Metabolic engineering – the improvement and addition, by genetic modification, of industrially relevant properties of microorganisms with respect to catalysis, transport and regulatory functions – is a well-established method for development of more cost-effective and ‘green’ industrial processes. Rapid depletion of oil reserves and a growing demand for sustainable, environmentally friendly processes provide incentives for efficient exploitation of new, renewable resources for the production of transport fuels and bulk chemicals. The narrow profit margins that are typical for such commodity products, impose a need to optimize processes in terms of kinetics of product formation and, especially, yield of product on feedstock. Metabolic engineering of microorganisms, with its continuously expanding toolbox, allows researchers to address the challenges involved in the development of biotechnological processes that can compete with petrochemical production. Due to its robustness in industrial fermentation processes and fast developments in yeast synthetic biology, Saccharomyces cerevisiae (a.k.a. baker’s yeast) has become one of the most popular metabolic engineering platforms in modern biotechnology. As a result, this microorganism, after having been used for ages in the production of alcoholic beverages and bread, is now recognized as multi-purpose microbial ‘workhorse’ with numerous industrial applications. Production of many natural and heterologous compounds with genetically modified strains of S. cerevisiae is under investigation or has already been implemented in industry. Many of those biochemicals, for example n-butanol, isoprenoids, lipids, flavonoids and 3-hydroxypropionic acid, require acetyl coenzyme A (acetyl-CoA) as a key precursor. The metabolism of this compound in S. cerevisiae cells is compartmentalised. The mitochondrial route, responsible for a substantial flux towards acetyl-CoA during respiratory growth on sugars, involves conversion of mitochondrial pyruvate into acetyl-CoA in the reaction catalysed by the pyruvate dehydrogenase complex. Since, in S. cerevisiae, acetyl-CoA cannot be exported from the mitochondria, another pathway is required to cover biosynthetic requirements for acetyl-CoA in the cytosolic compartment. This so-called ‘pyruvate dehydrogenase bypass’ requires the concerted activity of pyruvate decarboxylase, acetaldehyde dehydrogenase and acetyl-CoA synthetase. The latter two reactions are also required for growth on the C2-compounds – acetate (only acetyl-CoA synthetase) and ethanol. When ethanol is used as a carbon source, it is first converted to acetaldehyde in the reaction catalysed by alcohol dehydrogenases and further to acetyl-CoA, which is used to cover all biosynthetic and energetic requirements of ethanol-grown cells. Heterologous, acetyl-CoA-dependent product pathways expressed in the cytosol of S. cerevisiae exclusively depend on the cytosolic route for provision of this precursor. As a result, several previous metabolic engineering studies have been devoted to improving the synthesis of the cytosolic acetyl-CoA. Although successful in increasing the availability of this compound, by improving the capacity of the native route or by introduction of the heterologous pathways of cytosolic acetyl-CoA synthesis, those studies primarily focused on the kinetic parameters, while the energetic aspects of cytosolic acetyl-CoA synthesis received little attention. The maximum yield of biomass or of any other industrially relevant product on a substrate depends on the energy (ATP) cost of the biochemical pathways used for precursor and product formation. The synthesis of cytosolic acetyl-CoA in S. cerevisiae involves hydrolysis of ATP to AMP and pyrophosphate (PP i ) in the reaction catalysed by acetyl-CoA synthetase (ACS). The subsequent hydrolysis of PP i to inorganic phosphate (P i ) makes the energetic cost of cytosolic acetyl-CoA synthesis equivalent to hydrolysis of 2 ATP to 2 ADP and 2 P i . This ATP expenditure has a profound impact on the maximum yields of acetyl-CoA-dependent products that can be achieved in engineered yeast strains and, therefore, negatively influences the economy of the production process. Especially in the case of bulk chemicals and bio-fuels, this single ATP-consuming reaction can become cost-prohibitive. This thesis explores metabolic engineering strategies to address this key challenge in yeast biotechnology. After general introduction to metabolic engineering and S. cerevisiae, Chapter 1 describes the strategies that have hitherto been explored to increase the availability of cytosolic acetyl-CoA. Moreover, this Chapter discusses other pathways of cytosolic acetyl-CoA synthesis that occur in nature, some of which do not require an input of ATP and could therefore, upon expression in S. cerevisiae, lead to increase of maximum yields of acetyl-CoA-dependent products on substrate in engineered yeast strains. In S. cerevisiae, cytosolic acetyl-CoA synthesis and growth strictly depend on functional expression of either the Acs1 or Acs2 isoenzyme of acetyl-CoA synthetase (ACS). Before the research described in this thesis, viable S. cerevisiae strains in which both ACS1 and ACS2 had been deleted, had not been described in the literature. In addition to its anabolic and, under certain conditions, catabolic roles, cytosolic acetyl-CoA also plays key role in cellular regulation via acetylation of proteins, including histones. Chapter 2 explores the feasibility of replacing the native S. cerevisiae pathway for cytosolic acetyl-CoA synthesis by two alternative, ATP-independent pathways, and investigates their impact on growth and energetics of the engineered yeast strains. To this end, the native route of cytosolic acetyl-CoA synthesis was replaced by either an acetylating acetaldehyde dehydrogenase (A-ALD) or a pyruvate-formate lyase (PFL). Acetylating acetaldehyde dehydrogenase catalyses direct, ATP-independent oxidation of acetaldehyde to acetyl-CoA, while pyruvate-formate lyase converts cytosolic pyruvate into equimolar amounts of acetyl-CoA and formate. After evaluating the expression of different genes encoding acetylating acetaldehyde dehydrogenase and pyruvate-formate lyase, acs1? acs2? S. cerevisiae strains were constructed in which A-ALD or PFL functionally replaced ACS. In case of the A-ALD-dependent strains, also all acetaldehyde dehydrogenases were deleted, which resulted in complete replacement of the two-step conversion of acetaldehyde to acetyl-CoA, by a one-step reaction catalysed by A-ALD. The A-ALD-dependent strains showed aerobic growth rates of up to 79% of the reference strain, while anaerobic growth rates of PFL-dependent S. cerevisiae reached up to 73% of the growth rate of the reference strain. Physiological characterisation of the fastest growing A-ALD- and PFL-dependent strains was performed in glucose-limited chemostat cultures. Unexpectedly, the measured biomass yields on glucose of A-ALD- and PFL-dependent strains were 14% and 15% lower than those of the reference strain, respectively. Weak-acid uncoupling by formate, the formation of which was stoichiometrically coupled to growth of PFL-dependent strain, offers a plausible explanation for the reduced biomass yield of this strain. In A-ALD-dependent strain, the reduced biomass yield of the A-ALD-dependent strain was attributed to toxic effects of the acetaldehyde, which was present at higher levels in cultures of the engineered strain than in cultures of the reference strain. Changes in the synthesis of cytosolic acetyl-CoA might affect histone acetylation as well as central metabolism via acetylation of non-histone proteins and direct participation of this compound in key reactions. However, transcriptome analysis of A-ALD- and PFL-dependent strains revealed only small sets of genes with altered expression levels relative to the reference strain. Combined with their high growth rates, these observations suggested that these strains did not suffer from major limitations in acetyl-CoA provision. This conclusion was further supported by the minor differences in intracellular metabolite levels of an A-ALD-dependent strain relative to the control strain. Intracellular acetyl-CoA concentrations, which reflect the combination of mitochondrial and nucleocytosolic pools, were also not significantly different between A-ALD-dependent strain and the reference strain, which may suggest that intracellular concentrations of acetyl-CoA are subject to strong homeostatic regulations. Higher intracellular lysine concentrations in A-ALD-dependent strain might even be indicative for increased availability of cytosolic acetyl-CoA. The research presented in Chapter 2 demonstrated, for the first time, that the native pathway for cytosolic acetyl-CoA synthesis in S. cerevisiae can be entirely replaced by heterologous, ATP-independent pathways. During respiratory growth of S. cerevisiae on glucose, cytosolic acetyl-CoA is required to cover biosynthetic requirements of the cell, including the biosynthesis of lysine and of lipids. However, these biosynthetic fluxes are relatively small compared to the overall rate of ATP turnover and, therefore, have a relatively small impact on growth energetics (calculated impact on biomass yield in aerobic, glucose-limited cultures < 0.5%). A fundamentally different situation arises when ethanol is used as a carbon source and acetyl-CoA synthetase is the starting point for all biosynthetic and dissimilatory pathways in growing yeast cells. Under these conditions, saving of the 2 ATP required for synthesis of each molecule of cytosolic acetyl-CoA, should theoretically enable an increase of the biomass yield from 0.57 g/g ethanol for the native route up to 0.80 g/g ethanol for an ATP-independent pathway. Such a dramatic increase in the energetic efficiency of ethanol utilization could be highly relevant for industry, as it should result in higher yield on ethanol of any acetyl-CoA-dependent product. Therefore, Chapter 3 focuses on yeast strains in which the native pathway for acetyl-CoA synthesis was replaced by A-ALD. The A-ALD-dependent strains, however, did not show immediate growth on media with ethanol as the sole carbon source. Prolonged incubation, followed by long-term laboratory evolution experiments, yielded A-ALD-dependent yeast strains that were able to grow on ethanol with specific growth rates up to 0.11 h ?1 . Reverse engineering studies showed that mutations in ACS1, the gene that encodes one of the S. cerevisiae cytosolic acetyl-CoA synthetases, were essential for growth on ethanol of the A-ALD-dependent strains. Acquired mutations in A-ALD genes improved V MAX /K M for acetaldehyde of the encoded enzymes, but were not essential for growth on ethanol. Although the reverse engineered strains grew on ethanol, their growth rates were lower than the growth rates of the evolved strains. Further analysis of the growth of these strains suggested a limited availability of mitochondrial acetyl-CoA during growth of the A-ALD-dependent strains on ethanol. Out of five evolved strains tested in ethanol-limited chemostat cultures, only one evolved strain showed a 5% increase in the biomass yield on ethanol compared to the reference strain, which was far below the 40% theoretical prediction. Increased production of acetaldehyde and other byproducts was identified as possible cause for these suboptimal biomass yields. This study in Chapter 3 proves that the native yeast pathway for conversion of ethanol to acetyl-CoA can be replaced by an engineered pathway that has the potential to strongly improve biomass and product yields. Based on metabolic and evolutionary engineering, whole-genome resequencing, reverse engineering and physiological analysis of evolved strains, we identify intracellular acetaldehyde levels and provision of intramitochondrial acetyl-CoA as key targets for further optimization of ethanol conversion by eukaryotic cell factories. Pyruvate dehydrogenase (PDH) is a huge, multisubunit enzyme complex whose size is comparable to that of a ribosome. The PDH complex catalyzes the oxidative decarboxylation of pyruvate to acetyl-CoA which, in many organisms, is the key reaction at the interface of glycolysis and tricarboxylic acid cycle (TCA). In eukaryotes, the PDH complex occurs exclusively in the mitochondrial matrix and, therefore, cannot contribute to provision of cytosolic acetyl-CoA in S. cerevisiae. In Chapter 4, the challenge of the expression of a functional, heterologous PDH complex in the yeast cytosol is taken up. In order to determine if a heterologous PDH complex can replace the native route of cytosolic acetyl-CoA formation in S. cerevisiae, the PDH complex from the bacterium Enteroccocus faecalis was selected to be expressed in yeast. Three factors contributed to selecting the PDH complex from E. faecalis for this study: (i) bacterial PDH subunits presumably lack mitochondrial localization sequences; (ii) E. faecalis PDH is relatively insensitive to high NADH/NAD + ratios in comparison to other bacterial PDH’s; and (iii) in vitro experiments indicate that purified subunits of the E. faecalis PDH can self-assemble into a functional complex. All known PDH complexes require four cofactors for their activity: FAD + , NAD + , thiamine pyrophosphate (TPP) and lipoic acid. In S. cerevisiae, lipoic acid is synthesized and covalently attached to the PDH complex in the mitochondria. Therefore, cytosolic expression of a bacterial PDH complex was likely to require co-expression of proteins that catalyze lipoylation of PDH, as well as addition of lipoate to the media. The expression of genes encoding the four subunits of the E. faecalis PDH complex, together with two lipoate ligases from E. faecalis in the cytosol of S. cerevisiae, combined with deletion of ACS1 and ACS2, enabled growth of an acs1? acs2? strain of S. cerevisiae on glucose. The strict dependency of growth on the addition of lipoic acid confirmed the in vivo activity of the heterologous PDH complex. The aerobic growth rate on glucose of the obtained strain (0.36 h ?1 ) was comparable to the growth rate of the reference strain (0.42 h ?1 ) and independent of the concentration of lipoic acid in the range of 20 ng mL ?1 to 1000 ng mL ?1 . The Acs ? yeast strain expressing the bacterial PDH complex also showed a near-wild-type growth rate under anaerobic conditions (0.30 h ?1 and 0.33 h ?1 for engineered and wild-type strain, respectively). Functioning of the E. faecalis PDH complex in anaerobic yeast cultures indicates that it can also be applied in the design and construction of anaerobic product pathways. Enzyme activity assays indicated that, in an engineered strain, PDH from E. faecalis yielded higher specific activity (53 nmol min ?1 mg protein ?1 ) than the activity of the mitochondrial PDH of S. cerevisiae (12 nmol min ?1 mg protein ?1 ). The cytosolic localization of the heterologously expressed E. faecalis PDH complex in the yeast cytosol was confirmed by subcellular fractionation, combined with enzyme activity assay and mass-spectrometry-based proteomics. While the specific activities of the PDH in the mitochondrial fractions of engineered strain and the wild-type strain were not significantly different, the specific activity of PDH measured in the cytosolic fraction of Acs ? strain expressing PDH of E. faecalis was 32-fold higher than that of the wild-type strain. The cytosolic localization of the four subunits of the E. faecalis PDH complex, as well as of the two lipoate ligases of E. faecalis in the cytosol of S. cerevisiae was further confirmed by mass spectrometry-based proteome analysis of the cytosolic fraction. In E. faecalis, PDH occurs as a protein complex consisting of 210 subunits. Gel filtration of a cytosolic fraction, combined with enzyme activity measurements and proteomics analysis, demonstrated that E. faecalis PDH was present in the cytosol of S. cerevisiae as a complex with a size, a specific activity, and a relative abundance of the E1, E2, and E3 subunits comparable to those reported for native E. faecalis PDH. Chemostat-based physiological characterization in glucose-limited chemostat cultures showed comparable biomass yields and rates of sugar dissimilation for the Acs ? strain expressing PDH and a reference strain. These results, together with the absence of strong differences in the transcriptome of engineered and native strains, indicated that replacement of the cytosolic acetyl-CoA synthesis with PDH complex from E. faecalis did not lead to significant disturbances of the physiology of S. cerevisiae. The research described in this thesis demonstrates that complete replacement of the native route of cytosolic acetyl-CoA formation in S. cerevisiae with heterologous ATP-independent pathways can result in viable yeast strains. Moreover, it delivers proof of concept that implementation of the ATP-independent pathways may result in increased biomass yield on substrate. These strategies appear to be particularly suited for improving the product yield on sugar of any compound, produced in S. cerevisiae, that uses cytosolic acetyl-CoA as a precursor. Therefore, the results presented in this thesis provide metabolic engineers with new strategies to optimize the performance of Saccharomyces cerevisiae as a ‘cell factory’ for sustainable production of fuels and chemicals.