Pesticide exposure and inducible antipredator responses in the zooplankton grazer, Daphnia magna Straus João L.T. Pestana a, * , Susana Loureiro a , Donald J. Baird b , Amadeu M.V.M. Soares a a CESAM and Department of Biology, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal b Environment Canada at Canadian Rivers Institute, University of New Brunswick, Department of Biology, New Brunswick, Canada article info Article history: Received 23 June 2009 Received in revised form 20 October 2009 Accepted 27 October 2009 Keywords: Predator–prey interactions Imidacloprid Brown trout Infochemicals Multiple stressors abstract Risk assessment of toxic substances under ecologically-relevant scenarios which include the presence of natural stressors is essential to understand their indirect toxic effects and to improve prediction of the impacts of contamination on community structure and ecosystem function. Here, we study the effects of the pesticide imidacloprid on the responses of Daphnia magna to a com- bination of predator-release kairomones from trout and alarm cues from conspecifics, simulating differ- ent levels of perceived predation risk. The joint effects of simultaneous exposure to both types of stressors were assessed both by traditional analysis of variance and by employing conceptual models for the evaluation of contaminant mixture exposures. Results demonstrated that pesticide exposure can significantly increase the costs of inducible antipredator defences and impair life-history responses of daphnids to fish predation pressure. Since trait-mediated effects are well-known to play a key role in population dynamics, the combined direct and indirect effects of sub-lethal concentrations of pesti- cides could induce maladaptive responses in zooplankton populations in the field, reducing their long- term viability. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction In lakes and wetlands in areas of intense agricultural activity, zooplankton populations have to cope with both natural and anthropogenic structuring pressures acting simultaneously: e.g. fish predation and pesticide exposure (Hanazato, 2001). Predators directly affect prey populations by reducing their density, and by inducing energetically costly defensive traits (Preisser et al., 2005). In response to the presence of fish predators daphnids show changes in behavioural and life-history traits (Pija- nowska and Kowalczewski, 1997; Boersma et al., 1998; Tollrian and Harvell, 1999; Lass and Spaak, 2003). The presence of preda- tors has also been shown to reduce energy intake (Rose et al., 2003; Beckerman et al., 2007) place higher metabolic demands on daphnids (Pijanowska, 1997; Beckerman et al., 2007) and also to induce molecular mechanisms to cope with this predatory pres- sure (Pijanowska and Kloc, 2004; Pauwels et al., 2005). Pollutants can mimic the effects of predatory chemical cues or inhibit the induction of defences and have therefore the potential to disturb predator–prey interactions (Dodson et al., 1995; Barry, 1998; Hanazato, 1999; Hunter and Pyle, 2004). At the same time, responses of daphnids in the presence of predators could reduce daphnid tolerance to other environmental stressors and in turn affect their sensitivity to toxic substances (Hanazato, 2001; Kieu et al., 2001; Rose et al., 2001). Pesticides have been shown to cause changes in the behavioural (Dodson et al., 1995) and life-history (Hanazato and Dodson, 1992; Barry, 2000) responses of daphnids to predators. Focused research on multiple stressors is needed to improve understanding of their interactions, and to facilitate the assess- ment of additive or more-than-additive effects (Relyea and Hover- man, 2006). In this way, we can improve risk assessment by incorporating further realism, in a systematic fashion (Van Straa- len, 2003; Eggen et al., 2004; Relyea and Hoverman, 2006). More- over, the assessment of pesticide effects on predator–prey interactions and the incorporation of sub-lethal endpoints such as behaviour, physiology and life history would permit more pre- cise extrapolation of their effects on natural populations (Hanaz- ato, 2001; Fleeger et al., 2003; Relyea, 2005; Relyea and Hoverman, 2006). Imidacloprid is a systemic insecticide, belonging to a class of chloronicotinyl insecticides, acting on the nicotinic acetylcholine receptors (nAChRs) (Tomizawa and Casida, 2003). Prolonged acti- vation of the nAChRs by imidacloprid causes desensitization and receptor blocking, decreasing activity, reducing muscular coordi- nation, inducing tremors and ultimately, death (Moffat, 1993). Here, we examine how imidacloprid interacts with behavioural, physiological and life-history responses of Daphnia magna to pred- atory chemical cues from fish and we hypothesize that exposure to 0045-6535/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2009.10.066 * Corresponding author. Tel.: +351 234 370 792; fax: +351 234 372 587. E-mail address: [email protected](J.L.T. Pestana). Chemosphere 78 (2010) 241–248 Contents lists available at ScienceDirect Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Pesticide exposure and inducible antipredator responses in the zooplanktongrazer, Daphnia magna Straus
João L.T. Pestana a,*, Susana Loureiro a, Donald J. Baird b, Amadeu M.V.M. Soares a
a CESAM and Department of Biology, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugalb Environment Canada at Canadian Rivers Institute, University of New Brunswick, Department of Biology, New Brunswick, Canada
a r t i c l e i n f o
Article history:Received 23 June 2009Received in revised form 20 October 2009Accepted 27 October 2009
Risk assessment of toxic substances under ecologically-relevant scenarios which include the presence ofnatural stressors is essential to understand their indirect toxic effects and to improve prediction of theimpacts of contamination on community structure and ecosystem function.
Here, we study the effects of the pesticide imidacloprid on the responses of Daphnia magna to a com-bination of predator-release kairomones from trout and alarm cues from conspecifics, simulating differ-ent levels of perceived predation risk. The joint effects of simultaneous exposure to both types ofstressors were assessed both by traditional analysis of variance and by employing conceptual modelsfor the evaluation of contaminant mixture exposures. Results demonstrated that pesticide exposurecan significantly increase the costs of inducible antipredator defences and impair life-history responsesof daphnids to fish predation pressure. Since trait-mediated effects are well-known to play a key rolein population dynamics, the combined direct and indirect effects of sub-lethal concentrations of pesti-cides could induce maladaptive responses in zooplankton populations in the field, reducing their long-term viability.
� 2009 Elsevier Ltd. All rights reserved.
1. Introduction
In lakes and wetlands in areas of intense agricultural activity,zooplankton populations have to cope with both natural andanthropogenic structuring pressures acting simultaneously: e.g.fish predation and pesticide exposure (Hanazato, 2001).
Predators directly affect prey populations by reducing theirdensity, and by inducing energetically costly defensive traits(Preisser et al., 2005). In response to the presence of fish predatorsdaphnids show changes in behavioural and life-history traits (Pija-nowska and Kowalczewski, 1997; Boersma et al., 1998; Tollrianand Harvell, 1999; Lass and Spaak, 2003). The presence of preda-tors has also been shown to reduce energy intake (Rose et al.,2003; Beckerman et al., 2007) place higher metabolic demandson daphnids (Pijanowska, 1997; Beckerman et al., 2007) and alsoto induce molecular mechanisms to cope with this predatory pres-sure (Pijanowska and Kloc, 2004; Pauwels et al., 2005).
Pollutants can mimic the effects of predatory chemical cues orinhibit the induction of defences and have therefore the potentialto disturb predator–prey interactions (Dodson et al., 1995; Barry,1998; Hanazato, 1999; Hunter and Pyle, 2004). At the same time,responses of daphnids in the presence of predators could reducedaphnid tolerance to other environmental stressors and in turn
ll rights reserved.
x: +351 234 372 587.
affect their sensitivity to toxic substances (Hanazato, 2001; Kieuet al., 2001; Rose et al., 2001). Pesticides have been shown to causechanges in the behavioural (Dodson et al., 1995) and life-history(Hanazato and Dodson, 1992; Barry, 2000) responses of daphnidsto predators.
Focused research on multiple stressors is needed to improveunderstanding of their interactions, and to facilitate the assess-ment of additive or more-than-additive effects (Relyea and Hover-man, 2006). In this way, we can improve risk assessment byincorporating further realism, in a systematic fashion (Van Straa-len, 2003; Eggen et al., 2004; Relyea and Hoverman, 2006). More-over, the assessment of pesticide effects on predator–preyinteractions and the incorporation of sub-lethal endpoints suchas behaviour, physiology and life history would permit more pre-cise extrapolation of their effects on natural populations (Hanaz-ato, 2001; Fleeger et al., 2003; Relyea, 2005; Relyea andHoverman, 2006).
Imidacloprid is a systemic insecticide, belonging to a class ofchloronicotinyl insecticides, acting on the nicotinic acetylcholinereceptors (nAChRs) (Tomizawa and Casida, 2003). Prolonged acti-vation of the nAChRs by imidacloprid causes desensitization andreceptor blocking, decreasing activity, reducing muscular coordi-nation, inducing tremors and ultimately, death (Moffat, 1993).
Here, we examine how imidacloprid interacts with behavioural,physiological and life-history responses of Daphnia magna to pred-atory chemical cues from fish and we hypothesize that exposure to
242 J.L.T. Pestana et al. / Chemosphere 78 (2010) 241–248
imidacloprid can affect Daphnia–fish interactions by causing alter-ations in the direction or magnitude of induced responses to fishperceived predation risk.
2. Materials and methods
2.1. Experimental animals
D. magna, clone F (sensu (Baird et al., 1991)) was cultured undera 16:8 light dark photoperiod and 20� ± 2 �C in ASTM (AmericanSociety for Testing Materials) hard water enriched with organic ex-tract. Animals were fed the green alga Chlorella vulgaris Beijerinck,at a concentration of 3.0 � 105 cells/mL. Medium and food were re-newed every other day and neonates from the 3rd or 4th clutchwere used in all the experiments.
Salmo trutta Linnaeus, the Brown trout, was used as our modelvertebrate predator. Brown trout feed on aquatic and terrestrial in-sects, small fish and crustaceans including daphnids (Penczak andFormigo, 2000). Young fish were obtained from a fish-farm andmaintained in the laboratory at 17 ± 1 �C in 60 L plastic tanks withaerated artificial pond water (Naylor et al., 1989). After approxi-mately one month, healthy fish were selected for the preparationof chemical cues.
2.2. Preparation of predatory chemical cues
A combination of kairomones from brown trout (i.e. water con-ditioned by fish for 24 h) and also alarm cues from maceratedDaphnia were used to prepare the following experimental treat-ments (see (Pestana, 2008) for more details on chemical cuepreparation):
– No risk of predation = No chemical cues added– Low perceived risk of predation = (0.05 fish + 1 crushed Daphnia)
L�1
– High perceived risk of predation = (0.2 fish + 4 crushed Daphnia)L�1
2.3. Test chemical
Confidor�
200 SL was purchased from Bayer CropScience AG(Monheim, Germany) and was used to prepare the appropriatestock solutions of imidacloprid with distilled water. The concentra-tions of stock solutions used in the experiments were determinedat Terracon laboratories (Germany) from the original samples byHPLC–UV at 270 nm (see SI for details) and ranged from 427 mg/L to 444 mg/L. Stock solutions were stored at 4 �C protected fromlight. Imidacloprid concentrations were measured after 48 h in lifetable experiment to determine pesticide degradation in treatmentswith and without predatory chemical cues.
2.4. Acute toxicity experiments
Acute lethality (OECD, 2000) was estimated with organisms (10per treatment) exposed for 48 h to imidacloprid (0, 25, 50, 75, 100,125, 150, 175 and 200 mg/L) and no food. Mortality assessmentwas also conducted in the presence of predator exudates (high per-ceived predation risk treatment only), to find out if these alteredthe lethal sensitivity of D. magna to imidacloprid.
2.5. Feeding experiments
Three concentrations of imidacloprid (2.2, 4.4 and 8.8 mg/L)plus control were studied across three perceived predation risktreatments (zero, low and high perceived risk of predation) in a
total of 12 experimental treatments with five replicates each. Thefeeding experiments were performed in the dark with 25 4-d oldjuveniles per treatment in 150-mL glass vials containing 100 mLof solution plus C. vulgaris at 3 � 105 cells ml (Pestana, 2008).
2.6. Measurement of oxygen consumption
Two concentrations of imidacloprid (2.2 and 4.4 mg/L) plus con-trol were tested in a fully crossed design with three levels of per-ceived predation risk (zero, low and high). The oxygenconsumption experiment was performed using standard respirom-etry methods with 50-mL gastight syringes (Hamilton, USA). Threesyringes were used per treatment, each filled with 30 mL of testsolution (no food added) containing five 4-d old D. magna juve-niles. After expelling any remaining air, syringes were placed inthe dark in a water bath (20 �C) for 24 h. Oxygen consumptionwas calculated as the difference in the oxygen content of water be-fore and after the exposure period. A more complete description ofthis methodology is given elsewhere (Pestana, 2008).
2.7. Life table experiment
A 21-d experiment was performed (OECD, 1998), with daphnidneonates (<12 h old), with three concentrations of imidacloprid(2.2, 4.4 and 8.8 mg/L) plus control at each of three levels of per-ceived predation risk (zero, low and high). Ten replicates wererun for each treatment with one organism per replicate. Test solu-tion was renewed every other day and C. vulgaris was suppliedevery day at a concentration of 3.0 � 105 cells/ml. Age and size atfirst reproduction, number of broods, total number of offspringper female, and also the intrinsic rate of population increase (r)were the life-history parameters measured. Size of experimentalorganisms was also measured in the start (30 randomly selectedneonates) and at the end of the experimental period (21 d) to as-sess effects on somatic growth.
2.8. Data analysis
EC50 values for mortality (measured as immobilization) werecalculated using the probit method. For feeding, oxygen consump-tion and somatic growth (measured over 21 d), two-way ANOVAswere performed using perceived predation risk levels and imida-cloprid concentrations as fixed factors. Life historical data was ana-lyzed with factorial MANOVA followed by separated ANOVAs allwith imidacloprid concentration and perceived predation risk asfixed factors. When a significant interaction was observed, aone-way ANOVA testing for the effects of perceived predation risklevels in the different imidacloprid treatments was performed. Allstatistical analysis were performed with the Minitab statisticalsoftware (Minitab, 2000).
The intrinsic rate of population increase (r) was calculated iter-atively using the Euler’s equation (see Eq. SI-S1, Supporting infor-mation, for details).
To address mixture effects, the observed effect on feeding, res-piration, offspring production and intrinsic rate of population in-crease (r) were compared to the expected effects of mixturescalculated from effects of single compound exposures, based onexisting conceptual models: concentration addition (CA) and inde-pendent action (IA) (e.g. Backhaus et al., 2003; Altenburger et al.,2004; Jonker et al., 2004, 2005).
Both conceptual models were fitted to feeding and respirationdata to check the suitability of the different approaches. Althoughthe pharmacological mode of action of predatory chemical cues iscurrently unknown, both imidacloprid and perceived predationrisk impair food acquisition and increase respiration rates of daph-
Imidacloprid concentrations (mg/L)CTR 2.2 4.4 8.8
Feed
ing
rate
(cel
ls x
103 /i
nd/h
our)
0
35
70
105
140No perceived risk
Low perceived risk
High perceived risk
Fig. 1. Effects of imidacloprid and different levels of perceived risk of predation onD. magna feeding rate (mean + SE).
Imidcaloprid concentrations (mg/L)CTR 2.2 4.4
Res
pira
tion
rate
(ug
O2/o
rg/h
our)
0.0
0.2
0.4
0.6
0.8
1.0
No perceived risk
Low perceived risk
High perceived risk
Fig. 2. Effects of imidacloprid and different levels of perceived risk of predation onD. magna oxygen consumption (mean + SE).
J.L.T. Pestana et al. / Chemosphere 78 (2010) 241–248 243
nids. Thus, they share a common ecotoxicological mode of action(sensu Barata and Baird, 2000) (Barata et al., 2007).
For the IA conceptual model the fit to our data was made usingEq. (1) (Ferreira et al., 2008):
Y ¼ lmaxQn
i¼1qiðCiÞ ð1Þ
where Y denotes the biological response, Ci is the concentration ofchemical i in the mixture, qi(Ci) the probability of non-response,lmax the control response for the selected endpoint and
Qthe mul-
tiplication function.For the CA model fit, Eq. (2) (Ferreira et al., 2008) was applied,
where Ci is the dose used for stressor i in the mixture and ECxi isthe effect dose of stressor i that produces the same effect (x%) asthe whole mixture.
Xn
i¼1
Ci=ECi ¼ 1 ð2Þ
The procedure analysis suggested by Jonker et al. (2005) wasfollowed for the analysis of feeding and oxygen consumptionexperiments since it permitted significance testing of model fitfor both the independent action model and the concentration addi-tion model and also because the analysis accounts for differentnonlinear dose–response characteristics of stressors (Jonker et al.,2005). This procedure allows the evaluation of deviations from ref-erence models such as synergism, antagonism as well as dose ratioand dose level deviations. Dose level-dependent deviation meansthat the ‘‘deviation from either reference model at low dose levelsis different from the deviation at high dose levels. For instance,antagonism may be observed at low dose levels and synergism athigh dose levels” (Jonker et al., 2005). Dose ratio dependent devia-tions arises when ‘‘the deviation from either reference model de-pends on the composition of the mixture. In the case of twosubstances, antagonism can be observed where the toxicity ofthe mixture is caused mainly by toxicant 1, whereas synergismcan be observed where the toxicity is caused mainly by toxicant2” (Jonker et al., 2005).
These deviations from the reference models were obtained bythe addition of two parameters (a and b) and are tested within anested framework, see Jonker et al. (2005). For reference modelsand their deviations the fitting process was conducted through aseries of iterations performed in a spreadsheet environment usingthe built-in solver function (using the initial lmax as the average re-sponse in control treatment). The models are fitted to the datausing a maximum-likelihood method. The best fit is chosen usinga v2 test which minimizes the objective function based on thelog likelihood. The biological interpretation of these additionaldeviation parameters are described in Tables SI-S2 (Supportinginformation).
Since imidacloprid and fish chemical cues also caused con-trasting effects in terms of the reproductive parameters (revealinghere differences in terms of their ecotoxicological mode of ac-tion), it was decided to apply the IA conceptual model to analysethe effects of mixtures of both stressors on offspring productionand intrinsic rate of population increase (r). The analysis proce-dure for reproductive endpoints was different because the ap-proach described by Jonker et al. (2005) relies on dose–responsecurves and it is not suitable to estimate joint effects if the differ-ent stressors in the mixture that cause contrasting effects (Jonkeret al., 2005). For these endpoints, estimated mean values wereobtained directly from Eq. (1), using the maximum response(i.e. 0) in the case of neonate production and intrinsic rate ofpopulation increase (r). Effects estimated by the reference IAmodel were compared to the average values observed and theirrespective 95% confidence intervals to infer significant antago-
nism or synergism for each mixture treatment. This approachwas also applied to analyse feeding and respiration data to com-pare results in terms of the sensitivity of the different modellingapproaches used.
3. Results
3.1. Mortality
The imidacloprid 48-h EC50 (95% CI) for D. magna was 96.65 mg/L (95% CI: 87.83–105.60) with no predatory chemical cues and90.68 mg/L (95% CI: 82.04–99.30) when simultaneously exposedto high concentration of predation chemical cues.
3.2. Feeding and oxygen consumption
D. magna showed significant reduction in feeding(F3, 59 = 414.97, p < 0.001) and significantly higher respiration rates(F2, 26 = 141.10, p < 0.001) when exposed to sub-lethal concentra-tions of imidacloprid. Furthermore, D. magna also showed reducedfeeding rates (F2, 59 = 53.41, p < 0.001) and increased oxygen con-sumption (F2, 26 = 54.10, p < 0.001) when exposed to predatorychemical cues (Figs. 1 and 2). No significant interaction betweenimidacloprid and perceived predation risk was detected for feeding(F6, 59 = 1.38, p = 0.243) or for oxygen consumption (F4, 26 = 0.91,p = 0.477) by two-way ANOVA.
244 J.L.T. Pestana et al. / Chemosphere 78 (2010) 241–248
Using additive models for feeding, data showed an excellent fitto the IA reference model (SS = 2446.92; r2 = 0.958; p < 0.001). Asignificant dose level-dependent deviation from the IA modelwas also observed (SS = 2146.09; r2 = 0.963, v2 test, p = 0.02;a = 5.72; b = 1.96), where antagonism was shown for low doses ofboth stressors and a synergistic pattern for high doses of bothstressors. This switching from antagonism to synergism occurs atclose to lethal concentrations. Using the CA approach, there wasalso a good fit to the reference model (SS = 3137.09; r2 = 0.947p < 0.001) and the same dose level-dependent deviation was ob-served (SS = 2303.50; r2 = 0.961, v2 test, p < 0.001; a = 3.40;b = 2.05) with the switch from antagonism to synergism was alsonear the lethal threshold. The pattern of dose level-dependent
-1
CTR 2.2 4.4
Num
ber o
f bro
ods
0
3
4
5
6
CTR 2.2 4.4
neno
nate
s/fe
mal
e
0
20
40
60
80
100
120
Imidacloprid concen
No perceived risk
CTR 2.2 4.4
r (da
ys-1
)
0.00
0.20
0.30
0.35
0.40
0.45
0.25
Low perc
A
C
E
Fig. 3. Effects of imidacloprid and different levels of perceived risk of predation on D. mnumber of neonates produced per female; B – Size at maturity; C – Number of broods; Dgrowth rate.
deviations from additivity was observed, but was not significant,in the comparison between the observed data and the mean esti-mated responses calculated directly using the IA reference modelfor each mixture treatment (Fig. 4A). Analysis of variance alsofailed to reveal a significant interaction between stressors, againindicating additivity.
Respiration data adequately fitted the IA reference model(SS = 0.8719; r2 = 0.931; p < 0.001). On the other hand, and usingthe CA approach, a good fit to the reference model was ob-served (SS = 1.19855; p < 0.001) but also with a significant devi-ation for synergism (SS = 0.7891; r2 = 0.937, v2 test, p < 0.001;a = �5.43). A synergistic deviation from additivity was also ob-served in the comparison between observations and the mean
CTR 2.2 4.4
Som
atic
gro
wth
rate
(day
s )
0.000
0.045
0.050
0.055
0.060
0.065
0.070
CTR 2.2 4.4
AFR
(day
s)
0
7
8
9
10
11
trations (mg/L)
CTR 2.2 4.4
Bod
y le
ngth
(mm
)
0.0
1.6
2.0
2.4
2.8
3.2
eived risk High perceived risk
B
D
F
agna life-history parameters and somatic growth over 21 d (mean + SE): A – Total– Age at first reproduction; E – Intrinsic rate of natural increase, r; and F – Somatic
A
Feed
ing
rate
(cel
lsx1
03 / in
d/ho
ur)
Low risk
2.2 m
g IMI /L
0
20
40
60
80
100
High risk
2.2 m
g IMI /L
High risk
4.4 m
g IMI /L
Low risk
4.4 m
g IMI /L
Low risk
8.8 m
g IMI /L
High risk
8.8 m
g IMI /L
B
0.0
0.2
0.4
0.6
0.8
1.0
Res
pira
tion
rate
(µgO
2/ or
g/ h
our)
Low risk
2.2 m
g IMI /L
High risk
2.2 m
g IMI /L
High risk
4.4 m
g IMI /L
Low risk
4.4 m
g IMI /L
D
Intr
insi
cra
te o
fpop
ulat
ion
incr
ease
(r)
0.00
0.20
0.25
0.30
0.35
0.40C
0
20
40
60
80
Low risk
2.2 m
g IMI /L
High risk
2.2 m
g IMI /L
High risk
4.4 m
g IMI /L
Low risk
4.4 m
g IMI /L
Low risk
2.2 m
g IMI /L
High risk
2.2 m
g IMI /L
High risk
4.4 m
g IMI /L
Low risk
4.4 m
g IMI /L
Offs
prin
g
Fig. 4. Effects of combined exposures to imidacloprid different levels of perceived risk of predation on D. magna feeding (A), respiration rate (B), offspring (C) and intrinsic rateof population increase, r (D). Empty symbols denote observed responses (mean and 95% confidence intervals) and filled symbols represent effects predicted by independentaction reference model.
J.L.T. Pestana et al. / Chemosphere 78 (2010) 241–248 245
estimated responses calculated directly using the IA referencemodel for each mixture treatment (Fig. 4B) at low doses of bothstressors.
3.3. Life history
Over 21 d, mortality did not exceed 10% except in treatmentswith 8.8 mg/L imidacloprid, where mortality of 40–50% was ob-served. Furthermore there was an extremely low neonate produc-tion in the 8.8 mg/L treatment which biased the Jackniffe
Table 1Summary of results from MANOVA (A) and univariate ANOVAs (B) testing the effects of im
procedure and confounded calculation of r. For this reason, andbecause we wished to focus on sub-lethal effects, the 8.8 mg/Ltreatment was excluded from the statistical analysis of life-historyparameters.
The concentrations of imidacloprid at the end of day 2 were 65–71% of initial nominal concentrations (1.56, 3.03 and 5.67 mg/L atthe lowest medium and high imidacloprid concentrations respec-tively) which are in agreement with the low persistence of thispesticide in water. Similar concentrations of imidacloprid were ob-served in treatments with high levels of perceived predation risk
idacloprid and perceived predation risk on D. magna life history.
246 J.L.T. Pestana et al. / Chemosphere 78 (2010) 241–248
(2.99 and 5.72 mg/L for the medium and high imidacloprid concen-trations respectively) revealing no effect of predatory chemicalcues on imidacloprid stability.
D. magna exposed to predatory chemical cues showed increasedfecundity, reduction in size at maturity and in age at first reproduc-tion, and an increase in the intrinsic rate of population increase (r)relatively to the control treatment (Fig. 3A–E, Table 1).
Exposure to imidacloprid resulted in significant, contrasting ef-fects with daphnids exposed to sub-lethal concentration of thepesticide showing delayed maturation, production of less neonatesand thus significant reductions in intrinsic rate of natural increase,r, compared to control daphnids (Fig. 3A, D, and E, Table 1). A sig-nificant reduction in size at first reproduction and number ofbroods were observed for daphnids exposed to higher concentra-tions of imidacloprid (Fig. 3B and C, Table 1).
There was a significant interaction between imidacloprid andperceived predation risk on the life-history parameters of D. magna(Fig. 3A–E, Table 1). Exposure to imidacloprid inhibited D. magnaresponses to predatory chemical cues and this inhibition of induc-ible life-history responses was more pronounced at higher concen-trations of imidacloprid. Thus, there was a significant interaction interms of age at first reproduction (one-way ANOVA; perceived pre-dation risk in C0: p < 0.001, C1: p = 0.056, C2: p = 0.501; Fig. 3D),neonate production (one-way ANOVA; perceived predation riskin C0: p < 0.001, C1: p < 0.001, C2: p = 0.009, Fig. 3A), size at matu-rity (one-way ANOVA; perceived predation risk in C0: p < 0.001,C1: p = 0.042, C2: p = 0.661, Fig. 3B) and on the intrinsic rate of nat-ural increase (r) (one-way ANOVA; perceived predation risk in C0:p < 0.001, C1: p = 0.006, C2: p = 0.227; Fig. 3E).
Daphnids exposed to imidacloprid were also smaller than con-trols after 21 d (Fig. 3F, Table 1) but there was no significant effectof perceived predation risk on the growth of daphnids after 21 dnor was significant interaction between imidacloprid and per-ceived predation risk observed (Fig. 3F, Table 1).
Thus, in the absence of imidacloprid, D. magna under low andhigh perceived predation risk matured more rapidly and producedmore neonates than daphnids not exposed to predatory chemicalcues. However, daphnid populations exposed to imidacloprid un-der both low and high perceived predation risk levels showed sim-ilar r-values to daphnid populations not exposed to predation risk.By comparing experimental data with predicted values calculatedusing the IA reference model (Fig. 4B and C), it is shown that inhi-bition of life-history induced responses leads to synergistic effectsin the high perceived predation risk treatments. This is due to con-trasting effects of both stressors with imidacloprid inhibitingfecundity and increasing maturation time of daphnids under highperceived predation risk and thus greatly reducing the intrinsicrate of population increase.
4. Discussion
Imidacloprid is a relatively new pest control substance, whichdespite favourable characteristics of low persistence and low tox-icity towards vertebrates (Tomizawa and Casida, 2005) is generat-ing increasing concern regarding its potential impacts on naturalecosystems (Matsuda et al., 2001; Jemec et al., 2007; Pestanaet al., 2009b). D. magna has been shown to be less sensitive thanother freshwater crustaceans and insects (CCME, 2007). Here,EC50 values of daphnids exposed to imidacloprid and predatorychemical cues were not significantly different compared to daph-nids exposed to the pesticide alone revealing that for short termexposures perceived predation risk does not appear to increasethe sensitivity of D. magna to imidacloprid.
In contrast, short term exposure to sub-lethal concentrations ofimidacloprid significantly reduced feeding and increased metabolic
cost in D. magna. Altered activity of daphnids treated withimidacloprid may affect feeding due to abnormal swimmingbehaviour, while also increasing metabolic costs due to extra activ-ity and/or enhanced detoxification processes.
Exposure to increasing concentrations of predatory chemicalcues also led to reduced feeding and increased oxygen consump-tion. This is in concordance with studies of behavioural and phys-iological responses of daphnids to fish infochemicals in whichchanges in swimming behaviour, avoidance and alertness (Pija-nowska and Kowalczewski, 1997) while reducing the probabilityof encounters with fish predators (Lass and Spaak, 2003) can alsoreduce feeding and place higher metabolic demands on daphnidsreflecting the costs of vigilance (Beckerman et al., 2007).
Combined exposure to imidacloprid and predatory chemicalcues revealed that pesticide contamination increases the cost ofbehavioural antipredator defences of D. magna by further reducingenergy intake and increasing energy expenditure of daphnids un-der fish predation pressure. Assessment of joint effects of thesestressors on feeding and respiration showed significant deviationsfrom additivity reference models contrasting with the results ofanalysis of variance which showed no significant interaction be-tween the two stressors. The fact that models based on dose–re-sponse curves detected deviation patterns from additivity thatwere not limited to just synergism or antagonism (such as dose le-vel-dependent deviation for feeding) indicates the importance ofassessing responses at multiple concentrations of each stressorand of the mixture to improve model calibration.
Both reference models, IA and CA performed adequately interms of prediction of feeding and respiratory responses under acombination of imidacloprid and chemical cues simulating preda-tion risk. Further studies are necessary, but our results suggest thatthese approaches can be applied with respect to ecotoxicologicalmodes of action (Barata and Baird, 2000; Barata et al., 2007).
Long-term exposure of D. magna to imidacloprid resulted in sig-nificant reductions in growth and fecundity with delayed matura-tion. Consequently there was a significant reduction in r, theintrinsic rate of natural increase.
D. magna life-history responses to trout kairomones and alarmcues from conspecifics observed here were consistent with previ-ous studies of phenotypic plasticity of daphnids in response to fishperceived predation risk. The reduction in age and size at maturityand increased fecundity observed here show that daphnids underperceived risk of predation increase fitness through adaptivemechanisms that allow them to increase the probability of repro-ducing before being eaten by visual, size-selective predators suchas fish (Riessen, 1999; Tollrian and Harvell, 1999; Stibor and Nava-rra, 2000; Lass and Spaak, 2003).
Exposure of daphnids to both stressors showed that the effectsof predatory chemical cues were stronger in the absence of pesti-cide, showing that daphnids exposed to sub-lethal concentrationsof imidacloprid are unable to recognize, or respond to predatorychemical cues. Thus, these two stressors caused contrasting effectsin D. magna revealing that imidacloprid can inhibit the life-historyresponses induced by the risk of fish predation in D. magna.
The fact that the model approach suggested by Jonker et al.(2005) is based on dose–response curves was a limitation on ourstudy due to the ‘‘stimulation” effects of perceived predation riskin terms of daphnids reproductive output. It was therefore consid-ered inappropriate for assessment of joint effects of these stressorson reproductive endpoints.
Results demonstrated also that imidacloprid and predatorychemical cues yielded contrasting reproductive responses, but con-sistent feeding or respiration responses. This is concordance withother studies where reproductive responses to predation are con-sidered a physiological, inducible defence per se and not only areflection of energy intake and expenditure (Beckerman et al.,
J.L.T. Pestana et al. / Chemosphere 78 (2010) 241–248 247
2007). This means that although feeding and respiration may besensitive individual responses which are easily measured in thelaboratory, they may not always translate into population-level ef-fects in the field where organisms are exposed to complex mix-tures of contaminants and natural stressors like predation.Moreover and keeping in mind the concept of ecotoxicologicalmode of action, the combined effects of neurotoxic pesticidesand predatory chemical cues on reproductive endpoints are likelybest assessed by using a model of independent action.
Significant interactions detected by analysis of variance showedthat imidacloprid and predatory chemical cues can have more-than-additive effects for many life-history parameters and thecomparison between r mean values calculated based on IA ap-proach with the experimental data clearly confirms this synergismfor high perceived predation risk treatments. Therefore we canconclude that daphnid populations are at elevated risk from pesti-cides such as imidacloprid in the presence of predators.
Furthermore, an inhibition of phenotypic responses to preda-tion caused by pesticide exposure can have important conse-quences in terms of predation-related mortality when predatorsare more tolerant to the pesticide than their prey. The inhibitionof induced responses can in fact alter the survival rates of daphnidsunder fish predation. Taylor and Gabriel’s survival rate-functionmodels the effects of a prey’s threshold length on a predator’s abil-ity to capture prey (Taylor and Gabriel, 1992). For D. magna, be-cause fish are positively size-selective predators, the survival rateover an instar decreases in length under fish predation (Taylorand Gabriel, 1992; Barata et al., 2001). According to Taylor andGabriel ‘‘the threshold model survival rates for Daphnia shift fromhigh to low over a narrow range with 75% of the change occurringwith a 0.5 mm change in Daphnia length” (Taylor and Gabriel,1992). Applying this reasoning here, we would predict that in thefirst adult instar, daphnids exposed to high perceived predationrisk have elevated survival (up by 21.5%) since they were ca.0.14 mm smaller than daphnids not exposed to predatory chemicalcues. This increased survival arising from an induced life-historyresponse was itself greatly reduced in the treatments where daph-nids were also exposed to imidacloprid to 10% and less than 5% inthe lower and medium imidacloprid concentrations respectively,since the difference in length between predation risk-exposedand -unexposed daphnids was drastically diminished. Although itis difficult to estimate the benefits in terms of survival and fitnessof the prey’s induced life-history tactics in realistic scenarios due toall variables involved (food levels, presence of shelter, predatorspecies, competition, etc.), the effectiveness of life-history tacticsto perceived predation risk can be clearly observed, as can the det-rimental impacts that pesticides can have on these inducible life-history responses to predation.
Pesticides acting on the nervous system of organisms can inter-fere with predator–prey interactions, with additive or even syner-gistic effects (Dodson and Hanazato, 1995; Rose et al., 2001;Campero et al., 2007; Pestana et al., 2009a). Exposure to neurotoxicpesticides like imidacloprid, can induce the stimulation of cholin-ergic nerves that also innervate kairomone receptors or nerves thatinnervate endocrine glands responsible for releasing hormonesthat are responsible for the induced responses to perceived preda-tion risk (Barry, 1998). Thus, exposure to pesticides could disruptthe normal physiological and/or neuro-endocrine pathway of anti-predator responses (Hanazato, 1999; Barry, 2002). Nevertheless,imidacloprid concentrations tested here showed generalised toxic-ity with reductions of feeding and increased oxygen consumption,indicating that effects of pesticide on D. magna life-history re-sponses to the risk of predation can be explained based on theenergetic budget of organisms. The effects of imidacloprid werestronger under perceived predation risk because daphnids underfish predation risk allocate a greater proportion of their energy to
reproduction (Stibor and Navarra, 2000), while also reducingfeeding rates and experiencing higher metabolic costs (Beckermanet al., 2007), thus having less available energy for detoxificationand repair processes. Due to these energy constraints, pesticideexposure (which also reduces energy intake and elevates metaboliccosts) appears to increase the costs of Daphnia phenotypic plastic-ity, reducing or inhibiting its expression (Barry, 2000; Hanazato,2001).
5. Conclusions
Sub-lethal concentrations of pesticides can have significant det-rimental consequences on zooplankton populations not onlythrough their direct toxic effects but also through direct and indi-rect effects on predator–prey interactions. If we consider thatbehavioural, physiological, morphological and life-history plastic-ity can stabilize populations and community structure (Verschooret al., 2004; Miner et al., 2005), then inductions of maladaptive re-sponses in prey populations due to pesticide exposure can increasethe risk of local extinction in exposed populations.
We have also demonstrated that a unifying approach, based onphysiological mechanisms can be used to study the effects of dif-ferent stressors such as perceived predation and pesticides, aloneand in combination, using standard ecotoxicological methodolo-gies. In this way, we can further enhance the ecological relevanceof ecological risk assessments through the concept of ecotoxicolog-ical modes of action of stressors (Barata and Baird, 2000; Barataet al., 2007).
Acknowledgements
Financial support for this work was provided by PortugueseFoundation for Science and Technology through a PhD grant toJoão Pestana (SFRH/BD/9005/2002), and also through Canadian Na-tional Science and Engineering Research Council Discovery Grant312076-05 to Donald J. Baird. The study was also partly supportedby the EU Integrated project NoMiracle (Novel Methods for Inte-grated Risk assessment of Cumulative Stressors in Europe; http://nomiracle.jrc.it) Contract No. 003956 under the theme under theEU-theme ‘‘Global Change and Ecosystems” topic ‘‘Developmentof risk assessment methodologies”, coordinated by Dr. Hans Løkkeat NERI, DK-8600 Silkeborg, Denmark.
Appendix A. Supplementary material
Supplementary material associated with this article can befound, in the online version, at doi:10.1016/j.chemosphere.2009.10.066.
References
Altenburger, R., Walter, H., Grote, M., 2004. What contributes to the combined effectof a complex mixture? Environ. Sci. Technol. 38, 6353–6362.
Backhaus, T., Altenburger, R., Arrhenius, A., Blanck, H., Faust, M., Finizio, A.,Gramatica, P., Grote, M., Junghans, M., Meyer, W., Pavan, M., Porsbring, T.,Scholze, M., Todeschini, R., Vighi, M., Walter, H., Grimme, L.H., 2003. The BEAM-project: prediction and assessment of mixture toxicities in the aquaticenvironment. Cont. Shelf Res. 23, 1757–1769.
Baird, D.J., Barber, I., Bradley, M., Soares, A., Calow, P., 1991. A comparative-study ofgenotype sensitivity to acute toxic stress using clones of Daphnia magna Straus.Ecotoxicol. Environ. Safety 21, 257–265.
Barata, C., Baird, D.J., 2000. Determining the ecotoxicological mode of action ofchemicals from measurements made on individuals: results from instar-basedtests with Daphnia magna Straus. Aquat. Toxicol. 48, 195–209.
Barata, C., Baird, D.J., Soares, A., 2001. Phenotypic plasticity in Daphnia magnaStraus: variable maturation instar as an adaptive response to predationpressure. Oecologia 129, 220–227.
Barata, C., Baird, D.J., Nogueira, A.J.A., Agra, A.R., Soares, A., 2007. Life-historyresponses of Daphnia magna Straus to binary mixtures of toxic substances:
248 J.L.T. Pestana et al. / Chemosphere 78 (2010) 241–248
Pharmacological versus ecotoxicological modes of action. Aquat. Toxicol. 84,439–449.
Barry, M.J., 1998. Endosulfan-enhanced crest induction in Daphnia longicephala:evidence for cholinergic innervation of kairomone receptors. J. Plankton Res. 20,1219–1231.
Barry, M.J., 2000. Effects of endosulfan on chaoborus-induced life-history shifts andmorphological defenses in Daphnia pulex. J. Plankton Res. 22, 1705–1718.
Barry, M.J., 2002. Progress toward understanding the neurophysiological basis ofpredator-induced morphology in Daphnia pulex. Physiol. Biochem. Zool. 75,179–186.
Beckerman, A.P., Wieski, K., Baird, D.J., 2007. Behavioural versus physiologicalmediation of life history under predation risk. Oecologia 152, 335–343.
Boersma, M., Spaak, P., De Meester, L., 1998. Predator-mediated plasticity inmorphology, life history, and behavior of Daphnia: the uncoupling of responses.Am. Nat. 152, 237–248.
Campero, M., Slos, S., Ollevier, F., Stoks, R., 2007. Sublethal pesticide concentrationsand predation jointly shape life history: behavioral and physiologicalmechanisms. Ecol. Appl. 17, 2111–2122.
CCME, 2007. Canadian Water Quality Guidelines: Imidacloprid. ScientificSupporting Document. Canadian Council of Ministers of the Environment,Winnipeg.
Dodson, S.I., Hanazato, T., 1995. Commentary on effects of anthropogenic andnatural organic-chemicals on development, swimming behavior, andreproduction of Daphnia, a key member of aquatic ecosystems. Environ.Health Perspect. 103, 7–11.
Dodson, S.I., Hanazato, T., Gorski, P.R., 1995. Behavioral-responses of Daphnia-pulexexposed to carbaryl and chaoborus kairomone. Environ. Toxicol. Chem. 14, 43–50.
Eggen, R.I.L., Behra, R., Burkhardt-Holm, P., Escher, B.I., Schweigert, N., 2004.Challenges in ecotoxicology. Environ. Sci. Technol. 38, 58A–64A.
Ferreira, A.L.G., Loureiro, S., Soares, A.M.V.M., 2008. Toxicity prediction of binarycombinations of cadmium, carbendazim and low dissolved oxygen on Daphniamagna. Aquat. Toxicol. 89, 28–39.
Hanazato, T., 1999. Anthropogenic chemicals (insecticides) disturb natural organicchemical communication in the plankton community. Environ. Pollut. 105,137–142.
Hanazato, T., 2001. Pesticide effects on freshwater zooplankton: an ecologicalperspective. Environ. Pollut. 112, 1–10.
Hanazato, T., Dodson, S.I., 1992. Complex effects of a kairomone of chaoborus andan insecticide on Daphnia-pulex. J. Plankton Res. 14, 1743–1755.
Hunter, K., Pyle, G., 2004. Morphological responses of Daphnia pulex to chaoborusamericanus kairomone in the presence and absence of metals. Environ. Toxicol.Chem. 23, 1311–1316.
Jemec, A., Tisler, T., Drobne, D., Sepcic, K., Fournier, D., Trebse, P., 2007. Comparativetoxicity of imidacloprid, of its commercial liquid formulation and of diazinon toa non-target arthropod, the microcrustacean Daphnia magna. Chemosphere 68,1408–1418.
Jonker, M.J., Piskiewick, A.M., Castellà, N.I., Kammenga, J.E., 2004. Toxicity of binarymixtures of cadmium–copper and carbendazim–copper to the nematodeCaenorhabditis elegans. Environ. Toxicol. Chem. 23, 1529–1537.
Jonker, M.J., Svendsen, C., Bedaux, J.J.M., Bongers, M., Kammenga, J.E., 2005.Significance testing of synergistic/antagonistic, dose level-dependent, or doseratio-dependent effects in mixture dose–response analysis. Environ. Toxicol.Chem. 24, 2701–2713.
Kieu, N.D., Michels, E., De Meester, L., 2001. Phototactic behavior of Daphnia and thecontinuous monitoring of water quality: interference of fish kairomones andfood quality. Environ. Toxicol. Chem. 20, 1098–1103.
Minitab, I., 2000. Minitab Statistical Software 13.0.Moffat, A.S., 1993. Agriculture – new chemicals seek to outwit insect pests. Science
261, 550–551.Naylor, C., Maltby, L., Calow, P., 1989. Scope for growth in Gammarus-pulex, a fresh-
water benthic detritivore. Hydrobiologia 188, 517–523.OECD, 1998. Daphnia magna reproduction test. OECD – Organization for Economic
Cooperation and Development, Paris.OECD, 2000. Daphnia sp., acute immobilisation test. OECD – Organization for
Economic Cooperation and Development, Paris.Pauwels, K., Stoks, R., De Meester, L., 2005. Coping with predator stress: interclonal
differences in induction of heat-shock proteins in the water flea Daphnia magna.J. Evol. Biol. 18, 867–872.
Penczak, T., Formigo, N., 2000. Fish stocks in the Ancora River, north Portugal Foodconsumption estimated with a bioenergetics model. Fish. Res. 49, 117–128.
Pestana, J.L.T., 2008. Predation risk versus pesticide exposure: effects of fear andloathing on aquatic invertebrates. PhD thesis. University of Aveiro. Aveiro,Portugal, 167p. Available from: <http://biblioteca.sinbad.ua.pt/teses/2009001249>.
Pestana, J.L.T., Loureiro, S., Baird, D.J., Soares, A.M.V.M., 2009a. Fear and loathing inthe benthos: responses of aquatic insect larvae to the pesticide imidacloprid inthe presence of chemical signals of predation risk. Aquat. Toxicol. 93, 138–149.
Pestana, J.L.T., Alexander, A.C., Culp, J.M., Baird, D.J., Cessna, A.J., Soares, A.M.V.M.,2009b. Structural and functional responses of benthic invertebrates toimidacloprid in outdoor stream mesocosms. Environ. Pollut. 157, 2328–2334.
Pijanowska, J., 1997. Alarm signals in Daphnia? Oecologia 112, 12–16.Pijanowska, J., Kloc, M., 2004. Daphnia response to predation threat involves heat-
shock proteins and the actin and tubulin cytoskeleton. Genesis 38, 81–86.Pijanowska, J., Kowalczewski, A., 1997. Predators can induce swarming behaviour
and locomotory responses in Daphnia. Freshwater Biol. 37, 649–656.Preisser, E.L., Bolnick, D.I., Benard, M.F., 2005. Scared to death? The effects of
intimidation and consumption in predator–prey interactions. Ecology 86, 501–509.
Relyea, R.A., 2005. The impact of insecticides and herbicides on the biodiversity andproductivity of aquatic communities. Ecol. Appl. 15, 618–627.
Relyea, R., Hoverman, J., 2006. Assessing the ecology in ecotoxicology: a review andsynthesis in freshwater systems. Ecol. Lett. 9, 1157–1171.
Riessen, H.P., 1999. Predator-induced life history shifts in Daphnia: a synthesis ofstudies using meta-analysis. Can. J. Fish. Aquat. Sci. 56, 2487–2494.
Rose, R.M., Warne, M.S.J., Lim, R.P., 2001. The presence of chemicals exuded by fishaffects the life-history response of Ceriodaphnia cf dubia to chemicals withdifferent mechanisms of action. Environ. Toxicol. Chem. 20, 2892–2898.
Rose, R.M., Warne, M.S., Lim, R.P., 2003. Exposure to chemicals exuded by fishreduces the filtration and ingestion rates of Ceriodaphnia cf. Dubia.Hydrobiologia 501, 215–217.
Stibor, H., Navarra, D.M., 2000. Constraints on the plasticity of Daphnia magnainfluenced by fish-kairomones. Funct. Ecol. 14, 455–459.
Taylor, B.E., Gabriel, W., 1992. To grow or not to grow – optimal resource-allocationfor Daphnia. Am. Nat. 139, 248–266.
Tollrian, R., Harvell, C.D., 1999. The Ecology and Evolution of Inducible Defenses.Princeton Univ. Press, Princeton.
Tomizawa, M., Casida, J.E., 2003. Selective toxicity of neonicotinoids attributable tospecificity of insect and mammalian nicotinic receptors. Annu. Rev. Entomol.48, 339–364.
Verschoor, A.M., Vos, M., van der Stap, I., 2004. Inducible defences prevent strongpopulation fluctuations in bi- and tritrophic food chains. Ecol. Lett. 7, 1143–1148.
