Carbonic Anhydrase II: A Model System for Artificial Copper Center Design, Protein-guided Cycloadditions, Tethering Screenings and Fragment-based Lead Discovery

In this thesis a variety of quite different fragment-based lead discovery approaches have been applied to the target protein carbonic anhydrase II. The different projects were strongly supported and methodologically tailored towards protein crystallography; a method which is currently emerging as a routine analytical tool. This maturation mainly results from improved radiation sources and enhanced computing power. About 200-250 datasets were collected in due course of this thesis answering miscellaneous questions in the context of several ambitious projects to be answered by structural biology. How can interactions of proteins with small molecules be exploited for drug development? Are we able to, e.g. monitor in situ reactions using protein crystallography? Will a tethering approach fixing weak initial binders increase the success rate of fragment-based drug discovery? In order to cover this broad scope of diverse projects access to large amounts of CA II is a prerequisite. Therefore, the gene coding for CA II was cloned into a high-level expression system, the GST-tagged gene fusion system. The expression system was optimized yielding target protein of up to 30 mg from 1 L of original cell culture. Several CA II mutants were produced via site-directed mutagenesis and a mercury-free protein crystallization protocol had to be established for either the wild type and all mutants. Initial goal of the first project of this thesis was the accomplishment of a prominent click chemistry reaction, the Huisgen reaction – a Cu+ catalyzed 2+3 dipolar cycloaddition – within or next to the active site of CA II. Therefore, an artificial copper center had to be introduced at the surface of CA II which subsequently should be used by azide and alkyne building blocks to catalyze triazole formation. Unfortunately, the embarked rational design strategy remained unsuccessful which underlines that our still rather limited understanding of amino acid exchanges on the architecture of proteins might provoke unpredictable effects while tampering with complex biological systems such as a highly functionalized proteins. Finally, more by serendipity than design the formation of an artificial Cu center at the surface of CA II could be achieved. In parallel to the formerly described metal center design, the click chemistry project was also promoted by a new tether-assisted approach. By covalently attaching one building block to the protein surface and linking the second reactant via a reversible sulfonamide anchor to the active site Zn2+ ion, both components were brought into vicinity favorable to undergo cycloaddition reaction. The reaction could be induced by either the exposure to Cu+ ions or the steric constraints of the environment of CA II, respectively. Remarkably, by the protein environment the triazole formation was executed with regio- and stereoselectivity. The 1,5-S-triazole was selectively formed from a racemic building block mixture. Apart from this click chemistry approach, the tethering method was also applied to fragment-based lead discovery of CA II. Tethering allows the identification of fragments with rather weak affinity to the target protein. Suitable candidates for covalent fixation are extracted from a library of thiole containing compounds under reducing conditions. A computational docking study was applied to screen a virtual library and select a feasible number of compounds for synthesis. The "in solution" experiments monitored by HPLC-MS revealed many candidate fragments with promising affinity to CA II. The crystallographic analysis of several CA II-H64C tethered complexes showed that the Saccharin binding pocket at the protein surface could be successfully addressed by the tethering approach. Also a second surface exposed binding area, known from CA II activators could be screened. In an alternative fragment-based lead discovery approach, four novel head groups for zinc coordination have been tested on CA isoforms. A complex structure of CA II and 1,2-HOPTO revealed a novel Zn2+ binding mode with favorable interactions to Thr199 and Thr200. The fragment which is perfectly coordinated by a water network suggests a new class of putative CA inhibitors. The binding affinity can be improved by adding substituents to occupy to the hydrophilic and hydrophobic portion of the CA II binding pocket.