A high degree of taxonomic and chemical redundancy is a major limitation in sourcing microbial strain libraries for drug discovery. Currently, the creation of these libraries relies on outdated and costly methods, namely visual inspection of morphological differences of colonies from agar plates, or ribosomal RNA gene sequencing methods that are not indicative of a microbe?s capacity to produce specialized metabolites (SM). Despite the incredible potential of microorganisms to produce SMs, this redundancy remains a primary barrier to drug discovery efforts. In order to overcome this, we will develop a rapid, easy to use mass spectrometry (MS) technique that will maximize both the taxonomic and chemical diversity entering into microbial strain libraries. This will be coupled to our semi-automated, web-based bioinformatics pipeline that will be made available to the public. Our platform will employ matrix-assisted laser desorption/ionization time of flight MS (MALDI-TOF MS) to address a few major obstacles in the drug discovery process. First, we will develop a high-throughput MALDI- TOF MS method capable of gathering two distinct datasets from single colonies of bacteria from agar-based diversity plates: a) ribosomal protein fingerprints that are used to putatively identify the genus and species of the colony, and b) SM fingerprints of each colony to elucidate intra-species differences in SM production (Aim 1). Importantly, our MALDI-TOF MS platform is capable of processing and analyzing 384 strains in a 4-hour period, which is a significant advance when compared to other mass spectrometry or genomics-based profiling approaches. When applied to thousands of bacterial colonies of a cultivatable environmental microbiome, this platform will serve to maximize the taxonomic and chemical diversity entering a library, while minimizing the number of strains required to achieve this (e.g. addition of 300 strains as opposed to 3,000). Second, we will develop a facile fluorescence/MS-detection platform to interrogate the unmined biologically active chemical space of existing bacterial strain libraries (Aim 2). Using an existing Actinobacteria library as proof of concept, we will grow each strain under eight different cultivation conditions in 48-well agar plates. We will then develop and implement a series of antibiotic assays with fluorescent reporter strains of ESKAPE pathogens, and use MALDI-TOF MS to detect biologically active SMs that exist within zones of inhibition from each producing actinomycete. This method allows researchers to simultaneously observe growth inhibition via fluorescence imaging and to identify strains that produce bioactive SMs under single/unique cultivation conditions. This foregoes laborious liquid cultivation and chromatography steps of inactive bacteria (current practice). Data analysis will be facilitated through development of a web-based, semi-automated visualization pipeline that will be freely available to the scientific community (Aim 3). Successful completion of these aims will result in more a targeted, cost efficient, and accessible approach to microbial drug discovery, and represents a major innovation to front end microbial library generation that has arguably not seen an advance in decades.
A high degree of taxonomic and chemical redundancy is a major limitation in sourcing microbial strain libraries for drug discovery. In the current proposal, we will introduce a major innovation to the manner in which bacteria are classified and added to microbial libraries, and will provide a web-based bioinformatics tool that will help researchers overcome these shortcomings.
