"Our group has recently established that heme-containing proteins, in particular myoglobin and cytochrome P450s, constitute promising biocatalysts for the formation of carbon-nitrogen and carbon-carbon bonds via nitrene and carbene transfer reactions, a class of synthetically valuable transformations not occurring in nature. Building upon this work, a first goal of this research was to improve the scope and efficiency of these nitrene transfer biocatalysts for C-H amination reactions. To this end, we identified a novel, unusual P450-type enzyme named XplA, which catalyzes the intramolecular C-H amination of arylsulfonyl azide substrates with a significant enhancement of activity and chemoselectivity in comparison to other P450s, which efficiency in this reaction is limited by the generation of reduced byproducts. Primary kinetic isotope effect studies revealed that the mechanism proceeds with C-H bond activation as the rate limiting step and provided insights into the competition between productive vs. non-productive nitrene transfer pathways. Furthermore, we discovered that non-heme Rieske dioxygenases are viable C- H amination biocatalysts and explored their potential value in nitrene transfer reactions by means of protein engineering, large scale reactions in a bioreactor, and analysis of their reaction and substrate scope. In a second part, we expanded and modulated the reactivity of myoglobin biocatalysts in carbene transfer reactions. A series of artificial myoglobin-based metalloenzymes containing manganese, iron, cobalt, ruthenium, rhodium and iridium were investigated for cyclopropanation and Y-H (Y = N, S) carbene insertion reactions. Engineered variants containing a ruthenium cofactor were found to be excellent S-Hinsertion catalysts, while variants harboring an iridium cofactor were capable of C-H insertion reactions not supported by the parent protein. Next, we demonstrated how cofactor variation in combination with mutations of the proximal ligand anchoring the metalloporphyrin in the active site pocket drastically influences catalyst chemoselectivity. Specifically, we developed a serine-ligated cobalt-porphyrin variant that favors the more challenging olefin cyclopropanation reaction in the presence of competing functional groups. In contrast, the native protein with a histidine-ligated heme cofactor selectively undergoes the complementary Y-H (Y = N, Si) insertion reaction in the presence of unsaturated bonds. In a further study, we successfully extended the substrate scope of engineered myoglobin 'carbene transferases' for realizing N-H insertion reactions between benzyl- and alkylamines and different diazo precursors, which was previously not reported for other hemoproteins, thereby enabling access to valuable functionalized benzyland alkylamines. Finally, we devised a biocatalytic strategy for the asymmetric synthesis of chiral amines via myoglobin-catalyzed N-H insertion. Achieving high enantioselectivity in carbene-mediated N-H insertion reactions has been notoriously challenging. To this end, reactions involving a combination of evolved myoglobin variants with engineered diazo compounds led to the first report and highest enantioselectivity achieved by a biocatalyst in this reaction to date. In addition, stereodivergent biocatalysts were developed to obtain both mirror-image forms of chiral anilines. Altogether, these studies highlight how protein engineering provides a powerful strategy for expanding the biocatalytic toolbox toward synthetically useful yet challenging abiological biotransformations under environmentally friendly and sustainable conditions"--Pages xi-xii.