Screening of annotated compound collections across a panel of assays can provide bioactivity profiles which act as maps for compounds of unknown function (see “Biographies for Compounds,” Assay Drug Dev Technol 2012;10:309–310). In this article,* the mode of action for antibiotics were mapped across 15 bacterial strains (six gram-positive and nine gram-negative strains) using 72 commercially available antibiotics.
Profiling these antibiotics as dose–response curves provided the minimum inhibitory concentrations (MICs) for each antibiotic in every bacterial strain. Clustering of this data revealed 10 classes of antibiotics, and six of these (fluoroquinolones, rifamycins, resistomycins, amphenicols, aminoglycosides, and nitrofurans) had clusters in which the activity of cluster members showed a clear contiguous pattern. Certain drugs, like the sulfa drugs, were all too weak to measure in this assay, which used 100 μM as the highest testing concentration.
Compounds of similar structures clustered together, and compounds that target the same biological component with different structures were found in different clusters. For example, both aminoglycosides and tetracyclines target the 30S ribosomal subunit but were found in different clusters. This is thought to reflect the differences in compound uptake in bacterial strains; therefore, the term “BioMAP” reflects both the pharmacokinetic and mechanism of action properties of these antibiotics.
With BioMAP in hand, the investigators characterized the activity of partially purified microbial natural product extracts. One of the issues with identifying new antibiotics from natural product collections is that often the same antibiotics are identified after considerable effort to purify the active component from the extracts. Here, BioMAP was used to classify the MIC activity of the microbial extracts (Figure), which allowed subsequent labor-intensive purification efforts to focus on novel activity. The BioMAP was internally normalized to the MICs to make the profile concentration-independent, an important feature because the concentration of the active component in a natural product extract is unknown. Using this approach, the authors were able to identify a novel antibiotic, which they termed Arromycin, a unique napthoquinone.
The approach described provides a method to characterize natural product extract activity and prioritize fractions for purification and structure elucidation.
*Abstract from Chemistry & Biology 2012, Vol. 19: 1483–1495
Despite recognition of the looming antibiotic crisis by healthcare professionals, the number of new antibiotics reaching the clinic continues to decline sharply. This study aimed to establish an antibiotic profiling strategy using a panel of clinically relevant bacterial strains to create unique biological fingerprints for all major classes of antibiotics. Antibiotic mode of action profile (BioMAP) screening has been shown to effectively cluster antibiotics by structural class based on these fingerprints. Using this approach, we have accurately predicted the presence of known antibiotics in natural product extracts and have discovered a naphthoquinone-based antibiotic from our marine natural product library that possesses a unique carbon skeleton. We have demonstrated that bioactivity fingerprinting is a successful strategy for profiling antibiotic lead compounds and that BioMAP can be applied to the discovery of new natural product antibiotics leads.