ToxWatch pubs.acs.org/crt
Cite This: Chem. Res. Toxicol. XXXX, XXX, XXX−XXX
New in Silico Trends in Food Toxicology Francesca Cavaliere and Pietro Cozzini* Molecular Modelling Lab, Department of Food and Drug, University of Parma, Parco Area delle Scienze 11/A, 43124 Parma, Italy
Chem. Res. Toxicol. Downloaded from pubs.acs.org by 95.181.176.217 on 09/22/18. For personal use only.
ABSTRACT: Ever growing numbers of chemicals in food and drinking water make it impossible to address safety by classical approaches in toxicology. In silico chemical methods could be a first-line for hazard characterization, requiring food toxicology to expand the use of approaches currently well applied in medicinal chemistry.
C
between a small molecule and a protein at the atomic level and use a scoring function to predict the strength of the interaction considering the enthalpic and the entropic contributions to binding through the application of a specific force field. Molecular dynamics simulation, instead, allows one to study the atomic motions of a protein as a function of the time and is normally applied to study the conformational changes occurring in a protein over time and for analyzing the impact of a mutation or a polymorphism to the ligand−protein interaction.2 There are several practical considerations for an effective modeling approach. Protein structures are commonly derived by crystallographic data, and in this case the quality limits are defined in terms of resolution, R factor, and temperature, B factor. When the crystal structure is not available, homology modeling allows the prediction of 3D structures based on the assumption that proteins with similar sequence or belonging to the same protein family have a similar fold, and the accuracy of the data is dependent on the sequence identity. Another important point is the generation of correct ligand input files. A lot of software allows the assignment of correct threedimensional (3D) geometries, proper bond orders, tautomer and ionization state, and their energy minimization. However, it is also possible to retrieve directly the optimized 3D ligands structure from some Web sites or databases. Parametrization of ligand structures is often a laborious step of in silico methods, but software such as SwissParam, CGenFF, and ParaTool (just to cite the most famous) allows to speed-up this important step in computational approaches. Moreover, an important issue is the scoring function used for retrieving an estimation of the free energy of the protein−ligand complex. To date, different kinds of approaches (force field-based methods, semiempirical methods, empirical methods, and knowledge-based methods extensively treated by Spyrakis et al.)3 have emerged based on
hemicals present in food are compounds naturally occurring, intentionally added (additives, flavorings, supplements), or inadvertently present (pesticides, food contact materials, environmental pollutants). Although they are generally present at very low levels, we are daily exposed to these compounds. Thus, a prolonged exposure to them could adversely affect human health. Efficient risk assessment methods, such as in vitro and in vivo tests, are normally used for toxicological studies to define and/or predict the harmful potentially compounds. It has been estimated that more than 130 million compounds are currently on the market, and about 500−1000 chemicals are produced every year (source, Chemical Abstract Service (CAS)). This large number of chemicals is too big to be investigated by means of standard experimental approaches, and this value considerably increases if we consider the wide range of substances resulting from metabolic and degradation processes. Computational approaches could be a valuable alternative for toxicological studies. In silico approaches are defined as anything we can do with a computer and, in particular, to simulate and to predict the behavior of two or more molecules. In food science, in silico methodologies are commonly restricted to statistical analyses, structure−activity relationship (SAR) and quantitative structure−activity relationships (QSAR), based on the assumption that chemical structure or (sub)structure have properties that could be associated with a biological activity, despite the opportunity that many further computational techniques could be applied in the science of food toxicology.1 Ligand-based virtual screening method is a common application in drug design as a preliminary screening and/or when the crystal structure of the target is not available in the protein data bank (PDB: www.rcsb.org). It is used to allow one to predict putative binders for a target starting by large compound libraries in order to select a smaller number of compounds for biological testing. Molecular docking approaches allow to predict the putative binding mode(s) © XXXX American Chemical Society
A
DOI: 10.1021/acs.chemrestox.8b00133 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX
ToxWatch
Chemical Research in Toxicology
Sharjah, United Arab Emirates. https://doi.org/10.2174/ 97816080520111070301. (4) Ehrlich, V. A., Dellafiora, L., Mollergues, J., Dall’Asta, C., Serrant, P., Marin-Kuan, M., Lo Piparo, E., Schilter, B., and Cozzini, P. (2015) Hazard Assessment Through Hybrid In Vitro/In Silico Approach: The Case of Zearalenone. ALTEX 32 (4), 275−286. (5) Ginex, T., Dall’Asta, C., and Cozzini, P. (2014) Preliminary hazard evaluation of androgen receptor-mediated endocrine-disrupting effects of thioxanthone metabolites through structure-based molecular docking. Chem. Res. Toxicol. 27, 279−289.
the decomposition of the free energy in a variable number of different terms, that is, the sum of the interaction energy between protein and ligand, the solvation of the ligand, the protein and the complex, and the entropic and conformational changes. The question is how these computational approaches could be useful in toxicology food science. On the molecular scale, an interaction is an interaction, and thus by a chemical point of view, it is not important if a compound is a drug, a pesticide, a food contact material, an additive, and so on. Thus, these methods could enable us to explore and predict the interaction between food molecules and specific receptors in order to discriminate “interactor” from “non-interactor” compounds. One of the first applications of in silico methods in food science fields has been applied to discover new endocrine disrupting compounds (EDCs).1 Although EDCs can impact cellular hormonal pathways through different ways, most information is available about interference with the receptors belonging to the nuclear receptor family (NR), to which some EDCs can bind, enhancing or inhibiting the effect of a hormone.1 The imbalance in their regulation and/or function is related to a wide range of diseases, such as breast, endometrial, and prostate cancer. All these compounds are massively present in the environment, and their presence in food could have a large impact on public health. Thus, there is a great demand for fast methods that allow one to understand which compounds could be defined as safe and not safe. In silico methods are faster and cheaper than classical tests, and moreover, they permit reduction in the number of animal tests. In silico techniques have been used to decipher the mode and mechanism of action of some food molecules or to discover new xenoestrogens,1−5 food pollutants, mycotoxins and related metabolites, and food contact material5 that we may be significantly exposed to during our daily lives. However, in silico binding prediction is not an absolute certainty of the real activation of the receptor brought about by the ligand, thus compounds predicted as positive by computational methods require further in vitro analyses. We should consider in silico methods as a funnel that allows screening large libraries of compounds to retrieve only a limited number of compounds.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Pietro Cozzini: 0000-0002-4826-8108 Notes
Views expressed in this editorial are those of the authors and not necessarily the views of the ACS. The authors declare no competing financial interest.
■
REFERENCES
(1) Amadasi, A., Mozzarelli, A., Meda, C., Maggi, A., and Cozzini, P. (2009) Identification of Xenoestrogens in Food Additives by an Integrated in Silico and in Vitro Approach. Chem. Res. Toxicol. 22, 52−63. (2) Cavaliere, F., Montanari, E., Emerson, A., Buschini, A., and Cozzini, P. (2017) In silico pharmacogenetic approach: The natalizumab case study. Toxicol. Appl. Pharmacol. 330, 93−99. (3) Spyrakis, F., Kellogg, G. E., Alessio, A., and Cozzini, P. (2007) Scoring Functions for Virtual Screening. In Frontiers in Drug Design and Discovery, Vol 3, pp 317−379, Bentham Science: Emirate of B
DOI: 10.1021/acs.chemrestox.8b00133 Chem. Res. Toxicol. XXXX, XXX, XXX−XXX
