Dottorato in BIOCHIMICA

Titolo della tesi: Identification, analysis and inference of point mutations associated to drug resistance in bacteria: a lesson learnt from the resistance of Streptococcus pneumoniae to quinolones

Antibiotic resistance is one of the biggest public health challenges of our time. Bacterial chemoresistance is the phenomenon whereby bacteria develop the ability to survive and multiply in the presence of an antibacterial drugs; the expression of a resistant phenotype may be due to three fundamental mechanisms, including the expression of enzymes that inactivate the antibacterial drug, changes in the membrane permeability to antibiotics and the onset of point mutations causing the physical-chemical alteration of the antimicrobial targets. In recent decades, new antibiotic resistance mechanisms have emerged and are spreading globally, threatening human health and the ability to fight the most common infectious diseases. Quinolones, a new class of antibiotics that bind bacterial topoisomerases and inhibit cell replication, have been important in limiting the spread of penicillin- and macrolides-resistant Streptococcus pneumoniae. However, alarmingly, resistance to quinolones is spreading recently. Resistance is caused by the appearance of point mutations in the bacterial topoisomerase and gyrase. Some mutations are well known, but some are not and the information about known molecular mechanisms causing resistance is sparse and not systematically collected and organised. This means that it cannot be used to infer new mutations in newly sequenced bacterial genes and study how they may affect the drug binding. The lack of structured, organized and reusable information about point mutations associated with antibiotic resistance represents a critical issue and is a common pattern in the field. Here, we present a structural analysis of the mutations involved in the resistance to quinolones affecting the gyrase and topoisomerase genes in Streptococcus pneumoniae. Results, extended to other bacterial species, have been collected in a database and can now be used – through a web server - to analyse both known and yet unknown mutations occurring in bacterial topoisomerases and gyrases. Furthermore, structural data about point mutations associated with antibiotic resistance were used to train, test and validate a machine learning algorithm for the inference of still unknown mutations potentially involved in bacterial resistance to quinolone. As the performance of the algorithm, measured in terms of accuracy, sensitivity and specificity, is very promising, we plan to incorporate it in the web server to allow users to predict new mutations associated with bacterial resistance to quinolones.