Entwicklung von Inhibitoren wirtseigener Serinproteasen zur antiviralen Therapie

The development of novel strategies for antiviral therapy is of crucial importance, as many viruses represent a significant threat to global health. After entering their host, most viruses require the action of host proteases to replicate. Hence, the inhibition of such proteases can decrease the viral spread. This work focused on the design, synthesis, and characterization of inhibitors for the host proteases matriptase and S1P in the context of potential novel antiviral agents. Project A: Matriptase inhibitors The membrane-bound trypsin-like serine protease matriptase can cleave the hemagglutinins of certain influenza viruses, thereby potentially contributing to viral replication. In our working group, a multitude of matriptase inhibitors of the 3-amidinophenylalanine (Phe(3-Am)) type have been developed over the past years. In this doctoral thesis, the aim was to design and synthesize mainly monobasic matriptase inhibitors of the Phe(3-Am)-type with a high selectivity against the clotting proteases thrombin and factor Xa, as inhibiting the latter enzymes can potentially cause bleeding complications. This was supposed to be achieved by thorough analysis of the active sites of all three enzymes and the incorporation of docking studies. Previous compound series had provided two monobasic matriptase inhibitors (5 and 28) with a high selectivity against thrombin (SF 126 and 410, respectively) that was linked to a putative steric clash between the 60 insertion loop of thrombin with the inhibitors’ C-terminal cyclohexyl ureido group. In this work, the C-terminal piperidide as well as the ureido moiety were derivatized, ultimately revealing a piperidyl-4-tert-butyl urea as the optimal C-terminal group to achieve a high selectivity against thrombin. In combination with appropriate substituents on the N-terminal biphenyl system, several new monobasic derivatives with such selectivity were developed. The highest selectivity was achieved for monobasic compound I5 (SF 1533) and the dibasic inhibitor I20 (SF 3643). While the C-terminus was modified to decrease affinity for thrombin, the substituents on the N-terminal biphenyl system were intended to provide an enhanced selectivity against fXa. This clotting protease features a Phe174 residue in its S3/4 binding region, where matriptase carries a Gln in the analogous position (Gln175). Based on the crystal structure of a related Phe(3-Am) inhibitor in complex with matriptase (PDB: 2GV6), docking studies were performed to identify N-terminal substituents that are potentially able to address Gln175 of matriptase, thereby improving only the affinity for matriptase and not for fXa. These studies suggested 3- and 4-hydroxymethyl groups as appropriate substituents for this interaction with Gln175, providing the basis for all N-terminal variations performed in the context of this work. However, the obtained inhibitors did not exert the desired selectivity against fXa. Of all 24 new derivatives, only the dibasic compound I21 (SF 105) shows a decent selectivity. Nevertheless, four new crystal structures of three different inhibitors (I5, I12 and I20) in complex with matriptase, thrombin and trypsin were obtained, proving the presence of the predicted interaction between Gln175 and the 4-hydroxymethyl group of inhibitor I5. Moreover, the structural information provided by these complexes allowed the establishment of more detailed structure-activity-relationships that could aid the further development of this inhibitor type. Selected inhibitors demonstrated a strong inhibitory effect on the replication of a matriptase-dependent H9N2 influenza virus strain. Project B: S1P inhibitors The site-1 protease (S1P) is a member of the proprotein convertases and involved in the regulation of intracellular cholesterol and fatty acid levels. Furthermore, it cleaves the glycoprotein precursor of all known arenaviruses pathogenic to humans and therefore represents an interesting target for antiviral therapy. So far, only a single article about potent, reversible S1P inhibitors with in vitro activity has been published and all following studies concerning the effect of S1P inhibition on various biological systems have been performed with compound PF-429242 that was described in said article. In this work, peptidic boronic acids based on the substrate sequence RRLL↓ and deborylated analogues were established as a new S1P inhibitor class and further derivatives of compound PF-429242 were synthesized. The synthesis of peptidic boronic acid inhibitors was optimized and extensive studies were performed regarding the establishment of boronic acid SPPS. The enzyme kinetic characterization revealed peptidic boronic acids as highly potent S1P inhibitors and strongly suggests that the boronic acid moiety covalently interacts with the active-site serine. Compound I34 bearing a boro-cyclopentylglycine in P1 position binds to S1P with an IC50 value of 73 nM and therefore fivefold stronger than the reference inhibitor PF-429242 (IC50 = 374 nM). In case of PF-429242 derivatives, the reaction conditions for both reductive aminations required to synthesize such inhibitors were optimized. However, none of the new derivatives of the published compound PF-429242 exhibited a higher S1P affinity than the reference itself.