Current advances in industrial biotechnology are expanding the breadth of what is possible through the use of microbial cell factories. With global resources being exhausted at rapid rates and a heavy reliance on nonrenewable feedstocks, there is pressure to find alternatives using biorenewables. Application of these products is widespread, from reducing petroleum-based demand in the chemical industry, to synthesizing next-generation drug therapeutics at low cost. At the center of this biobased innovation is the need for a workhorse microorganism, capable of efficiently manufacturing these target molecules. In the context of biorenewable compounds, there are many cases where bacteria are not an option due to host toxicity or lack of processing ability, for example, post-translational modification. Baker’s yeast, Saccharomyces cerevisiae, is an ideal candidate as it is well characterized and frequently used industrially. Here we report its use in producing the polyketide triacetic acid lactone (TAL), an important polyketide which can be readily converted to replace fossil-fuel derived counterparts. Due to the complexity inherent in any host cell, metabolic pathway engineering was performed on central carbon metabolism to increase availability of acetyl-CoA and malonyl-CoA metabolites. The computational tool OptKnock in conjunction with other knockout interventions were used to create a strain (BY4741Δprb1Δpyc2Δnte1Δyia6) producing 2.2 g/L TAL, 26% of the theoretical yield on glucose, in fed-batch fermentation. In addition, enzyme engineering of the Gerbera hybrida 2-pyrone synthase (2-PS) catalyzing TAL formation was conducted and then evaluated in our strains. This effort screened 41 2-PS variants and elucidated key residues in the active site and the protein surface, improving catalytic efficiency and/or stability, leading to 2.5-fold increases in TAL levels in S. cerevisiae. The surface-modified variant 2-PS (C35S, C372S) exhibited the highest kcat/Km, over 40 times that of the native synthase. This variant was introduced into an engineered strain (BJ5464Δpyc2Δnte1) and cultivated in glucose or ethanol fed-batch to yield 2.7 g/L and 4.2 g/L TAL, respectively. Fermentation parameters were further improved and feed components modified for high titer (7.6 g/L) and yield (44% of theoretical) of this polyketide. Enhancement of precursor availability was also achieved by using a heterologous pathway from Escherichia coli. A modified E. coli bacterial pyruvate dehydrogenase (PDHm) was introduced in a cofactor-limited (Δzwf1) strain to couple NADPH cofactor balance with the synthesis of the TAL precursor acetyl-CoA. To prevent loss of acetyl-CoA and pyruvate by transport to the mitochondria, recently elucidated pathways were disrupted (Δpor2Δmpc2Δpda1Δyat2). The coupling of all three approaches (Δzwf1Δpor2Δmpc2Δpda1Δyat2+PDHm) elevated titers to 1.6 g/L, the highest by batch cultivation (and 35% of theoretical yield). These strategies are directly applicable to the synthesis of other acetyl-CoA derived compounds, and these studies provide a framework to develop design principles for robust, industrially relevant yeast for the synthesis of polyketides.

It is widely accepted that aggressive development of the bio-economy is needed to mitigate the effects of imminent resource shortages and climate change. The employment of microorganisms, such as yeasts, as biorefineries is one feasible route to meet this demand. Kluyveromyces marxianus is a non-conventional yeast commonly found in fermented dairy products that is exceptionally fast-growing, thermo- and pH-tolerant, metabolizes a wide range of low-cost carbon sources, and has great potential for biosynthesis of acetyl-CoA-based products. In this work we demonstrate the high native capacity of K. marxianus to produce triacetic acid lactone (TAL), a member of the diverse polyketide class of bioactive molecules, from waste substrates, achieving a yield comparable to other industrially relevant yeasts that required extensive engineering. Despite these advantageous characteristics, K. marxianus has not been widely engineered; this is due in part to the strong non-homologous end joining pathway which greatly hinders targeted genomic modification. The recent emergence of CRISPR-Cas9 genome editing technology has alleviated this roadblock, allowing metabolic engineering of K. marxianus to be more accessible. By constructing a highly efficient CRISPR-Cas9 system and disrupting the non-homologous end joining pathway, we have created a robust and efficient platform that routinely enables a high rate of targeted editing in K. marxianus, including genomic integrations with as little as 40 base pair homologies. Using this tool, we have conducted rationally and computationally guided metabolic engineering to significantly improve the biosynthesis of TAL and an additional polyketide, 6-methylsalisylic acid (6-MSA). Combining the best performing modifications led to 1.66- and 2.5-fold increases in TAL and 6-MSA biosynthesis, respectively, over the unengineered strains in defined media containing xylose or lactose. Altogether, our work demonstrates effective strategies for targeted metabolic engineering, as well as the great potential for high-titer biosynthesis of acetyl-CoA-based heterologous products in this under-studied and promising industrial yeast species. As an alternative to this bottom-up, genotype-to-phenotype approach to metabolic engineering, we have implemented high-throughput CRISPR-Cas9 genome-scale guide RNA libraries which instead enable phenotype-to-genotype gene discovery. Optimizing the library composition and expression system resulted in a final library with exceptionally high cutting activity. Application of this library has enabled the unbiased and direct identification of essential genes in our K. marxianus strains under prescribed growth environments. Pairing these libraries with a growth-coupled metabolite sensor also enabled direct identification of metabolic engineering targets that increase accumulation of critical polyketide precursors. This approach bypasses the current limitations of rational pathway analysis and use of computational models in less-characterized yeasts, by providing precise insights into the best interventions for optimal biosynthesis of heterologous products.

Advances in biotechnology have enabled potential solutions to the oil-dependency of the chemical industry through renewable feedstocks for the production of current compounds, the discovery and production of new compounds with novel properties, and in some cases, the synthesis of chemicals at a lower cost than with traditional processes. The yeast Saccharomyces cerevisiae has been widely used in industry to produce ethanol, proteins, nutritional supplements and pharmaceuticals. A recent major focus has been the use of microorganisms to produce biofuels, commodity chemicals and specialty chemicals. Long-chain fatty acids (LCFA), medium-chain fatty acids (MCFA), and polyketides are good targets for the biochemical industry as they supply platform chemicals that can be later derived to a library of final products by chemical catalysis. The focus of our work has been the metabolic engineering of S. cerevisiae for the synthesis of free fatty acids of defined length. We constructed a yeast strain devoid of its native fatty acid metabolism and expressing the heterologous type II fatty acid synthase (FAS) system from Escherichia coli and a C14-specific thioesterase. This FAS system increased total fatty acids and shifted the fatty acid profile, increasing the percentage of C14 fatty acids from less than 1% to 33%. Yeast β-oxidation was engineered for MCFA production by a series of pathway interventions. Disruption of two genes, the essential gene ECI1 for mono-unsaturated fatty acid (mUFA) degradation and FAA2 for the peroxisomal MCFA activator, led to an increase in C12:1 production from 50 mg/L to 1 g/L in oleic acid medium. CRISPR/Cas9 was used to swap the native FAS1 and FAS2 promoters to the strong constitutive TEF1 promoter to increase expression of the yeast FAS. This resulted in a 3.8-fold increase in the production of total MCFA from glucose media relative to the faa2∆ eci1∆ strain. LCFA production and secretion was increased by deleting three genes for enzymes involved in fatty acid activation and three genes for enzymes in the fatty acid degradation pathway. Most importantly, a novel approach overexpressing genes for a lipid particle-forming enzyme (DGA1) and a triacylglycerol lipase (TGL3) further increased the production of secreted LCFA in shake flack culture to 2.2 g/L, the highest by S. cerevisiae to date. Finally, we performed preliminary work to increase acetyl-CoA pools for the synthesis of the closely related polyketides by downregulating competing pathways, such as fatty acid synthesis and the glyoxylate shunt. We evaluated downregulation of FAS1 and FAS2 using the CRISPR interference (CRISPRi) system, and also a more traditional promoter engineering strategy.

In response to the exhaustion of carbon feedstock from fossilized sources, dependence on foreign energy supplies, and the threat of climate change, there is an increasing incentive for the production of biofuels and bio-based chemicals. Second-generation biorefineries provide a sustainable approach by relying on lignocellulosic biomass obtained from renewable resources. To increase economic viability and lower the cost of biomass processing there is a promising approach called Consolidated Bioprocessing (CBP). CBP may be accomplished by hydrolysis of lignocellulose’s main components—hemicellulose and cellulose—by displaying a xylanosome and a cellulosome on the surface of the microorganism responsible for synthesis of the desired chemical (e.g. ethanol, medium chain fatty acids, etc.). In the first part of this study, we describe strategies to increase the efficiency of CBP lignocellulosic biomass hydrolysis. First, we developed an expression system and a set of strains that showed enhancements in protein secretion of 5.8-fold to 11.5-fold. Next, we designed and characterized an extracellular glucose sensor that enables selective display of cellulose-degrading enzymes in the presence of the cellulosic fraction of lignocellulose. This is possible by leveraging cellulose-dependent signal amplification. The engineered version of the sensor showed up to 81% higher levels of expression, and a 55%-91% amplification of the signal was observed in the presence of cellobiose. Once saccharification is complete, CBP microorganisms will ideally synthesize chemicals that can functionally substitute those currently generated by the petrochemical industry. Fatty acids with chain lengths between 8 and 12 carbons have a wide variety of industrial applications, including as biofuels and precursors to commodity and fine chemicals. Our laboratory has successfully engineered Saccharomyces cerevisiae to produce hexanoic acid, octanoic acid (C8) and decanoic acid (C10). However, higher levels of production are hindered by the toxic effects of the fatty acids on the cells at increased concentrations. To overcome this limitation, we successfully explored two different strategies in the last part of this study: i) changing the membrane composition to alleviate the effects of C8, and ii) identifying and overexpressing efflux pumps that export C10. This allowed growth improvements of up to 10-fold (C8) and 11-fold (C10).

The yeast Saccharomyces cerevisiae has been an important host for the biosynthesis of valuable products in a diverse range of industrial sectors, ranging from energy to food and beverage to biopharmaceutical. With the advent of synthetic biology, several important molecular tools have been developed for engineering the yeast to be a more efficient microbial cell factory and to elucidate the vast interacting gene networks for better fundamental understanding. In this research, we have developed, validated, and applied two novel synthetic biology tools for S. cerevisiae; these can be used for the improvement of product synthesis and to examine yeast cellular processes. The first tool utilizes CRISPR-Cas9, a potent gene editing technique, to generate a medium-sized random combinatorial gene knockout yeast library. This was used to quickly and efficiently identify positive combinations of pathway and regulatory genes related to the accumulation and utilization of the acetyl-CoA and malonyl-CoA pool within the yeast cytosol. The generation of combinatorial gene knockout yeast libraries via CRISPR-Cas9 and the yeast culture conditions were first optimized to remove any bias and allow reproducibility, respectively. Using our new method, we identified yeast strains exhibiting up to 3-fold differences in specific titer of the polyketide triacetic acid lactone (TAL). This rapid low-cost approach allows generation and screening of yeasts cells containing random combinations of gene knockout with minimal bias, eliminating the need for the tedious process of sequentially knocking out potential gene targets. The second synthetic biology tool emulates natural enzyme colocalization as a strategy to increase the biosynthesis of heterologous products in yeast. We have taken advantage of the highly specific Cas6-RNA interaction and the predictability of RNA hybridizations to demonstrate Cas6-mediated RNA-guided protein assembly within the yeast cytosol. This scaffold was used to bring together a split luciferase in S. cerevisiae, resulting in a 3.6- to 20-fold increase in luminescence signal compared to the controls. Temporal control of the scaffold was also successfully demonstrated. Next, the scaffold was successfully applied to increase the amount of deoxyviolacein (desired) by 2-fold relative to prodeoxyviolacin (undesired) in a partial violacein pathway. To assess the generality of this colocalization method in other yeast systems, the split luciferase reporter system was evaluated in Kluyveromyces marxianus; RNA scaffold expression resulted in up to 1.9-fold increase in luminescence signal. Furthermore, this Cas6-mediated RNA scaffold has shown promise to colocalize the enzymes acetyl-CoA carboxylase and 2-pyrone synthase into a metabolon to improve synthesis of TAL, and for forming functional degradation complexes for targeted ubiquitination and protein degradation. These are the first studies that utilize RNA as synthetic protein scaffold within the yeast cytosol. The flexibility of the design suggest that this strategy can be used to create metabolons and for other applications in a wide range of recombinant hosts of interest.

Increased demand for petroleum resources coupled with political, environmental and supply-chain concerns has increased the investment and interest in bio-based chemical production. In particular, polyketide production in microorganisms facilitates the expansion of sustainable, local and scalable chemical production due to the versatility of polyketides as precursors in catalytic conversion to a range of high-value chemicals including pharmaceuticals, food additives and plastics. For these reasons, optimization of polyketide production in yeast aids the sustainability and affordability of various consumer products. This work engineers and evaluates the yeasts Saccharomyces cerevisiae and Kluyveromyces marxianus for the production of triacetic acid lactone (TAL), a platform chemical precursor to 1,3-pentadiene, sorbic acid and other commodity and high-value products. When considering the production of a specific product, one fundamental choice is host selection. In conjunction with traditional host S. cerevisiae, we explore development of the TAL production platform in a thermotolerant, rapidly growing and industrially relevant yeast K. marxianus. Prior to metabolic engineering, we demonstrate production of 1.24 g/L TAL from xylose, a low-cost substrate. The yield of this process on xylose (0.0295 mol TAL/mol carbon) is comparable to the state-of-the art in other hosts from glucose and demonstrates a breakthrough for xylose utilization in yeast.

To build on this versatile K. marxianus platform, and to improve the S. cerevisiae one, CRISPR-Cas9 was utilized to introduce two heterologous enzymes that divert carbon away from CO2 production, resulting in a 60% improvement in TAL yield. This pathway was then introduced into a genome scale model and flux balance analysis was performed using OptKnock, resulting in new gene knockout targets that are complementary to this new pathway. In parallel to both rationally designed and computational-based metabolic engineering strategies, we developed a novel growth selection method which is built from a genome-wide CRISPR-dCas9 library and a FapR/FapO malonyl-CoA sensor native to Bacillus subtilis. This method couples growth rate with intracellular malonyl-CoA levels such that over many generations, yeast strains with improved malonyl-CoA are enriched within a population and can be identified through next generation sequencing (NGS).

Due to significant developments in CRISPR-Cas9, metabolic modeling and sequencing affordability, metabolic engineering can advance via many different approaches including rational, computational and genome-scale methods. In this work, we incorporate each of these strategies to improve the understanding of both the S. cerevisiae and K. marxianus platforms for the production of polyketides and to develop a framework for enhanced production of other acetyl- and malonyl-CoA based products.

The production of fuels and chemicals from biorenewable resources is important to alleviate the instability in supply costs, growing environmental concerns and competing policy agendas associated with the use of petroleum feedstock. Many of these petroleum-derived fuels and chemicals can be directly or functionally substituted with bio-produced molecules. Among these compounds, short (SCFAs) and long chain fatty acids (LCFAs) can be utilized as free fatty acids or can fulfill a role as platform molecules using their functional group as a target for chemical catalysis. We have designed Saccharomyces cerevisiae strains to overproduce long and short chain fatty acids by combining novel enzyme and metabolic engineering strategies. LCFAs were increased by targeting three native intracellular processes: b-oxidation, activation (via acyl-CoA synthetases), and neutral lipid recycle. A sextuple mutant (faa1 faa4 fat1 faa2 pxa1 pox1) overexpressing the diacylglycerol acyltransferase, DGA1, and the triacylglycerol lipase, TGL3, yielded 2.2 g/L extracellular free fatty acids. SCFAs were overproduced by linking a Rattus norvegicus derived TEII to the Homo sapiens derived fatty acid synthase (hFAS). With the use of a Bacillus subtilis derived Sfp phosphopantetheine transferase, extracellular caprylic (C8) and total SCFA acid levels reached 63 mg/L and 68 mg/L, respectively. Combined over-expression of the phosphopantetheine transferase with the hFAS mutant resulted in C8 titers of up to 82 mg/L and total SCFA titers of up to 111 mg/L. Further enzyme and pathway engineering strategies were employed to increase SCFA production. By codon optimizing hFAS and targeting the specific degradation pathways for SCFAs in an engineered strain (faa2 ant1 pex11), we were able to increase hexanoic, octanoic, decanoic and total extracellular short chain fatty acids more than 75-, 220-, 120-, and 400-fold, respectively, relative to parent strain BY4741. Our novel combinations of pathway interventions have led to the highest extracellular SCFAs and LCFAs reported for S. cerevisiae. These engineering strategies should also prove useful for the increased production of other fatty acid derived biomolecules in yeast.

Saccharomyces cerevisiae is a promising microorganism for the production of ethanol for fuel, and the synthesis of precursors to industrial chemicals and natural products. The goal of this research was to engineer S. cerevisiae strains for the enhanced synthesis of these products. For the economical production of bioethanol, complete use of hemicellulosic sugars is necessary. Thus, the fungal arabinose pathway was imported and evaluated. The NADH-dependent L-xylulose reductase (ALX1) from Ambrosiozyma monospora was introduced to balance the use of redox cofactors. Three fungal arabinose pathway genes (XYL1, lad1, and lxr1 or ALX1) were codon and codon pair optimized for expression in S. cerevisiae. Various combinations and copy numbers of the three required genes were evaluated by measuring growth and xylitol production; use of ALX1 resulted in up to 9-fold higher xylitol titers relative to LXR1. Arabinose uptake, the rate-limiting step for arabinose utilization in S. cerevisiae, was also addressed by creating a chimeric protein of Gal2 and Hxt1. Polyketides are versatile molecules that can be industrial chemical precursors as well as drug precursors. Five different thioesterases were compared for the release of dihydromonacolin L (DML), the precursor to the cholesterol lowering agent lovastatin, from LovB in S. cerevisiae, and AptB was identified to be the best candidate. Most polyketides use acetate as starter unit and malonate or other malonate-based molecules as extender units. 6-methylsalicylic acid synthase (6-MSAS) was chosen as the model system for the engineering of S. cerevisiae to increase the intracellular availability of acetate and malonate. 6-MSA titer showed improvement by 50% via deletion of PYC1, by 3-fold via engineering of Acc1 (9-fold in activity), and by 90% via combined use of N-degron tagged URA3 and autoselection for the expression of 6-MSAS in complex medium. The strategies developed in this research contribute to the metabolic engineering of S. cerevisiae for the synthesis of ethanol, and biochemical and drug precursors.