Ribopyranosylierte Peptide als Bausteine für den reversiblen Aufbau von Bisubstrat-Inhibitoren für die Proteinkinase A

The development of bisubstrate inhibitors covers the aspect of the methods of fragment-based drug design as well as the transition state analog inhibitor synthesis. Two fragments, which are e. g. connected by a linker, simultaneously mimicking the natural substrates or inhibitors of an enzyme to increase the interactions by double addressing of both binding sites. Furthermore the intended increase of selectivity for a certain enzyme by using bisubstrate inhibitors has the advantage of a more specific binding to a certain protein kinase compared to the use of easier inhibitors. In this study, the development of such bisubstrate inhibitors was one aspect that was taken care of. Another aspect with opposing intentions concerning the increase of selectivity of the inhibitor was the use of covalent reversible chemistry for the linkage of both substrates/inhibitors to give the chance of transferability to other kinases, because inhibitors for different kinases usually bear huge similarities among each other. As a result the linkage between the substrate and the inhibitor was not incorporated in advance but they were intended to “find each other” during crystallization approach through their sterical proximity and affinity reasons. It was thought to find out, if the entropical gain caused by the release of water during the reaction of the diol and the boronic acid can overcome the unfavored esterification under aqueous conditions. Prior to performing crystallization approaches and affinity tests with these bisubstrate inhibitors in the pharmaceutical department in the working group of Prof. Klebe, suitable building blocks that form the linkage between them, had to be synthesized. This was done in dependence on the computer based calculations with the aid of already measured crystal structures of PKA. First the syntheses of different ribopyranosylated amino acid building blocks were performed by using amino acids bearing several hydroxy- and thiol-functions in the side chain as well as tetraacetyl protected ribopyranose, putting them together through lewis acid mediated glycosylation reactions. In this manner a total of six different RAA building blocks with partially good yields have been synthesized. The yields received in the respective syntheses were all providing enough product to be able to use it in manual as well as in automatic solid phase synthesis. Furthermore the transfer of the conditions used for the RAA-incorporation into short test-peptides to larger peptides was easily manageable, so several PKI peptides bearing RAA mutations at different positions near the active side were synthesized. For the synthesis of the inhibitor, Fasudil® was being used as chemical lead because of known KD-values for this inhibitor in a nanomolar range. Therefore simplified sulfonamide building blocks were initially synthesized to study their general ability for being linked to boronic acids. Beginning with a system where the linkage of the inhibitor was performed with the aldehyde function in formylphenylboronic acid via its imination with a free amine, the reaction was switched to another one providing a more stable linkage. Henceforth the linkage was being introduced by the reductive amination of formylphenylboronic acid with the corresponding primary or secondary amine. With this the second building block - the “borono-fasudil” with different variations was successfully finished and available for esterification reactions with the sugar building blocks. The linkage tendency with boronic acids was tested with each RAA building block in advance (using the corresponding tetrapeptides) using pyrenboronic acid on a NMR scale in DMSOd6. This was done to receive NMR spectra with lower complexity enabling detailed studies of the binding mode and the stability of each RAA building block linked to boronic acid. Finally the transfer of the linkage-experiments to the bisubstrate-inhibitor-system of the PKA in aqueous buffer was arranged, hoping for the formation of crystals to observe the esterification product. As a matter of fact crystal structures were received, which did not show the desired linkage but a distance between both components that could not be overcome by esterification (PDB-Codes: see unpublished dissertation Janis Müller). Similar results were obtained with the experiments with “borono-fasudil” of the second generation. The obtained crystal structures confirmed that no linkage between the sugar component and the boronic acid-modified inhibitor in the protein environment took place, because the participating functional groups are too far away from each other and the surrounding protein additionally decreases the mobility of both components. Consequently the examined system needs to be optimized, where the boronic acid component would be the more promising starting point as it holds more options for modification. A change of the ribose residue in the amino acid building block against another sugar would also be possible (experiments using fructose were performed) but would probably have a major impact on the general synthesis. Other protecting groups for the sugar moiety and the amino acids might become necessary or even a change of the glycosylation method. What basically could be proven is the tolerance of PKI sequences for mutations with RAA building blocks at different positions, which could be indicated by the obtained KD-values that still covered proper affinity ranges (except from PKI). At this time the crystallization approaches and affinity measurements had not been completely finished, therefore some results were still outstanding. Further studies would require the transfer of the whole system to aqueous conditions. While the reactivity in DMSOd6 was assigned to be quite high, no reactivity could be seen for the esterification in water. Since the desired linkage has to be performed inside the protein and water/buffer respectively, the described system needs optimization of both components to that effect that a transfer of the reaction into an aqueous system is possible. Therefore the ribose moiety inevitably has to be changed to another sugar, because it was shown in numerous experiments that the pyranoide conformer of ribose does not react with boronic acids in water, even though benzoboroxols were used, which exceptionally react in water. Moreover the tolerance of other peptide sequences referring to RAA building blocks were tested and experiments were initiated for the simultaneous incorporation of boronic acids and RAA building blocks into the same peptide for reasons of inducing intramolecular cyclisation. For example the β-hairpin of the foldon peptide showed a high tolerance for the incorporation of one as well as of two RbS-mutations into the peptide chain, which was indicated by high values for the folding population for both mutated peptides. Furthermore the filaggrin peptides having RbS mutations at both serine positions as well as the one with a RbT mutation at the threonine position showed positive results concerning their affinity towards autoantibodies of rheumatoid arthritis, which was observed with ELISA tests. The approach of intramolecular cyclisation reactions using different boronic acids with RbS building blocks in various heptapeptides and finally in β-hairpin peptides all showed the desired conversion leading to the cyclized products, but mostly with incomplete conversions. Generally spoken this method seems to be suitable to induce the intramolecular cyclisation in peptides. A similar approach covers the formation of macromolecules via the linkage of different peptides to each other. Therefore one of the peptides has to bear a boronic acid and another one a RAA building block. The tested system in this study consisted of a cutout of the Aβ sequence of prion peptides (with boronic acid coupled to side chain of lysine) and a cutout of the PKI sequence (with RbS mutation), which also gave the desired linkage, proven by NMR spectroscopy. In this way the combination of two different peptides was accomplished providing a peptide with a molecular weight of > 3000 g/mol just by combining two peptides in a NMR tube in DMSOd6. Both last-mentioned methods contain plenty of potential for the synthesis of cyclic peptides and the formation of macromolecules, because of the easily feasible incorporation of the linking-components into the corresponding peptides during SPPS and the reversibility of the esterification reaction. Another tracked project was the synthesis of phenylalanine derivatives via the Miyaura-borylation using aryl boronic acid compounds. In doing so two amino acid building blocks (4-iodo phenylalanine and 4,4’-bisphenylalanine) were successfully synthesized and were incorporated into the β-hairpin of the foldon and afterwards into the foldon sequence. The following NMR spectroscopic analysis partly showed an increased stability for the β-hairpin structure compared to the native sequence, which gave this topic a lot of potential for further investigations on the incorporation of such amino acid building blocks into the foldon peptide.