By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Comparing the effectiveness of chronic water column tests with the crustaceans Hyalella azteca (order: Amphipoda) and Ceriodaphnia dubia (order: Cladocera) in detecting toxicity of current-use insecticides
Linda A. Deanovic,
School of Veterinary Medicine, Department of Anatomy, Physiology, and Cell Biology, University of California, Davis, California, USA
Since the 1980s, standardized toxicity tests using the water flea Ceriodaphnia dubia has been used for ambient surface water monitoring in California and elsewhere. Data obtained with these tests have provided valuable information on water quality in freshwater environments 1, 2. Insecticides, in particular the organophosphates (OPs) diazinon and chlorpyrifos 2, 3, and more recently, pyrethroid insecticides have been identified as some of the dominant water pollutants 4.
While the chronic C. dubia water column test is generally considered the most sensitive method for detecting insecticide toxicity 1, conductivity in estuaries as well as some of California's more saline inland surface waters—such as the Colorado River, the New River, and the Alamo River—frequently exceeds this species' tolerance limit of 2,000 µS/cm (approximately 1 ppt) 5, 6 for dissolved salt concentrations. The U.S. Environmental Protection Agency (U.S. EPA) 5–7 recommends that test organisms other than standard freshwater species be considered for toxicity testing when salinity of water samples exceeds 2,000 µS/cm. For standard tests using marine species, a minimum salinity of 5 ppt is recommended. This leaves a significant gap (1–5 ppt), where few suitable standard water column toxicity tests exist.
The epibenthic amphipod species Hyalella azteca is euryhaline with a salinity tolerance range of approximately 0 to 15 ppt 8, thus bridging the gap between the U.S. EPA recommended freshwater and marine test species. In addition, H. azteca has been shown to be more sensitive than C. dubia to some insecticides 9–11. While standard protocols exist for a 10-d sediment test and a 96-h water column reference toxicant test 8, a water column test with acute and chronic end points has only recently been implemented and applied in California (State of California Surface Water Ambient Monitoring Program, 2008, http://www.waterboards.ca.gov/water_issues/programs/swamp/) 12. Before this method can be applied in routine water-quality monitoring, however, a detailed analysis is needed to establish the test's effectiveness in detecting water column toxicity. Test effectiveness is dependent on both the sensitivity of the test species as well as the variability of the end point measured. If comparable, the H. azteca 10-d water column test with mortality and growth as endpoints could be a valuable alternative for evaluating water column toxicity in water bodies where conductivity exceeds the tolerance limit of C. dubia.
The objective of the present study was to compare the effectiveness of the C. dubia and H. azteca water column tests for detecting toxicity due to four current-use insecticides: the OPs diazinon and chlorpyrifos and the pyrethroids bifenthrin and cyfluthrin. These insecticides have been found to be among the dominant toxicants in urban and agricultural runoff and sediments of California and elsewhere 2, 3, 13–15. In addition, tests were performed with synthetic control water as well as nontoxic ambient water samples to determine if matrix differences due to dissolved organic carbon (DOC) or salt composition would notably change effect levels of these relatively hydrophobic chemicals. Minimum significant differences (MSDs) were calculated to determine the statistical robustness of test methods and end points.
MATERIALS AND METHODS
Chronic toxicity tests with C. dubia were performed according to U.S. EPA protocols 6, with test duration ranging from 6 to 8 d depending on the time required to produce a third brood. We obtained C. dubia from cultures maintained in-house at the University of California, Davis, Aquatic Toxicology Laboratory according to U.S. EPA 6. The control water consisted of drinking water (DS Waters of America) amended with dry salts to moderately hard specifications. Test end points were mortality and reproduction.
Hyalella azteca were purchased from Aquatic Research Organisms. Immediately upon receipt, amphipods were moved to 10-L aquaria, fed, and acclimated to laboratory test conditions for 48 h. The 10-d testing procedure was based on protocols described in the Quality Assurance Program Plan for the California Surface Water Ambient Monitoring Program (2008, http://www.waterboards.ca.gov/water_issues/programs/swamp/) and the H. azteca sediment toxicity test protocol 8. At test initiation and on renewal days, water was warmed to test temperature (23 ± 1°C) in 600-ml beakers using a water bath and aerated at a rate of 100 bubbles/min until the dissolved oxygen (DO) concentration was 4.9 to 8.9 mg/L. Deionized water amended to moderately hard specifications 8 was used for controls. Tests were initiated with 9- to 14-d-old H. azteca. Each of four replicate 250-ml glass beakers contained 100 ml of water, a small piece of Nitex screen (approximately 4 cm2), and 10 organisms. Tests were conducted at a 16:8 h light:dark photoperiod. Mortality was recorded daily, and 80% of test water was renewed on days 2, 4, 6, and 8. Animals were fed a mixture of yeast, organic alfalfa, and trout chow (1 ml per replicate) at test initiation and every other day after water renewals. On day 10, the surviving H. azteca were dried to a constant weight at 103 to 105°C and weighed using a Mettler AE 163 balance.
Water-quality measurements—including pH, electrical conductivity, DO, and temperature—were recorded for all treatments at test initiation and termination. In addition, DO was measured on renewal water; pH and DO were measured on exposure water prior to renewal every 24 h (C. dubia) or 48 h (H. azteca). Conductivity and DO were measured using YSI 85 and a YSI 30 meters, respectively; temperature and pH were measured with a Beckman 255 pH meter. All meters were calibrated according to the manufacturer's instructions each day of use. Throughout the tests, water-quality parameters were within the physiologically optimal ranges of the test organisms.
Analytical-grade diazinon (CAS-No. 333-41-5, 99.5% purity), chlorpyrifos (CAS-No. 2921-88-2, 99.5% purity), bifenthrin (CAS-No. 82657-04-3, 99% purity), and cyfluthrin (CAS-No. 68359-37-5, 98% purity) were obtained from ChemService and spiked into two types of dilution water: (1) a moderately hard synthetic control water and (2) previously tested, nontoxic, ambient water from the Sacramento–San Joaquin delta collected near Byron, California, USA. Ambient samples were filtered prior to testing using a 1.0-µm pore filter capsule (Whatman, GE Healthcare). Both types of water were amended to a specific conductivity of 900 µS/cm using Instant Ocean (Aquarium Systems) and a pH of 7.9 using HCl or NaOH for test conditions that would be nonstressful to both species. Organisms were exposed to a minimum of five concentrations of each chemical, a solvent control, and a negative control. Pesticide-grade methanol was used as a carrier for the insecticides and did not exceed a final concentration of 0.05%.
Test data were analyzed using U.S. EPA standard multiple-concentration statistical protocols 5, 7. Fisher's exact test was used to detect differences in mortality in the C. dubia chronic test. Before analysis H. azteca mortality was arcsine square root–transformed. We examined C. dubia reproduction, H. azteca mortality, and H. azteca growth by homoschedastic (same variance) t tests, heteroschedastic (separate variance) t tests, or Wilcoxon tests, depending on the normality of the distributions and the homogeneity of variances. Lethal and sublethal effect concentrations were calculated using linear regression, nonlinear regression, or linear interpolation methods. Values for median lethal concentration (LC50) and effective concentration to produce lethality in 25% of the population (EC25) were calculated using CETIS v.1.1.2 (Tidepool Scientific Software). No observed effect concentration (NOEC) and lowest observed effect concentration (LOEC) values were calculated using U.S. EPA standard statistical protocols.
Variability of test end points
Fisher's exact test MSDs were calculated as the smallest difference in mortality between test sample and control that is significant given the sample size. The MSDs of homoschedastic and heteroschedastic t tests were determined 16, 17. Wilcoxon test MSDs were calculated based on the method proposed by Van der Hoeven 18. This method indicates the minimum difference between the means that would result in a significant Wilcoxon test result given the distribution of the data within each treatment. The MSDs of sublethal end points (fecundity, growth) were converted to percent MSDs (PMSDs) by expressing the MSD as a percentage of the mean control performance.
The MSD and PMSD values were calculated using data from previous monitoring studies performed in 2008 and 2009 at the University of California, Davis, Aquatic Toxicology Laboratory. For H. azteca, data from 30 (<2,000 µS/cm) or 38 (>2,000 µS/cm) tests were randomly selected from the results of a large-scale monitoring project in the Sacramento–San Joaquin Delta, California 12. For C. dubia, data from 32 tests were randomly selected from projects performed for the California Surface Water Ambient Monitoring Program (http://www.waterboards.ca.gov/water_issues/programs/swamp/). All tests met U.S. EPA acceptability criteria. Selected data included toxic and nontoxic samples whose conductivity fell within the optimal tolerance range of each test species. For C. dubia tests the optimal range was ≤2,000 µS/cm (approximately ≤1 ppt). Since this was not known for H. azteca, two sets of data were analyzed: (1) ≤2,000 µS/cm, to examine the performance of the H. azteca bioassay in the conductivity range suited for C. dubia, and (2) ≤10,000 µS/cm, reflective of the conductivity range where H. azteca shows no detectable decrease in survival or weight 12.
Samples for analytical chemistry were prepared identically to and simultaneously with test solutions at test initiation. They were preserved with dichloromethane (final concentration 0.1%) and stored in amber glass bottles in the dark at 4°C until successful completion of the test. For H. azteca tests, samples from both test series were analyzed; for C. dubia, only laboratory water samples were analyzed. Total aqueous pesticide and DOC concentrations were determined at the California Department of Fish and Game Water Pollution Control Laboratory. Pesticide samples were extracted within 7 d of sample receipt. Water extractions followed U.S. EPA Method 3510C (separatory funnel liquid–liquid extraction). One-liter samples were fortified with the surrogates triphenyl phosphate and dibromo-octaflurobiphenyl to monitor extraction efficiency and extracted twice with dichloromethane using a mechanical rotating extractor. Extracted materials were dried using sodium sulfate, concentrated and solvent-exchanged with petroleum ether using Kuderna-Danish evaporative glass water equipped with a three-ball Snyder column followed with a micro-Snyder apparatus, and adjusted to a final volume of 2 ml in iso-octane. Final extracts were analyzed for OP pesticides using U.S. EPA Method 8141B and for pyrethroids using U.S. EPA Method 8081B. The OP pesticides were analyzed using dual-column, high-resolution gas chromatography with flame photometric detectors in phosphorous mode. Pyrethroids were analyzed using dual-column, high-resolution gas chromatography equipped with an electron capture detector. Organic carbon samples were analyzed according to U.S. EPA Test Method 415.1.
Quality assurance/quality control
Acceptability criteria for the chronic C. dubia toxicity test requires 80% or greater average control survival, with at least 60% of surviving females having an average of 15 neonates and three broods. Acceptability criteria for the chronic H. azteca 10-d water column toxicity test requires 90% or greater control survival, with measurable weight of animals at the end of the test compared to the start of the test. To evaluate whether organism sensitivity was consistent throughout the project period, positive control reference toxicant tests were performed once a month using NaCl as the toxicant. Any deviations from test protocols were recorded. The following deviation occurred: the control treatment of the C. dubia test conducted with cyfluthrin did not produce the minimum number of offspring that the U.S. EPA requires for test acceptability. This test produced an average of only 11.4 neonates/surviving female in synthetic control water.
The H. azteca test was far more sensitive to pyrethroid insecticides than the C. dubia test for both mortality and sublethal end points (Tables 1 and 2 and Figure 1; Supplemental Data Table S1). The 10-d LC50 values for H. azteca (1.74 and 1.86 ng/L in synthetic and ambient water, respectively) for cyfluthrin were approximately 400 times lower than those for C. dubia (703.8 and 712.3 ng/L). Bifenthrin LC50 values for H. azteca (2.7 and 2.3 ng/L) were approximately 115 times lower than those for C. dubia (344.9 and 266.1 ng/L). Conversely, for the OP insecticides, the C. dubia test was more sensitive, with the 7-d LC50 being 18 (diazinon) and 4.4 (chlorpyrifos) times lower than the respective 10-d LC50 for H. azteca. Sublethal endpoints reflected these sensitivity differences between the two species; however, the EC25 for H. azteca growth could not be calculated for chlorpyrifos and cyfluthrin (Table 2). This was due to either significant mortality or test concentrations being too low to obtain sufficient data points for an estimate. For bifenthrin the EC25 for H. azteca growth was approximately 200 to 464 times lower than that for C. dubia fecundity, while for diazinon the EC25 for C. dubia fecundity was 7 to >15 times lower than that for H. azteca growth. Acute-to-chronic ratios, calculated as the ratio of the LC50 to chronic NOEC values 28, ranged from >1 to >6.6 with a mean of 1.9 for C. dubia and from <1 to 7.5 with a mean of 3 for H. azteca (Table 2).
Table 1. Lethal toxicity values of organophosphate (diazinon, chlorpyrifos) and pyrethroid (bifenthrin, cyfluthrin) insecticides to Ceriodaphnia dubia and Hyalella azteca
Lethal test end points and effect concentrations (ng/l)
C. dubia: 7-d mortality
H. azteca: 10-d mortality
NOEC = no observed effect concentration; LOEC = lowest observed effect concentration; LC50 = median lethal concentration; CI = confidence interval.
Filtered ambient water
Table 2. Sublethal effect concentrations of organophosphate (diazinon, chlorpyrifos) and pyrethroid (bifenthrin, cyfluthrin) insecticides to Ceriodaphnia dubia and Hyalella azteca
Sublethal test endpoints and effect concentrations (ng/l)
C. dubia: 7-d fecundity
H. azteca: 10-d growth
The C. dubia control did not meet the U.S. EPA's minimum reproductive requirement for test acceptability.
NOEC = no observed effect concentration; LOEC = lowest observed effect concentration; EC25 = observed effect concentration for 25%; CI = confidence interval; ACR = acute-to-chronic ratio .
The water matrix, synthetic or filtered ambient water, had little or no effect on LC50 values; however, sublethal test end points were more strongly influenced than mortality. Measured concentrations of the most important factor influencing bioavailability, DOC, were <1.00 to 1.04 mg/L in synthetic water and 6.19 to 7.61 mg/L in filtered ambient water. On average, ratios of effect concentrations derived in synthetic water versus filtered ambient water were 1.12 (mortality) and 1.43 (reproduction) for C. dubia and 1.05 (mortality) and 0.98 (growth) for H. azteca, suggesting that there was no matrix effect on toxicity despite the difference in DOC concentrations.
Variability of test end points
The statistical robustness of test methods and end points delineate as follows (Table 3). For tests performed at a conductivity of <2,000 µS/cm, H. azteca mortality was the most robust endpoint (MSD = 13.1%), followed by C. dubia reproduction (PMSD = 17.7%), C. dubia mortality (MSD = 42.2%) and H. azteca growth (PMSD = 59.8%). For H. azteca tests performed at 2,000 to 10,000 µS/cm, MSDs were nearly identical to those performed at <2,000 µS/cm. The MSD for C. dubia 7-d mortality (42.2%), one of the most commonly used end points in toxicity testing, was approximately three times higher than that of H. azteca mortality (13.1%). Even when C. dubia control survival was 100%, survival in contaminant treatments had to be reduced to about 60% or less to be significantly different. Thus, H. azteca mortality was a highly sensitive end point in tests conducted at low (<2,000 µS/cm) as well as medium (<10,000 µS/cm) conductivity and comparable to the PMSD of the C. dubia reproduction end point (17.7%). Both the C. dubia chronic mortality and the H. azteca weight end points (PMSD >40%) were less effective at detecting mild to moderate toxicity.
Table 3. Minimum significant differences of toxicity end points in Ceriodaphnia dubia 7-d and Hyalella azteca 10-d tests
MSD = minimum significant differences; PMSD = percent minimum significant differences. MSD in (%) for mortality end point and as PMSD (% of control performance) for sublethal end points.
C. dubia (≤2 mS/cm)
Heteroschedastic t test
Homoschedastic t test
H. azteca (<10 mS/cm)
Homoschedastic t test
Homoschedastic t test
H. azteca (<2 mS/cm)
Homoschedastic t test
Homoschedastic t test
Our data show that toxicity tests with two commonly used aquatic invertebrate species differ significantly in their effectiveness to detect two groups of insecticides, OPs and pyrethroids. Tests with the amphipod H. azteca were several hundred–fold more sensitive to pyrethroids than those with the water flea C. dubia, while C. dubia tests were up to 20-fold more sensitive to OPs. This difference may be due, in part, to differences in xenobiotic metabolism and nerve structure between the so-called higher crustaceans (Malacostraca), including amphipods such as H. azteca, and the lower crustaceans, among them the Branchiopoda, which include C. dubia19. Concentrations of 6.19 to 7.61 mg DOC/L in ambient water did not alter the toxicity of insecticides tested, even though the pyrethroids cyfluthrin and bifenthrin adsorb readily to organic carbon (log KOC = 5.1 and 5.4) 20. The results demonstrate that the choice of toxicity test, especially with respect to test species, can be crucially important for the detection of insecticide toxicity in surface water samples. With the increased use of pyrethroid pesticides in both urban and agricultural areas 4, 15, the exclusive use of C. dubia as a test species may therefore result in an underestimation of toxicity in freshwater environments.
Our findings confirm existing information on the sensitivity of the two test species for insecticides. Yang et al. 21 reported C. dubia 96-h LC50 values for cyfluthrin ranging from 93 to 210 ng/L. These concentrations are lower than the 7-d C. dubia LC50 concentrations determined in the present study but still much higher than the respective H. azteca LC50 values. Liu and Gan 22 reported a 96-h C. dubia LC50 for mixed isomers of cis-bifenthrin of 144 ng/L, which is approximately half of the 7-d LC50 determined for bifenthrin in the present study. Werner and Moran 23 summarized the effects of pyrethroid insecticides on numerous aquatic organisms and found that the pyrethroid sensitivity of H. azteca frequently exceeded that of C. dubia by more than an order of magnitude. Amphipods were sensitive to repeated application of 0.1 to 1 toxic units (the ratio of insecticide concentration and LC50) of pyrethroids in microcosm and mesocosm studies in 100% of the studies. Water fleas were impacted in only 50% of such studies 24. In contrast, C. dubia were more sensitive to the OP insecticides. In the present study, LC50 values for diazinon were approximately 20 times lower than those of H. azteca and EC25 values for C. dubia reproduction were approximately eight times lower than the EC25 for H. azteca growth. The results agree with available literature values. Reported 48-h LC50 values for diazinon were 0.92/0.49 and 15.07/22 µg/L for C. dubia and H. azteca, respectively 10, 25. For chlorpyrifos, the difference in sensitivity between the two species was smaller. Values reported elsewhere even suggest that H. azteca is more sensitive than C. dubia, with 96-h LC50 values of 0.04 µg/L for H. azteca9 and 0.05 and 0.06 µg/L for C. dubia25, 26. These data suggest that the two species may be similarly sensitive to chlorpyrifos.
To detect toxicity in ambient water samples, test end points must be statistically robust and sensitive. These traits largely depend on data variability, often expressed as MSD or PMSD. The MSD for mortality in the C. dubia 7-d chronic test was surprisingly high, which can be explained by the test protocol; this is unusual among commonly used aquatic toxicology tests in that every replicate includes only one animal. The mortality data generated by this test are therefore categorical in nature (each replicate was recorded as either alive or dead), and differences between treatments are evaluated using Fisher's exact test rather than the t tests and the Wilcoxon test used to analyze the results of other end points. In contrast, replicates in the H. azteca 10-d water column test contain 10 individuals each, and mortality data for each replicate are recorded as a percentage, making analysis by t test or Wilcoxon test appropriate. Mortality data for H. azteca were often nonnormally distributed and, therefore, usually analyzed using Wilcoxon tests. With mean MSDs of 13.1% (t tests) and 12.4% (Wilcoxon tests), the 10-d mortality end point of H. azteca was the most robust, least variable end point tested in the present study. On the contrary, H. azteca growth as a chronic, sublethal end point was far more variable (PMSD = > 60%) than C. dubia fecundity (PMSD = 17.7%). The upper bound of PMSD acceptability suggested by the U.S. EPA depends on the test endpoint but in no case exceeds 47%; thus, H. azteca growth is currently not a suitable toxicity endpoint. High variability and PMSD of growth measurements may be due to the variable size (7–14 d old) of organisms at the beginning of the test. Use of a more homogeneous organism age and size group could improve the sensitivity of this end point.
Based on the results of our study, we suggest that H. azteca is a suitable test species for fresh- to brackish-water samples (<2000–10,000 µS/cm) and that the 10-d test protocol presented in the present study is well suited for routine ambient water-quality monitoring. The ability to perform toxicity tests with a single species over a broad range of salinities is particularly valuable in estuarine systems or brackish surface waters. The present study results highlight the importance of carefully selecting appropriate test species and toxicity end points as test effectiveness is both species-dependent and end point–dependent. It may even be advisable to use multiple invertebrate tests given the frequency of different chemical classes in mixture found in surface waters. As shown here, environmental risk by pyrethroids may be seriously underestimated if surface water bodies are solely monitored using the C. dubia test. In addition, using appropriate tests can enable more timely initiation of further testing (e.g., toxicity identification evaluations), thus minimizing problems associated with chemical loss in samples during storage and testing 27. The 10-d H. azteca water column test detects toxicity of pyrethroid insecticides at much lower concentrations than the chronic C. dubia test; however, the latter is more effective in detecting toxicity due to OP insecticides. Even though validating and standardizing the 10-d water column H. azteca test protocol will require more effort, monitoring programs for surface water toxicity can benefit from the greater sensitivity of H. azteca, especially in samples where conductivity is above 2,000 µS/cm.
Table S1. (70 KB PDF).
We thank the staff of the University of California, Davis, Aquatic Toxicology Laboratory for conducting the toxicity tests and the California Department of Fish and Game Water Pollution Control Laboratory for conducting chemical analyses. Funding was provided by the Interagency Ecological Program, Sacramento, California (contract 4600008070 to I. Werner) and the Central Valley Regional Water Quality Control Board (contract 06-262-150 to I. Werner).