| dc.description.abstracteng | One of the major threats for the global health system is the increasing number of antibiotic resistant bacterial strains. The misuse of antibiotics during the Covid pandemic aggravates these problems. Therefore, the development of new antibiotics is mandatory to fight the rising threat of multi-resistant bacteria.
The essential second messenger c-di-AMP was discovered in 2008 and is mainly found in gram positive bacteria. It has been identified in several human pathogens like Listeria monocytogenes, Staphylococcus aureus or Enterococcus faecalis. It is a key player in the regulation of several pathways like DNA integrity scanning, cell wall metabolism or osmolyte homeostasis. Five different protein classes are able to synthesize c-di-AMP They all have the diadenylate cyclase domain (DAC) in common and need to dimerize to produce c-di-AMP in a metal-ion dependent manner from two ATP molecules. The protein class CdaA consists of three N-terminal transmembrane helices followed by the DAC domain and is often the sole diadenylate cyclase in several human pathogenic bacteria like Staphylococcus aureus, Streptococcus pneumoniae or Enterococcus faecium. The essentiality of c-di-AMP renders CdaA as potential target for the development novel antibiotic substances.
The major goal of this thesis was the establishment of a crystallization system of CdaA suitable for a fragment screening campaign in order to identify starting points for the development of inhibitors of CdaA. Therefore, the first part of this work is focused on the crystallization of CdaA from the soil bacterium Bacillus subtilis and the human pathogens Listeria monocytogenes, Staphylococcus aureus, Streptococcus pneumoniae and Enterococcus faecium. To obtain constructs well suited for crystallography, the CdaAs of these organisms were N- and C-terminal truncated to solely consist of the DAC domain. The purity and homogeneity of every purified CdaA were verified via SDS-PAGE and DLS. To prove the enzymatic functionality of the purified proteins, the coralyne activity assay was applied with varying divalent cations. All purified CdaAs exhibit activity in the presence of manganese ions while their activity in presence of magnesium- or cobalt-ions differs. To explore these variations, crystallization trials of the different CdaAs in presence of their favorable metal-ion and ATP were carried out.
As a result, the crystal structure of Enterococcus faecium CdaA in complex with c-di-AMP and manganese as well as the crystal structure of Streptococcus pneumoniae CdaA in complex with c-di-AMP and magnesium were solved at a resolution of 2.1 Å (EfCdaA) and 2.2 Å (SpCdaA). Unfortunately, both models represent a post catalytic state where the metal ion is not coordinated in a catalytically active way, giving no further insights into the structural basis for the differing metal ion specificity. Nevertheless, if both obtained structures are compared to the Listeria monocytogenes CdaA structure in complex with c-di-AMPa conserved interaction pattern with c-di-AMP could be observed, rendering the interacting amino acids as possible targets for inhibitor design.
In order to identify potential fragments which can reduce the activity of CdaA, the CdaAs from all previously mentioned organisms were applied to crystallization trials in their APO state. For the CdaA of Enterococcus faecium and Bacillus subtilis these trials were successful. However, the crystallization system for Enterococcus faecium CdaA does not matches the requirements for a crystallographic fragment screening campaign as the crystals diffracted just to 2.4 Å and required several months to grow. In contrast, the Bacillus subtilis APO crystals diffracted up to 1.6 Å and were highly reproduceable. Further optimization of the crystallization condition led to a crystallization system which is greatly suitable for a fragment screening campaign. Through the following fragment campaign using the F2X-Entry screen, 32 unique fragments were identified to interact with BsCdaA. Mapping these fragments on a Consurf model of CdaA reduced the number to six different fragments which bound in the highly conserved active/dimerization site. These fragments were also biochemically characterized using the coralyne and malachite green assay. As none of the fragments reduced the cyclase activity of BsCdaA, an alternative approach was carried out to identify an inhibitory compound.
Here, the structure of BsCdaA in complex with AMP, which were determined in this work served as starting point for the identification of an inhibitory compound. By a computational defragmentation approach employing SeeSar, adenine was identified as fragment of AMP which is largely responsible for the biding capability of AMP. Combing the methods of computer aided drug design and structural information gained by X-ray crystallography, the Janus kinase inhibitor Ruxolitinib was identified as potential inhibitory compound for CdaA. The conserved binding mode of Ruxolitinib towards CdaA was proven as Ruxolitinib exhibits the same interaction pattern for CdaA from Bacillus subtilis and Listeria monocytogenes as it can be concluded by the X-ray structures obtained in this work as Ruxolitinib binds in the active site of the protein, it most likely acts as competitive inhibitor for CdaA. Moreover, the IC50 value of Ruxolitinib was determined as 2.7 µM for BsCdaA. Determining the cyclase activity of CdaA from Listeria monocytogenes, Staphylococcus aureus, Streptococcus pneumoniae or Enterococcus faecium in presence of Ruxolitinib also show a reduced activity, underlining the conservation of the protein-ligand interaction. Besides the in vitro studies, also in vivo experiments utilizing different Bacillus subtilis strains were carried out. The results from these experiments suggest a high specificity of Ruxolitinib towards Bacillus subtilis CdaA in vivo. | de |