Evaluation of modes of action of pesticides to Daphnia magna based on QSAR, excess toxicity and critical body residues.

Agricultural pesticides serve as effective controls of unwanted weeds and pests. However, these same chemicals can exert toxic effects in non-target organisms. To determine chemical modes of action, the toxicity ratio (TR) and critical body residues (CBRs) of 57 pesticides were calculated for Daphnia magna. Results showed that the CBR values of inert compounds were close to a constant while the CBR values of pesticides varied over a wider range. Although herbicides are categorized as specifically-acting compounds to plants, herbicides did not exhibit excess toxicity to Daphnia magna and were categorized as inert compounds with an average logTR = 0.41, which was less than a threshold of one. Conversely, fungicides and insecticides exhibited strong potential for toxic effects to Daphnia magna with an average logTR >2. Many of these chemicals act via disruption of the nervous, respiratory, or reproductive system, with high ligand-receptor binding activity which leads to higher toxicity for Daphnia magna. Molecular docking using acetylcholinesterase revealed that fungicides and insecticides bind more easily with the biological macromolecule when compared with inert compounds. Quantitative structure-activity relationship (QSAR) analysis revealed that the toxicity of fungicides was mainly dependent upon the heat of formation and polar surface area, while the toxicity of insecticides was more related to hydrogen-bond properties. This comprehensive analysis reveals that there are specific differences in toxic mechanisms between fungicides and insecticides. These results are useful for determining relative risk associated with pesticide exposure to aquatic crustaceans, such as Daphnia magna.

Relationship between lethal toxicity in oral administration and injection to mice: effect of exposure routes.

The lethal toxicity (LD50) in oral administration, intravenous, intraperitoneal, intramuscular and subcutaneous injections were used to investigate relationships of log 1/LD50 from different exposure routes. Regression analysis showed that log 1/LD50 in oral route was related to the toxicity in injection route. This relationship in lethality between the two routes is apparently due to the same mechanisms of the compounds to the same species. However, the scatter in the correlation curve indicates that exposure route is an important factor that influences the relationship. Some compounds with low intestinal absorption exhibit much less toxicity in oral administration than that in the injection route. A systemic bias of log 1/LD50 between oral and injection routes indicates that tissue distribution of compounds between blood and target site is a very rapid process, leading to log 1/LD50 in injection greater than those in oral administration. Although compounds can be metabolized in the body both from oral and injection routes, first-pass metabolism occurs in oral route but not in injection route. This will result in decrease of toxicity in oral route for most compounds as compared with injection route. In addition, experimental uncertainty, differences in gender, and species can also affect relationships of log1/LD50 between exposure routes.

Theoretical consideration on the prediction of in vivo toxicity from in vitro toxicity: Effect of bio-uptake equilibrium, kinetics and mode of action.

Although in vitro assay is an ideal alternative method for the in vivo toxicity prediction, different in vivo-in vitro correlations have been observed for the toxicity endpoints obtained from different levels of species. In this paper, theoretical in vivo-in vitro toxicity correlations have been developed for cytotoxicity versus human, mammalian and fish toxicity, respectively. These theoretical models were then used to investigate the correlations and the influencing factors between in vivo and in vitro toxicity. Bio-uptake equilibrium theory can well explain why there is a significant correlation between fish and cell toxicity (R2 = 0.70); why human toxicity is very close to fish toxicity; and why hydrophobic compounds exhibit relatively greater toxicity than reactive or specifically-acting compounds to human and fish as compared to cells. The kinetic theory can well explain why there is a very poor relationship between mammal and cell toxicity (R2 = 0.44). This paper reveals that polar and ionized compounds can more easily pass through cell membrane and have greater bioconcentration potential. Increasing of hydrophobicity and ionization can increase the cytotoxicity. Inclusion of descriptors representing hydrophobicity, ionization, acidity and absorption into the correlation equations can significantly improve the correlations of cytotoxicity with human and fish toxicity (R2 > 0.8), but not with mammal toxicity (R2 = 0.49). These descriptors reflect the differences of the toxicodynamics and toxicokinetics between cells and organisms.

Comparison of Toxicities to Vibrio fischeri and Fish Based on Discrimination of Excess Toxicity from Baseline Level.

Investigations on the relationship of toxicities between species play an important role in the understanding of toxic mechanisms to environmental organisms. In this paper, the toxicity data of 949 chemicals to fish and 1470 chemicals to V. fischeri were used to investigate the modes of action (MOAs) between species. The results show that although there is a positive interspecies correlation, the relationship is poor. Analysis on the excess toxicity calculated from toxic ratios (TR) shows that many chemicals have close toxicities and share the same MOAs between the two species. Linear relationships between the toxicities and octanol/water partition coefficient (log KOW) for baseline and less inert compounds indicate that the internal critical concentrations (CBRs) approach a constant both to fish and V. fischeri for neutral hydrophobic compounds. These compounds share the same toxic mechanisms and bio-uptake processes between species. On the other hand, some hydrophilic compounds exhibit different toxic effects with greatly different log TR values between V. fischeri and fish species. These hydrophilic compounds were identified as reactive MOAs to V. fischeri, but not to fish. The interspecies correlation is improved by adding a hydrophobic descriptor into the correlation equation. This indicates that the differences in the toxic ratios between fish and V. fischeri for these hydrophilic compounds can be partly attributed to the differences of bioconcentration between the two species, rather than the differences of reactivity with the target macromolecules. These hydrophilic compounds may more easily pass through the cell membrane of V. fischeri than the gill and skin of fish, react with the target macromolecules and exhibit excess toxicity. The compounds with log KOW > 7 exhibiting very low toxicity (log TR < -1) to both species indicate that the bioconcentration potential of a chemical plays a very important role in the identification of excess toxicity and MOAs.

Discrimination of excess toxicity from baseline level for ionizable compounds: Effect of pH.

The toxic effect can be affected by pH in water through affecting the degree of ionization of ionizable compounds. Wrong classification of mode of action can be made from the apparent toxicities. In this paper, the toxicity data of 61 compounds to Daphnia magna determined at three pH values were used to investigate the effect of pH on the discrimination of excess toxicity. The results show that the apparent toxicities are significantly less than the baseline level. Analysis on the effect of pH on bioconcentration factor (BCF) shows that the log BCF values are significantly over-estimated for the strongly ionizable compounds, leading to the apparent toxicities greatly less than the baseline toxicities and the toxic ratios greatly less than zero. A theoretical equation between the apparent toxicities and pH has been developed basing on the critical body residue (CBR). The apparent toxicities are non-linearly related to pH, but linearly to fraction of unionized form. The determined apparent toxicities are well fitted with the toxicities predicted by the equation. The toxicities in the unionized form calculated from the equation are close to, or greater than the baseline level for almost all the strongly ionizable compounds, which are very different from the apparent toxicities. The studied ionizable compounds can be either classified as baseline, less inert or reactive compounds in D. magna toxicity. Some ionizable compounds do not exhibit excess toxicity at a certain pH, due not to their poor reactivity with target molecules, but because of the ionization in water.

Discrimination of excess toxicity from narcotic effect: influence of species sensitivity and bioconcentration on the classification of modes of action.

The toxicity data of 2624 chemicals to fish, Daphniamagna, Tetrahymenapyriformis and Vibriofischeri were used to investigate the effects of species sensitivity and bioconcentration on excess toxicity. The results showed that 47 chemical classes were identified as having the same modes of action (MOAs) to all four species, but more than half of the classes were identified as having different MOAs. Difference in chemical MOAs is one of the reasons resulting in the difference in toxic effect to these four species. Other important reasons are the difference in sensitivity and bioconcentration of species. Among the four species, V. fischeri has the most compounds identified as reactive MOA. This may be due to some compounds can be easily absorbed into the bacteria, react with the DNA or proteins, disrupt the normal function of the cell and exhibit significantly greater toxicity to the bacteria. On the other hand, the skin and lipid content of aqueous organisms can strongly inhibit the bio-uptake for some reactive compounds, resulting in a less toxic effect than expected. D. magna is the most sensitive species and T. pyriformis is the least sensitive species of the four species. For a comparison of interspecies toxicity, we need to use the same reference threshold of excess toxicity. However, some reactive compounds may be identified as baseline or less inert compounds for low sensitive species from the threshold developed from high sensitive species. The difference in the discrimination of excess toxicity to different species is not only because of the difference in MOAs for some compounds, but also due to the difference in sensitivity and bioconcentration.

Investigation on modes of toxic action to rats based on aliphatic and aromatic compounds and comparison with fish toxicity based on exposure routes.

The modes of toxic action (MOAs) play an important role in the assessment of the ecotoxicity of organic pollutants. However, few studies have been reported on the MOAs in rat toxicity. In this paper, the toxic contributions of functional groups in 1255 aromatic compounds were calculated from regression and were then compared with the toxic contributions in aliphatic compounds. The results show that some functional groups have same toxic contributions both in aromatic and aliphatic compounds, but some have not. To investigate the MOAs in rat toxicity, the distribution of toxic ratio (TR) was examined for well-known baseline and less inert compounds and thresholds of log TR=0.3 and 0.5 were used to classify baseline, less inert and reactive compounds. The results showed that some compounds identified as baseline compounds in fish toxicity were also classified as baseline compounds in rat toxicity. Except for phenols and anilines which were identified as less inert compounds in fish toxicity, aromatic compounds with functional groups such as ether, nitrile, nitrophenol, isocyanatoe and chloro were identified as less inert chemicals in rat toxicity. Reactive compounds identified in fish toxicity exhibit greater toxicity to rats. These compounds can undergo nucleophilic substitution, acylation and Schiff base formation with biological macromolecules. The critical body residues (CBRs) calculated from absorption and bioconcentration show that log 1/CBRs in rat toxicity are not equal to that in fish for some compounds. It suggests that the exposure route can affect the identification of MOAs between these two species for these compounds.

Investigation on the relationship between bioconcentration factor and distribution coefficient based on class-based compounds: The factors that affect bioconcentration.

Bioconcentration factor (BCF) is one of the most important parameters in the assessment of the potential hazard of new compounds in aquatic ecosystems. However, the factors that influence the estimation of BCFs for a large variety of chemicals have not been systemically investigated in the literature. In this paper, a large BCF data set containing 1088 nonionic and ionic organic compounds was used to study the relationship between BCF and molecular descriptors and influencing factors. Step-by-step analysis on the class-based compounds showed that nonlinear Gaussian and Sigmoid equations could well describe relationships between logBCF and distribution coefficient for the compounds over a wide range of structures and chloro or/and bromo substituted aromatics, respectively. The quality of fit from the nonlinear models is better than the BCFBAF method from the Epi Suite program for the class-based compounds. Systemic prediction deviations have been observed for some types of compounds. The reasons for systemic deviations for these compounds can be attributed to the difference in bioconcentration mechanism for hydrophilic compounds, transformation for hydroxyphenols and three-membered rings, physical barrier for long chain and large polycyclic compounds, difference in determining methods of BCF (kinetic and steady-state), bioavailability for highly hydrophobic compounds and accuracy of BCF measurements for compounds with extremely high or low BCFs. These factors are important and should be considered in any reliable bioconcentration prediction.

Linear and non-linear relationships between soil sorption and hydrophobicity: model, validation and influencing factors.

The hydrophobic parameter represented by the octanol/water partition coefficient (logP) is commonly used to predict the soil sorption coefficient (K(oc)). However, a simple non-linear relationship between logK(oc) and logP has not been reported in the literature. In the present paper, soil sorption data for 701 compounds was investigated. The results show that logK(oc) is linearly related to logP for compounds with logP in the range of 0.5-7.5 and non-linearly related to logP for the compounds in a wide range of logP. A non-linear model has been developed between logK(oc) and logP for a wide range of compounds in the training set. This model was validated in terms of average error (AE), average absolute error (AAE) and root-mean squared error (RMSE) by using an external test set with 107 compounds. Nearly the same predictive capacity was observed in comparison with existing models. However, this non-linear model is simple, and uses only one parameter. The best model developed in this paper is a non-linear model with six correction factors for six specific classes of compounds. This model can well predict logK(oc) for 701 diverse compounds with AAE = 0.37. The reasons for systemic deviations in these groups may be attributed to the difference of sorption mechanism for hydrophilic/polar compounds, low solubility for highly hydrophobic compounds, hydrolysis of esters in solution, volatilization for volatile compounds and highly experimental errors for compounds with extremely high or low sorption coefficients.

Interspecies correlations of toxicity to eight aquatic organisms: theoretical considerations.

Interspecies correlations allow the prediction of toxicity to a number of other species. However, little attention has been paid to the theoretical considerations of the interspecies relationship based on the differences of bio-uptake and toxic mechanism between species. This study examines the interspecies correlations of toxicity between species of Vibrio fischeri, river bacteria, algae, Daphnia magna, carp, Tetrahymena pyriformis, fathead minnow and guppy based on the theoretical background. The results show that there are good interspecies correlations between marine bacterium and fresh water bacteria or fish and fish. It is suggested that compounds share the same bio-uptake and toxic mechanism of action between the species. On the other hand, poor interspecies relationships were found between toxicities to algae and T. pyriformis or D. magna. It is suggested that compounds have different toxic mechanisms of action between these species. Interspecies relationships can be improved by inclusion of the octanol/water partition coefficient or the energy of the lowest unoccupied molecular orbital. They reflect the difference of bio-uptake or toxic mechanism of action between species for organic compounds. Benzoic acids show very different toxicity contributions to the three species, V. fischeri, D. magna and carp. They can be easily absorbed into the unicellular bacteria, V. fischeri. On the contrary, the skin and lipid content of multicellular organisms, such as D. magna and fish, can strongly inhibit the bio-uptake for ionizable compounds, which results in the different toxic effect between V. fischeri and D. magna or carp. Good correlation coefficients were observed between toxicities to V. fischeri and D. magna or fishes by inclusion of hydrophobic and ionization parameters. V. fischeri or D. magna can serve as a surrogate of fish toxicity for hydrophobic and ionizable compounds studied. Toxic mechanisms of action are discussed based on the theoretical background of the interspecies correlation.

Toxicity of organic chemicals to Tetrahymena pyriformis: effect of polarity and ionization on toxicity.

A large toxicity data set containing the toxicities of 250 phenols and 252 aliphatic compounds to Tetrahymena pyriformis was classified into different groups based on the structure and substituted functional groups. QSAR analysis was performed between the toxicity and calculated descriptors, expressed as hydrophobicity, polarity and ionization. Through an analysis of these class-based compounds, significant relationships were developed between the toxicity and hydrophobicity for non-polar and polar narcotic compounds. A single model for both non-polar and polar narcotics was developed by inclusion of a polar descriptor as well as the hydrophobic parameter logP. The highly hydrophobic polar narcotics can be treated as non-polar narcotics because their polar functional group(s) makes a relatively small contribution as compared to their hydrophobicity. A cut-off to classify the polar narcotics is difficult because polarity of a chemical not only depends on one or two functional groups (i.e. amino- or hydroxyl-) substituted on the compound, but also on the overall hydrophobicity of the compound. The toxicity increases with increasing the ionization by increasing the interaction between ionisable compounds and macromolecules at the target sites. However, the toxicity decreases with increasing the ionization by decreasing the bio-uptake for extremely ionisable compounds. A significant QSAR equation has been developed between the toxicity to T. pyriformis and the descriptors of hydrophobic, polarity/polarizability and ionization parameters for 457 compounds (R(2)=0.87). These compounds contain non-polar, polar and reactive compounds, and some of them are extremely ionisable. The models developed are simple, interpretable and transparent, using a small number of descriptors.

Classification of toxicity of phenols to Tetrahymena pyriformis and subsequent derivation of QSARs from hydrophobic, ionization and electronic parameters.

Phenolic compounds were classified into different groups based on the structure and functional groups of the phenol. Quantitative structure-activity relationship (QSAR) analysis was performed between the toxicity and octanol/water partition coefficient (logP) for these groups. The results showed that the toxicity of non-ionisable phenols is dependent on their hydrophobicity. Poor relationships were found between the toxicity and logP for ionisable compounds, and the use of methods based on logP to predict the toxicity of ionisable compounds can result in considerable errors. Ionized and unionized forms have different contributions to toxicity; the unionized form plays a more important role than the ionized form because the toxicity of organic acids and phenols decreases as the pH increases. In order to investigate the effect of ionization, the fraction of ionized and unionized forms of phenols at different pH values were calculated from the pK(a) values, and a corrected distribution partition coefficient (D(T)) was derived from QSAR analysis for ionisable compounds. The prediction of toxicity of non-reactive ionisable compounds was improved remarkably by using the D(T) parameter. Ionization not only affects the bio-uptake of ionisable compounds, but interaction with the receptor micromolecule can also depend on the electronic situation, which is also related to the ionization. Stepwise regression showed that the reactivity of ionisable phenols was strongly correlated with the fraction of negatively charged form (F(-)). Interpretable QSAR equations with good statistical fits were developed from hydrophobic, ionization and electronic parameters for 207 phenols.