Article pubs.acs.org/est
Phenanthrene Bioaccumulation in the Nematode Caenorhabditis elegans Nicole Spann,*,† Willem Goedkoop,§ and Walter Traunspurger† †
Department of Animal Ecology, Bielefeld University, Bielefeld, Germany Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences, Uppsala, Sweden
§
S Supporting Information *
ABSTRACT: The contribution of food to the bioaccumulation of xenobiotics and hence toxicity is still an ambiguous issue. It is becoming more and more evident that universal statements cannot be made, but that the relative contribution of foodassociated xenobiotics in bioaccumulation depends on species, substance, and environmental conditions. Yet, small-sized benthic or soil animals such as nematodes have largely been disregarded so far. Bioaccumulation of the polycyclic aromatic hydrocarbon phenanthrene in the absence and presence of bacterial food was measured in the nematode Caenorhabditis elegans. Elimination of phenanthrene in the nematodes was biphasic, suggesting that there was a slowly exchanging pool within the nematodes or that biotransformation of phenanthrene took place. Even with food present, dissolved phenanthrene was still the major contributor to bioaccumulated compound in nematode tissues, whereas the diet only contributed about 9%. Toxicokinetic parameters in the treatment without food were different from the ones of the treatment with bacteria, possibly because nematodes depleted their lipid reserves during starvation.
■
INTRODUCTION Within their habitat, animals are exposed to chemical pollutants via multiple routes. Chemicals can be taken up from the surrounding aqueous medium, i.e., water for pelagic organisms or pore water for sediment- or soil-inhabiting ones, and from ingested particles, i.e., food, sediment/soil particles. The importance of both routes is ambiguous in many situations,1−5 and highly dependent on the chemical characteristics of the pollutants (e.g., hydrophobicity). Especially for sedimentand soil-inhabiting invertebrates, particles can be a major contributor to accumulation and toxicity.6,7 The commonly used test organisms for assessing bioaccumulation in sediment and soil environments are macroinvertebrate annelid worms,6,8−10 nonbiting midge larvae (Chironomus spec.),8 and several species of crustaceans (e.g., gammarids, amphipods).8,11,12 However, these habitats also harbor a suite of smaller animals (meiofauna) such as, for example, nematodes and rotifers that frequently occur in high abundances and have important ecosystem functions that could be affected by the exposure to pollutants. Nematodes are an extremely diverse taxon and the dominant group within the meiofauna.13 Their well-studied species Caenorhabditis elegans has been frequently used to quantify the deleterious effects of trace metals and organic pollutants on a wide range of end points including mortality, reproduction, growth, behavior, population growth, gene expression, and metabolism. C. elegans feeds on bacteria by drawing liquid with suspended particles into its pharynx (pharyngeal pumping). In © XXXX American Chemical Society
this process the particles get transported to the intestine whereas most of the liquid then is expelled.14 In the absence of a food stimulus the pharyngeal pumping is very much reduced.15 The species also has a thick, but permeable, cuticle. C. elegans can therefore take up substances via its body surface and through the intestine via particulate food and liquid uptake.1 Several recent articles have addressed the question of the relative importance of aqueous and dietary pathways for metal toxicity by providing the animals exclusively with contaminated medium, with contaminated food in a clean medium, or vice versa, or with both contaminated medium and food.1−3 These studies concurred that bacteria facilitate metal uptake and therefore enhance toxicity. Yet, these studies used toxicity measurements such as reproduction, growth, gene expression, and antioxidant enzymes as assessment tools, while only a single study quantified body burdens at two points in time during the experiment.3 In the chosen test-setups the test substance will always partition into both phases (dissolved and particle-associated).1,2,16 The results of these studies are therefore difficult to interpret because both phases will simultaneously contribute to accumulation and toxicity and the relative contributions cannot be separated clearly by assessing toxicity. Received: September 17, 2014 Revised: January 8, 2015 Accepted: January 9, 2015
A
DOI: 10.1021/es504553t Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
treatment, but also including 5.5 × 108 E. coli/mL. To allow partitioning of phenanthrene to the bacteria in the Aq+D treatment (assuming that E. coli does not metabolize phenanthrene23), age-synchronous adult nematodes were only added 24 h after preparing the test vials with the above exposure solutions (50−250 nematodes per vial; in the case of the Aq+D treatment FUdR was also added at this later point). The test vials with the nematodes were kept in the dark at 20 °C. After 24 h of phenanthrene uptake, the nematodes in the samples allocated for the elimination phase were transferred to fresh medium containing no phenanthrene and methanol (medium either without, Aq treatment, or with bacteria, Aq+D treatment; details in SI). At each sampling occasion, actual exposure concentrations in the medium (dissolved and bacteria-associated) as well as nematode internal phenanthrene concentrations were measured by liquid scintillation counting (LSC). In the Aq treatment, a 0.5 mL medium subsample was used to determine the dissolved phenanthrene concentration (CW in mg/L). In the Aq+D treatment, the LSC measurement of a 0.5 mL medium sample yielded the total phenanthrene concentration (CM, mg/L) in the medium (dissolved and bacteria-bound). Another 1 mL subsample was centrifuged to pellet bacteria, and radioactivity was determined in 0.5 mL of the bacteria-free supernatant (giving the dissolved phenanthrene concentration CW, mg/L). Bacterial biomass in the medium was quantified in a 1 mL subsample from each vial of the Aq+D treatment by spectrophotometry (600 nm) and converted to bacterial wet mass in kgbac per L of medium. The difference between the total and dissolved phenanthrene concentrations together with the bacteria biomass concentration (kgbac/L) was used to calculate the concentration of phenanthrene in bacteria (CB expressed as mg phenanthrene per kgbac; based on wet mass of bacteria). Nematodes were removed from the test vials and left for 25−30 min to clear their guts in clean K-medium,24 before 20 animals were transferred to a new scintillation vial, followed by LSC. Nematode phenanthrene concentrations, Ctot, were expressed in mg phenanthrene per kgnem wet mass of nematodes (nematode biomass determination is described in the SI). Ten samples were below the limit of detection (LOD; six nematode samples at t = 0 h, four bacteria samples from the elimination period of the Aq+D treatment, see the SI), but the calculated concentration values were used nonetheless and were not replaced by arbitrary values. Additional details on sample processing, radioactivity analysis, and detection limits can be found in the SI. Other details that were needed for describing the toxicokinetics in C. elegans were determined in separate experiments. These were the ingestion rate (mass of bacteria consumed per nematode biomass and hour, 0.0142 kgbac/ (kgnem × h), all based on wet mass), and the growth rate (assuming exponential growth) of nematodes in the Aq (−0.0029 1/h) and Aq+D treatment (0.0059 1/h). Details of both experiments can also be found in the SI. Toxicokinetic Modeling. At first, we modeled phenanthrene uptake by the nematodes using a simple onecompartment model. However, especially for the Aq treatment, this model overestimated elimination and the corresponding curve did not fit the measured values (SI). Therefore, we modified the model to comprise a quickly exchanging central and a slowly exchanging peripheral compartment (e.g., storage compartment) in the nematodes (Figure 1).
In this study we therefore apply a toxicokinetic approach, where concentrations in the nematodes are measured at certain points in time and the resulting data combined with kinetic modeling. This is a standard technique in studies on largersized animals (e.g., annelids, fish, mussels),6−8,17,18 and recently also models for smaller-sized crustaceans were put forward.19,20 However, for animals as small as nematodes this technique has not been applied before. As the model substance we chose 14Cradiolabeled-phenanthrene, a polycyclic aromatic hydrocarbon with an intermediate octanol−water partition coefficient (log KOW 4.5221), that will partition into both, dissolved and particulate, phases. We exposed adult C. elegans to either dissolved phenanthrene only or to dissolved and bacteriaassociated phenanthrene, and we measured nematode internal concentrations over a 24 h uptake phase followed by 72 h of elimination. We expected a significant contribution of the bacteria-associated fraction to the phenanthrene uptake flux within the nematodes as well as a higher uptake rate from the dissolved phase in the combined treatment (stimulated pharyngeal pumping due to the presence of bacteria and therefore more liquid intake into the intestine). Both mechanisms will lead to higher internal concentrations in the combined treatment after the 24 h uptake phase. We supplemented the toxicokinetic approach with an experiment on the effects of the dissolved-only and combined exposure to phenanthrene on the reproduction of C. elegans. We assumed that the dissolved-only phenanthrene would result in lower toxicity due to lower accumulation of phenanthrene in the absence of food.
■
MATERIALS AND METHODS Bioaccumulation Experiment. Age-synchronous, infertile young adult C. elegans were grown in K-medium supplemented with Escherichia coli OP50, cholesterol, and 5-fluoro-2′deoxyuridine (FUdR; additional details in Supporting Information, SI).22 Before their use in the experiment, they were left to depurate for 1−2.5 h. Toxicokinetic parameters for phenanthrene in C. elegans were determined in two experiments both consisting of a 24 h uptake phase (nematodes exposed to radiolabeled 14C-phenanthrene; dissolved in methanol, specific activity 1.924 × 109 Bq/mmol, Moravek Biochemicals, Brea, CA), followed by a 72 h elimination phase in phenanthrene-free medium. Subsamples for phenanthrene analysis (medium and nematode samples) were taken after 0, 2, 4, 7, 12, and 24 h during the uptake phase (exact times are in SI), and after 28, 36, 48, 72, and 96 h after the start of the experiment in the elimination phase. The nominal exposure concentration of 14C-phenanthrene used during the 24 h exposure phase was 0.02 mg/L, which is well below the acute toxicity of phenanthrene (SI; mortality of C. elegans after 24 h was always ≤10% in concentrations up to 100 mg/L). A concentration of 0.02 mg/L phenanthrene also showed no negative effects on the reproduction of C. elegans (details in SI). Tests were conducted with the following two treatments (three replicates per sampling occasion): aqueous (Aq) treatment and aqueous plus dietary (Aq+D) treatment. Test vials (glass scintillation vials) for the Aq treatment contained 2 mL of K-medium with 0.12 mM FUdR, and 0.02 mg/L 14Cphenanthrene (phenanthrene was added via a stock solution with methanol, final concentration of methanol in test vial was 0.25%). For the Aq+D treatment, glass vessels contained 4 mL of exposure solution with the same composition as in the Aq B
DOI: 10.1021/es504553t Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
in the whole organism (Ctot) the following is therefore applicable:
C tot = CC + C P
(4)
The contribution of the dietary uptake pathway (via bacteriaassociated phenanthrene) to the total uptake flux can be calculated as percentage originating from food AE × IR × C B = × 100 k u × C W + AE × IR × C B
We used the numerical routines in the software program OpenModel version 2.2.025 to estimate the nine parameters for both treatments simultaneously in one fitting procedure (two separate values with one each for ku, ke, kCP, kPC in the Aq and Aq+D treatments; one value for AE, constrained to be ≤1 in the Aq+D treatment). Concentrations of phenanthrene in water (CW) and bacteria (CB) as the average value for each point in time were linearly interpolated between sampling points. Best-fit parameter estimates were obtained by minimizing the residual sum of squares (RSS) between measured nematode phenanthrene concentrations and predicted total nematode concentrations Ctot using the Marquardt algorithm (without weighting) in OpenModel. Subsequently, we tested if the model could be simplified by sharing parameters between treatments (e.g., using the same value for ku for both treatments). In the first round of model simplification, four new models were set up (each one having only eight parameters to be estimated), where in each new model one parameter (ku, ke, kCP, or kPC) was shared between treatments, and new parameter estimates were generated with OpenModel. If more than one reduced model provided an adequate fit to the data (no significant increase in the RSS compared to the RSS of the original nine-parameter model; Ftest,26 p > 0.05), the model with the lowest RSS was chosen as the new best-fitting model. This new eight-parameter model was adopted as the reference model against which further simplified models (having only seven parameters) were tested in a second round of model simplification. This simplification procedure was repeated until the simplest model, which still explained the data well, was found. Details of the derivation of initial starting values for the Marquardt method and the settings for the software program as well as the data sets used can be found in the SI. Bioconcentration factors (BCF, L/kgnem), bioaccumulation factors (BAF, L/kgnem, Aq+D treatment), and biomagnification factors (BMF, kgbac/kgnem, Aq+D treatment) were calculated using the best-fit parameters and simulating nematode tissue concentrations up to steady state (the fit of simulated total tissue concentrations reaching a plateau and running parallel to the x-axis). For the calculation of the BAF, we divided the simulated tissue concentration by the measured average total phenanthrene concentration in the whole exposure medium during the uptake phase (CM in mg/L, aqueous and bacteriaassociated phenanthrene).
Figure 1. Schematic representation of the two-compartment model used for modeling phenanthrene accumulation in C. elegans. For abbreviations, please consult the main text.
Phenanthrene can enter the nematode via the dissolved phase (ingestion of liquid medium or diffusion through the body wall) and the bacteria-associated phase (ingestion of food bacteria), where it will first arrive in the central compartment. From there, phenanthrene can be transferred to the peripheral (storage) compartment and back again. Elimination to the outside medium only takes place via the central compartment. In addition, growth of nematodes during the experiment dilutes tissue concentrations (or in case of the Aq treatment enriched tissue phenanthrene concentrations because nematodes lost biomass during the experiment). In the Aq treatment the concentration of phenanthrene in the central compartment of the nematodes CC (mg/kgnem) is governed by uptake from the liquid phase with uptake rate constant ku (L/(kgnem × h)) and the concentration in the water phase CW (mg/L), transfer from the peripheral compartment CP (mg/kgnem) with rate constant kPC (1/h), and by elimination to the outside medium (rate constant ke, 1/h), growth dilution (assuming exponential growth; growth rate g 1/h), and transfer to the peripheral compartment with rate constant kCP (1/h): dC C = k u × C W − (ke + k CP + g ) × CC + kPC × C P dt (1)
For the case of the Aq+D treatment, the equation has to be extended to also include uptake from bacterial food with the nematodes’ ingestion rate for bacteria IR (kgbac/(kgnem × h)), the efficiency with which phenanthrene can be assimilated from the bacterial food resource AE, and the concentration of bacteria-associated phenanthrene CB (mg/kgbac): dC C = k u × C W + AE × IR × C B − (ke + k CP + g ) dt × CC + kPC × C P
(2)
For the peripheral compartment, only the transfer from the central compartment, transfer back to the central compartment, and growth dilution are important: dC P = k CP × CC − (kPC + g ) × C P dt
(5)
BCF =
C tot mean C W uptake phase
(6)
BAF =
C tot mean CM uptake phase
(7)
(3)
Because CC and CP are both expressed relative to the total mass of the nematodes, for the concentration of phenanthrene C
DOI: 10.1021/es504553t Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 2. Toxicokinetics of phenanthrene in the nematode C. elegans exposed to it via an aqueous phase only (Aq) or via both aqueous and dietary sources (Aq+D). Nematode phenanthrene concentrations are shown for the Aq treatment in part a and the Aq+D treatment in part b (n = 3; black solid circles: measured values; black open circles: outlier at t = 12 h, not included in the shown model fit; black crosses: values that were
Phenanthrene Bioaccumulation in the Nematode Caenorhabditis elegans Nicole Spann,*,† Willem Goedkoop,§ and Walter Traunspurger† †
Department of Animal Ecology, Bielefeld University, Bielefeld, Germany Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences, Uppsala, Sweden
§
S Supporting Information *
ABSTRACT: The contribution of food to the bioaccumulation of xenobiotics and hence toxicity is still an ambiguous issue. It is becoming more and more evident that universal statements cannot be made, but that the relative contribution of foodassociated xenobiotics in bioaccumulation depends on species, substance, and environmental conditions. Yet, small-sized benthic or soil animals such as nematodes have largely been disregarded so far. Bioaccumulation of the polycyclic aromatic hydrocarbon phenanthrene in the absence and presence of bacterial food was measured in the nematode Caenorhabditis elegans. Elimination of phenanthrene in the nematodes was biphasic, suggesting that there was a slowly exchanging pool within the nematodes or that biotransformation of phenanthrene took place. Even with food present, dissolved phenanthrene was still the major contributor to bioaccumulated compound in nematode tissues, whereas the diet only contributed about 9%. Toxicokinetic parameters in the treatment without food were different from the ones of the treatment with bacteria, possibly because nematodes depleted their lipid reserves during starvation.
■
INTRODUCTION Within their habitat, animals are exposed to chemical pollutants via multiple routes. Chemicals can be taken up from the surrounding aqueous medium, i.e., water for pelagic organisms or pore water for sediment- or soil-inhabiting ones, and from ingested particles, i.e., food, sediment/soil particles. The importance of both routes is ambiguous in many situations,1−5 and highly dependent on the chemical characteristics of the pollutants (e.g., hydrophobicity). Especially for sedimentand soil-inhabiting invertebrates, particles can be a major contributor to accumulation and toxicity.6,7 The commonly used test organisms for assessing bioaccumulation in sediment and soil environments are macroinvertebrate annelid worms,6,8−10 nonbiting midge larvae (Chironomus spec.),8 and several species of crustaceans (e.g., gammarids, amphipods).8,11,12 However, these habitats also harbor a suite of smaller animals (meiofauna) such as, for example, nematodes and rotifers that frequently occur in high abundances and have important ecosystem functions that could be affected by the exposure to pollutants. Nematodes are an extremely diverse taxon and the dominant group within the meiofauna.13 Their well-studied species Caenorhabditis elegans has been frequently used to quantify the deleterious effects of trace metals and organic pollutants on a wide range of end points including mortality, reproduction, growth, behavior, population growth, gene expression, and metabolism. C. elegans feeds on bacteria by drawing liquid with suspended particles into its pharynx (pharyngeal pumping). In © XXXX American Chemical Society
this process the particles get transported to the intestine whereas most of the liquid then is expelled.14 In the absence of a food stimulus the pharyngeal pumping is very much reduced.15 The species also has a thick, but permeable, cuticle. C. elegans can therefore take up substances via its body surface and through the intestine via particulate food and liquid uptake.1 Several recent articles have addressed the question of the relative importance of aqueous and dietary pathways for metal toxicity by providing the animals exclusively with contaminated medium, with contaminated food in a clean medium, or vice versa, or with both contaminated medium and food.1−3 These studies concurred that bacteria facilitate metal uptake and therefore enhance toxicity. Yet, these studies used toxicity measurements such as reproduction, growth, gene expression, and antioxidant enzymes as assessment tools, while only a single study quantified body burdens at two points in time during the experiment.3 In the chosen test-setups the test substance will always partition into both phases (dissolved and particle-associated).1,2,16 The results of these studies are therefore difficult to interpret because both phases will simultaneously contribute to accumulation and toxicity and the relative contributions cannot be separated clearly by assessing toxicity. Received: September 17, 2014 Revised: January 8, 2015 Accepted: January 9, 2015
A
DOI: 10.1021/es504553t Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
treatment, but also including 5.5 × 108 E. coli/mL. To allow partitioning of phenanthrene to the bacteria in the Aq+D treatment (assuming that E. coli does not metabolize phenanthrene23), age-synchronous adult nematodes were only added 24 h after preparing the test vials with the above exposure solutions (50−250 nematodes per vial; in the case of the Aq+D treatment FUdR was also added at this later point). The test vials with the nematodes were kept in the dark at 20 °C. After 24 h of phenanthrene uptake, the nematodes in the samples allocated for the elimination phase were transferred to fresh medium containing no phenanthrene and methanol (medium either without, Aq treatment, or with bacteria, Aq+D treatment; details in SI). At each sampling occasion, actual exposure concentrations in the medium (dissolved and bacteria-associated) as well as nematode internal phenanthrene concentrations were measured by liquid scintillation counting (LSC). In the Aq treatment, a 0.5 mL medium subsample was used to determine the dissolved phenanthrene concentration (CW in mg/L). In the Aq+D treatment, the LSC measurement of a 0.5 mL medium sample yielded the total phenanthrene concentration (CM, mg/L) in the medium (dissolved and bacteria-bound). Another 1 mL subsample was centrifuged to pellet bacteria, and radioactivity was determined in 0.5 mL of the bacteria-free supernatant (giving the dissolved phenanthrene concentration CW, mg/L). Bacterial biomass in the medium was quantified in a 1 mL subsample from each vial of the Aq+D treatment by spectrophotometry (600 nm) and converted to bacterial wet mass in kgbac per L of medium. The difference between the total and dissolved phenanthrene concentrations together with the bacteria biomass concentration (kgbac/L) was used to calculate the concentration of phenanthrene in bacteria (CB expressed as mg phenanthrene per kgbac; based on wet mass of bacteria). Nematodes were removed from the test vials and left for 25−30 min to clear their guts in clean K-medium,24 before 20 animals were transferred to a new scintillation vial, followed by LSC. Nematode phenanthrene concentrations, Ctot, were expressed in mg phenanthrene per kgnem wet mass of nematodes (nematode biomass determination is described in the SI). Ten samples were below the limit of detection (LOD; six nematode samples at t = 0 h, four bacteria samples from the elimination period of the Aq+D treatment, see the SI), but the calculated concentration values were used nonetheless and were not replaced by arbitrary values. Additional details on sample processing, radioactivity analysis, and detection limits can be found in the SI. Other details that were needed for describing the toxicokinetics in C. elegans were determined in separate experiments. These were the ingestion rate (mass of bacteria consumed per nematode biomass and hour, 0.0142 kgbac/ (kgnem × h), all based on wet mass), and the growth rate (assuming exponential growth) of nematodes in the Aq (−0.0029 1/h) and Aq+D treatment (0.0059 1/h). Details of both experiments can also be found in the SI. Toxicokinetic Modeling. At first, we modeled phenanthrene uptake by the nematodes using a simple onecompartment model. However, especially for the Aq treatment, this model overestimated elimination and the corresponding curve did not fit the measured values (SI). Therefore, we modified the model to comprise a quickly exchanging central and a slowly exchanging peripheral compartment (e.g., storage compartment) in the nematodes (Figure 1).
In this study we therefore apply a toxicokinetic approach, where concentrations in the nematodes are measured at certain points in time and the resulting data combined with kinetic modeling. This is a standard technique in studies on largersized animals (e.g., annelids, fish, mussels),6−8,17,18 and recently also models for smaller-sized crustaceans were put forward.19,20 However, for animals as small as nematodes this technique has not been applied before. As the model substance we chose 14Cradiolabeled-phenanthrene, a polycyclic aromatic hydrocarbon with an intermediate octanol−water partition coefficient (log KOW 4.5221), that will partition into both, dissolved and particulate, phases. We exposed adult C. elegans to either dissolved phenanthrene only or to dissolved and bacteriaassociated phenanthrene, and we measured nematode internal concentrations over a 24 h uptake phase followed by 72 h of elimination. We expected a significant contribution of the bacteria-associated fraction to the phenanthrene uptake flux within the nematodes as well as a higher uptake rate from the dissolved phase in the combined treatment (stimulated pharyngeal pumping due to the presence of bacteria and therefore more liquid intake into the intestine). Both mechanisms will lead to higher internal concentrations in the combined treatment after the 24 h uptake phase. We supplemented the toxicokinetic approach with an experiment on the effects of the dissolved-only and combined exposure to phenanthrene on the reproduction of C. elegans. We assumed that the dissolved-only phenanthrene would result in lower toxicity due to lower accumulation of phenanthrene in the absence of food.
■
MATERIALS AND METHODS Bioaccumulation Experiment. Age-synchronous, infertile young adult C. elegans were grown in K-medium supplemented with Escherichia coli OP50, cholesterol, and 5-fluoro-2′deoxyuridine (FUdR; additional details in Supporting Information, SI).22 Before their use in the experiment, they were left to depurate for 1−2.5 h. Toxicokinetic parameters for phenanthrene in C. elegans were determined in two experiments both consisting of a 24 h uptake phase (nematodes exposed to radiolabeled 14C-phenanthrene; dissolved in methanol, specific activity 1.924 × 109 Bq/mmol, Moravek Biochemicals, Brea, CA), followed by a 72 h elimination phase in phenanthrene-free medium. Subsamples for phenanthrene analysis (medium and nematode samples) were taken after 0, 2, 4, 7, 12, and 24 h during the uptake phase (exact times are in SI), and after 28, 36, 48, 72, and 96 h after the start of the experiment in the elimination phase. The nominal exposure concentration of 14C-phenanthrene used during the 24 h exposure phase was 0.02 mg/L, which is well below the acute toxicity of phenanthrene (SI; mortality of C. elegans after 24 h was always ≤10% in concentrations up to 100 mg/L). A concentration of 0.02 mg/L phenanthrene also showed no negative effects on the reproduction of C. elegans (details in SI). Tests were conducted with the following two treatments (three replicates per sampling occasion): aqueous (Aq) treatment and aqueous plus dietary (Aq+D) treatment. Test vials (glass scintillation vials) for the Aq treatment contained 2 mL of K-medium with 0.12 mM FUdR, and 0.02 mg/L 14Cphenanthrene (phenanthrene was added via a stock solution with methanol, final concentration of methanol in test vial was 0.25%). For the Aq+D treatment, glass vessels contained 4 mL of exposure solution with the same composition as in the Aq B
DOI: 10.1021/es504553t Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
in the whole organism (Ctot) the following is therefore applicable:
C tot = CC + C P
(4)
The contribution of the dietary uptake pathway (via bacteriaassociated phenanthrene) to the total uptake flux can be calculated as percentage originating from food AE × IR × C B = × 100 k u × C W + AE × IR × C B
We used the numerical routines in the software program OpenModel version 2.2.025 to estimate the nine parameters for both treatments simultaneously in one fitting procedure (two separate values with one each for ku, ke, kCP, kPC in the Aq and Aq+D treatments; one value for AE, constrained to be ≤1 in the Aq+D treatment). Concentrations of phenanthrene in water (CW) and bacteria (CB) as the average value for each point in time were linearly interpolated between sampling points. Best-fit parameter estimates were obtained by minimizing the residual sum of squares (RSS) between measured nematode phenanthrene concentrations and predicted total nematode concentrations Ctot using the Marquardt algorithm (without weighting) in OpenModel. Subsequently, we tested if the model could be simplified by sharing parameters between treatments (e.g., using the same value for ku for both treatments). In the first round of model simplification, four new models were set up (each one having only eight parameters to be estimated), where in each new model one parameter (ku, ke, kCP, or kPC) was shared between treatments, and new parameter estimates were generated with OpenModel. If more than one reduced model provided an adequate fit to the data (no significant increase in the RSS compared to the RSS of the original nine-parameter model; Ftest,26 p > 0.05), the model with the lowest RSS was chosen as the new best-fitting model. This new eight-parameter model was adopted as the reference model against which further simplified models (having only seven parameters) were tested in a second round of model simplification. This simplification procedure was repeated until the simplest model, which still explained the data well, was found. Details of the derivation of initial starting values for the Marquardt method and the settings for the software program as well as the data sets used can be found in the SI. Bioconcentration factors (BCF, L/kgnem), bioaccumulation factors (BAF, L/kgnem, Aq+D treatment), and biomagnification factors (BMF, kgbac/kgnem, Aq+D treatment) were calculated using the best-fit parameters and simulating nematode tissue concentrations up to steady state (the fit of simulated total tissue concentrations reaching a plateau and running parallel to the x-axis). For the calculation of the BAF, we divided the simulated tissue concentration by the measured average total phenanthrene concentration in the whole exposure medium during the uptake phase (CM in mg/L, aqueous and bacteriaassociated phenanthrene).
Figure 1. Schematic representation of the two-compartment model used for modeling phenanthrene accumulation in C. elegans. For abbreviations, please consult the main text.
Phenanthrene can enter the nematode via the dissolved phase (ingestion of liquid medium or diffusion through the body wall) and the bacteria-associated phase (ingestion of food bacteria), where it will first arrive in the central compartment. From there, phenanthrene can be transferred to the peripheral (storage) compartment and back again. Elimination to the outside medium only takes place via the central compartment. In addition, growth of nematodes during the experiment dilutes tissue concentrations (or in case of the Aq treatment enriched tissue phenanthrene concentrations because nematodes lost biomass during the experiment). In the Aq treatment the concentration of phenanthrene in the central compartment of the nematodes CC (mg/kgnem) is governed by uptake from the liquid phase with uptake rate constant ku (L/(kgnem × h)) and the concentration in the water phase CW (mg/L), transfer from the peripheral compartment CP (mg/kgnem) with rate constant kPC (1/h), and by elimination to the outside medium (rate constant ke, 1/h), growth dilution (assuming exponential growth; growth rate g 1/h), and transfer to the peripheral compartment with rate constant kCP (1/h): dC C = k u × C W − (ke + k CP + g ) × CC + kPC × C P dt (1)
For the case of the Aq+D treatment, the equation has to be extended to also include uptake from bacterial food with the nematodes’ ingestion rate for bacteria IR (kgbac/(kgnem × h)), the efficiency with which phenanthrene can be assimilated from the bacterial food resource AE, and the concentration of bacteria-associated phenanthrene CB (mg/kgbac): dC C = k u × C W + AE × IR × C B − (ke + k CP + g ) dt × CC + kPC × C P
(2)
For the peripheral compartment, only the transfer from the central compartment, transfer back to the central compartment, and growth dilution are important: dC P = k CP × CC − (kPC + g ) × C P dt
(5)
BCF =
C tot mean C W uptake phase
(6)
BAF =
C tot mean CM uptake phase
(7)
(3)
Because CC and CP are both expressed relative to the total mass of the nematodes, for the concentration of phenanthrene C
DOI: 10.1021/es504553t Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 2. Toxicokinetics of phenanthrene in the nematode C. elegans exposed to it via an aqueous phase only (Aq) or via both aqueous and dietary sources (Aq+D). Nematode phenanthrene concentrations are shown for the Aq treatment in part a and the Aq+D treatment in part b (n = 3; black solid circles: measured values; black open circles: outlier at t = 12 h, not included in the shown model fit; black crosses: values that were
