Structure-toxicity modeling of pesticides to honey bees

Chemometric modeling of plant protection products (PPPs) for the prediction of acute contact toxicity against honey bees (A. mellifera): A 2D-QSAR approach

Honey bees (Apis mellifera) are vital for economic, viable agriculture and for food safety. Although Plant Protection Products (PPPs) are of undeniable importance in the global agricultural system, these have become potential threats for non-target organisms like pollinators (e.g., honey bees etc.), resulting in the disruption of the ecological balance. In the current work, we have used the 113 PPP analogs to develop a 2D-QSAR model and explored the structural features modulating the toxic effects on honey bees, following the Organization for Economic Co-operation and Development (OECD) guidelines. The extensive validation of the developed model has been performed using internal and external validation metrics to make sure that the model is statistically sound and interpretable enough to be acceptable. The obtained results (R² = 0.666, Q² = 0.594, Q²_F1 = 0.647 and Q²_F2 = 0.646) determine the predictability and reliability of the developed model. This model should be useful for the predictions (acute contact toxicity (LD₅₀)) of the new and untested compounds located inside the applicability domain of the developed model. Moreover, we have performed the in-silico prediction of toxicity against honey bees of a total of 709 compounds obtained from the pesticide properties database (PPDB) using the developed model.

Effects of glyphosate exposure on honeybees

Honeybees show an important pollination ability and play vital roles in improving crop yields and increasing plant genetic diversity, thereby generating tremendous economic benefits for humans. However, honeybee survival is affected by a number of biological and abiotic stresses, including the effects of fungi, bacteria, viruses, parasites, and especially agrochemicals. Glyphosate, a broad-spectrum herbicide that is primarily used for weed control in agriculture, has been reported to have lethal and sublethal effects on honeybees. Here, we summarize recent advances in research on the effects of glyphosate on honeybees, including effects on their behaviors, growth and development, metabolic processes, and immune defense, providing a detailed reference for studying the mechanism of action of pesticides. Furthermore, we provide possible directions for future research on glyphosate toxicity to honeybees.

Integrating QSAR models predicting acute contact toxicity and mode of action profiling in honey bees (A. mellifera): Data curation using open source databases, performance testing and validation

Honey bees (Apis mellifera) provide key ecosystem services as pollinators bridging agriculture, the food chain and ecological communities, thereby ensuring food production and security. Ecological risk assessment of single Plant Protection Products (PPPs) requires an understanding of the exposure and toxicity. In silico tools such as QSAR models can play a major role for the prediction of structural, physico-chemical and pharmacokinetic properties of chemicals as well as toxicity of single and multiple chemicals. Here, the first integrative honey bee QSAR model has been developed for PPPs using EFSA's OpenFoodTox, US-EPA ECOTOX and Pesticide Properties DataBase i) to predict acute contact toxicity (LD₅₀) and ii) to profile the Mode of Action (MoA) of pesticides active substances. Three different classification-based and four regression-based models were developed and tested for their performance, thus identifying two models providing the most reliable predictions based on k-NN algorithm. The two-category QSAR model (toxic/non-toxic; n = 411) was validated using sensitivity (=0.93), specificity (=0.85), balanced accuracy (=0.90), and Matthews correlation coefficient (MCC = 0.78) as statistical parameters. The regression-based model (n = 113) was validated for its reliability and robustness (R² = 0.74; MAE = 0.52). Current study proposes the MoA profiling for 113 pesticides active substances and the first harmonised MoA classification scheme for acute contact toxicity in honey bees, including LD_50s data points from three different databases. The classification allows to further define MoAs and the target site of PPPs active substances, thus enabling regulators and scientists to refine chemical grouping and toxicity extrapolations for single chemicals and component-based mixture risk assessment of multiple chemicals. Relevant future perspectives are briefly addressed to integrate MoA, adverse outcome pathways (AOPs) and toxicokinetic information for the refinement of single-chemical/combined toxicity predictions and risk estimates at different levels of biological organization in the bee health context.

Antifungal residues analysis in various Romanian honey samples analysis by high resolution mass spectrometry

Given that the pesticide and fungicide residues determination in honey is not a routine analysis in Romania, information on these emerging contaminants is useful for consumer's safety. High resolution mass spectrometry technique was applied by Q-Exactive Orbitrap LC-MS/MS to identify and quantify environmental contaminants in honey. A list of 25 compounds, biocides and antifungals was selected for the method development, based on the occurrence in the Romanian environment and their potential usage in agriculture. The method was applied for 18 various honey samples collected in different geographic regions of Romania. Eleven compounds were present in the honey samples: carbendazim, enilconazole, hexaconazole, penconazole, tebuconazole, flusilazole, thiabendazole, terconazole, cyproconazole, propiconazole, metalaxyl. Targeted MS/MS analyses were performed for confirmation. The measured quantities ranged from 1.7-7.2 μg kg^-1, lower than MRLs established by the legislation. The most abundant compound was enilconazole (imazalil), which was detected in fourteen samples. To the best of our knowledge, the present study is the first one concerning antifungal contamination of honey in Romania. The results proved that the tested honey samples are safe for human consumption.

Acute thiamethoxam toxicity in honeybees is not enhanced by common fungicide and herbicide and lacks stress-induced changes in mRNA splicing

Securing food supply for a growing population is a major challenge and heavily relies on the use of agrochemicals to maximize crop yield. It is increasingly recognized, that some neonicotinoid insecticides have a negative impact on non-target organisms, including important pollinators such as the European honeybee Apis mellifera. Toxicity of neonicotinoids may be enhanced through simultaneous exposure with additional pesticides, which could help explain, in part, the global decline of honeybee colonies. Here we examined whether exposure effects of the neonicotinoid thiamethoxam on bee viability are enhanced by the commonly used fungicide carbendazim and the herbicide glyphosate. We also analysed alternative splicing changes upon pesticide exposure in the honeybee. In particular, we examined transcripts of three genes: (i) the stress sensor gene X box binding protein-1 (Xbp1), (ii) the Down Syndrome Cell Adhesion Molecule (Dscam) gene and iii) the embryonic lethal/abnormal visual system (elav) gene, which are important for neuronal function. Our results showed that acute thiamethoxam exposure is not enhanced by carbendazim, nor glyphosate. Toxicity of the compounds did not trigger stress-induced, alternative splicing in the analysed mRNAs, thereby leaving dormant a cellular response pathway to these man-made environmental perturbations.

Machine Learning Methods in Computational Toxicology

Various methods of machine learning, supervised and unsupervised, linear and nonlinear, classification and regression, in combination with various types of molecular descriptors, both "handcrafted" and "data-driven," are considered in the context of their use in computational toxicology. The use of multiple linear regression, variants of naïve Bayes classifier, k-nearest neighbors, support vector machine, decision trees, ensemble learning, random forest, several types of neural networks, and deep learning is the focus of attention of this review. The role of fragment descriptors, graph mining, and graph kernels is highlighted. The application of unsupervised methods, such as Kohonen's self-organizing maps and related approaches, which allow for combining predictions with data analysis and visualization, is also considered. The necessity of applying a wide range of machine learning methods in computational toxicology is underlined.

Insights into the Molecular Basis of the Acute Contact Toxicity of Diverse Organic Chemicals in the Honey Bee

Use of chemical pollutants, including pesticides and other industrial chemicals, has resulted in significant risks to the whole ecosystem. Therefore, ecological risk assessment of chemicals is vital and necessary. Since the honey bee (Apis mellifera) is probably among the most exposed species to the polluting chemicals, we focused on the in silico estimation of honey bee toxicity (HBT) of chemicals and the analysis of the relevance of chemical HBT and several key physical-chemical properties and structural characteristics. A total of 40 classification models were developed by combination of five machine learning methods along with seven kinds of fingerprints and a set of molecular descriptors. After 5-fold cross validation and external validation, several models showed good predictive power. The relevance of 12 key physical-chemical properties and chemical HBT was also investigated. Five properties, including AlogP, logD, molecular weight (MW), molecular surface area (MSA), and the number of rotatable bonds (nRTB), indicated positive correlation coefficients with HBT, while molecular solubility (logS) and the number of hydrogen bond donors (nHBD) indicated negative correlation coefficients. Finally, seven privileged substructures responsible for chemical HBT were identified from KRFP and SubFP fingerprints. The results of this study should provide critical information and useful tools for chemical HBT estimation in environmental risk assessment.

Uncovering the novel characteristics of Asian honey bee, Apis cerana, by whole genome sequencing

Background

The honey bee is an important model system for increasing understanding of molecular and neural mechanisms underlying social behaviors relevant to the agricultural industry and basic science. The western honey bee, Apis mellifera, has served as a model species, and its genome sequence has been published. In contrast, the genome of the Asian honey bee, Apis cerana, has not yet been sequenced. A. cerana has been raised in Asian countries for thousands of years and has brought considerable economic benefits to the apicultural industry. A cerana has divergent biological traits compared to A. mellifera and it has played a key role in maintaining biodiversity in eastern and southern Asia. Here we report the first whole genome sequence of A. cerana.

Results

Using de novo assembly methods, we produced a 238 Mbp draft of the A. cerana genome and generated 10,651 genes. A.cerana-specific genes were analyzed to better understand the novel characteristics of this honey bee species. Seventy-two percent of the A. cerana-specific genes had more than one GO term, and 1,696 enzymes were categorized into 125 pathways. Genes involved in chemoreception and immunity were carefully identified and compared to those from other sequenced insect models. These included 10 gustatory receptors, 119 odorant receptors, 10 ionotropic receptors, and 160 immune-related genes.

Conclusions

This first report of the whole genome sequence of A. cerana provides resources for comparative sociogenomics, especially in the field of social insect communication. These important tools will contribute to a better understanding of the complex behaviors and natural biology of the Asian honey bee and to anticipate its future evolutionary trajectory.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-16-1) contains supplementary material, which is available to authorized users.

Use of artificial neural networks in the QSAR prediction of physicochemical properties and toxicities for REACH legislation

With the introduction of the REACH legislation in the European Union, there is a requirement for property and toxicity data on chemicals produced in or imported into the EU at levels of 1 tonne/year or more. This has meant an increase in the in silico prediction of such data. One of the chief predictive approaches is QSAR (quantitative structure-activity relationships), which is widely used in many fields. A QSAR approach that is increasingly being used is that of artificial neural networks (ANNs), and this chapter discusses its application to the range of physicochemical properties and toxicities required by REACH. ANNs generally outperform the main QSAR approach of multiple linear regression (MLR), although other approaches such as support vector machines sometimes outperform ANNs. Most ANN QSARs reported to date comply with only two of the five OECD Guidelines for the Validation of (Q)SARs.

Enantioselective behavior of malathion enantiomers in toxicity to beneficial organisms and their dissipation in vegetables and crops

The dissipation behavior of the two enantiomers of malathion was elucidated in five plant species using enantioselective high performance liquid chromatography (HPLC), and the acute toxicity of the individual enantiomers toward earthworms and honeybees was studied. The calculated LC(50) values of the R-, S- and rac-malathion to earthworms were 0.3869, 25.17, and 19.19 μg/cm(2), respectively, while the calculated LC(50) values of R-, S- and rac-malathion to bees were 2.15, 36.67, and 7.11 μg/mL, respectively. This indicated that the R-enantiomer was more toxic than S-enantiomer. The results of the degradation of racemate in Chinese cabbage and rape showed that the inactive S-(-)-enantiomer degraded faster than the active R-(+)-enantiomer. Inversely, we found a preferential degradation of the R-(+)-enantiomer in sugar beet. However, the degradation of malathion in paddy rice and wheat were nonenantioselectivity. In all plants, malathion was degraded to levels <10% after 5 days, and the calculated t(½) values of the enantiomers ranged from 0.83 to 1.43 days in these five plants. In conclusion, our findings of enantioselectivity in the environmental fate and acute toxicity of the malathion enantiomers may have implications for better environmental and ecological risk assessment for chiral pesticides in general.

Interpretation of honeybees contact toxicity associated to acetylcholinesterase inhibitors

The widespread use of different pesticides generates adverse effects on non target organisms like honeybees. Organophosphorous and carbamates kill honeybees through the inactivation of acetylcholinesterase (AChE), thereby interfering with nerve signaling and function. For this class of pesticides, it is fundamental to understand the relationship between their structures and the contact toxicity for honeybees. A Quantitative Structure-Activity Relationship (QSAR) study was carried out on 45 derivatives by a genetic algorithm approach starting from more than 2500 descriptors. In parallel, a new 3D model of AChE associated to honeybees was defined. Physicochemical properties of the receptor and docking studies of the derivatives allow understanding the meaningful of three descriptors and the implication of several amino acids in the overall toxicity of the pesticides.

Linear and non-linear QSAR modelling of juvenile hormone esterase inhibitors

A tight control of juvenile hormone (JH) titre is crucial during the life cycle of a holometabolous insect. JH metabolism is made through the action of enzymes, particularly the juvenile hormone esterase (JHE). Trifluoromethylketones (TFKs) are able to inhibit this enzyme to disrupt the endocrine function of the targeted insect. In this context, a set of 96 TFKs, tested on Trichoplusia ni for their JHE inhibition, was split into a training set (n = 77) and a test set (n = 19) to derive a QSAR model. TFKs were initially described by 42 CODESSA (Comprehensive Descriptors for Structural and Statistical Analysis) descriptors, but a feature selection process allowed us to consider only five descriptors encoding the structural characteristics of the TFKs and their reactivity. A classical and spline regression analysis, a three-layer perceptron, a radial basis function network and a support vector regression were experienced as statistical tools. The best results were obtained with the support vector regression (r(2) and r(test)(2) = 0.91). The model provides information on the structural features and properties responsible for the high JHE inhibition activity of TFKs.

Metabolic enzymes associated with xenobiotic and chemosensory responses in Nasonia vitripennis

The numbers of glutathione S-transferase, cytochrome P450 and esterase genes in the genome of the hymenopteran parasitoid Nasonia vitripennis are about twice those found in the genome of another hymenopteran, the honeybee Apis mellifera. Some of the difference is associated with clades of these families implicated in xenobiotic resistance in other insects and some is in clades implicated in hormone and pheromone metabolism. The data support the hypothesis that the eusocial behaviour of the honeybee and the concomitant homeostasis of the nest environment may obviate the need for as many gene/enzyme systems associated with xenobiotic metabolism as are found in other species, including N. vitripennis, that are thought to encounter a wider range of potentially toxic xenobiotics in their diet and habitat.

Additive SMILES-based optimal descriptors in QSAR modelling bee toxicity: Using rare SMILES attributes to define the applicability domain

The additive SMILES-based optimal descriptors have been used for modelling the bee toxicity. The influence of relative prevalence of the SMILES attributes in a training and test sets to the models for bee toxicity has been analysed. Avoiding the use of rare attributes improves statistical characteristics of the model on the external test set. The possibility of using the probability of the presence of SMILES attributes in training and test sets for rational definition of the applicability domain is discussed.

Artificial neural network modeling in environmental toxicology

Artificial neural networks are increasingly used in environmental toxicology to find complex relationships between the ecotoxicity of xenobiotics and their structure and/or physicochemical properties. The raison d'etre of these nonlinear tools is their ability to derive powerful QSARs for molecules presenting different mechanisms of action. In this chapter, the main QSAR models derived for aquatic and terrestrial species are reviewed. Their characteristics and modeling performances are deeply analyzed.

SMILES as an alternative to the graph in QSAR modelling of bee toxicity

Simplified Molecular Input Line Entry System (SMILES) nomenclature has been used as elucidating the molecular structure in construction of the quantitative structure-activity relationships (QSAR) for predicting bee toxicity. On the basis of the symbols used in the SMILES notation numerical parameters have been obtained, which are simple and fast to calculate. The method has been used to develop a QSAR model to predict toxicity of pesticides on bees. Results on a heterogeneous set of pesticides are good. Statistical characteristics of this model are: n=85, R2=0.68, s=0.82, F=180 (training set); n=20, R2=0.72, s=0.68, F=46 (test set).

A deficit of detoxification enzymes: pesticide sensitivity and environmental response in the honeybee

The honeybee genome has substantially fewer protein coding genes ( approximately 11 000 genes) than Drosophila melanogaster ( approximately 13 500) and Anopheles gambiae ( approximately 14 000). Some of the most marked differences occur in three superfamilies encoding xenobiotic detoxifying enzymes. Specifically there are only about half as many glutathione-S-transferases (GSTs), cytochrome P450 monooxygenases (P450s) and carboxyl/cholinesterases (CCEs) in the honeybee. This includes 10-fold or greater shortfalls in the numbers of Delta and Epsilon GSTs and CYP4 P450s, members of which clades have been recurrently associated with insecticide resistance in other species. These shortfalls may contribute to the sensitivity of the honeybee to insecticides. On the other hand there are some recent radiations in CYP6, CYP9 and certain CCE clades in A. mellifera that could be associated with the evolution of the hormonal and chemosensory processes underpinning its highly organized eusociality.

SAR and QSAR modeling of endocrine disruptors

A number of xenobiotics by mimicking natural hormones can disrupt crucial functions in wildlife and humans. These chemicals termed endocrine disruptors are able to exert adverse effects through a variety of mechanisms. Fortunately, there is a growing interest in the study of these structurally diverse chemicals mainly through research programs based on in vitro and in vivo experimentations but also by means of SAR and QSAR models. The goal of our study was to retrieve from the literature all the papers dealing with structure-activity models on endocrine disruptor xenobiotics. A critical analysis of these models was made focusing our attention on the quality of the biological data, the significance of the molecular descriptors and the validity of the statistical tools used for deriving the models. The predictive power and domain of application of these models were also discussed.

A new strategy for using supervised artificial neural networks in QSAR

A new type of environmental QSAR model is presented for the common situation in which the biological activity of molecules mainly depends on their 1-octanol/water partition coefficient (log P). In a first step, a classical regression equation with log P is derived. The residuals obtained with this simple linear equation are then modeled from a supervised artificial neural network including different molecular descriptors as input neurons. Finally, results produced by the linear and nonlinear models are both considered for calculating the activity values, which are compared with the initial actual activity values. A heterogeneous database of 569 organic compounds with 96-h LC50s measured to the fathead minnow (Pimephales promelas), randomly divided into a training set of 484 chemicals and a testing set of 85 chemicals, was used as illustrative example to show the potentialities of this new modeling strategy Finally, practical suggestions are given for designing this type of hybrid QSAR model.

Linear versus nonlinear QSAR modeling of the toxicity of phenol derivatives to Tetrahymena pyriformis

Quantitative structure-activity relationship (QSAR) models were derived from a structurally heterogeneous set of 200 phenol derivatives for which the 50% growth inhibition concentration (IGC(50)) values to the ciliated protozoan Tetrahymena pyriformis were available. Each molecule was described by means of physicochemical descriptors and structural features. Partial least squares (PLS) regression analysis and a three-layer perceptron were used as statistical engine. The performances of the linear and nonlinear models were estimated from an external testing set of 50 chemicals. Despite hard constraints voluntarily imposed in the design of the neural network models, they provided better simulation results than the PLS models.

Partial least squares modelling of the acute toxicity of aliphatic compounds to Tetrahymena pyriformis

The aim of this study was to evaluate a multivariate statistical model, utilising Partial Least Squares (PLS) analysis, for the prediction of the acute toxicity of aliphatic chemicals to the ciliate Tetrahymena pyriformis. A model was developed that was capable of making a prediction regardless the mechanism of toxic action. The toxicity of 476 compounds, possessing different mechanisms of toxic action was considered. A set of 74 descriptors, including the octanol-water partition coefficient, molecular-orbital descriptors, geometrical, topological and connectivity indices, was generated. A three-component, eight-descriptor PLS model was developed. It was validated by a Y-permutation test and by simulation of external prediction for complementary subsets. A comparison with existing class or mechanism-based models, derived on the same data set, was made.