Metabolomic Profiling of Soil Streptomyces: Open-Source Datasets for Microbial Interaction Studies

Abstract

The bacterial genus Streptomyces is renowned for its prolific production of bioactive secondary metabolites, playing a crucial role in soil microbial ecosystems. However, the molecular mechanisms underpinning interactions between these microbes remain poorly understood. With the aim of establishing a protocol for the characterisation of the molecular interplay between members of the same microbial community, we examine the molecular interplay between two model soil bacteria: Streptomyces coelicolor M145 and Streptomyces avermitilis MA-4680. We observed a distinct morphological response of M145 to the presence of MA-4680. To characterise the underlying signalling mechanism, we conducted metabolite profiling over a time course sampled at 3, 5 and 7 days. Samples were analysed with high-resolution mass spectrometry (HRMS), in both positive and negative ionisation mode, and converted into the open-source.mzML format. This resulted in 59 files for each ionisation mode. In addition, we examined knockout mutants deficient in primary signalling molecules—butyrolactones in M145 and avenolide in MA-4680. While butyrolactone deletion had no observable impact, the loss of avenolide production in MA-4680 altered the phenotype of M145 and corresponding metabolite profiles. The resulting HRMS datasets are presented here as 98 .mzML files (49 for each ionisation mode). All data were processed through our established computational pipeline for feature extraction and metabolite annotation, and the resulting data are made available as easily accessible files, together with the associated metadata. Co-culturing Streptomyces coelicolor with Streptomyces avermitilis leads to observable morphological changes, when compared with axenic cultures. Untargeted metabolomics data show clear metabolic changes in response to co-culture. The disruption of the intra-cellular signalling systems in the two species had different effects. While the disruption of scbA in S. coelicolor had no observable effect on the interaction, the disruption of aco in S. avermitilis changed the phenotype observed in the interaction plates and the associated metabolic signatures. The obtained datasets are publicly available in the Zenodo database [LINK once made public], adhering to the FAIR principles to support reanalysis, future compound identification, and benchmarking of untargeted metabolomics annotation workflows. For this purpose, the datasets include raw and processed data, including .csv files containing all the probabilistically annotated features as well as all the code used for the data processing, together with the instructions for the exploration of the datasets which have been made publicly available to support reusability. These datasets will serve as valuable resources for training and evaluating machine learning models in the context of microbial interactions in soil environments.