivers of estrogenicity all sites exceeded a CI of 1.0 for at least two endpoints. The CI profiles, in river water. Neale et al. (2015), who assessed estrogenicity in 22 as well as the cumulative concentration and the number of detected hy- river water samples, reported that 0.31 to 61% of the observed po- drophobic compounds, differed less between individual sites in extracts tency could be explained by target chemicals and that E1 along of SPMD than in extracts of POCIS. Extracts of POCIS from S1 elicited with a phytoestrogen genistein were the main drivers. At site S2 (Sa- dioxin-like potency, which exceeded the response in the extract from rajevo DS), with BEQchem based on detected estrogens exceeding the site S5, which resulted in a CI of 0.36. Contamination indices indicate BEQbio by N300%, antiestrogenic effects might be present and mask that the most contaminated sites were S2 and S3. The greatest cumula- part of the estrogenic potency elicited by the detected estrogens. tive concentrations and numbers of detected compounds were Antiestrogenic effects have been commonly detected in river waters observed for the extract of the POCIS at S3. A complete analysis for and they are believed to result from the combined action of a multi- S2 was not available. This result implies that the city of Sarajevo consti- tude of chemicals present in complex environmental mixtures tuted the major source of contaminants relevant for the observed AhR- (Gehrmann et al., 2016; Oh et al., 2006). mediated potency. The trend of decreasing cumulative concentrations Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612 1609 in extracts of POCIS samplers downstream of S3 cannot be clearly seen 3.5. Limitations of the research and its environmental implications in CIs and no patterns between CIs and cumulative concentrations in extracts of SPMD were observed, despite extensive, multi-residue analy- Beside advantages of passive sampling techniques compared to grab ses. Observed potencies could not be assigned to specific compounds, sampling techniques such as time integrative sampling of bioavailable which is a common case for complex environmental matrices (Weller, contaminants and lower achievable detection limits, passive sampling 2012). Novel approaches proposed for future monitoring schemes, in- also suffers from several limitations. As mentioned earlier, toxic poten- cluding a combination of non-target identification of chemicals (Peng cies measured in passive samplers can be translated into equivalent et al., 2016), screening of effects (Sun et al., 2017, 2016), mixture toxic- toxic potencies in water only when making assumptions of fully inte- ity modelling and effect-directed analysis (Altenburger et al., 2015) will grative uptake of all compounds present in the sampled mixtures and facilitate identification of compounds responsible for adverse effects in when calculations are done with averaged sampling rates over a broad aquatic ecosystems. range of compound properties. These approximations are necessary since the identity of compounds causing the observed effects remains largely unknown. The application of models for improvement of mea- 3.4. Hazard assessment surement accuracy that relate sampling parameters to physicochemical compound properties is thus precluded. The approximation of sampling Hazard quotients could not be calculated for all compounds. parameters is ultimately associated with an increased uncertainty of re- NORMAN lowest PNEC values were available for 167 of 168 target ported data. The uncertainty of SPMD-derived aqueous concentrations compounds (99.4%). Concentrations of 7 compounds exceeded is generally lower than that of POCIS data, since for SPMD site specific PNECs at least at 2 sampling sites (Table 3). Hazard quotient (HQ) sampling rates can be derived using PRC approach, and sampling rates of the insecticide, diazinon, exceeded 1.0 at all 9 sites, where it was of nonpolar compounds sampled by SPMDs only weakly depend on mo- detected and its HQs, which were as great as 4.4, were greatest of lecular structure (Lohmann et al., 2012). In our study, only a limited set all compounds. HQs of diclofenac, a NSAID, and two estrogens, E1 of five deuterated PAHs as PRCs was applied for estimation of SPMD and E2 2, exceeded HQ of 1.0 at 2 sites. In this study, the PNEC for sampling rates. The accuracy of estimation can be improved by applica- EE2 was less than its LOQ so no HQ could be calculated. Therefore, tion of a broader range of PRC compounds, as has been shown by Booij EE2 might still pose a potential risk to aquatic biota even though it and Smedes (2010). POCIS data has to be considered semi-quantitative was never detected. In the class of PAHs, HQs of benzo[b]fluoran- since the uptake mechanism of polar compounds from water is not fully thene exceeded 1.0 at 6 sites and those of fluoranthene and benzo understood, PRC approach cannot be applied for in situ sampling rate [k]fluoranthene exceeded 1.0 at 2 sites. Complete results of hazard correction, and also a larger variability of sampling rates on physico- assessment and a list of the lowest NORMAN PNEC values are chemical compound properties and environmental factors (water shown in SM2- Table S8. In a previous study that applied a similar flow, pH, temperature) is expected than for SPMDs (Miège et al., methodology of assessment of samples collected in 4 European 2015). The application of a single constant POCIS sampling rate value river basins, diazinon, diclofenac and fluoranthene were also identi- of 0.2 L d−1 for all compounds was thus chosen as a compromise in a sit- fied and prioritized as most hazardous (Tousova et al., 2017). uation when the effect of environmental variables and compound prop- Diazinon was ranked among the most hazardous compounds in sev- erties on sampling rate could not be fully controlled or quantified. The eral Iberian, North European and US rivers (Kuzmanovic et al., 2014; elevated measurement uncertainty can be accepted as long as it is von der Ohe et al., 2011). Diclofenac was identified as a driver of haz- lower than the variability of environmental concentrations, which ard in Greek rivers (Thomaidi et al., 2015 and Kosma et al., 2014). may be dramatic in dynamic river systems such as the Bosna river inves- Those authors also found several compounds from the class of antibi- tigated in our study. The obtained semi-quantitative data cannot be, in otics, including sulfamethoxazole, trimethoprim and clarithromycin general, directly applied for checking compliance with environmental to exceed HQs of 1.0 to aquatic biota. These antibiotics were also de- quality criteria, however, they are very suitable for screening of areas tected in the study, results of which are presented here. However, and pollutants of concern and identification of areas, where a focused their concentrations did not exceed their, respective PNECs. Of monitoring can be performed at a later stage using conventional moni- PAHs, benzo[k]fluoranthene (Smital et al., 2013)andfluoranthene toring methods. (Von der Ohe et al., 2011) have been prioritized previously the most hazardous. The overall hazard index (HI), resulting from the 4. Conclusions summation of all HQs at each sampling site, indicates that all sites downstream of the reference site S1 might cause adverse effects to The study which assessed water quality of the Bosna River found aquatic biota as their HIs exceed 1.0. concentrations of contaminants or observed potency of mixtures that

Table 3 Target compounds with hazard quotient (HQ) values exceeding 1 (in bold) and overall hazard index (HI) at individual sampling sites.

Compound PNEC [ng S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 Frequency name L−1 ] of PNEC Spring of Sarajevo, Visoko, Visoko, Lasva Zepce, Maglaj, Doboj, Modrica, Modrica, exceedance Bosna DS US DS conﬂuence, US US US US DS [%] US

17β-estradiol 4.0 × 10−1 bLOD 2.0 1.1 7.5 × 10−1 bLOD 3.0 × 10−1 1.8 × 10−1 1.5 × 10−1 1.3 × 10−1 1.0 × 10−1 20 Benzo[b] 1.7 × 10−1 bLOD 1.2 1.3 1.2 8.5 × 10−1 1.2 1.6 9.0 × 10−1 9.0 × 10−1 1.7 60 fluoranthene Benzo[k] 1.7 × 10−1 bLOD 8.8 × 10−1 1.2 8.6 × 10−1 6.0 × 10−1 9.7 × 10−1 1.1 6.8 × 10−1 7.1 × 10−1 7.4 × 10−1 20 fluoranthene Diazinon 1.2 × 10+1 bLOD 4.4 2.2 2.2 2.5 3.7 3.6 2.7 1.6 1.6 90 Diclofenac 5.0 × 10+1 bLOD n.a. 1.3 1.6 9.3 × 10−1 6.9 × 10−1 5.6 × 10−1 4.0 × 10−1 3.1 × 10−1 2.0 × 10−1 22 Estrone 3.6 × 100 bLOD 1.6 7.0 × 10−1 1.1 bLOD 2.3 × 10−1 1.4 × 10−1 1.3 × 10−1 1.0 × 10−1 8.1 × 10−2 20 Fluoranthene 6.3 × 100 1.3 × 10−2 9.1 × 10−1 9.3 × 10−1 6.7 × 10−1 5.0 × 10−1 9.7 × 10−1 9.5 × 10−1 1.2 8.5 × 10−1 1.4 20 Total hazard 3.9E-01 1.5E + 01 1.5E + 01 1.4E + 01 9.7E + 00 1.3E + 01 1.3E + 01 1.0E + 01 7.9E + 00 1.0E + 01 index 1610 Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612 did not differ significantly from those measured in other European rivers, Monitoring. Elsevier, Amsterdam, pp. 141–169 https://doi.org/10.1016/S0166-526X (06)48007-7. even in those areas with more advanced wastewater treatment technolo- Boussen, C., 2012. Key factors influencing contraceptive use in Eastern Europe and Central gies and infrastructure. Chemical analyses revealed frequent occurrences Asia [WWW Document]. https://eeca.unfpa.org/en/publications/key-factors- of pesticides, which were banned for use in plant protection products in influencing-contraceptive-use-eastern-europe-and-central-asia, Accessed date: 3 September 2018. the EU, e.g. diazinon, carbendazim, isoproturon, diuron, prometryn, Brumovský, M., Bečanová, J., Kohoutek, J., Thomas, H., Petersen, W., Sørensen, K., Sáňka, metolachlor. Diazinon occurred at most sites at concentrations which O., Nizzetto, L., 2016. Exploring the occurrence and distribution of contaminants of might cause adverse effects on aquatic biota. These compounds are of emerging concern through unmanned sampling from ships of opportunity in the – concern and should be included in regular monitoring and possible North Sea. J. Mar. Syst. 162, 47 56. https://doi.org/10.1016/j.jmarsys.2016.03.004. Buerge, I.J., Poiger, T., Buser, H., Wa, C., 2003. Caffeine, An Anthropogenic Marker for mitigation measures. With the exception of estrogenicity, potencies of Wastewater Contamination of Surface Waters. 37, pp. 691–700. https://doi.org/ endocrine effects observed in vitro, by use of bioassays, could not be 10.1021/es020125z. explained by targeted compounds. Natural estrogens were found to be Calamari, D., Zuccato, E., Castiglioni, S., Bagnati, R., Fanelli, R., 2003. Strategic survey of therapeutic drugs in the rivers Po and lambro in northern Italy. Environ. Sci. Technol. largely responsible for the observed estrogenic potency. These results em- 37, 1241–1248. https://doi.org/10.1021/es020158e. phasize the need to apply bioassays as a complementary tool in routine Chen, K.Y., Chou, P.H., 2016. Detection of endocrine active substances in the aquatic envi- monitoring of water quality, because chemical analysis alone cannot ronment in southern Taiwan using bioassays and LC-MS/MS. Chemosphere 152, 214–220. https://doi.org/10.1016/j.chemosphere.2016.02.115. indicate effects elicited by mixtures of compounds occurring at small Clara, M., Strenn, B., Kreuzinger, N., 2004. Carbamazepine as a Possible Anthropogenic concentrations in environmental mixtures. It should be further enhanced Marker in the Aquatic Environment: Investigations on the Behaviour of Carbamazepine in future studies by application of progressive approaches combining in Wastewater Treatment and during Groundwater Infiltration. 38, pp. 947–954. https://doi.org/10.1016/j.watres.2003.10.058. effect-based screening, modelling of mixtures responses and effect- Connon, R.E., Geist, J., Werner, I., 2012. Effect-based tools for monitoring and predicting directed analysis together with non-target identification of chemicals to the ecotoxicological effects of chemicals in the aquatic environment. Sensors 12, prioritize the most relevant toxicants and effect drivers. 12741–12771. https://doi.org/10.3390/s120912741 (Switzerland). Creusot, N., Tapie, N., Piccini, B., Balaguer, P., Porcher, J.M., Budzinski, H., Aït-Aïssa, S., 2013. Distribution of steroid- and dioxin-like activities between sediments, POCIS and Acknowledgements SPMD in a French river subject to mixed pressures. Environ. Sci. Pollut. Res. 20, 2784–2794. https://doi.org/10.1007/s11356-012-1452-5. Demirpence, E., Duchesne, M.-J., Badia, E., Gagne, D., Pons, M., 1993. MVLN cells: a biolu- The study was supported by the EDA-EMERGE project (FP7-PEOPLE- minescent MCF-7-derived cell line to study the modulation of estrogenic activity. – 2011-ITN, grant agreement number 290100), SOLUTIONS project J. Steroid Biochem. Mol. Biol. 46, 355 364. https://doi.org/10.1016/0960-0760(93) 90225-L. funded by the European Union Seventh Framework Programme (FP7, Dulio, V., van Bavel, B., Brorström-Lundén, E., Harmsen, J., Hollender, J., Schlabach, M., grant agreement no. 603437) and NATO ESP.EAP.SFP 984073 project. Slobodnik, J., Thomas, K., Koschorreck, J., 2018. Emerging pollutants in the EU: The authors thank Simone Milanolo and Melina Džajić-Valjevac from 10 years of NORMAN in support of environmental policies and regulations. Environ. Sci. Eur. 30, 5. https://doi.org/10.1186/s12302-018-0135-3. the Hydro-Engineering Institute, Sarajevo, Bosnia and Herzegovina for Ebele, A.J., Abou-Elwafa Abdallah, M., Harrad, S., 2017. Pharmaceuticals and personal care deployment and retrieval of passive samplers, Alena Otoupalíková and products (PPCPs) in the freshwater aquatic environment. Emerg. Contam. 3, 1–16. Jiří Kohoutek from Masaryk University for conducting the chemical https://doi.org/10.1016/j.emcon.2016.12.004. Emelogu, E.S., Pollard, P., Dymond, P., Robinson, C.D., Webster, L., McKenzie, C., Dobson, J., analyses. The authors thank South Bohemian Research Center of Bresnan, E., Moffat, C.F., 2013. Occurrence and potential combined toxicity of dis- Aquaculture and Biodiversity of Hydrocenoses (CENAKVA) for analysis solved organic contaminants in the Forth estuary and Firth of Forth, Scotland of pharmaceuticals. Prof. Giesy was supported by the Canada assessed using passive samplers and an algal toxicity test. Sci. Total Environ. “ ” 461–462, 230–239. https://doi.org/10.1016/j.scitotenv.2013.05.011. Research Chair program, the 2012 High Level Foreign Experts Escher, B.I., Leusch, F., 2012. Bioanalytical Tools in Water Quality Assessment. IWA Pub- (#GDT20143200016) program, funded by the State Administration of lishing https://doi.org/10.1002/ieam.1340. Foreign Experts Affairs, the P.R. China to Nanjing University and the Escher, B.I., Allinson, M., Altenburger, R., Bain, P.A., Balaguer, P., Busch, W., Crago, J., Einstein Professor Program of the Chinese Academy of Sciences and a Denslow, N.D., Dopp, E., Hilscherova, K., Humpage, A.R., Kumar, A., Grimaldi, M., Jayasinghe, B.S., Jarosova, B., Jia, A., Makarov, S., Maruya, K.A., Medvedev, A., Distinguished Visiting Professorship in the School of Biological Sciences Mehinto, A.C., Mendez, J.E., Poulsen, A., Prochazka, E., Richard, J., Schifferli, A., of the University of Hong Kong. Schlenk, D., Scholz, S., Shiraishi, F., Snyder, S., Su, G., Tang, J.Y.M., van der Burg, B., van der Linden, S.C., Werner, I., Westerheide, S.D., Wong, C.K.C., Yang, M., Yeung, B.H.Y., Zhang, X., Leusch, F.D.L., 2014. Benchmarking organic micropollutants in Appendix A. Supplementary data wastewater, recycled water and drinking water with in vitro bioassays. Environ. Sci. Technol. 48, 1940–1956. https://doi.org/10.1021/es403899t. Supplementary data to this article can be found online at https://doi. European Commission, 2016. EU - Pesticides Database (WWW Document). Fedorova, G., Randak, T., Lindberg, R.H., Grabic, R., 2013. Comparison of the quantitative org/10.1016/j.scitotenv.2018.08.336. performance of a Q-Exactive high-resolution mass spectrometer with that of a triple quadrupole tandem mass spectrometer for the analysis of illicit drugs in wastewater. Rapid Commun. Mass Spectrom. 27, 1751–1762. https://doi.org/10.1002/rcm.6628. References Fedorova, G., Randak, T., Golovko, O., Kodes, V., Grabicova, K., Grabic, R., 2014. A passive sampling method for detecting analgesics, psycholeptics, antidepressants and illicit Altenburger,R.,Ait-Aissa,S.,Antczak,P.,Backhaus,T.,Barceló,D.,Seiler,T.-B.,Brion,F.,Busch, drugs in aquatic environments in the Czech Republic. Sci. Total Environ. 487, W., Chipman, K., de Alda, M.L., de Aragão Umbuzeiro, G., Escher, B.I., Falciani, F., Faust, 681–687. https://doi.org/10.1016/j.scitotenv.2013.12.091. M.,Focks,A.,Hilscherova,K.,Hollender,J.,Hollert,H.,Jäger,F.,Jahnke,A.,Kortenkamp, Gehrmann, L., Bielak, H., Behr, M., Itzel, F., Lyko, S., Simon, A., Kunze, G., Dopp, E., Wagner, A.,Krauss,M.,Lemkine,G.F.,Munthe,J.,Neumann,S.,Schymanski,E.L.,Scrimshaw,M., M., Tuerk, J., 2016. (Anti-)estrogenic and (anti-)androgenic effects in wastewater Segner, H., Slobodnik, J., Smedes, F., Kughathas, S., Teodorovic, I., Tindall, A.J., Tollefsen, during advanced treatment: comparison of three in vitro bioassays. Environ. Sci. K.E., Walz, K.-H., Williams, T.D., Van den Brink, P.J., van Gils, J., Vrana, B., Zhang, X., Pollut. Res., 1–11 https://doi.org/10.1007/s11356-016-7165-4. Brack, W., 2015. Future water quality monitoring - adapting tools to deal with mixtures Geissen, V., Mol, H., Klumpp, E., Umlauf, G., Nadal, M., van der Ploeg, M., van de Zee, S.E.a. of pollutants in water resource management. Sci. Total Environ. 512–513C, 540–551. T.M., Ritsema, C.J., 2015. Emerging pollutants in the environment: a challenge for https://doi.org/10.1016/j.scitotenv.2014.12.057. water resource management. Int. Soil Water Conserv. Res. 3, 57–65. https://doi.org/ Alvarez, D.A., Petty, J.D., Huckins, J.N., Jones-Lepp, T.L., Getting, D.T., Goddard, J.P., 10.1016/j.iswcr.2015.03.002. Manahan, S.E., 2004. Development of a passive, in situ, integrative sampler for hydro- Ginebreda, A., Kuzmanovic, M., Guasch, H., de Alda, M.L., López-Doval, J.C., Muñoz, I., philic organic contaminants in aquatic environments. Environ. Toxicol. Chem. 23, Ricart, M., Romaní, A.M., Sabater, S., Barceló, D., 2014. Assessment of multi-chemical 1640. https://doi.org/10.1897/03-603. pollution in aquatic ecosystems using toxic units: compound prioritization, mixture Bain, P., Williams, M., Kumar, A., 2014. Assessment of multiple hormonal activities in characterization and relationships with biological descriptors. Sci. Total Environ. wastewater at different stages of treatment. Environ. Toxicol. Chem. 33, 2297–2307. 468–469, 715–723. https://doi.org/10.1016/j.scitotenv.2013.08.086. https://doi.org/10.1002/etc.2676. Grabic, R., Fick, J., Lindberg, R.H., Fedorova, G., Tysklind, M., 2012. Multi-residue method Booij, K., Smedes, F., 2010. An improved method for estimating in situ sampling rates of for trace level determination of pharmaceuticals in environmental samples using liq- nonpolar passive samplers. Environ. Sci. Technol. 44, 6789–6794. https://doi.org/ uid chromatography coupled to triple quadrupole mass spectrometry. Talanta 100, 10.1021/es101321v. 183–195. https://doi.org/10.1016/j.talanta.2012.08.032. Booij, K., Vrana, B., Huckins, J.N., 2007. Theory, modelling and calibration of passive sam- Guha, N., Guyton, K.Z., Loomis, D., Barupal, D.K., 2016. Prioritizing chemicals for risk as- plers used in water monitoring. In: Greenwood, R., Mills, G., Vrana, B. (Eds.), Compre- sessment using chemoinformatics: examples from the IARC onographs on pesticides. hensive Analytical Chemistry 48. Passive Sampling Techniques in Environmental Environ. Health Perspect. 124, 1823–1829. https://doi.org/10.1289/EHP186. Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612 1611

Hamers, T., Leonards, P.E.G., Legler, J., Dick Vethaak, A., Schipper, C.A., 2010. Toxicity Marsik, P., Rezek, J., Židková, M., Kramulová, B., Tauchen, J., Vaněk, T., 2017. Non-steroidal profiling: an integrated effect-based tool for site-specific sediment quality as- anti-inflammatory drugs in the watercourses of Elbe basin in Czech Republic. sessment. Integr. Environ. Assess. Manag. 6, 761–773. https://doi.org/10.1002/ Chemosphere 171, 97–105. https://doi.org/10.1016/j.chemosphere.2016.12.055. ieam.75. Miège, C., Gabet, V., Coquery, M., Karolak, S., Jugan, M.L., Oziol, L., Levi, Y., Chevreuil, M., Harman, C., Allan, I.J., Vermeirssen, E.L.M., 2012. Calibration and use of the polar organic 2009. Evaluation of estrogenic disrupting potency in aquatic environments and chemical integrative sampler-a critical review. Environ. Toxicol. Chem. 31, urban wastewaters by combining chemical and biological analysis. TrAC Trends 2724–2738. https://doi.org/10.1002/etc.2011. Anal. Chem. 28, 186–195. https://doi.org/10.1016/j.trac.2008.11.007. Harman, C., Grung, M., Djedjibegovic, J., Marjanovic, A., 2013. Screening for Stockholm Miège, C., Mazzella, N., Allan, I., Dulio, V., Smedes, F., Tixier, C., Vermeirssen, E., Brant, J., convention persistent organic pollutants in the Bosna River (Bosnia and O'Toole, S., Budzinski, H., Ghestem, J.-P., Staub, P.-F., Lardy-Fontan, S., Gonzalez, J.-L., Herzogovina). Environ. Monit. Assess. 185, 1671–1683. https://doi.org/10.1007/ Coquery, M., Vrana, B., 2015. Position paper on passive sampling techniques for the s10661-012-2659-0. monitoring of contaminants in the aquatic environment - achievements to date Huckins, J.N., Manuweera, G.K., Petty, J.D., Mackay, D., Lebo, J.A., 1993. Lipid-containing and perspectives. Trends Environ. Anal. Chem. 8. https://doi.org/10.1016/j. semipermeable membrane devices for monitoring organic contaminants in water. teac.2015.07.001. Environ. Sci. Technol. 27, 2489–2496. Milinovic, Z., 2013. Preliminary Results Of the 2013 Census of Population, Households and Huckins, J.N., Petty, J.D., Lebo, J.A., Almeida, F.V., Booij, K., Alvarez, D.A., Cranor, W.L., Clark, Dwellings in Bosnia and Herzegovina [WWW Document]. Agency Stat, Bosnia R.C., Mogensen, B.B., 2002. Development of the permeability/performance reference Herzegovina. compound (PRC) approach for in situ calibration of semipermeable membrane de- Moschet, C., Vermeirssen, E.L.M., Seiz, R., Pfefferli, H., Hollender, J., 2014. Picogram per vices (SPMDs). Environ. Sci. Technol. 36, 85–91. liter detections of pyrethroids and organophosphates in surface waters using passive Jalova, V., Jarosova, B., Blaha, L., Giesy, J.P.P., Ocelka, T., Grabic, R., Jurcikova, J., Vrana, B., sampling. Water Res. 66, 411–422. https://doi.org/10.1016/j.watres.2014.08.032. Hilscherova, K., Jálová, V., Jarošová, B., Bláha, L., Giesy, J.P.P., Ocelka, T., Grabic, R., Nagy, S.R., Sanborn, J.R., Hammock, B.D., Denison, M.S., 2002. Development of a green Jurčíková, J., Vrana, B., Hilscherová, K., 2013. Estrogen-, androgen- and aryl hydrocar- fluorescent protein-based cell bioassay for the rapid and inexpensive detection and bon receptor mediated activities in passive and composite samples from municipal characterization of Ah receptor agonists. Toxicol. Sci. 65, 200–210. https://doi.org/ waste and surface waters. Environ. Int. 59, 372–383. https://doi.org/10.1016/j. 10.1093/toxsci/65.2.200. envint.2013.06.024. Neale, P.A., Ait-Aissa, S., Brack, W., Creusot, N., Denison, M.S., Deutschmann, B., Jarosova, B., Blaha, L., Vrana, B., Randak, T., Grabic, R., Giesy, J.P., Hilscherova, K., 2012. Hilscherova, K., Hollert, H., Krauss, M., Novák, J., Schulze, T., Seiler, T.B., Serra, H., Changes in concentrations of hydrophilic organic contaminants and of endocrine- Shao, Y., Escher, B.I., 2015. Linking in vitro effects and detected organic disrupting potential downstream of small communities located adjacent to headwa- micropollutants in surface water using mixture toxicity modeling. Environ. Sci. ters.Environ.Int.45,22–31. Technol. 49, 14614–14624. https://doi.org/10.1021/acs.est.5b04083. Jarošová, B., Erseková, A., Hilscherová, K., Loos, R., Gawlik, B.M., Giesy, J.P., Bláha, L., 2014. Oh, S.M., Park, K., Chung, K.H., 2006. Combination of in vitro bioassays encompassing dif- Europe-wide survey of estrogenicity in wastewater treatment plant effluents: ferent mechanisms to determine the endocrine-disrupting effects of river water. Sci. the need for the effect-based monitoring. Environ. Sci. Pollut. Res. Int. 21, Total Environ. 354, 252–264. https://doi.org/10.1016/j.scitotenv.2005.01.041. 10970–10982. https://doi.org/10.1007/s11356-014-3056-8. Pavlíková, N., Bláhová, L., Klán, P., Bathula, S.R., Sklenář, V., Giesy, J.P., Bláha, L., 2012. Jones, L., Ronan, J., McHugh, B., McGovern, E., Regan, F., 2015. Emerging priority sub- Enantioselective effects of alpha-hexachlorocyclohexane (HCH) isomers on androgen stances in the aquatic environment: a role for passive sampling in supporting WFD receptor activity in vitro. Chemosphere 86, 65–69. https://doi.org/10.1016/j. monitoring and compliance. Anal. Methods 7, 7976–7984. https://doi.org/10.1039/ chemosphere.2011.08.052. C5AY01059D. Peng, H., Chen, C., Cantin, J., Saunders, D.M.V., Sun, J., Tang, S., Codling, G., Hecker, M., Jugan, M.L., Oziol, L., Bimbot, M., Huteau, V., Tamisier-Karolak, S., Blondeau, J.P., Lévi, Y., 2009. Wiseman, S., Jones, P.D., Li, A., Rockne, K.J., Sturchio, N.C., Cai, M., Giesy, J.P., 2016. In vitro assessment of thyroid and estrogenic endocrine disruptors in wastewater treat- Untargeted screening and distribution of organo-iodine compounds in sediments ment plants, rivers and drinking water supplies in the greater Paris area (France). Sci. from Lake Michigan and the Arctic Ocean. Environ. Sci. Technol. 50, 10097–10105. Total Environ. 407, 3579–3587. https://doi.org/10.1016/j.scitotenv.2009.01.027. https://doi.org/10.1021/acs.est.6b03221. Kasprzyk-Hordern, B., Dinsdale, R.M., Guwy, A.J., 2008. The occurrence of pharmaceuti- Prokeš, R., Vrana, B., Klánová, J., 2012. Levels and distribution of dissolved hydrophobic or- cals, personal care products, endocrine disruptors and illicit drugs in surface water ganic contaminants in the Morava river in Zlín district, Czech Republic as derived in South Wales, UK. Water Res. 42, 3498–3518. https://doi.org/10.1016/j. from their accumulation in silicone rubber passive samplers. Environ. Pollut. 166, watres.2008.04.026. 157–166. https://doi.org/10.1016/j.envpol.2012.02.022. Könemann, S., Kase, R., Simon, E., Swart, K., Buchinger, S., Schlüsener, M., Hollert, H., Reichenberg, F., Mayer, P., 2006. Two complementary sides of bioavailability: accessibility Escher, B.I., Werner, I., Aït-Aïssa, S., Vermeirssen, E., Dulio, V., Valsecchi, S., Polesello, and chemical activity of organic contaminants in sediments and soils. Environ. S., Behnisch, P., Javurkova, B., Perceval, O., Di Paolo, C., Olbrich, D., Sychrova, E., Toxicol. Chem. 25, 1239–1245. Schlichting, R., Leborgne, L., Clara, M., Scheffknecht, C., Marneffe, Y., Chalon, C., Rusina, T., Smedes, F., Koblizkova, M., Klanova, J., 2010. Calibration of silicone rubber pas- Tušil, P., Soldàn, P., von Danwitz, B., Schwaiger, J., San Martín Becares, M.I., Bersani, sive samplers: experimental and modeled relations between sampling rate and com- F., Hilscherová, K., Reifferscheid, G., Ternes, T., Carere, M., 2018. Effect-based and pound properties. Environ. Sci. Technol. 44, 362–367. chemical analytical methods to monitor estrogens under the European water frame- Schirmer, K., Chan, A.G.J., Greenberg, B.M., Dixon, D.G., Bols, N.C., 1998. Ability of 16 prior- work directive. TrAC Trends Anal. Chem. https://doi.org/10.1016/J.TRAC.2018.02.008. ity PAHs to be photocytotoxic to a cell line from the rainbow trout gill. Toxicology König, M., Escher, B.I., Neale, P.A., Krauss, M., Hilscherová, K., Novák, J., Teodorović, I., 127, 143–155. https://doi.org/10.1016/S0300-483X(98)00031-6. Schulze, T., Seidensticker, S., Kamal Hashmi, M.A., Ahlheim, J., Brack, W., 2017. Impact Sima, L., Amador, J., Silva, A.D.K., Miller, S.M., Morse, A.N., Pellegrin, M.-L., Rock, C., Wells, of untreated wastewater on a major European river evaluated with a combination of M.J.M., 2014. Emerging pollutants – part I: occurrence, fate and transport. Water En- in vitro bioassays and chemical analysis. Environ. Pollut. 220, 1220–1230. https://doi. viron. Res. 86, 1994–2035. https://doi.org/10.2175/106143014X14031280668731. org/10.1016/j.envpol.2016.11.011. Skodova, A., Prokes, R., Simek, Z., Vrana, B., 2016. In situ calibration of three passive sam- Kosma, C.I., Lambropoulou, D.A., Albanis, T.A., 2014. Investigation of PPCPs in wastewater plers for the monitoring of steroid hormones in wastewater. Talanta 161, 405–412. treatment plants in Greece: occurrence, removal and environmental risk assessment. https://doi.org/10.1016/j.talanta.2016.08.068. Sci. Total Environ. 466–467, 421–438. https://doi.org/10.1016/j.scitotenv.2013.07.044. Smital, T., Terzić, S., Lončar, J., Senta, I., Žaja, R., Popović, M., Mikac, I., Tollefsen, K.-E., Kunz,P.Y.,Simon,E.,Creusot,N.,Jayasinghe,B.S.,Kienle,C.,Maletz,S.,Schifferli,A.,Schönlau, Thomas, K.V., Ahel, M., 2013. Prioritisation of organic contaminants in a river basin C., Aït-Aïssa, S., Denslow, N.D., Hollert, H., Werner, I., Vermeirssen, E.L.M., 2017. Effect- using chemical analyses and bioassays. Environ. Sci. Pollut. Res. 20, 1384–1395. based tools for monitoring estrogenic mixtures: evaluation of five in vitro bioassays. https://doi.org/10.1007/s11356-012-1059-x. Water Res. 110, 378–388. https://doi.org/10.1016/J.WATRES.2016.10.062. Sun, J., Tang, S., Peng, H., Saunders, D.M.V., Doering, J.A., Hecker, M., Jones, P.D., Giesy, J.P., Kuzmanovic, M., Ginebreda, A., Barceló, D., 2014. Risk assessment and prioritization of Wiseman, S., 2016. Combined transcriptomic and proteomic approach to identify tox- pollutants in continental Mediterranean waters based on hazard quotients. Contrib. icity pathways in early life stages of Japanese medaka (Oryzias latipes) exposed to to Sci. 10, 125–134. https://doi.org/10.2436/20.7010.01.197. 1,2,5,6-tetrabromocyclooctane (TBCO). Environ. Sci. Technol. 50, 7781–7790. Lin, Y.-H., Chen, C.-Y., Wang, G.-S., 2007. Analysis of steroid estrogens in water using https://doi.org/10.1021/acs.est.6b01249. liquid chromatography/tandem mass spectrometry with chemical derivatiza- Sun, J., Peng, H., Alharbi, H.A., Jones, P.D., Giesy, J.P., Wiseman, S.B., 2017. Identification of tions. Rapid Commun. Mass Spectrom. 21, 1973–1983. https://doi.org/10.1002/ chemicals that cause oxidative stress in oil sands process-affected water. Environ. Sci. rcm.3050. Technol. 51, 8773–8781. https://doi.org/10.1021/acs.est.7b01987. Liscio, C., Abdul-Sada, A., Al-Salhi, R., Ramsey, M.H., Hill, E.M., 2014. Methodology for pro- Szekacs, A., Mortl, M., Darvas, B., 2015. Monitoring pesticide residues in surface and filing anti-androgen mixtures in river water using multiple passive samplers and ground water in Hungary: surveys in 1990–2015. J. Chem. 2015. bioassay-directed analyses. Water Res. 57, 258–269. https://doi.org/10.1016/j. Thomaidi, V.S., Stasinakis, A.S., Borova, V.L., Thomaidis, N.S., 2015. Is there a risk for the watres.2014.03.039. aquatic environment due to the existence of emerging organic contaminants in Lohmann, R., Booij, K., Smedes, F., Vrana, B., 2012. Use of passive sampling devices for treated domestic wastewater? Greece as a case-study. J. Hazard. Mater. 283, monitoring and compliance checking of POP concentrations in water. Environ. Sci. 740–747. https://doi.org/10.1016/j.jhazmat.2014.10.023. Pollut. Res. 19, 1885–1895. https://doi.org/10.1007/s11356-012-0748-9. Tiedeken, E.J., Tahar, A., McHugh, B., Rowan, N.J., 2017. Monitoring, sources, receptors, and Long, M., Bonefeld-Jørgensen, E.C., 2012. Dioxin-like activity in environmental and human control measures for three European Union watch list substances of emerging con- samples from Greenland and Denmark. Chemosphere 89, 919–928. https://doi.org/ cern in receiving waters – a 20 year systematic review. Sci. Total Environ. 574, 10.1016/j.chemosphere.2012.06.055. 1140–1163. https://doi.org/10.1016/j.scitotenv.2016.09.084. Loos, R., Carvalho, R., António, D.C., Comero, S., Locoro, G., Tavazzi, S., Paracchini, B., Ghiani, Tousova, Z., Oswald, P., Slobodnik, J., Blaha, L., Muz, M., Hu, M., Brack, W., Krauss, M., Di M., Lettieri, T., Blaha, L., Jarosova, B., Voorspoels, S., Servaes, K., Haglund, P., Fick, J., Paolo, C., Tarcai, Z., Seiler, T.-B., Hollert, H., Koprivica, S., Ahel, M., Schollée, J.E., Lindberg, R.H., Schwesig, D., Gawlik, B.M., 2013. EU-wide monitoring survey on Hollender, J., Suter, M.J.-F., Hidasi, A.O., Schirmer, K., Sonavane, M., Ait-Aissa, S., emerging polar organic contaminants in wastewater treatment plant effluents. Creusot, N., Brion, F., Froment, J., Almeida, A.C., Thomas, K., Tollefsen, K.E., Tufi, S., Water Res. 47, 6475–6487. https://doi.org/10.1016/j.watres.2013.08.024. Ouyang, X., Leonards, P., Lamoree, M., Torrens, V.O., Kolkman, A., Schriks, M., 1612 Z. Toušová et al. / Science of the Total Environment 650 (2019) 1599–1612

Spirhanzlova, P., Tindall, A., Schulze, T., 2017. European demonstration program on seasonal and spatial variability of dissolved hydrophobic organic contaminants and the effect-based and chemical identification and monitoring of organic pollutants in metals in the Danube river. Environ. Pollut. 184, 101–112. https://doi.org/10.1016/j. European surface waters. Sci. Total Environ. 601–602, 1849–1868. https://doi.org/ envpol.2013.08.018. 10.1016/j.scitotenv.2017.06.032. Wang, J., Song, G., Li, A., Henkelmann, B., Pfister, G., Tong, A.Z., Schramm, K.W., 2014. Com- US EPA, 2015. Interactive Chemical Safety for Sustainability (iCSS) Dashboard v2 [WWW bined chemical and toxicological long-term monitoring for AhR agonists with SPMD- Document]. based virtual organisms in drinking water Danjiangkou reservoir, China. von der Ohe, P.C., Dulio, V., Slobodnik, J., De Deckere, E., Kühne, R., Ebert, R.-U., Ginebreda, Chemosphere 108, 306–313. https://doi.org/10.1016/j.chemosphere.2014.01.056. A., De Cooman, W., Schüürmann, G., Brack, W., 2011. A new risk assessment approach Weller, M.G., 2012. A unifying review of bioassay-guided fractionation, effect-directed for the prioritization of 500 classical and emerging organic microcontaminants as po- analysis and related techniques. Sensors 12, 9181–9209. https://doi.org/10.3390/ tential river basin specific pollutants under the European water framework directive. s120709181. Sci. Total Environ. 409, 2064–2077. https://doi.org/10.1016/j.scitotenv.2011.01.054. Wilson, V.S., Bobseine, K., Lambright, C.R., Gray, L.E., 2002. A novel cell line, MDA-kb2, that Vörösmarty, C.J., McIntyre, P.B., Gessner, M.O., Dudgeon, D., Prusevich, a, Green, P., stably expresses an androgen- and glucocorticoid-responsive reporter for the detec- Glidden, S., Bunn, S.E., Sullivan, C. a, Liermann, C.R., Davies, P.M., 2010. Global threats tion of hormone receptor agonists and antagonists. Toxicol. Sci. 66, 69–81. to human water security and river biodiversity. Nature 467, 555–561. https://doi.org/ Working Group on Prioritisation of Emerging Substances, 2013. NORMAN Prioritisation 10.1038/nature09549. Framework for Emerging Substances. [WWW Document]. URL. http://www.nor- Vrana, B., Allan, I.J., Greenwood, R., Mills, G.A., Dominiak, E., Svensson, K., Knutsson, J., man-network.net/sites/default/files/norman_prioritisation_manual_15 April2013_ Morrison, G., 2005. Passive sampling techniques for monitoring pollutants in water. final_for_website_0.pdf. TrAC Trends Anal. Chem. 24, 845–868. Vrana, B., Klučárová, V., Benická, E., Abou-Mrad, N., Amdany, R., Horáková, S., Draxler, A., Humer, F., Gans, O., 2014. Passive sampling: an effective method for monitoring

ANNEX IV Science of the Total Environment 581–582 (2017) 350–358

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Assessment of a novel device for onsite integrative large-volume solid phase extraction of water samples to enable a comprehensive chemical and effect-based analysis

Tobias Schulze a,⁎,MarijanAhelb, Jörg Ahlheim a, Selim Aït-Aïssa c, François Brion c, Carolina Di Paolo d, Jean Froment a,e,f, Anita O. Hidasi g, Juliane Hollender g,h, Henner Hollert d,MengHua,d,AnettKloßa, Sanja Koprivica b, Martin Krauss a, Melis Muz a,d, Peter Oswald i,MargitPetrea, Jennifer E. Schollée g,h, Thomas-Benjamin Seiler d, Ying Shao d, Jaroslav Slobodnik i, Manoj Sonavane c, Marc J.-F. Suter g, Knut Erik Tollefsen e,j,ZuzanaTousoval,K, Karl-Heinz Walz l,WernerBracka,d a UFZ Helmholtz Centre for Environmental Research, Permoserstrasse 15, 04318 Leipzig, Germany b Ruđer Bošković Institute, Division for Marine and Environmental Research, Bijenička cesta 54, 10000 Zagreb, Croatia c Institut National de l'Environnement Industriel et des Risques INERIS, Unité d'Ecotoxicologie, 60550 Verneuil-en-Halatte, France d RWTH Aachen University, Department of Ecosystem Analyses, Institute for Environmental Research, Worringerweg 1, 52074 Aachen, Germany e Norwegian Institute for Water Research (NIVA), Gaustadalléen 21, N-0349 Oslo, Norway f Department of Chemistry, University of Oslo (UiO), PO Box 1033, Blindern, N-0316 Oslo, Norway g Eawag: Swiss Federal Institute for Aquatic Science and Technology, 8600 Dubendorf, Switzerland h ETH Zurich, Institute of Biogeochemistry and Pollutant Dynamics, 8092 Zurich, Switzerland i Environmental Institute, s.r.o., Okružná 784/42, 972 41 Koš,SlovakRepublic j Norwegian University of Life Sciences (NMBU), PO Box 5003, N-1432 Ås, Norway k Masaryk University, Faculty of Science, RECETOX, Kamenice 753/5, 625 00 Brno, Czech Republic l MAXX Mess- u. Probenahmetechnik GmbH, Hechinger Straße 41, 72414 Rangendingen, Germany

HIGHLIGHTS GRAPHICAL ABSTRACT

• A novel solid phase extraction device for chemical and effect-based analysis was developed • Good recoveries for organic contaminants in a large log D range were obtained for 159 out of 251 compounds • Samples were successfully evaluated using a set of seven different bioassays for ten endpoints • The device is applicable of sampling of up to 50 L of water

article info abstract

Article history: The implementation of targeted and nontargeted chemical screening analysis in combination with in vitro and Received 19 November 2016 organism-level bioassays is a prerequisite for a more holistic monitoring of water quality in the future. For chem- Received in revised form 20 December 2016 ical analysis, little or no sample enrichment is often sufﬁcient, while bioanalysis often requires larger sample vol- Accepted 20 December 2016 umes at a certain enrichment factor for conducting comprehensive bioassays on different endpoints or further Available online 4 January 2017

⁎ Corresponding author. E-mail address: [email protected] (T. Schulze).

http://dx.doi.org/10.1016/j.scitotenv.2016.12.140 0048-9697/© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). T. Schulze et al. / Science of the Total Environment 581–582 (2017) 350–358 351

effect-directed analysis (EDA). To avoid logistic and technical issues related to the storage and transport of large Editor: D. Barcelo volumes of water, sampling would benefit greatly from onsite extraction. This study presents a novel onsite large volume solid phase extraction (LVSPE) device tailored to fulfill the requirements for the successful effect-based Keywords: and chemical screening of water resources and complies with available international standards for automated Automated water sampler sampling devices. Laboratory recovery experiments using 251 organic compounds in the log D range from Chemical analysis − Bioassay analysis 3.6 to 9.4 (at pH 7.0) spiked into pristine water resulted in acceptable recoveries and from 60 to 123% for Effect-based analysis 159 out of 251 substances. Within a European-wide demonstration program, the LVSPE was able to enrich com- −1 −1 Applicability domain pounds in concentration ranges over three orders of magnitude (1 ng L to 2400 ng L ). It was possible to dis- Large-volume solid phase extraction criminate responsive samples from samples with no or only low effects in a set of six different bioassays (i.e. LVSPE acetylcholinesterase and algal growth inhibition, androgenicity, estrogenicity, fish embryo toxicity, glucocorticoid activity). The LVSPE thus proved applicable for onsite extraction of sufficient amounts of water to investigate water quality thoroughly by means of chemical analysis and effect-based tools without the common limitations due to small sample volumes. © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction (co-)polymers are commercially available (Fontanals et al., 2007; Fontanals et al., 2011; Hennion, 1999). A combination of complementa- In Europe, the protection of natural water resources is regulated by ry sorbents to cover a broad range of compounds with different proper- the Water Framework Directive (WFD; European Union, 2000)and ties has been successfully applied to surface water samples (Kern et al., the Groundwater Daughter Directive to WFD (GWD; European Union, 2009). It is an advantage of SPE to capture and stabilize the compounds 2006) that are implemented in European member states' legislations on the sorbents when sampled (Hillebrand et al., 2013). Different ap- and international river basin management. The monitoring and regula- proaches and devices for the sampling of large volumes of water have tion of the chemical status of surface and ground waters refer to the pri- been developed since the 1970s (CIAgent, 2012; Coes et al., 2014; ority substances listed in WFD and amended by the GWD and the Dawson et al., 1976; de Lappe et al., 1983; Dean et al., 2009; Ehrhardt Environmental Quality Standards (EQS) Directive (European Union, and Bums, 1990; Ellis et al., 2008; Gomez-Belinchon et al., 1988; 2008, 2013). However, it has been demonstrated that monitoring of pri- Green et al., 1986; Hanke et al., 2012; Lakshmanan et al., 2010; ority pollutants is not sufficient, because mixtures of many more known McKenzie-Smith et al., 1994; Petrick et al., 1996; Reineke et al., 2002; and unknown chemicals contribute to adverse environmental effects Roll et al., 2016; Sarkar and Sen, 1989; SEASTAR INSTRUMENTS, 1984; (Malaj et al., 2014; Moschet et al., 2014; Neale et al., 2015; von der Sturm et al., 1998; Suarez et al., 2006; Supowit et al., 2016; Thomas et Ohe et al., 2009). al., 2004; Thomas et al., 2001; Weigel et al., 2001; Yunker et al., 1989). The combination of targeted and nontargeted chemical screening Briefly, many of the devices were best suited for low water volumes analysis with in vitro and organism-level bioassays has been recom- (for analytical purposes), are not (anymore) commercially available or mended for the identification of (eco-)toxicologically active compounds do not operate in a fully automated mode (see Supporting material for and mixtures by a number of more recent studies to supplement the detailed information). existing concepts towards a holistic effect-based and chemical analyses Since none of the existing devices and approaches satisfies all of the approach (Altenburger et al., 2012; Brack et al., 2015; Creusot et al., above-mentioned requirements, a novel device for the onsite large-vol- 2013; Di Paolo et al., 2016; Krauss et al., 2010; Silva et al., 2002; ume SPE (LVSPE) was developed. It fulfills the following technical Wernersson et al., 2015). Generally, the amount of sample enrichment characteristics: required for chemical analyses and bioassay depends on the sensitivity • of individual methods as well as the physicochemical properties, bio- Automated device for the unattended and representative sampling ac- availability, exposure concentrations, toxic potentials and mixture tox- cording to international standards (e.g., ISO 5667-1, 2006); • fi icity effects of the compounds contained in the sample. Modern Combination of SPE with a pre- ltration cartridge to separate chemical analytical instrumentation allows for the analysis of small suspended particulate matter (SPM) from the water phase; • water volumes with no or only low sample enrichment for most of the Tailor-made columns that allow customizable selection and combina- typical water pollutants (Bahlmann et al., 2015; Berset et al., 2010; tion of sorbents to focus on chemical properties and quantities as de- Brack et al., 2015, 2016; Dyer et al., 2004; Fernández-Ramos et al., termined by the goals of the research question; • 2014; Seitz et al., 2006), while the analysis of some priority substances Implementation of a pressurized system to force the water through with very low EQS values as well as in vivo and in vitro tests may require the extraction columns; • greater enrichment and larger water volumes (Neale et al., 2015; OECD, Usage of 12 V electronic components (controller, pumps, valves) and 2004; OECD, 2012). low energy consumption, in such a way that the device can run with a The implementation of integrated chemical and effect-based moni- car battery or a battery-buffered fuel cell, solar panel or wind turbine. toring strategies (Brack et al., 2017) would greatly benefit from automated onsite sampling techniques for efficient and successful real- The successful implementation and application of sampling ap- time collection and extraction of large water volumes. Such techniques proaches in the chemical and biological assessment of complex environ- can prevent logistic, technical, economic and scientific issues related to mental mixtures requires the assurance of the representativeness and the storage and transport of large volumes of water to the laboratory. integrity of the samples with minimized alteration and bias (Brack et Furthermore, this approach allows time-integrated sampling of a al., 2016; Schulze et al., 2011). The aim was to assess whether the water body over days or weeks to yield representative samples (Roll LVSPE device: and Halden, 2016). The most powerful sampling and enrichment approach for complex 1. Is able to capture a wide-ranging set of known organic water con- mixtures of known and unknown contaminants is solid phase extrac- taminants (among them pesticides, biocides, pharmaceuticals, and tion (SPE). Several well-tested and widely used solid phases that trap artificial sweeteners) with good recoveries and high repeatability? organic compounds with a broad range of properties (nonpolar to 2. Can enrich a sufficient volume of water to perform a set of different polar, neutral to charged) based on C18 or polystyrene-divinylbenzene bioassays even for minimally contaminated waters? 352 T. Schulze et al. / Science of the Total Environment 581–582 (2017) 350–358

3. Does provide blank samples containing no or very low contamina- the half quantity those were considered. The cartridges were assembled, tion and deriving no or minimal toxicological effects to be able to un- filled with the solid phase sorbents and conditioned separately accord- equivocal distinguish the chemical and effect signals from ing to Table 1. To account for a swelling of the sorbents, the amounts background levels? were slightly reduced to fit into the columns. After conditioning and sampling, the openings were covered with 2. Material and methods aluminum foil to avoid contamination and drying of the wet sorbent. The columns were stored and transported at 4 °C before and after sam- 2.1. Technical description of the LVSPE device pling. Later, the cartridges were connected separately to a nitrogen gas stream for 1 h to purge residual water and subsequently subjected to The design of the LVSPE device allows for the collection of up to 50 L freeze drying for around 8 h. The extraction was carried out according water (Fig. 1, LVSPE50). The main parts of the devices are the pre-filter, to Table 1. The extracts of the different cartridges were kept separate the sampling and dosing chamber, the ball valve, the pressure chamber for further analysis with HR-XAW and HR-XCW extracts being neutral- and the controller. The devices are built into a Storm Case iM 2750 ized by adding formic acid or 7 N ammonia in methanol (MeOH) before (L × W × D: 62.5 × 50 × 36.6 cm) purchased from Peli Products (Barce- storage. All extracts were reduced in volume using rotary evaporation lona, Spain). An apparatus following the same principle but designed for and adjusted to a final enrichment factor of 1:250 (HR-X) and 1:500 the extraction of up to 1000 L is presented in SM. (HR-XAW, HR-XCW) using a mixture of MeOH:ethyl acetate Briefly, water is sucked by vacuum into the borosilicate glass dosing (EtAc;1:1,v:v) before preparation of aliquots for chemical and biological system (1). The water enters the Sartopure GF+ MidiCap pre-filter analyses. (Sartorius) (2) in the inflow pipe to remove suspended particulate matter. A conductivity sensor controls the maximal water level in the glass 2.3. Laboratory and field performance of the LVSPE50 device tube (volume: 600 mL) and a dip tube allows exact dosing of the sample volume (500 mL). The ball valve (3) keeps the water in the dosing sys- Recoveries were tested under laboratory conditions. A 60 L grab tem and releases it into the pressure chamber (4) when opened. After water sample of a pristine creek (Wormsgraben, Harz Mountains, Ger- release, the ball valve closes and the water is pumped with a positive many; N 51.770167, E 10.696444) was collected on 14 January 2014 pressure of approximately 100 kPa through one cartridge (5) or a se- and stored in a clean stainless steel drum at 4 °C. The sample was divid- quence of cartridges with different sorbents (Fig. 1a). The cartridges ed into 6 × 10 L sub-samples in 10-L borosilicate glass beakers. Three are filled from the bottom to avoid preferential flow paths through the out of six samples were spiked using a mixture of 251 organic com- solid phase bed. pounds (500 ng each; Table S1) in the log D range of −3.6 to 9.7 The controller allows a customized programming of the sampling fre- (pH 7). The substances in the spike mix cover different compounds clas- quency and the total number of sub-samples of 500 mL each until the de- ses such as pharmaceuticals, pesticides, industrial chemicals and other sired total volume is reached (e.g., 50 L). The extraction cartridge of the chemicals of emerging concern which are typically analyzed in surface LVSPE50 device is built of polyvinylidene fluoride (PVDF) (Fig. 1b). Car- waters and wastewater treatment plant effluents (e.g., Hug et al., tridges made of stainless steel can also be used, but fine threads in such 2014; Loos et al., 2013a, 2013b; Richardson and Ternes, 2014; Ruff et parts are prone to malfunctions due to the brittleness of this material. al., 2015). The recoveries were calculated as the ratio between the The cartridges are available in different sizes (4 to 10 g of sorbent). The amount of substance found in the extracts and the amount of substance solid phases are packed between the glass filter plates, and the cartridges spiked to the water samples. Beakers were coated and wrapped with are closed with two screw caps with O-ring type silicone tights. aluminum foil to protect from light and contamination. The remaining three samples were used as unspiked ambient field controls in order 2.2. Preparation, conditioning and extraction of sampler cartridges and pro- to check for background concentrations of the targeted analytes. The cessing of samples samples were extracted using the LVSPE50 with the HR-X, HR-XAW and HR-XCW sorbents in sequence (Table 1). The beakers were rinsed The quantity of sorbents used was up-scaled from an amount of 0.2 g with 1 L of original Wormsgraben water, which was extracted using of sorbent, which is commonly used to extract 1–2 L of water in case of the same cartridges to remove residual compounds from the glass walls. Chromabond® HR-X (Macherey Nagel). Since the cartridges with ion Subsequently, the LVSPE50 device was applied on 35–50 L surface exchange sorbents Chromabond® HR-XAW and Chromabond® HR- water samples collected at 18 sampling sites in six European countries XCW were grouped in flow direction behind the column with HR-X, (Croatia, Czech Republic, Germany, Hungary, Slovakia, Switzerland;

Fig. 1. (a) Picture of the LVSPE50 device; (1): Dosing system (500 mL), (2): pre-filter (3): ball valve, (4): pressure chamber (550 mL), (5): extraction cartridge, (6): controller (Photo by MAXX GmbH); (b) Scheme of the LVSPE50 cartridge, body and screw caps, made of polyvinylidene fluoride; (1): inlet fitting, (2) lower and upper screw caps with mortises to take in the (3) silicone tights (O-rings), (4) outlet fitting, (5) glass filter disc, (6) body containing the sorbent. T. Schulze et al. / Science of the Total Environment 581–582 (2017) 350–358 353

Table 1 separation was used. To account for compound losses, we used matrix- Settings for sampler preparation, conditioning and extraction; MeOH: methanol, EtAc: matched calibration and processed calibration standards using a down- ethyl acetate, LC-MS: liquid chromatography–mass spectrometry; HR-X: hydrophobic scaled SPE method corresponding to that for the LVSPE samples. To polystyrene-divinylbenzene copolymer; HR-XAW: weakly basic secondary and tertiary ammonium polymeric anion exchanger based on HR-X; HR-XCW: weak carboxylic acid demonstrate the applicability of the LVSPE approach for the effect- modified polymeric cation exchanger for SPE; during sampling the sorbents are assembled based analysis of surface water samples, aliquots of the LVSPE samples in the order HR-X, HR-XAW and HR-XCW. were subjected to a set of in vitro and organism-level bioassays (Table LVSPE50 2). The results of the bioassays were reported as relative enrichment factors (REF) which express the enrichment of the mixture of organic Solid phases HR-X (10 g) fi HR-XAW (4 g) pollutants in a sample to achieve a speci c effects in a bioassay HR-XCW (4 g) (Escher and Leusch, 2011; Escher et al., 2006; Escher et al., 2014). The Conditioning HR-X methods for chemical and biological analysis are detailed in the Supple- −200 mL EtAc mentary data (Section S2.4). −200 mL MeOH −100 mL water (LC-MS grade) HR-XAW −200 mL MeOH 2.6. Data analysis −100 mL water (LC-MS grade) HR-XCW Log D values at pH 7.0 and other physicochemical descriptors were − 200 mL MeOH fi −100 mL water (LC-MS grade) calculated using the PhysChem Pro ler of ACD/Percepta (ACD, 2015). Extraction HR-X Open Babel v2.3.2 (O'Boyle et al., 2011) was used to generate InChIKey −100 mL EtAc for the compound identification. Statistical analysis was performed with −100 mL MeOH R 3.3.0 (R Core Team, 2016). The Venn diagram was drawn with the R HR-XAW package VennDiagram (Chen, 2016). The elbow method retrieved the −100 mL MeOH with 2% 7 N ammonia in MeOH HR-XCW optimal number of clusters used in k-means clustering (Ketchen and −100 mL MeOH with 1% formic acid Shook, 1996). Descriptive curve functions were calculated using the R package e1071 (Meyer et al., 2015). Processing of the bioassay data and calculation of concentration-response curves was performed Table S2) as part of the European Demonstration Program (EDP) of the with GraphPad Prism v6.07 (2015). The estrogenic assay data was EDA-EMERGE project (Brack et al., 2013). assessed using the REGTOX Excel™ Macro (http://www.normalesup. org/~vindimian/fr_index.html) as previously described (Kinani et al., 2.4. Ambient unspiked field control and laboratory blank 2010).

Unspiked ambient field controls and a laboratory blank were processed in parallel to the laboratory recovery test and the field sampling 3. Results and discussion campaign in the Saale river basin, respectively. The control samples were subjected to the whole preparation and elution procedure without 3.1. Chemical and biological analysis of the circulation blank any enrichment step. For the ambient unspiked control related to the recovery test, three sub-samples of each 10 L of a 60 L pristine water from The extraction procedure was tested for any undesired chemical Wormsgraben were assessed by chemical analysis to account for possi- contamination as well as toxicological effects to exclude false positives ble interference with spiked compounds. For the blank sample related during monitoring. This step included the recovery and the circulation to the field sampling during the EDP, the concept of a circulation blanks blanks. None of the targeted compounds (N=251)weredetectedinei- was used to evaluate leaching of compounds from the sampling device, ther blank. For the HR-X extract of the circulation blank, the lowest ob- filters, tubing and sorbents and to reduce the efforts of regular process- served effect concentrations (LOEC) elicited a REF of 100 for the ER- and ing of control samples. Typically, the extraction of larger volumes of AR-mediated activity (expressed as cytotoxicity at this LOEC) and a water to obtain blank samples is very expensive with regards to obtain LOEC at REF 250 and REF 500, respectively, for the AChE inhibition large amounts of clean laboratory water or may be affected by back- and the (sub-)lethal endpoints in FET. For the algal growth inhibition ground contamination and artifacts originating from the water sample assay, the no observed effect concentration (NOEC) was at REF 100 for itself, even of high quality such as LC-MS grade. To check for background all three sorbents used. The (sub-)lethal effects in the FET showed a contamination or leaching form the machine, one circulation blank LOEC and NOEC of REF 500, respectively, for the HR-XAW and HR- (sample EDP4091) was processed using 5 L of liquid LC-MS grade XCW. These minor effects of the circulation blank appeared only at water (Chromasolv, Sigma-Aldrich) mineralized with analytical grade high REFs and hence they are unlikely to interfere with the evaluation sodium chloride (0.2 g L−1, Merck) to avoid problems with the conduc- of effects of environmental water samples. However, a thorough tivity-based dosing system. The water was stored in a 5 L brown glass cleaning and conditioning (Table 1) of the sorbents used is highly rec- bottle as a reservoir. From this reservoir, LVSPE50 extracted a 500 mL ommended to remove production residues and contamination due to sub-sample per cycle and discharged the extracted water back into absorption of background air contaminants. the reservoir. Overall, 100 cycles resulted in a blank sample The concept of the circulation blank was based on the assumption representing 50 L of water. that contamination originates from the device, filters, sorbent or tubing and not from the “pure” high-grade water used for the processing of the 2.5. Chemical and biological analysis blank. This approach allowed testing the potential mobilization of prob- lematic contamination from filters, sorbents and tubing under realistic Briefly, liquid chromatography high-resolution mass spectrometry conditions. As a compromise, the circulation blank allowed simulating (LC-HRMS) analysis was carried out using an Agilent 1200 LC coupled the extraction of for instance, 50 L of water by pumping 5 L of LC-MS to a Thermo LTQ Orbitrap XL mass spectrometer with electrospray ion- grade water ten times through the instrument. Nevertheless, the circu- ization (ESI) according to Hug et al. (2015). A Kinetex™ core-shell C18 lation blank of 5 L is a simulation rather than an actual extraction of a column (100 mm × 3.0 mm; 2.6 μm; Phenomex) with a linear gradient 50 L “pure” water sample. If contamination results from the enriched with water and methanol (both containing 0.1% formic acid to account water, it may mask the contaminants leached from the device and for anionic species) at a flow rate of 0.2 mL min−1 for chromatographic consumables. 354 T. Schulze et al. / Science of the Total Environment 581–582 (2017) 350–358

Table 2 Bioassays used for assessment of LVSPE samples; AChE: inhibition of acetylcholinesterase, AR: androgen receptor-mediated activity, ER: estrogen receptor-mediated activity, GR: glucocorticoid receptor mediated activity.

Bioassay Type Target compound groups Endpoint Reference

AChE inhibition Enzymatic Insecticides, miscellaneous Inhibition of AChE enzyme activity (Ellman et al., 1961; Froment et al., 2016; Galgani reaction and Bocquene, 1991) Algal growth inhibition with Organism-level Herbicides, disinfectants, Inhibition of algal growth (OECD, 2011; Rojíčková et al., 1998) Raphidocelis subcapitata miscellaneous Ames ﬂuctuation assay with TA98 In vitro Natural and synthetic Induction of reverse mutations (Ames et al., 1975; Reifferscheid et al., 2011; mutagenic compounds Reifferscheid et al., 2012) AR-mediated activity - MDA-kb2 In vitro Natural and synthetic (Anti-) androgenic response (Creusot et al., 2015; Wilson et al., 2002) cells (anti)androgens ER-mediated activity - MELN cells In vitro Natural and synthetic (Anti-) estrogenic response (Balaguer et al., 1999; Creusot et al., 2015; Kinani (anti)estrogens et al., 2010) GR-CALUX® In vitro Natural and synthetic (Anti-) glucocorticoid receptor (Sonneveld et al., 2005) (anti)glucocorticoids mediated response Zebraﬁsh embryo acute toxicity Organism-level Biocides, pharmaceuticals, Survival, sublethal responses (e.g., (ISO 15088, 2007; OECD, 2013) miscellaneous heartbeat)

3.2. Chemical assessment of spiked water samples XCW (Fig. 2, Table S4). The entire repeatability of the recoveries was 11%, 3% and 2%, respectively, for the HR-X, HR-XAW, and HR-XCW sor- In the recovery test, three replicates of each 10 L of a pristine natural bents (with N=3 replicates of spiked water samples). Two compounds, water sample spiked with 251 compounds were subjected to extraction ethion and triclocarban were not found in any of the three fractions. The with LVSPE50 and analysis with LC-HRMS, to assess the extraction effi- reason was maybe a strong irreversible adsorption to surfaces or the ciency and accuracy of LVSPE. sorbents for which the solvation power of the solvents used was not The Venn diagram in Fig. 2 shows the distribution of the compounds sufficient. between the three different solid phases. The majority of compounds Fig. 3 depicts the distribution of recoveries for the HR-X sorbent. The were recovered from the HR-X (98%; 246 out of 251), the first material recoveries exceeded 50% for 204 out of 251 spiked chemicals. The den- in flow direction. For most chemicals in the intersection of the three sity function retrieved a slightly super Gaussian (kurtosity = 0.3) and solid phases, the main part of spiked substances was found in the HR- left-skewed (skewness = −0.3) distribution (see insert in Fig. 3). The X(N = 48 out of 69) with b10% of recovery in HR-XAW and HR-XCW, calculation of the distribution and density functions for the HR-XAW respectively, the second and third material in flow direction. Only few and HR-XCW sorbents was impossible due to many observations with substances recovered mainly in the HR-XAW (e.g., benzenesulfonic tiny recoveries and thus low variances of the values. acid, chloridazon-desphenyl, perfluorobutanoic acid, salicylic acid) or To evaluate the relationship between the recoveries and the physi- in the HR-XCW (e.g., gabapentin, metformin). The average recoveries cochemical properties of the compounds, regression analysis and k- of the spiked compounds were 88 ± 43% (average and standard devia- means clustering (with k=3 centers) was performed (Fig. 4, Fig. S5, tion; median: 96%; N = 246 out of 251) for the HR-X, 9 ± 21% (N =59 Table S4). Regression analysis did not resolve any systematic dependen- out of 251) for the HR-XAW and 4 ± 6% (N = 49 out of 251) for the HR- cy between the recoveries and the log D and other descriptors (e.g., pKa, Kd, log P; data not shown). Since other analytical factors such as chromatography, ionization or irreversible adsorption to the sorbents or surfaces may affect the recoveries, this result might be different in another experimental setting.

Fig. 3. Histogram of the recoveries (in %) of compounds (N = 251) spiked in a pristine Fig. 2. Venn diagram of spiked compounds recovered in the three different sorbents in water sample of Wormsgraben (Harz Mountains, Germany) and extracted with the ﬂow direction: HR-X: neutral solid phase material, HR-XAW: anionic exchanger solid LVSPE50 device using the neutral HR-X sorbent; the insert shows the density function of phase material, HR-XCW: cationic exchanger solid phase material. the distribution. T. Schulze et al. / Science of the Total Environment 581–582 (2017) 350–358 355

competitive ionic interactions of the ionic exchangers with inorganic + + + + 2+ 2+ − − cations (Li bNa bNH4 bK bMg bCa ) and anions (Cl bBr b − 2− 2− NO3 bSO4 bClO4 )(Bäuerlein et al., 2012), (2) the chemical analysis due to matrix effects (Mallet et al., 2004; Wu et al., 2010), and (3) the results of bioassays due to salinity intolerances of the test species (Gonçalves et al., 2007; Dinnel et al., 1987; Haque et al., 2014; Sawant et al., 2001). Therefore, proper washing of the cartridges with ultra- clean water after extraction is recommended to avoid the carryover of a higher load of inorganic salts to the organic extract (Loos et al., 2013a, 2013b; Wu et al., 2010).

3.3. Biological assessment of ﬁeld samples

A major reason for developing the LVSPE approach was the lack of appropriate sampling equipment for the effect-based screening analysis and monitoring of water resources. Enrichment of a larger volume of Fig. 4. Scatterplot of the total recoveries (in %) of compounds (N = 251) spiked in a water is required to deliver enough extract for the subsequent testing pristine water sample of Wormsgraben (Harz Mountains, Germany) and extracted with the LVSPE50 device versus the water-octanol partition coefficient at pH 7.0 corrected for in a set of different bioassays or even to perform effect-directed analysis. the speciation (log D); the dashed lines express the limits of the clusters derived from k- To investigate whether the LVSPE approach is applicable for effect- means cluster analysis (k = 3); the plot shows only the data for HR-X. based analysis, extracts of samples collected during the EDP were assessed using seven in vitro and organism-level bioassays representing The resulting three groups of k-means clustering include (1) one diverse modes of action (MOA) and adverse effects of pollutants (Table group of compounds with low recoveries in HR-X (b60%) and a larger 2). Since HR-XAW and HR-XCW extracts of those samples were only ef- overlap with HR-XAW and HR-XCW (56 out of 251 compounds), (2) fective in a few assays and endpoints, only the results for the HR-X ex- one group with recoveries in HR-X in the range of 60% to 123% with tracts are represented in this study (Tousova et al., unpublished data). only small overlap with both other sorbents (159 out of 251 com- Using the observation of a biological response at a REF of 100 as a crite- pounds), and (3) one group with recoveries in HR-X N123% with only rion of decision, 8 out of 10 toxic endpoints (Table 2) allowed a discrim- very small overlap with the ion exchanging phases (36 out 251 com- ination of active from non-active surface water samples with 5% pounds, Fig. 4, Fig. S5, Table S4). (endpoint mutagenicity) to 77% (endpoint estrogen receptor mediated Among the causes for recoveries assigned to the first or third group are activity) of the samples exhibiting significant responses. A REF of 100 chromatographic reasons such as elution during dead time and matrix ef- is an enrichment level that can be easily achieved in effect-based and fects in ESI-MS analysis. The matrix effect is caused by co-extracted dis- chemical monitoring using LVSPE in a reasonable period and without solved organic matter (DOM). The DOM is a heterogeneous mixture of any problems of blank toxicity (Fig. 5). Anti-AR activity and AChE inhi- compounds with a wide range of different structures and hence a higher bition did not respond to any of the samples, the latter up to a REF of load of DOM related compounds with affinity to polystyrene- 500. For range finding and avoidance of masking effects of the targeted divinylbenzene co-polymers can be expected (Raeke et al., 2016; specific endpoints, the occurrence of cytotoxicity was tested in all cell- Swenson et al., 2014) that co-elute with similar compounds in LC. Howev- based tests beforehand. er, correction with spiked internal standards and matrix-matched calibration often cannot compensate matrix effects. In the case of very nonpolar or hydrophilic compounds, an irreversible adsorption to surfaces and the sorbents or breakthrough is reasonable, respectively. The latter was observed for 4-aminobenzamide, acetaminophen, chloridazon-desphenyl, chlormequat, mepiquat, and N,N-dimethylsulfamide, which were qualita- tively detected in the effluent water after extraction. The chemical assessment of spiked water samples revealed that the LVSPE approach using the hydrophobic sorbent HR-X was suitable to capture a larger number of the spiked compounds with good recoveries between 60% and 123% without apparent dependency on their physicochemical properties. The usage of any other general purpose solid phase (e.g., Oasis® HLB or Amberlite® XAD) or resins with specificfunctional groups such as ionic exchangers might be an opportunity for tailored applications. However, in this study, the latter considerably enhanced the recoveries of only a few compounds (e.g., benzenesulfonic acid, benzothiazole, gabapentin, metformin, N-nitrosomorpholine, perfluorobutanoic acid, salicylic acid). Certainly, in the setting of the recovery experiment using a relative low volume of spiked water (10 L) and a low expected content of dissolved organic carbon (DOC), the amount of 10 g of HR-X (or similar sorbents) as the first sorbent in flow direction could be enough to trap large amounts of spiked compounds. In another setting with larger volumes of spiked water with higher content of DOC, a larger breakthrough and distribution over the Fig. 5. Occurrence of responses in bioassays to 18 LVSPE samples collected during the three phases is possible. European demonstration program; most samples were tested in most bioassays up to a REF of 100, except AChE (up to REF 500); REF: relative enrichment factor; AChE: In marine applications, the salt content of the water can be an issue acetylcholinesterase enzymatic inhibition, AR: androgenic mediated activity, ER: to be considered. Higher salinity caused by co-extracted inorganic salts estrogenic mediated activity, GR: glucocorticoid receptor mediated signalling, FET: can effect (1) the extraction of charged organic compounds due to zebrafish embryo test (Tousova et al., unpublished data). 356 T. Schulze et al. / Science of the Total Environment 581–582 (2017) 350–358

3.4. Chemical assessment of field samples first screening in a broad set of bioassays and afterwards used for effect-directed analysis in specific assays to unravel cause-effect relation- Fig. 6 shows a selection of concentrations of chemicals analyzed in ships for the prioritization of effects and pollutants. the EDP water samples extracted using the HR-X sorbent. The analytes cover a wide range of substance classes such as pesticides, pharmaceu- Conflicts of interest ticals or industrial chemicals and their transformation products. The −1 concentration levels were in the range from 0.2 ng L to The authors declare no conflicts of interests. However, we empha- −1 2360 ng L . The minimal and maximal concentration levels span sized that the described device is considered for market release and over one to two orders of magnitude for most compounds. Once widely commercial application. used legacy pesticides such as atrazine or simazine were among the identified substances. The overall concentration levels were comparable Acknowledgements to those found frequently in European surface waters (Loos et al., 2013a, 2013b; Ruff et al., 2015; ter Laak et al., 2010). The chemical assessment We feel particularly grateful to Arnold Bahlmann, Birgit Beck, of real water samples showed that the LVSPE approach was applicable Norbert Gockner, Ulrich Haid, Christine Hug, Igor Liska, Riccardo Massei, to water samples containing compounds in a wide span of Emma Schymanski, Peter Tarabek and Tobias Wannenmacher for fruit- concentrations. ful discussions as well as logistical and technical support. This research was supported by the German Federation of Industrial Research Associ- 4. Conclusions ations (grant agreement no. KF2081009MK0), the project TOX-BOX (grant agreement no. 02WRS1282C; http://www.bmbf.riskwa.de), the This study demonstrated LVSPE as a promising tool for the high qual- European Marie Curie Initial Training Network EDA-EMERGE (grant ity sampling and extraction of pollutants for chemical and effect-based agreement no. 290100; http://www.ufz.de/eda-emerge), the European screening of water resources in field applications. LVSPE allows for FP7 Collaborative Project SOLUTIONS (grant agreement no. 603437; onsite extraction of large volumes of water up to 50 L from natural or ar- http://www.solutions-project.eu), the Joint Danube Survey 2013 tificial water sources and thus provides sufficient sample volumes at the (http://www.danubesurvey.org), the NORMAN Association (http:// required enrichment factors for biological screening in a set of different www.norman-network.net) and TERENO (http://teodoor.icg.kfa- bioassays and for chemical screening. Unequivocal distinction between juelich.de/overview-de). The data set of LVSPE50 field measurements likely effects of a blank sample and the effects of even only marginally is available on request. CDP (ESR1), JF (ESR12), AH (ESR5), MH polluted surface water samples was possible in this investigation. Fur- (ESR9), MM (ESR10), MS (ESR2), SK (ESR7), ZT (ESR8) and JeS thermore, LVSPE appears to be suitable for the enrichment of complex (ESR10) were supported by EDA-EMERGE. For a visual impression of mixtures of known water contaminants with no or only low systematic the LVSPE sampling approach, please visit http://www.youtube.com/ dependence from physicochemical properties with “good” recoveries. watch?v=-Zk3GlFYfRw. The flexible concept of the device allows for tailoring the configuration to the user's needs to reach the goals of a particular study. The device Appendix A. Supplementary data will facilitate the development of holistic effect-based and chemical assessment strategies to supplement the existing concepts of water qual- Supplementary data associated with this article can be found in the ity assessment manifested in, e.g., the European Union Water online version, at http://dx.doi.org/10.1016/j.scitotenv.2016.12.140. Framework Directive. For example, the samples can be subjected to a These data include the Google map of the most important areas described in this article.

References

ACD, 2015. ACD/Percepta. Advanced Chemistry Development, Inc., Toronto, ON, Canada. Altenburger, R., Scholz, S., Schmitt-Jansen, M., Busch, W., Escher, B.I., 2012. Mixture toxicity revisited from a toxicogenomic perspective. Environ. Sci. Technol. 46, 2508–2522. Ames, B.N., McCann, J., Yamasaki, E., 1975. Methods for detecting carcinogens and mutagens with the salmonella/mammalian-microsome mutagenicity test. Mutat. Res./En- viron. Mutagen. Relat. Sub. 31, 347–363. Bahlmann, A., Lochen, T., Schulze, T., Kirschner, A., Brack, W., Schneider, R.J., et al., 2015. Chemical and immunochemical analysis of anthropogenic markers and contaminants. In: Liska, I., Wagner, F., Sengl, M., Deutsch, K., Slobodník, J. (Eds.), Joint Danube Survey 3 - A Comprehensive Analysis of Danube Water Quality. ICDPR - International Commission for the Protection of the Danube River, Vienna, pp. 277–283. Balaguer, P., François, F., Comunale, F., Fenet, H., Boussioux, A.-M., Pons, M., et al., 1999. Reporter cell lines to study the estrogenic effects of xenoestrogens. Sci. Total Environ. 233, 47–56. Bäuerlein, P.S., ter Laak, T.L., Hofman-Caris, R.C.H.M., de Voogt, P., Droge, S.T.J., 2012. Re- moval of charged micropollutants from water by ion-exchange polymers – effects of competing electrolytes. Water Res. 46, 5009–5018. Berset, J.-D., Brenneisen, R., Mathieu, C., 2010. Analysis of llicit and illicit drugs in waste, surface and lake water samples using large volume direct injection high performance liquid chromatography–electrospray tandem mass spectrometry (HPLC–MS/MS). Chemosphere 81, 859–866. Brack, W., Govender, S., Schulze, T., Krauss, M., Hu, M., Muz, M., et al., 2013. EDA-EMERGE: an FP7 initial training network to equip the next generation of young scientists with the skills to address the complexity of environmental contamination with emerging pollutants. Environ. Sci. Eur. 25, 1–7. Brack, W., Altenburger, R., Schüürmann, G., Krauss, M., López Herráez, D., van Gils, J., et al., 2015. The SOLUTIONS project: challenges and responses for present and future Fig. 6. Concentrations of frequently found organic compounds in 18 samples collected emerging pollutants in land and water resources management. Sci. Total Environ. −1 during the European demonstration program with the LVSPE50 device (in ng L ) 503–504, 22–31. grouped by occurrence of ﬁndings given in brackets; HR-X: neutral solid phase; data Brack, W., Ait-Aissa, S., Burgess, R.M., Busch, W., Creusot, N., Di Paolo, C., et al., 2016. Ef- shown as box and whisker plot (box: median, 5th/95th-percentiles; whiskers: minimum fect-directed analysis supporting monitoring of aquatic environments — an in- and maximum; dot in box: mean), pluses represent outliers. depth overview. Sci. Total Environ. 544, 1073–1118. T. Schulze et al. / Science of the Total Environment 581–582 (2017) 350–358 357

Brack,W.,Dulio,V.,Ågerstrand,M.,Allan,I.,Altenburger,R.,Brinkmann,M.,etal., analysis of hydrocarbons, polychlorobiphenyls, and fatty acids dissolved in seawater. 2017. Towards the review of the European Union Water Framework Directive: Environ. Sci. Technol. 22, 677–685. recommendations for more efficient assessment and management of chemical Gonçalves, A.M.M., Castro, B.B., Pardal, M.A., Gonçalves, F., 2007. Salinity effects on surviv- contamination in European surface water resources. Sci. Total Environ. 576, al and life history of two freshwater cladocerans (Daphnia magna and Daphnia 720–737. longispina). Ann. Limnol. 43, 13–20. Chen, H., 2016. VennDiagram: Generate High-Resolution Venn and Euler Plots. GraphPad, 2015. GraphPad Prism 6.07. GraphPad Software Inc., La Jolla, CA, USA. CIAgent, 2012. C.I.Agent(R) C.L.A.M Continuous Low-level Aquatic Monitoring. Green, D.R., Stull, J.K., Heesen, T.C., 1986. Determination of chlorinated hydrocarbons in Productsheet, Louisville. coastal waters using a moored in situ sampler and transplanted live mussels. Mar. Coes, A.L., Paretti, N.V., Foreman, W.T., Iverson, J.L., Alvarez, D.A., 2014. Sampling trace or- Pollut. Bull. 17, 324–329. ganic compounds in water: a comparison of a continuous active sampler to continu- Hanke, G., Mariani, G., Comero, S., Loos, R., Bidoglio, G., Polesello, S., et al., 2012. Chemical- ous passive and discrete sampling methods. Sci. Total Environ. 473–474, 731–741. monitoring on-site exercises to harmonize analytical methods for priority substances Creusot, N., Budzinski, H., Balaguer, P., Kinani, S., Porcher, J.-M., Aït-Aïssa, S., 2013. Effect- in the European Union. TrAC Trends Anal. Chem. 36, 25–35. directed analysis of endocrine-disrupting compounds in multi-contaminated sedi- Haque, F., Farhana, T., Amin, F.B., Islam, M.S., 2014. Effect of different salinity exposures on ment: identification of novel ligands of estrogen and pregnane X receptors. Anal. the embryonic development of zebrafish (Danio rerio). Proceedings of the 5th Inter- Bioanal. Chem. 405, 2553–2566. national Conference on Environmental Aspects of Bangladesh. Creusot,N.,Brion,F.,Piccini,B.,Budzinski,H.,Porcher,J.M.,Aït-Aïssa,S.,2015.BFCOD activity Hennion, M.-C., 1999. Solid-phase extraction: method development, sorbents, and cou- in fish cell lines and zebrafish embryos and its modulation by chemical ligands of pling with liquid chromatography. J. Chromatogr. A 856, 3–54. human aryl hydrocarbon and nuclear receptors. Environ. Sci. Pollut. Res. 22, Hillebrand, O., Musallam, S., Scherer, L., Nödler, K., Licha, T., 2013. The challenge of sam- 16393–16404. ple-stabilisation in the era of multi-residue analytical methods: a practical guideline Dawson, R., Riley, J.P., Tennant, R.H., 1976. Two samplers for large-volume collection of for the stabilisation of 46 organic micropollutants in aqueous samples. Sci. Total En- chlorinated hydrocarbons. Mar. Chem. 4, 83–88. viron. 454–455, 289–298. de Lappe, B.W., Risebrough, R.W., Walker, I.W., 1983. A large-volume sampling assembly Hug, C., Ulrich, N., Schulze, T., Brack, W., Krauss, M., 2014. Identification of novel for the determination of synthetic organic and petroleum compounds in the dis- micropollutants in wastewater by a combination of suspect and nontarget screening. solved and particulate phases of seawater. Can. J. Fish. Aquat. Sci. 40, s322–s336. Environ. Pollut. 184, 25–32. Dean, K.E., Suarez, M.P., Rifai, H.S., Palachek, R.M., Koenig, L., 2009. Bioaccumulation of Hug, C., Sievers, M., Ottermanns, R., Hollert, H., Brack, W., Krauss, M., 2015. Linking muta- polychlorinated dibenzodioxins and dibenzofurans in catfish and crabs along an estu- genic activity to micropollutant concentrations in wastewater samples by partial arine salinity and contamination gradient. Environ. Toxicol. Chem. 28, 2307–2317. least square regression and subsequent identification of variables. Chemosphere Di Paolo, C., Ottermanns, R., Keiter, S., Ait-Aissa, S., Bluhm, K., Brack, W., et al., 2016. Bio- 138, 176–182. assay battery interlaboratory investigation of emerging contaminants in spiked water ISO 15088, 2007. Determination of the acute toxicity of waste water to zebrafish Eggs extracts – towards the implementation of bioanalytical monitoring tools in water (Danio rerio). quality assessment and monitoring. Water Res. 104, 473–484. ISO 5667-1, 2006, Genève. Guidance on the design of sampling programmes and sam- Dinnel, P.A., Link, J.M., Stober, Q.J., 1987. Improved methodology for a sea urchin sperm pling technique. cell bioassay for marine waters. Arch. Environ. Contam. Toxicol. 16, 23–32. Kern, S., Fenner, K., Singer, H.P., Schwarzenbach, R.P., Hollender, J., 2009. Identification of Dyer, R.A., Balaam, J.L., Thomas, K.V., 2004. The development of a solid phase extraction transformation products of organic contaminants in natural waters by computer- (SPE) system for environmental monitoring. USA-Baltic Internation Symposium. aided prediction and high-resolution mass spectrometry. Environ. Sci. Technol. 43, 2004, pp. 1–4. 7039–7046. Ehrhardt, M.G., Bums, K.A., 1990. Petroleum-derived dissolved organic compounds con- Ketchen, D.J., Shook, C.L., 1996. The application of cluster analysis in strategic manage- centrated from inshore waters in Bermuda. J. Exp. Mar. Biol. Ecol. 138, 35–47. ment research: an analysis and critique. Strateg. Manag. J. 17, 441–458. Ellis, S.G., Booij, K., Kaputa, M., 2008. Comparison of semipermeable membrane device Kinani, S., Bouchonnet, S., Creusot, N., Bourcier, S., Balaguer, P., Porcher, J.-M., et al., 2010. (SPMD) and large-volume solid-phase extraction techniques to measure water con- Bioanalytical characterisation of multiple endocrine- and dioxin-like activities in sed- centrations of 4,4′-DDT, 4,4′-DDE, and 4,4′-DDD in Lake Chelan, Washington. iments from reference and impacted small rivers. Environ. Pollut. 158, 74–83. Chemosphere 72, 1112–1117. Krauss, M., Singer, H., Hollender, J., 2010. LC–high resolution MS in environmental analy- Ellman, G.L., Courtney, K.D., Andres, V., Featherstone, R.M., 1961. A new and rapid color- sis: from target screening to the identification of unknowns. Anal. Bioanal. Chem. 397, imetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 7, 88–95. 943–951. Escher, B.I., Leusch, F.L., 2011. Bioanalytical Tools in Water Quality Assessment. IWA Pub- Lakshmanan, D., Howell, N.L., Rifai, H.S., Koenig, L., 2010. Spatial and temporal variation of lishing, London. polychlorinated biphenyls in the Houston Ship Channel. Chemosphere 80, 100–112. Escher, B.I., Pronk, W., Suter, M., Maurer, M., 2006. Monitoring the removal efficiency of Loos, R., Carvalho, R., António, D.C., Comero, S., Locoro, G., Tavazzi, S., et al., 2013a. EU- pharmaceuticals and hormones in different treatment processes of source-separated wide monitoring survey on emerging polar organic contaminants in wastewater urine with bioassays. Environ. Sci. Technol. 40, 5095–5101. treatment plant effluents. Water Res. 17, 6475–6487. Escher, B.I., Allinson, M., Altenburger, R., Bain, P.A., Balaguer, P., Busch, W., et al., 2014. Loos, R., Tavazzi, S., Paracchini, B., Canuti, E., Weissteiner, C., 2013b. Analysis of polar or- Benchmarking organic micropollutants in wastewater, recycled water and drinking ganic contaminants in surface water of the northern Adriatic Sea by solid-phase ex- water with in vitro bioassays. Environ. Sci. Technol. 48, 1940–1956. traction followed by ultrahigh-pressure liquid chromatography–QTRAP® MS using European Union, 2000. Directive 2000/60/EC of the European Parliament and of the a hybrid triple-quadrupole linear ion trap instrument. Anal. Bioanal. Chem. 405, Council of 23 October 2000 establishing a framework for community action in the 5875–5885. field of water policy. In: Community, E. (Ed.), Official Journal of the European Malaj, E., von der Ohe, P.C., Grote, M., Kühne, R., Mondy, C.P., Usseglio-Polatera, P., et al., Union. Directive 2000/60/EC. European Community, pp. 1–72. 2014. Organic chemicals jeopardize the health of freshwater ecosystems on the con- European Union, 2006. Directive 2006/118/EC of the European Parliament and of the tinental scale. Proc. Natl. Acad. Sci. 111, 9549–9554. Council of 12 December 2006 on the protection of groundwater against pollution Mallet, C.R., Lu, Z., Mazzeo, J.R., 2004. A study of ion suppression effects in electrospray and deterioration. Official Journal of the European Union. Directive 2006/118/EC. Eu- ionization from mobile phase additives and solid-phase extracts. Rapid Commun. ropean Community, pp. 19–31. Mass Spectrom. 18, 49–58. European Union, 2008. Directive 2008/105/EC of the European Parliament and the Coun- McKenzie-Smith, F., Tiller, D., Allen, D., 1994. Organochlorine pesticide residues in water cil of 16 December 2008 on environmental quality standards in the field of water pol- and sediments from the Ovens and King Rivers, North-East Victoria, Australia. Arch. icy, amending and subsequently repealing Council Directives 82/176/EEC, 83/513/ Environ. Contam. Toxicol. 26, 483–490. EEC, 84/156/EEC, 84/491/EEC, 86/280/EEC and amending Directive 2000/60/EC of Meyer, D., Dimitriadou, E., Hornik, K., Weingessel, A., Leisch, F., 2015. e1071: Misc Func- the European Parliament and of the Council. Official Journal of the European Union. tions of the Department of Statistics. Probability Theory Group. Directive 2008/105/EC. European Community, pp. 84–97. Moschet, C., Wittmer, I., Simovic, J., Junghans, M., Piazzoli, A., Singer, H., et al., 2014. How a European Union, 2013. Directive 2013/39/EU of the European Parliament and the Council complete pesticide screening changes the assessment of surface water quality. Envi- amending Directives of 12 August 2013 amending Directives 2000/60/EC and 2008/ ron. Sci. Technol. 48, 5423–5432. 105/EC as regards priority substances in the field of water policy. Official Journal of Neale, P.A., Ait-Aissa, S., Brack, W., Creusot, N., Denison, M.S., Deutschmann, B., et al., 2015. the European Union. 2013/39/EU. European Community, pp. 1–17. Linking in vitro effects and detected organic micropollutants in surface water using Fernández-Ramos, C., Šatínský, D., Šmídová, B., Solich, P., 2014. Analysis of trace organic mixture toxicity modeling. Environ. Sci. Technol. 49, 14614–14624. compounds in environmental, food and biological matrices using large-volume sample O'Boyle, N., Banck, M., James, C., Morley, C., Vandermeersch, T., Hutchison, G., 2011. Open injection in column-switching liquid chromatography. TrAC Trends Anal. Chem. 62, babel: an open chemical toolbox. J. Cheminformat. 3, 33. 69–85. OECD, 2004. Test No. 202: Daphnia sp., Acute Immobilisation Test. OECD Publishing, Paris. Fontanals, N., Marcé, R.M., Borrull, F., 2007. New materials in sorptive extraction tech- OECD, 2011. Test No. 201: Freshwater Alga and Cyanobacteria, Growth Inhibition Test. niques for polar compounds. J. Chromatogr. A 1152, 14. OECD Publishing, Paris. Fontanals, N., Marcé, R.M., Borrull, F., 2011. Overview of the novel sorbents available in OECD, 2012. Test No. 211: Daphnia magna Reproduction Test. OECD Publishing, Paris. solid-phase extraction to improve the capacity and selectivity of analytical determi- OECD, 2013. Test No. 236: Fish Embryo Acute Toxicity (FET) Test. OECD Publishing, Paris. nations. Contributions to Science. 6, pp. 199–213. Petrick, G., Schulz-Bull, D.E., Martens, V., Scholz, K., Duinker, J.C., 1996. An in-situ filtra- Froment, J., Thomas, K.V., Tollefsen, K.E., 2016. Automated high-throughput in vitro tion/extraction system for the recovery of trace organics in solution and on particles screening of the acetylcholine esterase inhibiting potential of environmental samples, tested in deep ocean water. Mar. Chem. 54, 97–105. mixtures and single compounds. Ecotoxicol. Environ. Saf. 130, 74–80. Raeke, J., Lechtenfeld, O.J., Wagner, M., Herzsprung, P., Reemtsma, T., 2016. Selectivity of Galgani, F., Bocquene, G., 1991. Semi-automated colorimetric and enzymatic assays for solid phase extraction of freshwater dissolved organic matter and its effect on ultra- aquatic organisms using microplate readers. Water Res. 25, 147–150. high resolution mass spectra. Environ. Sci. Process. Impacts 18, 918–927. Gomez-Belinchon, J.I., Grimalt, J.O., Albaiges, J., 1988. Intercomparison study of liquid- R Core Team, 2016. R: A Language and Environment for Statistical Computing. 2016. R liquid extraction and adsorption on polyurethane and Amberlite XAD-2 for the Foundation for Statistical Computing, Vienna. 358 T. Schulze et al. / Science of the Total Environment 581–582 (2017) 350–358

Reifferscheid, G., Buchinger, S., Cao, Z., Claus, E., 2011. Identification of mutagens in fresh- Suarez, M.P., Rifai, H.S., Palachek, R., Dean, K., Koenig, L., 2006. Distribution of water sediments by the Ames-fluctuation assay using Nitroreductase and acetyl- polychlorinated dibenzo-p-dioxins and dibenzofurans in suspended sediments, dis- transferase overproducing test strains. Environ. Mol. Mutagen. 52, 397–408. solved phase and bottom sediment in the Houston Ship Channel. Chemosphere 62, Reifferscheid, G., Maes, H.M., Allner, B., Badurova, J., Belkin, S., Bluhm, K., et al., 2012. In- 417–429. ternational round-robin study on the Ames fluctuation test. Environ. Mol. Mutagen. Supowit, S.D., Roll, I.B., Dang, V.D., Kroll, K.J., Denslow, N.D., Halden, R.U., 2016. Active 53, 185–197. sampling device for determining pollutants in surface and pore water – the in situ Reineke, N., Bester, K., Huhnerfuss, H., Jastorff, B., Weigel, S., 2002. Bioassay-directed sampler for biphasic water monitoring. Sci. Rep. 6, 21886. chemical analysis of River Elbe surface water including large volume extractions Swenson, M.M., Oyler, A.R., Minor, E.C., 2014. Rapid solid phase extraction of dissolved or- and high performance fractionation. Chemosphere 47, 717–723. ganic matter. Limnol. Oceanogr. Methods 12, 713–728. Richardson, S.D., Ternes, T.A., 2014. Water analysis: emerging contaminants and current ter Laak, T.L., van der Aa, M., Houtman, C.J., Stoks, P.G., van Wezel, A.P., 2010. Relating en- issues. Anal. Chem. 86, 2813–2848. vironmental concentrations of pharmaceuticals to consumption: a mass balance ap- Rojíčková, R., Dvořáková, D., Maršálek, B., 1998. The use of miniaturized algal bioassays in proach for the river Rhine. Environ. Int. 36, 403–409. comparison to the standard flask assay. Environ. Toxicol. Water Qual. 13, 235–241. Thomas, K.V., Hurst, M.R., Matthiessen, P., Waldock, M.J., 2001. Characterization of estro- Roll, I.B., Halden, R.U., 2016. Critical review of factors governing data quality of integrative genic compounds in water samples collected from United Kingdom estuaries. Envi- samplers employed in environmental water monitoring. Water Res. 94, 200–207. ron. Toxicol. Chem. 20, 2165–2170. Roll, I.B., Driver, E.M., Halden, R.U., 2016. Apparatus and method for time-integrated, ac- Thomas, K.V., Balaam, J., Hurst, M.R., Thain, J.E., 2004. Identification of in vitro estrogen tive sampling of contaminants in fluids demonstrated by monitoring of hexavalent and androgen receptor agonists in North Sea offshore produced water discharges. En- chromium in groundwater. Sci. Total Environ. 556, 45–52. viron. Toxicol. Chem. 23, 1156–1163. Ruff, M., Mueller, M.S., Loos, M., Singer, H.P., 2015. Quantitative target and systematic von der Ohe, P.C., De Deckere, E., Prüß, A., Muños, I., Wolfram, G., Villagrasa, M., et al., non-target analysis of polar organic micro-pollutants along the river Rhine using 2009. Toward an integrated assessment of the ecological and chemical status of Euro- high-resolution mass-spectrometry – identification of unknown sources and com- pean river basins. Integr. Environ. Assess. Manag. 5, 50–61. pounds. Water Res. 87, 145–154. Weigel, S., Bester, K., Hühnerfuss, H., 2001. New method for rapid solid-phase extraction Sarkar, A., Sen, G.R., 1989. Determination of organochlorine pesticides in Indian coastal of large-volume water samples and its application to non-target screening of North water using a mooredin-situ sampler. Water Res. 23, 975–978. Sea water for organic contaminants by gas chromatography–mass spectrometry. Schulze, T., Streck, G., Paschke, A., 2011. Sampling and conservation. In: Frimmel, F.H. J. Chromatogr. A 912, 151–161. (Ed.), Aquatic Chemistry and Biology. 3. Academic Press, Oxford, pp. 131–152. Wernersson, A.-S., Carere, M., Maggi, C., Tusil, P., Soldan, P., James, A., et al., 2015. The Eu- Sawant, M., Zhang, S., Li, L., 2001. Effect of salinity on development of zebrafish, ropean technical report on aquatic effect-based monitoring tools under the water Brachydanio rerio. Curr. Sci. 81, 1347–1349. framework directive. Environ. Sci. Eur. 27, 1–11. SEASTAR INSTRUMENTS LTD, 1984. SEASTAR in situ water sampler (advertisement). En- Wilson, V.S., Boseine, K., Lambright, C.R., Gray, L.E., 2002. A novel cell line, MDA-kb2, that viron. Sci. Technol. 18 230A-230A. stably expresses an androgen- and glucocorticoid-responsive reporter for the detec- Seitz, W., Schulz, W., Weber, W.H., 2006. Novel applications of highly sensitive liquid tion of hormone receptor agonists and antagonists. Toxicol. Sci. 66, 69–81. chromatography/mass spectrometry/mass spectrometry for the direct detection of Wu, J., Qian, X., Yang, Z., Zhang, L., 2010. Study on the matrix effect in the determination of ultra-trace levels of contaminants in water. Rapid Commun. Mass Spectrom. 20, selected pharmaceutical residues in seawater by solid-phase extraction and ultra- 2281–2285. high-performance liquid chromatography–electrospray ionization low-energy colli- Silva, E., Rajapakse, N., Kortenkamp, A., 2002. Something from “nothing” - eight weak es- sion-induced dissociation tandem mass spectrometry. J. Chromatogr. A 1217, trogenic chemicals combined at concentrations below NOECs produce significant 1471–1475. mixture effects. Environ. Sci. Technol. 36, 1751–1756. Yunker, M.B., McLaughlin, F.A., Macdonald, R.W., Cretney, W.J., Fowler, B.R., Smyth, T.A., Sonneveld, E., Jansen, H.J., Riteco, J.A.C., Brouwer, A., van der Burg, B., 2005. Development 1989. Measurement of natural trace dissolved hydrocarbons by in situ column ex- of Androgen- and Estrogen-Responsive Bioassays, Members of a Panel of Human Cell traction. An intercomparison of two adsorption resins. Anal. Chem. 61, 1333–1343. Line-Based Highly Selective Steroid-Responsive Bioassays. Toxicol. Sci. 83, 136–148. Sturm, B., Knauth, H.-D., Theobald, N., Wünsch, G., 1998. Hydrophobic organic micropollutants in samples of coastal waters: efficiencies of solid phase extraction in the presence of humic acid. Fresenius J. Anal. Chem. 361, 803–810.

ANNEX V Article

pubs.acs.org/ac

Rapid Screening of Acetylcholinesterase Inhibitors by Eﬀect-Directed Analysis Using LC × LC Fractionation, a High Throughput in Vitro Assay, and Parallel Identiﬁcation by Time of Flight Mass Spectrometry † † ‡ § ‡ † Xiyu Ouyang,*, Pim E. G. Leonards, Zuzana Tousova, , Jaroslav Slobodnik, Jacob de Boer, † and Marja H. Lamoree † Institute for Environmental Studies (IVM), VU University Amsterdam, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands ‡ Environmental Institute (EI), Okruzna 784/42, 972 41 Kos, Slovak Republic § Faculty of Science, Masaryk University, RECETOX, Kamenice 753/5, 625 00 Brno, Czech Republic

*S Supporting Information

ABSTRACT: Effect-directed analysis (EDA) is a useful tool to identify bioactive compounds in complex samples. However, identification in EDA is usually challenging, mainly due to limited separation power of the liquid chromatography based fractionation. In this study, comprehensive two-dimensional liquid chromatography (LC × LC) based microfractionation combined with parallel high resolution time of flight (HR-ToF) mass spectrometric detection and a high throughput acetylcholinesterase (AChE) assay was developed. The LC × LC fractionation method was validated using analytical standards and a C18 and pentafluorophenyl (PFP) stationary phase combination was selected for the two-dimensional separation and fractionation in four 96-well plates. The method was successfully applied to identify AChE inhibitors in a wastewater treatment plant (WWTP) effluent. Good orthogonality (>0.9) separation was achieved and three AChE inhibitors (tiapride, amisulpride, and lamotrigine), used as antipsychotic medicines, were identified and confirmed by two-dimensional retention alignment as well as their AChE inhibition activity.

− ne of the major challenges nowadays in the field of logical effect.2 4 Yet, when dealing with an extremely complex O environmental analysis is sample complexity. The huge sample matrix such as WWTP (wastewater treatment plant) number of compounds present in the environment due to effluent, sediment, or biological tissue, it usually happens that cumulative human activities is increasing. A significant part of still tens or even hundreds of MS peaks are detected in each of these compounds is believed to have a potential to cause the active fractions, usually due to limited separation power adverse effects on the ecosystem. Besides, complementing provided in the fractionation procedure. Straightforward extensive programs in many countries that monitor target identification of bioactive chemicals in these active fractions compounds to protect environmental quality, there is a need to therefore still remains challenging unless two or more develop, refine, and apply integrated analytical and effect-based additional fractionation steps would be performed.2,4,5 tools in a so-called effect-directed analysis (EDA) approach.1 Furthermore, as environmental contaminants originate from Conventionally, after sample extraction EDA includes a a variety of sources such as industrial activities, agricultural fractionation based on high performance liquid chromatog- production or urban sewage, their physicochemical properties raphy (HPLC) where usually dozens of fractions are collected, also significantly differ. The persistent organic pollutants followed by a biological screening of the fractions by a specific (POPs) including polycyclic aromatic hydrocarbons (PAHs) bioassay and analysis by either liquid or gas chromatography and dioxin-like compounds are usually nonpolar or very weakly (LC or GC) coupled to high resolution mass spectrometry polar compounds. Pesticides are usually more polar than POPs, (HRMS) of the active fractions to identify the compounds while some of the pharmaceuticals and personal care products responsible for the observed effects. By applying different bioassays covering a broad range of toxicological end points, it Received: November 13, 2015 was proven that EDA is an important tool for environmental Accepted: January 12, 2016 quality assessment to identify contaminants having a toxico- Published: January 12, 2016

© 2016 American Chemical Society 2353 DOI: 10.1021/acs.analchem.5b04311 Anal. Chem. 2016, 88, 2353−2360 Analytical Chemistry Article (PPCPs) and their corresponding metabolites emerging in the ■ EXPERIMENTAL SECTION environment may have a strongly polar character. Therefore, Chemicals. Acetonitrile was HPLC grade supplied by one single HPLC separation mechanism for sample fractiona- ffi Sigma-Aldrich (Zwijndrecht, The Netherlands). Water was tion may not be su cient to fully address the chemical obtained from a Milli-Q Reference A+ purification system properties of complex environmental samples. (Millipore, Bedford, MA, USA). Formic acid and the reagents To further simplify the complexity of the LC fractions, used in the AChE inhibition assay were purchased from Sigma- developing and applying an innovative comprehensive Fluka (Zwijndrecht, The Netherlands). The EPA 531.1 analytical tool capable of delivering higher resolution and Carbamate Pesticide Calibration Mixture was purchased from × multiselective fractionation in EDA is required. LC LC is an Restek (Bellefonte, PA, USA). The standards of tiapride, emerging technique where an extra, independent LC separation amisulpride and lamotrigine were purchased from Sigma-Fluka. fi is implemented after the rst conventional separation to All standards were diluted in 50:50 acetonitrile/water (v/v) to achieve greater peak capacity. Theoretically, the peak capacity 1−10 μg/mL prior to injection. of LC × LC is the product of the peak capacities of the Sampling Site. The WWTP in Brno-Modrice,̌ Czech individual dimensions, provided that the two separation Republic (49.12447N, 16.62697E), serves the urban agglom- mechanisms are truly orthogonal.6 Furthermore, stationary eration of Brno with a population of over 400,000. The WWTP phase combinations of the two dimensions in LC × LC may collects mainly municipal wastewater (WW). However, WW provide multiselective separation of the contents of a sample. In from several industries, hospitals or street runoff contributes to recent years, LC × LC has been successfully applied to complex the overall load of emerging contaminants because the sewage sample matrices in diverse fields such as proteomics, system does not differentiate between these sources. The − pharmaceuticals, food and polymer science, etc.7 10 The WWTP was modernized in 2004 and equipped with tertiary application of LC × LC in environmental analysis is also treatment including nitrogen and phosphorus removal. promising; however, there are only very few studies on the use WWTP Effluent Sampling and Sample Preparation. of LC × LC in this field.11,12 In comparison with GC based The WWTP effluent sample was collected in mid-August 2013 techniques, another major advantage of LC × LC in EDA is the using a large volume solid phase extraction (LVSPE) device straightforward applicability for fractionation. By implementing (UFZ, Leipzig, Germany; Maxx GmbH company, Rangendin- 18 a post column fraction collector, direct microfractionation in 96 gen, Germany). The device enabled extraction of 50 L of or 384 well plates after LC × LC separation can be easily water within 4 h and primary on-site fractionation into 3 ffi ff achieved. In addition, the high resolution fractionation in fractions based on the a nity of di erent sorbents for distinct fi microplates enables high throughput screening of bioactivity. compound groups. Water entering the device was ltered with a fi fi μ Acetylcholinesterase (AChE) inhibitors are chemicals that glass ber lter (0.63 m, Sartopure GF+ MidiCap, Sartorius ̈ inhibit the AChE enzyme that hydrolyses the neurotransmitter AG, Gottingen, Germany) and then pressurized through three fi acetylcholine (ACh) to choline, which may cause accumulation sorbent cartridges mounted in sequence. The rst cartridge of ACh in the synaptic cleft and result in overstimulation of containing the neutral sorbent polystyrene-divinylbenzene cholinergic receptors.13 AChE inhibitors such as organo- copolymer (PS-DVB; Chromabond HR-X, Macherey Nagel, ̈ − phosphates and carbamates have been widely used as Duren, Germany 8 g) to capture neutral and semipolar insecticides. They kill insects by inhibiting the AChE enzyme compounds was followed by the second cartridge with a weak 14 anionic exchanger based on PS-DVB sorbent (Chromabond in their nervous system. Some carbamates have also been − applied to attenuate presynaptic cholinergic deficits related to HR-XAW 3.5 g) to capture acidic compounds. The third 15 cartridge with a weak cationic exchanger also based on PS-DVB Alzheimer’s disease. In this study, the AChE inhibition was sorbent (Chromabond HR-XCW − 3.5 g) capture basic investigated in LC × LC fractions of WWTP effluents using the compounds that are cationic at a water pH of 6−8. classic Ellman’s method with modifications.16,17 Prior to the field sampling, all equipment parts were cleaned Hyphenation of LC × LC to a high resolution mass with methanol (MeOH) and the sorbent cartridges were spectrometer (HRMS) was established for identification of fl preconditioned with a mixture of MeOH and Milli-Q water. In environmental contaminants. Of all MS detectors, time of ight the field, the weather conditions as well as the main (ToF) MS is considered to be the most suitable to couple with × physicochemical parameters of the sampled water (temper- LC LC due to its fast scan rate and moderate to high ature, pH, dissolved oxygen, conductivity) were recorded. After resolving power. However, the high flow rate commonly used fi × the eld sampling, the sorbent cartridges were freeze-dried and in the second dimension of LC LC is not optimal for most of eluted separately with different solvent mixtures (neutral the ionization sources. By applying a postcolumn splitter, a cartridge: MeOH and ethyl acetate, 1:1, v/v; anionic cartridge: small portion of the eluent was introduced to the MS while the 2% v/v 7N ammonia in MeOH; cationic cartridge: 1.7% v/v rest was collected in 96 or 384 well plates using a fraction formic acid in MeOH). The eluates were filtered through glass collector. This parallel approach not only provides an optimal microfiber filters (Whatman GF/F: 0.7 μm, VWR, Vienna, flow rate for the MS interface but also enables a direct Austria) and evaporated to dryness using nitrogen. Each dried correlation of activity in the wells with MS identification. For eluate was stored at −20 °C until reconstitution in 1 mL of identification of active chemicals, the accurate mass and Milli-Q water and acetonitrile (1:1, v/v) for further analysis. In isotopic pattern were first evaluated to determine the molecular this study only the eluate from the neutral cartridge was formula. Then the molecular formulas were searched in online investigated. An instrument blank was prepared by extraction of databases for likely candidates, allowing narrow mass errors and 2 L mineralized LC-grade distilled water according to the isotopic pattern fits. Eventually the tentatively identified procedure used for the field sample. compounds were confirmed by two-dimensional retention LC × LC-ToF MS Instrumentation. The LC × LC system alignment and bioactivity in the AChE inhibition assay.11 integrates an Agilent 1100 auto sampler, an Agilent 1100 HPLC

2354 DOI: 10.1021/acs.analchem.5b04311 Anal. Chem. 2016, 88, 2353−2360 Analytical Chemistry Article binary pump (first dimension), an Agilent 1290 Infinity UHPLC binary pump (second dimension), and an Agilent 1290 Infinity thermostated column compartment (TCC) with a 2-position/4-port duo valve and two sampling loops (80 μL) installed as the 2D interface (Agilent Technologies, Waldbronn, Germany). The LC × LC module was controlled by the Chemstation version B.04.03 (Agilent Technologies) with 2D- LC add-on. In the first dimension, a ZORBAX Eclipse Plus (1.8 μm, 2.1 × 150 mm ID) C18 Rapid Resolution HD column (Agilent Technologies, Santa Clara, CA, USA) was used and a Phenomenex Kinetex (2.6 μm, 50 × 4.6 mm ID) PFP column (Phenomenex, Torrance, CA, USA) was deployed for the second dimension. After the second column, the flow was split using a QuickSplit adjustable flow splitter (Richmond, CA, USA): 20% was directed to a Bruker micrOTOF time of flight (ToF, resolving power ∼10000) mass spectrometer with an electrospray interface (ESI, Bruker Daltonics, Bremen, Germany) while the remaining 80% went through an Agilent 1260 Infinity variable wavelength detector (VWD) followed by an Agilent 1260 fraction collector (Agilent Technologies). Figure 1 illustrates the configuration of the integrated LC × LC system (valve position 1).

Figure 1. Instrumentation of the developed LC × LC system with min 20% B, shiftedmL/min. to 35% in 25 min and then 80% at 55 min; 0.5 min 40% B, shifted to 85% in 48 min. Flow rate: 2.0 min 45% B, shiftedmin. to Flow 50% in rate: 30 2.0 min mL/min. and then 90% at 55 min; 0.5 min 65% B, shifted to 75% in 30 min and then 95% at 40 Mobile phase: (A) water with 0.1% formic acid; (B) acetonitrile with 0.1% formic acid. Modulation time: 0.6 min. Gradient: 0 parallel ESI-ToF-MS and UV detection followed by fraction collection Mobile phase: (A) water with 0.1% formic acid; (B) acetonitrile with 0.1% formic acid. Modulation time: 0.6 min. Gradient: 0 for EDA.

LC × LC Conditions. The chromatographic conditions of the LC × LC experiments are listed in Table 1 and are the optimized conditions for the carbamate mixture or the WWTP sample. MS Settings. The start and stop signals of the micrOTOF were initiated by external control via serial ports. The MS data were recorded by Bruker OtofControl 3.0 (Bruker Daltonics), using a scan frequency of 5 Hz in order to collect enough data points for fast separation in the second dimension. The ion rst dimension LC conditions second dimension LC conditions source and transfer settings of the MS were adjusted to achieve fi optimum sensitivity in the selected mass range (50−1000, m/ z). The capillary voltage of the ESI was 4500 V with end plate offset −500 V. Due to the relatively high flow rate (400 μL/ 20% B, 30 minrate: 45% 0.1 B, mL/min. 55 min 80% B, 60 min 80% B. Flow min) in the second dimension, the nebulizer gas (N2) was 45% B, 35 minrate: 55% 0.1 B, mL/min. 55 min 90% B, 60 min 90% B. Flow operated at 4.0 bar and the drying gas was set to 8 L/min at a L Mobile phase: (A) water; (B) acetonitrile. Gradient: 0 min ° L Mobile phase: (A) water; (B) acetonitrile. Gradient: 0 min μ temperature of 200 C. To enable best detection of the most μ 20 20 interesting molecules, the capillary exit was set at 100 V with a volume injection skimmer voltage of 33.3 V, the hexapole RF was regulated to 90 Vpp and lens 1 prepulse storage was set to 1 μs. PFP, PFP, C C ° ff ° × Assessment of Orthogonality. Two di erent algorithms × column 25 were applied to evaluation the orthogonality of the separation 25 LC Conditions combination C18 to characterize the performance of the LC × LC. The C18 × normalized surface coverage was first estimated using the 11 uent ffl ̌ surface coverage method described in a previous study. The ice details of the calculation can be found in SI. The orthogonality was also evaluated by measuring the spread of peaks within the samples mate pesticide mixture 19 WWTP e extract EPA 531.1 carba- Brno-Modr separation space, according to a recent approach. Table 1. LC

2355 DOI: 10.1021/acs.analchem.5b04311 Anal. Chem. 2016, 88, 2353−2360 Analytical Chemistry Article LC × LC ToF-MS Data Analysis. First, the base peak ■ RESULTS AND DISCUSSION chromatograms (BPCs) obtained were calibrated by creating a Validation of the LC × LC Microfractionation calibration segment before the analysis using the calibration Approach. To evaluate the applicability of LC × LC for tunemix solution and high precision calibration (HPC) in the microfractionation, a carbamate mixture was fractionated into instrument software package DataAnalysis version 4.1 (Bruker four 96-well plates followed by the AChE inhibition assay to Daltonics). The calibrated data were then saved as netCDF. test the applicability of this workflow before applying it to a real Second, the data (both from UV and MS) were evaluated and sample. In virtue of the postcolumn flow splitter shown in the two-dimensional counter plots were generated by GC Figure 1, the LC × LC fractionations were performed in parallel Image 2.3b4 (Lincoln, NE, USA) using the linear interpolation to the ToF-MS analysis, in triplicate. The AChE screening algorithm. Third, tentative compound identifications were results in the four plates are presented in Figure 2 as a heatmap carried out using DataAnalysis. The SmartFormula function in DataAnalysis was used to determine possible chemical formulas corresponding to the observed accurate masses and isotopic patterns. Afterward, the CompoundCrawler function was applied to search for known compounds from a large number of databases, such as ChemSpider, NIST and METLIN. The data analysis of the semiquantitative experiments was performed directly in DataAnalysis by integrating the peak areas of a series of concentrations of standards injected. Fractionation Settings. The fractionation was established after the UV detector by an Agilent 1260 fraction collector. The delay volume (55 μL) was determined according to the Agilent Figure 2. Average (n = 3) percentage of AChE inhibition of the 384 technical note (Part Nr. G1364−90104). For each fractiona- fractions in 4 96-well plates of the EPA 531.1 carbamate mixture (2 μg/mL of each component) after LC × LC fractionation. The tion, four 96-well microplates (polystyrene F-bottom, Greiner numbers on the top are the column numbers (1−12) of the plates and Bio-One, Alphen a/d Rijn, The Netherlands) were placed in the letter and number combinations to the left of the graph show the the fractionation collector. The total run time (60 min) was plate number (P1−P4) together with the row number (A-H) of the evenly divided into 384 (4 × 96) parts, giving 9.375 s collection plates. time for each fraction. The fractionation was performed row by row following the shortest pathway, which is illustrated in the Supporting Information (Figure S1). The needle distance after comparison with a blank fractionation run. According to above the wells was 3.5 mm. After the fractionation, volatile the retention times of the compounds in the mixture recorded by the MS, after dead volume correction, it is possible to solvents were evaporated using a CentriVap concentrator fi ° correlate the observed activity to speci c compounds in the (Labconco, Kansas City, KS, USA) at 40 C in 6 h, to remove mixture. AChE inhibition was observed for each of the fractions the solvents for the following bioassay screening. Prior to the μ known to contain one of the carbamates present in the mixture. fractionation process, 10 L of glycerol solution (10% in water, Although very good separation was achieved in both m/m, Sigma-Fluka) was added to each well as a solvent keeper 20 dimensions (see Figure S2 in the SI), several compounds to enhance the recovery. caused a response in more than two fractions, which could be AChE Inhibition Assay. The AChE inhibition assay was attributed to the high concentrations present in the mixture. In 17 performed based on Hamers et al. with modifications. The addition, the metabolite aldicarb-sulfoxide (produced by method used S-acetylthiocholine-iodide (ATC) as a substrate cytochrome P450) is roughly 200 times more potent to inhibit to evaluate the AChE activity. The ATC hydrolysis was AChE than aldicarb itself21 and therefore led to significant measured by reaction of ATC with 5, 5′-dithiobis(2-nitro- activities in several fractions such as P1D12, P1D11, P1D10, benzoic acid) (DTNB, Ellman’s reagent) to generate the yellow P1C10 and P1C11. The activities in the wells P1C10 and 5-thio-2-nitro-benzoic acid anion, using a UV−vis spectrometer P1C11 also indicated minor wrap-arounds occurring in the LC μ ff × LC system, also illustrating the high sensitivity of the assay. at 412 nm. First, 50 L of 0.1 M phosphate bu er (KH2PO4/ μ fi The outstanding performance of the LC × LC system in terms K2HPO4; pH 7.5) and 50 L of puri ed AChE from electric eel of retention time stability in both dimensions was demon- (Electrophorus electricus; 125 mU/mL in 0.02 M sodium 11 phosphate buffer; pH 7.0) were added to each well of the strated in our previous work, ensuring highly repeatable and accurate fractionation. fractions in the 96-well plates. The plate was incubated for 30 × μ LC LC Separation of the LVSPE Extract of the WWTP min at room temperature. Then 50 L of 5 mM DTNB and 50 ffl × μ E uent. The LC LC separation was established in L of 0.8 mM ATC were added to each well. Afterward, a accordance with our previous work.11 Because of good kinetic measurement was immediately performed at 412 nm for orthogonality and solvent compatibility, C18 and PFP columns 5 min (interval 20 s) in a Multiskan FC micro plate photometer were selected for the two chromatographic dimensions. A fast fi fi (Thermo Scienti c, Waltham, MA, USA). For each identi ed gradient in the second dimension was optimized. The contour compound (tiapride, amisulpride, and lamotrigine) a dose plot of the LC × LC-ESI (+)-ToF MS chromatogram is shown response curve was determined in triplicate, using a series of 8 in Figure 3. Using this system, the peak capacity was concentrations. The dose response curves were plotted and the significantly enhanced compared with 1DLC as extensive IC50 values were calculated using GraphPad Prism 6 (GraphPad separation is achieved in the second dimension. In total, 92 Software, San Diego, CA, USA). visible peaks were matched using the template matching

2356 DOI: 10.1021/acs.analchem.5b04311 Anal. Chem. 2016, 88, 2353−2360 Analytical Chemistry Article

directly link the active fractions to the chromatographic peaks containing high resolution MS information. The identification of compounds that may have contributed to the activity was therefore rather straightforward. In contrast to the conventional EDA approach in which fractionation into larger fractions is applied, high resolution microfractionation greatly reduced the number of candidate chemicals in the active fractions. Merely 2−4 major MS peaks (intensity >10000) were detected in each active fraction and tentative identification of these peaks was performed according to their accurate masses and isotopic patterns (Table 2). The maximum mass error was set at 10 Figure 3. Contour plot of the LC × LC-ESI (+)-ToF MS analysis of ppm and the maximum mSigma value (lower indicates a better the Brno WWTP effluent extract. MS data were acquired at a fit) was 20. For the peaks for which SmartFormula and frequency of 5 Hz. Details of the chromatographic conditions and MS CompoundCrawler returned more candidates, the log Kow settings are given in the Experimental Section . values of the candidate compounds were also taken into account to rank the results. In well P1B2, the major peaks were function after blob detection (minimum peak 1000, absolute observed at m/z 125.9875 and m/z 110.0095 but no chemical value). The estimated orthogonality was 0.937 (93.7%) using 11 formula was generated by SmartFormula. In the other six active the surface coverage approach and 0.945 (or 94.5%) using the wells, several pharmaceuticals and their metabolites could be 19 ff spread of peaks method. The results from the two di erent tentatively identified, as shown in Table 2. Among all the methods were comparable and both showed great orthogon- tentatively identified compounds, tiapride and lamotrigine were ality of the LC × LC separation. The enhanced peak capacity − reported to be weak AChE inhibitors with IC50 values of 1 10 and near perfect orthogonality provide the optimal basis for μM and 50−150 μM, respectively.22,23 The report of the high resolution microfractionation in a relatively short time and ffl 24 fi presence of tiapride in WWTP e uent was rather limited, supporting fast and accurate identi cation of active compounds while lamotrigine was found widely in environmental water after bioassay screening. samples, including drinking water, surface water and WWTP AChE Activity in LC × LC Fractions of Brno WWTP ffl 25 ffl e uent. Although no AChE inhibiting activity was reported E uent Extract. For the assessment of AChE activity in the in the literature for amisulpride, the structurally fairly similar extract of the Brno WWTP effluent, the extract and the × antipsychotic drug sulpride, was reported as a weak AChE corresponding blank were fractionated with LC LC using the inhibitor.22 Therefore, in addition to tiapride and lamotrigine, same procedure as for the carbamate mixture. Figure 4 amisulpride was also selected for further analytical and bioassay confirmation. Two dimensional retention alignments were performed to confirm the presence of tiapride (m/z 207.1493, [M + H]+), lamotrigine (m/z 242.1438, [M + H]+) and amisulpride (m/z 305.1094, [M + H]+) in the WWTP effluent extract. In Figure 5,LC× LC runs of the three standards are shown at the top, while the EICs of the compounds are shown at the bottom. For all three tentatively identified compounds, the retention times in the first dimension perfectly matched with those of their corresponding standard. In the second dimension, the retention time differences of tiapride, lamotrigine and amisulpride between the standards of and the sample were 1.23 s, 0.00 11 Figure 4. AChE inhibition (%, n = 3) of the 384 fractions of the Brno and 0.21 s, respectively. In earlier work, a maximum retention WWTP effluent extract after LC × LC fractionation. The numbers on time deviation of 5% in the second dimension was proposed as the top are the column numbers (1−12) of the plates and the letter a criterion for confirmation, meaning that with a second and number combinations to the left of the graph show the plate dimension run time of 36 s the maximum deviation is 1.8 s. number (P1−P4) together with the row number (A-H) of the plates. Therefore, the presence of the three candidate compounds was confirmed. demonstrates the AChE activity in the 384 fractions of the Bioactivity Confirmation of Candidate AChE Inhib- extract after bioassay screening. In total, there were 7 active itors. To confirm the bioactivity in the active fractions, fractions when a threshold value of 10% inhibition was applied. semiquantification based on LC × LC-ToF MS of tiapride, All active fractions were found in the first well plate containing lamotrigine and amisulpride was performed. Semiquantification the more polar compounds, with activities in the range of 11 to was performed here as EDA is not particularly suitable for 24% inhibition (P1B2, 11.9%; P1C2, 16.9%; P1C3, 24.2%; quantitative analysis due to the impossibility to include internal P1C6, 11.4%; P1C7, 16.1%; P1C10, 15.1%; P1D11, 12.1%). standards. The concentrations in the well of the three identified The observation of the concentration of the activity in the more compounds that exhibited AChE inhibition were determined polar fractions of the extract is supported by the fact that the using calibration curves. In fraction P1C3, the concentration of AChE inhibition assay is performed exclusively in the aqueous tiapride was estimated to be 2.3 μM. Due to wrap-around, phase. tiapride could also be found in fraction P1C6 at an estimated Effect-Directed Analysis of the Active Fractions Using concentration of 1.0 μM. In fraction P1C7, where amisulpride Parallel LC × LC-ESI (+)-ToF MS. Using the experimental and lamotrigine were identified, the estimated concentrations setup with the postcolumn splitter (Figure 1) it was possible to were 1.0 μM and 1.5 μM, respectively. The concentrations of

2357 DOI: 10.1021/acs.analchem.5b04311 Anal. Chem. 2016, 88, 2353−2360 a Chemistry Analytical Table 2. Tentative Identiﬁcation of Compounds Present in the Active Fractions

logKow known m/z possible [M + m/z error (KOWWIN AChE fraction measured H]+ calculated (ppm) mSigma candidate v1.67) known biological role activity − P1C2 232.1080 C12H14N3O2 232.1081 0.2 5.9 Isocarboxazid 1.32 nonselective, irreversible monoamine oxidase inhibitor (MAOI) no Triaziquone 0.51 react with DNA to form intrastrand cross-links; used in chemotherapy no 4- 0.50 metabolite of aminophenazone. no formylaminoantipyrine − 246.1238 C13H16N3O2 246.1237 0.5 15.0 4-acetamidoantipyrine 0.13 metabolite of metamizole no

P1C3 309.1810 C16H25N2O4 309.1809 0.4 8.3 Ubenimex 0.18 competitive protease inhibitor, aminopeptidase B inhibitor, leukotriene A4 hydrolase no and aminopeptidase N; used in treatment of myelocytic leukemia − β 250.1792 C15H24NO2 250.1802 3.7 12.6 Alprenolol 2.81 nonselective blocker; used in the treatment of angina pectoris no − * 22 329.1528 C15H25N2O4S 329.1530 0.6 10.5 Tiapride 0.53 selective D2/D3 dopamine receptor antagonist, used in neurological and psychiatric yes disorders P1C6 250.1801 wrap-around from fraction P1C3 329.1535 β 337.2124 C18H29N2O4 337.2122 0.8 14.0 Acebutolol 1.19 selective 1 receptor blocker for the treatment of hypertension and arrhythmias no * P1C7 370.1809 C17H28N3O4S 370.1795 3.8 10.7 Amisulpride 1.11 an atypical selective D2/D3 dopamine receptor antagonist; used as antipsychotic drugs no β 268.1908 C15H26NO3 268.1907 0.2 6.1 Metoprolol 1.69 selective 1 receptor blocker used in treatment of high blood pressure no * 23 256.0151 C9H8Cl2N5 256.0151 0.0 10.0 Lamotrigine 0.99 anticonvulsant drug used in the treatment of epilepsy and bipolar disorder yes P1C11 268.1910 wrap-around from fraction P1C7

P1D11 134.0723 C7H8N3 134.0713 8.1 5.7 2-aminobenzimidazole 0.88 metabolite of benomyl and carbendazim no

2358 a *Conﬁrmed compounds by analytical standards. O:10.1021/acs.analchem.5b04311 DOI: nl Chem. Anal. 06 8 2353 88, 2016, Article − 2360 Analytical Chemistry Article

Figure 5. Two dimensional retention alignment of the tentatively identified compounds tiapride (A), lamotrigine (B), and amisulpride (C). the three compounds at the sampling site (concentration factor By effect-directed analysis using LC × LC fractionation, a =50,000) were therefore approximately 22 ng/l for tiapride, 7 high throughput in vitro AChE inhibition assay and parallel ng/l for amisulpride and 8 ng/l for lamotrigine. For the identification by ToF-MS, the identification and confirmation compounds tiapride and amisulpride, the dose response curves of three pharmaceuticals with applications in neurology/ to derive the IC50 values in the AChE inhibition assay are psychiatry was achieved. The compounds causing the effects presented in Figure 6. Due to the relatively limited solubility of in the other four active fractions could not be identified, possibly due to the limited sensitivity and/or suitability of the ESI-ToF-MS instrumentation used. For instance, organophosphate pesticides are strong AChE inhibitors, but the preferred interface for their analysis is atmospheric pressure chemical ionization (APCI).26 Ideally, to achieve a higher identification success rate the suspect screening based on accurate mass, as shown here using ToF-MS instrumentation, should be complemented with MS/MS and in silico fi Figure 6. AChE inhibition dose response curves (n = 3) of tiapride fragmentation based approaches for the identi cation of 27 (A) and amisulpride (B). unknowns.

■ CONCLUSION lamotrigine in the test medium, a dose response curve could A high throughput EDA method was established for fast not be reliably obtained for this compound using the current screening of AChE inhibitors in WWTP effluents, building AChE protocol. further on our previous LC × LC-ToF MS application for μ − μ 11 For tiapride, an IC50 value of 4.8 M (3.0 7.6 M, 95% environmental analysis. The EDA approach comprised LC × fi con dence interval) was found which compares well to the LC separation, ESI-ToF MS detection, and parallel micro- − μ range of 1 10 M reported in the literature. The IC50 of μ μ fi fractionation followed by high throughput bioassay screening. amisulpride was 18 M (11- 30 M, 95% con dence interval), The analytical performance of the system was evaluated for which is in the range of the structurally similar compound carbamate pesticides. The bioactivity observed in the 4 × 96 sulpride (10−100 μM).22 It is the first time to identify ffl wells correlated excellently with the identity of the active amisulpride as an AChE inhibitor from WWTP e uent. compounds in these fractions. To evaluate the contribution of the identified compounds to The presence of tiapride, amisulpride and lamotrigine, all the observed activities in the “hot” fractions, the percentage three weak AChE inhibitors that are used as antipsychotic inhibition could be derived from the dose response curves by medicines in a WWTP effluent was demonstrated. For interpolation using the estimated concentrations obtained fi through the semiquantitative chemical analysis. In the fractions amisulpride this was the rst time that AChE inhibiting activity P1C3 and P1C6, the respective concentrations of 2.3 and 1.0 was observed. μM of tiapride were able to cause 25% and 12% inhibition, The greater peak capacity and excellent orthogonality (0.937 which is similar to the measured value 24% and 11%, and 0.945, respectively, estimated by two methods) using a respectively. Similarly, in fraction P1C7, 1.0 μM of amisulpride stationary phase combination of C18 and PFP enabled high × showed an inhibition of 12% of the AChE activity, while the resolution postcolumn fractionation. The LC LC separation fi ff measured response is 16%. Care should be taken to interpret also resulted in a signi cantly reduced matrix e ect, which fi this slight discrepancy not too strictly with regard to the supported a fast and simple identi cation of AChE inhibitors in percentage of the activity that can be explained by identified the active fractions after bioassay screening. In addition, highly compounds, as most EDA studies have a somewhat semi- accurate two-dimensional retention alignment in combination quantitative aspect. However, considering that in the same with bioactivity provided further confidence in the ToF-MS fraction the weak AChE inhibitor lamotrigine was present at an identification using positive ion mode ESI. To compensate for estimated concentration of 1.5 μM, this may add to the the loss in sensitivity associated with flow splitting, the use of explanation of the observed effect. concentrated extracts is required.

2359 DOI: 10.1021/acs.analchem.5b04311 Anal. Chem. 2016, 88, 2353−2360 Analytical Chemistry Article

Other assays with different toxicological end points in Hollert, H.; Tarabek, P.; Tousova, Z.; Slobodnik, J.; Walz, K.; Brack, microplate format may easily be implemented in this approach. W. Jt. Danube Surv. 3 2014, 284−295. (19) Camenzuli, M.; Schoenmakers, P. J. Anal. Chim. Acta 2014, 838, Other chromatographic column combinations may provide a − highly orthogonal fractionation system. For instance, for very 93 101. polar compounds hydrophilic interaction liquid chromatog- (20) Booij, P.; Vethaak, A. D.; Leonards, P. E. G.; Sjollema, S. B.; ff Kool, J.; de Voogt, P.; Lamoree, M. H. Environ. Sci. Technol. 2014, 48 raphy (HILIC) may be of interest. Di erent ionization sources (14), 8003−8011. such as APCI and APPI (atmospheric pressure photoioniza- (21) Perkins, E. J.; El-Alfy, A.; Schlenk, D. Toxicol. Sci. 1999, 48 (1), tion) which are more efficient at ionizing nonpolar compounds 67−73. may also be used. (22) Fontaine, J.; Reuse, J. Eur. J. Pharmacol. 1980, 68,55−60. (23) Waldmeier, P. C.; Baumann, P. A.; Wicki, P.; Feldtrauer, J. J.; ■ ASSOCIATED CONTENT Stierlin, C.; Schmutz, M. Neurology 1995, 45 (10), 1907−1913. (24) Wode, F.; van Baar, P.; Dünnbier, U.; Hecht, F.; Taute, T.; *S Supporting Information Jekel, M.; Reemtsma, T. Water Res. 2015, 69, 274−283. The Supporting Information is available free of charge on the (25) Ferrer, I.; Thurman, E. M. Anal. Chem. 2010, 82 (19), 8161− ACS Publications website at DOI: 10.1021/acs.anal- 8168. chem.5b04311. (26) Thurman, E. M.; Ferrer, I.; Barcelo,D.́ Anal. Chem. 2001, 73 − Additional information as noted in the text (PDF) (22), 5441 5449. (27) Schymanski, E. L.; Singer, H. P.; Longree,́ P.; Loos, M.; Ruff, M.; Stravs, M. a.; Ripolleś Vidal, C.; Hollender, J. Environ. Sci. Technol. ■ AUTHOR INFORMATION 2014, 48 (3), 1811−1818. Corresponding Author *E-mail: [email protected]; [email protected]. Tel.: +31 205989571. Fax: +31 205988941. Notes The authors declare no competing financial interest. ■ ACKNOWLEDGMENTS Authors are grateful to EU FP7 EDA-EMERGE project (EU contract 290100) for supporting this study. ■ REFERENCES (1) Brack, W. Anal. Bioanal. Chem. 2003, 377 (3), 397−407. (2) Legler, J.; van Velzen, M.; Cenijn, P. H.; Houtman, C. J.; Lamoree, M. H.; Wegener, J. W. Environ. Sci. Technol. 2011, 45 (19), 8552−8558. (3) Thomas, K. V.; Langford, K.; Petersen, K.; Smith, a. J.; Tollefsen, K. E. Environ. Sci. Technol. 2009, 43 (21), 8066−8071. (4) Yue, S.; Ramsay, B. A.; Brown, R. S.; Wang, J.; Ramsay, J. A. Environ. Sci. Technol. 2015, 49 (1), 570−577. (5) Bandow, N.; Altenburger, R.; Streck, G.; Brack, W. Environ. Sci. Technol. 2009, 43 (19), 7343−7349. (6) Li, X.; Stoll, D. R.; Carr, P. W. Anal. Chem. 2009, 81 (2), 845− 850. (7) Nagele,̈ E.; Vollmer, M.; Hörth, P. J. Chromatogr. A 2003, 1009 (1−2), 197−205. (8) Stoll, D. R.; Talus, E. S.; Harmes, D. C.; Zhang, K. Anal. Bioanal. Chem. 2014,13−15. (9) Tranchida, P.; Dugo, P.; Dugo, G.; Mondello, L. J. Chromatogr. A 2004, 1054 (1−2), 3−16. (10) van der Horst, A.; Schoenmakers, P. J. J. Chromatogr. A 2003, 1000 (1−2), 693−709. (11) Ouyang, X.; Leonards, P.; Legler, J.; van der Oost, R.; de Boer, J.; Lamoree, M. J. Chromatogr. A 2015, 1380, 139−145. (12) Haun, J.; Leonhardt, J.; Portner, C.; Hetzel, T.; Tuerk, J.; Teutenberg, T.; Schmidt, T. C. Anal. Chem. 2013, 85 (21), 10083− 10090. (13) Soreq, H.; Seidman, S. Nat. Rev. Neurosci. 2001, 2 (4), 294−302. (14) Fukuto, T. R. Environ. Health Perspect. 1990, 87, 245−254. (15) Bartus, R.; Dean, R.; Beer, B.; Lippa, A. Science (Washington, DC, U. S.) 1982, 217 (4558), 408−414. (16) Ellman, G. L.; Courtney, K. D.; Andres, V.; Featherstone, R. M. Biochem. Pharmacol. 1961, 7 (2), 88−95. (17) Hamers, T. Toxicol. Sci. 2000, 58 (1), 60−67. (18) Schulze, T.; Krauss, M.; Novak, J.; Hilscherova, K.; Ait-aissa, S.; Creusot, N.; Macova, M.; Neale, P.; Escher, B. I.; Gomes, T.; Tollefsen, K. E.; Tarcai, Z.; Shao, Y.; Deutschmann, B.; Seiler, T.;

2360 DOI: 10.1021/acs.analchem.5b04311 Anal. Chem. 2016, 88, 2353−2360