The Degree of Synergistic Interactions in Springtails - American

Jul 5, 2014 - •S Supporting Information. ABSTRACT: Organisms in the environment are exposed to multiple stressors. However, for terrestrial inverteb...
0 downloads 22 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Simultaneous Control of Phenanthrene and Drought by Dual Exposure System: The Degree of Synergistic Interactions in Springtails was Exposure Dependent Stine N. Schmidt,†,§ Martin Holmstrup,‡ Christian Damgaard,‡ and Philipp Mayer*,†,§ †

Department of Environmental Science, Faculty of Science and Technology, Aarhus University, Frederiksborgvej 399, 4000 Roskilde, Denmark ‡ Department of Bioscience, Faculty of Science and Technology, Aarhus University, Vejlsøvej 25, 8600 Silkeborg, Denmark S Supporting Information *

ABSTRACT: Organisms in the environment are exposed to multiple stressors. However, for terrestrial invertebrates, it remains difficult to study the effects of combined stressors under welldefined exposure conditions. Thus, the current study develops a new dual exposure system for the simultaneous and independent control of chemical and drought exposure in bioassays with terrestrial organisms: Passive dosing from silicone controlled the chemical activity of phenanthrene (chemical stress), while saline solutions controlled the water activity (drought stress) in the closed exposure system. The dual exposure system was then applied in a full factorial experiment with seven exposure levels (72), which aimed at determining the combined effects of phenanthrene and drought on the survival of the terrestrial springtail Folsomia candida after 7 d exposure. Fitting an “independent action” model to the complete data set revealed statistically significant synergy between phenanthrene and drought (p < 0.0001). However, the degree of synergy was exposure dependent with some synergy at higher and only minor synergy at lower exposure levels. This emphasizes the need for taking exposure levels into account when extrapolating synergy observations from (eco)toxicological studies done at high exposure levels.

1. INTRODUCTION Soil biodiversity is affected by a wide range of physical, chemical, and biological stressors acting simultaneously, such as, extreme temperatures, low humidity levels, nutrient deficiency, anthropogenic pollution, and intra- and interspecies competition for food, water, and space.1−3 In contrast to reallife environmental conditions, many ecotoxicological studies with terrestrial organisms are conducted in the laboratory under “standard” and (near) optimal conditions with regards to nonchemical stressors,4 which is also the case in the context of chemical risk assessments.5 In these cases, possible synergistic or antagonistic interactions between chemical and nonchemical stress are neglected, thereby introducing the possibility of either under- or overestimating the actual risk to organisms in the environment.6,7 Previously, the effects of combined chemical and drought stress have been investigated in several species of terrestrial invertebrates.8−12 In this regard, two experimental challenges exist, namely (1) how to expose the test organism to the test compound and drought simultaneously and independently and (2) how to control the levels of the two stressors accurately and precisely throughout the experiment. These challenges have not been solved in previous studies, where the chemical and © 2014 American Chemical Society

drought exposure to test organisms have been successive rather than simultaneous8−11,13−15 and/or with nonrigorously controlled exposure levels.12,16 Specifically, in terrestrial ecotoxicological studies the total concentration of the test compound(s) in soil is most often known rather than the effective exposure, since the latter depends on a range of system properties, such as the type of soil, soil humidity, and soil temperature.17 Moreover, it is not feasible to accurately control soil humidity by adjusting the soil water content in experiments investigating the effects of drought.12,16 Therefore, the current study first develops and then applies a new dual exposure system for the simultaneous and independent control of chemical and drought exposure in bioassays with terrestrial organisms. The dual exposure system combines two well-established methods, namely passive dosing for controlling the chemical exposure18−20 and saline solutions for controlling the relative humidity (and thus the drought exposure).21−23 In passive dosing of hydrophobic organic compounds (HOCs), a Received: Revised: Accepted: Published: 9737

March 18, 2014 June 27, 2014 July 5, 2014 July 5, 2014 dx.doi.org/10.1021/es501331m | Environ. Sci. Technol. 2014, 48, 9737−9744

Environmental Science & Technology

Article

temperature26 (Tm, K) and the experimental temperature (T, K) according to Yalkowsky et al. (1979),27 assuming the entropy of melting to be 56 J mol−1K−1 (i.e., Walden’s rule);

biocompatible silicone polymer is loaded with the test compound(s) and subsequently applied to establish and maintain the chemical exposure by continuous equilibrium partitioning. In this way, the method provides well-defined and constant chemical exposure during experiments. In the dual exposure system, the chemical exposure is controlled from HOC loaded silicone in passive dosing vials. The test organisms are added to the passive dosing vials, and the vials are covered by nylon net. Then, the vials are placed inside glass jars holding saline solution, which controls the relative humidity in the closed exposure system and thus the drought exposure of the test organisms. Gas exchange between passive dosing vials and glass jars is ensured through the nylon net, and the test organisms are thus simultaneously exposed to the test compound(s) and drought. Applying this dual exposure system, the terrestrial springtail Folsomia candida was simultaneously exposed to the polycyclic aromatic hydrocarbon (PAH) phenanthrene and soil-relevant drought for 7 d in a full factorial experiment with seven exposure levels (72, Figure 1).

amax = e

⎛ T ⎞ 6.8 × ⎜1 − m ⎟ T ⎠ ⎝

(1)

In the current study, drought exposure is expressed by the water activity, which corresponds to the relative humidity. The water activity notation is well-established in food science for quantifying the moisture in food products.28 In addition, water activity is used to quantify water availability in soil sciences and is also used within plant and animal physiology.29 The water activity is also defined between 0 and 1, but in contrast to the chemical activity of phenanthrene, drought stress increases with decreasing water activity. Both passive dosing for controlling chemical activity and saline solutions for controlling water activity have been documented to accurately and precisely control the exposure of the given stressor.18,23,30 Therefore, it is hypothesized that the dual exposure system is able to accurately and precisely control chemical activity and water activity simultaneously and independently throughout a 7-d experiment, which is not possible with previously presented methods. Moreover, the full factorial test design with several exposure levels will provide a sound basis for the data analysis and allow the determination of possible antagonistic or synergistic interactions between phenanthrene and drought. Several studies with terrestrial springtails suggest synergistic interactions between PAHs and drought,10,11,14,15 and it is therefore hypothesized that the combined effects of phenanthrene and drought is also characterized by synergy in the current study.

2. EXPERIMENTAL SECTION 2.1. Materials and Chemicals. The 10 mL clear glass vials and airtight screw caps (with nonfastened Teflon-coated septa) were from Mikrolab Aarhus A/S, Denmark. Medical grade polydimethylsiloxane (PDMS) silicone was prepared using the MDX4-4210 kit from Factor II, Inc., U.S.A. Phenanthrene (≥99.5%) was from Aldrich, U.S.A. The 160 mL amber glass jars and plastic screw lids were from Nordic Pack, Denmark. Sodium chloride (NaCl, ≥99.5%) was from Sigma-Aldrich, Denmark. . Ethanol (96%, Kemetyl A/S, Denmark), Milli-Q water (Super Q treated, Millipore, U.S.A.), methanol (99.9%, Merck, Germany), and lint free tissue (Assistent, Germany) were used as described below. 2.2. Controlling Combined Stressors by Dual Exposure System. A new dual exposure system was developed in order to accurately and precisely control chemical activity (chemical stressor) and water activity (drought stressor) simultaneously and independently in toxicity experiments with terrestrial invertebrates (Figure 1). A full factorial experiment was conducted with seven chemical activities of phenanthrene (including a control treatment) and seven water activities, giving a total of 49 different combinations of the two stressors. Two pilot studies were conducted to determine the water activity levels of the final experiment, and the results were in general agreement with the final experiment (see the Supporting Information, SI). 2.2.1. Phenanthrene Exposure (Inner Vials). Chemical exposure, in terms of the chemical activity of phenanthrene, was controlled from silicone in 10 mL passive dosing vials (aPHE, Figure 1). The air in the vials equilibrated with the silicone and thereby adjusted the chemical activity of the

Figure 1. Conceptual drawing of the new dual exposure system for the simultaneous and independent control of chemical and drought exposure in toxicity experiments with terrestrial invertebrates. The chemical activity of phenanthrene (aPHE) is controlled from phenanthrene loaded silicone in 10 mL passive dosing vials (inner vials), while the water activity (awater) is controlled from saline solution in 160 mL amber glass jars (outer jar). The springtails are indicated on the silicone surface in the inner passive dosing vials, and gas exchange between vials and jars is ensured through nylon net. The arrows indicate equilibrium partitioning.

Then, a modified dose−response function is fitted to the observed survival data from the full factorial experiment, allowing the determination of possible antagonistic or synergistic interactions between phenanthrene and drought.24 Here, phenanthrene exposure is expressed by the chemical activity of the compound (denoted a), which quantifies the energetic level of a hydrophobic organic compound relative to the energetic level of its subcooled liquid (reference state, a = 1).25 The chemical activity is then defined between 0 and 1, with chemical stress increasing with increasing chemical activity.25 The maximum chemical activity (amax, dimensionless) of phenanthrene can be estimated from its melting 9738

dx.doi.org/10.1021/es501331m | Environ. Sci. Technol. 2014, 48, 9737−9744

Environmental Science & Technology

Article

exceeded the volume ratio between the saline solution and the PDMS (17−25 L L−1). Thus, the phenanthrene loaded PDMS controlled the chemical activity of phenanthrene in the entire system, and the phenanthrene loss to the saline solution was negligible. The closed system was pre-equilibrated for approximately 36 h before test start. Age synchronized F. candida were cultured and maintained as has been described earlier.18,30 The springtails had a fresh weight of 234 ± 20 μg and a dry weight of 106 ± 7 μg (46−49 d old, n = 138). To begin the experiment, ten springtails (42− 45 d old) were transferred to each inner passive dosing vial. Nylon nets (with a mesh size of 100 μm) were placed to cover the vial openings and then fixed by the screw caps, from which the Teflon-coated septa had been removed. In this way, the springtails were prevented from leaving the vials while at the same time allowing gaseous exchange between vials and jars (Figure 1). The springtails could move freely within the vials and most springtails were observed on the PDMS surface, leading to phenanthrene exposure through headspace and via direct contact with silicone. The outer glass jars were opened for approximately 1 min at days 2, 4, and 6 in order to prevent oxygen depletion in the system. After 7 d exposure, the springtails were transferred to Petri dishes with a clean and moistened charcoal/plaster-of-Paris mixture and recovered overnight at room temperature. Then, springtail survival was determined with springtails being characterized as living when able to walk in a coordinated manner, if necessary, after gentle stimulation with a fine brush. The experiment was conducted at 20 °C with a 12:12 h light/dark photoperiod, and without feeding the springtails. Their relatively low metabolic rate makes them capable of starving for months,33 and it is therefore unlikely that starvation for 7 d had any significant influence on the results. 2.2.4. Exposure Confirmation and System Response Time. The phenanthrene exposure was confirmed analytically after ending the experiment in order to investigate the ability of the dual exposure system to accurately and precisely control chemical exposure for 7 d. The passive dosing vials were cleaned with small volumes of Milli-Q water and lint free tissues, in order to remove springtail waste. An amount of 1000 μL Milli-Q water was added to each vial, and the vials were closed and stored for at least 24 h for phenanthrene to equilibrate.19 After equilibration, 500 μL water was transferred to a 1.5 mL amber HPLC-vial, and 500 μL methanol was added in order to stabilize the sample. The water/methanol samples were stored at −18 °C until chemical analysis. The samples were analyzed by high pressure liquid chromatography (HPLC, Agilent 1100 Series) with multiband fluorescence detection (G1321A FLD with Ex: 260 nm and Em; 350, 420, 440, and 500 nm). The HPLC was equipped with a CP-EcoSpher PAH column (Varian, Inc.), and operated with a flow rate of 1.000 mL min−1 (30 μL injection, 28 °C). Methanol was used as solvent, and the isocratic method had a run time of 10 min. Phenanthrene concentrations were quantified by external calibration. In this way, the freely dissolved phenanthrene concentration in the equilibrated water was determined in each passive dosing vial (n = 210), and no phenanthrene was detected in controls loaded with methanol (n = 39). Six of the 210 exposure measurements were identified as outliers using the Generalized Extreme Studentized Deviate test.34 Two of these outlier measurements were significantly below the remaining measurements at the given exposure level (p < 0.05), which is likely due to evaporative losses during the

system, with increasing chemical activity inducing increasing chemical stress. The passive dosing vials were made, loaded, and cleaned as earlier described in details,30 and the procedures are therefore only briefly outlined below. In short, 500 ± 5 mg silicone (ρPDMS ≈ 1 kg L−1) was cast and cured in the bottom of each 10 mL glass vial, and the cured PDMS was subsequently cleaned with ethanol and Milli-Q water. A saturated, methanol stock solution was made with phenanthrene, i.e., a solution with the maximum chemical activity of phenanthrene (amax = 0.165 at 21 °C, eq 1). The stock solution was then diluted with methanol to obtain loading solutions with chemical activities of 0.082, 0.041, 0.027, 0.021, 0.010, and 0.005, as these levels were expected to result in springtail lethality from 0 to 100%.30 Loading of the silicone was done by equilibrium partitioning of phenanthrene between 1000 μL loading solution and PDMS in closed passive dosing vials. After at least 48 h, the loading solutions were renewed in order to compensate for the phenanthrene depletion in the first volume of loading solution. After another 48 h, the spent loading solutions were removed, the silicone was cleaned four times with Milli-Q water to quantitatively remove methanol, and then dried with lint free tissue. Each of the six treatments with phenanthrene was done in 35 replicates (five replicates for each of the seven drought levels), while 39 controls were loaded with pure methanol and cleaned as described above (aPHE = 0.000). Loading of the passive dosing vials and storage until use was done at room temperature (approximately 21 °C). 2.2.2. Drought Exposure (Outer Jars). Drought exposure, in terms of water activity, was controlled in the water vapor above saline solutions in 160 mL jars (awater, Figure 1). The air in the system equilibrated with the saline solution and thereby adjusted the water activity of the system, with increasing salinity resulting in decreasing water activity and thereby increasing drought stress.23,31 Saline solutions were prepared to obtain water activities of 0.998, 0.993, 0.988, 0.983, 0.978, 0.973, and 0.968, as these levels were expected to result in springtail lethality from 0 to 100%8,13 (see also SI). The concentrations of the saline solutions were calculated from the following linear correlation (r2 = 1.00, 1.000 ≥ awater ≥ 0.960) between water activity (awater, dimensionless) and sodium chloride concentrations (CNaCl, g L−1) based on data from Weast (1989);32 a water = 1 − 0.000556 × C NaCl

(2)

Earlier, the relationship between sodium chloride concentration and water activity has been shown to be largely independent of temperature within the range 15−50 °C.22 2.2.3. The Dual Exposure System Used in Springtail Experiments. The screw caps were removed from the loaded passive dosing vials, and the dual exposure system was assembled by placing four vials in an empty jar before adding 25 mL saline solution to the bottom of the jar (Figure 1). It was ensured that no saline solution ended up in the passive dosing vials, and the jars were tightly closed with aluminum foil and plastic lids. In this way, the chemical activity of phenanthrene (aPHE) was controlled from silicone in the inner passive dosing vials simultaneously with the water activity (awater) being controlled from the saline solutions in the outer jars (Figure 1). The five replicate vials were distributed between two jars (two and three vials, respectively). Empty glass vials were added to the jars in order to prevent the passive dosing vials from tipping over but otherwise not used in the experiment. The PDMS to water partition ratio of phenanthrene (9338 L L−1)19 greatly 9739

dx.doi.org/10.1021/es501331m | Environ. Sci. Technol. 2014, 48, 9737−9744

Environmental Science & Technology

Article

cellular membranes by narcosis,39 whereas dehydration (if not countered or tolerated) leads to increased concentrations of osmolytes preventing the maintenance of homeostasis.29 In this way, the model is a multiplicative interaction model, assuming a multiplicative probability for survival with regards to chemical stress (C), drought stress (D), and their possible interaction (CD);24

exposure confirmation. Contrarily, four of the outlier measurements significantly exceeded the remaining measurements at the given exposure level (p < 0.05), which is likely due to phenanthrene sorbing to small amounts of dust or biological material, meaning that the measured aqueous concentration exceeded the freely dissolved concentration. In order to compare to estimated exposure levels, the freely dissolved phenanthrene concentrations (Cfree, μg L−1) were translated into chemical activities (a, dimensionless) using the phenanthrene subcooled liquid solubility (SL, μg L−1): a=

Cfree SL

f (x ; b , x0 , λ) = ⎡⎣(1 + e−bCx0,C)(1 + e−bDx0,D) (1 + e−bCDx0,CD)(1 − λ)⎤⎦ −1 × ⎡⎣(1 + ebC(xC − x0,C))(1 + ebD(xD− x0,D))(1 + ebCD(xCxD− x0,CD))⎤⎦

(3)

(5)

The subcooled liquid solubility was estimated as the ratio between the water solubility (μg L−1) at 20 °C35 and the amaxvalue at 21 °C (eq 1). The ability of saline solutions to accurately and precisely control water activity is well-established,21−23 and the water activity was therefore not confirmed by measurements in the current experiment. The accuracy and precision of the method is supported by the fact that saline solutions are used to calibrate hygrometers, which are then used to measure water activity.36,37 To the contrary, it was necessary to determine the time needed to equilibrate the air in the inner passive dosing vials to a given water activity (i.e., the system response time). Thus, a simple system was constructed to mimic the dual exposure system: A humidity sensor with data logger (EL-USB2-LCD+, Lascar Electronics Inc.) measured the relative humidity (i.e., the water activity) in an inner aluminum tube inserted in an aluminum container, holding either pure Milli-Q water (awater = 1.00) or a saturated saline solution (awater ≈ 0.75).38 During the 6-d experiment, humidity was recorded every 5 min. 2.3. Toxicity of Combined Stressors. The experimental design included some degree of pseudoreplication, with 5 replicate passive dosing vials distributed in 2 jars. Thus, for the observed survival data, the average standard deviation for replicates within the same jar was determined (6.6%, 2 or 3 replicates in 1 jar), and compared to the average standard deviation for all replicates within a treatment (8.0%, 5 replicates in 2 jars). This comparison showed that the jar to jar variation only had a secondary contribution to the variation in the lethality data set, and the 5 vials were thus treated as replicates within the subsequent statistical analysis. A modified dose− response function with two stressors was fitted to observed survival data, in order to determine whether the effects of combined chemical and drought stress were characterized by independent effects, antagonism, or synergism. Additionally, the effective activities resulting in 50% lethality for phenanthrene and drought, respectively, were estimated.24 For each stressor, the model is based on a two-parameter sigmoidal dose−response function defined between 0 and 1;24 f (x ; b , x 0 , λ ) =

The model was fitted to the binomial survival data using the maximum likelihood approach in the Mathematica software (Version 9, Wolfram Research, Inc., U.S.A.).24 The robustness of the fitting procedure and the parameter estimates were checked using Bayesian Markov chain Monte Carlo (MCMC) methods.15 The reader is referred to Damgaard et al. (2002) for further description and evaluation of the model.24 Model output was applied to make lethality probability contour plots illustrating springtail lethality as a function of combined chemical and drought stress. Additionally, one plot was made to show the springtail’s tolerance to phenanthrene alone (expressed by the effective lethal chemical activity of phenanthrene causing 50% lethality, LaPHE 50) as a function of decreasing water activity. In the same way, a second plot was made to show the springtail’s tolerance to drought alone (expressed by the effective lethal water activity causing 50% lethality, Lawater 50) as a function of increasing chemical activity of phenanthrene.

3. RESULTS AND DISCUSSION 3.1. Controlling Combined Stressors by Dual Exposure System. Chemical activities, calculated from measured phenanthrene concentrations, were plotted against the water activity in the given treatment, and the results showed high precision in controlled exposure levels (Figure 2). Exposure levels were precisely controlled between replicates (n = 5, e.g., aPHE = 0.005 and awater = 0.998) with relative standard deviations (RSDs) between replicates of 0.7 to 26.0% (average 3.9%) for the entire data set and 0.7 to 4.2% (average 2.3%) when outliers were omitted. More importantly, exposure levels were also precisely controlled between treatments (n = 5 × 7, e.g., aPHE = 0.005 and awater = 0.998−0.968) with RSDs between treatments of 2.6 to 11.0% (average 5.9%) for the entire data set and 2.4 to 3.0% (average 2.7%) when outliers were omitted (see SI Table S1 for the RSDs of each treatment). Finally, the exposure in dilution series (n = 5 × 6, e.g., aPHE = 0.005−0.082 and awater = 0.998) was well produced and controlled by passive dosing (Figure 2). Overall, the phase partitioning of phenanthrene from the PDMS provided a precise exposure to phenanthrene with RSD values for chemical activities in the low percentage range, in accordance with recent research.20,40,41 In addition, the chemical activities were plotted against estimated chemical activities, which were calculated from eq 1 and appropriate dilution factors (SI Figure S1). The chemical activities were moderately higher than the corresponding estimated levels with average differences of 20.4% for the entire data set and 19.9% when outliers were omitted, in agreement with previous results.20,30,41 The analytical measurements further indicated that phenanthrene exposure was kept

(1 + e−bx0)(1 − λ) 1 + eb(x − x0)

(4)

where b is the shape parameter of the function, x0 is the point of inflection at a given stress level (x), and λ is the control/ residual lethality.24 The model is based on the concept of independent action,24 since the effects of phenanthrene and drought are believed to be independent, i.e., the stressors cause different and unrelated injuries to the organism. Thus, phenanthrene perturb the physical properties and function of 9740

dx.doi.org/10.1021/es501331m | Environ. Sci. Technol. 2014, 48, 9737−9744

Environmental Science & Technology

Article

Figure 2. Exposure confirmation of phenanthrene at the end of the 7-d experiment. Chemical activities of phenanthrene (aPHE, final) are plotted against the water activity (awater) in the given treatment, and each symbol represents one replicate vial (n = 210, 6 outliers in red). The average of the measurements (n = 5 × 7) is indicated by a broken line and given to the right of the figure, with estimated levels given in parentheses. The relative standard deviations between treatments (n = 5 × 7, e.g., aPHE = 0.005 and awater = 0.998−0.968) are given in SI Table S1.

constant during the experiment, which is also in accordance with previous results.18,20,30 The system response time, with regards to alterations in water activity, was determined in a 6-d experiment with data shown in SI Figure S2. In turns, the system was equilibrated with pure Milli-Q water and with saturated saline solution, resulting in three alterations in water activity. In each case, the system response time was less than 4 h (SI Figure S2), and preequilibration of the system for approximately 36 h before test start was thus sufficient. The dual exposure system is very flexible when it comes to format and to some degree also when it comes to test organisms and test compounds. Both the dimension of the passive dosing vials and the glass jars with saline solution can be adjusted to meet given requirements. In this way, e.g., the number and types of replicates and the size of the test organism can be easily accommodated. When it comes to test organisms, water activities of the test system can be adjusted to have relevance for a number of terrestrial organisms ranging from hydrophilic soil invertebrates (1.000 > awater > 0.950) to more xerophilic organisms. Thus, it is likely that the system is applicable to a wide variety of terrestrial arthropods and possibly also to worms and snails/slugs in various life stages, although this would need further testing and adjustments. When it comes to test compounds, the passive dosing method for terrestrial organisms has previously been found applicable for hydrophobic organic compounds with octanol to air partition ratios (log Koa-values) of about 5 to 8.5.30 On the one hand, the dual exposure system is artificial due to the lack of, e.g., soil and food. On the other hand, this system provides the opportunity to perform experiments with simultaneously and independently controlled levels of a wide range of hydrophobic organic compounds (even their mixtures) and

drought to various taxonomic groups of terrestrial invertebrates, which is hardly possible in experimental systems containing soil. In this way, the dual exposure system creates a new and necessary foundation for studying the effects of combined chemical and drought stress. 3.2. Toxicity of Combined Stressors. The springtail lethality caused by phenanthrene and drought stress is presented in Figure 3. Survival in the least stressed organisms, exposed to aPHE = 0.000 (controls) and awater = 0.998, was 87.8 ± 9.7% (n = 89 springtails). As expected, springtail lethality increased with increasing chemical activity and decreasing water activity (Figure 3), and full lethality was observed at the highest chemical activity of 0.082 and the lowest water activity of 0.968. The average lethality ± standard deviation of each treatment is given in SI Table S2. Dose−response functions with two stressors (eq 5) were fitted to the observed survival data. Data were fitted (1) to a function with interaction term that accounts for possible antagonism or synergism and (2) to a function without interaction term (simple independent effects). When fitting the functions to the entire 72 data set, the interaction term was found to significantly exceed zero (p < 0.0001), revealing statistically significant synergistic interactions between phenanthrene and drought. For visualization, the synergistic interactions between phenanthrene and drought can be seen when comparing the lethality probability contour plots obtained for data analysis with and without the interaction term of eq 5 (Figure 4). As hypothesized, the identified synergy is in line with previous results from experiments testing the effects of a prior exposure to PAH on drought tolerance in terrestrial springtails.10,11,14,15 The identified synergy was investigated as a function of exposure levels by focusing on the levels of stress causing 50% 9741

dx.doi.org/10.1021/es501331m | Environ. Sci. Technol. 2014, 48, 9737−9744

Environmental Science & Technology

Article

Figure 3. Springtail lethality caused by phenanthrene and drought stress. Each bar represents the average lethality of five replicates (n = 50 springtails). The average lethality ± standard deviation is given in SI Table S2. Seven levels of chemical and drought exposure were included in a full factorial design giving a total of 49 treatments and 249 passive dosing vials, which altogether contained 2484 springtails. aPHE is short for the chemical activity of phenanthrene, while awater is short for the water activity.

Figure 5. (A) The effective lethal chemical activity of phenanthrene resulting in 50% lethality (LaPHE 50) as a function of drought stress for springtails surviving the given drought level. (B) The effective lethal water activity resulting in 50% lethality (Lawater 50) as a function of chemical stress for springtails surviving the given phenanthrene level. With increasing stress, these decreasing lines express the effect of synergy on the effective lethal activities. A horizontal line would represent results characterized by independent effects, while an increasing line would represent results characterized by antagonism. For visualization, the degree of synergy is indicated by shading, and the red lines indicate the limits for sublethal stress conditions (i.e., at aPHE < 0.01 and awater > 0.990 and aPHE < 0.01).

drought stress level (awater = 0.998). This value slightly exceeds a previous result of 0.026 determined for F. candida in a passive dosing experiment without the drought stressor.30 This difference is most likely owing to the longer period for the springtails to recover from sublethal stress in the current study (overnight) compared to the much shorter recovery period of 2 h in the previous study.30 With decreasing water activity (i.e., with increasing drought stress) the LaPHE 50-value decreased for the organisms surviving the given drought stress level, and the model thereby expressed the synergy in the data set (Figure 5A). However, the degree of synergy was clearly exposure dependent with only minor synergy under sublethal drought stress conditions (down to about awater = 0.990),8,13,15,31,42 where LaPHE 50 dropped a maximum of just 14% relative to control conditions at awater = 0.998 (Figure 5A). More pronounced synergy was expressed at higher drought stress levels, but even when applying a wide humidity range the LaPHE 50-values still remained within the expected chemical

Figure 4. Springtail 25%, 50%, and 75% lethality probability contour plots as a function of chemical and drought stress, fitted by the maximum likelihood approach. The full lines illustrate the lethality probability when the interaction term of eq 5 is included in the model (r2 = 0.97), while the broken lines illustrate lethality probability without the interaction term (r2 = 0.94). For visualization, the degree of synergy is expressed by the difference between respective contour lines. The residuals for the two models are given in SI Figure S3. aPHE is short for the chemical activity of phenanthrene, while awater is short for the water activity.

lethality (the middle full line in Figure 4). The fitted model was applied to plot the effective lethal chemical activity of phenanthrene (LaPHE 50) as a function of decreasing water activity (Figure 5A). The LaPHE 50-value was 0.043 at the lowest 9742

dx.doi.org/10.1021/es501331m | Environ. Sci. Technol. 2014, 48, 9737−9744

Environmental Science & Technology

Article

Present Address

activity range of 0.01 to 0.1 reported for the initiation of baseline toxicity.18,19,30 The same conclusion was reached when plotting the effective lethal water activity (Lawater 50) as a function of increasing chemical activity of phenanthrene (Figure 5B). The Lawater 50value was 0.982 at a chemical activity of 0.000 (no chemical stress, Figure 5B). This result is in line with the result of a pilot water activity experiment (Lawater 50 = 0.980, SI) and results of previous studies of 0.970−0.980.8,13,15,31 With increasing chemical activity (i.e., with increasing chemical stress) the Lawater 50-value decreased for the organisms surviving the given chemical stress level (Figure 5B). However, up to the lower chemical activity limit for baseline toxicity (a = 0.01),18,19,30 the Lawater 50 only increased from 0.9819 to 0.9830, which again expressed only a minor degree of synergy at lower exposure levels. At first glance, the current results support the conclusions of two recent reviews emphasizing the need for including nonchemical stressors in ecotoxicological studies or even risk assessments.4,5 In one review, a statistically significant interaction between chemical and nonchemical stressors was found in 62% of 61 studies,5 while another review found that more than 50% of the reviewed studies documented synergistic interactions between stressors, whereas antagonistic effects were reported in fewer cases.4 However, a closer look at the data revealed the degree of synergy to be exposure dependent with only minor synergy under sublethal stress conditions (i.e., at awater > 0.990 and aPHE < 0.01, Figure 5). These results support a recent study by Holmstrup and co-workers (2014)41 reporting independent effects without statistically significant interactions in physiological and molecular responses in springtails when exposed to low levels of phenanthrene and drought (aPHE = 0.010 and awater = 0.988). The exposure dependent synergy, indentified in the current study, emphasizes the need for taking exposure levels into account when extrapolating synergy observations from (eco)toxicological studies done at high exposure levels. In addition, multiple stressor experiments conducted at lower exposure levels could help clarify aspects on exposure dependent interactions. The developed dual exposure system is expected to provide a practical, stable, and precise experimental platform for future ecotoxicological experiments investigating the combined effects of hydrophobic organic compounds and drought.

■

§

Department of Environmental Engineering, DTU Environment, Technical University of Denmark, Miljøvej 113, 2800 Kgs. Lyngby, Denmark Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We greatly thank Karina V. Fisker and Stine Slotsbo for invaluable help during the toxicity experiment. We also thank Margit M. Fernqvist for her guidance and assistance with passive dosing, and Elin Jørgensen and Zdenek Gavor for their guidance and assistance with Folsomia candida. Finally, we thank Britta Munter and Tinna Christensen for assistance with graphic material. This research project was financially supported by the European Commission (OSIRIS, COGE037017), The Danish Council for Independent Research (Contract No. 10-084579), and Unilever UK Central Resources Limited (Contract CH-2013-0093). Additionally, S.N.S. was supported by the PhD research program STAiR.

■

ASSOCIATED CONTENT

* Supporting Information S

Descriptions and results of two pilot studies conducted to determine the water activity levels of the final experiment. Table S1, exposure confirmation of phenanthrene (precision); Figure S1, exposure confirmation of phenanthrene (accuracy); Figure S2, determination of the system response time; Table S2, the average lethality ± standard deviation to Folsomia candida exposed to combined stressors; and Figure S3, the residual plots when fitting modified dose−response functions to the observed survival data. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

(1) Larcher, W. Physiological Plant Ecology: Ecophysiology and Stress Physiology of Functional Groups; Springer: New York, 2003. (2) Bardgett, R. D. The Biology of Soil; Oxford University Press Inc.: New York, 2005. (3) Wardle, D. A. The influence of biotic interactions on soil biodiversity. Ecol. Lett. 2006, 9 (7), 870−886. (4) Holmstrup, M.; Bindesbol, A. M.; Oostingh, G. J.; Duschl, A.; Scheil, V.; Kohler, H. R.; Loureiro, S.; Soares, A.; Ferreira, A. L. G.; Kienle, C.; Gerhardt, A.; Laskowski, R.; Kramarz, P. E.; Bayley, M.; Svendsen, C.; Spurgeon, D. J. Interactions between effects of environmental chemicals and natural stressors: A review. Sci. Total Environ. 2010, 408 (18), 3746−3762. (5) Laskowski, R.; Bednarska, A. J.; Kramarz, P. E.; Loureiro, S.; Scheil, V.; Kudlek, J.; Holmstrup, M. Interactions between toxic chemicals and natural environmental factorsA meta-analysis and case studies. Sci. Total Environ. 2010, 408 (18), 3763−3774. (6) Heugens, E. H. W.; Hendriks, A. J.; Dekker, T.; van Straalen, N. M.; Admiraal, W. A review of the effects of multiple stressors on aquatic organisms and analysis of uncertainty factors for use in risk assessment. Crit. Rev. Toxicol. 2001, 31 (3), 247−284. (7) Stahl, R. G.; Hooper, M. J.; Balbus, J. M.; Clements, W.; Fritz, A.; Gouin, T.; Helm, R.; Hickey, C.; Landis, W.; Moe, S. J. The influence of global climate change on the scientific foundations and applications of Environmental Toxicology and Chemistry: Introduction to a SETAC international workshop. Environ. Toxicol. Chem. 2013, 32 (1), 13−19. (8) Holmstrup, M. Drought tolerance in Folsomia candida Willem (Collemhola) after exposure to sublethal concentrations of three soilpolluting chemicals. Pedobiologia 1997, 41 (4), 361−368. (9) Holmstrup, M.; Petersen, B. F.; Larsen, M. M. Combined effects of copper, desiccation, and frost on the viability of earthworm cocoons. Environ. Toxicol. Chem. 1998, 17 (5), 897−901. (10) Sjursen, H.; Sverdrup, L. E.; Krogh, P. H. Effects of polycyclic aromatic compounds on the drought tolerance of Folsomia f imetaria (collembola, isotomidae). Environ. Toxicol. Chem. 2001, 20 (12), 2899−2902. (11) Sjursen, H.; Holmstrup, M. Cold and drought stress in combination with pyrene exposure: Studies with Protaphorura armata (Collembola: Onychiuridae). Ecotox. Environ. Safe. 2004, 57 (2), 145− 152. (12) Long, S. M.; Reichenberg, F.; Lister, L. J.; Hankard, P. K.; Townsend, J.; Mayer, P.; Wright, J.; Holmstrup, M.; Svendsen, C.; Spurgeon, D. J. Combined chemical (fluoranthene) and drought effects on Lumbricus rubellus demonstrate the applicability of the

AUTHOR INFORMATION

Corresponding Author

*Phone: +45 45251569; e-mail: [email protected]. 9743

dx.doi.org/10.1021/es501331m | Environ. Sci. Technol. 2014, 48, 9737−9744

Environmental Science & Technology

Article

independent action model for multiple stressor assessment. Environ. Toxicol. Chem. 2009, 28 (3), 629−636. (13) Hojer, R.; Bayley, M.; Damgaard, C. F.; Holmstrup, M. Stress synergy between drought and a common environmental contaminant: Studies with the collembolan Folsomia candida. Glob. Change Biol. 2001, 7 (4), 485−494. (14) Sorensen, T. S.; Holmstrup, M. A comparative analysis of the toxicity of eight common soil contaminants and their effects on drought tolerance in the collembolan Folsomia candida. Ecotox. Environ. Safe. 2005, 60 (2), 132−139. (15) Skovlund, G.; Damgaard, C.; Bayley, M.; Holmstrup, M. Does lipophilicity of toxic compounds determine effects on drought tolerance of the soil collembolan Folsomia candida? Environ. Pollut. 2006, 144 (3), 808−815. (16) Friis, K.; Damgaard, C.; Holmstrup, M. Sublethal soil copper concentrations increase mortality in the earthworm Aporrectodea caliginosa during drought. Ecotox. Environ. Safe. 2004, 57 (1), 65−73. (17) Hurdzan, C. M.; Lanno, R. P. Determining exposure dose in soil: The effect of modifying factors on chlorinated benzene toxicity to earthworms. Chemosphere. 2009, 76 (7), 946−951. (18) Mayer, P.; Holmstrup, M. Passive dosing of soil invertebrates with polycyclic aromatic hydrocarbons: Limited chemical activity explains toxicity cutoff. Environ. Sci. Technol. 2008, 42 (19), 7516− 7521. (19) Smith, K. E. C.; Dom, N.; Blust, R.; Mayer, P. Controlling and maintaining exposure of hydrophobic organic compounds in aquatic toxicity tests by passive dosing. Aquat. Toxicol. 2010, 98 (1), 15−24. (20) Schmidt, S. N.; Holmstrup, M.; Smith, K. E. C.; Mayer, P. Passive dosing of polycyclic aromatic hydrocarbon (PAH) mixtures to terrestrial springtails: Linking mixture toxicity to chemical activities, equilibrium lipid concentrations, and toxic units. Environ. Sci. Technol. 2013, 47 (13), 7020−7027. (21) Lang, A. R. G. Osmotic coefficients and water potentials of sodium chloride solutions from 0 to 40°C. Aust. J. Chem. 1967, 20 (9), 2017−2023. (22) Chirife, J.; Resnik, S. L. Unsaturated solutions of sodium chloride as reference sources of water activity at various temperatures. J. Food Sci. 1984, 49 (6), 1486−1488. (23) Holmstrup, M.; Westh, P. Effects of dehydration on water relations and survival of lumbricid earthworm worm egg capsules. J. Comp. Physiol. B-Biochem. Syst. Environ. Physiol. 1995, 165 (5), 377− 383. (24) Damgaard, C.; Hojer, R.; Bayley, M.; Scott-Fordsmand, J. J.; Holmstrup, M. Dose-response curve modeling of excess mortality caused by two forms of stress. Environ. Ecol. Stat. 2002, 9 (2), 195− 200. (25) Reichenberg, F.; Mayer, P. Two complementary sides of bioavailability: Accessibility and chemical activity of organic contaminants in sediments and soils. Environ. Toxicol. Chem. 2006, 25 (5), 1239−1245. (26) Lide, D. R. Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 2003. (27) Yalkowsky, S. H.; Orr, R. J.; Valvani, S. C. Solubility and partitioning. 3. The solubility of halobenzenes in water. Ind. Eng. Chem. Fundam. 1979, 18 (4), 351−353. (28) Grant, W. D. Life at low water activity. Philos. Trans. R. Soc. London Ser. B-Biol. Sci. 2004, 359 (1448), 1249−1266. (29) Hadley, N. Water Relations of Terrestrial Arthropods; Academic Press: San Diego, CA, 1994. (30) Schmidt, S. N.; Smith, K. E. C.; Holmstrup, M.; Mayer, P. Uptake and toxicity of polycyclic aromatic hydrocarbons in terrestrial springtailsStudying bioconcentration kinetics and linking toxicity to chemical activity. Environ. Toxicol. Chem. 2013, 32 (2), 361−369. (31) Hilligsoe, H.; Holmstrup, M. Effects of starvation and body mass on drought tolerance in the soil collembolan Folsomia candida. J. Insect Physiol. 2003, 49 (1), 99−104. (32) Weast, R. C. Handbook of Chemistry and Physics; CRC Press: Cleveland, OH, 1989.

(33) Hopkin, S. P. Biology of the Springtails; Oxford University Press: Oxford, 1997. (34) Rosner, B. Percentage points for a generalized ESD many-outlier procedure. Technometrics 1983, 25 (2), 165−172. (35) Shaw, D. G. IUPAC Solubility Data Series. Hydrocarbons (C8− C36) with Water and Seawater; Pergamon Press: Oxford, U.K., 1989. (36) Brown, R. W.; Oosterhuis, D. M. Measuring plant and soil water potentials with thermocouple psychrometers: Some concerns. Agron. J. 1992, 84 (1), 78−86. (37) Wescor. Instruction/Service Manual HR-33T Dewpoint Microvoltmeter; Wescor Inc.: Logan, 1986. (38) Greenspan, L. Humidity fixed points of binary saturated aqueous solutions. J. Res. Nat. Bur. StandA Phys. Chem. 1977, 81 (1), 89−96. (39) van Wezel, A. P.; Opperhuizen, A. Narcosis due to environmental pollutants in aquatic organisms: Residue-based toxicity, mechanisms, and membrane burdens. Crit. Rev. Toxicol. 1995, 25 (3), 255−279. (40) Gouliarmou, V.; Smith, K. E. C.; de Jonge, L. W.; Mayer, P. Measuring binding and speciation of hydrophobic organic chemicals at controlled freely dissolved concentrations and without phase separation. Anal. Chem. 2012, 84 (3), 1601−1608. (41) Holmstrup, M.; Slotsbo, S.; Schmidt, S. N.; Mayer, P.; Damgaard, C.; Sorensen, J. G. Physiological and molecular responses of springtails exposed to phenanthrene and drought. Environ. Pollut. 2014, 184, 370−376. (42) Sjursen, H.; Bayley, M.; Holmstrup, M. Enhanced drought tolerance of a soil-dwelling springtail by pre-acclimation to a mild drought stress. J. Insect Physiol. 2001, 47 (9), 1021−1027.

9744

dx.doi.org/10.1021/es501331m | Environ. Sci. Technol. 2014, 48, 9737−9744