Assessment of sublethal endpoints after chronic exposure of the nematode Caenorhabditis elegans to palladium, platinum and rhodium

Environmental Pollution 230 (2017) 31e39

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Assessment of sublethal endpoints after chronic exposure of the nematode Caenorhabditis elegans to palladium, platinum and rhodium* Gerhard Schertzinger*, Sonja Zimmermann, Daniel Grabner, Bernd Sures €tsstraße 5, 45144 Essen, Department of Aquatic Ecology and Centre for Water and Environmental Research, University of Duisburg-Essen, Universita Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 December 2016 Received in revised form 12 June 2017 Accepted 13 June 2017

The aim of this study was to investigate chronic effects of the platinum-group elements (PGE) palladium (Pd), platinum (Pt) and rhodium (Rh) on the nematode Caenorhabditis elegans. Aquatic toxicity testing was carried out according to ISO 10872 by determining 96 h EC50 values for sublethal endpoints, including growth, fertility and reproduction. Single PGE standard solutions were used as metal source. Based on the EC50 values for Pt, reproduction (96 h EC50 ¼ 497 mg/L) was the most sensitive endpoint followed by fertility (96 h EC50 ¼ 726 mg/L) and growth (96 h EC50 ¼ 808 mg/L). For Pd, no precise EC50 values could be calculated due to bell-shaped concentration response curves, but the 96 h EC50 for reproduction ranged between 10 and 100 mg/L. Pd and Pt had effects on all endpoints. With raising element concentrations reproduction was inhibited first. At a certain concentration, fertility was also affected, which in turn had an additional effect on reproduction. Growth inhibition can also lead to a loss of fertility if the worms do not reach an appropriate body size to become fertile. Rhodium showed no inhibition of any endpoint between concentrations of 100 to 10,000 mg Rh/L. The results of this study allow the following order of PGE with respect to decreasing toxicity to C. elegans: Pd > Pt » Rh. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Growth Fertility Reproduction Metal toxicity ISO 10872

1. Introduction The platinum group elements (PGE) palladium (Pd), platinum (Pt) and rhodium (Rh) are used in dentistry, jewelry, manufacturing and electrical as well as petroleum industries (Zereini and Wiseman, 2015), but their major consumer is the automobile industry, where these metals are used in catalytic converters to reduce the release of unburnt hydrocarbons, carbon monoxide and nitrogen oxides in engine fumes (Taylor, 1987). The positive effects of automobile catalytic converters, namely the reduction of pollutant concentrations in exhaust gases as well as reduced lead (Pb) emissions, are accompanied by the release of particles containing PGE. This was already noticed shortly after the introduction €rtsch and Schlatter, 1988; of automobile catalytic converters (Ba € nig et al., 1992). As a consequence, the Hertel et al., 1990; Ko

* This paper has been recommended for acceptance by Dr. Harmon Sarah Michele. * Corresponding author. E-mail address: [email protected] (G. Schertzinger).

http://dx.doi.org/10.1016/j.envpol.2017.06.040 0269-7491/© 2017 Elsevier Ltd. All rights reserved.

environmental concentrations of the three PGE increased within the last three decades (summarized in Zereini and Wiseman, 2015). Following deposition on the road, the metals accumulate in soils and aquatic sediments (summarized by Ruchter et al., 2015) and subsequently become available for flora and fauna (Sch€ afer et al., 1998; Moldovan et al., 2001; Zimmermann and Sures, 2004; Zimmermann et al., 2015). Accordingly, it was suggested that environmental monitoring of PGE should be established in order to prevent an increase of PGE concentrations to critical levels (Sures et al., 2015). Therefore, appropriate tools are required to investigate the biological availability and toxicity of PGE. However, only relatively few studies assessed the uptake and toxicity of PGE so far (Osterauer et al., 2009; summarized by Sures et al., 2015). No data on the effect of PGE on terrestrial invertebrates is available to date and toxicity tests with aquatic organisms are scarce. In addition, little is known about chronic effects of Pt, Pd and Rh on organisms as well as their effects on the population level (summarized by Sures et al., 2015). The available data suggests that a couple of adverse effects occur if aquatic organisms are exposed to PGE. However, usually relatively high exposure concentrations were applied so far exceeding current environmental concentrations of

32

G. Schertzinger et al. / Environmental Pollution 230 (2017) 31e39

PGE and most studies focus on relatively short exposure periods in terms of the life cycle of organisms, which is contrary to longlasting natural exposure conditions. Up to now, the range of organisms tested for possible adverse effects of PGE comprises five plant species and different taxa of animals including parasites (acanthocephalans), gastropods and bivalves (molluscs), annelids (worms), crustaceans and two taxa of vertebrates, i.e. fish and amphibians (summarized in Sures et al., 2015). Bioaccumulation studies with these organisms usually show highest uptake rates for Pd followed by Pt, whereas Rh accumulation is always least (Zimmermann and Sures, 2004; Zimmermann et al., 2015). Subchronic effect studies using aquatic animals were only rarely performed (Sures et al., 2015), but indicate effects during ontogenesis and early embryonic development (e.g. Biesinger and Christensen, €hler, 2009; Osterauer 1972; Monetti et al., 2003; Sawasdee and Ko et al., 2009, 2010, 2011). As traffic-related PGE mainly accumulate €fer and in soils and sediments especially along roads (e.g. Scha Puchelt, 1998; Ruchter and Sures, 2015; summarized in Zereini and Wiseman, 2015), it would be advantageous to use nematodes as test organisms since they are very abundant in nearly every type of soil or sediment. Nematodes, such as Caenorhabditis elegans, have been widely used for assessing metal toxicity due to the advantages of this species over others such as its easy maintenance, short testing period, low expenses, transparent body and the possibility to perform life cycle tests in a rather short period of time (Williams and Dusenbery, 1988, 1990; Traunspurger et al., 1997; Boyd and Williams, 2003; Roh et al., 2006; Rudel et al., 2013). Moreover, a comparison of the LC50 values of C. elegans obtained in aquatic systems with those of other benthic organisms (Amphipoda, Zygoptera, Diptera and Gastropoda) showed that C. elegans is the most sensitive species in exposure tests using different metal salts (Williams and Dusenbery, 1990). These facts have promoted the development of several standardized assays for environmental toxicity testing using C. elegans as a test organism (Williams and Dusenbery, 1990; Traunspurger et al., 1997; Leung et al., 2008; ASTM, 2012; International Organization for Standardization, € ss et al. (2012), the standardized toxicity 2010). According to Ho test with C. elegans, based on the international standard ISO 10872 (International Organization for Standardization, 2010), is an appropriate tool for assessing the toxicity of aquatic and terrestrial systems using growth, fertility and reproduction as endpoints. Due to an exposure period of 96 h at 20  C, which covers the whole life span of C. elegans (development over four larval stages (L1 e L4) to adult worms and subsequent reproduction), chronic effects on the individual worms (growth and fertility) as well as effects on the population level (reproduction) can be determined using this test

system. Accordingly, the aim of the present study was a first assessment of the toxicity of the platinum-group elements Pd, Pt and Rh for C. elegans using single PGE standard solutions. The toxicity was evaluated in terms of growth, fertility (presence of at least one egg within a hermaphrodite) and reproduction (number of offspring per hermaphrodite). These endpoints have been shown to be adequate and sensitive endpoints in the assessment of metal €ss et al., 2012) and toxicity on nematodes (Anderson et al., 2001; Ho add information on possible adverse effects of PGE during the different stages within the life cycle of organisms. 2. Materials and methods 2.1. Maintenance and age synchronization of nematodes The toxicity tests were performed according to ISO 10872 (International Organization for Standardization, 2010) with some modifications. The wild type strain N2 of C. elegans var. Bristol was cultivated from a Dauer larvae stock on nematode growth medium agar plates (NGM-agar plates) containing a lawn of Escherichia coli (OP50, uracil deficient strain) as food source. Both organisms were obtained from the Caenorhabditis Genetics Center (CGC) at the University of Minnesota. Deviating from ISO 10872, a method according to Stiernagle (2006) called “Egg-Prep” was applied for age synchronization of the nematodes to obtain only first-stage larvae (L1). To this end, C. elegans eggs were axenized from gravid hermaphrodites by lysis through a bleaching solution containing sodium hypochlorite and sodium hydroxide. Remaining eggs were transferred to a petri dish with a thin layer of M9-medium and allowed to hatch overnight. 2.2. Test validity and sensitivity Previous to metal toxicity tests, the sensitivity and validity of the test system was checked (Table 1). Two tests wereperformed using exposure concentrations of the reference material benzyldimethylhexadecylammonium chloride (BAC-C16, C25H46ClNxH2O, p.a.) as given in the ISO guideline. According to the latter, the 96 h EC50 value (growth) for BAC-C16 should be in the range of 8e22 mg/L. The sensitivity tests were also necessary to derive the concentration of BAC-C16 for the positive control, which should be close to the 96 h EC50 value (growth) for BAC-C16. Additionally, all validity criteria of ISO 10872 were investigated in negative controls tested in parallel. A test is considered to be valid, if the average recovery of the exposed test organisms in the negative controls is  80% and

Table 1 Test conditions, validity criteria and 96 h EC50 (growth) for BAC-C16. Parameter Test conditions

Unit

Bacterial density FAU  Temperature C Initial body length þ SD mm Validity criteria of the negative control according to ISO 10872 1. RecoveryC % 2. MaleC % 3. MaleR (max) % 4. FertilityC % (% CVi) 5. ReproductionC O/H (% CVi) Validity criteria 1.-5. e GrowthC mm (% CVi) Reference material 96 h EC50 (growth) mg/L BAC-C16

Test 1

Test 2

Specifications according to ISO 10872

755 19.1e25.0 221 ± 7 96.2 0 0 100 (0) 39 (10.3) þ/þ/þ/þ/þ 1037 (2.3) 9.1

773 19.2e22.2 228 ± 9 89.6 4,2 12,5 100 (0) 54 (7.4) þ/þ/þ/þ/þ 960 (4.0) 10.8

1000 ± 50 20.0 ± 0.5 e 80%e120% 10% 20% 80% 30

8.0e22.0

FAU ¼ formazin attenuation unit, RecoveryC/MaleC/FertilityC/ReproductionC/Growthc ¼ mean value of the respective parameter from three replicates of the control, MaleR (max) ¼ control replicate with the highest percentage of males, CVi ¼ test internal coefficient of variance, O/H ¼ offspring per hermaphrodite, þ ¼ validity criteria fulfilled, SD ¼ standard deviation.

G. Schertzinger et al. / Environmental Pollution 230 (2017) 31e39

33

Table 2 Nominal and quantified Pd concentrations measured in two aliquots of the respective exposure medium and their percentage of the nominal concentration for both tests. Pd

Quantified Concentration (Test 1)

Quantified Concentration (Test 2)

Nominal concentration (mg/L)

Mean (mg/L)

Mean (mg/L)

10 100 200 400 600 800 1000 1200

< LoD < LoQ 160 471 568 806 1085 1331

% of nominal concentration

< LoD < LoQ 161 440 635 933 1062 1368

e e 80 118 95 101 109 111

% of nominal concentration e e 80 110 106 117 106 114

LoD ¼ Limit of Detection, LoQ ¼ Limit of Quantification.

120%. The average percentage of males in the negative controls should be  10% and the occurrence of males in a single replicate of the negative controls has to be  20%. Average fertility and average reproduction in the negative controls should be  80% and 30 progenies per exposed test organism. 2.3. PGE toxicity tests Range finding tests were performed in order to determine the concentration range at which Pd, Pt and Rh causes a response between 0% and 100% (inhibition of the respective endpoint related to the control) in C. elegans. The nominal concentrations used for Pd and Pt are shown in Tables 2 and 3. For Rh, no effects on the endpoints growth, fertility and reproduction were observed within the tested concentration range between 100 and 10,000 mg/L. As environmental concentrations of Rh are usually much lower (summarized in Zereini and Wiseman, 2015), no further toxicity tests with higher concentrations of Rh were carried out. In total, two test series with three replicates of each nominal Pd/ Pt exposure concentration were conducted. Parallel to each precious metal test series, three replicates of negative and positive controls were prepared and used for the verification of validity parameters according to ISO 10872 (Table 4). Prior to the test, 2 ml of the nematode suspension containing 10 ± 3 L1 were added to each well of 24 well plates. For each control group, a 4 ml mix of the exposure medium was prepared, containing 2 ml of E. coli OP50 feed medium (775 ± 35 FAU instead of 1000 ± 50 FAU as described in ISO 10872), 1.96 ml deionized water (MiliPore) and 40 ml of the respective metal standard solution. The bacterial density of the feed medium is expressed as formazin attenuation unit (FAU) according to ISO 7027 (International Organization for Standardization, 2016) and was measured with the Infinite plate reader M200 by Tecan, at a wavelength of 860 nm.

The metal standard solutions, stabilized in 1% HCl, were prepared by using 1 g/L Pd or Pt stock solutions, p.a. and 37% HCl, p.a. Finally, 1 ml of the mix was added to each well of the three replicates, and the remaining solution was used for metal quantification. The concentration of the positive controls was 10 mg BAC-C16 per liter, which is close to the 96 h EC50 value (growth) for BAC-C16 (see Table 1). According to the ISO guideline, inhibition of growth in the positive control should be between 20% and 80% compared to the control. To check for possible effects of the acid solvent, HCl controls containing 500 ml feed medium, 490 ml deionized water and 10 ml 1% hydrogen chloride solution per well were tested in parallel to each Pd/Pt test series. The multi-well plates were sealed with Parafilm®, covered and incubated under constant horizontal shaking at 20  C for 96 h in the dark. After exposure, the tests were terminated by adding 0.5 ml of a 0.3 g/L Rose Bengal solution to each well in order to stain the worm's cuticle to improve the visibility for counting. Subsequently, the multi-well plates were covered and heated for 10 min at 80  C. This procedure induced the stretching of the nematodes to allow determination of their length. The multi-well plates were stored until evaluation for a maximum of 3 weeks at 4  C. The endpoints growth, fertility and reproduction were evaluated according to ISO 10872. Nematode body size was determined from the head to the tip of the tail using a microscope with the camera Moticam 2300 (Motic) and the software Motic Images Plus 2.0 (Motic) (Fig. S1, supplement). To calculate growth, the mean initial body length of 30 randomly selected first-stage larvae (L1) was subtracted from the mean body size of all hermaphrodites from each replicate. A nematode was considered as gravid if at least one egg was visible within its body. For the endpoint reproduction, the offspring in each replicate was counted. To calculate EC50 values with a 95% confidence interval, all replicate data were transferred into inhibition values according to ISO 10872 (inhibition of the

Table 3 Nominal and quantified Pt concentrations measured in two aliquots of the respective exposure medium and their percentage of the nominal concentration for both tests. Pt

Quantified concentration (Test 1)

Nominal concentration (mg/L)

Mean (mg/L)

100 250 500 600 700 800 900 1000 1200

< LoD < LoQ 481 570 658 746 839 996 1182

LoD ¼ Limit of Detection, LoQ ¼ Limit of Quantification.

Quantified concentration (Test 2) % of nominal concentration e e 96 95 94 93 93 100 99

Mean (mg/L) < LoD < LoQ 462 565 659 761 868 996 1202

% of nominal concentration e e 92 94 93 95 97 100 100

34

G. Schertzinger et al. / Environmental Pollution 230 (2017) 31e39

Table 4 Test conditions, validity criteria and inhibition values of positive and HCl controls for each test series.

Test conditions

Validity criteria of the negative control according to ISO 10872

Positive control HCl control

Parameter

Unit

Pd/Pt Test 1

Pd/Pt Test 2

Bacterial density Temperature Initial body length ±SD 1. RecoveryC 2. MaleC 3. MaleR (max) 4. FertilityC 5. ReproductionC Validity criteria 1.e5. GrowthC Inhibition of growth Inhibition of growth, fertility and reproduction

FAU  C mm % % % % (% CVi) O/H (% CVi) e mm (% CVi) % %

791 19.5e21.3 239 ± 9 100.0 0 0 97.2 (3.2) 31 (3.2) þ/þ/þ/þ/þ 1089 (4.4) 38.4 <10

809 19.8e21.4 253 ± 11 93.0 0 0 100 (0) 34 (8.8) þ/þ/þ/þ/þ 1041 (1.2) 45.3 <10

FAU¼ Formazin Attenuation Unit, RecoveryC/MaleC/FertilityC/ReproductionC/GrowthC ¼ mean value of the respective parameter from three replicates of the control, MaleR (max) ¼ control replicate with the highest percentage of males, CVi ¼ test internal coefficient of variance, O/H ¼ offspring per hermaphrodite, þ ¼ validity criteria fulfilled, SD ¼ standard deviation.

respective endpoint related to the control). The inhibition values of the replicate data were fitted as nonlinear regression with variable € ss et al. slope using the software GraphPad Prism 6. According to Ho (2012), significant differences from the control could be expected at inhibition values > 10%. Therefore, in this study all inhibition values of the single replicates lower than 10% were set to zero prior to calculation by GraphPad Prism 6. Since stimulation was not considered, negative inhibition values were also set to zero. In this context, inhibition of all endpoints in the HCl controls should be below 10% in relation to the negative control, to ensure that no effects occurred due to the acid solvent used in the metal standards. 2.4. Nematode density effect As the number of individuals differed between 7 and 13 larvae per well, a possible effect of nematode density on the test results was analyzed. To this end, the nematode growth was determined in dependence on the number of individuals. In total, three experimental groups with 12 replicates each and 7e13 larvae/well were investigated, a negative control group and two positive control groups with exposure concentrations of the reference material BAC-C16 of 7.5 mg/L and 15 mg/L, respectively. Afterwards, the nematodes’ mean body lengths of all 12 replicates of each experimental group were analyzed in a first step for significant differences due to the number of individuals (One-way ANOVA). Additionally, differences in the mean body lengths due to the number of individuals within one experimental group were investigated (Tukey's Multiple Comparison Test). 2.5. Metal quantification in the exposure media Metal concentrations in the exposure medium were quantified using microwave digestion combined with graphite furnace atomic absorption spectrometry (GF-AAS). The digestion of the exposure medium samples (two 400 ml aliquots) was carried out with a microwave digestion system (MARS 6, CEM GmbH) as described by Zimmermann et al. (2001). For metal quantification, PGE were analyzed in the digestion solution in triplicate by GF-AAS (AAnalyst™ 600, PerkinElmer) equipped with Zeeman effect background correction according to Zimmermann et al. (2003). Limit of detection (LoD), limit of quantification (LoQ), precision and analytical recovery were also determined according to Zimmermann et al. (2003). 2.6. Data analysis The software GraphPad Prism 6 was used to create the graphs and to perform the statistical analysis.

3. Results and discussion 3.1. Test validity and sensitivity following adaptation of the test procedure Despite slight methodological changes, the modified test system fulfils all validity criteria with regard to the negative controls given by ISO 10872 (Table 1). The sensitivity criteria for BAC-C16 were met as the 96 h EC50 values for the endpoint growth (9.1 mg/L in test 1 and 10.8 mg/L in test 2) were in the range of 8e22 mg/L for both tests (Table 1 and Fig. S2, supplement). As the changes to the methodology did not influence the validity criteria and the sensitivity of the test system, all performed tests are valid. The applied Egg-Prep-method was highly reliable and repeatable in terms of the initial body length of the L1, as the standard deviation between the freshly hatched and starved progenies were only up to 11 mm (Tables 1 and 4). According to ISO 10872, age synchronization is carried out by filtration of a nematode suspension through a filter cascade (10 and 5 mm). However, this method did not work using polycarbonate filters as no L1 passed the 5 mm filter. According to Hanna et al. (2016), age synchronization by filtration is possible using filter gauze from Hepfinger J. (Munich, Germany), but L2 may also pass the filters and fungal contamination could occur and affect the test results, which is nearly excluded by the bleaching method. The study of Hanna et al. (2016) also showed that the size of the test well (24 vs. 12 well plates) had no significant effect on the outcome of the test. The validity criteria for reproduction with 30 progenies per hermaphrodite were reached in all tests and the average growth of all control worms in this study was 1032 ± 53 mm. For comparison, reproduction in an inter-laboratory comparison reached 130 ± 25 progenies per adult worm and the €ss et al., average growth of control worms was 1242 ± 124 mm (Ho 2012). The lower reproductive output as well as the decreased body length in the present study is in accordance with findings published by Hanna et al. (2016), who showed that shaking the plates during exposure is a crucial point in the assay. The reproductive output was decreased by approximately 70% and the body length of the control worms was decreased by more than 300 mm when plates were shaken. Another explanation could be that the nematodes were not sufficiently fed, as bacterial densities in the present study were 775 ± 35 FAU and therefore ranged below the density recommended in the ISO guideline (1000 ± 50 FAU). Hanna et al. (2016) revealed that growth inhibition in the positive control (15 mg BAC-C16/L) decreases with raising feed densities. It is still unclear whether this may occur due to different amounts of feed or due to a change in bioavailability of the test substance with varying feed densities. Accordingly, growth was affected by approximately

G. Schertzinger et al. / Environmental Pollution 230 (2017) 31e39

60% and 65% at feed densities of 400 and 350 FAU in the test wells and concentrations of the reference material BAC-C16 of 15 mg/L. In the present study, under nearly the same conditions, growth was already affected by 80e86%. Thus, the lower feed density alone is not sufficient to explain the low 96 h EC50 value (growth) for BACC16 of around 10 mg/L determined in the present study. However, we suggest that age synchronization by bleaching could also increase the sensitivity of the test system. Using the filtration method, the L1 already feed on the bacterial lawn on the agar plate before exposure and might have a better developed cuticle and physiological mechanisms against toxic stressors in contrast to the test organisms obtained by the “Egg-Prep” method in the present study. The latter were exposed directly after hatching and therefore might be more sensitive to toxic stressors, which could explain why the 96 h EC50 values (growth) for BAC-C16 were at the lower border of the acceptable range of 8e22 mg/L (Table 1). As the “Egg-Prep” method provided equally developed organisms at the beginning of the experiment (body length, age after hatching, no feed before the start of exposure), it would be worth to analyze if this synchronization method may improve the reproducibility of the test results as well as the sensitivity of the test system.

35

3.5. Palladium toxicity The concentration response curves for the endpoints growth, fertility and reproduction were not sigmoidal over the investigated concentration range between 10 and 1200 mg Pd/L (Fig. 1 aec). Accordingly, the calculation of EC50 values for the different endpoints was not possible though clear adverse effects were visible in nearly each test.

3.2. Nematode density effect No significant differences (p < 0.05, One-way ANOVA) between the mean body lengths of all replicates of the respective experimental group were detected. In a second step, the single replicates of each experimental group were tested pairwise for significant differences. Again, no differences (p < 0.001, Tukey's Multiple Comparison Test) were observed. Thus, the fluctuating number of 7e13 individuals per test well had no significant effect on the test results. 3.3. Metal concentration in the exposure media The limit of detection (LoD) based on the concentration in the exposure medium for Pd and Pt was 39 mg/L and 118 mg/L, respectively. The limit of quantification (LoQ) based on the concentration in the exposure medium for Pd and Pt was 92 mg/L and 276 mg/L, respectively. The precision expressed as relative standard deviation was 3.8% for Pd and 4.1% for Pt. An analytical recovery of 99.8% for Pd and 97.6% for Pt was determined demonstrating that nearly no loss of the analytes occurred during the microwave digestion procedure. Thus, the analytical procedure to determine the metal concentrations in the exposure medium was highly accurate. Usually, at least 80% of the nominal Pd concentrations and 92% of the nominal Pt concentrations were detected in the exposure media except for the lowest applied PGE concentrations. The sensitivities were too low to determine the concentrations of 10 and 100 mg Pd/L as well as 100 and 250 mg Pt/L in the exposure media. Considering the low differences between the nominal and the quantified concentrations (Tables 2 and 3), calculations for concentration-response curves and effect concentrations were based on the nominal concentrations. 3.4. Validity of PGE toxicity tests All validity criteria with regard to the negative controls given by ISO 10872 were met for both Pd/Pt test series (Table 4). Additionally, in both test series the validity criteria for the positive control as well as for the HCl control were fulfilled as the inhibition of growth in the positive control (10 mg BAC-C16/L) was between 20 and 80% and the inhibition of all endpoints for the HCl control was below 10%. Accordingly, the results of the toxicity tests were valid.

Fig. 1. aec: Concentration-response plots for the inhibition of the respective endpoint by Pd in two independent nematode tests. Points represent means and error bars standard deviations of three replicates.

36

G. Schertzinger et al. / Environmental Pollution 230 (2017) 31e39

For the inhibition of growth, only in test 2 a slight inhibition was detected, which occurred at exposure concentrations between 400 and 800 mg Pd/L. Interestingly, at higher concentrations (1000 and 1200 mg Pd/L), no growth inhibition was recognized. In test 1, no growth inhibition occurred at all. Although an overall stronger inhibition of fertility was observed in test 2 compared to test 1, the general response was similar in both tests, with an initial increase of inhibition and a sharp decline at the highest concentrations. First effects on fertility occurred at exposure concentration of 100 mg Pd/L (test 1) and 200 mg Pd/L (test 2). As a result, 17% of the nematodes in test 1 and 48% in test 2 were not able to develop eggs. The differences between the tests for the endpoint fertility can be explained by the fact that inhibition of fertility is not always a direct effect of the toxicant. It may rather be a result of growth inhibition, as hermaphrodites start laying eggs at temperatures between 20 and 25  C as soon as they reach a size of 1060 mm and 1110 mm, respectively (Byerly et al., 1976). Thus, if growth is inhibited and hermaphrodites are not able to reach the required body size to become fertile, inhibition of fertility is a consequence due to growth inhibition. Therefore, the stronger inhibition of fertility in test 2 might be an additional effect due to growth inhibition observed in the same test. But there still seems to be a direct effect of Pd on fertility as inhibition also occurred in test 2 and at concentrations that did not affect growth in test 1 (200 mg Pd/L). This is underlined by the observation that the gravid nematodes exposed to the intermediate concentrations (100e1000 mg Pd/L) only had 1 to 3 eggs within their body, whereas the gravid nematodes exposed to the lowest and highest concentrations of Pd contained more than 5 eggs. Both were rated equally as gravid without further differentiation as a hermaphrodite is according to ISO 10872 counted as gravid if at least one egg occurs. Therefore, it can be suggested that the interpretation of the results could be improved by using egg numbers as an additional, potentially more sensitive measure for fertility. Inhibition of reproduction showed a similar pattern as the inhibition of growth and fertility. A clear and comparable inhibition of reproduction was observed for both tests with a concentrationdependent increase and a decrease at higher concentrations. At the lowest exposure concentration of 10 mg Pd/L, reproduction was already inhibited by 40% (test 1) and 17% (one replicate in test 2). An inhibition of nearly 100% was determined over a concentration range from 200 to 600 mg Pd/L for test 1 and over a concentration range from 200 to 800 mg Pd/L for test 2. This inhibition of reproduction at the intermediate concentrations cannot be explained as a consequence of the inhibition of fertility, because fertility was not completely inhibited. Thus, not all of the nematodes in the test wells were infertile. Furthermore, in test 1 only a slight inhibition of fertility was determined, whereas reproduction was completely inhibited. Additionally, reproduction was inhibited in the test wells with the lowest (10 mg Pd/L) and the highest (1200 mg Pd/L) Pd concentrations, where no effect on fertility was observed. These findings underline the assumption that a complete inhibition of reproduction of C. elegans must be a combination of a direct effect due to exposure with Pd and an additional effect due to the inhibition of fertility. The fact that in both test series (as well as in the range finding tests) inhibition of all endpoints was decreased at the highest exposure concentration compared to intermediate Pd concentrations allows the conclusion that either a loss of bioavailable Pd species or an induction of protective mechanisms against metal stress in C. elegans took place at higher exposure concentrations. The first conclusion could be explained by precipitation of Pdphosphate due to the exposure in M9 medium. For the latter, it is possible that an expression of metallothioneins (low molecular weight, cysteine-rich metal binding proteins) in C. elegans was

induced more efficiently at the highest concentration of Pd present in the exposure medium. As a consequence, Pd was bound by the proteins and its toxic effects on C. elegans were reduced at the higher exposure concentrations. Induction of metallothioneins was described to occur in C. elegans by Freedman et al. (1993). In other organisms such as the mussel Dreissena polymorpha, metallothionein induction was also described to occur after exposure to Pd (Singer et al., 2005; Frank et al., 2008). However, there is also evidence for a possible detoxification mechanism through E. coli in the test medium (Deplanche et al., 2010). The authors demonstrate a biological reduction of Pd ions to Pd0 due to E. coli. This biological reduction resulted in Pd nanoparticles on the cell surfaces of the bacteria, which can be seen by transmission electron microscopy (TEM) as black dots. A nearly complete reduction of Pd2þ to Pd0 within 180 min was observed for a solution containing approximately 200 mg/L Pd2þ, which is much higher than the concentrations used in the present study. If this effect is the explanation for the observed phenomenon, we must assume that the reduction only takes place when the Pd concentration exceeds a certain threshold. The fact that E. coli did not reduce all of the present Pd2þ might support this idea, but a minimum Pd concentration, which triggers the biological reduction is unknown (Deplanche et al., 2010). However, a study of De Windt et al. (2006) showed that the biological reduction of Pd2þ to Pd0 nanoparticles is dependent on the amount of Pd2þ per bacterium. A higher Pd concentration per cell resulted in more and bigger nanoparticles on the cell surface. To prove if a biological reduction of Pd2þ by E. coli took place during exposure, samples of the exposure media could be investigated by TEM. However, due to the low concentrations of Pd in this study, the detection of nanoparticles by TEM is unlikely. In summary, for Pd no EC50 values could be calculated for the investigated endpoints. The lowest concentrations at which an effect on growth, fertility and reproduction could be observed were 400, 100 and 10 mg Pd/L, respectively. Furthermore, it can be estimated that the 96 h EC50 value for reproduction of C. elegans due to exposure to Pd ranges between 10 and 100 mg Pd/L. Although other endpoints, exposure periods and test organisms were used, the estimated sensitivity for Pd in the applied test system lies in the same range as determined in other toxicity tests. For example, the 96 h EC50 value (immobilization) for the annelid Tubifex tubifex was found to be 92 mg Pd/L (Khangarot, 1991). Moreover, the Daphnia sp. acute immobilization test (OECD Guideline 202) showed similar values for 24 h EC50, 48 h EC50 and 48 h LC50 values, which were around 16 mg Pd/L (±1 mg/L; quantified concentration) (Zimmermann et al., 2017). The algae Pseudokirchneriella subcapitata showed molecular effects, which resulted in morphological changes, modifications of the photosystem as well as growth inhibition starting at the lowest exposure concentrations of 30 mg Pd/ L (Vannini et al., 2011). For the crustacean Hyalella azteca, Borgmann et al. (2005) observed a LC50 value of 570 mg Pd/L after 7 days of exposure (nominal concentration in soft water). Nevertheless, in terms of Pd toxicity, reproduction of C. elegans seems to be one of the most sensitive sublethal endpoints tested so far, as first significant effects can be detected at concentrations as low as 10 mg Pd/L. All in all, the inhibition of all investigated endpoints exemplify that the nematodes are affected on the level of individuals (growth and fertility) as well as on the population level (reproduction), when chronically exposed to Pd. 3.6. Platinum toxicity For the exposure of C. elegans to Pt over a concentration range from 100 to 1200 mg Pt/L, a concentration-dependent increase of all inhibition endpoints was observed. Therefore, sigmoidal-shaped concentration response curves could be fitted (Fig. 2 aec) and

G. Schertzinger et al. / Environmental Pollution 230 (2017) 31e39

37

First significant signs of inhibition (>10% inhibition) of fertility and growth were observed at exposure concentrations of 600 mg Pt/ L (test 1) and 800 mg Pt/L (test 2), indicating that some of the exposed worms did not reach the required size to become fertile (Byerly et al., 1976). Inhibition of reproduction could already be detected in both tests at 500 mg Pt/L. At this concentration level, no inhibition of fertility was observed. Therefore, Pt seems to impair egg survival and larval development, even at comparably low concentrations. At higher exposure concentrations (600 mg Pt/L in test 1 and 800 mg Pt/L test 2), reproduction is additionally affected due to the inhibition of fertility. Comparing toxicity values from different assays is always challenging due to the use of different test organisms, endpoints and exposure periods. However, the toxicity of Pt towards fish and snail embryos is up to 10,000 times higher compared to the toxicity of C. elegans towards Pt, as first significant effects on nematode reproduction occurred at 381 mg Pt/L (mean 96 h EC10, reproduction). As an example, Osterauer et al. (2009) observed that the heart rate in embryos of the fish Danio rerio and the snail Marisa cornuarietis was affected at the lowest tested aqueous concentrations of 0.202 mg Pt/L and 0.038 mg Pt/L, respectively (nominal concentration for both organisms was 0.1 mg PtCl2/L). Additionally, the hatching rate was affected in both organisms. The hatching success of Danio rerio was significantly reduced at exposure concentrations of 44.2 mg Pt/L (50 mg PtCl2/L), 84 h after fertilization. Hatching of Marisa cornuarietis was almost completely inhibited at 74.2 mg Pt/L (100 mg PtCl2/L) during the whole test period of 14 days. At day 12, snail hatching was significantly affected even at exposure concentrations of 3.6 mg Pt/L (10 mg PtCl2/L). Toxicity tests using frog embryos of Xenopus laevis revealed a 120 h LC50 value of 589 mg Pt/L (3.02 mM Pt). Growth of the frog embryos was already inhibited at a concentration of 78 mg Pt/L (0.4 mM Pt) (Monetti et al., 2003). For the annelid Tubifex tubifex, the 96 h EC50 value (immobilization) was found to be 61 mg Pt/L (Khangarot, 1991). In the Daphnia sp. acute immobilization test (OECD Guideline 202), a 48 h EC50 value for the endpoint immobilization of 117 mg Pt/L and a 48 h LC50 value of 168 mg Pt/L were found (Zimmermann et al., 2017). Borgmann et al. (2005) determined that 50% of individuals of the crustacean Hyalella azteca died after one week of exposure to 221 mg Pt/L (nominal concentration in soft water). Thus, Pt seems to be less toxic towards C. elegans compared to other test organisms. Nevertheless, it can be stated that Pt exposure can result in chronic effects on C. elegans individuals as growth and fertility were inhibited. Additionally, Pt exposure can affect populations of C. elegans as reproduction was inhibited at even lower concentrations. 3.7. Rhodium toxicity

Fig. 2. aec: Concentration-response curves for the inhibition of the respective endpoint by Pt in two independent nematode tests. Points represent means and error bars standard deviations of three replicates. Lines are fitted by a sigmoidal model.

mean 96 h EC50 values of both tests were calculated for the investigated endpoints, showing that reproduction (96 h EC50 ¼ 497 mg/L) is the most sensitive endpoint followed by fertility (96 h EC50 ¼ 725 mg/L) and growth (96 h EC50 ¼ 808 mg/L). 96 h EC50 values of the single tests as well as confidence intervals and the coefficient of correlation (R2) between the measured values and the sigmoidal fit are given in the supplementary material (Table S1).

For Rh, no effect on the endpoints growth, fertility and reproduction was observed within the concentration range from 100 to 10,000 mg Rh/L. As environmental concentrations of Rh were not known to exceed the highest range finding concentration of 10,000 mg Rh/L, no further toxicity tests with higher Rh concentrations were carried out. These results were in accordance with former toxicity studies (see Sures et al., 2015). All available data showed that Rh acts only in very high concentrations as a toxicant. In the Daphnia sp. acute immobilization test (OECD Guideline 202), a 48 h EC50 value for the endpoint immobilization of 15 mg Rh/L and a 48 h LC50 value of 57 mg Rh/L were found (Zimmermann et al., 2017). Borgmann et al. (2005) determined a LC50 value for Hyalella azteca after one week of exposure to 980 mg Rh/L (nominal concentration in soft water). Due to the long exposure period, this value is relatively high compared to the 48 h EC50 value for Daphnia magna. However, LC50 values are known to decrease with longer exposure periods (Williams and Dusenbery, 1990).

38

G. Schertzinger et al. / Environmental Pollution 230 (2017) 31e39

The lower toxicity of Rh can be explained by the fact that Rh is not known to induce metallothioneins (sulfhydryl group-rich metal binding proteins, produced for metal detoxification) in contrast to Pt and Pd (Frank et al., 2008). Accordingly, the affinity of Rh to sulfhydryl groups of functional enzymes is assumed to be very low. This could also explain that it may not bind strongly to sulfhydryl groups of enzymes and thus do not alter their function. 4. Conclusion Our results show that C. elegans is a suitable test organism to assess the toxicity of PGE in liquid media. First effects on the reproduction of C. elegans could be expected at Pd concentrations below 10 mg/L. For Pt, 96 h EC50 values of 497 mg Pt/L for the endpoint reproduction, 725 mg Pt/L for fertility and 808 mg Pt/L for growth were determined. Following this study, first significant effects of Pt on C. elegans were identified at 381 mg Pt/L (mean 96 h EC10, reproduction). This concentration by far exceeds environmentally relevant Pt concentrations in pore water of natural sediments, which are in the upper ng per liter level and therefore do not pose a risk to free-living individuals of C. elegans. However, considering that reproduction of C. elegans most probably starts to be affected at Pd concentrations below 10 mg Pd/L and the highest pore water concentrations along roadsides reach up to 2 mg Pd/L, it is obvious that environmental monitoring of PGE is important for the future, taking the constant increase of PGE emissions in the environment over the last three decades into account. The sensitivity of the applied test system for Pd and Pt was similar (Pd) or lower (Pt) compared to other aquatic toxicity tests. Nevertheless, the advantage of the C. elegans test is the suitability to test also other matrices like natural or spiked artificial sediments and soils, enabling more realistic scenarios, including mixture toxicity and the presence of organic matter. Furthermore, we found an unusual response pattern of Pd with decreasing effects of all endpoints at the highest exposure concentrations, observed in two independent tests. This effect cannot be explained to date, but underlines the need for further studies with PGE (and other substances) using the C. elegans test system. Moreover, age synchronization by Egg-Prep is suitable to enhance the reproducibility and sensitivity of the test results, due to homogenization of the test individuals (age after hatching, body length) and prevention of feeding before exposure, compared to the filtration method described by ISO 10872. Acknowledgments The authors want to thank Mark Schumann and Thomas Knura for their assistance in the lab, regarding metal detection. Special thanks to Dr. Michelle Pahl for language and grammar help. Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2017.06.040. References Anderson, G.L., Boyd, W.A., Williams, P.L., 2001. Assessment of sublethal endpoints for toxicity testing with the nematode Caenorhabditis elegans. Environ. Toxicol. Chem. 20, 833e838. http://dx.doi.org/10.1002/etc.5620200419. ASTM, 2012. Standard Guide for Conducting Laboratory Soil Toxicity Tests with the Nematode Caenorhabditis elegans E2172-01(2008). ASTM Annual Book of Standards, vol 1106. American Society for Testing and Materials, Westconshohocken, PA, USA.

€rtsch, A., Schlatter, C., 1988. Platinemissionen aus automobil-katalysatoren. Ba Schriftenr. Umweltschutz (Bern) 95, 1e57. Biesinger, K.E., Christensen, G.M., 1972. Effects of various metals on survival, growth, reproduction, and metabolism of Daphnia magna. J. Fish. Res. Board Can. 29, 1691e1700. Borgmann, U., Couillard, Y., Doyle, P., Dixon, D.G., 2005. Toxicity of sixty-three metals and metalloids to Hyalella azteca at two levels of water hardness. Environ. Toxicol. Chem. 24, 641e652. http://dx.doi.org/10.1897/04-177R.1. Boyd, W.A., Williams, P.L., 2003. Comparison of the sensitivity of three nematode species to copper and their utility in aquatic and soil toxicity tests. Environ. Toxicol. Chem. 22, 2768e2774. http://dx.doi.org/10.1897/02-573. Byerly, L., Cassada, R.C., Russell, R.L., 1976. The life cycle of the nematode Caenorhabditis elegans, I. wild-type growth and reproduction. Dev. Biol. 51, 23e33. Deplanche, K., Caldelari, I., Mikheenko, I.P., Sargent, F., Macaskie, L.E., 2010. Involvement of hydrogenases in the formation of highly catalytic Pd(0) nanoparticles by bioreduction of Pd(II) using Escherichia coli mutant strains. Microbiology 156, 2630e2640. De Windt, W., Boon, N., Van den Bulcke, J., Rubberecht, L., Prata, F., et al., 2006. Biological control of the size and reactivity of catalytic Pd(0) produced by Shewanella oneidensis. Antonie Leeuwenhoek 90, 377e389. http://dx.doi.org/ 10.1007/s10482-006-9088-4. Frank, S., Singer, C., Sures, B., 2008. Metallothionein (MT) response after chronic palladium exposure in the zebra mussel, Dreissena polymorpha. Environ. Res. 108, 309e314. Freedman, J.H., Slice, L.W., Dixon, D., Fire, A., Rubin, C.S., 1993. The novel metallothionein genes of Caenorhabditis elegans. Structural organization and inducible, cell-specific expression. J. Biol. Chem. 268, 2554e2564. Hanna, S.K., Cooksey, G.A., Dong, S., Nelson, B.C., Mao, L., Elliott, J.T., Petersen, E.J., 2016. Feasibility of using a standardized Caenorhabditis elegans toxicity test to assess nanomaterial toxicity. Environ. Sci. Nano 3, 1080e1089. €nig, H.P., Inacker, O., Malessa, R., 1990. Nachweis der Freisetzung und Hertel, R.F., Ko Identifizierung von Edelmetallen im Abgasstrom von Katalysatorfahrzeugen. Edelmetallemissionen: Gesellschaft für Strahlen- und Umweltforschung mbH (GSF), pp. 16e21. €ss, S., Ahlf, W., Bergtold, M., Bluebaum-Gronau, E., Brinke, M., et al., 2012. Ho Interlaboratory comparison of a standardized toxicity test using the nematode Caenorhabditis elegans (ISO 10872). Environ. Toxicol. Chem. 31, 1525e1535. International Organization for Standardization, 2010. Water Qualityddetermination of the Toxic Effect of Sediment and Soil Samples on Growth, Fertility and Reproduction of Caenorhabditis elegans (Nematoda), vol 10872. ISO, Geneva, Switzerland, p. 2010. International Organization for Standardization, 2016. Water Qualityddetermination of TurbiditydPart 1: Quantitative Methods, vols 7027e1. ISO, Brussels, Belgium, p. 2016. Khangarot, B.S., 1991. Toxicity of metals to a freshwater tubificid worm Tubifex tubifex (Müller). Environ. Contam. Toxicol. 46, 906e912. €nig, H.P., Hertel, R.F., Koch, W., Rosner, G., 1992. Determination of platinum Ko emissions from a three-way catalyst-equipped gasoline engine. Atmos. Environ. 26A, 741e745. Leung, M.C., Williams, P.L., Benedetto, A., Au, C., Helmcke, K.J., et al., 2008. Caenorhabditis elegans: an emerging model in biomedical and environmental toxicology. Toxicol. Sci. 106, 5e28.  mez, M.M., Palacios, M.A., Morrison, G.M., 2001. BioMoldovan, M., Rauch, S., Go accumulation of palladium, platinum and rhodium from urban particulates and sediments by the freshwater isopod Asellus aquaticus. Water Res. 35, 4175e4183. Monetti, C., Bernardini, G., Vigetti, D., Prati, M., Fortaner, S., Sabbioni, E., Gornati, R., 2003. Platinum toxicity and gene expression in Xenopus embryos: analysis by FETAX and differential display. Altern. Lab. Anim. 31, 401e408. €hler, H.R., 2009. Uptake of platinum by zebrafish Osterauer, R., Haus, N., Sures, B., Ko (Danio rerio) and ramshorn snail (Marisa cornuarietis) and resulting effects on early embryogenesis. Chemosphere 77, 975e982. Osterauer, R., Marschner, L., Betz, O., Gerberding, M., Sawasdee, B., Cloetens, P., €hler, H.R., 2010. Turning snails into slugs: induced Sures, B., Triebskorn, R., Ko body plan changes and formation of an internal shell. Evol. Dev. 12, 474e483. €hler, H.R., 2011. Genotoxicity of platOsterauer, R., Faßbender, C., Braunbeck, T., Ko inum in embryos of zebrafish (Danio rerio) and ramshorn snail (Marisa cornuarietis). Sci. Total Environ. 409, 2114e2119. Roh, J.-Y., Lee, J., Choi, J., 2006. Assessment of stress-related gene expression in the heavy metal-exposed nematode Caenorhabditis elegans: a potential biomarker for metal-induced toxicity monitoring and environmental risk assessment. Environ. Toxicol. Chem. 25, 2946e2956. http://dx.doi.org/10.1897/05-676R.1. Ruchter, N., Sures, B., 2015. Distribution of Pt and other traffic related metals in sediments and clams (Corbicula sp.). Water Res. 70, 313e324. Ruchter, N., Zimmermann, S., Sures, B., 2015. Field studies on PGE in aquatic ecosystems. Environ Sci. In: Zereini, F., Wiseman, C.L.S. (Eds.), 2015. Platinum Metals in the Environment. Springer-Verlag, Berlin Heidelberg. ISBN 978-3662-44558-7, 351-360. Rudel, D., Douglas, C.D., Huffnagle, I.M., Besser, J.M., Ingersoll, C.G., 2013. Assaying environmental nickel toxicity using model nematodes. PLoS One 8/10, 1e17. http://dx.doi.org/10.1371/journal.pone.0077079. € hler, H.R., 2009. Embryo toxicity of pesticides and heavy metals to Sawasdee, B., Ko the Ramshorn snail, Marisa cornuarietis (Prosobranchia). Chemosphere 75, 1539e1547. Sch€ afer, J., Puchelt, H., 1998. Platinum-group metal (PGM) emitted from automobile

G. Schertzinger et al. / Environmental Pollution 230 (2017) 31e39 catalytic converters and their distribution in roadside soils. Geochem Explor 64, 307e314. €fer, J., Hannker, D., Eckhardt, J.D., Stüben, D., 1998. Uptake of traffic-related Scha heavy metals and platinum group elements (PGE) by plants. Sci. Total Environ. 215, 59e67. Singer, C., Zimmermann, S., Sures, B., 2005. Induction of heat shock proteins (hsp70) in the Zebra mussel (Dreissena polymorpha) following exposure to platinum group metals (platinum, palladium and rhodium): comparison with lead and cadmium exposures. Aquat. Toxicol. 75, 65e75. Stiernagle, T., 2006. WormBook. In: Maintenance of C. elegans, Chapter 6: Cleaning Contaminated C. elegans Stocks, Protocol 4: Egg Prep: Removing Bacterial or Yeast Contaminants from C. elegans Stock Plates. WormBook. The C. elegans Research Community. http://dx.doi.org/10.1895/wormbook.1.101.1. http://www. wormbook.org (18.11.2016). Sures, B., Ruchter, N., Zimmermann, S., 2015. Biological effects of pge on aquatic organisms. Environ Sci. In: Zereini, F., Wiseman, C.L.S. (Eds.), 2015. Platinum Metals in the Environment. Springer-Verlag, Berlin Heidelberg. ISBN 978-3662-44558-7, 383-399. Taylor, K.C., 1987. Automobile catalytic converters. In: Studies in Surface Science and Catalysis, vol 30, pp. 97e116. http://dx.doi.org/10.1016/S0167-2991(09)60416-X. €ss, S., Beier, S., Ahlf, W., et al., 1997. Ecotoxicological Traunspurger, W., Haitzer, M., Ho assessment of aquatic sediments with Caenorhabditis elegans (Nematoda) e a method for testing liquid medium and whole sediment samples. Environ. Toxicol. Chem. 16/2, 245e250. Vannini, C., Domingo, G., Marsoni, M., Fumagalli, A., Terzaghi, R., Labra, M., De Mattia, F., Onelli, E., Bracale, M., 2011. Physiological and molecular effects associated with palladium treatment in Pseudokirchneriella subcapitata. Aquat.

39

Toxicol. 102, 104e113. Williams, P.L., Dusenbery, D.B., 1988. Using the nematode Caenorhabditis elegans to predict mammalian acute lethality to metallic salts. Toxicol. Ind. Health 4, 469e478. Williams, P.L., Dusenbery, D.B., 1990. Aquatic toxicity testing using the nematode, Caenorhabditis elegans. Environ. Toxicol. Chem. 9, 1285e1290. Zereini, F., Wiseman, C.L.S., 2015. Platinum Metals in the Environment. Environ Sci. Springer-Verlag, Berlin Heidelberg. ISBN 978-3-662-44558-7. Zimmermann, S., Menzel, C., Berner, Z., Eckhardt, J.D., Stüben, D., et al., 2001. Trace analysis of platinum in biological samples: a comparison between sector field ICP-MS and adsorptive cathodic stripping voltammetry following different digestion procedures. Anal. Chim. Acta 439/2, 203e209. Zimmermann, S., Messerschmidt, J., von Bohlen, A., Sures, B., 2003. Determination of the platinum group metals Pt, Pd and Rh in biological samples by electrothermal atomic absorption spectrometry as compared with adsorptive cathodic stripping voltammetry and total reflection X-ray fluorescence analysis. Anal. Chim. Acta 498, 93e104. Zimmermann, S., Sures, B., 2004. Significance of platinum group metals emitted from automobile exhaust gas converters for the biosphere. Environ Sci Pollut Res 11, 194e199. http://dx.doi.org/10.1007/BF02979675. Zimmermann, S., Sures, B., Ruchter, N., 2015. Laboratory studies on the uptake and bioaccumulation of PGE by aquatic plants and animals. Environ Sci. In: Zereini, F., Wiseman, C.L.S. (Eds.), 2015. Platinum Metals in the Environment. Springer-Verlag, Berlin Heidelberg. ISBN 978-3-662-44558-7, 361-381. Zimmermann, S., Wolff, C., Sures, B., 2017. Toxicity of platinum, palladium and rhodium to Daphnia magna in single and binary metal exposure experiments. Environ. Pollut. 222, 368e376.