Novel strategies for engineering redox metabolism in Saccharomyces cerevisiae

Abstract

In its search to decrease the environmental impact of the production of materials and food, and for other socio-economic reasons, mankind has recently taken the first steps into a paradigm shift from a petrochemical-based society to a new, sustainable and to a significant extent bio-based society. In this new scenario, biomass derived from agriculture and forestry industry and/or its residues are used for the generation of chemicals and materials that are currently derived from oil. To produce these chemicals, microorganisms and/or enzyme-catalyzed reactions can be used to convert the carbohydrates present in the biomass into a wide variety of products. Industrial biotechnology studies these conversion processes and aims to improve them. Nowadays, the single largest process in industrial biotechnology, in terms of product volume, is the production of ethanol from sugars using baker’s yeast (Saccharomyces cerevisiae). The study of this model organism in terms of physiology, metabolism and genetics has contributed to significant improvements in industrial biotechnology. However, to compete with the efficiency and economics of petrochemical industry, also for products other than ethanol, further major breakthroughs are necessary. To achieve these advances, a knowledge-based redesign, followed by implementation of genetic changes via state-of-the-art molecular biology tools is used to improve cellular activities of microbial cell factories. This discipline in applied science is called Metabolic Engineering. An important constraint in the design of metabolic engineering strategies is the balancing of reactions in metabolism that involve the transfer of electrons by the redox cofactor couples NADH/NAD+ and NADPH/NADP+. For instance, redox cofactor balancing plays a central role in the formation of glycerol as the major by-product of alcoholic fermentation by bakers’ yeast. Under aerobic conditions, oxygen is used as the final electron acceptor. However, under the anaerobic conditions that are required for cost effective production of ethanol, the yeast S. cerevisiae produces glycerol to reoxidize the “excess”-NADH derived from biosynthetic reactions. This PhD thesis aimed to explore new strategies to increase the flexibility of the (anaerobic) redox metabolism of S. cerevisiae for the production of fuels and chemicals. After a general introduction to industrial biotechnology and S. cerevisiae, Chapter 1 describes the concept of metabolic engineering and how it has been applied in the past to reduce the formation of glycerol formation in alcoholic fermentation. The ability of yeast cells to use glycerol as a redox sink can be eliminated by the double deletion of the genes encoding for glycerol-3-phosphate dehydrogenase (GPD1 and GPD2 in S. cerevisiae). In order to grow anaerobically, such mutants depend on the availability of an external electron acceptor, such as for instance externally added acetoin. However, addition of such a compound is too expensive to use on industrial scale. Of course, this is not a problem when the external electron acceptor is already present in the industrial feedstock. Acetic acid is present in small quantities in first generation feedstocks for ethanol production and in larger amounts as product from hydrolysis processes in lignocellulosic biomass (second generation feedstocks). Since, acetic acid is more oxidized than ethanol, its NADH-dependent reduction to ethanol could theoretically obviate the need for glycerol production in anaerobic cultures of S. cerevisiae. In practice, this does not occur because acetic acid is almost fully dissociated at the near-neutral pH inside yeast cells. Other microorganisms can use acetic acid as electron acceptor. These microorganisms use a linear pathway that first activates acetic acid to acetyl-CoA in a reaction that costs two ATP equivalents. In two subsequent reactions coupled to NADH oxidation, acetyl-CoA is first reduced to acetaldehyde, which is then reduced to ethanol. Only the reaction that reduces acetyl-CoA towards acetaldehyde, catalyzed by an (acetylating) acetaldehyde dehydrogenase, is not naturally present in S. cerevisiae. In Chapter 2, the replacement of glycerol formation as redox sink by acetic acid as external electron acceptor was studied in anaerobic cultures of a gpd1? gpd2? S. cerevisiae expressing the Escherichia coli mhpF gene encoding an (acetylating) acetaldehyde dehydrogenase. Growth of the constructed strain (gpd1? gpd2? mhpF) at a maximum specific growth rate of 0.14 h-1 was dependent on the presence of acetic acid. Under these conditions, the strain did not produce glycerol and showed a 13% higher ethanol yield on glucose than the isogenic reference strain (GPD1 GPD2). These results are a proof of concept for this metabolic engineering strategy, where glycerol formation was replaced by the removal of, otherwise inhibitory, acetic acid from lignocellulosic hydrolysates, leading to a significant increase in ethanol yield on glucose. Besides its role in redox-cofactor balancing, glycerol is also the main compatible solute of S. cerevisiae, accumulating inside yeast cells when they face high extracellular osmotic pressure. This type of stress is especially important in first generation alcoholic fermentation, where a high initial sugar concentration is present at the start of the production process. In Chapter 3, the tolerance to high osmotic pressure of a strain lacking functional glycerol-3-phosphate dehydrogenases (Gpd-) and able to use acetate as electron acceptor was studied. Based on these findings, evolutionary engineering for anaerobic growth at high sugar concentrations (1 M glucose) was used to obtain an osmotolerant Gpd- S. cerevisiae strain expressing an (acetylating) acetaldehyde dehydrogenase. After the desired phenotype was obtained, single colonies were isolated and characterized under relevant conditions. An isolated evolved Gpd- strain grew anaerobically at 1 M glucose at a maximum specific growth rate of 0.12 h-1 in the presence of acetic acid (2 g l-1). Surprisingly, formation of glycerol was observed again towards the end of the fermentation, albeit at much lower concentrations than in the Gpd+ reference strain under identical conditions. Moreover, the evolved strain exhibited an apparent higher ethanol yield on glucose than the reference strain, reaching a value of 1.84 mol mol-1 (92% of the theoretical ethanol yield on glucose), when cultured in the presence of 3 g l-1 of acetic acid and 1 M glucose. Genetic analysis of the evolved strain revealed that this evolved phenotype was the consequence of one dominant chromosomal mutation, and one mutation in the plasmid-borne mhpF gene for anaerobic growth. In industrial biotechnology, the use of evolutionary engineering coupled to reverse metabolic engineering is a powerful tool in the development of strains with desired phenotypes and in the transfer of these characteristics to different strains. However, during laboratory or industrial evolution, also undesired phenotypes are observed, which are equally important to comprehend: only by understanding their molecular basis, such phenotypes can be removed from production strains. Cell flocculation and/or aggregation were frequently observed during laboratory evolution in sequential batch cultures under a wide variety of selective culture conditions. Although flocculation can be a desired phenotype in beer brewing, where it may facilitate separation of the yeast, it is not desirable during evolutionary engineering, where it diverts the selective pressure away from the target of interest and furthermore complicates the generation of single cell lines. This phenomenon, which was also observed in some of the evolutionary engineering lines with Gpd- S. cerevisiae strains that were performed in the context of this thesis, was further studied in Chapter 4 with a strain obtained from laboratory evolution for faster consumption of glucose-galactose mixtures. Reverse metabolic engineering of ‘multicellular’ strains of S. cerevisiae obtained from this study revealed that genome duplication and deregulation of the cell cycle were key elements in the development of a multicellular/agglomeration phenotype. Whole genome sequencing of two single colony isolates from independent laboratory evolutions showed that multicellular phenotype resulted from different point mutations in ACE2 gene, a key transcriptional regulator in the separation between bud and mother cell in yeast. Moreover, both final evolved strains became diploids, whereas the original parental strain was haploid. The multicellular phenotype was reverted by introduction of a functional copy of the original ACE2 allele in the evolved strains. Introduction of the mutant allele and doubling the genome size by mating in the parent strain led to the same multicellular, fast-sedimenting phenotype that was observed in the evolved strains. These results do not only shed light on the mutations that underlie the evolution of multicellular yeast strains, but can also be applied to induce or eliminate cell aggregation in industrial strains. Major breakthroughs in industrial biotechnology benefit from rapidly developing techniques in synthetic biology. These advantages allow scientists to use bolder and more creative metabolic engineering strategies. An unexplored metabolic engineering strategy in redox metabolism is the use of carbon dioxide as electron acceptor via enzymes from autotrophic microorganisms. Carbon dioxide is a by-product of yeast fermentation and therefore abundantly present in industrial fermentation processes. The Calvin cycle, present in different autotrophic organisms, uses ATP and NADPH for carbon dioxide fixation and provides metabolic building blocks required in biosynthesis. Phosphoribulokinase and ribulose-1,5-bisphosphate carboxylase/oxidase (Rubisco) are the two key enzymes of the Calvin cycle and have been involved in the fixation of the majority of the organic carbon available in nature. Using state-of-the-art synthetic biology techniques, we functionally expressed both enzymes in a S. cerevisiae strain and studied its physiology in anaerobic carbon-limited chemostat cultures at a dilution rate of 0.05 h-1 (Chapter 5). Functional expression of a single-subunit Rubisco from the chemolithoautotrophic bacterium Thiobacillus denitrificans required the co-expression of chaperones proteins GroEL and GroES from E. coli, and resulted in the first functional expression of Rubisco in a heterotrophic eukaryotic microorganism. Anaerobic chemostat cultures sparged with pure nitrogen gas showed a 68% lower glycerol yield on sugars, and 11% and 12% higher ethanol and biomass yields on sugars, respectively, than an isogenic reference strain. Increasing the concentration of dissolved carbon dioxide by purging the fermenters with a gas mixture of 10% v/v CO2 and 90% v/v nitrogen resulted in a further decrease of the glycerol yield on sugars to less than 10% of that observed in the reference strain. To study the performance of the engineered strain using a fermentation setup more relevant to industrial application, anaerobic batch cultures were run on 20 g l-1 galactose in laboratory fermenters sparged with a 10%-90 % mixture of CO2 and nitrogen. Under these conditions, the constructed strain showed no differences in growth kinetics when compared to its isogenic reference, while its glycerol yield on galactose decreased by 60% and its ethanol yield on galactose increased by 8% relative to the reference strain. This strategy not only demonstrates the potential of using carbon dioxide as electron acceptor in the metabolic engineering of yeasts and other microorganisms, but also illustrates how co-expression of chaperone proteins can aid the functional expression of bacterial proteins in yeast cytosol. To conclude, this thesis presents two novel metabolic engineering strategies (Chapter 2 & 5) that increased the flexibility of redox metabolism in anaerobic fermentation by implementing novel redox sinks in baker’s yeast. Both strategies have high potential to substantially contribute to optimizing product formation by S. cerevisiae in anaerobic industrial biotechnological production processes. In such processes even a small increase in product yield on substrate, given the large volumes of production, can result in large economic benefits without the introduction of new process steps. Even though novel strategies can show high potential for industrial application, an integrative view on process optimization should also consider potential negative effects on strain robustness. Evolutionary engineering (Chapter 3) and reverse metabolic engineering (Chapter 4) of desired and undesired phenotypes, provide additional powerful tools in the implementation of metabolic engineering strategies. The chapters of this thesis represent small but significant steps in the vigorous research needed to enable wide-spread, economically viable and sustainable production of transport fuels and chemicals in a bio-based society.