A virtual library of one million new macrolide scaffolds could speed up drug discovery

Researchers at North Carolina State University have created the largest virtual library of publicly available macrolide scaffolds. The library – called V1M – contains chemical structures and properties calculated for 1 million macrolide scaffolds that can be used as antibiotics or anticancer drugs.

"As chemists, we can only examine a tiny portion of the current chemical universe," says Denis Fourches, an assistant professor of chemistry at NC State and a corresponding author of an article describing the work . "If we tried to synthesize and test all the chemicals of interest, it would take too much time and cost too much. So we have to use computers to explore these unknown parts of the chemical space. "

Chemists use computers by listing or virtually generating new molecules and comparing their predicted properties with those of existing drugs. This in silico screening method quickly and inexpensively identifies the compounds with the desired properties that experimental chemists can then synthesize and test.

Macrolides are a family of chemicals mainly used as antibiotics and anticancer drugs. Their unique ring structure allows them to bind to difficult protein targets. Some of them are considered last-resort medicines, especially for drug-resistant bacteria.

"Macrolides are natural products," says Fourches. "These chemicals are produced by bacteria to kill other bacteria, but the synthesis of these highly complex compounds takes 20 to 25 chemical steps, which is a long and costly process." If you want to find new compounds , simulation is by far the fastest way to do it. "

Forks and his colleagues have created a computer program called PKS Enumerator, which generates very large libraries of virtual chemical analogues of macrolide drugs. The software uses chemical building blocks extracted from a set of 18 known bioactive macrolides, breaking them down into its chemical components, then rearranging them to create new compounds based on a series of rules and constraints imposed on the user. The resulting library of new macrolides – V1M – classifies new compounds by size, weight, topology, and donors and acceptors of hydrogen bonds.

"We wanted to create a virtual library of completely new chemicals that no one probably had ever synthesized, but these compounds still needed to be sufficiently similar to the known macrolide drugs for this library to be relevant to the research community," says Fourches. "V1M is the first public domain library of these new macrolides, all of which are chemically similar to the 18 known bioactive macrolides we have analyzed, so hopefully other researchers will be able to use this library to analyze and identify certain compounds that may be useful. to the discovery of drugs. "

###

The search appears in the Journal of Cheminformatics. Phyo Phyo Kyaw Zin, a graduate student from the State of NC State, is the first author. Gavin Williams, an associate professor of chemistry at NC State, also contributed to the work.

Note to editors: A summary of the article follows.

"Enumeration and analysis of large libraries of macrolide scaffolds based on pathformatics"

DOI: 10.1186 / s13321-018-0307-6

Authors: Phyo Phyo Kyaw Zin, Gavin Williams, Denis Fourches, North Carolina State University

published: Journal of Cheminformatics

We report on the development of a counting technology in computer chemistry and the analysis of a large dataset resulting from sca? Olds virtual macrolides. Although macrolides have been shown to have interesting biological properties, there is no ready-to-use virtual library of various macrolides in the public domain. Molecular modeling (in particular virtual screening) of these complex molecules is very relevant because the organic synthesis of these compounds, where feasible, generally requires many synthetic steps and therefore considerably slows down the discovery of new bioactive macrolides. We present here an approach of computation and associated software allowing to design and to generate libraries of sca? Old virtual macrocycle / macrolide with user-defined constitutional and structural constraints (eg, types and number of structural motifs to include in the macrocycle, ring size, maximum number of compounds generated). To study the chemical diversity of such generated molecules, we have listed the V1M library (Virtual 1 million Macrolide sca? Olds), each containing twelve common structural motifs. For each macrolide, we calculated several key properties, such as molecular weight, hydrogen bond donors / acceptors, topological polar surface. In this study, we discuss (1) the initial concept and current characteristics of our investigator software PKS (polyketides), (2) chemical diversity and distribution of structural motifs in the V1M library, and ( 3) unique opportunities for future virtual projection. such listed sets of macrolides. It is important to note that V1M is provided in the additional material in this article, which allows other researchers to perform any type of molecular modeling and virtual screening studies. Therefore, this technology for counting very large libraries of macrolide scaffolds could offer unique potential in the field of computer chemistry and drug discovery for the rational design of new antibiotics and anticancer agents.

Warning: AAAS and EurekAlert! are not responsible for the accuracy of the news releases published on EurekAlert! contributing institutions or for the use of any information via the EurekAlert system.