Sampling area

Sampling was performed in August and September 2015 at 12 sites along the river. Position of the sites is indicated in the Fig 1, while the list of analyzed parameters is summarized in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Sampling sites along the Sava River.

https://doi.org/10.1371/journal.pone.0162450.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Overview of the parameters analyzed within the Sava survey 2015.

https://doi.org/10.1371/journal.pone.0162450.t001

Indication of presence of wastewaters–faecal indicators

From each site a sample of 500 mL of water was taken in sterile glass bottles and transferred to a laboratory at 4°C. For the quantification of E. coli, Defined Substrate Technology (DST) was used, which detects E. coli by enzymatic hydrolysis of specific substrates [15]. Briefly, two dilutions of water samples (1:10 and 1:1000) were prepared and Colilert-18 powder was added, stirred and left at room temperature for a couple of minutes to dissolve. Suspensions were poured into a Colilert Quanti-Tray 2000 and sealed. Incubation was carried out at 37°C for 24 h and quantification was performed by the Most Probable Number (MPN) result, based on the color change and fluorescence in 97 wells.

Mutagenic potency of water samples—SOS/umuC

From each site a sample of 50 mL of water was taken and stored at -20°C. The SOS/umuC assay was applied on stored water samples filtrated through 0.2 μm pore size filters using the protocol described by Žegura et al. [16]. The overnight culture of S. typhimurium TA1535/pSK1002 was diluted 10 times with fresh TGA (tryptone, glucose and ampicilin medium) and incubated at 37°C for 1.5 h with aeration until the bacteria reached the exponential growth phase. Treatment was performed in microtiter plates by adding 180 μL of water sample, 20 μL of 10 x TGA and 70 μL of bacterial culture or in the case of metabolic activation 180 μL water sample, 20 μL 10 x TGA with cofactors and 70 μL of S9 bacterial culture mixture prepared as described in the ISO standards [16]. 4-Nitroquinoline (4-NQO, final concentration 0.5 μg/mL) was used as a positive control in experiments without metabolic activation while benzo(a)pyren (final concentration 10 μg/mL) was used in experiments with metabolic activation. Sterile bidistilled water was used as the negative control. The microtiter plate was incubated at 37°C for 2 h with aeration. After the treatment, the incubation mixture was then diluted 10 times with fresh TGA medium in new microtiter plates and incubated for a further 2 h. The bacterial growth rate was determined by measuring absorbance at 600 nm (OD₆₀₀) at microtitar plate reader. ß-Galactosidase activity was determined after using o-nitrophenyl-ß-D-galactopyranoside (ONPG) as a substrate for 20 min at 25°C. Absorption was measured at 405 nm using a reference solution without bacteria. The bacterial growth rate was calculated using the following formula: G = sample OD600/control OD600. A growth ratio less than 0.75, that represents 25% inhibition of biomass was considered to be an indication of cytotoxicity.

Induction ratio (IR) was calculated by the formula: sample OD405/(control OD405 x G). An induction ratio 1.5 was taken as the threshold at which the sample was considered as genotoxic [17]. All treatments were performed in triplicates in three individual experiments.

Determination of soluble concentrations of metals and metalloids in water

For the determination of the soluble element content in water samples from the Sava River, samples were filtered through a 0.45 μm filter and concentrations were determined by inductively coupled plasma mass spectrometry (ICP-MS). For the determination of the total Hg concentration, all water samples were analyzed to prevent losses of dissolved gaseous Hg during the filtrations. The results are given in S1 Table. The data revealed that the concentrations of elements in water are low.

Fish specimen collection

Whole body metal content and genotoxicological parameters were assessed in specimens of spirlin and bleak based on the on site species availability as indicated in Table 1.

All animal procedures were in compliance with Directive 2010/63/EU on the protection of animals used for experimental and other scientific purposes, and were approved by the Ethical Committee for the Use of Laboratory Animals of the Institute for Biological Research "Siniša Stanković", University of Belgrade. Fish sampling was approved by the Ministry of the Environment and Spatial Planning of the Republic of Slovenia, Ministry of Agriculture of Republic of Croatia and Ministry of Agriculture and Environmental Protection of Republic of Serbia. Specimens were collected by combining electro-fishing (Aquatech DC electro fisher IG 1300, 2.6 kW, 80–470 V) and fishing nets (mesh size 1 cm). The condition factor of the collected specimens was calculated according to the formula: CF = W/L³ × 100, where W is weight (g) and L is total length of fish (mm) [18]. Immediately after sampling, fish specimens were anesthetized with clove oil prior to dissection.

Assessment of concentration of metals and metalloids in fish tissue

Each individual separately was grinded in a Laboratory homogenizer Sterilmixer (International P.B.I. S.p.A.) and whole body composite (wbc) samples were weighed using an electronic balance (± 0.1 g) and stored at -20°C prior to analysis.

Wbc samples were submitted to the Analytical Chemistry Laboratory within the Institute of Chemistry at the Faculty of Chemistry, University of Belgrade, Serbia, for chemical analysis. The element concentrations (Al, As, Cd, Co, Cr, Cu, Fe, Hg, Mn, Ni, Pb, Se, Sn and Zn) were determined by inductively coupled plasma optical emission spectrometry (ICP-OES), using a Thermo Fisher Scientific iCAP 6500 Duo ICP (Cambridge, United Kingdom) equipped with a RACID86 Charge Injector Device (CID) detector, concentric type nebulizer, quartz torch, and alumina injector. The following wavelengths of the ICP-OES analysis were used (nm): Al 394.401, As 189.042, Cd 228.802, Co 228.616, Cr 205.552, Cu 324.754, Fe 259.941, Hg 184.950, Mn 259.373, Ni 231. 604, Pb 220.353, Se 196.090, Sn 246.161 and Zn 206.191. The limits of detection (LOD) were: Al 0.076, As 0.013, Cd 0.028, Co 0.65, Cr 0.081, Cu 0.19, Fe 0.16, Hg 0.016, Mn 0.034, Ni 0.15, Pb 0.56, Se 0.12, Sn 0.54, Zn 0.057 μg/L. The limits of quantification (LOQ) were Al 0.381, As 0.045, Cd 0.091, Co 0.95, Cr 0.029, Cu 0.67, Fe 0.54, Hg 0.063, Mn 0.115, Ni 0.59, Pb 1.78, Se 0.45, Sn 1.64, Zn 0.180 μg/L.

Fish samples (~1.5 g) were dried in a lyophilizer (Christ Alpha 2–4 LD, Harz, Germany), and then digested in an Advanced Microwave Digestion System (ETHOS 1, Milestone, Italy) using a mixture of 65% nitric acid and 30% hydrogen peroxide (Suprapur®, Merck, Darmstadt, Germany, 10:2, v/v) at 200°C for 20 min. After cooling to room temperature and without filtration, the solution was diluted to a fixed volume (volumetric flask, 25 ml) with ultra-pure water with a conductivity of 0.055 μS/cm (Barnstead™ GenPure™ Pro, Thermo Scientific, Germany), before being analyzed by ICP-OES.

Blanks with no fish tissue were run with each batch of samples to monitor contamination by the reagents used. The standards for the instrument calibration were prepared on the basis of the multi-element (SS-Low Level Elements ICV Stock, 10 mg/L) and mono-element (Hg Calibration Stock, 10 mg/L Hg; Sn Calibration Stock, mg/L Sn) certified reference solutions ICP Standard (VHG Labs, Inc-Part of LGC Standards, Manchester, NH 03103 USA) and analyzed to support quality assurance and control. The muscle standard reference material (DORM-4; National Research Council of Canada) was digested in triplicate and analyzed to support quality assurance and control.

Mean values and standard deviations were calculated for each group and elements concentrations were expressed as mgkg^-1 wet weight (ww). Kolmogorov-Smirnov test was used to determine if data were normally distributed. Differences in concentrations of elements in fishes from different sampling stations were analyzed by non-parametric Kruskal-Wallis test. Post hoc inter-group comparisons of element levels were performed by the Dunn’s Multiple Comparison Test. Statistical analysis of data was carried out using SPSS 16.0 statistical package programs for Windows (SPSS Inc., Chicago, IL, USA). The metal pollution index (MPI) was calculated to compare the total metal content in the different sampling stations using the following equation [19]: MPI = (cf1 x cf2 x cf3 x …cfn)^1/n, where cf_n = concentration of the metal n in the sample.

Blood sample collection and preservation

Blood samples were cryopreserved based on methodology described in Akcha et al. [20] with slight modifications. Blood was collected directly from the heart with 3 mL syringes (21 G needle rinsed with sodium heparin) and one drop of blood of each specimen was diluted 20x in 4°C cooled medium (RPMI 1640 supplemented with 25% fetal bovine serum). Due to storage limitations, blood samples from two specimens were pooled together to form a single sample. To all samples, cryoprotective agent was added (DMSO final concentration 20%) and samples were immediately frozen in liquid nitrogen until the analysis (up to 3 weeks).

In preliminary experiments, the effects of cryopreservation on cell viability and the level of DNA damage were assessed in 8 specimens of bleak collected at the site situated in the Sava River near Belgrade prior to Sava River survey.

Comet assay

Samples were taken from liquid nitrogen and immediately thawed at 21°C in water bath. Cell viability was assessed by acridine orange/ethidium bromide differential staining described in details in Gačić et al. [21]. Afterwards, samples were diluted in 1xPBS to obtain approximately 50,000 cells/mL. The suspensions were centrifuged (2,000 rpm, 10 min, 4°C), the supernatants were discharged and pellets were suspended in 100 μL of residual supernatant. For each sample three slides were prepared: one for alkaline and two for Fpg–modified comet assay.

The comet procedure was performed under yellow light as described in Kolarević et al. [22] with slight modifications. Briefly, microscope slides were pre-coated with 1% normal melting point (NMP) agarose and air dried for 24 h. The second, supportive layer was formed of 80 μL of 1% NMP agarose. The final layer was formed of 30 μL of cells suspension (prepared as described earlier) gently mixed with 70 μL of 1% low melting point agarose (37°C). The slides were held in freshly made cold (4°C) lysis buffer (2.5 M NaCl, 100 mM ethylenediaminetetraacetic acid disodium salt dyhidrate (EDTA), 10 mM Tris, 10% dimethyl sulfoxide (DMSO), 1.5% Triton X-100, pH 10) for 2 h. To allow DNA unwinding, slides were placed in an electrophoresis chamber containing cold (4°C) alkaline electrophoresis buffer (300 mM NaOH, 1 mM EDTA, pH 13) for 20 min. Electrophoresis was performed with a voltage gradient 0.75 V/cm and amperage 300 mA for 20 min at (4°C). Afterwards, neutralisation was carried out in freshly made cold (4°C) neutralizing buffer (0.4 M Tris, pH 7.5) for 15 min. Slides were preserved by fixation in cold methanol at 4°C for 15 min. Staining was performed with 20 μL per slide of acridine orange (2 μg/mL). The slides were examined with a fluorescence microscope (Leica, DMLS, Austria, under magnification 400 X, excitation filter 510–560 nm, barrier filter 590 nm). Microscopic images of comets were scored using Comet IV Computer Software (Perceptive Instruments, UK). Tail intensity (TI %—percentage of DNA in the tail of the comet) was chosen as a measure of DNA damage. For each sample 100 nucleoids were scored. As the possible indication of apoptosis, excessively damaged nuclei or so called hedgehogs (HH) were counted for each slide using an hedgehog tool available in the Comet IV Software.

Fpg–modified comet assay

For each sample, two slides were prepared for Fpg—modified assay as described in section 2.8, one for buffer and one for the enzyme. After one hour of lysis slides were washed 3 times in cold (4°C) washing buffer (100 mM KCl 100 mM, 10 mM Na₂EDTA and 10 mM HEPES, adjusted pH 7.2). At slides prepared for buffer, 45 μL of buffer (100 mM KCl, 10 mM EDTA, 10 mM HEPES, 0.1 mg/mL BSA, adjusted pH 7.2) was added while on slides prepared for enzyme 45 μL of 300 x diluted Fpg enzyme (Trevigen, Maryland) was added and covered with coverslips. Slides were incubated for 30 min at 37°C in humidity chamber. Afterwards slides were held for 5 min at 4°C, coverslips were removed and slides were subjected to denaturation step as for the standard comet assay protocol described in the section 2.7. The net contribution of the 8-hydroxy-2′-deoxyguanosine (8-oxoG) in final DNA damage evaluated by Fpg—modified comet assay was calculated by subtraction of the mean TI% values obtained from slides exposed to buffer only from the mean TI% values obtained from the slides exposed to Fpg enzyme [23].

Micronucleus assay

Slides for micronucleus assay were prepared as described in Štraser et al. [24], air dried for 24 h, stained with acridine orange 25 μg/mL and examined at 1000x magnification. For each sample at least 3,000 cells was examined. Nuclear aberrations were scored by criteria of Fenech [25]. Diameter for micronuclei was between 1/3 and 1/16 of the main nuclei.

Statistical analyses of data in gentoxicological bioassays

Statistical analysis of the results obtained in the experiments was carried out using Statistica 6.0 Software (StatSoft, Inc.) and SPSS 20.0 (Inc., Chicago, IL, USA). Kolmogorov-Smirnov test was used to determine if data were normally distributed. Data on MN frequency were analyzed by one-way ANOVA followed by Tukey’s post-hoc test. Comet assay data were analyzed by Kruskal-Wallis one-way ANOVA followed by Dunn’s Multiple Comparison Test since they were not normally distributed. The level of significance for all comparisons was set at p < 0.05. Correlation analyses were carried out using Pearson’s correlation test with significance level p < 0.05.

Ranking of the sites by integrated biomarker response (IBR)

The IBR ranking of the sites was performed based on parameters–E. coli numbers (EC), metal pollution index (MPI), condition factor (CF), mutagenicity (SOS with metabolic activation), comet assay (CA), oxidative stress approximated with net contribution of 8-oxoG sites (OS) and micronucleus assay (MN). IBR was assessed as described by Beliaeff and Burgeot [26]. Briefly, the value of each parameter (X_i) was standardized by the formula Y_i = (X_i -mean)/SD, where Y_i is the standardized parameter response, mean and SD are calculated based on values for the selected parameter for all sites. Z_i was then calculated as Z_i = Y_i if the studied parameter respond to contamination by induction or Z_i = -Y_i if the parameter respond to contamination by inhibition. The minimum value Z_i for each parameter was marked (min) and the scores for the studied parameters were computed as S_i = Z_i + |min|. Scores for each parameter (S_i) for particular site were used as radius coordinates of the studied parameters of the star plots. Individual areas Ai of the star plot were calculated according to the formula: Ai = S_i x S_i+1 x sin (51.43°) / 2, where S_i and S_i+1 represent the individual parameter scores and their successive star plot radius coordinates. The IBR value is calculated as following: IBR = sum of all A_i, where A_i is the area represented by two consecutive indicators on the star plot, and n is the number of indicators used in the IBR calculation. Scores for each parameter (S_i) were used for ranking of the parameters while the IBR values were used for final ranking of the sites. The site with the lowest rank was considered as the site with the lowest level of stressors.

Indication of presence of wastewaters

Numbers of E. coli were assessed in river water samples and samples collected at wastewaters discharge points (S1 Fig). Majority of river samples were slightly and moderately polluted. The highest numbers of E. coli were recorded in a river sample collected at the site Županja (32,300 MPN/100 mL). Wastewaters discharge points were found at the sites Zagreb (WWZ), S. Mitrovica (WWSM) and Šabac (WWS). Samples collected from the WW discharge points at the sites S. Mitrovica and Šabac indicated excessive faecal pollution affecting water quality at downstream situated sites (S. Mitrovica 2 and Šabac 2 respectively). Therefore, the sites Županja, S. Mitrovica 2 and Šabac 2 are recognized as hotspots of faecal pollution.

SOS/umuC

The results are summarized at the Fig 2. In investigated water samples cytotoxic effect were not recorded and threshold value 1.5 of induction was not breached in any case. The highest induction was detected at the sites Litija and Vrhovo. Positive controls for experiment (4NQO and benzo(a)pyrene) affirmed validity of experimental system.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. SOS induction rate (mean ± SE) in SOS/umuC assay.

Red line represents threshold induction value (1.5); PC—positive controls: 4NQO (0.5 μg/mL) in experiments without metabolic activation and benzo(a)pyrene (10 μg/mL) in experiments with metabolic activation.

https://doi.org/10.1371/journal.pone.0162450.g002

Assessment of concentration of metals and metalloids in fish tissue

During the chemical analysis of metals and metalloids, a replicate analysis of the reference material showed good accuracy, with recovery ranging from 89.80% to 103.75%. The average concentrations in wbc samples of spirlin and bleak and MPI values are presented in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. The average element concentrations in whole body composite (wbc) samples of A.alburnus and A. bipunctatus (marked with *) in mgkg^-1 ww, and MPI values for each sampling station; different letters denote significant differences among studied sites (p < 0.05).

https://doi.org/10.1371/journal.pone.0162450.t002

A pattern was observed at all sampling sites with the highest Zn concentrations (23.09 ± 6.65 to 47.67 ± 9.73) and at almost all of the sampling sites with the lowest Co concentrations (0.0001 ± 0.0004 to 0.009 ± 0.016) (Table 2). The highest concentrations and numbers of metals (Al, Co, Fe, Mn, Ni and Sn) were recorded at Županja sampling site. On the other hand, the lowest concentrations of metals and metalloids (Al, Cu, Fe, Hg, Se, Zn) were in the highest number determined at the Beograd sampling station.

These results affected MPI values, and therefore the highest MPI was recorded for the Županja sampling station (0.33) and the lowest for the Belgrade sampling station (0.14). However, when comparing the values for concentrations of metals and metalloids at these two sites statistically significant difference was observed only for Cd and Ni.

When comparing the data of faecal pollution and data on concentration of metals and metalloids in fish tissue, significant correlation (r = 0.78, p = 0.008) was observed between the numbers of E. coli and MPI.

Effects of cryopreservation and data validation

As indicated in Fig 3, the effects of cryopreservation on cell viability and DNA damage induction were assessed. Cell viability was reduced for 18% while cryopreservation did not additionally increased DNA damage level.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Cryopreservation effects on the cell viability and DNA damage level.

*statistical significance in comparison with fresh sample (p < 0.05).

https://doi.org/10.1371/journal.pone.0162450.g003

Although the values were within the same range, the level of DNA damage after cryopreservation was significantly lower comparing to the fresh sample. However, significant increase of HH frequency was noticed.

Assessment of cell viability in cryopreserved samples.

Average cell viability in cryopreserved samples was about 80% (Table 3). The lowest viability was observed in a sample collected at the site Vrhovo (65 ± 6%) indicating possible cytotoxic effect. Significant negative correlation was observed between the cell viability and HH frequency (r = -0.61, p = 0.0361).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Cell viability, frequency of hedgehogs and frequency of micronuclei in fish blood samples (mean ± SE).

https://doi.org/10.1371/journal.pone.0162450.t003

Assessment of the DNA damage in blood cells

Alkaline comet assay was performed for the assessment of DNA damage while Fpg -modified comet assay was performed for the assessment of oxidative stress. Data obtained by alkaline comet assay indicated variation of DNA damage within the studied sites (Fig 4). The sites with the highest levels of DNA damage were Litija, Vrhovo, Jasenovac and Šabac 2. Impact of wastewater discharges was evident at the site Šabac 2 which had significantly higher TI% values in comparison with upstream situated Šabac 1.

Relation between the data obtained in the standard comet assay, HH% and cell viability was investigated. There was neither significant correlation between the level of DNA damage and HH% (r = 0.20; p = 0.53) nor DNA damage and cell viability (r = -0.36; p = 0.25).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. The values of tail intensity % obtained in standard alkaline comet assay in fish blood cells; values are represented as mean ± SE; different letters denote significant differences among studied sites (p < 0.05).

https://doi.org/10.1371/journal.pone.0162450.g004

Fpg–modified comet assay

Data obtained in Fpg—modified assay are summarized in the Table 4. TI% values from slides exposed to buffer only were within the same range as values in the standard alkaline comet assay, moreover significant positive correlation between the values was observed (Fig 5A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Correlation between the levels of DNA damage obtained in different assays.

Correlation of the values obtained in standard comet assay and (A) buffer exposed slides, (B) Fpg exposed slides and (C) net 8-oxoG sites; full line—regression line; dashed line—95 confidence level.

https://doi.org/10.1371/journal.pone.0162450.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Fpg—modified comet assay, tail intensity values for buffer and Fpg—exposed slides and net contribution of 8-oxoG sites (mean ± SE); different letters denote significant differences among studied sites (p < 0.05).

https://doi.org/10.1371/journal.pone.0162450.t004

In all cases digestion with Fpg enzyme led to significant increase of DNA damage in comparison with corresponding standard alkaline and buffer treated control. Significant correlation was observed between the mean TI% values obtained from slides for standard alkaline comet assay and slides exposed to Fpg enzyme (Fig 5B). The highest net contribution of 8-oxoG sites was detected in specimens from Čatež, S. Mitrovica 1 and Belgrade (Table 4). When plotting the values of TI% obtained in the standard alkaline comet assay and the values on net contribution of 8-oxoG sites in DNA damage, significant correlation was not observed (Fig 5C).

Micronucleus

The highest frequency of MN was detected in specimens collected at the sites Vrhovo and Zagreb (Table 3). However, in comparison with the reference site there was no significant difference in MN frequency at any of investigated sites.

Correlation of genotoxicological parameters and concentrations of metals in fish

Significant positive correlation was observed only between the values of TI% in alkaline comet assay and concentration of Hg in tissue (r = 0.75, p = 0.013). Significant negative correlation was observed between the frequency of 8-oxoG sites and concentration of Cu (r = -0.66, p = 0.038) (Table 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Correlation between the monitored genotoxicological parameters and concentration of metals and metalloids in fish tissue; marked correlations are significant (p < 0.05).

https://doi.org/10.1371/journal.pone.0162450.t005

Ranking of the sites by IBR

The final IBR values as well as ranking for each studied marker are graphically presented at Fig 6. The site Šabac 1 had the lowest IBR rank (1.03) while the Županja and Vrhovo had the highest ranks (8.15 and 7.97 respectively) (Table 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Graphical presentation of IBR.

https://doi.org/10.1371/journal.pone.0162450.g006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Integrated Biomarker Response (IBR) ranking of the studied sites.

https://doi.org/10.1371/journal.pone.0162450.t006

General

In this study we have performed comprehensive genotoxicological survey by applying the battery of assays in prokaryotes and aquatic eukaryotes. When constructing the battery of bioassays we were focused on types of assays which have already been employed earlier in ecogenotoxicological studies of the Sava River [8–14, 27]. Also, we have used experience gained during the Joint Danube Survey 3 –JDS3 [22, 28, 29, 30] for updating the assay list as usage of the same methodology would enable comparison of water quality of the Sava and Danube in further research. As the assays employed in the study showed different sensitivity in detection of the stressors at investigated sites we have finally employed IBR approach to identify the most critical spots on the river but also to recommend which site has potential to be a reference site for Sava River. As the reference site is defined as the site which is minimally exposed to the stressors of interest [31], the Šabac 1 was chosen as the best available option.

Indication of presence of wastewaters

The whole Sava River Basin is receiving high amounts of untreated or improperly treated wastewaters originating from the various size settlements that lie on the banks of the Sava River and its tributaries [1]. Therefore, our primary goal was identification of hotspots of contamination with wastewaters as possible sources of genotoxic pollution. As an indication of wastewater effluents, we have decided to use the numbers of E. coli in water. The results of the JDS3 confirmed that this parameter and concentration of caffeine in the Danube River are the most reliable indicators of contamination related to wastewaters [28, 32]. Indeed, wastewater outlets were recorded at three sampling sites (Zagreb, S. Mitrovica and Šabac), but based on concentrations of E. coli, throughout the full river course, only two hotspots of faecal contamination (critical faecal pollution) were identified (Županja and Šabac 2).

Mutagenicity of water samples

The SOS/umuC was included in our research as the study of Žegura et al. [16] and Kittinger et al. [6] indicated the assays high potency in detection of genotoxic potential in wastewaters and surface waters. We have assumed that the assay would provide preliminary screening of genotoxic potential, but the induction ratio has not exceeded value 1.5 (threshold value) in any of the investigated samples. Similarly, in the study of Kittinger et al. [30] performed on the water samples of the Danube River, only four sites out of 68 investigated within the JDS3 showed mutagenic potential, indicating either to low sensitivity of the assay or low genotoxic potential of water.

Indication of possible industrial pollution

In the moment of sampling, concentrations of metals and metalloids in water at the sampling sites were generally low. Therefore we focused on the concentrations of the elements accumulated in the tissues of selected species as an indication of prolonged exposure. Recent studies regarding metals and metalloids contamination in the Sava River were obtained in the European chub (Squalius cephalus) [33–35]. Based on the dataset obtained in our earlier survey of the Sava River [36], we have selected two minnow species, which are ecological equivalents in different communities. Spirlin is characteristic for the upper section and bleak is present in middle and lower sections of the river. Both spirling and bleak are active and fast swimmers, of the short life-span, with consequently high metabolic rates which can lead to high accumulation of metals and metalloids. Also, they are both epipelagic fish, not exposed to the contaminants in the benthic zone. Based on these features, bleak was successfully used for estimation of metals and metalloids in studies of Uysal et al. [37], Al Sayegh Petkovšek et al. [38], Mercai et al. [39] and many others.

We estimated the highest industrial pollution with the highest observed MPI for the Županja sampling site and of lowest for the Belgrade sampling site. This result is in accordance with Dragun et al. [40] who previously reported that Sava River water reflects a certain anthropogenic impact in the Croatian section of the river while, on the other hand, Vuković et al. [41] stated that the industrial activity in Serbia slowed down during the past two decades, which is the reason for the weakly noticeable anthropogenic input of heavy metals in the Sava River system (water and sediment) from the nearby environment.

Although we have assessed the presence of wastewaters by two aspects (indicators of faecal pollution and industrial pollution) it is important to emphasize that significant correlation was observed between the numbers of E. coli and MPI clearly showing that at the majority of the sites pollution from a single source prevails.

When making the parallel of the data on metal accumulation in tissue and the data from genotoxicological assays, we would emphasize that the positive correlation was observed between the concentration of mercury and TI% values. Mercury is listed as a priority substance by Directive 2008/105/EC [42] whose genotoxic potential in aquatic environment is well known [43]. Negative correlation was detected between the oxidative stress and concentration of copper which is understandable as copper can have a protective effect as a constitutive of superoxide dismutase [44]. Correlations which were not significant were not taken into consideration because of the possible causative effect.

Assessment of genotoxic potential along the river