High-throughput bioassay for mechanism of action determination of antibacterial drugs

While the 'war' on infectious diseases has been considered won, antibiotic-resistant bacteria are currently responsible for 25,000 death's yearly in Europe. No new broad-spectrum antibiotic has been introduced since the 1960s, and the last new class was discovered in 1986. As the antibiotic pipeline is clearly exhausted, new tools to advance antibiotic research are required. The current work explored Fourier-transform infrared spectroscopy to classify the mechanism of action of 13 antibiotics, acting by 3 distinct Mode-Of-Action (MOA) and belonging to 7 different classes. After optimization of a biological assay and pre-processing techniques, principal component analysis and partial least squares discriminant analysis were applied in a multi-level approach, including the MOA, antibiotic class and ultimately individual antibiotics acting on very specific molecular targets. Overall results indicate that the proposed method presents metabolic resolution to identify antibiotics at three levels of classification (i.e. different MOA, classes and even acting on specific targets). Interestingly, the resolution capacity obtained at these three levels of classification depended on the antibiotic type, which highlights the importance of the multi-level approach taken. Ultimately the present work reinforces the applicability of the method has a metabolic fingerprinting tool for antibiotic discovery.