Chemosphere 120 (2015) 292–298
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Mechanistic modelling of toxicokinetic processes within Myriophyllum spicatum S. Heine a,⇑, W. Schmitt b, A. Schäffer a, G. Görlitz b, H. Buresová c, G. Arts d, T.G. Preuss a a
Institute for Environmental Research, RWTH Aachen University, Worringerweg 1, 52074 Aachen, Germany Bayer Crop Science AG, Alfred-Nobel-Straße 50, 40789 Monheim am Rhein, Germany c Department of Environmental Chemistry, Institute of Chemical Technology, Prague, Technicka 5, 166 28 Praha 6, Czech Republic d Alterra, Wageningen University and Research Centre, P.O. Box 47, 6700 AA Wageningen, The Netherlands b
h i g h l i g h t s A toxicokinetic model approach for M. spicatum. Analyzing time-variable exposure of chemicals within M. spicatum. Methods to estimate chemical-specific, toxicokinetic model parameters. A model-based method to estimate cuticular permeabilities.
a r t i c l e
i n f o
Article history: Received 4 April 2014 Received in revised form 8 July 2014 Accepted 18 July 2014
Handling Editor: A. Gies Keywords: Toxicokinetics Modelling Myriophyllum spicatum Time-variable exposure Risk assessment
a b s t r a c t Effects of chemicals are, in most cases, caused by internal concentrations within organisms which rely on uptake and elimination kinetics. These processes might be key components for assessing the effects of time-variable exposure of chemicals which regularly occur in aquatic systems. However, the knowledge of toxicokinetic patterns caused by time-variable exposure is limited, and gaining such information is complex. In this work, a previously developed mechanistic growth model of Myriophyllum spicatum is coupled with a newly developed toxicokinetic part, providing a model that is able to predict uptake and elimination of chemicals, as well as distribution processes between plant compartments (leaves, stems, roots) of M. spicatum. It is shown, that toxicokinetic patterns, at least for most of the investigated chemicals, can be calculated in agreement with experimental observations, by only calibrating two chemical- specific parameters, the cuticular permeability and a plant/water partition coefficient. Through the model-based determination of the cuticular permeabilities of Isoproturon, Iofensulfuron, Fluridone, Imazamox and Penoxsulam, their toxicokinetic pattern can be described with the model approach. For the use of the model for predicting toxicokinetics of other chemicals, where experimental data is not available, equations are presented that are based on the log (Poct/wat) of a chemical and estimate parameters that are necessary to run the model. In general, a method is presented to analyze time-variable exposure of chemicals more in detail without conducting time and labour intensive experiments. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Time-variable exposure of chemicals, such as plant protection products, regularly occurs in edge-of-field water bodies (Reinert et al., 2002). The toxicological assessment of these complex chemical exposure patterns is hardly possible. Most of the conducted toxicological tests dealing with aquatic organisms are mainly used with static concentrations of the respective toxicants and do not ⇑ Corresponding author. E-mail address: [email protected] (S. Heine). http://dx.doi.org/10.1016/j.chemosphere.2014.07.065 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.
account for time-variable exposure (Fairchild et al., 1997; Hanson et al., 2003; Kulkarni et al., 2013). To gain general knowledge of effects caused by time-variable exposure of toxicants, it is essential to consider toxicokinetics, as they describe the internal concentration within an organism (Ashauer and Escher, 2010). However, in aquatic risk assessment, toxicokinetics are hardly considered, because of the time and labour that specially designed toxicokinetic experiments demand. A less costly alternative are mechanistic models, capable of predicting time-dependent concentrations within aquatic organisms. Such models are already available for invertebrates (Ashauer et al., 2007;
S. Heine et al. / Chemosphere 120 (2015) 292–298
Preuss et al., 2009) and also for the macrophyte Lemna spp. (Schmitt et al., 2013). In the current research, a previously developed growth model of M. spicatum (Heine et al., in press) is coupled with an easy to parameterize toxicokinetic submodel, providing a mechanistic model which can be used to simulate time-dependent toxicokinetics of chemicals for different plant compartments including roots. Toxicokinetic predictions are based on plant characteristics, as well as physicochemical properties of the chemicals. The aim of this work is to develop a toxicokinetic model of M. spicatum and to prove that this model is suitable to predict toxicokinetics for aquatic macrophytes, in general. Furthermore, a method should be provided, how chemical-specific parameters, necessary to run the model, can be estimated without conducting toxicokinetic experiments for each chemical. 2. Materials and methods 2.1. Kinetic experiments 2.1.1. Plant culturing To obtain plants throughout the year, a stock culture of M. spicatum was established, using the procedure described by Maletzki et al. (2010). Prior to toxicokinetic experiments, several plants were removed from the stock culture. Plant parts containing roots or shoot tips were removed. Remaining plant parts were further divided into approximately five centimeter long fragments. The fragments were washed with deionized water for 30 min to remove the nutrient solution and placed in Fernbach flasks containing tap water. Tap water was changed every three days to prevent algae growth. During a period of four weeks, the plant fragments were grown to developed new shoots. For toxicokinetic experiments, the newly developed shoots with a length of four centimeters were used. 2.1.2. Chemicals [phenyl-UL-14C] – Isoproturon (3-(4-Isopropylphenyl)-1, 1-dimethylurea) (specific radioactivity 114 lCi mg1) and [phenyl-UL-14C] – Iofensulfuron (1-(2-Iodophenylsulfonyl)-3–3-(4methoxy-6-methyl-1,3,5-triazin-2-YL)urea) (specific radioactivity 101.1 lCi mg1) with a radiochemical purity > 89% were obtained from Bayer CropScience AG, Isotope Chemistry, Wuppertal, Germany. 2.1.3. Uptake and elimination experiments Uptake and elimination experiments were conducted in gas wash bottles, each with a capacity of approximately 100 mL (height: 20 cm; diameter: three cm). Each bottle was filled with 50 mL tap water containing 40 lg L1 radioactive labeled Isoproturon or 47 lg L1 radioactive labeled Iofensulfuron, corresponding to a radioactivity of 10 000 dpm mL1. Before sealing the gas wash bottles with ground glass stoppers, preventing gas exchange, one shoot was placed into each of the bottles. For uptake experiments, shoots were removed from the bottles at different sampling dates and dabbed with extra soft facial tissues to absorb water, attached on the outer surface of the shoots. Before drying the shoots for at least 24 h at 50 °Celsius and recording the dry weight, the shoot fresh weight was measured. In order to investigate, if the uptake of chemicals into M. spicatum is influenced by active, energy dependent processes, in addition to passive transport, some shoots were exposed to sodium azide for at least 72 h prior to the experiments. Sodium azide inhibits the Adenosine-50 -triphosphate (ATP) production, thus, kills plant cells and prevents any active processes influencing the toxicokinetics. A concentration of 2 g L1 azide was used to inactivate the plants in
293
accordance with Pavlostathis et al. (1998). In the following, plants exposed to sodium azide are referred to as inactivated plants. The uptake of Isoptroturon into shoots was measured after 1, 2, 72 and 120 h for normal shoots and after 1, 4, 24, 48, 72 and 192 h for inactivated shoots. Iofensulfuron uptake was measured after 1, 4, 24, 48, 96, 168 and 192 h for normal, as well as inactivated shoots. The Iofensulfuron uptake experiment with normal shoots was repeated once. For elimination experiments, the water in the bottles, containing the radioactive labeled chemical, was removed after the accumulation phase. The gas wash bottles, each including one shoot, were briefly flushed with five milliliter of water to remove chemical residuals from the glass surface. Afterwards, the bottles were filled up with 100 mL tap water to its maximum capacity, to reduce the re-uptake of chemicals and were sealed again. The water in the bottles was entirely replaced with clean water every 24 h. On each sampling date, shoots were removed from the gas wash bottles and the same procedure as in the uptake experiments was used to measure dry and fresh weight of each shoot. Elimination of Isoproturon from normal shoots was measured (after an accumulation phase of 120 h) after 24, 72, 144, 192 and 240 h. Elimination of Isoproturon from inactivated shoots was measured (after an accumulation phase of 192 h) after 24, 72, 168, 192 and 240 h. Elimination of Iofensulfuron was measured after an accumulation phase of 192 h after 4, 24, 72, 96, 144, 192, 312 and 456 h. At each measurement, three replicates were analyzed and the mean was calculated. 2.1.4. Water and plant analysis One milliliter of solution of each gas wash bottle was analyzed by liquid scintillation on each sampling date (solvent: Irga Safe Plus solving liquid; counter: LS6500 Multi-Purpose Scintillation Counter). For analyzing plants, cellulose cones, each including one shoot, were combusted for three minutes using a sample oxidizer (OX-500, Perkin Elmer). 14-C carbon dioxide, formed during combustion, was trapped in solving liquid and measured using a scintillation counter (solvent: Oxysolve C-400; counter: LS6500 Multi-Purpose Scintillation Counter). Measured values were related to fresh and dry weight of single shoots.
3. Model development The toxicokinetic model is intended to extend an earlier, independently developed growth model of M. spicatum (Heine et al., in press), being able to dynamically calculate toxicokinetics of organic chemicals in M. spicatum. The model should provide a method to analyze transport processes in terms of distribution and accumulation. 3.1. Model concept Fig. 1 shows the model concept and the interaction of the toxicokinetic submodel and the previously developed growth model. Parentheses on the left hand-side in Fig. 1 illustrate the different compartments and transport processes (arrows) which are considered in the toxicokinetic submodel. The plant is subdivided into upper plant compartments (leaves and stems) and a belowground compartment (roots). Changes of concentrations in the different compartments are calculated simultaneously by a system of first order differential equations. These equations are used to calculate two general transport processes, one describing the uptake from external solution to plant compartments and the elimination from compartments to external solution, the other one describing upward transport from compartment to compartment.
294
S. Heine et al. / Chemosphere 120 (2015) 292–298
Fig. 1. Model concept where symbols in parentheses represent considered processes and grey boxes to the left and to the right input data necessary to run the submodels (DIC: dissolved inorganic carbon).
3.2. Model approach The change of internal concentration in each compartment dC ( compartment ) is calculated by Eqs. (1)–(3), considering uptake and dt elimination rates, as well as internal transport processes. Uptake from external solution is dependent on the uptake rate constant, elimination on the elimination rate constant and the plant/water partition coefficient (Kpw) of a chemical. Uptake, as well as elimination rates, are defined by the cuticular permeability (PMcut), the surface area (A) and the internal or external water concentration. Next to transport processes from external to internal, Eqs. (1)–(3) consider internal transport processes representing the xylem sap flow in plants from roots to aboveground compartments. These internal transport processes are dependent on the xylem sap flow rate (QXF) and the plant/water partition coefficient (Kpw). By linking the calculated internal amount of a chemical to the total plant volume (Vcompartment) internal concentrations are obtained.
PMcut AC roots roots Q XFKC K pw dC roots PMcut A C ext pw ¼ dt V roots dC metðrootsÞ ð1Þ dt Q XF C stems PMcut AC stems roots þ Q XFKC K pw K dC stems PM cut A C ext pw pw ¼ dt V stems dC metðstemsÞ dt ð2Þ dC leaves ¼ dt
PM cut A C ext
PMcut AC leaves K pw
þ
Q XF C stems K pw
V leaves
w
þ V car: K car: w þ V pr: K pr: V total
ð5Þ
logðK pr: w Þ ¼ 0:7 logðPoct=wat Þ
ð6Þ
The equations describe the equilibrium partitioning of a chemical to carbohydrates or proteins in dependence of its lipophilicity in terms of the respective octanol/water partition coefficient. In addition to the already described equations, a sink term describing metabolism, was integrated into the model (Eq. (7)) with kmet as the metabolic rate of the metabolite, PMmet as the cuticular permeability of the metabolite, C met representing the concentration of the metabolite inside the plant, C int as the concentration of the parent chemical in the plant and K mw being the metabolite-compartment/water partition coefficient. C
Þ dC metðcompartmentÞ kmet C int ðPMmet A Kmet mw ¼ dt V compartment
ð7Þ
3.3. M. spicatum specific parameterization Plant volume (Vi) is calculated from fresh weight, assuming a density of one. If necessary, dry weight is converted to fresh weight by multiplying it with a factor of 6.5 (derived from own experiments). Volumes of different plant constituents were calculated using available data of single constituents and their percentage of the total plant weight based on Boyd0 s (1968) measurements.
ð3Þ
The plant/water partition coefficient (K pw ) is calculated by Eq. (4), considering the partition coefficient of different chemical constituents of the plant (water, lipids, proteins and carbohydrates) and their proportional volume in relation to the overall plant volume (Schmitt, 2008).
V lipid K lipids
logðK car: w Þ ¼ 0:741 logðP oct=wat Þ 1:86
3.4. Determination of chemical-specific cuticular permeabilities
dC metðleavesÞ dt
K pw ¼
K lipids w is set to be equal the octanol/water partition coefficient because of similar properties of lipids and octanol (lipophilic and non-polar) and K w is set to one. K car: w (carbohydrates) and K pr: w (proteins) are calculated using Eqs. (5) and (6) (Hung et al., 2010; Schwarzenbach et al., 2005).
w
þ V w Kw ð4Þ
The model was used to analyze toxicokinetic pattern of different chemicals. To describe chemical-specific toxicokinetic patterns two parameters (cuticular permeability (PMcut ) and plant/water partition coefficient (K pw )) are required. If sufficient experimental data were available, the plant/water partition coefficient was based on the external water concentration and the last measured internal concentration during uptake experiments, assuming that at this point exchange processes were in equilibrium. If experimental data were not available, Eq. (2) was used to predict the plant/water partition coefficient. Cuticular permeability was determined by
295
S. Heine et al. / Chemosphere 120 (2015) 292–298
non-linear, least-squares optimization based on uptake data using the function lsqnonlin in MATLAB (MATLAB, 2012). Measurements up to the first four hours were neglected, to avoid establishing cuticular permeabilities which were dominated by sorption patterns. It is known, that uptake through a cuticle can be divided into a first, rapid phase, characterizing sorption onto the cuticle surface and a second, slower phase, representing cuticle penetration (Schreiber and Schonherr, 1992). Besides determining the cuticular permeability by least square optimization, available data describing the uptake into M. spicatum was used to derive further chemical-specific cuticular permeabilities. Data of Vassios (2012) were used to derive cuticular permeabilities for Fluridone, Penoxsulam and Imazamox, using the previously stated approach. Uptake rates (k) for several chlorinated aromatic hydrocarbons provided by Gobas et al. (1991) were transformed into cuticular permeabilities (see Eq. (8)).
PMcut ¼
kV A
ð8Þ
Plant volumes (V) were set to be equal to the plant fresh weight. Surface area (A) was calculated as stated in Table 1.
the shoots, was analyzed using data of Buresova et al. (2013). The published pore water concentration in the sediment (Buresova et al., 2013). was used as model input. Because of the Linuron concentration in roots, which were not measured during the experiments, but were needed for applying the model in order to analyze root to shoot translocation, the root concentration in the model was set to be in equilibrium with the concentration in pore water. This is in regard to the absence of a cuticle in roots, acting as a barrier. 4. Results Table 2 lists the determined cuticular permeabilities of M. spicatum, the predicted and experimental determined plant/ water partition coefficients and the octanol/water partition coefficients for each chemical. Cuticular permeabilities increase with increasing lipophilicity of a chemical. Obviously, some differences in cuticular permeabilities between living and inactivated plants occurred but confidence intervals showed that no significant differences exist since the ranges of the confidence intervals overlap. The predicted plant/ water partition coefficient is in approximate agreement with the experimental determined, but differences increase for chemicals with decreasing lipophilicity.
3.5. Analyzing root translocation 4.1. Uptake and elimination experiments To test whether the reported speed of xylem sap flow (Tuth, 1932) is suitable to explain the root-to-shoot translocation of chemicals, the root uptake of Linuron and, consequently, translocation to
Toxicokinetics of Isoproturon in living and inactivated plants are not entirely equal (Fig. 2). Uptake can be divided into a first
Table 1 Parameterization of the toxicokinetic model. Parameter
Description
Value
Unit
Reference
A Cext Ci Cint Cmet kmet Kmw Kpw PMcut PMroots PMmet QXF Vcar. Vlipids Vpr. Vtotal Vw
Surface area External concentration Internal concentration compartment i Internal concentration entire plant Internal concentration of metabolite Metabolic rate Metabolite/water partition coefficient Plant/water partition coefficient Cuticular permeability Root permeability Cuticular permeability of metabolite Speed of xylem sap flow Volume carbohydrates Volume lipids Volume proteins Total volume Volume water
2100
cm2 g dry weight1 lg mL1 lg g fresh weight1 lg g fresh weight1 lg g fresh weight1 d1 – – cm d1 cm d1 cm d1 cm3 h1 cm3 cm3 cm3 cm3 cm3
Nielsen and Sandjensen (1991) – – – – – – – – – – Tuth (1932) Boyd (1968) Boyd (1968) Boyd (1968) Boyd (1968) Boyd(1968)
0.0035 0.024064 0.002317 0.012544 0.910925 0.872
Table 2 Derived cuticular permeabilities, experimental and predicted plant/water partition coefficients (Kpw), chemical names and octanol/water partition coefficients (Poct/wat). The gray shaded area represents experimental results using inactivated plants. Name
log(Poct/wat)
PM (cm s1) (95% CI)
log(Kpw) exp.
log(Kpw) pred.
Reference
Fluridone Penoxsulam Imazamox Iofensulfuron Isoproturon Trichlorobenzene Tetrachlorobenzene Pentachlorobenzene Hexachlorobenzene Tetrachlorobiphenyl Octachlorostyrene Hexachlorobiphenyl Octachlorobiphenyl Decachlorobiphenyl Iofensulfuron Isoproturon
3.16 0.602 0.78 0.8 2.87 4.02 4.51 5.03 5.47 6.1 6.29 7 7.8 8.26 0.8 2.87
1.89 106 (1.30 107–3.89 106) 6.39 108 (2.63 108–1.02 107) 3.24 109 (1.39 109–4.98 109) 1.44 107 (1.08 107–1.85 107) 3.54 107 (8.91 107–1.62 x 106) 6.81 107 3.17 106 9.36 106 5.11 106 1.53 105 1.37 105 1.70 105 1.69 105 5.52 106 4.05 108 (2.19 108–1.03 107) 1.36 107 (4.63 108–2.20 107)
1.30 0.62 1.00 0.59 1.24 1.52 2.24 3.14 3.03 3.70 3.79 4.40 5.79 5.73 0.47 0.64
0.80 0.44 0.44 0.44 0.67 1.44 1.89 2.40 2.84 3.47 3.66 4.36 5.16 5.62 0.44 0.67
Based on Vassios (2012) Based on Vassios (2012) Based on Vassios (2012) This work This work Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) This work This work
296
S. Heine et al. / Chemosphere 120 (2015) 292–298
Fig. 2. Uptake and elimination pattern of Isoproturon with inactivated plants (left) and living plants (right) (black symbols: experimental data, blue lines: model results, dotted lines: Isoproturon concentration in water, dashed line: model approach considering metabolism). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
rapid uptake phase followed by a slower one. This uptake pattern is the same for living, as well as for inactivated plants, although, the total amount that is taken up is four times higher in living plants. Elimination of Isoproturon is characterized by a rapid decline of internal Isoproturon amounts in both experiments. Model results are in agreement with the experimental data when the cuticular permeability and the experimentally determined plant/water partition coefficient for Isoproturon (Table 2) are used (Fig. 2, solid lines). Using a model approach including metabolism to explain differences between living and inactivated plants by optimizing the metabolic parameters, does not increase the agreement between model predictions and experimental data (Fig. 2, dashed line). The squared differences between the measured and predicted data are smallest when metabolism is not considered. Toxicokinetics of Iofensulfuron differs only slightly between living and inactivated plants (Fig. 3). While both diagrams show, equal to the toxicokinetics of Isoproturon, a first rapid uptake phase followed by a second, slower one, the differences of Iofensulfuron uptake between living and inactivated plants are small. Alike, the difference in the total amount of Iofensulfuron that is taken up in living and inactivated plants is, compared to that of Isoproturon, small. Elimination of Iofensulfuron is characterized by a first, rapid decrease of Iofensulfuron. Afterwards considerable amounts of Iofensulfuron stay within M. spicatum shoots. Model results are only in agreement with the experimental data when the cuticular permeability of Iofensulfuron and the plant/water partition coefficient (Table 2) are used, in addition to metabolic parameters that are calibrated to the uptake and elimination data. The metabolic parameters are intended to describe the supposed persistent binding of Iofensulfuron.
Fig. 3. Uptake and elimination pattern of Iofensulfuron (black points: experimental data, blue line: model results, grey points: inactivated plants, dotted line: Iofensulfuron concentration in water). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
4.2. Analyzing root to shoot translocation By calibrating the speed of the xylem sap flow (within the maximum and average values measured by Tuth (1932)), which is the carrier for root to shoot movement, and assuming that shoot to water elimination of Linuron does not occur (as observed during the experiments) model predictions are in agreement with the experimental data for the Linuron treatments of 0.29, 2.37 and 31.37 lg L1 (Fig. 4). The respective xylem sap flow speed is 0.0065 cm3 h1 which is within the range of the average (0.0035 cm3 h1) and the maximum sap flow speed (0.01 cm3 h1) recorded by Tuth (1932). However, model predictions for the two highest Linuron treatments of 451.83 and 1100.5 lg L1 overestimated the amount of Linuron translocated to the shoots. 4.3. Correlation of cuticular permeability and physicochemical properties Fig. 5 (left side) correlates the determined chemical-specific cuticular permeabilities of M. spicatum (Table 2) with the octanol/water partition coefficient of each chemical. The correlation of the chemical-specific cuticular permeabilities of M. spicatum and the chemical lipophilicities, in terms of the octanol/water partition coefficient, can be described by Eq. (9).
logðPMcut Þ ¼ 0:3254 logðPoct=wat Þ 7:1529 R2 : 0:81
ð9Þ
Fig. 4. Experimental (data from Buresova et al. (2013)) and predicted root-to-shoot translocation of Linuron.
S. Heine et al. / Chemosphere 120 (2015) 292–298
297
Fig. 5. Logarithm of the cuticular permeabilities (PMcut) of chemicals plotted against the respective octanol/water partition coefficients (left) [black points: experimental data, gray squares: data of Vassios (2012), white squares: data of Gobas et al. (1991)] and correlation of cuticular permeabilities and octanol/water partition coefficients for different plant species (right) [equations for Hedera helix taken from Popp et al. (2005); for Citrus auranticum from Kerler and Schonherr (1988)].
Fig. 5 (right side) shows the relationship for chemical-specific cuticular permeabilities and their octanol/water partition coefficients for M. spicatum and two terrestrial plant species. The relationship for Hedera helix are taken from Popp et al. (2005), the ones for Citrus auranticum from Kerler and Schonherr (1988). Both sources are based on measurements with isolated cuticles. While the intercept of the equation established in this work, that correlates chemical-specific cuticular permeability and lipophilicity, is within the range of the equations established by others for terrestrial plant cuticles, the slope for M. spicatum is considerably lower. 5. Discussion 5.1. Cuticular permeabilities and plant/water partition coefficients Most cuticular permeablities of M. spicatum were within ranges known from terrestrial plant cuticles. Some, especially lipophilic compounds, showed cuticular permeabilities outside the known ranges of terrestrial plant cuticles. These differences might be related to different, species-specific structures of plant cuticles, which can be divided into six types (Jeffree, 2007). For amphibious plants it was shown that the cuticles of submerged leaves were in average three orders of magnitude more permeable for oxygen than aerial leaves of the same species (Frost-Christensen et al., 2003). Also Schönherr (1976) discovered that the cuticle of Potamogeton lucens was more permeable for water than the cuticle of terrestrial species. The experimentally determined cuticular permeabilities of Iofensulfuron and Isoproturon were different for living and inactivated M. spicatum plants. Since the chemical migrates through the cuticle, consisting of dead plant material, permeability should be unaffected by plant processes. However, concentration gradients, which are a major factor of diffusion, might be altered by plant processes. For example, by incorporating a chemical in a plant compartment where it does not account for concentration gradients between the external phase (water) and the plant zone behind the cuticle. It is well established, that in plants chemicals are transported to specialized compartments as the vacuole (Coleman et al., 1997) and, therefore, might influence concentration gradients and permeation. For Iofensulfuron, it was shown that some of the observed uptake and elimination pattern might be explained by assuming that plants indirectly influence cuticular permeability through metabolism. For Isoproturon, this approach did not increase the agreement between model predictions and experimental data.
Supposed that cuticular permeability of M. spicatum is affected by movement of compounds within its tissues, the determined cuticular permeabilities comprise additional plant and chemical specific information. The hypothesis that differences between living and inactivated plants might be caused by plant induced processes, is supported by differences in plant/water partition coefficients of Isoproturon between living and inactivated plants. Metabolism, in our experiment, can be excluded as reason for higher plant/water partition coefficients in living plants compared to inactivated ones, since metabolites are, in most cases, less lipophilic than the parent chemical. However, some metabolites might bind to lignin and cellulose in the apoplast (Coleman et al., 1997) and, thus, cause apparent higher plant/water partition coefficients. However, overlapping confidence intervals for the cuticular permeability of Iofensulfuron and Isoproturon for living and inactivated plants showed that no clear differences between both exist. Therefore, the question whether the cuticular permeabilities determined in this work are affected by active chemical movement processes within the plants cannot be answered. The intrinsic parameter of the uptake model (the volume flow), which was inferred from the experimental data, is the product of cuticular permeability and surface area (PMcut A). The reported cuticular permeabilities (PMcut) were calculated from this, by using surface area (A) values estimated from the plant biomass. Therefore, the PMcut-values might be biased by a systematic error in the calculation of the plant surface area from the plant weight. This potential error can, however, easily be corrected by rescaling the PMcut-values if a more accurate estimation of the surface area is available. 5.2. Analyzing root-to-shoot translocation Root-to-shoots translocation is driven by the speed of xylem sap flow. Within our simulations it had a value of 0.0065 cm3 h1 under the prerequisite that exchange from shoots to the external water phase does not occur. This is in accordance to the experiments of Buresova et al. (2013) on Linuron plant interaction which revealed that nearly no Linuron was released from plant shoots. The disagreement between our model results and the experimental data at the two highest treatments might be caused by toxic effects since EC50 values of Linuron are 137 lg L1 (Kemp et al., 1985). Model results indicate that the xylem sap flow collapses at the highest treatments which is in accordance with first observed
298
S. Heine et al. / Chemosphere 120 (2015) 292–298
changes in plant appearance (smaller and chlorosis) and effect concentrations calculated from the experiments (Buresova et al., 2013). 5.3. Correlation of cuticular permeability and physicochemical properties Cuticular permeability of M. spicatum is proportional to the lipophilicity of a chemical. This relationship was also identified for terrestrial plant cuticles (Kerler and Schonherr, 1988) which, however, is slightly different from that of M. spicatum, potentially due to species-specific differences of the cuticle structures (Riederer, 2006). Furthermore, the equation established to describe the dependency between cuticular permeability and lipophilicity is based on chemicals with log(Poct/wat) values covering the large range of 0.8 to 8.26. Equations from other authors focus either only on lipophilic chemicals (Kerler and Schonherr, 1988) or use two equations, one describing the dependency for lipophilic, the other one the dependency for hydrophilic chemicals (Popp et al., 2005). It also has to be noted, that in contrast to most other research, where cuticles were isolated from plants, we used a model approach to determine cuticular permeability which might introduce some uncertainty. 6. Conclusion This study demonstrated that the presented mechanistic model approach is suitable to describe toxicokinetics of Isoproturon, Iofensulfuron, Fluridone, Imazamox and Penoxsulam within M. spicatum. Through the established equations that estimate toxicokinetic parameters, that are necessary to run the model, the model can also be used to predict toxicokinetics of other chemicals just with the respective octanol/water partition coefficient. Special toxicokinetic experiments, as conducted within this study, are not mandatory to use the model approach. The model analysis indicates that the speed of the root-to-shoot translocation of chemicals is realistically described by the model approach. Furthermore, a general, model-based method is presented how cuticular permeabilities can be estimated without isolating plant cuticles, which is, at least for some plant species, hardly possible. As a final conclusion, it can be stated, that the model approach can be used as a tool to analyze time-variable chemical exposure. In a future perspective, by integrating toxicodynamics, the model might even be used to predict the effects caused by time-variable exposure of toxicants, such as plant protection products. Acknowledgement This research was funded by Bayer CropScience. References Ashauer, R., Escher, B.I., 2010. Advantages of toxicokinetic and toxicodynamic modelling in aquatic ecotoxicology and risk assessment. J. Environ. Monit. 12, 2056–2061. Ashauer, R., Boxall, A.B.A., Brown, C.D., 2007. New ecotoxicological model to simulate survival of aquatic invertebrates after exposure to fluctuating and sequential pulses of pesticides. Environ. Sci. Technol. 41, 1480–1486. Boyd, C.E., 1968. Fresh-water plants – a potential source of protein. Econ. Bot. 22, 359–367.
Buresova, H., Crum, S.J.H., Belgers, J.D.M., Adriaanse, P.I., Arts, G.H.P., 2013. Effects of linuron on a rooted aquatic macrophyte in sediment-dosed test systems. Environ. Pollut. 175, 117–124. Coleman, J.O.D., Blake-Kalff, M.M.A., Davies, T.G.E., 1997. Detoxification of xenobiotics by plants: chemical modification and vacuolar compartmentation. Trends Plant Sci. 2, 144–151. Fairchild, J.F., Ruessler, D.S., Haverland, P.S., Carlson, A.R., 1997. Comparative sensitivity of Selenastrum capricornutum and Lemna minor to sixteen herbicides. Arch. Environ. Contam. Toxicol. 32, 353–357. Frost-Christensen, H., Bolt Jorgensen, L., Floto, F., 2003. Species specificity of resistance to oxygen diffusion in thin cuticular membranes from amphibious plants. Plant Cell Environ. 26, 561–569. Gobas, F.A.P.C., McNeil, E.J., Lovett-Doust, L., Haffner, G.D., 1991. Bioconcentration of chlorinated aromatic hydrocarbons in aquatic macrophytes. Environ. Sci. Technol. 25, 924–929. Hanson, M.L., Sibley, P.K., Mabury, S.A., Muir, D.C.G., Solomon, K.R., 2003. Field level evaluation and risk assessment of the toxicity of dichloroacetic acid to the aquatic macrophytes Lemna gibba, Myriophyllum spicatum, and Myriophyllum sibiricum. Ecotoxicol. Environ. Saf. 55, 46–63. Heine, S., Schmitt, W., Görlitz, G., Schäffer, A., Preuss, T.G., in press. Effects of light and temperature fluctuations on the growth of Myriophyllum spicatum in toxicity tests - A model based analysis. Environ. Sci. Pollut. R. Hung, H., Lin, T., Chiou, C.T., 2010. Partition coefficients of organic contaminants with carbohydrates. Environ. Sci. Technol. 44, 5430–5436. Jeffree, C.E., 2007. The fine structure of the plant cuticle. In: Riederer, M., Müller, C. (Eds.), Biology of the Plant Cuticle. Annual Plant Reviews, vol. 23. springer, berlin, pp. 11–110. Kemp, W.M., Boynton, W.R., Cunningham, J.J., Stevenson, J.C., Jones, T.W., Means, J.C., 1985. Effects of atrazine and linuron on photosynthesis and growth of the macrophytes Potamogeton perfoliatus L. and Myriophyllum spicatum L. in an estaurine environment. Mar. Environ. Res. 16, 255–280. Kerler, F., Schonherr, J., 1988. Accumulation of lipophilic chemicals in plant cuticles – prediction from octanol water partition-coefficient. Arch. Environ. Con. Toxicol. 17, 1–6. Kulkarni, D., Daniels, B., Preuss, T.G., 2013. Life-stage-dependent sensitivity of the cyclopoid copepod Mesocyclops leuckarti to triphenyltin. Chemosphere 92, 1145–1153. Maletzki, D., Hunken, A., Hubner, I., Kussatz, C., 2010. Mynophyllum spicatum as a test organism development of a sediment free test system and first results with 3,5-dichlorophenol. Umweltwiss. und Schadst.-Forsch. 22, 702–710. MATLAB version 7.14.0.739. 2012. Natick, Massachusetts: The MathWorks Inc. Nielsen, S.L., Sandjensen, K., 1991. Variation in growth rates of submerged rooted macrophytes. Aquat. Bot. 39, 109–120. Pavlostathis, S.G., Comstock, K.K., Jacobson, M.E., Saunders, F.M., 1998. Transformation of 2,4,6-trinitrotoluene by the aquatic plant Myriophyllum spicatum. Environ. Toxicol. Chem. 17, 2266–2273. Popp, C., Burghardt, M., Friedmann, A., Riederer, M., 2005. Characterization of hydrophilic and lipophilic pathways of Hedera helix L. cuticular membranes: permeation of water and uncharged organic compounds. J. Exp. Bot. 56, 2797– 2806. Preuss, T.G., Hammers-Wirtz, M., Hommen, U., Rubach, M.N., Ratte, H.T., 2009. Development and validation of an individual based Daphnia magna population model: the influence of crowding on population dynamics. Ecol. Model. 220, 310–329. Reinert, K.H., Giddings, J.M., Judd, L., 2002. Effects analysis of time-varying or repeated exposures in aquatic ecological risk assessment of agrochemicals. Environ. Toxicol. Chem. 21, 1977–1992. Riederer, M., 2006. Thermodynamics of the water permeability of plant cuticles: characterization of the polar pathway. J. Exp. Bot. 57, 2937–2942. Schmitt, W., 2008. General approach for the calculation of tissue to plasma partition coefficients. Toxicol. In Vitro 22, 457–467. Schmitt, W., Bruns, E., Dollinger, M., Sowig, P., 2013. Mechanistic TK/TD-model simulating the effect of growth inhibitors on Lemna populations. Ecol. Model. 255, 1–10. Schönherr, J., 1976. Water permeability of isolated cuticular Membranes: the effect of cuticular waxes on diffusion of water. Planta (Berl.) 131, 159–164. Schreiber, L., Schonherr, J., 1992. Uptake of organic chemicals in conifer needles: surface adsoprtion and permeability of cuticles. Environ. Sci. Technol. 26, 153– 159. Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M., 2005. Environmental Organic Chemistry. Wiley-Interscience, p. 341. Tuth, H.F., 1932. The movement of water through some submerged plants. Amer. J. Bot. 19, 693–709. Vassios, J.D., 2012. Herbicide absorption and translocation by Eurasian Watermilfoil and hydrilla. PhD thesis. Colorado State University, Colorado.
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Mechanistic modelling of toxicokinetic processes within Myriophyllum spicatum S. Heine a,⇑, W. Schmitt b, A. Schäffer a, G. Görlitz b, H. Buresová c, G. Arts d, T.G. Preuss a a
Institute for Environmental Research, RWTH Aachen University, Worringerweg 1, 52074 Aachen, Germany Bayer Crop Science AG, Alfred-Nobel-Straße 50, 40789 Monheim am Rhein, Germany c Department of Environmental Chemistry, Institute of Chemical Technology, Prague, Technicka 5, 166 28 Praha 6, Czech Republic d Alterra, Wageningen University and Research Centre, P.O. Box 47, 6700 AA Wageningen, The Netherlands b
h i g h l i g h t s A toxicokinetic model approach for M. spicatum. Analyzing time-variable exposure of chemicals within M. spicatum. Methods to estimate chemical-specific, toxicokinetic model parameters. A model-based method to estimate cuticular permeabilities.
a r t i c l e
i n f o
Article history: Received 4 April 2014 Received in revised form 8 July 2014 Accepted 18 July 2014
Handling Editor: A. Gies Keywords: Toxicokinetics Modelling Myriophyllum spicatum Time-variable exposure Risk assessment
a b s t r a c t Effects of chemicals are, in most cases, caused by internal concentrations within organisms which rely on uptake and elimination kinetics. These processes might be key components for assessing the effects of time-variable exposure of chemicals which regularly occur in aquatic systems. However, the knowledge of toxicokinetic patterns caused by time-variable exposure is limited, and gaining such information is complex. In this work, a previously developed mechanistic growth model of Myriophyllum spicatum is coupled with a newly developed toxicokinetic part, providing a model that is able to predict uptake and elimination of chemicals, as well as distribution processes between plant compartments (leaves, stems, roots) of M. spicatum. It is shown, that toxicokinetic patterns, at least for most of the investigated chemicals, can be calculated in agreement with experimental observations, by only calibrating two chemical- specific parameters, the cuticular permeability and a plant/water partition coefficient. Through the model-based determination of the cuticular permeabilities of Isoproturon, Iofensulfuron, Fluridone, Imazamox and Penoxsulam, their toxicokinetic pattern can be described with the model approach. For the use of the model for predicting toxicokinetics of other chemicals, where experimental data is not available, equations are presented that are based on the log (Poct/wat) of a chemical and estimate parameters that are necessary to run the model. In general, a method is presented to analyze time-variable exposure of chemicals more in detail without conducting time and labour intensive experiments. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Time-variable exposure of chemicals, such as plant protection products, regularly occurs in edge-of-field water bodies (Reinert et al., 2002). The toxicological assessment of these complex chemical exposure patterns is hardly possible. Most of the conducted toxicological tests dealing with aquatic organisms are mainly used with static concentrations of the respective toxicants and do not ⇑ Corresponding author. E-mail address: [email protected] (S. Heine). http://dx.doi.org/10.1016/j.chemosphere.2014.07.065 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.
account for time-variable exposure (Fairchild et al., 1997; Hanson et al., 2003; Kulkarni et al., 2013). To gain general knowledge of effects caused by time-variable exposure of toxicants, it is essential to consider toxicokinetics, as they describe the internal concentration within an organism (Ashauer and Escher, 2010). However, in aquatic risk assessment, toxicokinetics are hardly considered, because of the time and labour that specially designed toxicokinetic experiments demand. A less costly alternative are mechanistic models, capable of predicting time-dependent concentrations within aquatic organisms. Such models are already available for invertebrates (Ashauer et al., 2007;
S. Heine et al. / Chemosphere 120 (2015) 292–298
Preuss et al., 2009) and also for the macrophyte Lemna spp. (Schmitt et al., 2013). In the current research, a previously developed growth model of M. spicatum (Heine et al., in press) is coupled with an easy to parameterize toxicokinetic submodel, providing a mechanistic model which can be used to simulate time-dependent toxicokinetics of chemicals for different plant compartments including roots. Toxicokinetic predictions are based on plant characteristics, as well as physicochemical properties of the chemicals. The aim of this work is to develop a toxicokinetic model of M. spicatum and to prove that this model is suitable to predict toxicokinetics for aquatic macrophytes, in general. Furthermore, a method should be provided, how chemical-specific parameters, necessary to run the model, can be estimated without conducting toxicokinetic experiments for each chemical. 2. Materials and methods 2.1. Kinetic experiments 2.1.1. Plant culturing To obtain plants throughout the year, a stock culture of M. spicatum was established, using the procedure described by Maletzki et al. (2010). Prior to toxicokinetic experiments, several plants were removed from the stock culture. Plant parts containing roots or shoot tips were removed. Remaining plant parts were further divided into approximately five centimeter long fragments. The fragments were washed with deionized water for 30 min to remove the nutrient solution and placed in Fernbach flasks containing tap water. Tap water was changed every three days to prevent algae growth. During a period of four weeks, the plant fragments were grown to developed new shoots. For toxicokinetic experiments, the newly developed shoots with a length of four centimeters were used. 2.1.2. Chemicals [phenyl-UL-14C] – Isoproturon (3-(4-Isopropylphenyl)-1, 1-dimethylurea) (specific radioactivity 114 lCi mg1) and [phenyl-UL-14C] – Iofensulfuron (1-(2-Iodophenylsulfonyl)-3–3-(4methoxy-6-methyl-1,3,5-triazin-2-YL)urea) (specific radioactivity 101.1 lCi mg1) with a radiochemical purity > 89% were obtained from Bayer CropScience AG, Isotope Chemistry, Wuppertal, Germany. 2.1.3. Uptake and elimination experiments Uptake and elimination experiments were conducted in gas wash bottles, each with a capacity of approximately 100 mL (height: 20 cm; diameter: three cm). Each bottle was filled with 50 mL tap water containing 40 lg L1 radioactive labeled Isoproturon or 47 lg L1 radioactive labeled Iofensulfuron, corresponding to a radioactivity of 10 000 dpm mL1. Before sealing the gas wash bottles with ground glass stoppers, preventing gas exchange, one shoot was placed into each of the bottles. For uptake experiments, shoots were removed from the bottles at different sampling dates and dabbed with extra soft facial tissues to absorb water, attached on the outer surface of the shoots. Before drying the shoots for at least 24 h at 50 °Celsius and recording the dry weight, the shoot fresh weight was measured. In order to investigate, if the uptake of chemicals into M. spicatum is influenced by active, energy dependent processes, in addition to passive transport, some shoots were exposed to sodium azide for at least 72 h prior to the experiments. Sodium azide inhibits the Adenosine-50 -triphosphate (ATP) production, thus, kills plant cells and prevents any active processes influencing the toxicokinetics. A concentration of 2 g L1 azide was used to inactivate the plants in
293
accordance with Pavlostathis et al. (1998). In the following, plants exposed to sodium azide are referred to as inactivated plants. The uptake of Isoptroturon into shoots was measured after 1, 2, 72 and 120 h for normal shoots and after 1, 4, 24, 48, 72 and 192 h for inactivated shoots. Iofensulfuron uptake was measured after 1, 4, 24, 48, 96, 168 and 192 h for normal, as well as inactivated shoots. The Iofensulfuron uptake experiment with normal shoots was repeated once. For elimination experiments, the water in the bottles, containing the radioactive labeled chemical, was removed after the accumulation phase. The gas wash bottles, each including one shoot, were briefly flushed with five milliliter of water to remove chemical residuals from the glass surface. Afterwards, the bottles were filled up with 100 mL tap water to its maximum capacity, to reduce the re-uptake of chemicals and were sealed again. The water in the bottles was entirely replaced with clean water every 24 h. On each sampling date, shoots were removed from the gas wash bottles and the same procedure as in the uptake experiments was used to measure dry and fresh weight of each shoot. Elimination of Isoproturon from normal shoots was measured (after an accumulation phase of 120 h) after 24, 72, 144, 192 and 240 h. Elimination of Isoproturon from inactivated shoots was measured (after an accumulation phase of 192 h) after 24, 72, 168, 192 and 240 h. Elimination of Iofensulfuron was measured after an accumulation phase of 192 h after 4, 24, 72, 96, 144, 192, 312 and 456 h. At each measurement, three replicates were analyzed and the mean was calculated. 2.1.4. Water and plant analysis One milliliter of solution of each gas wash bottle was analyzed by liquid scintillation on each sampling date (solvent: Irga Safe Plus solving liquid; counter: LS6500 Multi-Purpose Scintillation Counter). For analyzing plants, cellulose cones, each including one shoot, were combusted for three minutes using a sample oxidizer (OX-500, Perkin Elmer). 14-C carbon dioxide, formed during combustion, was trapped in solving liquid and measured using a scintillation counter (solvent: Oxysolve C-400; counter: LS6500 Multi-Purpose Scintillation Counter). Measured values were related to fresh and dry weight of single shoots.
3. Model development The toxicokinetic model is intended to extend an earlier, independently developed growth model of M. spicatum (Heine et al., in press), being able to dynamically calculate toxicokinetics of organic chemicals in M. spicatum. The model should provide a method to analyze transport processes in terms of distribution and accumulation. 3.1. Model concept Fig. 1 shows the model concept and the interaction of the toxicokinetic submodel and the previously developed growth model. Parentheses on the left hand-side in Fig. 1 illustrate the different compartments and transport processes (arrows) which are considered in the toxicokinetic submodel. The plant is subdivided into upper plant compartments (leaves and stems) and a belowground compartment (roots). Changes of concentrations in the different compartments are calculated simultaneously by a system of first order differential equations. These equations are used to calculate two general transport processes, one describing the uptake from external solution to plant compartments and the elimination from compartments to external solution, the other one describing upward transport from compartment to compartment.
294
S. Heine et al. / Chemosphere 120 (2015) 292–298
Fig. 1. Model concept where symbols in parentheses represent considered processes and grey boxes to the left and to the right input data necessary to run the submodels (DIC: dissolved inorganic carbon).
3.2. Model approach The change of internal concentration in each compartment dC ( compartment ) is calculated by Eqs. (1)–(3), considering uptake and dt elimination rates, as well as internal transport processes. Uptake from external solution is dependent on the uptake rate constant, elimination on the elimination rate constant and the plant/water partition coefficient (Kpw) of a chemical. Uptake, as well as elimination rates, are defined by the cuticular permeability (PMcut), the surface area (A) and the internal or external water concentration. Next to transport processes from external to internal, Eqs. (1)–(3) consider internal transport processes representing the xylem sap flow in plants from roots to aboveground compartments. These internal transport processes are dependent on the xylem sap flow rate (QXF) and the plant/water partition coefficient (Kpw). By linking the calculated internal amount of a chemical to the total plant volume (Vcompartment) internal concentrations are obtained.
PMcut AC roots roots Q XFKC K pw dC roots PMcut A C ext pw ¼ dt V roots dC metðrootsÞ ð1Þ dt Q XF C stems PMcut AC stems roots þ Q XFKC K pw K dC stems PM cut A C ext pw pw ¼ dt V stems dC metðstemsÞ dt ð2Þ dC leaves ¼ dt
PM cut A C ext
PMcut AC leaves K pw
þ
Q XF C stems K pw
V leaves
w
þ V car: K car: w þ V pr: K pr: V total
ð5Þ
logðK pr: w Þ ¼ 0:7 logðPoct=wat Þ
ð6Þ
The equations describe the equilibrium partitioning of a chemical to carbohydrates or proteins in dependence of its lipophilicity in terms of the respective octanol/water partition coefficient. In addition to the already described equations, a sink term describing metabolism, was integrated into the model (Eq. (7)) with kmet as the metabolic rate of the metabolite, PMmet as the cuticular permeability of the metabolite, C met representing the concentration of the metabolite inside the plant, C int as the concentration of the parent chemical in the plant and K mw being the metabolite-compartment/water partition coefficient. C
Þ dC metðcompartmentÞ kmet C int ðPMmet A Kmet mw ¼ dt V compartment
ð7Þ
3.3. M. spicatum specific parameterization Plant volume (Vi) is calculated from fresh weight, assuming a density of one. If necessary, dry weight is converted to fresh weight by multiplying it with a factor of 6.5 (derived from own experiments). Volumes of different plant constituents were calculated using available data of single constituents and their percentage of the total plant weight based on Boyd0 s (1968) measurements.
ð3Þ
The plant/water partition coefficient (K pw ) is calculated by Eq. (4), considering the partition coefficient of different chemical constituents of the plant (water, lipids, proteins and carbohydrates) and their proportional volume in relation to the overall plant volume (Schmitt, 2008).
V lipid K lipids
logðK car: w Þ ¼ 0:741 logðP oct=wat Þ 1:86
3.4. Determination of chemical-specific cuticular permeabilities
dC metðleavesÞ dt
K pw ¼
K lipids w is set to be equal the octanol/water partition coefficient because of similar properties of lipids and octanol (lipophilic and non-polar) and K w is set to one. K car: w (carbohydrates) and K pr: w (proteins) are calculated using Eqs. (5) and (6) (Hung et al., 2010; Schwarzenbach et al., 2005).
w
þ V w Kw ð4Þ
The model was used to analyze toxicokinetic pattern of different chemicals. To describe chemical-specific toxicokinetic patterns two parameters (cuticular permeability (PMcut ) and plant/water partition coefficient (K pw )) are required. If sufficient experimental data were available, the plant/water partition coefficient was based on the external water concentration and the last measured internal concentration during uptake experiments, assuming that at this point exchange processes were in equilibrium. If experimental data were not available, Eq. (2) was used to predict the plant/water partition coefficient. Cuticular permeability was determined by
295
S. Heine et al. / Chemosphere 120 (2015) 292–298
non-linear, least-squares optimization based on uptake data using the function lsqnonlin in MATLAB (MATLAB, 2012). Measurements up to the first four hours were neglected, to avoid establishing cuticular permeabilities which were dominated by sorption patterns. It is known, that uptake through a cuticle can be divided into a first, rapid phase, characterizing sorption onto the cuticle surface and a second, slower phase, representing cuticle penetration (Schreiber and Schonherr, 1992). Besides determining the cuticular permeability by least square optimization, available data describing the uptake into M. spicatum was used to derive further chemical-specific cuticular permeabilities. Data of Vassios (2012) were used to derive cuticular permeabilities for Fluridone, Penoxsulam and Imazamox, using the previously stated approach. Uptake rates (k) for several chlorinated aromatic hydrocarbons provided by Gobas et al. (1991) were transformed into cuticular permeabilities (see Eq. (8)).
PMcut ¼
kV A
ð8Þ
Plant volumes (V) were set to be equal to the plant fresh weight. Surface area (A) was calculated as stated in Table 1.
the shoots, was analyzed using data of Buresova et al. (2013). The published pore water concentration in the sediment (Buresova et al., 2013). was used as model input. Because of the Linuron concentration in roots, which were not measured during the experiments, but were needed for applying the model in order to analyze root to shoot translocation, the root concentration in the model was set to be in equilibrium with the concentration in pore water. This is in regard to the absence of a cuticle in roots, acting as a barrier. 4. Results Table 2 lists the determined cuticular permeabilities of M. spicatum, the predicted and experimental determined plant/ water partition coefficients and the octanol/water partition coefficients for each chemical. Cuticular permeabilities increase with increasing lipophilicity of a chemical. Obviously, some differences in cuticular permeabilities between living and inactivated plants occurred but confidence intervals showed that no significant differences exist since the ranges of the confidence intervals overlap. The predicted plant/ water partition coefficient is in approximate agreement with the experimental determined, but differences increase for chemicals with decreasing lipophilicity.
3.5. Analyzing root translocation 4.1. Uptake and elimination experiments To test whether the reported speed of xylem sap flow (Tuth, 1932) is suitable to explain the root-to-shoot translocation of chemicals, the root uptake of Linuron and, consequently, translocation to
Toxicokinetics of Isoproturon in living and inactivated plants are not entirely equal (Fig. 2). Uptake can be divided into a first
Table 1 Parameterization of the toxicokinetic model. Parameter
Description
Value
Unit
Reference
A Cext Ci Cint Cmet kmet Kmw Kpw PMcut PMroots PMmet QXF Vcar. Vlipids Vpr. Vtotal Vw
Surface area External concentration Internal concentration compartment i Internal concentration entire plant Internal concentration of metabolite Metabolic rate Metabolite/water partition coefficient Plant/water partition coefficient Cuticular permeability Root permeability Cuticular permeability of metabolite Speed of xylem sap flow Volume carbohydrates Volume lipids Volume proteins Total volume Volume water
2100
cm2 g dry weight1 lg mL1 lg g fresh weight1 lg g fresh weight1 lg g fresh weight1 d1 – – cm d1 cm d1 cm d1 cm3 h1 cm3 cm3 cm3 cm3 cm3
Nielsen and Sandjensen (1991) – – – – – – – – – – Tuth (1932) Boyd (1968) Boyd (1968) Boyd (1968) Boyd (1968) Boyd(1968)
0.0035 0.024064 0.002317 0.012544 0.910925 0.872
Table 2 Derived cuticular permeabilities, experimental and predicted plant/water partition coefficients (Kpw), chemical names and octanol/water partition coefficients (Poct/wat). The gray shaded area represents experimental results using inactivated plants. Name
log(Poct/wat)
PM (cm s1) (95% CI)
log(Kpw) exp.
log(Kpw) pred.
Reference
Fluridone Penoxsulam Imazamox Iofensulfuron Isoproturon Trichlorobenzene Tetrachlorobenzene Pentachlorobenzene Hexachlorobenzene Tetrachlorobiphenyl Octachlorostyrene Hexachlorobiphenyl Octachlorobiphenyl Decachlorobiphenyl Iofensulfuron Isoproturon
3.16 0.602 0.78 0.8 2.87 4.02 4.51 5.03 5.47 6.1 6.29 7 7.8 8.26 0.8 2.87
1.89 106 (1.30 107–3.89 106) 6.39 108 (2.63 108–1.02 107) 3.24 109 (1.39 109–4.98 109) 1.44 107 (1.08 107–1.85 107) 3.54 107 (8.91 107–1.62 x 106) 6.81 107 3.17 106 9.36 106 5.11 106 1.53 105 1.37 105 1.70 105 1.69 105 5.52 106 4.05 108 (2.19 108–1.03 107) 1.36 107 (4.63 108–2.20 107)
1.30 0.62 1.00 0.59 1.24 1.52 2.24 3.14 3.03 3.70 3.79 4.40 5.79 5.73 0.47 0.64
0.80 0.44 0.44 0.44 0.67 1.44 1.89 2.40 2.84 3.47 3.66 4.36 5.16 5.62 0.44 0.67
Based on Vassios (2012) Based on Vassios (2012) Based on Vassios (2012) This work This work Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) Based on Gobas et al. (1991) This work This work
296
S. Heine et al. / Chemosphere 120 (2015) 292–298
Fig. 2. Uptake and elimination pattern of Isoproturon with inactivated plants (left) and living plants (right) (black symbols: experimental data, blue lines: model results, dotted lines: Isoproturon concentration in water, dashed line: model approach considering metabolism). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
rapid uptake phase followed by a slower one. This uptake pattern is the same for living, as well as for inactivated plants, although, the total amount that is taken up is four times higher in living plants. Elimination of Isoproturon is characterized by a rapid decline of internal Isoproturon amounts in both experiments. Model results are in agreement with the experimental data when the cuticular permeability and the experimentally determined plant/water partition coefficient for Isoproturon (Table 2) are used (Fig. 2, solid lines). Using a model approach including metabolism to explain differences between living and inactivated plants by optimizing the metabolic parameters, does not increase the agreement between model predictions and experimental data (Fig. 2, dashed line). The squared differences between the measured and predicted data are smallest when metabolism is not considered. Toxicokinetics of Iofensulfuron differs only slightly between living and inactivated plants (Fig. 3). While both diagrams show, equal to the toxicokinetics of Isoproturon, a first rapid uptake phase followed by a second, slower one, the differences of Iofensulfuron uptake between living and inactivated plants are small. Alike, the difference in the total amount of Iofensulfuron that is taken up in living and inactivated plants is, compared to that of Isoproturon, small. Elimination of Iofensulfuron is characterized by a first, rapid decrease of Iofensulfuron. Afterwards considerable amounts of Iofensulfuron stay within M. spicatum shoots. Model results are only in agreement with the experimental data when the cuticular permeability of Iofensulfuron and the plant/water partition coefficient (Table 2) are used, in addition to metabolic parameters that are calibrated to the uptake and elimination data. The metabolic parameters are intended to describe the supposed persistent binding of Iofensulfuron.
Fig. 3. Uptake and elimination pattern of Iofensulfuron (black points: experimental data, blue line: model results, grey points: inactivated plants, dotted line: Iofensulfuron concentration in water). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
4.2. Analyzing root to shoot translocation By calibrating the speed of the xylem sap flow (within the maximum and average values measured by Tuth (1932)), which is the carrier for root to shoot movement, and assuming that shoot to water elimination of Linuron does not occur (as observed during the experiments) model predictions are in agreement with the experimental data for the Linuron treatments of 0.29, 2.37 and 31.37 lg L1 (Fig. 4). The respective xylem sap flow speed is 0.0065 cm3 h1 which is within the range of the average (0.0035 cm3 h1) and the maximum sap flow speed (0.01 cm3 h1) recorded by Tuth (1932). However, model predictions for the two highest Linuron treatments of 451.83 and 1100.5 lg L1 overestimated the amount of Linuron translocated to the shoots. 4.3. Correlation of cuticular permeability and physicochemical properties Fig. 5 (left side) correlates the determined chemical-specific cuticular permeabilities of M. spicatum (Table 2) with the octanol/water partition coefficient of each chemical. The correlation of the chemical-specific cuticular permeabilities of M. spicatum and the chemical lipophilicities, in terms of the octanol/water partition coefficient, can be described by Eq. (9).
logðPMcut Þ ¼ 0:3254 logðPoct=wat Þ 7:1529 R2 : 0:81
ð9Þ
Fig. 4. Experimental (data from Buresova et al. (2013)) and predicted root-to-shoot translocation of Linuron.
S. Heine et al. / Chemosphere 120 (2015) 292–298
297
Fig. 5. Logarithm of the cuticular permeabilities (PMcut) of chemicals plotted against the respective octanol/water partition coefficients (left) [black points: experimental data, gray squares: data of Vassios (2012), white squares: data of Gobas et al. (1991)] and correlation of cuticular permeabilities and octanol/water partition coefficients for different plant species (right) [equations for Hedera helix taken from Popp et al. (2005); for Citrus auranticum from Kerler and Schonherr (1988)].
Fig. 5 (right side) shows the relationship for chemical-specific cuticular permeabilities and their octanol/water partition coefficients for M. spicatum and two terrestrial plant species. The relationship for Hedera helix are taken from Popp et al. (2005), the ones for Citrus auranticum from Kerler and Schonherr (1988). Both sources are based on measurements with isolated cuticles. While the intercept of the equation established in this work, that correlates chemical-specific cuticular permeability and lipophilicity, is within the range of the equations established by others for terrestrial plant cuticles, the slope for M. spicatum is considerably lower. 5. Discussion 5.1. Cuticular permeabilities and plant/water partition coefficients Most cuticular permeablities of M. spicatum were within ranges known from terrestrial plant cuticles. Some, especially lipophilic compounds, showed cuticular permeabilities outside the known ranges of terrestrial plant cuticles. These differences might be related to different, species-specific structures of plant cuticles, which can be divided into six types (Jeffree, 2007). For amphibious plants it was shown that the cuticles of submerged leaves were in average three orders of magnitude more permeable for oxygen than aerial leaves of the same species (Frost-Christensen et al., 2003). Also Schönherr (1976) discovered that the cuticle of Potamogeton lucens was more permeable for water than the cuticle of terrestrial species. The experimentally determined cuticular permeabilities of Iofensulfuron and Isoproturon were different for living and inactivated M. spicatum plants. Since the chemical migrates through the cuticle, consisting of dead plant material, permeability should be unaffected by plant processes. However, concentration gradients, which are a major factor of diffusion, might be altered by plant processes. For example, by incorporating a chemical in a plant compartment where it does not account for concentration gradients between the external phase (water) and the plant zone behind the cuticle. It is well established, that in plants chemicals are transported to specialized compartments as the vacuole (Coleman et al., 1997) and, therefore, might influence concentration gradients and permeation. For Iofensulfuron, it was shown that some of the observed uptake and elimination pattern might be explained by assuming that plants indirectly influence cuticular permeability through metabolism. For Isoproturon, this approach did not increase the agreement between model predictions and experimental data.
Supposed that cuticular permeability of M. spicatum is affected by movement of compounds within its tissues, the determined cuticular permeabilities comprise additional plant and chemical specific information. The hypothesis that differences between living and inactivated plants might be caused by plant induced processes, is supported by differences in plant/water partition coefficients of Isoproturon between living and inactivated plants. Metabolism, in our experiment, can be excluded as reason for higher plant/water partition coefficients in living plants compared to inactivated ones, since metabolites are, in most cases, less lipophilic than the parent chemical. However, some metabolites might bind to lignin and cellulose in the apoplast (Coleman et al., 1997) and, thus, cause apparent higher plant/water partition coefficients. However, overlapping confidence intervals for the cuticular permeability of Iofensulfuron and Isoproturon for living and inactivated plants showed that no clear differences between both exist. Therefore, the question whether the cuticular permeabilities determined in this work are affected by active chemical movement processes within the plants cannot be answered. The intrinsic parameter of the uptake model (the volume flow), which was inferred from the experimental data, is the product of cuticular permeability and surface area (PMcut A). The reported cuticular permeabilities (PMcut) were calculated from this, by using surface area (A) values estimated from the plant biomass. Therefore, the PMcut-values might be biased by a systematic error in the calculation of the plant surface area from the plant weight. This potential error can, however, easily be corrected by rescaling the PMcut-values if a more accurate estimation of the surface area is available. 5.2. Analyzing root-to-shoot translocation Root-to-shoots translocation is driven by the speed of xylem sap flow. Within our simulations it had a value of 0.0065 cm3 h1 under the prerequisite that exchange from shoots to the external water phase does not occur. This is in accordance to the experiments of Buresova et al. (2013) on Linuron plant interaction which revealed that nearly no Linuron was released from plant shoots. The disagreement between our model results and the experimental data at the two highest treatments might be caused by toxic effects since EC50 values of Linuron are 137 lg L1 (Kemp et al., 1985). Model results indicate that the xylem sap flow collapses at the highest treatments which is in accordance with first observed
298
S. Heine et al. / Chemosphere 120 (2015) 292–298
changes in plant appearance (smaller and chlorosis) and effect concentrations calculated from the experiments (Buresova et al., 2013). 5.3. Correlation of cuticular permeability and physicochemical properties Cuticular permeability of M. spicatum is proportional to the lipophilicity of a chemical. This relationship was also identified for terrestrial plant cuticles (Kerler and Schonherr, 1988) which, however, is slightly different from that of M. spicatum, potentially due to species-specific differences of the cuticle structures (Riederer, 2006). Furthermore, the equation established to describe the dependency between cuticular permeability and lipophilicity is based on chemicals with log(Poct/wat) values covering the large range of 0.8 to 8.26. Equations from other authors focus either only on lipophilic chemicals (Kerler and Schonherr, 1988) or use two equations, one describing the dependency for lipophilic, the other one the dependency for hydrophilic chemicals (Popp et al., 2005). It also has to be noted, that in contrast to most other research, where cuticles were isolated from plants, we used a model approach to determine cuticular permeability which might introduce some uncertainty. 6. Conclusion This study demonstrated that the presented mechanistic model approach is suitable to describe toxicokinetics of Isoproturon, Iofensulfuron, Fluridone, Imazamox and Penoxsulam within M. spicatum. Through the established equations that estimate toxicokinetic parameters, that are necessary to run the model, the model can also be used to predict toxicokinetics of other chemicals just with the respective octanol/water partition coefficient. Special toxicokinetic experiments, as conducted within this study, are not mandatory to use the model approach. The model analysis indicates that the speed of the root-to-shoot translocation of chemicals is realistically described by the model approach. Furthermore, a general, model-based method is presented how cuticular permeabilities can be estimated without isolating plant cuticles, which is, at least for some plant species, hardly possible. As a final conclusion, it can be stated, that the model approach can be used as a tool to analyze time-variable chemical exposure. In a future perspective, by integrating toxicodynamics, the model might even be used to predict the effects caused by time-variable exposure of toxicants, such as plant protection products. Acknowledgement This research was funded by Bayer CropScience. References Ashauer, R., Escher, B.I., 2010. Advantages of toxicokinetic and toxicodynamic modelling in aquatic ecotoxicology and risk assessment. J. Environ. Monit. 12, 2056–2061. Ashauer, R., Boxall, A.B.A., Brown, C.D., 2007. New ecotoxicological model to simulate survival of aquatic invertebrates after exposure to fluctuating and sequential pulses of pesticides. Environ. Sci. Technol. 41, 1480–1486. Boyd, C.E., 1968. Fresh-water plants – a potential source of protein. Econ. Bot. 22, 359–367.
Buresova, H., Crum, S.J.H., Belgers, J.D.M., Adriaanse, P.I., Arts, G.H.P., 2013. Effects of linuron on a rooted aquatic macrophyte in sediment-dosed test systems. Environ. Pollut. 175, 117–124. Coleman, J.O.D., Blake-Kalff, M.M.A., Davies, T.G.E., 1997. Detoxification of xenobiotics by plants: chemical modification and vacuolar compartmentation. Trends Plant Sci. 2, 144–151. Fairchild, J.F., Ruessler, D.S., Haverland, P.S., Carlson, A.R., 1997. Comparative sensitivity of Selenastrum capricornutum and Lemna minor to sixteen herbicides. Arch. Environ. Contam. Toxicol. 32, 353–357. Frost-Christensen, H., Bolt Jorgensen, L., Floto, F., 2003. Species specificity of resistance to oxygen diffusion in thin cuticular membranes from amphibious plants. Plant Cell Environ. 26, 561–569. Gobas, F.A.P.C., McNeil, E.J., Lovett-Doust, L., Haffner, G.D., 1991. Bioconcentration of chlorinated aromatic hydrocarbons in aquatic macrophytes. Environ. Sci. Technol. 25, 924–929. Hanson, M.L., Sibley, P.K., Mabury, S.A., Muir, D.C.G., Solomon, K.R., 2003. Field level evaluation and risk assessment of the toxicity of dichloroacetic acid to the aquatic macrophytes Lemna gibba, Myriophyllum spicatum, and Myriophyllum sibiricum. Ecotoxicol. Environ. Saf. 55, 46–63. Heine, S., Schmitt, W., Görlitz, G., Schäffer, A., Preuss, T.G., in press. Effects of light and temperature fluctuations on the growth of Myriophyllum spicatum in toxicity tests - A model based analysis. Environ. Sci. Pollut. R. Hung, H., Lin, T., Chiou, C.T., 2010. Partition coefficients of organic contaminants with carbohydrates. Environ. Sci. Technol. 44, 5430–5436. Jeffree, C.E., 2007. The fine structure of the plant cuticle. In: Riederer, M., Müller, C. (Eds.), Biology of the Plant Cuticle. Annual Plant Reviews, vol. 23. springer, berlin, pp. 11–110. Kemp, W.M., Boynton, W.R., Cunningham, J.J., Stevenson, J.C., Jones, T.W., Means, J.C., 1985. Effects of atrazine and linuron on photosynthesis and growth of the macrophytes Potamogeton perfoliatus L. and Myriophyllum spicatum L. in an estaurine environment. Mar. Environ. Res. 16, 255–280. Kerler, F., Schonherr, J., 1988. Accumulation of lipophilic chemicals in plant cuticles – prediction from octanol water partition-coefficient. Arch. Environ. Con. Toxicol. 17, 1–6. Kulkarni, D., Daniels, B., Preuss, T.G., 2013. Life-stage-dependent sensitivity of the cyclopoid copepod Mesocyclops leuckarti to triphenyltin. Chemosphere 92, 1145–1153. Maletzki, D., Hunken, A., Hubner, I., Kussatz, C., 2010. Mynophyllum spicatum as a test organism development of a sediment free test system and first results with 3,5-dichlorophenol. Umweltwiss. und Schadst.-Forsch. 22, 702–710. MATLAB version 7.14.0.739. 2012. Natick, Massachusetts: The MathWorks Inc. Nielsen, S.L., Sandjensen, K., 1991. Variation in growth rates of submerged rooted macrophytes. Aquat. Bot. 39, 109–120. Pavlostathis, S.G., Comstock, K.K., Jacobson, M.E., Saunders, F.M., 1998. Transformation of 2,4,6-trinitrotoluene by the aquatic plant Myriophyllum spicatum. Environ. Toxicol. Chem. 17, 2266–2273. Popp, C., Burghardt, M., Friedmann, A., Riederer, M., 2005. Characterization of hydrophilic and lipophilic pathways of Hedera helix L. cuticular membranes: permeation of water and uncharged organic compounds. J. Exp. Bot. 56, 2797– 2806. Preuss, T.G., Hammers-Wirtz, M., Hommen, U., Rubach, M.N., Ratte, H.T., 2009. Development and validation of an individual based Daphnia magna population model: the influence of crowding on population dynamics. Ecol. Model. 220, 310–329. Reinert, K.H., Giddings, J.M., Judd, L., 2002. Effects analysis of time-varying or repeated exposures in aquatic ecological risk assessment of agrochemicals. Environ. Toxicol. Chem. 21, 1977–1992. Riederer, M., 2006. Thermodynamics of the water permeability of plant cuticles: characterization of the polar pathway. J. Exp. Bot. 57, 2937–2942. Schmitt, W., 2008. General approach for the calculation of tissue to plasma partition coefficients. Toxicol. In Vitro 22, 457–467. Schmitt, W., Bruns, E., Dollinger, M., Sowig, P., 2013. Mechanistic TK/TD-model simulating the effect of growth inhibitors on Lemna populations. Ecol. Model. 255, 1–10. Schönherr, J., 1976. Water permeability of isolated cuticular Membranes: the effect of cuticular waxes on diffusion of water. Planta (Berl.) 131, 159–164. Schreiber, L., Schonherr, J., 1992. Uptake of organic chemicals in conifer needles: surface adsoprtion and permeability of cuticles. Environ. Sci. Technol. 26, 153– 159. Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M., 2005. Environmental Organic Chemistry. Wiley-Interscience, p. 341. Tuth, H.F., 1932. The movement of water through some submerged plants. Amer. J. Bot. 19, 693–709. Vassios, J.D., 2012. Herbicide absorption and translocation by Eurasian Watermilfoil and hydrilla. PhD thesis. Colorado State University, Colorado.
