Most bacteria and fungi produce secondary metabolites, including nonribosomally synthesized peptides, some of which possess potent pharmacological activities. These peptidic compounds comprise 3 to 15 ‘amino acids’ and often contain unusual building blocks such as aryl or fatty acids or non-proteinogenic amino acids, respectively. These modifications strongly add to their structural as well as functional diversity and are made possible by means of large enzymes or enzyme complexes, the so-called nonribosomal peptide synthetases (NRPS). NRPS exhibit a modular organization with each module containing specific domains catalyzing the basic reactions required for peptide assembly. The goal of this work is the elucidation of protein structures of catalytic domains involved in nonribosomal peptide synthesis with the main focus being set on thioesterase domains, condensation domains as well as epimerization domains. Thioesterase domains catalyze the release of the enzyme-bound, full-length peptide and mostly the concomitant cyclization of the product. Condensation domains catalyze peptide bond formation between two aminoacyl intermediates or an aminoacyl and a peptidyl intermediate, respectively. Epimerization domains primarily alter the stereochemical configuration of amino acid residues in the enzyme-bound intermediates. In bacterial synthetase clusters, they are also involved in intermolecular interactions leading to mutual recognition of cognate synthetases. After the elucidation of the 1.8 Å crystal structure of the thioesterase domain involved in fengycin biosynthesis, a model of a putative enzyme-substrate complex was derived in silico using a combination of docking and molecular dynamics. The model suggests an edge-on binding mode of the peptide in the enzyme’s binding pocket. Finally the binding mode was tested biochemically through analysis of the enzyme’s acceptance towards altered substrates as well as by molecular dynamics simulations for stability of the complex. The 1.8 Å crystal structure of the PCP-C bidomain protein of the tyrocidin synthetase TycC from B. brevis was solved using multiple anomalous dispersion (MAD). This structure provides insight not only into the architecture of an internal condensation domain but also into the orientation and interactions of the linker connecting the two adjacent domains. The conformation of the PCP domain corresponds to the A/H-conformation. The condensation domain comprises two chloramphenicol acetyl transferase (CAT)-like subdomains with its active site being located at the two subdomains‘ interface. Interactions between these two subdomains are mainly observed in two regions, the bridge region and the floor loop. The former bridges the active site, the latter forms the bottom of the domain’s active site and conributes to the active site’s structural integrity through an intricate network of hydrogen bonds and salt bridges. The exact mechanism of the condensation reaction remains unknown. However, structure-based calculations contradict the model in which the active site histidine acts as a catalytic base. Lead by further analysis of the active site’s surroundings a new model for the catalytic mechanism is described. The structure of an NRPS-epimerization domain from the tyrocidin synthetase TycA from B. brevis was solved at 1.65 Å resolution and gives a first impression of the structural organization of these cofactor-independent epimerases. Due to their evolutionary relatedness to condensation domains they also consist of two CAT-like subdomains. Direct comparison of condensation and epimerization domains indicates profound differences in their active sites, their bridge regions and their floor loops. In addition to the conserved, catalytically active histidine residue, H146, the active site of the epimerization domain contains a likewise conserved glutamate residue, E285. As for the condensation domain, calculations regarding the protonation states result in a protonated H146, thus precluding a role as a catalytic base as described in earlier models. While the detailed mechanism remains unknown, the epimerization domain is compared to structurally and functionally similar enzymes in order to improve our understanding of the epimerization mechanism.