journa l homepage: www.e lsev ier .com/ locate /aquatox
Direct and indirect effects of pollutants on algae and algivorousciliates in an aquatic indoor microcosm
Markus Liebiga,∗, Gunnar Schmidta,b, Daniel Bontjec, Bob W. Kooic,Georg Streckd, Walter Traunspurgerb, Thomas Knackera
a ECT Oekotoxikologie GmbH, Boettgerstrasse 2-14, D-65439 Floersheim/Main, Germanyb feld, G
De BoeAnaly
sm ws (Cryes andor comoncehe m
pareEC foo reddia aiderer risk
Bielefeld University, Department of Animals Ecology, Morgenbreede 45, D-33615 Bielec Vrije Universiteit, Faculty of Earth & Life Sciences, Department of Theoretical Biology,d UFZ - Helmholtz-Zentrum fur Umweltforschung GmbH, Department of Effect-Directed
a r t i c l e i n f o
Article history:Received 30 November 2007Received in revised form 7 March 2008Accepted 14 March 2008
An aquatic indoor microcoon phototrophic flagellatecaused effects to flagellatwere exposed separatelyflagellates at low �g L−1 ca NOEC of 15.2 �g L−1 in tthe multi-species test comprometryn. The lower NOby an indirect effect due tcentrations in the test meindoor microcosm is consprovide data for higher tie
1. Introduction
Single-species toxicity tests are advantageous when regulatoryneeds require repeatability and comparability of effects data. How-ever, these test systems do not reflect interactive components ofnatural systems such as predation or nutrient recycling. The use ofindoor multi-species or defined microcosm test systems representsa link between single-species tests and indoor microcosm or meso-cosm field studies containing natural assemblages of organisms.Indoor multi-species tests can be useful in higher tier risk assess-ment as they can assist in the design of field studies or help to clarifycause and effect relationships of specific responses observed inmesocosms tests (Campbell et al., 1999). The integration of differenttrophic levels in multi-species test systems permits the assessmentof potential indirect effects, including cascading effects (Lawler,1993; Kooi, 2003).
In the present study, single-species and multi-species toxi-city tests were performed with aquatic organisms representingthree trophic levels: producers (autotrophic flagellates), consumers
as used to study effects of the pesticides parathion-methyl and prometrynptomonas sp.) and predatory ciliates (Urotricha furcata). Parathion-methylciliates at the range of low mg L−1, regardless of whether the organismsbined in the multi-species test system. Prometryn caused effects on the
ntrations, resulting in a NOEC of 6.9 �g L−1 in the single-species test andulti-species microcosm. For ciliates the NOEC decreased by factor 145 ind to the NOEC of 2.2 mg L−1 in the single-species test when exposed to
r ciliates exposed to prometryn in the microcosm was most likely causeduced availability of flagellates as food. The measurement of nutrient con-nd organisms facilitated the modelling of effects. The presented aquaticd as a tool which could be standardised and applied as an instrument toassessment.
(algivorous ciliates) and decomposer (unspecified bacterial com-munity). The combination of these organisms in the same aquatic
medium under defined conditions is defined as indoor multi-species microcosm test system which represents a canonicalcommunity (can-com). Canonical in this context means ‘the sim-plest representative that still has all essential properties of themicrocosm system’, e.g. nutrient assimilation, growth, degradationand nutrient recycling. In toxicity studies, using a can-com providesthe opportunity to study direct and indirect effects of toxicants, i.e.to observe the propagation of effects between trophic levels. Hence,the research presented in this paper reaches beyond the establish-ment of direct and indirect effects of contaminants on two or morespecies in laboratory scale microcosms (see review of Fleeger etal., 2003). Instead the objectives are (a) to generate effects data onspecies interacting within three trophic levels in a simple micro-cosm and (b) to design the studies in a way that the data can be usedfor modelling the effects of biological and toxicological stressorsbased on assumptions of the Dynamic Energy Budget (DEB) theory(Kooijman, 2000; Koelmans et al., 2001). Therefore, not only growthparameters like cell number and abundance but also additionalparameters like nutrient contents in cells and medium were deter-mined. Among others, the generated data will be used for computersimulations of effects of toxicants on population growth in canon-
ical communities, whereby the observed effects are translated topredicted effects in the environment.
2. Material and methods
2.1. Test organisms
In the can-com, the trophic level of producers is repre-sented by the mixotrophic phytoflagellate Cryptomonas sp. (strainSAG 26.80) which was purchased as a non-axenic stock cul-ture from the Experimental Phycology and Culture Collection ofAlgae, Gottingen, Germany. It is cultivated in frequently dilutedbatch cultures under permanent illumination at approximately33 ± 3 �mol photons m−2 s−1 at 20 ± 1.5 ◦C. In preliminary experi-ments with Cryptomonas sp. using undiluted WC medium (Guillardand Lorenzen, 1972) an average cellular content of 5.8 ± 0.5(n = 13) pmol C and 0.75 ± 0.19 (n = 14) pmol N was measured in upto 23 days old flagellate cultures.
Mixotrophy is a wide-spread phenomenon in aquatic systemswhereby numerous plankton species combine a heterotrophic andautotrophic feeding mode. Cryptophyte ingestion rates are gener-ally low and extreme environmental conditions may be necessaryto induce phagotrophy (Sanders and Porter, 1988). At the describedculturing and test conditions, Cryptomonas sp. is considered to pre-vail exclusively in the autotrophic mode.
The consumers are represented by the widely distributedplanktonic prostomatid ciliate Urotricha furcata (SCHEWIAKOFF).Prostomatid ciliates play a major role in the microbial foodweb of many freshwater lakes and reservoirs (e.g. Muller, 1989;Schonberger, 1994). U. furcata feeds on bacteria and autotrophicflagellates (Foissner et al., 1999). However, it does not feed uponbacteria when cryptophytes are present abundantly (Weisse andFrahm, 2001). The stock cultures of U. furcata were kindly providedby Th. Weisse (Austrian Academy of Sciences, Mondsee, Austria)and originally derived from the mesotrophic Lake Mondsee, Austria.
Decomposers were represented by an unspecified bacterialcommunity in the can-com which facilitate nutrient recycling ofassimilation and maintenance products or dead organisms. Thedecomposers originated from the non-axenic stock cultures of flag-ellates and ciliates. Single-species experiments were not performedwith bacteria.
2.2. Test medium
Modified WC medium according to Guillard and Lorenzen (1972)was used for all stock cultures and test systems. Since for mod-elling purposes experimental data on both dynamic and staticphases of growth are preferred, the nitrate content of the mediumwas five-fold reduced compared to the composition described byGuillard and Lorenzen (1972) with a final nitrate concentrationof 0.2 mM.
2.3. Test compounds
Two pesticides with different modes of action were selectedas model compounds for the toxicity tests: prometryn (CAS-No. 7287-19-6), a selective systemic methylthiotriazine herbicide,and parathion-methyl (CAS-No. 298-00-0), a non-systemic phenylorganothiophosphate insecticide, nematicide and acaricide. Bothcompounds produced by Riedel-de Haen were purchased fromSigma–Aldrich Chemie GmbH, Taufkirchen, Germany, with puritiesof 99.2% and 99.8%, respectively. The authorisation for parathion-methyl for the use as plant protection product was withdrawn in2003 (EC, 2003). However, both pesticides are included in govern-mental monitoring programmes of surface waters of the catchment
logy 88 (2008) 102–110 103
of the River Elbe. In these surface waters prometryn and parathion-methyl were measured at maximum concentrations of 49 and18 ng L−1, respectively (ARGE Elbe, 2006, 2007).
2.4. Test conditions
All tests were performed as static test systems in 300 mL Erlen-meyer flasks filled with 100–150 mL cell suspension and closedwith air-permeable lids. At each work day the test vessels wereslightly shaken in order to keep the cell suspensions homoge-nous and to reduce the development of a biofilm on the vessels’wall. The test medium and abiotic conditions were the same asdescribed above for the cell cultivation. An exception was thereduced light intensity to 5.7 ± 0.6 �mol photons m−2 s−1 duringthe ciliate single-species tests. The applied test concentrationranges were decided based on the outcomes of previously con-ducted range-finding tests.
The flagellate single-species test procedure followed the revisedOECD guideline 201 (OECD, 2006). These tests were performed over14 days with prometryn (Csp-P) and parathion-methyl (Csp-PM) atinitial cell numbers of 3 × 104 cells mL−1. Controls and five test con-centrations in a geometric series were applied with four replicatesfor each treatment (nominal concentrations: Csp-P: 7–35.4 �g L−1;Csp-PM: 0.21–2.5 mg L−1).
Ciliate short-term toxicity tests named Uf-P (with prometryn)and Uf-PM (with parathion-methyl) were performed over 48 hwith initial cell numbers of 450 and 350 ciliates mL−1, respec-tively. The ciliate tests were evaluated after 24 h, since after 48 hcells were affected by starvation. Geometric series of five test con-centrations were applied at nominally 0.94–15 mg L−1 for Uf-Pand 0.5–8 mg L−1 for Uf-PM. In the ciliate cultures, the flagellatesserving as prey cannot be separated without harming the cells.Therefore, the ciliate single-species tests were started with a mini-mum flagellate cell density whereas the reduced light intensity didnot allow growth of flagellates.
The multi-species tests using flagellates and ciliates togetheras test organisms (MS-P with prometryn, MS-PM with parathion-methyl) were performed over 10 days (MS-P) and 13 days(MS-PM) with initial cell numbers of 105 flagellates mL−1 and100 ciliates mL−1. The selected exposure times allowed completegrowth cycles of the ciliate populations consisting of a short ini-tial lag phase, a log phase showing exponential growth and a deathphase of the populations. No stationary phases were observed dur-ing growth of the ciliates. The applied nominal test concentrations
were 8.75, 17.5 and 35 �g L−1 (MS-P) and 0.4, 1.26 and 4 mg L−1 (MS-PM) with four replicates in each concentration and six replicates forcontrols.
Sampling frequencies were each 24 h for the multi-species andsingle-species ciliate tests and 48–72 h for the flagellate single-species tests. The assessed biological parameters were cell number,area below the growth curve and average specific growth rate deter-mined according to OECD (1984, 2006) in the multi-species andflagellate single-species tests. In the ciliate short-term tests sur-vival was assessed by determining cell numbers only. Results arereported as mean ± standard deviation (S.D.).
2.5. Measurement of cell densities
For measuring cell numbers, samples were taken periodicallyfrom the test vessels using a sterile pipette. Before sampling thecell suspensions were homogenised thoroughly by gentle agita-tion. During the single-species tests the cell number of Cryptomonassp. was determined photometrically by measuring the extinctionof the cell suspension at 680 nm. During the multi-species teststhe flagellate cell number was determined using a Thoma count-
ing chamber at 150-fold magnification after Lugol’s fixation ofthe samples (1%, v/v). For the determination of the ciliate cellnumber a Sedgewick-Rafter settling chamber of 1 mL sample vol-ume was used. The bacterial cell density was determined at thebeginning and at the end of the tests in formalin fixed sam-ples (2%, v/v) using DAPI staining as described by Porter and Feig(1980).
2.6. Evaluation of nutrient content in cells and test medium
The nutritional status of the organisms and of the surroundingmedium is important for modelling based on the DEB theory. Inorder to evaluate the nutrient content of the test organisms dur-ing different growth phases, as well as of the test medium, thetotal organic carbon (TOC) content and the total nitrogen contentwere determined. Therefore, samples of cells and filtrated medium(0.2 �m) were taken at the beginning and at the end of the tests,and at an intermediate sampling point when highest cell num-bers of Cryptomonas sp. during the flagellate single-species testsor of U. furcata during the multi-species tests were expected. Dur-ing the test period of 48 h of the ciliate short-term toxicity testsa significant change of the nutritional status was not expected;therefore, analyses of nutrient content were not performed in thesetests. TOC content was determined photometrically by the differ-ence method using a cuvette test-kit (LCK 380, Hach Lange GmbH,Germany). Total nitrogen content was also determined photomet-rically using the test-kit LCK 138 (Hach Lange GmbH, Germany).Nutrient contents (total nitrogen and TOC) of the medium weremeasured in sterile filtrates (0.2 �m) of the samples. For the mea-surement of the nutrient content in the cells the samples werefirst centrifuged triply for 15 min at 2533 × g in order to con-centrate the cells and purge them from the test medium. Aftereach centrifugation step approximately 80% of the supernatantwas discarded and the cells were resuspended in demineralisedwater. Following this purging procedure, cells were solubilised byultrasonication.
2.7. Supporting chemical analysis
Supporting chemical analysis is needed to determine the actualtest item concentrations in comparison to the nominally appliedconcentrations. Therefore, samples from the controls and from
the lowest and highest test concentration levels were taken atthe beginning and at the end of the single and multi-speciestests and additionally at an intermediate day during the multi-species tests. The analysis of the pesticides was performed usinga solid phase extraction (SPE) method with C18-cartridges (0.5 g)purchased from Waters (Eschborn, Germany). All samples weredrawn through PTFE-filters (0.2 �m). A known amount of atrazine(Promochem, Wesel, Germany) dissolved in ethyl acetate wasadded to each sample to assess the accuracy of the analyti-cal procedure. Samples of prometryn and parathion-methyl testsolutions were adjusted to pH 7.0 and 6.0, respectively. Ana-lytes were eluted with 10 mL ethyl acetate, reduced with N2 toa volume of 200 �L and finally analysed on a HP6890 capil-lary column gas chromatograph equipped with a HP5973 massselective detector (Agilent Technologies). A HP5-MS capillarycolumn (length 30 m, inner diameter 0.32 mm, film thickness0.25 �m, Agilent Technologies) and helium as carrier gas wasused. All injections were done in the splitless mode with avolume of 1 �L. The analyses were conducted in SIM mode.Concentrations of the target analytes were calculated using anexternal calibration and corrected by means of the injection stan-dards.
logy 88 (2008) 102–110
2.8. Growth of cell populations and statistical evaluation
Growth of the test organisms was determined as (a) absolutecell number (N), (b) average specific growth rates (�) and (c) thearea under the growth curves (A) using the formulas given in theOECD guideline 201 (OECD, 1984, 2006). The mean value of cellnumbers for each treatment was plotted against time to producegrowth curves. � was calculated for each replicate and samplingperiod, i.e. between day 0 and each sampling date.
One-way analysis of variance (ANOVA) was performed for eachtest and each parameter (N, A and �) at the significance level˛ = 0.05. In most cases the p-value was <0.001. In some cases thisvalue was higher and at two sampling occasions in two tests thep-value was >0.05 (Csp-PM, all parameters at day 7 and MS-PM,parameter A for flagellates at day 5). Correspondence with normaldistribution was seen in the data of all tests applying Kolmogorov-Smirnov test on normal distribution (significance level ˛ = 0.05).The Cochran’s test was used for testing homogeneity of variancesand thereafter the Student t-test with Bonferroni-adjustment wasapplied for NOEC (no observed effect concentration) and LOEC(lowest observed effect concentration) determination. In the caseswhen variances homogeneity was not given, the Kruskal–Wallis-test procedure (non-parametric ANOVA) was applied. Followingthat procedure, the Welch t-test for inhomogeneous varianceswith Bonferroni-adjustment was used for NOEC and LOEC deter-mination. ECX values were derived by probit analysis using linearmaximum likelihood estimation. Goodness of fit was assessed bythe parameter Chi2. For all statistical evaluations the softwareToxRat Professional V.2.09 (ToxRat Solutions GmbH, 2001–2006,Germany) was used.
3. Results
3.1. Abiotic parameters
The pH of the test solutions varied between all tests in the rangeof 6.8–7.5. However, a concentration–effect relationship was appar-ently not established. Within each test the pH did not change bymore than 0.5 units, except the flagellate single-species test withprometryn (Csp-P) where the pH increased from initially 7.1–7.3 to7.8–8.3 at the end of the test at all treatment levels and the control.In the multi-species and ciliate single-species tests the oxygen con-tent was above 90% saturation. During the flagellate single-speciestests oxygen was not measured. The temperature of 20 ± 1.5 ◦C was
maintained during all tests. An exception was the transient increaseof the temperature at day 4 of the multi-species test with prometryn(MS-P) up to a maximum of 28.6 ◦C resulting in slightly modifiedgrowth behaviour of the organisms.
3.2. Chemical analyses
In the tests Csp-P, Csp-PM, Uf-PM and MS-P the chemical analy-sis showed recoveries within a range of ±20% of the nominal valuesthroughout the test periods. In the tests Uf-P and MS-PM the mea-sured concentrations were considerably lower and reached valuesof 59.8% and 64.7% of the nominal values, respectively (geomet-ric means, Uf-P: n = 4; MS-PM: n = 6). However, a relevant decreaseduring the course of tests was not observed. Hence, for all teststhe calculated effect concentrations were corrected by the meanrecoveries.
3.3. Nutrient contents in medium and cells
The total nitrogen content in the medium decreased almost con-stantly due to the consumption of nitrate by the test organisms.
n.d. = not determined.a Cellular nutrient content at day 0 was determined in the pre-cultures before in
In the fresh WC medium the bioavailable nitrogen of nominally2.8 mg nitrogen L−1 is derived from NaNO3 and vitamins. This wasalmost depleted at day 7 (Table 1). The non-bioavailable nitro-gen is derived from TES-buffer and the Na2-EDTA contained inthe medium. The TOC content of the medium increased duringthe course of the tests. A concentration–effect relationship withregard to nutrient content of the medium was not observed in bothflagellate single-species tests.
Table 1 shows further the shift of nutrient contents in the flag-
ellate cells during the tests. In the test Csp-P, the total nitrogencontent in cells decreased at all test concentrations and the con-trols, except for the highest test item concentration C5. The shift ofTOC content in the cells was not as pronounced as the total nitro-gen content. The largest decrease of TOC content occurred at thetest concentrations C2 and C5, where the lowest shift of nitrogencontent per cell was observed. In the test Csp-PM, the total nitro-gen content in cells tended to increase slightly until the end of thetest, except for cells of the controls and the concentration level C2.Until the end of the test the average TOC content per cell increasedby 35% compared to the initial TOC content measured in the pre-culture. However, there is no consistency of the derived data whichwould imply an effect of the test items on the nutrient content inthe flagellate cells or in the medium.
In the multi-species tests the initial total nitrogen content in thetest medium was lower than in the single-species test (Table 2). Inthe test with prometryn (MS-P), after a decline of the total nitrogencontent until day 5 there was a subsequent increase until the end ofthe test. The initial TOC content of the medium started somewhathigher compared to the single-species tests. A slight increase ofthe TOC content was observed during the test periods. However,
Table 2Total nitrogen and TOC contents of filtrated medium during both multi-species tests MS-
Test code Concentration code Total nitrogen (mg L−1)
no concentration effect relationship could be detected in any of thetests. The measurements of nutrient contents in the cells showedinconsistent data which are not shown.
During the first week the growth curves of Cryptomonas sp.exposed to prometryn (Csp-P) show treatment related effects
(Fig. 1). After the first week of exposure the declining nutrient con-tent in the test medium limited cell growth (cf. Table 1), whicheventually at the end of the test period led to similar cell numbersin all treatments and control. The parameter A proved to be sta-tistically more sensitive than N and �. The NOEC related to A was6.9 �g L−1 (p < 0.001) at day 7 and 15.5 �g L−1 at day 14 (p = 0.031)(Table 3).
In the test with parathion-methyl (Csp-PM) growth of Cryp-tomonas sp. was not affected during the first seven days of exposure(ANOVA: p = 0.09). The highest test item concentration C5 showedthe lowest cell numbers, but the next lower test item concentra-tion C4 reached the same cell numbers as observed in the control.However, between day 7 and day 14 when nutrients in the testmedium became limited, cell growth in the controls and concen-tration levels C1–C3 reached the static phase at approximately26 × 104 cells mL−1, whereas at the highest concentration levels C4and C5 the total cell numbers decreased. The parameter A proved tobe statistically less sensitive than N and �. The lowest NOEC relatedto N and � was 0.7 mg L−1 (C3) (p < 0.001).
Results of the probit for both flagellate tests are shown in Table 3.In the test Csp-P, again, the most sensitive parameter was A with an
106 M. Liebig et al. / Aquatic Toxicology 88 (2008) 102–110
Fig. 1. Growth curves for cell numbers (A), area below the growth curves (B), and averaflagellate growth inhibition tests with prometryn Csp-P (C0 = control, C1–C5 = 6.9, 10.3, 150.70, 1.30, 2.40 mg L−1); (n = 4, standard deviation; statistical analyses at days 7 and 14: *sign
Table 3Flagellate growth inhibition tests Csp-P and Csp-PM: statistical evaluation of aver-age cell number N, area below the growth curve A and average specific growth rate� (n = 4) after 7 and after 14 days of exposure based on measured substance concen-trations (values in squared brackets are outside of the tested concentration range;upper/lower 95% confidence limits in parentheses)
Parameter NOEC LOEC EC10 EC50
Csp-P: effect concentrations at day 7 of exposure (�g L−1)N 23.2 34.8 18.2 (14.9/20.54) 31.5 (29.5/34.1)A 6.9 10.3 7.9 (2.8/11.4) 22.9 (18.0/32.1)� 23.2 34.8 21.6 (19.3/23.3) [39.3] (37.1/42.4)
Csp-P: effect concentrations at day 14 of exposure (�g L−1)N 34.8 [>34.8] [37.9] (n.d.)b [100.0] (n.d.)b
a Not determined since p > 0.05 (ANOVA).b Not determined due to mathematical reasons.
ge specific growth rates (C) dependent on measured test item concentrations for.5, 23.2, 34.8 �g L−1) and parathion-methyl Csp-PM (C0 = control, C1–C5 = 0.20, 0.38,ificantly different from control, p < 0.05, Student t-test with Bonferroni-adjustment).
EC10 of 7.9 �g L−1 and 16.6 �g L−1 at day 7 and day 14, respectively,and with an EC50 of 22.9 �g L−1 and 36.3 �g L−1 at day 7 and day 14,respectively. It is pointed out that the EC50 values calculated at day14 are outside of the tested concentration range and therefore of
higher uncertainty. In the test Csp-PM, due to the low effect levelsobserved, an EC50 could not be determined with acceptable accu-racy. Only at day 14 the EC10 could be determined with the lowestvalue for the parameter A at 0.68 mg L−1.
In both ciliate tests, Uf-P and Uf-PM, within the first 24 h thecell numbers increased in the control and the concentration lev-els C1–C3. Obviously, the cells were still able to reproduce by celldivision due to their internal reserves. After 48 h a strong reduc-tion of cell numbers was observed at all test concentrations, thatwas probably at least partially caused by starvation. Therefore, theresults of the statistical evaluation are given only for the time point24 h (Table 4).
3.6. Multi-species toxicity tests (MS-P and MS-PM)
In the multi-species tests with both pesticides (Figs. 2 and 3)the ciliates experienced an initial lag phase of about 1–2 days.
M. Liebig et al. / Aquatic Toxicology 88 (2008) 102–110 107
Fig. 2. Growth curves for cell numbers (A), area below the growth curves (B), and average specific growth rates (C) dependent on measured test item concentrations forthe multi-species tests with prometryn MS-P (C0 = control, C1–C3 = 7.6, 15.2, 30.3 �g L−1; controls: n = 6, treatments: n = 4; standard deviation; statistical analyses at day 6(flagellates) and 8 (ciliates): *significantly different from control, p < 0.05, Student t-test with Bonferroni-adjustment).
Thereafter, the exponential growth phase of the ciliates started.The flagellate populations of all treatments and controls decreasedduring the first three days of exposure.
In the multi-species test with prometryn (MS-P, Fig. 2), consid-ering the specific mode of action of prometryn as an herbicideand the results of the single-species test Uf-P, direct effects onthe ciliates should not be expected at the applied concentra-tion levels. Nevertheless, growth of the ciliates was significantlyreduced at the highest concentration during the exponentialgrowth phase, but this effect follows the flagellate growth inhi-bition. However, at all treatment levels and the control, thecell numbers of flagellates as well as the cell numbers of cil-
Table 4Effect concentrations for the ciliate short-term toxicity tests Uf-P and Uf-PM assess-ing survival of Urotricha furcata based on measured substance concentrations(upper/lower 95% confidence limits in parentheses)
Test code Effect concentrations after 24 h of exposure (mg L−1)
Table 5Multi-species tests MS-P and MS-PM: effect concentrations for the parameter cellnumber N, area below the growth curve A and average specific growth rate � ofCryptomonas sp. and Urotricha furcata at their maximum of growth based on mea-sured substance concentrations (value in squared brackets is outside of the testedconcentration range; upper/lower 95% confidence limits in parentheses; C0: n = 6,C1–C3: n = 4, ˛ = 0.05)
Parameter Cryptomonas sp. day 6 (�g L−1) U. furcata day 8 (�g L−1)
a Not determined due to mathematical reasons.b Not determined since p > 0.05 (ANOVA).
108 M. Liebig et al. / Aquatic Toxicology 88 (2008) 102–110
verage, 2.59 mdent t-
Fig. 3. Growth curves for cell numbers (A), area below the growth curves (B), and amulti-species tests with parathion-methyl MS-PM (C0 = control, C1–C3 = 0.26, 0.82day 5 (flagellates) and 7 (ciliates): *significantly different from control, p < 0.05, Stu
iates reached the same densities at the end of the exposureperiod.
In the test with parathion-methyl (MS-PM, Fig. 3) the ciliate con-trol populations showed significantly higher growth (cell numbersN, p < 0.001) and hence higher feeding activity than the treatedsolutions. Nevertheless, the control flagellates showed also thehighest cell numbers compared to the treatments. A pronouncedconcentration–effect relationship for both test organisms remainedthroughout the test period.
In both multi-species tests effects were most prominent dur-ing the growth phases and at the maximum of growth, wherebythe ciliates reached their growth maximum two days after theflagellates. The effect concentrations were statistically derived atthat time point when the organisms reached their growth peak(Table 5).
4. Discussion and conclusions
For comparison with effects data for algae derived from scien-tific literature only those data that were determined at exposureday 7 in the present study for the flagellates should be used. Afterseven days the flagellate populations were suffering from limited
specific growth rates (C) dependent on measured test item concentrations for theg L−1; controls: n = 6, treatments: n = 4; standard deviation; statistical analyses at
test with Bonferroni-adjustment).
nutrient availability and therefore these effects data are not com-parable with those data derived according to standard procedures
providing unlimited nutrient conditions. Similarly, the ciliate pop-ulations in the single-species tests were influenced at least partiallyby starvation after 48 h of exposure. Thus, for comparison with lit-erature data the 24 h effect concentrations for the ciliates shouldbe considered.
The EC50 (7d) of 22.8 �g prometryn L−1 for the Cryptomonas sp. issimilar to values from literature for other green algae like the EC50of 12 �g L−1 calculated for growth inhibition of Selenastrum capri-cornutum (Office of Pesticide Programs, 2005). Gaggi et al. (1995)determined an EC50 of 21 �g L−1 and 53 �g L−1 for S. capricornutumand Dunaliella tertiolecta, respectively. Likewise, the NOEC(7d) of1.3 mg parathion-methyl L−1 observed for Cryptomonas sp. is com-parable with the NOEC values of 0.2–1.5 mg L−1 determined for thegreen alga Chlamydomonas reinhardi by Schafer et al. (1994).
For the ciliates only one value from literature was found forParamecium aurelia exposed for 36 min to prometryn resulting ina LC100 of 10 mg L−1 (U.S.EPA, 2007). This is in the same range asthe EC50 (24h) of 4.2 mg L−1 determined in our study. Schafer et al.(1994) derived effects data for parathion-methyl using the ciliatespecies Tetrahymena pyriformis with a NOEC(48h) of 2.1 mg L−1 andan EC50 (48h) of 4.7 mg L−1. However, the observed endpoint for
the bacterivorous ciliate T. pyriformis was growth inhibition andnot survival as assessed for the algivorous U. furcata in our study.Twagilimana et al. (1998) determined an EC50 (24h) of 8.2 mg L−1
for another bacterivorous ciliate, Spirostomum teres, also assessinggrowth inhibition caused by parathion-methyl.
The assessment of bacterial density did not show aconcentration–effect relationship in any case (data not shown).Hence, it can be concluded that during the tests the bacterialcommunities could fulfil their function as mineraliser and recyclerof nutrients without limitation.
For the multi-species tests corresponding data are not availablefrom literature. However, the effects observed for the flagellatesin both multi-species tests were in the same range as in the single-species tests. Likewise, the effect concentration for ciliates observedin the multi-species test with parathion-methyl MS-PM was in thesame range as in the single-species test. Since in the test MS-PMboth the flagellates and the ciliates of the control showed the high-est cell numbers compared to the treatments, a direct effect ofparathion-methyl on both test organisms can be assumed. If theflagellates were not affected directly at the higher concentrationlevels, they would have grown equal or even superior to the con-trol, due to less feeding pressure. In the multi-species test withprometryn MS-P, effects on ciliates occurred at a concentration afactor 149 lower compared to the single-species test. This highersensitivity was possibly caused by an indirect effect due to reducedavailability of flagellates as food, i.e. the propagation of effects fromone trophic level to the next.
Indirect effects, with and without toxic stress, were alsoobserved in similar test systems, e.g. in chemostats, using othertest organisms (e.g. Lawrence et al., 1989; Lawler, 1993). In somecases the impact of the toxicant was expressed stronger in itsindirect than in its direct effects (e.g. van den Brink et al., 1995). Fur-thermore, although not a general rule (Traunspurger et al., 1996),community parameters in multi-species tests such as abundanceand species composition can indicate higher sensitivity towardstoxicity than parameters observed in single-species tests (Lampertet al., 1989).
The design of the can-com tests was chosen to provide data thatcan be used for effects modelling based on the DEB-Tox-module ofthe Dynamic Energy Budget (DEB) theory (Kooijman, 2000). Onespecific feature was the measurement of nutrient contents in cellsand test media. The analyses of total nitrogen and TOC contentsallow assessing the availability of nutrients and the physiologicalstate of the test organisms. Due to the overlapping cell size of Cryp-
tomonas and Urotricha, the separation of the organisms from eachother in the multi-species test turned out to be not feasible withoutunacceptable loss of accuracy. The data obtained from the single-species flagellate tests illustrate that the nutrient content in cellsmight fluctuate strongly within the same batches during the testcourse. Although the cells used as inoculum were always derivedfrom the exponential growth phase, the physiological and hencenutritional state of the cells may develop distinct during the differ-ent test courses. The inherent high variance within and betweentreatment groups was hiding any possible effect on nutrient con-tent caused by the toxicant. Therefore, this parameter was not usedfor the evaluation of sublethal effects on the flagellates.
Another feature of the test design was the reduced nitrate con-tent of the modified WC medium to a level that the decrease ofthe bioavailable nitrogen until day 7 caused a stationary phase inthe flagellate single-species tests. The lower initial nitrogen andthe higher initial TOC content of the medium in the multi-speciestests was caused by the higher portion of cell inoculum neededto reach the higher initial cell densities compared to the single-species tests. Therefore, the decrease of total nitrogen was not aspronounced as during the single-species test with Cryptomonas. A
logy 88 (2008) 102–110 109
stationary phase was required for modelling purposes according tothe DEB-theory. In a stable algae population individuals die and arereplaced by new individuals. Thus, net growth of a population canbe zero, while there remains a (large) turn-over of biomass. Thisturn-over is determined by the individual growth and death rates.A toxicant could influence one or both rates, and then a new balancebetween population growth and decline will be found. This new sta-tionary phase would be at a lower absolute cell number. In addition,the growth curve would show retardation compared to the controlgroup. Thus, both the growth phase and stationary phase containinformation on how and which biological processes are affected,e.g. growth related nutrient assimilation or hazard rate.
The advantage of the DEB-Tox-models is the integration of allthe data produced during the test period resulting in an overall‘no effect concentration’ (NEC) which is independent of the evalu-ated time point (Kooijman and Bedaux, 1996; Kooijman et al., 1996;Jager et al., 2006). In contrast, the NOEC is derived statistically foran evaluated parameter and for a certain time point of exposure.An exception is the parameter area below the growth curve (A)which integrates all values measured during an exposure period. Anexample for the time dependence of the NOEC was demonstrated inthe multi-species test with prometryn. During exponential growthafter 6 days of exposure significant effects on the absolute cellnumber N were observed, whereas after 13 days of exposure theseeffects were mitigated and not significant anymore.
The simple multi-species test described here can be used toassess direct and indirect effects caused by toxic stress on popu-lation density and growth parameters across several trophic levels,which are adequate markers for the evaluation of the function-ing of ecosystems (Sugiura, 1992; Kooi et al., 2008). In short-termand single-species tests such interactions among species are notdetectable. In order to assess the potential of community changeslong-term toxicity studies are required, preferentially over sev-eral generations. Indoor multi-species tests are useful in highertier environmental risk assessment when potential risks to severalpopulations and/or a community have been identified. Compara-ble tests in literature indicate that such indoor-defined microcosmstests may be used as a tool to better define the designs of complexfield studies (Campbell et al., 1999).
However, since the outcome of multi-species or microcosmstudies varies strongly depending on the applied test design andon the ecological abiotic and biotic besides the toxicological testconditions, such test systems should be at least partially standard-ised in order to use the results in environmental risk assessment. In
this respect, the multi-species test system described here has sev-eral advantages: initial test conditions such as species assemblageand relative abundance as well as the nature of the test medium canbe controlled and adjusted; the test organisms fulfil important eco-logical roles and hence represent environmentally relevant speciesof natural freshwater ecosystems; the test system is favourablein respect of its easy implementation, simple handling and lowcosts.
Acknowledgements
Many thanks to Anja Coors for valuable support in the statisticalanalyses and to two anonymous reviewers for their helpful com-ments on the manuscript. This study was financially supported bythe EU-project MODELKEY (Contract No. 511237-GOCE) within theSixth Framework Programme.
References
ARGE Elbe, 2006. Gewassergutebericht der Elbe 2005. Arbeitsgemeinschaft fur dieReinhaltung der Elbe (Eds.). Wassergutestelle, Hamburg, pp. 1–64.
ARGE Elbe, 2007. Wassergutedaten der Elbe von Schmilka bis zur See – Zahlentafeln2005. Arbeitsgemeinschaft fur die Reinhaltung der Elbe (Eds.). Wassergutestelle,Hamburg, pp. 1–205.
Campbell, P.J., Arnold, D.J.S., Brock, T.C.M., Grandy, N.J., Heger, W., Heimbach, F.,Maund, S.J., Streloke, M., 1999. Guidance document on higher-tier aquatic riskassessment for pesticides (HARAP). Report from the SETAC-Europe/OECD/ECWorkshop, 19–22 April 1998, Lacanau Ocean, France. SETAC-Europe, Brussels.
EC—European Commission, 2003. Commission decision of 10 March 2003 con-cerning the non-inclusion of parathion-methyl in Annex I to Council Directive91/414/EEC and the withdrawal of authorisations for plant protection productscontaining this active substance. 2003/166/EC. Official Journal L67, pp. 18–19.
Fleeger, J.W., Carman, K.R., Nisbet, R.M., 2003. Indirect effects of contaminants in
aquatic ecosystems. Sci. Total Environ. 317, 207–233.
Foissner, W., Berger, H., Schaumburg, J., 1999. Identification and ecology of limneticplankton ciliates. Bayerisches Landesamt fur Wasserwirtschaft, Munchen.
Gaggi, C., Sbrilli, G., El Naby, A.M.H., Bucci, M., Duccini, M., Bacci, E., 1995. Toxicityand hazard ranking of s-triazine herbicides using Microtox®, two green algalspecies and a marine crustacean. Environ. Toxicol. Chem. 14, 1065–1069.
Guillard, R.R.L., Lorenzen, C.J., 1972. Yellow-green algae with chlorophyllide c. J.Phycol. 8, 10–14.
Jager, T., Heugens, E.H.W., Kooijman, S.A.L.M., 2006. Making sense of ecotoxicologi-cal test results: towards application of process-based models. Ecotoxicology 15,305–314.
Koelmans, A., Van der Heijde, A., Knijff, L., Aalderink, R., 2001. Integrated modellingof eutrophication and organic contaminant fate & effects in aquatic ecosystems.A review. Water Res. 35, 3517–3536.
Kooi, B.W., 2003. Numerical bifurcation analysis of ecosystems in a spatially homo-geneous environment. Acta Biotheor. 51, 189–222.
Kooi, B.W., Bontje, D., Voorn van, G.A.K., Kooijman, S.A.L.M., 2008. Sublethal toxiceffects in a simple aquatic food chain. Ecol. Model. 212, 304–318.
Kooijman, S.A.L.M., Bedaux, J.J.M., 1996. Some statistical properties of estimates ofno-effect concentrations. Water Res. 30, 1724–1728.
Kooijman, S.A.L.M., 2000. Dynamic Energy and Mass Budgets in Biological Systems,vol. 2, Press Syndicate of the University of Cambridge, Cambridge.
Lampert, W., Fleckner, W., Pott, E., Schober, U., Storkel, K.U., 1989. Herbicide effects onplanktonic systems of different complexicity. Hydrobiologia 188/189, 415–424.
Lawler, S.P., 1993. Direct and indirect effects in microcosm communities of protists.Oecologia 93, 184–190.
logy 88 (2008) 102–110
Lawrence, S.G., Holoka, M.H., Hamilton, R.D., 1989. Effects of cadmium on a microbialfood chain, Chlamydomonas reinhardii and Tetrahymena vorax. Sci. Total Environ.87/88, 381–395.
Muller, H., 1989. The relative importance of different ciliate taxa in the pelagic foodweb of Lake Constance. Microb. Ecol. 18, 261–273.
OECD, 1984. OECD Guideline for Testing of Chemicals—201. Alga, Growth InhibitionTest. OECD, Organisation for Economic Co-operation and Development, Paris.
OECD, 2006. OECD Guidelines for the Testing of Chemicals—201. Freshwater Algaand Cyanobacteria, Growth Inhibition Test. OECD, Organisation for EconomicCo-operation and Development, Paris.
Office of Pesticide Programs, 2000. Pesticide Ecotoxicity Database. EnvironmentalFate and Effects Division, U.S.EPA, Washington, D.C. Cited at: ECOTOX—
Ecotoxicology Database: http://www.epa.gov/cgi-bin/ecotox quick searc(accessed: 19 April 2005).
Porter, K.G., Feig, Y.S., 1980. The use of DAPI for identifying and counting aquaticmicroflora. Limnol. Oceanogr. 25, 943–948.
Sanders, R.W., Porter, K.G., 1988. Phagotrophic phytoflagellates. In: Marshall, K.C.(Ed.), Advances in Microbial Ecology. Plenum Press, New York, pp. 167–182.
Schafer, H., Hettler, H., Fritsche, U., Pitzen, G., Roderer, G., Wenzel, A., 1994. Biotestsusing unicellular algae and ciliates for predicting long-term effects of toxicants.Ecotoxicol. Environ. Saf. 27, 64–81.
Schonberger, M., 1994. Planktonic ciliated protozoa of Neusiedler See (Aus-tria/Hungary): a comparison between the turbid open lake and a reedlessbrown-water pond. Marine Microb. Food Webs 8, 251–263.
Sugiura, K., 1992. A multi-species laboratory microcosm for screening ecotoxicolog-ical impacts of chemicals. Environ. Toxicol. Chem. 11, 1217–1226.
Traunspurger, W., Schafer, H., Remde, A., 1996. Comparative investigation on theeffect of a herbicide on aquatic organisms in single-species tests and aquaticmicrocosms. Chemosphere 33, 1129–1141.
Twagilimana, L., Bohatier, J., Groliere, C.-A., Bonnemoy, F., Sargos, D., 1998. A newlow-cost microbiotest with the protozoan Spirostomum teres: culture conditionsand assessment of sensitivity of the ciliate to 14 pure chemicals. Ecotoxicol.Environ. Saf. 41, 231–244.
U.S.EPA, 2007. ECOTOX database: http://cfpub.epa.gov/ecotox (accessed: May 2007).Van den Brink, P.J., van Donk, E., Gylstra, R., Crum, S.J.H., Brock, T.C.M., 1995. Effects of
chronic low concentrations of the pesticides chlorpyrifos and atrazine in indoorfreshwater microcosms. Chemosphere 31, 3181–3200.
Weisse, T., Frahm, A., 2001. Species-specific interactions between small plank-tonic ciliates (Urotricha spp.) and rotifers (Keratella spp.). J. Plankton Res. 23,1329–1338.
