In vitro and in vivo characterization of darobactin derivatives

Abstract

Antibiotics are our major weapon to fight infectious diseases. Many of today’s standard medical procedures, such as organ transplants or cancer treatments, are only possible through the use of antibiotics, and they have drastically increased the average human lifespan. But we are about to lose this huge advantage as we are facing an antimicrobial resistance crisis. Some microorganisms are no longer susceptible to antibiotics, making infectious diseases very difficult to treat. Alongside to tuberculosis, the most problematic pathogens are Gram-negative bacteria, like Acinetobacter baumannii, Escherichia coli, Klebsiella pneumoniae and Pseudomonas aeruginosa. Since not many novel antibiotics have been discovered since the so-called ‘golden age’ of antibiotic discovery, we are in urgent need of new treatment options. A lot of the antibiotics discovered during this period were natural products, which still represent a source of great potential for the discovery of new antibiotics. The discovery of the natural product darobactin A (DAR A) is a good example, marking a major step forward in the fight against the looming antimicrobial resistance crisis. The compound shows activity against all the important Gram-negative pathogens and has a novel mode of action with no known cross-resistance to antibiotics available on the market. By targeting BamA, a part of a complex in the outer membrane of Gram-negative bacteria responsible for the folding and integrating of outer membrane proteins, DAR A acts on the outside of the bacterial cell wall. This avoids the major problem of antibiotics having to cross both the inner and outer cell membrane to reach a target inside of Gram-negative bacteria. With DAR A opening up a new class of BamA inhibitors, many derivatization studies were initiated to optimize the structure of this compound. In addition to exchanging amino acids in the sequence of the cyclic heptamer, genome mining approaches have been used to identify other DAR producers and find natural derivatives. In the first project of this work, this approach led to the discovery of the DAR biosynthetic gene cluster in P. luteoviolacea H33, which was found to contain additional genes next to the ones encoding for DAR. This strain produces three new DAR variants: bromodarobactin, dehydrodarobactin and dehydrobromodarobactin. One of these genes revealed to be a flavin-dependent tryptophan halogenase with a novel fold. This halogenase, DarH, brominates the N-terminal tryptophan of DAR A to form bromodarobactin. The enzyme was purified and further characterized. In vitro and heterologous expression studies in vivo showed that DarH can also brominate a DAR derivative and catalyze the iodination of tryptophan. Furthermore, activity assays with bromodarobactin revealed increased activity and plasma binding compared to DAR A. To further increase the activity of DAR A or to optimize the pharmacokinetics of the compound, in the second project of this dissertation a technique was developed to replace the natural amino acids by non-canonical amino acids. This was achieved by adapting the amber stop codon suppression technique for the production of DAR A derivatives. An altered aminoacyl-tRNA synthase that accepts non-canonical amino acids is used in combination with a tRNA that recognizes a nonsense codon. This nonsense codon is integrated into the biosynthetic gene cluster at the desired position of exchange by mutation. In the case of DAR A, the codon at position seven of the heptapeptide has been changed, resulting in the production of darobactin A F7F with a 4-fluoro-L-phenylalanine, darobactin A F7I (4-iodo-L-phenylalanine), darobactin A F7F5 (2,3,4,5,6-pentafluoro-L-phenylalanine) and darobactin A F7OMe (4-methoxy-L-phenylalanine). Darobactin A F7F could be purified in large scale, was characterized by NMR experiments and showed a similar activity compared to DAR A in activity assays.