Leitstruktursuche mithilfe eines kristallographischen Fragment-Screenings zur Inhibierung des Klasse-II-Chaperons IpgC aus Shigella flexneri

Shigellosis is an inflammatory disease of the colon caused by a bacterial infection with Shigella. To develop pathogenicity, the bacteria have to penetrate the epithelial cells of the colon. IpgC – a class II chaperone – is essential for the active infection of endothelial cells due to its multiple interactions with invasion proteins. The focus of this thesis was to find a lead structure for the inhibition of IpgC function. The first step was to find a crystallization condition for IpgC. Initially a protein variant shortened by four amino acids at the C-terminus was used. With this variant a crystallization condition could be found which however resulted in crystals that had a too low scattering power, most likely due to high solvent content. Another variant, which was additionally shortened by nine amino acids at the N terminus, was used for the crystallization experiments. With this protein variant crystals with high scattering power could finally be obtained. With a resolution of 1.58 Å a first high-resolution crystal structure of IpgC was obtained. Compared to already deposited IpgC-structures, a previously unobserved dimer arrangement of the protein was obtained. This could be of importance in the future, because IpgC also exists in solution as a dimer. With the same crystallization condition, it was possible to perform a crystallographic fragment screening using an in-house fragment library comprising 96 structurally diverse fragments. Ten fragment hits were identified which had bound to the protein at a total of five different sites. Most of the fragments had bound in a pocket in the interface of the dimer. The 96 fragments were again used to perform a thermal shift assay. No fragments were found that increased the melting point of the protein that would have indicated a stabilizing effect. Instead, 26 fragments were identified that lowered the melting point of the protein. These included some of the crystallographic fragment hits which had bound to the dimer interface. This was explained by the fact that some of these fragments destabilize the dimer by breaking hydrogen bonds between chain A and chain B of the protein. The fragments found in the dimer interface were considered as a good starting point for further drug development. In silico docking was used in order to identify potential follow-up compounds. By using only commercially available compounds for docking, synthesis of the compounds could initially be avoided. Based on good docking results and using three different fragment hits as a starting point, 17 follow-up compounds were selected. On the basis of soaking experiments and subsequent crystal structure determination, the follow-up compounds that actually form a complex with IpgC were to be determined. Two compounds could be clearly identified as hits. Another compound represents a further potential hit. Two of these compounds (U08 and U11) had bound in interface pocket 1, as expected. One of these two follow-up compounds (U11) confirmed the binding mode observed for fragment J20 and filled the hydrophobic region of the binding pocket much better compared to J20. Contrary to expectations based on docking, the third compound (U09) occupied a completely different binding site. It is the only known binding site of an interaction partner so far, namely of IpaB. Spa13 – part of the type III secretion system – was to be cloned, recombinantly produced and purified in order to determine the three-dimensional structure of the IpgC-Spa13 complex by X ray crystallography. A protocol for heterologous production and purification of Spa13 was established. Using an MBP tag, Spa13 could be produced and partially purified in soluble form. However, Spa13 was found to be present in a highly aggregated form that made purification difficult. The coproduction of Spa13 together with its interaction partner IpgC was not successful either. To still obtain information about the interaction of IpgC with Spa13, a peptide microarray was performed. This was to determine the binding epitope of Spa13. However, the results were not conclusive. Only one of three positive controls was positive. For two of the Spa13 peptides a very weak signal could be obtained. However, these peptide sequences could not be confirmed as binding epitopes in initial experiments by MST. Co-crystallization experiments were performed to determine the crystal structure of IpgC together with the binding epitopes of MxiE and IpaC. Concerning the co-crystallization of IpgC with the binding epitope of MxiE crystals with a good scattering power could be obtained several times, but in each case the respective structure revealed that the peptide had not bound. Using the published condition for co-crystallization of IpgC with the IpaB binding epitope, the co-crystallization of IpgC with the binding epitope of IpaC was successful. Here, however, the collected data set to 2.62 Å resolution could not be resolved for unexplained reasons.