Insights into the Molecular Basis of the Acute Contact Toxicity of

Nov 21, 2017 - Use of chemical pollutants, including pesticides and other industrial chemicals, has resulted in significant risks to the whole ecosyst...
3 downloads 10 Views 2MB Size
Subscriber access provided by READING UNIV

Article

Insights into Molecular Basis of Acute Contact Toxicity of Diverse Organic Chemical in Honey Bee Xiao Li, Yuan Zhang, Hongna Chen, Huanhuan Li, and Yong Zhao J. Chem. Inf. Model., Just Accepted Manuscript • DOI: 10.1021/acs.jcim.7b00476 • Publication Date (Web): 21 Nov 2017 Downloaded from http://pubs.acs.org on November 22, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Chemical Information and Modeling is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Information and Modeling

Insights into Molecular Basis of Acute Contact Toxicity of Diverse Organic Chemical in Honey Bee Xiao Li1,2*, Yuan Zhang2, Hongna Chen3, Huanhuan Li2, Yong Zhao1,2* 1. Beijing Computing Center, Beijing Academy of Science and Technology, 7 Fengxian road, Beijing 100094, China 2. Beijing Beike Deyuan Bio-Pharm Technology Co.Ltd., 7 Fengxian road, Beijing 100094, China 3. Tigermed Consulting Co., Ltd., 20 Chaowai Street, Beijing 100020, China * To whom correspondence should be addressed. Corresponding authors Yong Zhao Phone: +86-10-5934-1764; Fax: +86-10-5934-1855 E-mail: [email protected] Xiao Li Phone: +86-10-5934-1890 E-mail: [email protected] or [email protected]

ACS Paragon Plus Environment

Journal of Chemical Information and Modeling 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract The chemical pollutants, including pesticides and other industrial chemicals, have resulted in significant risks to the whole ecosystem. Therefore, the ecological risk assessment of chemicals is vital and necessary. Since 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 chemical 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 fingerprint 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 were 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.

Keywords: honey bee toxicity of chemical; machine learning; physical-chemical properties; molecular fingerprints; structural alerts.

ACS Paragon Plus Environment

Page 2 of 31

Page 3 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Information and Modeling

Introduction Chemicals are an integral part of modern life and they are used in a wide variety of products.

However, misuse and improper disposal of chemicals has also resulted

in significant risks to the environment. The hazardous chemicals, especially pesticides and other toxic industrial chemicals, continued to affect all aspects of natural resources, including the water, soil, atmosphere and wildlife. The honey bee (Apis mellifera) is among the most important insects because they transfer pollen between plants, and always be exposed to chemical pollutants when visiting the plants.1 In the past decades, people have seen a dramatic drop in honeybee numbers due to chemical toxicity in various countries.2-6 Therefore, the special attention should be paid to assess the potential risk of chemicals in honey bee. Toxicity of chemicals in honey bee should always be assessed with 24h or 48h acute toxicity experimental bioassays to determine the median lethal concentration that induces 50 percent death (LC50) or lethal dose that induces 50 percent death (LD50). Since a full evaluation of the honey bee toxicity (HBT) of a large number of organic chemicals by experimental test is costly, time consuming, and poses an ethical problem, there is a very urgent need to develop alternative methods and tools for HBT estimation. In silico techniques, such as quantitative structure–activity relationship (QSAR) have been widely used to reduce the cost of the environmental risk assessment.7, 8 QSAR models have already been used for the estimation of ecotoxicological endpoints in various ecologically significant species including T. pyriformis,7, 9 quail10

ACS Paragon Plus Environment

Journal of Chemical Information and Modeling 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and fathead minnow11-13 in the past decades. Besides, there were also several QSAR models for HBT prediction. In 1991, Vighi et al.14 developed a QSAR model for HBT prediction with organophosphorus pesticides. Later, James et al.15 also proposed QSAR models for the prediction of HBT with 100 organophosphorus pesticides (89 in training set and 11 in test set) using multi-layer feedforward neural network method. The root mean square residual value for the training set was 0.43, and that for test set was 0. 386. Since Vighi and James's models were only generated to simulate the HBT of organophosphorus pesticides, the usefulness of these models is very limited. In 2014, Singh, et al. established global models for qualitative and quantitative HBT prediction with more than 200 structurally diverse pesticides. The models were developed to predict the qualitative and quantitative HBT of new organic pesticides for regulatory purposes.5 More recently, Como, et al. also proposed a series of global model for HBT prediction of pesticides using k-Nearest Neighbor (kNN) algorism based on 256 pesticides with acute contact toxicity data collected from different sources.16 The models were validated with good prediction (accuracy of 70% for all compounds and of 65% for highly toxic compounds), which suggested that these models could be used to reliably predict the HBT of structurally diverse pesticides. However, the physical-chemical properties and structural characteristics of toxic and none toxic compounds were not analyzed in these studies, and the usefulness of most published models was restricted because of poor availability. In this study, we focused on: (1) the development of chemical acute contact HBT models combining different machine learning algorisms with substructure pattern

ACS Paragon Plus Environment

Page 4 of 31

Page 5 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Information and Modeling

recognition methods based on structurally diverse organic chemicals; (2) the analysis of the difference of key physical-chemical properties and structural characteristics between the chemicals with high HBT and those without HBT.

ACS Paragon Plus Environment

Journal of Chemical Information and Modeling 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Materials and methods Data preparation In this study, we addressed acute contact toxicity of organic chemicals in honeybees. The training data was extracted from Como's work.16 Como carried out an outstanding work to collect chemical HBT data from different sources, including the DEMETRA project,2 the Terrestrial US EPA ECOTOX database present in the OECD QSAR Toolbox V3.3 (www.qsartoolbox.org) and the European Food Safety Authority (EFSA, http://www.efsa.europa.eu/it/). The compounds with LD50 values presented in terms of µg/bee resulted from the mortality of honey bees recorded after 48h of contact exposure were proposed in Como's work. Besides, an external validation set were extracted from Singh's work. Both the data sets were carefully prepared by the followed steps: (1) removing mixtures, inorganic and organometallic compounds; (2) salts were converted to their parent forms. Then, The duplicate substances in external from training set were removed. The U.S. EPA has established HBT categories on the basis of LD50, chemicals with LD50 < 2 µg/bee were categorized as highly toxic, 2 µg/bee ≤ LD50 < 11 µg/bee as moderately toxic, and chemicals with LD50 ≥ 11 µg/bee were classified as practically non-toxic.17 Considering that the U.S. EPA criteria was a little liberal for HBT risk assessment, 100 µg/bee was also used as thresholds for HBT model building in the previous studies.5, 11, 16 Herein, we grouped the compounds into three classes with LD50 thresholds 11 µg/bee and 100 µg/bee, compounds with LD50