In total 150 samples, 75 from bottom sediments, and 75 from recent sediments or sediment traps were collected from five different sampling sites (Malenovice, Belov, Spytihnev, Certak, and Certak oxbow lake) in the Morava River and its tributary Drevnice River, both located in the south-eastern part of the Czech Republic (Figure 1). Samples were periodically collected every 28 days from June 2007 to July 2008, with 15 samplings throughout the season. Near to Zlín city, Malenovice locality is situated on the Drevnice River in an area with strong presence of industries in rubber, footwear and mechanical engineering. Belov site is located in an agricultural area and reflects the situation in Morava River upstream from the confluence with Drevnice River. Spytihnev locality is situated downstream and integrates the contamination from both Morava and Drevnice rivers. Certak and Certak oxbow lake are both located on Morava River. Certak represents the normal watercourse, while Certak oxbow lake is part of a channel that remained abandoned since 1930s, when new channel regulations were implemented.

Two types of samples, i.e., bottom sediments and fresh recent sedimented material were collected. Bottom sediments samples (top 0–10 cm layer) were collected using the plastic trowels. Recent sediments were taken using glass sediment traps placed close to the river bed but avoiding direct contact with bottom sediments. Unwanted debris such as large pieces of wood, leaves and stones were removed from all samples. The samples were then homogenized and subsequently freeze-dried (Gamma 1–16 LSC, Christ). Dry sediments were then sieved (2 mm mesh). Samples were clustered according to the four hydrological seasons, i.e., spring (March–May), summer (June–August), autumn (September–November), and winter (December–February), being October–March and April–September the colder and warmer seasons, respectively.

Sediment samples were characterized by measuring the content of total organic carbon (TOC) using high-temperature LiquiTOC II analyser (Elementar Analysensysteme). The grain size distribution was determined using a Retsch AS 200 sieving machine (Retsch, 0.036–4 mm fractions) and a Cilas 1064 laser diffraction granulometer (IVA Industrieberatung, 0.0004–0.5 mm fractions).

Freeze-dried sediments (10 g) were extracted with dichloromethane (DCM) using automated Büchi B-811 extractor (Büchi). The concentrated extracts were cleaned up on silica column (for PAHs) and H₂SO₄ modified silica column (organohalogens). Copper powder was used to remove sulfur. Samples were analyzed using GC-MS (Agilent 6890N GC/Agilent 5973N MS) supplied with a J&W Scientific fused silica column DB-5MS for 16 U.S. Environmental Protection Agency polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs, congeners 28, 52, 101, 118, 138, 153, and 180), organochlorine pesticides (OCPs, namely α-, β-, γ-, and δ-isomers of hexachlorocyclohexane, HCH, p,p′-dichlorodiphenyltrichloroethane, p,p′-DDT, and metabolites p,p′-DDE and p,p′-DDD, and finally hexachlorobenzene, HCB). Concentrations of chemicals were quantified using Pesticide Mix 13 (Dr. Ehrenstorfer, GmbH) and PAH Mix 27 (Promochem) standard mixtures.

Concentrations of metals that are known to be toxic and represent environmental hazards in waters and sediments (V, Cr, Co, Ni, Cu, Zn, As, Mo, Cd, Sb, Pb, and Hg) were determined according to ISO 11466 method adapted to analytical instrumentation. Nitro-hydrochloric acid (2.3 ml HNO₃ and 7.0 ml HCl) was used to leaching metals from sediments (1 g dry weight). Analyses were performed with using ICP-MS on Agilent 7500ce instrument (Agilent). Mercury in samples was determined with a thermo-oxidation method using AMA-254 analyser (Altec).

Kinetic bioassays were performed in microplates as previously described (Bláha et al., 2010), with minor modifications. Given that the flash protocol is compatible with colored sediment matrices, toxicity in sediment samples was tested by direct contact between V. fischeri and the solid phase (Lappalainen et al., 1999, 2001). Sediment suspensions and aqueous elutriates were prepared according to European Norm (EN 12920:2006+A1:2008).

Sediment suspensions (100 mg dry wt sed./ml of distilled water, pH 7.0 ± 0.2) were pre-prepared by shaking (vortex, 2000 rpm) for 5 min and tested immediately or stored at 4°C for no more than 24 h before testing. Suspensions were then shaken once again (30 s) immediately before testing, and 0.15 ml were pipetted into micro-plate wells along with 0.01 ml of 32% w/v sodium chloride (NaCl) to reach the final concentration of 2% NaCl necessary for marine V. fisheri. Serial dilutions (1:1) for each sample were directly prepared in 2% w/v NaCl (pH 7.0 ± 0.2) in the microplate. Final concentrations tested (considering dilution with bacterial inoculum) were 75, 37.5, 18.75, 9.38, and 4.69 mg dry wt sed./ml. Elutriates were prepared in microtubes with sediment suspensions (100 mg dry wt sed./ml) shaken at 15 rpm for 24 h using a slow multi rotator. Suspensions were then centrifuged at 7000 rpm for 10 min and supernatants were diluted as previously described (serial dilutions 1:1 in 2% NaCl, 5 concentrations).

Microplates containing sample aliquots (0.08 ml of serial dilutions of sediment suspensions or elutriate supernatants), positive controls (K₂Cr₂O₇, 380 mg/ml of 2% NaCl) and negative controls (2% w/v NaCl) were temperate at 15°C and toxicity tested as described below.

Freeze-dried luminescent bacteria V. fischeri NRRL B-11177 (Institute of Microbiology, Czech Academy of Sciences) were reconstituted according to ISO 11348-3 (1998) in ice-cold 2% w/v NaCl and kept on ice. Prior to testing, aliquots of bacterial suspension were diluted in 2% NaCl and tempered at 15°C in a water bath for 30 min. 0.02 ml of bacteria suspension were injected into each well of the microplates with studied samples, and the luminescence was immediately monitored for 2 s (initial peak value). The luminescence was measured using a microplate Luminoscan Ascent luminometer (Thermo) equipped with computer-controlled injectors. The microplate was then shaken inside the luminometer incubated at 15°C and the final toxicity signal recorded after 30 s (S30-value). The inhibition of luminescence (percentage of control) was calculated as follows:

where CF is a correction factor (the S30/peak ratio in negative controls) reflecting natural attenuation of bacterial luminescence during 30 s exposures.

Concentrations of sediment suspensions that caused 50% inhibition of luminescence (IC₅₀ in mg dry weight/ml) were derived from the four-parametric logistic curve calculated in Graph-Pad™ Prism (GraphPad Software). Because sediment elutriates rarely caused 50% decreases in luminescence, the IC₅₀-value was replaced in those cases by the INH₇₅-value, which was defined as a decrease in the light emission observable after 30 s in a non-diluted aqueous extract corresponding to 75 mg dw sed./ml.

Standard summary statistics were used to describe distributions of the primary data. Non-parametric analyses were applied, because assumptions of parametric analyses (normality of data distribution) were not found in all variants. Correlations between sediment parameters and toxicity were tested using Spearman's rank correlation, differences were tested by Wilcoxon paired test. For all of the tests, p < 0.05 were considered statistically significant. Calculations were performed in Statistica (StatSoft, Tulsa, OK, USA).

To better visualize the results, the individual data points within each analyzed parameters (for chemical analyses results as well as toxicity) were categoried into four groups (I–IV). This approach allowed for better relative comparisons of different analyzed parameters, where the actual numerical values ranged across orders of magnitude. Group I represented the smallest levels of contamination (or the lowest toxicity, respectively), and the group IV contained the highest values of contamination or toxicity, respectively. The groups I–IV were formed as follows: the minimum (MIN0) and maximum (MAX0) of all detected values of a certain parameter were cut off to eliminate extremes. The values falling between the second minimal (MIN) and the second maximal (MAX) values were then divided into four groups (I–IV) based on the statistical distributions. For each parameter (i.e., toxicity value, organic carbon-TOC, or concentrations of individual contaminants), the upper limits (L) for each of the 4 groups were calculated:

where a = 1 or 2 or 3 or 4 for Groups I–IV, respectively. Examples of criteria (limits, L) for groups I, II, III, and IV for selected parameters (TOC, sum of PAHs, total Cd concentrations) are shown in Supplementary Table S1.

In general, aquatic elutriates were significantly less toxic compared to the suspensions from the same sampling site. In elutriates the 50% inhibitory effects were observed only rarely, and the toxicity could be described only as inhibitions of luminescence (%) at the highest tested concentration of sediment elutriate (corresponding to 75 mg dw/mL, Supplementary Table S3). Sediment suspensions showed in general more pronounced effects allowing for calculation of IC₅₀ (Supplementary Table S2 and Table 1). However, the toxic responses of V. fisheri to elutriates and suspensions were significantly correlated (Spearman rank coefficient for all samples N = 121, Rs = 0.849, p < 0.0001).

Detailed investigation of the toxicity results of suspended materials revealed that recent sediments from traps (IC₅₀ median values ranging from 7.8 to 53.7 mg dry weight/ml) were generally more toxic than bottom sediments (IC50 23.6 to >75 mg dry weight/ml; Table 1). The difference was confirmed by Wilcoxon Matched Pairs Test (N = 47, Z = 5.46, p < 0.0001).

Toxicity of both bottom and freshly trapped sediments showed a clear seasonal pattern for all sampling sites (Table 1, Supplementary Table S2, Figure 2). The toxicity increased during the winter period and reached the lowest IC50-values at the end of February or beginning of March. Systematically lowest toxicity was detected during summer months.

With respect to bottom sediments, the lowest toxicity was systematically found at Certak and Certak oxbow lake, the two sites further downstream away from the city of Zlín (the highest IC50-values in Table 2 and Figure 3A). The pattern is also clear from Figure 4A—bottom sediments), where distributions of 1/IC50 toxicity values are shown as pie charts. On the other hand relatively higher toxicities of bottom sediments were found at Malenovice and Belov sites. For fresh sediments from traps (Figures 3B, 4B), toxicity results did not show as apparent spatial differences between sites as for bottom sediments. The site with the highest toxicity was again Malenovice and interestingly, the second highest toxicities were found at Certak oxbox-lake.

The complete results of analysis of total 17 different contaminant groups—five organic (sums of PAHs, PCBs, HCHs, DDTs, and HCB) and 12 toxic metals in studied bottom and fresh sediments are shown in Supplementary Table S4 (EXCEL). To better visualize and compare the chemical loads, the values were transferred to relative groups I–IV as described in Materials and Methods Section, and the aggregate results per individual sites are shown in Figure 4A (bottom sediments) and Figure 4B (sediment traps). Bottom sediments were most polluted at Spytihnev site (almost 60% of the values were in the highest contamination Groups III or IV—cf Figure 4A—Spytihnev pie charts for contamination). Interestingly, the samples from this site did not elicit major toxic effects (cf Figure 4A—piecharts for 1/IC50-values). On the contrary, the lowest levels of bottom sediment contamination were detected in Certak oxbow lake, and the findings corresponded to the pattern of toxicities (cf Figure 4A). With regard to freshly trapped sediments (Figure 4B) no major differences were apparent among the sites in contaminant levels or toxicity (as apparent from the distribution patterns shown in pie-charts).

Correlation analyses between the toxicity values (1/IC₅₀ for solid phase test and INH₇₅ for elutriates) and 17 contaminant parameters and total organic carbon (TOC) are shown in Table 2 (bottom sediment samples) and Table 3 (sediment traps).

In bottom sediments (Table 2), 16 out of 18 variables correlated with the IC₅₀-values from tests with solid phase suspensions when considering data from all sites (N = 75). The toxicity of elutriates (INH75-values) correlated with only eight parameters. In the sediments from Malenovice and Belov (where the toxicity values were variable and widely distributed—cf Figure 2A) the highest number of significant correlations was observed for both 1/IC50 and INH75 (especially between the toxicity and metal concentrations, Table 2). Only rare correlations were detected at the other sites. For recent sediments from traps (Table 3), generally lower number of significant correlations between toxicity and matrix composition was recorded. Correlations in samples from Malenovice site highlighted links to concentrations of PAHs, PCBs, HCB, and some metals.

Further, principal component analysis (PCA) defined two principal factors that explained the 70 and 59.6% of the total observed variability in bottom sediments and trap samples, respectively (Figure 5). Toxicity in bottom sediments appeared to be intimately related to TOC, although close associations were also found for HCB and PAHs. Heavy metals seemed to not contribute to the toxic activity. The relationship between toxicity and TOC was also apparent in sediment traps, but this association was slightly less pronounced compared to bottom sediments. Persistent pollutants such as PCBs, DDT, and HCH seemed to also contribute to toxicity. Similar to bottom sediments, toxicity seemed not to be related to the presence of metals.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.