Major AhR-active chemicals in sediments of Lake Sihwa, South Korea: Application of effect-directed analysis combined with full-scan screening analysis

This study utilized effect-directed analysis (EDA) combined with full-scan screening analysis (FSA) to identify aryl hydrocarbon receptor (AhR)-active compounds in sediments of inland creeks flowing into Lake Sihwa, South Korea. The specific objectives were to (i) investigate the major AhR-active fractions of organic extracts of sediments by using H4IIE-luc in vitro bioassay (4 h and 72 h exposures), (ii) quantify known AhR agonists, such as polycyclic aromatic hydrocarbons (PAHs) and styrene oligomers (SOs), (iii) identify unknown AhR agonists by use of gas chromatography-quadrupole time-of-flight mass spectrometry (GC-QTOFMS), and (iv) determine contributions of AhR agonists to total potencies measured by use of the bioassay. FSA was conducted on fractions F2.6 and F2.7 (aromatics with log Kow 5–7) in extracts of sediment from Siheung Creek (industrial area). Those fractions exhibited significant AhR-mediated potency as well as relatively great concentrations of PAHs and SOs. FSA detected 461 and 449 compounds in F2.6 and F2.7, respectively. Of these, five tentative candidates of AhR agonist were selected based on NIST library matching, aromatic structures and numbers of rings, and available standards. Benz[b]anthracene, 11H-benzo[a]fluorene, and 4,5-methanochrysene exhibited significant AhRmediated potency in the H4IIE-luc bioassay, and relative potencies of these compounds were determined. Potency balance analysis demonstrated that these three newly identified AhR agonists explained 1.1% to 67% of total induced AhR-mediated potencies of samples, which were particularly great for industrial sediments. Follow-up studies on sources and ecotoxicological effects of these compounds in coastal environments would be


Introduction
Currently available biological tests to detect toxicants remain limited and are failing to provide sufficient information on the compounds responsible for measurable toxic effects (Brack, 2003). Instrumental analyses do not consider how chemicals interact in complex mixtures; thus, are providing little information on potential effects on biota (Brack, 2003;Rijk et al., 2009;Wang et al., 2014). Effect-directed analysis (EDA) represents a powerful tool for identifying chemicals in environmental samples, such as sediments, crude oil, wastewater, and biota (Brack, 2003;Brack et al., 2016;Hong et al., 2016a;Muschket et al., 2018). A typical EDA study consists of a bioassay, fractionation, and instrument analyses. The approach has been successfully utilized to identify and address key toxicants in environmental matrices (Brack, 2003;Brack et al., 2016;Muschket et al., 2018). For instance, EDA was used to investigate major aryl hydrocarbon receptor (AhR) and estrogen receptor (ER) agonists in sediments of the west coast of South Korea (Jeon et al., 2017). Another EDA study using combined in vitro and in vivo assays identified toxicologically active compounds in wetland sediments from North-Eastern Spain (Regueiro et al., 2013). EDA is useful for identifying causative agents in environmental samples, use of targeted chemical analysis alone, cannot fully explain responses of bioassays (Hong et al., 2016b;Jeon et al., 2017;Lee et al., 2017).
Full-scan screening (untargeted) analysis (FSA) has become a widely used technique to find previously unidentified compounds in environmental samples by use of high-resolution mass spectrometry (HRMS) such as time-of-flight mass spectrometry (TOFMS) (Gomez et al., 2009;Ibáñez et al., 2008;Schymanski et al., 2015;Tousova et al., 2018;Zedda and Zwiener, 2012). Because samples might include hundreds to thousands of chemicals, HRMS isolates neighboring peaks that have not been separated by low-resolution mass spectrometry because of the similar m/z; and thus increases selectivity for screening unknown toxicants (Hernandez et al., 2012;Zedda and Zwiener, 2012). However, it is laborious finding key toxicants by use of FSA. Thus, a stepwise approach is necessary, i.e., checking the matching score, identifying molecular formulas in libraries or databases, and then confirmation by use of authentic standards (Booij et al., 2014;Hollender et al., 2017;Muz et al., 2017). When supported by EDA, FSA became more powerful to identify candidate compounds and previously undescribed toxicants in complex environmental media. For example, this approach successfully identified AhR agonists in sediments of the Three Gorges Reservoir in China (Xiao et al., 2016), and previously undescribed anti-androgenic compounds in surface water of a nearby upstream wastewater treatment plant (WWTP) of Silstedt, Germany (Muschket et al., 2018).
Lake Sihwa is an artificial lake surrounded by the cities of Siheung, Ansan, and Hwaseong on the west coast of South Korea (Khim et al., 1999;Khim and Hong, 2014;Lee et al., 2014Lee et al., , 2017. Sihwa industrial complexes include a variety of businesses, such as metal, petrochemical, biochemical, and engineering manufacturing. During the past 20 years, various chemicals originating from industrial areas have been discharged to inland creeks, which has resulted in deterioration of the environment of Lake Sihwa (Hong et al., 2016b;Lee et al., 2014Lee et al., , 2017Yoo et al., 2006). To improve the environment around Lake Sihwa, in 2011, the Korean Government designated a special coastal management zone, and in 2013, implemented a system to manage total loadings of contaminants (Hong et al., 2016b;Lee et al., 2014). Although, in recent years, management efforts have been applied to the environment of Lake Sihwa, studies over the last five years have repeatedly reported that sediments of inland creeks remain polluted by various persistent toxic substances, such as polycyclic aromatic hydrocarbons (PAHs), styrene oligomers (SOs), and alkylphenols (APs) (Hong et al., 2016bJeon et al., 2017;Lee et al., 2017;Meng et al., 2017). Results of several studies have shown that concentrations of PAHs and APs in sediments of inland creeks of Lake Sihwa exceeded interim sediment quality guidelines (ISQGs) established by the Canadian Council of Ministers of the Environment (CCME) (CCME, 2002;Hong et al., 2016b;Lee et al., 2017). Among these classes of compounds, PAHs and SOs are AhR agonists (Eichbaum et al., 2014;Hong et al., 2016b;Xiao et al., 2017).
The AhR is a ligand-activated transcription factor that mediates a wide range of biological and toxicological effects (Giesy et al., 2002;Mitchell and Elferink, 2009). Binding of xenobiotics to the AhR initiates a variety of biochemical, physiological, and toxicological effects, including carcinogenicity, developmental toxicity, and the development of tumors in some organisms (Mimura and Fujii-Kariyama, 2003;Quintana, 2013). Due to such complexity in toxicity assessment relating to PAHs, AhR binding affinity in organic extracts of sediments contaminated by AhR-active PAHs could be screened in vitro. Furthermore, identification of unknown AhR-mediated potency in environmental matrices addresses overall sample toxicity which gives important information for protection of aquatic life as well as human health. Screening for AhR-mediated potency by use of in vitro bioassay can be useful to evaluate contamination by toxic substances and potential adverse effects in environments.
The present study investigated AhR-active compounds in sediments of an industrialized area of Lake Sihwa. Specific objectives were to: (i) investigate AhR-active fractions in organic extracts of sediments by use of the H4IIE-luc bioassay; (ii) measure concentrations of known AhR agonists, such as PAHs and SOs; (iii) identify previously untargeted AhR agonists by use of a combination of the H4IIE-luc bioassay and GC-QTOFMS; and (iv) determine contributions of traditional and newly identified AhR agonists to total induced potencies.

Collection and preparation of samples
Sediments were collected from inland creeks (tributaries) of industrial (C1-C5) and urban (C6) areas in April 2015 and from rural areas (C7-C8) in September 2017 at Lake Sihwa, South Korea (Fig. 1). Although samples were collected in different years, sediments record a relatively long history of contamination by persistent toxic substances, including the target compounds of the present study, compared to that of water samples. Thus, in this study, sampling time was not considered during the interpretation of data. Surface sediments were collected with

Sihwa & Banwol Industrial Complex
Fig. 1. Sites from which sediments were collected from the inland creeks (tributaries) of Lake Sihwa, South Korea. J. Cha, et al. Environment International 133 (2019) 105199 a hand shovel and transferred to pre-cleaned glass jars. Sediments were transported to the laboratory and stored at −20°C until analysis. Detailed descriptions of preparation of samples for bioassay and chemical analysis have been described previously (Hong et al., 2015(Hong et al., , 2016bLee et al., 2017). In brief, sediments were freeze-dried, passed through a 1mm sieve, and homogenized. Approximately 60 g of homogenized sediments were placed into the thimble and extracted on a Soxhlet extractor with 350 mL dichloromethane (DCM, J.T. Baker, Phillipsburg, NJ) for 16 h. To remove elemental sulfur from organic extracts, activated copper was added for about 1 h. Organic extracts were concentrated to 4 mL with a rotary evaporator and N 2 gas flow (~15 g sediment equivalent (SEq) mL −1 ). Four milliliters of raw extract (REs) were divided into 2 mL volumes for silica gel column fractionation and H4IIE-luc bioassay. The REs separated for bioassay were exchanged into dimethyl sulfoxide (DMSO, Sigma-Aldrich, Saint Louis, MO).

Silica gel and RP-HPLC fractionations
To perform fractionation based on polarity, approximately 8 g of activated silica gel (70-230 mesh, Sigma-Aldrich) was added to a glass column. Two milliliters of organic extract was passed through the column and divided into non-polar (F1), aromatics (F2), and polar (F3) fractions. F1 was eluted with 30 mL hexane (Honeywell, Charlotte, NC). F2 was collected with 60 mL of 20% DCM in hexane. F3 was eluted with 50 mL of 60% DCM in acetone (J.T. Baker). Eluted fractions were evaporated on a rotary evaporator and concentrated to 2 mL using N 2 gas flow. Of the 2 mL silica gel fraction samples, 1 mL was further fractionated into ten finer fractions based on log K ow values using reverse-phase (RP)-HPLC (Agilent 1260 HPLC; Agilent Technologies, Santa Clara, CA) (for details, see Hong et al., 2016aHong et al., , 2016b.

H4IIE-luc in vitro bioassay
AhR-mediated potencies were measured by use of the H4IIE-luc bioassay in REs, the silica gel fractions, and RP-HPLC fractions of sediments. The bioassay was performed following the methods of previous studies (Hong et al., 2016b;Lee et al., 2017). In brief, trypsinized cells (~7.0 × 10 4 cells mL −1 ) were seeded on a 96-well plate at 250 µL per well. After seeding, the cells were incubated at 37°C in a 5% CO 2 incubator for 24 h. Dosing was carried out by adding the sample to the well-cultured cells. Plates contained a positive control, such as benzo[a] pyrene (BaP) or 2,3,7,8-tetrachloro dibenzo-p-dioxin (TCDD), a solvent control (0.1% DMSO), and a media control. BaP (for 4 h) and TCDD (for 72 h) were diluted three times with 50 nM (=100 %BaP max ) and 300 pM (=100 %TCDD max ) as the first concentration, respectively. Dosing was performed at six concentrations. Luciferase luminescence was quantified using a Victor X3 multi-label plate reader (PerkinElmer, Waltham, MA) after 4 h or 72 h of exposure. Responses were converted to percentages of the maximum responses of BaP and TCDD, respectively. Finally, magnitude-based BaP-EQ values and potency-based BaP-EQ values (ng BaP-EQ g −1 dm) for AhR-mediated potency at 4 h were calculated. Potency-based BaP-EQ values were obtained from sample dose-response relationships elicited by sediment samples at six dilutions.

Full-scan screening analysis
FSA using GC-QTOFMS was performed on F2.6 and F2.7 of the organic extract of sediment from Siheung Creek (C4), where AhR-mediated potency was greater. The gas chromatograph Agilent 7890B coupled with a 7200 QTOFMS (Agilent Technologies) was used for FSA. The carrier gas was 1.2 mL min −1 He. A DB-5MS UI column (30 m × 0.25 mm i.d. × 0.25 µm film) was used for separation. Instrumental conditions are detailed in Table S1. The selection criteria of candidates for AhR agonists from the GC-QTOFMS analysis consisted of four steps. The first step involved matching the compounds with the NIST library (ver. 2014) (Booij et al., 2014;Zedda and Zwiener, 2012), and removing targeted PAHs and SOs. The second step involved selecting only compounds with a score of ≥70 by identifying the library matching factors (Muz et al., 2017). The third step involved identifying aromatic compounds. A previous study showed that compounds with the structure of aromatic rings had AhR binding affinity (Mekenyan et al., 1996). The fourth step involved selecting compounds with three or more aromatic rings. Because AhR-active PAHs have more than three benzene rings, compounds with three or more rings were selected (Louiz et al., 2008;Xiao et al., 2016). From the FSA, five commercially available compounds including benz[b]anthracene (BbA), cyclopenta [cd]pyrene (CcdP), 11H-benzo[a]fluorene (11BaF), 4,5-methanochrysene (4,5MC), and 1-methylpyrene (1MP) were selected to conduct chemical and toxicological confirmations. BbA and 1MP were obtained from Sigma-Aldrich and 11BaF and CcdP were purchased from Accustandard (NewHaven, CT). 4,5MC was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The five putative AhR agonists in fractions were quantified using GC-MSD.

Relative potency values of putative AhR-active compounds
To determine relative potencies (RePs) for AhR-mediated potencies of the five candidate compounds, H4IIE-luc bioassay was performed at 4 h. Each compound was prepared at a total of eight concentrations (viz., 1000, 333, 111, 37, 12, 4.1, 1.4, and 0.46 ng mL −1 ), and was analyzed using the in vitro bioassay method described in Section 2.3. Effective concentrations (EC) and RePs were measured based on previous studies (Villeneuve et al., 2000), with minor modifications. In brief, ReP 20 , ReP 50 , and ReP 80 of the candidate compounds were calculated based on the EC 20 , EC 50 , and EC 80 , respectively. If the deviation between ReP 20 , ReP 50 , and ReP 80 was within an order of magnitude, ReP 50 was generally considered reliable (Horii et al., 2009).

Potency balance analysis
Potency balance analyses were performed between instrument-derived BEQ and bioassay derived BaP-EQs (potency-based concentrations) to determine contributions of each compound to AhR-mediated potency. Instrument-derived BEQ concentrations were calculated as the sum of products of measured concentrations of individual compounds multiplied by their RePs. RePs of AhR-active PAHs and SOs are presented in Table S2 (Hong et al., 2016b;. The BEQ concentrations of newly identified AhR agonists were calculated using chemical concentrations, with this study delineating the ReP values. The contribution of instrument-derived BEQs of individual compounds (PAHs, SOs, and newly identified AhR agonists) to bioassay derived BaP-EQs was determined.

QSAR modeling using VEGA
The VEGA platform is an in silico program that contains tens of quantitative structure-activity relationship (QSAR) models for various endpoints (Marzo et al., 2016). In silico techniques are used to predict the toxicological endpoints of a chemical-based on its physicochemical properties and structure (Pizzo et al., 2013). Five candidate compounds identified by FSA were evaluated for mutagenicity, carcinogenicity, developmental toxicity, and estrogen receptor activity .

AhR-mediated potencies in sediments
Detectable AhR-mediated potencies were found in extracts of sediments from inland creeks at both 4 h and 72 h exposures in the H4IIEluc transactivation bioassay with luciferase as the reporter gene ( Fig.  S2). All REs of sediments reached saturation efficiency (≥100 % BaP max ) after 4 h exposure, while the AhR-mediated potencies of REs at 72 h exposure varied among sites. In the three silica gel fractions of eight sediment REs, F2 (aromatics) and F3 (polar) generally exhibited greater AhR-mediated potencies than did F1 (non-polar) at both 4 h and 72 h exposure durations. This result was expected because F2 mainly contains compounds capable of binding to the AhR, including PCDD/Fs, coplanar PCBs, and PAHs (Hong et al., , 2016bKinani et al., 2010;Lee et al., 2017;Louiz et al., 2008). Industrial areas of Siheung Creek (C4) and Singil Creek (C5), an urban area of Ansan Creek (C6), and a rural area of Samhwa Creek (C7) were then subjected to RP-HPLC fractionation, in which the AhR response of F2 was the greatest in each area. Significant AhR-mediated potencies were commonly observed in F2.6-F2.8 at 4 h exposure duration at sites C4-C7 (Fig. 2). These fractions contained aromatic compounds with 5-8 log K ow values, indicating that these compounds are major AhR agonists. For example, the well-known AhR agonists (such as PAHs and SOs) were detected in these fractions. BaA, Chr, SD1, and SD3 occurred in F2.6, and BbF, BkF, BaP, IcdP, DbahA, and ST2 occurred in F2.7 (Hong et al., 2016b;Lee et al., 2017). Similar patterns have been observed in previous studies conducted in Lake Sihwa and the west coast of South Korea (Hong et al., 2016b;Jeon et al., 2017). However, it was not possible to determine whether the greater AhR responses found in F2.6 and F2.7 in different sediment samples were caused by the same substances across regions, or to identify the common physico-chemical characteristics of AhR-active compounds.
Simultaneous tests at durations of exposure of 4 h or 72 h in the H4IIE-luc bioassay provided metabolic information on the AhR agonists present in sediments. Labile compounds (such as PAHs) tended to be metabolized as a function of increasing exposure time; however, refractory compounds (such as PCDD/Fs and coplanar-PCBs) were relatively stable during the exposure of 72 h (Xiao et al., 2017). The present study identified strong AhR-mediated potencies after exposure for either 4 h or 72 h in the H4IIE-luc bioassay. Due to the relatively strong AhR-mediated potency, this study focused on identifying causative compounds during 4 h exposures. In addition, F3 exhibited significant AhR-mediated potency. In the RP-HPLC fractions of F3, significant AhRmediated potencies were observed in F3.5-F3.8 after 4 h exposure, while weaker responses were observed after 72 h exposure (Fig. S3). Thus, AhR agonists occurred in polar fractions and were actively metabolized during the longer exposure. Because it is difficult and challenging to identify AhR-active compounds in the polar fraction (Creusot et al., 2013;Hong et al., 2016b;Lubcke-von Varel et al., 2011), further studies focusing on polar AhR agonists in sediments are needed.
The full dose-response curves of selected fractions (F2.6 and F2.7 of C4-C7 sediment extracts) were obtained (Fig. S4). Responses of some fraction samples were insufficient (< 50 %BaP max ) to calculate EC 50 values (Lee et al., 2013). Thus, potency-based BaP-EQ concentrations were calculated based on EC 20 values. These concentrations of potencybased BaP-EQs were used for the potency balance analysis (vs. instrument-derived BEQ concentrations).

Absolute and relative concentrations of PAHs and SOs
PAHs and SOs were detected in extracts of all sediments collected from inland creeks of Lake Sihwa. Concentrations of sedimentary PAHs ranged from 160 to 1400 ng g −1 dm (mean: 460 ng g −1 dm) in industrial areas (C1-C5), 310 ng g −1 dm in more urban areas (C6), and from 14 to 120 ng g −1 dm in rural areas (C7-C8) ( Table S3). Concentrations of PAHs in extracts of sediments from industrial areas were generally greater than those in extracts of sediments from more urban and rural areas. Thus, contamination by sedimentary PAHs seems to be associated with land use and other surrounding activities at nearby industrial complexes. Concentrations of PAHs measured in sediments from inland creeks of Lake Sihwa were compared with several existing SQGs, such as effect-range-low and -median values (ERL and ERM) , threshold and probable effect concentrations (TEC and PEC) , and ISQGs and probable effect levels (PEL) (CCME, 2002). Some PAHs including Ace, Flu, Phe, Fl, Py, and DbahA in the C4 sediment exceeded threshold effect concentration guidelines (i.e., ERL, TEC, and ISQG), but none exceeded probable effect concentration guidelines (i.e., ERM, PEC, and PEL) (Fig. S5). To assess potential sources of PAHs, diagnostic paired ratios were applied, including Ant/(Ant + Phe), Fl/(Fl + Py), and IcdP/(IcdP + BghiP) (data not shown). Ratios calculated indicated that sources of sedimentary PAHs from the inland creeks of Lake Sihwa were pyrogenic, including grass, wood, coal, and fossil fuel combustion (Yang et al., 2014). Patterns of distributions of sedimentary SOs were generally similar to those of PAHs. Concentrations of SOs ranged from 89 to 870 ng g −1 dm (mean: 590 ng g −1 dm) in more industrial areas, 230 ng g −1 dm in urban areas and from 150 to 300 ng g −1 dm in rural areas, respectively (Table S4). In industrial areas, C5 sediment had the greatest concentration of SOs, followed by C4, C1, C2, and C3. Unlike the unwanted by-product PAHs, SOs are mainly derived from polystyrene plastic products; thus, the types of industries present around the inland creek might have influenced the distribution of SOs .
Greater concentrations of AhR-active PAHs and SOs were found in sediments of inland creeks in the industrial areas (C4 and C5) compared to those from urban (C6) and rural (C7) areas (Fig. 3a). To determine contributions of known AhR agonists, instrument-derived concentrations of BEQs were calculated using concentrations of AhR-active PAHs and SOs and their ReP values ( Fig. 3b and Table S5). Potency balance analysis between instrument-derived BEQs and bioassay-derived BaP-EQs was performed on F2.6 and F2.7 of C4-C7 sediments (Fig. 3b). The results showed that known AhR agonists (such as PAHs and SOs) explained only a small portion of total AhR-mediated activities, ranging from 1.1 to 39% for F2.6 and 1.2 to 6.4% for F2.7, respectively. Unknown (i.e., untargeted) AhR-active compounds were largely present in the inland creek sediments of Lake Sihwa. In F2.6 of C4, Chr (18%) and BaA (5.6%) showed the greatest contribution to AhR-mediated potencies. In comparison, BbF (2.1%), BkF (1.9%), and BaP (1.9%) were the major AhR agonists in F2.7 of C4. Although concentrations of SOs were comparable to concentrations of PAHs, their contributions were small (< 1%) because of their small RePs. Overall, potency balance analysis showed that known AhR-active PAHs and SOs did not fully explain the bioassay results for sediments. Thus, FSA was performed to search for previously unidentified AhR agonists present in more potent fractions.

Full-scan screening analysis of highly potent fractions
FSA was performed on F2.6 and F2.7 of C4 sediment (Siheung Creek) using GC-QTOFMS. The process for selecting the causative compounds consisted of four steps (Fig. 4a). In the first step (NIST library matching, ver. 2014), 461 and 449 compounds were detected in F2.6 and F2.7 of C4 sediment extracts, respectively. Among them, the number of compounds with a matching score of ≥70 (i.e., second step) was 267 in F2.6 and 214 in F2.7. The third step involved identifying aromatics, with 129 and 93 compounds being selected, respectively.  The fourth step involved selecting compounds with three or more aromatic rings, with a total of 13 and 27 compounds being selected, respectively (Table S6). Six compounds were detected both in F2.6 and in F2.7. Thus, a total of 34 compounds were selected as tentative candidates of AhR agonists from the organic extracts of sediment (Table S6 lists these compounds). The common characteristics of these compounds are (i) molecular mass 200-300 (average 240), (ii) log K ow 5-7, and (iii) methyl-substituted structures. Due to the lack of authentic standards, among the 34 compounds, five commercially available compounds including BbA, CcdP (in F2.6), 11BaF, 4,5MC, and 1MP (in F2.7) were tested for identification of compounds and toxicological confirmations. Overall, this study successfully applied FSA to select candidate AhR agonists in more potent fractions. However, the approach had several limitations, including being dependent on library matching. Using a high-resolution mass spectrometer, the molecular formula is somewhat reliable; however, the chemical structure cannot be characterized. To solve this problem, it would be necessary to confirm the structure using standards, but available standards are limited.

Chemical and toxicological confirmation
Chemical and toxicological confirmation of the five compounds (BbA, CcdP, 11BaF, 4,5MC, and 1MP) were conducted to determine concentrations in sediments and AhR binding potencies ( Fig. 4b and Table 1). GC retention time and mass fragment ions of the candidate compounds with those of samples were matched using GC-MSD. Concentrations of BbA, CcdP, 11BaF, 4,5MC, and 1MP in inland creek sediments were quantified. For toxicological confirmation, dose-response tests for five candidates were performed after exposure for 4 h of H4IIEluc cells (Fig. 4b). Among the five candidates, three compounds (BbA, 11BaF, and 4,5MC) showed significant AhR-mediated potencies (significant level = 5%BaP max ), thus ReP values relative to that of BaP could be obtained. RePs of BbA, 11BaF, and 4,5MC were 10.6, 1.2, and 1.0, respectively (Fig. 4b). Variation among ReP 20 , ReP 50 , and ReP 80 was relatively small, and thus, the use of ReP 50 was considered reliable (Table S7) (Lee et al., 2013). BbA was a strong AhR agonist, with its ReP being 10-fold greater than that of BaP. ReP values of 11BaF and 4,5MC were comparable to that of BaP. BbA and 11BaF were previously reported as AhR-active chemicals (Larsson et al., 2014), but have not been reported in environmental samples to date. BbA and 11BaF showed AhR-mediated potencies along exposure durations of 24, 48, and 72 h in the H4IIE-luc bioassay, which indicated that these chemicals are not easily metabolized .

Distribution and sources of newly identified AhR agonists in sediments
In this study, three AhR-active chemicals were newly identified from sediment fractions through EDA combined with FSA, which were all detected in sediments from the inland creeks of Lake Sihwa (Table  S8). Relatively great concentrations of 11BaF (253 ng g −1 dm), 4,5MC (49 ng g −1 dm), and BbA (19 ng g −1 dm) were detected in sediment of C4, followed by C5, C6, and C7. Similar to distributions of PAHs and SOs, these newly identified AhR agonists predominated in industrial areas. Of the three AhR agonists, 11BaF was the most frequently detected with the greatest concentrations among sampling sites. In comparison, BbA was primarily detected in sediments from industrial areas. Greatest concentrations of 11BaF were observed in the industrial, urban, and rural areas, and mainly originated from gasoline engines, tobacco smoke, and oil-fired heating (Snook et al., 1978). Previous studies reported that 4,5MC is derived from urban particulate matter and tobacco smoke (Agarwal et al., 1999;. 4,5MC concentrations were comparable in C4 and C5 sediments (industrial area) and were 6-8 times greater than those in urban and rural areas (C6 and C7). BbA is an essential element for the active layer of highperformance organic field-effect transistors (OFETs) and organic lightemitting diodes (OLEDs). This compound was identified as the active material, because of its high hole OFET mobility in the single-crystal form (Gundlach et al., 2002;Takahashi et al., 2007;Yamamoto and Takimiya, 2007). BbA was primarily distributed in the C4 and C5 sediments of the industrial areas compared to urban and rural areas.
Studies on occurrences and distributions of 11BaF, 4,5MC, and BbA in sediments remain limited globally (Kishida et al., 2007). The concentration of BbA in sediment from C4 was greater than those recorded in the suburban and rural areas of Vietnam, but was lesser than those observed in some industrial areas (Kishida et al., 2007). Unlike 4,5MC and 11BaF, which originate from combustion processes, BbA originates from transistor film, which was widely used in the industrial area. BbA has a very high affinity to bind AhR and contributes to the total induced  AhR-mediated potency (see Section 3.6 for more details); thus regulations and monitoring would be of immediate concern.

Contributions of newly identified AhR agonists to total induced potencies
To address relative contributions of traditional and newly identified AhR agonists to total AhR-mediated potencies, instrument-derived BEQs using all the identified AhR-active chemicals were calculated ( Fig. 5 and Table S5). The potency balance suggested varying contribution among chemicals and sites. For example, BbA was the greatest contributor, explaining 67% of the total AhR-mediated potency in F2.6 of the C4 sediment organic extract. 11BaF and 4,5MC explained 22% and 2.9% AhR-mediated potency of F2.7 in the C4 organic sediment extract. Thus, these three newly identified AhR agonists contributed significantly to the total induced AhR-mediated potencies compared to the known traditional AhR agonists, such as BaA, Chr, BkF, etc. Addition of these three AhR agonists increased the explanatory power of total induced AhR-mediated potencies for the F2.6 and F2.7 fractions of inland creek sediments. Meantime, it should be noted that compositions and contributions of AhR agonists are site-specific. The three AhR agonists were observed in F2.6 and F2.7 fractions of the C4 sediment extract (industrial area). These compounds occurred predominantly in industrial areas and appeared to act as major AhR agonists. However, BbA, 11BaF, and 4,5MC showed relatively minor contributions (< 3%) to the fractions of organic sediment extracts from urban and rural areas.  Although large AhR-mediated potencies were detected in the sediments of urban (C6) and rural (C7) areas, a large proportion of such elevated AhR potencies has yet to be identified.
The five candidate compounds selected in this study (BbA,CcdP,11BaF,4,5MC,and 1MP) occurred widely in sediments from inland creeks (Table S8). Whether these compounds had other potential toxicities, such as mutagenicity, carcinogenicity, developmental toxicity, and estrogen activity using VEGA QSAR was further evaluated (Table  S9). Consequently, it was postulated that these compounds have other potential toxicities in aquatic environments. Specifically, it was likely that all these compounds exhibited mutagenicity, carcinogenicity, and developmental toxicity. CcdP and 1MP were prominent in sediments of the inland creeks, mainly originated from automobile exhaust, atmospheric soot (Eisenstadt and Gold, 1978), crude oil, and meat and charcoal smoke (Dyremark and Westerholm, 1995). In addition, these substances have been reported to have specific potential toxicities in several previous studies (Table S9). For example, BbA is known to have potential AhR activity  this study) and mutagenicity , CcdP has potential genotoxicity and tumorgenicity , 11BaF has AhR activity  this study), 4,5MC has AhR activity (this study) and mutagenicity (Lee-Ruff et al., 1987), and 1MP has mutagenicity and tumorgenicity . Since these emerging substances are widely distributed in sediments near industrial complexes, further studies on their sources, fate, potential adverse effects, and reduction measures are necessary.

Conclusions
Results of the present study indicated that EDA combined with FSA enhanced the understanding of the distributions of bioactive chemicals and potential toxicities associated with environmental mixture samples. We have identified and addressed the contributions of known and newly identified AhR-active compounds from sediment extracts in a highly complexed area of varying land uses and activities. This study demonstrated that three newly identified AhR agonists from sediments act as strong AhR agonists and contributed a large proportion of sample activities, particularly for the samples collected from highly industrialized areas. However, the results of the H4IIE-luc bioassay used in the present study do not allow direct extrapolation to ecotoxicological responses, but provide baseline information on pollution of AhR-active compounds in sediments for further risk evaluation. Thus, further studies are needed to confirm whether the newly identified AhR agonists have toxic effects on living aquatic organisms, which will provide better understanding of ecologically relevant predictions of risk in the contaminated sediments. Meanwhile, the stepwise screening approach throughout non-target analysis applied in this study would benefit to identify and address unknown causative agents in environmental samples, elsewhere. In addition, candidate compounds that are not AhR agonists might exhibit other toxicity mechanisms; thus it is crucial to characterize all the aspects of sources, distribution, fate, and ecotoxicological effects of target compounds in the aquatic environment.

Sites
Concentrations of five candidates for AhR agonists (ng g -1 dm)  S1. Instrumental conditions of GC-MSD for PAH and SO analyses.  ; TEC and PEC: threshold and probable effect concentrations ; and ISQG: interim sediment quality guidelines, PEL: probable effect levels (CCME, 2002)).