The integrity of our genome is continuously endangered by DNA damaging factors. Several cellular mechanisms have evolved to recognize and remove different types of DNA lesions. Despite the wealth of information on the three-dimensional structure and the catalytic mechanism of DNA repair enzymes, the essential process of target site search and identification remains more elusive. How can a small number of repair proteins find and detect the rare sites of damage rapidly and efficiently over an excess of millions of undamaged bases? To address this pivotal question in DNA repair, I focused on the central players from the two DNA damage excision repair pathways in my studies: nucleotide excision repair (NER) and base excision repair (BER). As examples for completely different approaches of damage search, recognition and verification, I compared the NER protein Xeroderma pigmentosum group D (XPD) with the BER proteins human thymine DNA glycosylase (hTDG) and human 8-oxoguanine glycosylase (hOgg1). In particular, the single molecule approach of atomic force microscopy (AFM) imaging and complementary biochemical and biophysical techniques were applied. I established a simple, optimized preparation approach, which yields homogeneous and pure samples of long (several hundreds to thousands of base pairs) DNA substrates suitable for the AFM studies with DNA repair proteins. Via this sample preparation, a single target site of interest can be introduced into DNA at a known position, which allows separate analysis of specific protein-DNA complexes bound to the lesion site and nonspecific complexes bound to non-damaged DNA. The first part of the thesis investigates the XPD protein involved in eukaryotic NER. In general, the NER mechanism removes helix-distorting lesions – carcinogenic UV light induced photoproducts, such as cyclobutane pyrimidine dimers (CPDs) as well as bulky DNA adducts. The 5’-3’ helicase XPD has been proposed to be one of the key players in DNA damage verification in eukaryotic NER, which is still a matter of hot debate. In the studies, I focused on XPD from the archaeal species Thermoplasma acidophilum (taXPD), which shares a relatively high sequence homology with the sequence of the human protein and may serve as a good model for its eukaryotic counterpart. Based on AFM experiments and accompanying DNA binding affinity measurements with the biosensor technology Biolayer Interferometry (BLI), a clear role of XPD in damage verification was deciphered. Specifically, the data suggested that the ATP-dependent 5’-3’ helicase activity of XPD was blocked by the presence of damage leading to stalled XPD-DNA damage verification complexes at the lesion sites. Successful damage verification led to ATP-dependent conformational changes visible by a significant transition in DNA bend angles from ~ 50° to ~ 65° at the site of the bound protein. Remarkably, this DNA bend angle shift was observed both in the presence of ATP and ATPγs (non-hydrolyzable ATP analog) indicating that ATP-binding instead of ATP hydrolysis was sufficient to induce repair competent conformational changes of XPD. Most importantly, detailed protein binding position and DNA bend angle analyses revealed for the first time that XPD preferably recognizes a bulky fluorescein lesion on the translocated strand, whereas a CPD lesion is preferentially detected on the opposite, non-translocated strand. Despite the different recognition strategies for both types of damages, they share a common verification complex conformation, which may serve as a signal for the recruitment of further NER factors. In the second part of the thesis, AFM imaging and a 2-Aminopurine fluorescence-based base-flipping assay were combined to investigate damage search and recognition by DNA glycosylases in BER. Exemplarily, I chose to study hTDG as a representative of the vast glycosylase family. hTDG excises thymine and uracil from mutagenic G:T and G:U mispairs contributing to cancer and genetic disease. The AFM data suggested that hTDG uses the intrinsic flexibility of G:T and G:U wobble pairs for initial damage sensing, while scanning DNA as a search complex (SC, slightly bent DNA). Remarkably, hTDG has been indicated to continuously switch between the search and interrogation conformation (IC, stronger bent DNA) during damage search. In the IC, target bases are interrogated by extrahelical base flipping, which is facilitated by protein-induced DNA bending and enhanced DNA flexibility at mismatches. AFM and fluorescence analyses revealed that the flipped base is stabilized via hTDG’s arginine finger. Correct target bases are perfectly stabilized within the enzyme’s catalytic pocket resulting in prolonged residence time and enhanced excision probability. To test for the generalizability of the proposed hTDG damage search model to BER glycosylases, identical studies were performed with a second glycosylase, hOgg1. The data on hOgg1, which removes structurally more stable 8-oxoguanine lesions, supported the hypothesis developed for lesion recognition by hTDG as a common strategy employed by BER glycosylases

The oncogene MYC is deregulated and overexpressed in a high variety of human cancers and is considered an important driver in tumorigenesis. The MYC protein binds to virtually all active promoters of genes which are also bound by the RNA Polymerase II (RNAPII). This results in the assumption that MYC is a transcription factor regulating gene expression. The effects of gene expression are weak and often differ depending on the tumor entities or MYC levels. These observations could argue that the oncogene MYC has additional functions independent of altering gene expression. In relation to this, the high diversity of interaction partners might be important. One of them is the RNAPII associated Factor I complex (PAF1c). In this study, direct interaction between PAF1c and MYC was confirmed in an in-vitro pulldown assay. ChIP sequencing analyses revealed that knockdown of PAF1c components resulted in reduced MYC occupancy at active promoters. Depletion or activation as well as overexpression of MYC led to reduced or enhanced global occupancy of PAF1c in the body of active genes, arguing that MYC and PAF1c bind cooperatively to chromatin. Upon PAF1c knockdown cell proliferation was reduced and additionally resulted in an attenuation of activation or repression of MYC-regulated genes. Interestingly, knockdown of PAF1c components caused an accumulation in S-phase of cells bearing oncogenic MYC levels. Remarkably, enhanced DNA damage, measured by elevated gH2AX and pKAP1 protein levels, was observed in those cells and this DNA damage occurs specifically during DNA synthesis. Strikingly, MYC is involved in double strand break repair in a PAF1c-dependent manner at oncogenic MYC levels. Collectively the data show that the transfer of PAF1c from MYC onto the RNAPII couples the transcriptional elongation with double strand break repair to maintain the genomic integrity in MYC-driven tumor cells.

The high infection rates and recent emergence of extremely drug resistant forms of Mycobacterium tuberculosis pose a significant challenge for global health. The NADH- dependent enoyl-ACP-reductase InhA of the type II mycobacterial fatty acid biosynthesis pathway is a well-validated target for inhibiting mycobacterial growth. InhA has been shown to be inhibited by a variety of compound series. Prominent classes of InhA inhibitors from literature include diaryl ethers, pyrrolidine carboxamides and arylamides which can be subjected to further development. Despite the progress in this area, very few compounds are in clinical development phase. The present work involves a detailed computational investigation of the binding modes and structure-based optimisation of pyrrolidine carboxamides as InhA inhibitors. With substituents of widely varying bulkiness, the pyrrolidine carboxamide dataset presented a challenge for prediction of binding mode as well as affinity. Using advanced docking protocols and in-house developed pose selection procedures, the binding modes of 44 compounds were predicted. The poses from docking were used in short molecular dynamics (MD) simulations to ascertain the dominant binding conformations for the bulkier members of the series. Subsequently, an activity-based classification strategy could be developed to circumvent the affinity prediction problems observed with this dataset. The prominent motions of the bound ligand and the active site residues were then ascertained using Essential Dynamics (ED). The information from ED and literature was subsequently used to design a total of 20 compounds that were subjected to extensive in-silico evaluations. Finally, the molecular determinants of rapid-reversible binding of pyrrolidine carboxamides were investigated using long MD simulations.

Gene expression and transfer of the genetic information to the next generation forms the basis of cellular life. These processes crucially rely on DNA, thus the preservation, transcription and translation of DNA is of fundamental importance for any living being. The general transcription factor TFIIH is a ten subunit protein complex, which consists of two subcomplexes: XPB, p62, p52, p44, p34, and p8 constitute the TFIIH core, CDK7, CyclinH, and MAT1 constitute the CAK. These two subcomplexes are connected via XPD. TFIIH is a crucial factor involved in both, DNA repair and transcription. The central role of TFIIH is underlined by three severe disorders linked to failure of TFIIH in these processes: xeroderma pigmentosum, Cockayne syndrome, and trichothiodystrophy. Only limited structural and functional data of TFIIH are available so far. Here, the model organism Chaetomium thermophilum was utilized with the aim to structurally and functionally characterize TFIIH. By combining the expression and purification of single TFIIH subunits with the co-expression and co-purification of dual complexes, a unique and powerful modular system of the TFIIH core subunits could be established, encompassing all proteins in high quality and fully functional. This system permits the step-wise assembly of TFIIH core, thereby making it possible to assess the influence of the intricate interaction network within TFIIH core on the overall enzymatic activities of TFIIH, which has not been possible so far. Utilizing the single subunits and dual complexes, a detailed interaction network of TFIIH core was established, revealing the crucial role of the p34 subunit as a central scaffold of TFIIH by linking the two proteins p44 and p52. Our studies also suggest that p62 constitutes the central interface of TFIIH to the environment rather than acting as a scaffold. TFIIH core complexes were assembled and investigated via electron microscopy. Preliminary data indicate that TFIIH adopts different conformational states, which are important to fulfill its functions in transcription and DNA repair. Additionally, a shortened construct of p62 was used to develop an easy-to-use, low cost strategy to overcome the crystallographic phase problem via cesium derivatization.

The number of fungal infections is rising in Germany and worldwide. These infections are mainly caused by the opportunistic fungal pathogen C. albicans, which especially harms immunocompromised people. With increasing numbers of fungal infections, more frequent and longer lasting treatments are necessary and lead to an increase of drug resistances, for example against the clinically applied therapeutic fluconazole. Drug resistance in C. albicans can be mediated by the Multidrug resistance pump 1 (Mdr1), a membrane transporter belonging to the major facilitator family. However, Mdr1-mediated fluconazole drug resistance is caused by the pump’s regulator, the transcription factor Mrr1 (Multidrug resistance regulator 1). It was shown that Mrr1 is hyperactive without stimulation or further activation in resistant strains which is due to so called gain of function mutations in the MRR1 gene. To understand the mechanism that lays behind this constitutive activity of Mrr1, the transcription factor should be structurally and functionally (in vitro) characterized which could provide a basis for successful drug development to target Mdr1-mediated drug resistance caused by Mrr1. Therefore, the entire 1108 amino acid protein was successfully expressed in Escherichia coli. However, further purification was compromised as the protein tended to form aggregates, unsuitable for crystallization trials or further characterization experiments. Expression trials in the eukaryote Pichia pastoris neither yielded full length nor truncated Mrr1 protein. In order to overcome the aggregation problem, a shortened variant, missing the N-terminal 249 amino acids named Mrr1 ‘250’, was successfully expressed in E. coli and could be purified without aggregation. Similar to the wild type Mrr1 ‘250’, selected gain of function variants were successfully cloned, expressed and purified with varying yields and with varying purity. The Mrr1 `250’ construct contains most of the described regulatory domains of Mrr1. It was used for crystallization and an initial comparative analysis between the wild type protein and the variants. The proposed dimeric form of the transcription factor, necessary for DNA binding, could be verified for both, the wild type and the mutant proteins. Secondary structure analysis by circular dichroism measurements revealed no significant differences in the overall fold of the wild type and variant proteins. In vitro, the gain of function variants seem to be less stable compared to the wild type protein, as they were more prone to degradation. Whether this observation holds true for the full length protein’s stability in vitro and in vivo remains to be determined. The crystallization experiments, performed with the Mrr1 ‘250’ constructs, led to few small needle shaped or cubic crystals, which did not diffract very well and were hardly reproducible. Therefore no structural information of the transcription factor could be gained so far. Infections with M. tuberculosis, the causative agent of tuberculosis, are the leading cause of mortality among bacterial diseases. Especially long treatment times, an increasing number of resistant strains and the prevalence of for decades persisting bacteria create the necessity for new drugs against this disease. The cholesterol import and metabolism pathways were discovered as promising new targets and interestingly they seem to play an important role for the chronic stage of the tuberculosis infection and for persisting bacteria. In this thesis, the 3-ketoacyl-CoA thiolase FadA5 from M. tuberculosis was characterized and the potential for specifically targeting this enzyme was investigated. FadA5 catalyzes the last step of the β-oxidation reaction in the side-chain degradation pathway of cholesterol. We solved the three dimensional structure of this enzyme by X-ray crystallography and obtained two different apo structures and three structures in complex with acetyl-CoA, CoA and a hydrolyzed steroid-CoA, which is the natural product of FadA5. Analysis of the FadA5 apo structures revealed a typical thiolase fold as it is common for biosynthetic and degradative enzymes of this class for one of the structures. The second apo structure showed deviations from the typical thiolase fold. All obtained structures show the enzyme as a dimer, which is consistent with the observed dimer formation in solution. Thus the dimer is likely to be the catalytically active form of the enzyme. Besides the characteristic structural fold, the catalytic triad, comprising two cysteines and one histidine, as well as the typical coenzyme A binding site of enzymes belonging to the thiolase class could be identified. The two obtained apo structures differed significantly from each other. One apo structure is in agreement with the characteristic thiolase fold and the well-known dimer interface could be identified in our structure. The same characteristics were observed in all complex structures. In contrast, the second apo structure followed the thiolase fold only partially. One subdomain, spanning 30 amino acids, was in a different orientation. This reorientation was caused by the formation of two disulfide bonds, including the active site cysteines, which rendered the enzyme inactive. The disulfide bonds together with the resulting domain swap still permitted dimer formation, yet with a significantly shifted dimer interface. The comparison of the apo structures together with the preliminary activity analysis performed by our collaborator suggest, that FadA5 can be inactivated by oxidation and reactivated by reduction. If this redox switch is of biological importance requires further evaluation, however, this would be the first reported example of a bacterial thiolase employing redox regulation. Our obtained complex structures represent different stages of the thiolase reaction cycle. In some complex structures, FadA5 was found to be acetylated at the catalytic cysteine and it was in complex with acetyl-CoA or CoA. These structures, together with the FadA5 structure in complex with a hydrolyzed steroid-CoA, revealed important insights into enzyme dynamics upon ligand binding and release. The steroid-bound structure is as yet a unique example of a thiolase enzyme interacting with a complex ligand. The characterized enzyme was used as platform for modeling studies and for comparison with human thiolases. These studies permitted initial conclusions regarding the specific targetability of FadA5 as a drug target against M. tuberculosis infection, taking the closely related human enzymes into account. Additional analyses led to the proposal of a specific lead compound based on the steroid and ligand interactions within the active site of FadA5.