Download - Toxicokinetic Model Describing Bioconcentration and Biotransformation of Diazinon in Daphnia magna

Published: May 11, 2011

r 2011 American Chemical Society 4995 dx.doi.org/10.1021/es104324v | Environ. Sci. Technol. 2011, 45, 4995–5002

ARTICLE

pubs.acs.org/est

Toxicokinetic Model Describing Bioconcentration andBiotransformation of Diazinon in Daphnia magnaAndreas Kretschmann,†,‡ Roman Ashauer,† Thomas G. Preuss,§ Piet Spaak,†,|| Beate I. Escher,†,^ andJuliane Hollender*,†,‡

†Eawag, Swiss Federal Institute of Aquatic Science and Technology, 8600 D€ubendorf, Switzerland‡Institute of Biogeochemistry and Pollutant Dynamics and )Institute of Integrative Biology, ETH Z€urich, CH-8092, Z€urich, Switzerland^The University of Queensland, National Research Centre for Environmental Toxicology (Entox), 39 Kessels Road,Brisbane, Qld, 4108 Australia

§Institute for Environmental Research (Biology V), RWTH Aachen University, 52062 Aachen, Germany

bS Supporting Information

’ INTRODUCTION

Toxicity of a compound toward an organism can in general bedescribed as a function of exposure concentration, exposureduration, and toxicokinetic (TK) and toxicodynamic (TD)processes.1 TK processes describe the time course of the pathwayof a toxicant from the surrounding environment to the target siteinside the organism and consist of uptake, partitioning to targetand nontarget sites, excretion, and biotransformation. TD pro-cesses link the internal concentration of a toxicant with the toxiceffect observed at the organism level by describing the time-dependent build-up and recovery of injury at the molecular,organ, and organism level.1�3 Only the fraction of a pollutantthat reaches the target site is toxicologically active.3

The toxic potency of a chemical is commonly assessed viaexternal effect concentrations, for example, its LC50 value. Whatreflects the toxic potential in a more accurate way is the con-centration inside the organism.3 Quantitative structure�activity

relationships (QSARs) have been developed for estimationof the internal concentration of lipophilic chemicals in aquaticorganisms via their octanol�water partitioning coefficient(Kow).

4,5 A drawback of these QSARs is that they are validonly under steady-state conditions and do not account forbiotransformation.

For phosphorothionate insecticides like diazinon, chlorpyr-ifos, and parathion, it is crucial to consider biotransformationprocesses in order to accurately assess their toxicity. Thesecompounds are metabolically activated by cytochrome P450monooxygenases (P450) to their respective oxon analogues in anoxidative desulfuration.6 In contrast to the thioesters, the oxons

Received: December 23, 2010Accepted: April 15, 2011Revised: April 12, 2011

ABSTRACT: A toxicokinetic model for Daphnia magna, which simulatesthe internal concentration of the insecticide diazinon, its detoxificationproduct 2-isopropyl-6-methyl-4-pyrimidinol, and its active metabolitediazoxon, is presented. During in vivo exposure to diazinon with andwithout inhibition of cytochrome P450 by piperonyl butoxide, the parentcompound as well as its metabolites were quantified with high-perfor-mance liquid chromatography�tandemmass spectrometry (LC-MS/MS) inextracts ofD.magna. Rate constants of all relevant toxicokinetic steps wereobtained by modeling the time course of the internal concentrations witha multicomponent first-order kinetics model. When cytochrome P450was inhibited, the kinetic bioconcentration factor (BCF) of diazinonincreased from 17.8 to 51.0 mL 3 gww

�1. This clearly indicates thatdiazinon is biotransformed to a high degree by cytochrome P450 in D. magna. The dominant elimination step of diazinon wasshown to be its oxidative dearylation to pyrimidinol (62% of total elimination) with a corresponding rate constant of 0.16 h�1. Incontrast, oxidative activation to diazoxon with a rate constant of 0.02 h�1 amounted to only 8% of the total elimination. Duringexposure to diazinon, the active metabolite diazoxon could be detected only in very low concentrations (approximately 0.5% of theparent compound), presumably due to a very fast reaction with the target site acetylcholinesterase. During the exposure experiments(no feeding of daphnids), an exponential decline of the lipid content in D. magna with a first-order rate constant of 0.013 h�1 wasobserved. For short exposure times (e24 h), this had only a minor influence on the determined TK parameters. Such a TK modelcontaining detailed biotransformation processes is an important tool for estimation of the toxic potential of chemicals, particularly,when active metabolites are formed inside an organism.

4996 dx.doi.org/10.1021/es104324v |Environ. Sci. Technol. 2011, 45, 4995–5002

Environmental Science & Technology ARTICLE

are very potent inhibitors of acetylcholinesterase (AChE).6 Inaddition to activation, phosphorothionates and their oxons canalso be metabolically detoxified by several other biotransforma-tion reactions. These reactions include oxidative dearylation ofthe thio compound catalyzed by P450,7 hydrolysis of the oxonanalog by A-esterases,8,9 and dealkylation and dearylation of boththe thio compound and the oxon analogue catalyzed by glu-tathione-S-transferase (GST).9,10 Phenol derivatives formedduring hydrolysis can further be conjugated with endogeneoussubstrates like sulfate, glucose, or glucuronide.11 For different aquaticorganisms a multitude of phopshorothionate metabolites could bedetected in vivo,12 and the extent of the different biotransformationprocesses was highly species- and compound-dependent.9,13

Tools that can predict the time course of the internal con-centration of a chemical as function of absorption, distribution,excretion, and especially metabolism are TK models. For orga-nothiophosphates, several TK models were developed for aqua-tic organisms, which describe activation and/or detoxification ofthe parent compound inside the organism on a time-resolvedbasis.14�16 These studies did not further specify TK processes,for example, via the inhibition of enzymes catalyzing certainbiotransformation pathways or via exposure to metabolites itselfwithin in vivo experiments. Additionally, to our knowledge thevariation in lipid content during exposure experiments and itsinfluence on TK processes has not been analyzed yet in such TKstudies.

This study focuses on the development of a process-based TKmodel forDaphnia magna and the phosphorothionate insecticidediazinon, which includes biotransformation processes and ac-counts for the formation of active and inactive metabolites invivo on a time-resolved basis. In order to identify relevant TKprocesses and to determine the respective kinetic parameters, D.magna was exposed to diazinon (with or without inhibtion ofP450) as well as to its metabolites diazoxon and 2-isopropyl-6-methyl-4-pyrimidinol (pyrimidinol) in vivo. After different ex-posure times, internal concentrations of the parent compoundand its metabolites were quantified with LC-MS/MS. Sincediazinon with a log Kow of 3.8117 is likely to accumulate mainlyin the lipid phase of D. magna, the time course of the lipid con-tent during the exposure experiments and its influence on TKparameters was elucidated.

’THEORY: TOXICOKINETIC PHASE OF DIAZINON IND. MAGNA

The following TK processes were incorporated into the TKmodel and are depicted in Figure 1:(A) Unchanged diazinon is taken up by passive diffusion.(B) Diazinon is eliminated via passive diffusion. Other elim-

ination processes like dealkylation/dearylation catalyzedby GST, which are not separately considered, are in-cluded in this pathway.

(C) Diazinon is activated to the oxon analog diazoxon byoxidative desulfuration catalyzed by P450.

(D) Diazinon is detoxified under formation of pyrimidinolvia oxidative dearylation catalyzed by P450.

(E) Diazoxon is eliminated by irreversible binding to thetarget site AChE. Since AChE in D. magna was found tobe very sensitive toward diazoxon,18 this reaction is assumedto be very fast, so that other elimination processes ofdiazoxon can be considered as negligible.

(F) Pyrimidinol itself can be eliminated by passive diffusionor by conjugation, for example, with sulfate or glucose.Both processes are lumped together in the model.

The interaction of diazoxon with the target site AChE(process E) is actually a TD process. Nevertheless, this pathwayhas to be considered within the TK phase since it reduces theinternal concentration of diazoxon. Toxic effects in D. magnaafter exposure to acutely toxic diazinon concentrations wereanalyzed in a separate TD study.18 The formation of pyrimidinolduring AChE inhibition in step E was neglected.

If the organism is regarded as a single homogeneous compart-ment and first-order kinetics are assumed for uptake andelimination of a chemical into and out of the organism, respec-tively, the time course of the internal concentration can bedescribed by a one-compartment first-order kinetics model.4

The time course of diazinon and its metabolites diazoxon andpyrimidinol in D. magna can therefore be described as follows(equations are based on ref 16 and modified to our needs):

diazinon:dcdiazinonint ðtÞ

dt¼ kdiazinonin 3 c

diazinonext ðtÞ � kdiazinonel, tot 3 c

diazinonint ðtÞ

ð1Þ

Figure 1. Incorporated toxicokinetic (TK) processes of diazinon inD. magna into the TKmodel (PBO= piperonyl butoxide as inhibitor of cytochromeP450).



with

kdiazinonel, tot ¼ kdiazinonact þ kdiazinondearyl þ kdiazinonel ð2Þ

diazoxon:dcdiazoxonint ðtÞ

dt¼ kdiazinonact 3 c

diazinonint ðtÞ

� ki 3 cdiazoxonint ðtÞ 3 cAChEint ðtÞ ð3Þ

pyrimidinol:dcpyrint ðtÞ

dt¼ kdiazinondearyl 3 c

diazinonint ðtÞ � kpyrel 3 c

pyrint ðtÞ

ð4Þ

cextdiazinon (M) is the diazinon concentration in the exposure water,cint (mol 3 gww

�1) is the internal concentration of the respec-tive compound. kin

diazinon (L 3 gww�1

3 h�1) is the uptake rate

constant of diazinon. kel,totdiazinon (h�1) is the cumulative rate

constant summarizing all possible elimination processes ofdiazinon as mentioned above. kdearyl

diazinon (h�1) and kactdiazinon

(h�1) describe the dearylation and activation of diazinoncatalyzed by P450, kel

diazinon (h�1) the remaining eliminationprocesses. The term after the minus sign in eq 3 describes theinhibition of AChE (irreversible inhibition in one step;mechanism applies if low internal diazoxon concentrationsare assumed; see ref 18). ki (gww 3mol�1

3 h�1) is the second-

order inhibition rate constant. cintAChE (mol 3 gww

�1) is theconcentration of active AChE in the organism. kel

pyr (h�1) isthe elimination rate constant of pyrimidinol.

If the activation and dearylation process of diazinon isinhibited by a specific P450 inhibitor, such as piperonyl butoxide(PBO), eq 2 reduces to

kdiazinonel, tot, inh ¼ kdiazinonel, tot � ðkdiazinonact þ kdiazinondearyl Þ ¼ kdiazinonel ð5Þ

In this case kindiazinon is termed kin,inh

diazinon.The time required to reach effective steady state (cint =

0.99 3 cintss at tss) and the time to eliminate 99% of the chemical

in the organism (tel) after transfer into toxicant free medium aregiven by eq 6:

tss ¼ tel ¼ ln 100=kel, tot ð6Þ

’MATERIALS AND METHODS

Chemicals and Test Organism. A list of chemicals andprepared stock solutions used in the experiments can be foundin the Supporting Information. Reported diazinon, diazoxon, andpyrimidinol concentrations are measured; PBO concentrationsare nominal. For experiments, D. magna Straus (clone 5) with anage of 5�6 or 7�8 days were used. Daphnids were cultured inartificial M4 medium (according to ref 19) at a temperature of20 ( 2 �C and with a 16 h light and 8 h dark cycle and fed with0.025 mg of particulated organic carbon 3 daphnid

�13 day

�1.Mothers were derived from the culture of the Institute forEnvironmental Research at the RWTH Aachen University(Germany), originally obtained from stock cultures at the Uni-versity of Sheffield (U.K.).In Vivo Exposure Experiments. For in vivo experiments,

D. magna was exposed to diazinon, pyrimidinol, and diazoxon.Analysis of uptake and elimination of the respective compoundduring exposure and after transfer into toxicant-free medium,respectively, was performed in separate experiments (experi-ments I�VII; for details see Table 1). Diazinon experimentswere performed with or without the presence of the P450inhibitor PBO. Exposure concentrations were chosen in orderto have high internal concentrations for analysis but that animalsstill show no effect (immobilization). EC50(48 h) values forimmobilization determined in a preexperiment for 7�10 day olddaphnids were 5.6 nM for diazinon and 10.1 nM for diazoxon.PBO concentrations were chosen according to Ankley et al.20 Anuptake study with diazinon without PBO was performed twice(experiments IV and V). In experiment III (uptake of diazinon inthe presence of PBO), an approximately 14 times higher diazinonconcentration (66.8 nM) was applied compared to the otherexperiments in order to properly assess the relevance of pyrimi-dinol formation in the case of inhibited P450. No effect wasexpected due to inhibited diazoxon formation. In experiment II,daphnids were transferred in diazinon-free medium containingPBO after exposure to diazinon and PBO.During the uptake and elimination experiments, daphnids

were sampled after different exposure times out of exposure ortoxicant-free medium, respectively. Two replicates per treatmentand two for each control (pure medium, acetone, and acetoneþPBO controls) were used. At each sampling time, 20 daphnidswere sampled in duplicate. No immobilization or mortality was

Table 1. Experimental Details of Separately Performed in Vivo Uptake and Elimination Studies with D. magnaa

exposure model fit

expt

age of daphnids,

days agent

concn, nM

(PBO, μM)

duration, h

(PBO, h) eqs used

fitted TK

parameters best-fit results

I 5�6 pyrimidinol, elimination 231.7 16.7 4 kelpyr 0.78 ( 0.01 h�1

II 5�6 diazinon (PBO), elimination 4.8 (1.5) 19.1 1, 5 keldiazinon 0.08 ( 0.02 h�1

III 7�8 diazinon (PBO), uptake 66.8 (3.0) 61.5 (3.3)b 1, 5 kin,inhdiazinon 4.08 ( 0.11 mL 3 gww

�13 h

�1

IV 7�8 diazinon, uptake 5.2 20.8 1, 3, 4 kindiazinon 4.62 ( 0.14 mL 3 gww

�13 h

�1

V 5�6 diazinon, uptake 4.8 20.8 1, 3, 4 kel,totdiazinon 0.26 ( 0.01 h�1

VI 5�6 diazinon, elimination 4.5 18.5 1, 3, 4 kdearyldiazinon; kact

diazinon 0.16 ( 0.02 h�1; 0.02 h�1

VII 7�8 diazoxon, uptake 10.3 17.8 c

a Experiments were performed with diazinon with and without inhibition of cytochrome P450 by piperonyl butoxide (PBO), diazoxon, and pyrimidinol.Best-fit results of the TK model fitted to measured internal concentrations are also shown. Numbers in parentheses indicate PBO concentrations used .Equations 1, 3, and 4 were fitted simultaneously to data from experiments IV, V, and VI. bDaphnids were preexposed to 3.0 μM PBO for 3.3 h. cNomodel fit was available since internal diazoxon concentrations were too low for clear identification and quantification.



observed in the controls. Mortality/immobilization was observedonly in diazinon treatments at the end of experiments III (7.5%of daphnids), IV (5%), and V (5%). At the beginning of theexperiments in daphnids, no eggs and ovaries were visible underthe microscope apart from experiment II, where 75% of thedaphnids possessed ovaries.Measured oxygen content and pH during the experiments

were between 5 and 8 mg 3 L�1 and 7.2 and 8.4, respectively.

Compound concentrations in the medium were measured induplicate at the beginning and the end of the experiments.Compound concentrations remained stable during the exposurephase (loss e5% of initial concentration).Chemical Analysis. Prior to compound analysis, daphnid

samples were homogenized in MeOH using a FastPrep FP120Bio 101 (Savant Instruments, Inc.). Homogenates were filteredand extracts were subsequently diluted with H2O (maximalMeOH content 5%). For details see Supporting Information.Diazinon and its metabolites diazoxon and pyrimidinol in

extracts of D. magna and in the exposure medium were analyzedwith online solid-phase extraction (SPE)�HPLC�ESI-MS/MSusing a TSQ Quantum Ultra mass spectrometer from ThermoFisher Scientific (SPE material, ENVþ from International Sor-bent Technology Ltd.; particle size, 40�70 μm; HPLC column,Atlantis T3 column 3.0 � 150 mm, C18, particle size3 μm, pore size 100 Å from Waters). A linear gradient withMeOH and water (2 mM ammonium acetate) was applied (totalflow 400 μL 3min

�1). The limit of detection (LOD) and the limitof quantification (LOQ) in daphnid matrix determined as asignal-to-noise ratio of 3:1 and 10:1 were 0.3 and 1 for diazinon,0.2 and 1 for diazoxon, and 11 and 48 pg absolute for pyrimidinol,respectively. For quantification of diazinon and diazoxon and ofpyrimidinol diazinon-diethyl-D10 and pyrimidinol-13C4 served asinternal standards, respectively. Data were evaluated with Xcali-bur (Thermo Scientific) in the Qual and Quan Browser. Mea-sured internal concentrations are referred to the wet weight ofthe animals at the beginning of the experiment. Measuredinternal diazinon concentrations used for further evaluationswere allg LOQ, and pyrimidinol and diazoxon data were allgLOD. Diazoxon data were partly below the calibration range.Characterization of D. magna. The time course of length,

wet weight, dry weight, and total lipid content ofD. magna underculturing conditions (as described in Supporting Information)was followed over a period of 3 days. In another experiment thetime course of total lipid and wet weight under exposureconditions without feeding was determined. The age of theanimals at the beginning of each experiment was 5�6 days.Lipids were extracted according to the method developed by

Smedes21 by applying a mixture of cyclohexane, 2-propanol, andwater. In the case of animals fed during the experiment, extractedlipids were determined gravimetrically. For measurement of therelative lipid amount in animals not fed, the sulfophosphovanil-line method developed by Z€ollner and Kirsch22 adapted to 96-well plates was used, as it has a higher sensitivity than thegravimetric method.Data Evaluation and Toxicokinetic Parameter Estimation.

Model equations and TK parameters were fitted to the internalconcentrations of the respective experiments as listed in Table 1.The inhibition rate constant ki was determined in vitro in aprevious TD study18 and amounts to 0.080 gww 3 pmol�1

3 h�1.

The fit of the TK model was performed for internal concen-trations normalized to the wet weight and lipid content of D.magna at the beginning of the experiments as well as to the lipid

content at the respective sampling time point. External concen-trationsmeasured at the beginning and end of the exposure phasewere included into the fit. Compound amounts detected in theelimination medium were e3% of the exposure concentrationand therefore set to 0.For linear and nonlinear regression and statistical tests, the

program GraphPad Prism version 4.03 for Windows (GraphPadSoftware, San Diego, CA; www.graphpad.com) was used.Parameter estimation of the rate constants was performedwith ModelMaker version 4.0 (Cherwell Scientific Ltd.,Oxford, U.K.; www.modelkinetix.com). Differential equationswere solved with the Runge�Kutta method. Best-fit valueswere obtained by least-squares optimization by the Levenberg�Marquardt method.

’RESULTS AND DISCUSSION

Parametrization of the Toxicokinetic Model. Figure 2shows the measured internal concentrations of diazinon, diazox-on, and pyrimidinol during the exposure and subsequent elim-ination phase of (a) pyrimidinol, (b) diazinon in the presence ofPBO, and (c) diazinon in the absence of PBO and the corre-sponding best-fit curves. Experimental details and best-fit valuesare summarized in Table 1. In a first step, kel

pyr and keldiazinon were

determined via fit of eqs 4 and 1 to the internal pyrimidinol anddiazinon concentrations measured during the elimination phaseof experiments I and II (exposure to pyrimidinol and diazinonþPBO, see Figure 2 a,b), respectively (for details of the fitprocedure see Supporting Information). Values for kel

pyr andkeldiazinon were kept fixed in further fit procedures. In a next step,the uptake rate constant with inhibited P450 kin,inh

diazinon wasobtained by the fit of eq 1 to data from experiment III (seeFigure 2 b). In a last step, eqs 1, 3, and 4 were fitted simulta-neously to internal diazinon, diazoxon, and pyrimidinol concen-trations obtained within experiments IV, V, and VI (exposure todiazinon without PBO, see Figure 2 c). Here, TD parameters(fixed) and equations necessary to describe the time course ofdiazoxon and AChE in eq 3 were taken from a previous study.18

The activation rate constant kactdiazinon was calculated by applying

eq 2. Besides measured internal concentrations, diazoxon calcu-lated via the time course of AChE measured in a previousexperiment18 during in vivo exposure to diazinon (cext

diazinon =9.1 nM) was included in the fit (see below).Internal Diazinon Concentration. The time course of the

internal diazinon concentration during uptake and elimination ofdiazinon in the absence of PBO is well described with first-orderkinetics (overall fit: r2 = 0.93). The best-fit values for kin

diazinon andkel,totdiazinon (pathways A and BþCþD, respectively, in Figure 1)amount to 4.62 ( 0.14 mL 3 gww

�13 h

�1 and 0.26 ( 0.01 h�1,respectively. Steady state during exposure and complete elimina-tion after transfer into clean medium, respectively, were reachedafter approximately 18 h (eq 6). The kinetic BCF = kin/kel was17.8 mL 3 gww

�1 in the absence of PBO and 51.0 mL 3 gww�1 in

the presence of the inhibitor. The elimination rate constant ofdiazinon with inhibited P450 is reduced by approximately 70%.This indicates that diazinon is biotransformed to a high degree inD. magna and that reactions catalyzed by P450 are the majorbiotransformation processes. This is consistent with a reducedBCF attributed to metabolic degradation observed for severalorganophosphates in fish.23

When calculated via a QSAR according to ref 5 with a log Kow

of diazinon of 3.81,17 a BCF value of 137.6 mL 3 gww�1 is



Figure 2. Internal concentration of (a) pyrimidinol during the elimination phase after exposure to pyrimidinol (experiment I), (b) diazinon during the exposureand elimination phases of diazinon in the presence of piperonyl butoxide (experiments II and III), and (c) diazinon, diazoxon, and pyrimidinol during theexposure and elimination phase of diazinon (experiments IV, V, and VI). Discrete diazoxon values in the lower right panel were calculated via theacetylcholinesterase activity measured during in vivo exposure to diazinon18 and included in the fit. Solid and dashed lines indicate the best fit of the TKmodel tothe internal concentrations during the exposure and elimination phases, respectively (for equations used, seeTable 1). Arrows indicate theTKpathway accordingto Figure 1.



obtained. In comparison to the experimental BCF, this high valueconfirms the high susceptibility of diazinon to biotransformationsince this relationship is valid only for chemicals that are notmetabolized.4

Internal Pyrimidinol Concentration. During exposure todiazinon, the internal concentration of pyrimidinol reachedapproximately 20% of the parent compound in steady state(determined via the best-fit curves of experiments IV and V).The time course was linearly dependent on diazinon. The first-order rate constant for the formation of pyrimidinol via oxidativedearylation of diazinon kdearyl

diazinon (pathway D in Figure 1) was0.16 ( 0.02 h�1. This dearylation is therefore the majorelimination process of diazinon (62% of total elimination). kel

pyr

(0.78 ( 0.01 h�1) determined after direct exposure to pyrimi-dinol is fast compared to diazinon, presumably due to its betterwater solubility and/or fast biotransformation. A possible reac-tion could be the conjugation with sulfate as it was reported forD. magna exposed to pyrene.24

Similar to our findings, fish exposed to diazinon in vivo alsoexhibited lower internal pyrimidinol than diazinon levels, but theratio of the internal concentrations differed between different fishspecies.15 In contrast to our study, the freshwater shrimpGammaruspulex exhibited internal pyrimidinol concentrations 4 times higherthan diazinon.16

Internal Diazoxon Concentration. The toxic metabolitediazoxon could only be detected in very low concentrations nearthe LOD and LOQ during exposure to diazinon. It amounted toapproximately 0.5% of the internal diazinon concentration after21 h of exposure in experiment IV. In experiments II, III, V andVI, as well as when directly exposed to diazoxon in experimentVII, the internal diazoxon concentrations were in most samplestoo low for clear identification and quantification. According toour TK model, the activation of diazinon to diazoxon (kact

diazinon =0.02 h�1, calculated via eq 2) contributes only 8% to the totalelimination of diazinon. Since the standard deviations of kel,tot

diazinon,keldiazinon, and kdearyl

diazinon, which were used for the calculation ofkactdiazinon, were of the same order of magnitude as kact

diazinon, thisvalue possesses a relatively high uncertainty. However, deter-mined values for kact

diazinon lay within the relatively narrow intervalof 0.01�0.02 h�1 when kel

diazinon was varied in the range of0.06�0.10 h�1. This supports the robustness of the determinedactivation rate constant. In combination with a very reactiveAChE in D. magna (as shown in vitro in previous TDexperiments),18 this may explain the very low diazoxon concen-trations found. As for pyrimidinol, the amount of the toxicmetabolite diazoxon present during in vivo exposure to diazinonseems to be strongly species-dependent: For G. pulex the steady-state concentration was approximately 20% compared todiazinon,16 whereas in fish no diazoxon could be detected.15

In contrast to pyrimidinol, the time course of diazoxon was notlinearly dependent on diazinon: The amount of diazoxonincreased continuously, but this conclusion is subject to highuncertainty due to the very low levels found in the organism.Preexperiments showed that these low levels are unlikely to beformed due to nonphysiological factors during the LC-MS/MSmeasurement or due to contamination. For verification, mea-sured diazoxon data were compared with the internal diazoxonconcentration calculated via the time course of AChE activityduring in vivo exposure to 9.1 nM diazinon (see Figure 2 c;experiments and calculations are described in ref 18). Thecalculated diazoxon concentration exhibits an analog timecourse and is of the same order of magnitude (considering an

approximately 2-fold higher external diazinon concentration as inexperiment IV) and therefore supports our data.The TKmodel is able to explain the data satisfactorily only for

the higher calculated diazoxon concentrations. In our model weassume elimination of diazoxon solely via a fast reaction withAChE. The increase of diazoxon can therefore be explained by adepletion of nonoccupied target sites with ongoing exposure.The misfit, especially for the low measured diazoxon concentra-tions, indicates that in addition to the depletion of free AChE afurther factor might have an influence on the time course ofdiazoxon, which might be the induction of P450. This factor ismore pronounced for low diazoxon concentrations where AChEis saturated slowly. Expression of P450 genes was reported, forexample, forD. pulex after exposure to environmental inducers.25

Since we have no data about the time course of P450 inD. magnaduring our exposure experiments, induction of P450 is notconsidered in our model.Discussion of Model Assumptions. When P450 was inhib-

ited by PBO, the internal amounts of diazoxon and pyrimidinolwere too low for clear identification and quantification (eLOQ)or showed no clear time pattern despite the fact that the externaldiazinon concentration was approximately 14� higher (66.8 nM)during the uptake phase in experiment III. Furthermore, no toxiceffect (immobilization) could be observed even after 60 h ofexposure and even though an LC50 (48 h) value for diazinon ofapproximately 5.6 nM was observed. These findings clearlyindicate the inhibition of the activation step and the inhibitionof pyrimidinol formation and confirm the model assumption thatdiazinon is activated to diazoxon and detoxified under formationof pyrimidinol, both catalyzed by P450 (pathways C and D inFigure 1). Furthermore, hydrolysis of diazinon by A-esterases(not inhibited by PBO) to pyrimidinol was negligible. Thatactivation of diazinon to diazoxon (oxidative desulfuration) anddetoxification to pyrimidinol (oxidative dearylation) are cata-lyzed both by P450, as assumed in our model, was reportedby Fabrizi et al.7 in vitro for cell preparations from rat liver.Diazoxon and pyrimidinol were detected as products of anoxidative diazinon metabolism not only for mammals but alsofor fish.9 Here, monooxygenase activities with respect to diazox-on and pyrimidinol formation were highly species-dependent.9

In general for phosphorothionates, the ratio of oxidative dearyla-tion and desulfuration is dependent on the different P450 iso-zymes involved7,26 and differs also between different compounds.26

Since different isozymes were identified in theDaphnia species,27

the presence of both oxidative desulfuration and oxidativedearylation in D. magna is a reasonable assumption supported(albeit not proven) by experimental evidence.Within our model, the formation of pyrimidinol during

inhibition of AChE by diazoxon was neglected because of thevery low internal diazoxon concentrations. Estimated internalconcentrations of pyrimidinol formed via the reaction of diazox-on with AChE were low compared to the concentrationsmeasured in the in vivo experiments. Furthermore, hydrolysisof diazoxon to pyrimidinol by A-esterases was regarded asnegligible if a very fast reaction with AChE was assumed. Thisdetoxification process indeed might rather play a minor role inD.magna, as is supported by the fact that pyrimidinol was presentonly in very low amounts (<LOQ) and without a clear increasewith time when directly exposed to diazoxon (10 nM, datanot shown). Also, for different fish species in in vitro experiments,only very low or no A-esterase activity at all toward diazinonwas reported.9



Influence of Lipid Content on the Bioaccumulation ofDiazinon. When D. magna was cultured with constant foodsupply, the lipid content increased proportionally to the wet weightand amounted to 1.7( 0.1% of the wet weight. If daphnids werekept under exposure conditions (no feeding), the lipid contentdeclined according to an exponential decay. The best-fit value forthe first-order depletion rate constant was 0.013 h�1 (95%confidence interval 0.012�0.015, r2 = 0.94). After 24 and 48 h,approximately 70% and 50%, respectively, of the initial total lipidwas left in the animals. In contrast to total lipid, the wet weightincreased with time. Details are given in the Supporting Infor-mation.As a moderately hydrophobic chemical, diazinon will accu-

mulate in the lipids of D. magna. Therefore, dilution of internalconcentrations by growth of the organism was not incorporatedinto the TK model. The influence of lipid depletion on TKparameters was estimated by fitting the TK model to internaldiazinon concentrations normalized to the lipid content at thebeginning of the experiment as well as at the respective samplingtime point (in units of nanomoles per gram of lipid) (Figure 3 a).When eq 1 was fitted simultaneously to experiments IV, V, andVI, the rate constants for uptake of diazinon, kin

diazinon, with andwithout correction by the lipid time course were almost identical.The elimination rate constant kel,tot

diazinon was lower by approxi-mately 20% if the decline in lipid content was considered. If thetwo uptake experiments were regarded separately during the fit(IV þ VI and V þ VI, respectively) without correction by thetime course of lipid content, the observed differences betweenthe rate constants were larger compared to the simultaneous fitwith and without correction via the lipid time course (for best-fitresults see Supporting Information). It can therefore be con-cluded that for short exposure times the depletion in lipidcontent has only minor relevance compared to interexperimentalvariability. For this reason, experimental data were not correctedvia the lipid time course. Furthermore, the lipid phase might notbe the primary storage compartment for the less hydrophobicmetabolites, which were of major interest in our study.For longer exposure times, the influence of the lipid content in

D. magna is more significant. Figure 3 b compares the internalconcentrations during exposure to diazinon with inhibited P450(experiment III) with and without correction by the lipid time

course. In the first case, diazinon is continuously increasingand is higher by 75% and 122% after 43 and 62 h, respectively,compared to the noncorrected data. This is in agreement withfindings for fish with varying lipid content.28 Our results suggeststhat for longer exposure times the lipid content inD. magna has amore significant influence on bioaccumulation rate constants ofdiazinon, and it is therefore recommended to account for thevariability in lipid content over the course of the experiment ifDaphnia are fasted for more than 24 h. That a change in lipidcontent might indeed have a significant influence on the tox-icokinetics was, for example, shown for the amphipod Pontopor-eia hoyi: here the elimination rate constant of polycyclic aromatichydrocarbons was inversely proportional to the lipid content.29

Implications for Risk Assessement. TK pathways involvedin the toxicity of phosphorothionate insecticides are manifoldand highly species-dependent. The presented TK model isable to combine the complex TK pattern of phosphorothio-nates and particularly includes detailed information aboutbiotransformation processes. This model enables the time-dependent prediction of active metabolites inside the organ-ism, which triggers the toxic effect. Such an approach is ofgreat importance for the estimation and prediction of thetoxic potency of xenobiotics where biotransformation has astrong influence on the toxic effect. This is a significantadvance to conventional tools where, for example, the bioac-cumulation potential of chemicals is estimated via their Kow

by application of QSARs. These tools cannot account forcomplex TK patterns as is the case for phosphorothionateinsecticides.

’ASSOCIATED CONTENT

bS Supporting Information. Additional text, nine tables, andfive figures with details about culturing of D. magna and experi-mental procedures, including a list of chemicals used as well asdetails on workup and analytical methods. This material is availablefree of charge via the Internet at http://pubs.acs.org.

’AUTHOR INFORMATION

Corresponding Author*Phone: þ41587655493; e-mail: [email protected].

Figure 3. Internal concentrations of diazinon during the exposure and elimination phases of diazinon, (a) without (experiments IV, V, and VI, best fit ofeq 1) and (b) with (experiment III) inhibition of cytochrome P450monooxygenase with piperonyl butoxide. Internal concentrations (2) are referred tothe lipid content at the beginning of the respective experiment. Data (4) are corrected via the time course of lipid content during the experiment (daphnids notfed). Solid lines indicate the best-fit curves of the uptake phase; dashed lines indicate the best-fit curves of the elimination phase of diazinon.



’ACKNOWLEDGMENT

We thank Kristina Hitzfeld for experimental support; KathrinHoffmann, Christine Dambone, and Esther Keller for their helpand advice on handling and culturing D. magna; Merle Richter,Heinz Singer, and Alfred L€uck for their help with the chemicalanalysis; and Kathrin Fenner and Tjalling Jager for inspiringdiscussions.

’REFERENCES

(1) Rozman, K. K.; Doull, J. Dose and time as variables of toxicity.Toxicology 2000, 144, 169–178.(2) Ashauer, R.; Boxall, A. B. A.; Brown, C. D. New ecotoxicological

model to simulate survival of aquatic invertebrates after exposure tofluctuating and sequential pulses of pesticides. Environ. Sci. Technol.2007, 41, 1480–1486.(3) Escher, B. I.; Hermens, J. L. M. Internal exposure: Linking

bioavailability to effects. Environ. Sci. Technol. 2004, 38, 455A–462A.(4) Hawker, D. W.; Connell, D. W. Bioconcentration of lipophilic

compounds by some aquatic organisms. Ecotoxicol. Environ. Saf. 1986,11, 184–197.(5) Geyer, H. J.; Scheunert, I.; Bruggemann, R.; Steinberg, C.; Korte,

F.; Kettrup, A. QSAR for organic-chemical bioconcentration inDaphnia,algae, and mussels. Sci. Total Environ. 1991, 109, 387–394.(6) Forsyth, C. S.; Chambers, J. E. Activation and degradation of the

phosphorothionate insecticides parathion and EPN by rat brain. Bio-chem. Pharmacol. 1989, 38, 1597–1603.(7) Fabrizi, L.; Gemma, S.; Testai, E.; Vittozzi, L. Identification of

the cytochrome P450 isoenzymes involved in the metabolism ofdiazinon in the rat liver. J. Biochem. Mol. Toxicol. 1999, 13, 53–61.(8) Poet, T. S.; Wu, H.; Kousba, A. A.; Timchalk, C. In vitro rat

hepatic and intestinal metabolism of the organophosphate pesticideschlorpyrifos and diazinon. Toxicol. Sci. 2003, 72, 193–200.(9) Keizer, J.; Dagostino, G.; Nagel, R.; Volpe, T.; Gnemi, P.;

Vittozzi, L. Enzymological differences of AChE and diazinon hepaticmetabolism - correlation of in vitro data with the selective toxicity ofdiazinon to fish species. Sci. Total Environ. 1995, 171, 213–220.(10) Fujioka, K.; Casida, J. E. Glutathione S-transferase conjugation

of organophosphorus pesticides yields S-phospho-, S-aryl-, andS-alkylglutathione derivatives. Chem. Res. Toxicol. 2007, 20, 1211–1217.(11) Johnston, J. J.; Corbett, M. D. The uptake and in vivo metabo-

lism of the organophosphate insecticide fenitrothion by the blue crab,Callinectes sapidus. Toxicol. Appl. Pharmacol. 1986, 85, 181–188.(12) Takimoto, Y.; Ohshima, M.; Miyamoto, J. Comparative meta-

bolism of fenitrothion in aquatic organisms. 3. Metabolism in thecrustaceans, Daphnia pulex and Palaemon paucidens. Ecotoxicol. Environ.Saf. 1987, 13, 126–134.(13) Chambers, J. E.; Ma, T.; Scott Boone, J.; Chambers, H. W. Role

of detoxication pathways in acute toxicity levels of phosphorothionateinsecticides in the rat. Life Sci. 1994, 54, 1357–1364.(14) Nuutinen, S.; Landrum, P. F.; Schuler, L. J.; Kukkonen, J. V. K.;

Lydy, M. J. Toxicokinetics of organic contaminants in Hyalella azteca.Arch. Environ. Contam. Toxicol. 2003, 44, 467–475.(15) Keizer, J.; D’Agostino, G.; Vittozzi, L. The importance of

biotransformation in the toxicity of xenobiotics to fish. I. Toxicity andbioaccumulation of diazinon in guppy (Poecilia reticulata) and zebra fish(Brachydanio rerio). Aquat. Toxicol. 1991, 21, 239–254.(16) Ashauer, R.; Hintermeister, A.; Caravatti, I.; Kretschmann, A.;

Escher, B. I. Toxicokinetic and toxicodynamic modeling explains carry-over toxicity from exposure to diazinon by slow organism recovery.Environ. Sci. Technol. 2010, 44, 3963–3971.(17) Bowman, B. T.; Sans, W. W. Determination of octanol-water

partitioning coefficients (K(ow)) of 61 organophosphorus and carba-mate insecticides and their relationship to respective water solubility (S)values. J. Environ. Sci. Health, Part B: Pestic., Food Contam., Agric. Wastes1983, 18, 667–683.

(18) Kretschmann, A.; Ashauer, R.; Hitzfeld, K.; Spaak, P.; Hollen-der, J.; Escher, B. I. Mechanistic toxicodynamic model for receptor-mediated toxicity of diazoxon, the active metabolite of diazinon inDaphnia magna. Environ. Sci. Technol. 2011, in press.

(19) OECD Guideline 202 for testing of chemicals. Daphnia sp.,acute immobilization test, 2004.

(20) Ankley, G. T.; Dierkes, J. R.; Jensen, D. A.; Peterson, G. S.Piperonyl butoxide as a tool in aquatic toxicological research withorganophosphate insecticides. Ecotoxicol. Environ. Saf. 1991, 21, 266–274.

(21) Smedes, F. Determination of total lipid using non-chlorinatedsolvents. Analyst 1999, 124, 1711–1718.

(22) Z€ollner, N.; Kirsch, K.::Uber die quantitative Bestimmung von

Lipoiden (Mikromethode) mittels der vielen nat€urlichen Lipoiden(allen bekannten Plasmalipoiden) gemeinsamen Sulfophosphovanillin-Reaktion. Z. Gesamte Exp. Med. 1962, 135, 545–561.

(23) de Bruijn, J.; Hermens, J. Uptake and elimination kinetics oforganophosphorous pesticides in the guppy (Poecilia reticulata): Corre-lations with the octanol/water partition coefficient. Environ. Toxicol.Chem. 1991, 10, 791–804.

(24) Ikenaka, Y.; Eun, H.; Ishizaka, M.; Miyabara, Y. Metabolism ofpyrene by aquatic crustacean, Daphnia magna. Aquat. Toxicol. 2006,80, 158–165.

(25) David, P.; Dauphin-Villemant, C.; Mesneau, A.; Meyran, J. C.Molecular approach to aquatic environmental bioreporting: Differentialresponse to environmental inducers of cytochrome P450 monooxygen-ase genes in the detritivorous subalpine planktonic Crustacea, Daphniapulex. Mol. Ecol. 2003, 12, 2473–2481.

(26) Mutch, E.;Williams, F.M.Diazinon, chlorpyrifos and parathionare metabolised by multiple cytochromes P450 in human liver. Toxicol-ogy 2006, 224, 22–32.

(27) Baldwin, W. S.; Leblanc, G. A. Identification of multiple steroidhydroxylases in Daphnia magna and their modulation by xenobiotics.Environ. Toxicol. Chem. 1994, 13, 1013–1021.

(28) Van den Heuvel, M. R.; McCarty, L. S.; Lanno, R. P.; Hickie,B. E.; Dixon, D. G. Effect of total body lipid on the toxicity andtoxicokinetics of pentachlorophenol in rainbow trout (Oncorhynchusmykiss). Aquat. Toxicol. 1991, 20, 235–252.

(29) Landrum, P. F. Toxicokinetics of organic xenobiotics in theamphipod, Pontoporeia hoyi: Role of physiological and environmentalvariables. Aquat. Toxicol. 1988, 12, 245–271.