Bioinformatic strategies to understand the complexities of medicinal natural product mixtures

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Lindsay K. Caesar (Creator)
The University of North Carolina at Greensboro (UNCG )
Web Site:
Nadja Cech

Abstract: Compounds from natural sources, as well as those inspired by them, represent the majority of small molecule drugs on the market today. Plants, owing to their complex biosynthetic pathways, are poised to synthesize diverse secondary metabolites that selectively target biological macromolecules. Despite the vast chemical landscape of botanicals and other natural products, drug discovery programs from these sources have diminished due to the costly and time-consuming nature of standard practices and high rates of compound rediscovery. Additionally, natural product mixtures are incredibly complex, and the standard reductionist approaches often ignore the presence of combination effects such as synergy and antagonism. Bioinformatics tools can be used to integrate biological and chemical datasets, and statistical analyses of these datasets are broadly termed “biochemometrics.” Biochemometric approaches enable researchers to predict active constituents early in the fractionation process and to tailor isolation efforts toward the most biologically relevant compounds. Throughout the course of this project, bioinformatics approaches were used to (1) discover biologically active constituents from the botanical medicines, (2) develop and improve data filtering, data transformation, and model simplification parameters to optimize biochemometrics models, and (3) produce a new approach capable of predicting mixture constituents that contribute to synergy, additivity, and antagonism in complex mixtures. The first goal was achieved by applying bioassay-guided fractionation, biocheomometric selectivity ratio analysis, and molecular networking to comprehensively evaluate the antimicrobial activity of the botanical Angelica keiskei Koidzumi against Staphylococcus aureus. This approach enabled the identification of putative active constituents early in the fractionation process, and provided structural information for these compounds. A subset of chalcone analogs were prioritized for isolation, yielding antimicrobial compounds 4-hydroxyderricin, xanthoangelol, and xanthoangelol K. This approach successfully identified a low abundance compound (xanthoangelol K) that has not been previously reported to possess antimicrobial activity. Two studies were undertaken to achieve the second goal. First ,we demonstrated the effectiveness of hierarchical cluster analysis (HCA) of replicate injections (technical replicates) as a methodology to identify chemical interferents and reduce their contaminating contribution to metabolomics models. Pools of metabolites were prepared from the A. keiskei and analyzed in triplicate using ultraperformance liquid chromatography coupled to mass spectrometry (UPLC-MS). Before filtering, HCA failed to cluster replicates in the datasets. To identify contaminant peaks, we developed a filtering process that evaluated the relative peak area variance of each variable within triplicate injections. This filtering process identified 128 ions that did not show consistent peak area from injection to injection that likely originated from the UPLC-MS system. When interferents were removed, replicates clustered in all datasets, highlighting the importance of technical replication in mass spectrometry-based studies and providing tool for evaluating the effectiveness of data filtering prior to statistical analysis. As a follow up study, the impact of data acquisition and data processing parameters on selectivity ratio models were assessed using an inactive botanical mixture spiked with known antimicrobial compounds. Selectivity ratio models were used to identify active constituents that were intentionally added to the mixture, as well as an additional antimicrobial compound, randainal, which was masked by the presence of antagonists in the mixture. This study revealed that data processing approaches, particularly data transformation and model simplification tools using a variance cutoff, had significant impacts on the models produced, either masking or enhancing the ability to detect active constituents in samples. This study emphasized the importance of data processing for obtaining reliable information from metabolomics models and demonstrates the strengths and limitations of selectivity ratio analysis to comprehensively assess complex botanical mixtures. Often, analytical tools aimed to assess biological mixtures ascribe the activity to a few known components. Although researchers recognize this as an oversimplification, research methodologies to address this problem have not been developed. To overcome this and to achieve the third goal of this project, a new approach called Simplify was developed that can both identify mixture components that contribute to biological activity and characterize the nature of their interactions prior to isolation. As a test case, this approach was applied to the botanical Salvia miltiorrhiza and successfully utilized to identify both additive and synergistic compounds. These findings illustrate the efficacy of this approach for understanding how natural product mixtures work in concert and are expected to serve as a launching point for the comprehensive evaluation of mixtures in future studies.

Additional Information

Language: English
Date: 2019
Biochemometrics, Drug discovery, Mass spectrometry, Metabolomics, Natural products, Synergy
Medicinal plants
Natural products
Mass spectrometry

Email this document to