Engineering the E. coli UDP-Glucose Synthesis Pathway for Oligosaccharide Synthesis Zichao Mao, Hyun-Dong Shin, and Rachel Ruizhen Chen* School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0100 A metabolic engineering strategy was successfully applied to engineer the UDP-glucose synthesis pathway in E. coli. Two key enzymes of the pathway, phosphoglucomutase and UDP-glucose pyrophosphorylase, were overexpressed to increase the carbon flux toward UDP-glucose synthesis. When additional enzymes (a UDP-galactose epimerase and a galactosyltransferease) were introduced to the engineered strain, the increased flux to UDP-glucose synthesis led to an enhanced UDP-galactose derived disaccharide synthesis. Specifically, close to 20 mM UDP-galactose derived disaccharides were synthesized in the engineered strain, whereas in the control strain only 2.5 mM products were obtained, indicating that the metabolic engineering strategy was successful in channeling carbon flux (8-fold more) into the UDP-glucose synthesis pathway. UDP-sugar synthesis and oligosaccharide synthesis were shown to increase according to the enzyme expression levels when inducer concentration was between 0 and 0.5 mM. However, this dependence on the enzyme expression stopped when expression level was further increased (IPTG concentration was increased from 0.5 to 1 mM), indicating that other factors emerged as bottlenecks of the synthesis. Several likely bottlenecks and possible engineering strategies to further improve the synthesis are discussed. Introduction Although carbohydrates are well-known for their structural roles and functions in energy storage, only recently have their crucial roles as information carriers in many biological processes been uncovered. Oligosaccharide moieties of the glycoconju- gates (glycoproteins and glycolipids) serve as recognition sites for hormones, antibodies, toxins, viruses, and bacteria. They also play key roles in cell-cell recognition, inflammation, can- cer, metastasis, and many other biological processes. This recent revelation of carbohydrates as medically relevant biomolecules has generated great interest in developing carbohydrate-based therapies for various diseases. In fact, several carbohydrate- based anti-infective drugs and cancer vaccines are in advanced human clinical trials and prospects of carbohydrate-based thera- pies for a wide range of diseases are evidently very bright (1). Unfortunately, this development is greatly hampered by the difficulties in generating these molecules to have the required specific linkage and anomerity, due to the intrinsic carbohydrate structure containing multiple hydroxyl groups with similar reac- tivity (2). Despite recent progress, the chemical synthesis of carbohydrate still relies on a laborious scheme of protection and deprotection to differentiate the subtle differences in hy- droxyl groups, making the chemical synthetic route a lengthy, low-yielding process. Enzyme-based strategies toward complex carbohydrates represent emerging technologies that have the potential to greatly simplify the process (3), because enzymes or whole-cell-catalyzed reactions afford high regioselectivity and stereoselectivity and can therefore avoid the protection and de- protection steps in assembling complex carbohydrate molecules. Glycosyltransferases are group-transfer enzymes catalyzing the transfer of a sugar molecule from its activated form to an acceptor (which can be another saccharide or an aglcone). The most commonly used activated substrates are sugar nucleotides such as UDP-glucose and UDP-galactose. The glycosyltrans- ferase-catalyzed reactions are highly regioselective and stereo- selective, and almost no side products are formed. However, enzyme substrate, the donor sugar nucleotide, is prohibitively expensive to use in a stoichiometric ratio. Therefore, using whole cells as catalysts where sugar nucleotide cofactors are regener- ated in situ is therefore preferred. Several whole-cell-based synthesis strategies were developed in the past few years, showing promises in deriving practical solutions to the challenges in large-scale synthesis of oligosac- charides. The elegant strategy using sucrose synthase allowed utilizing UDP (coproduct of the reaction) to regenerate sugar nucleotides without additional input of a high-energy compound (4). Microbial cell coupling employing three or more different types of recombinant microbes (e.g., Corynebacterium ammo- niagenes and two types of plasmid-carrying E. coli cells) was impressive in achieving product concentrations over 100 g/L (5-7). Encouraging results were also obtained with a single- strain-based approach (8-12). However, due to the enormous diversity of this class of molecules, it is likely that many varied strategies would be necessary. Further improvements in overall production efficiency (product concentration, productivity, and yield) are needed to bring biocatalytic synthesis to the realm of common practice. UDP-glucose is the starting point of the synthesis of other UDP-sugars (such as UDP-galactose and UDP-glucuronic acid). Its availability is thus a prerequisite for high-rate synthesis of any oligosaccharides containing these sugars. As shown in Figure 1, UDP-glucose is a multigene, multipathway product. The synthesis of UDP-glucose from glucose interfaces the glycolysis pathway, the pentose phosphate pathway, nucleotide synthesis, and energy production. In this paper, we report a metabolic engineering strategy aimed to maximize the partition of carbon flux flow toward UDP-glucose synthesis at the branch point of glucose-6-phosphate (G6P) (Figure 1). Particularly, the two enzymes directly involved in the synthesis of UDP-glucose 369 Biotechnol. Prog. 2006, 22, 369-374 10.1021/bp0503181 CCC: $33.50 © 2006 American Chemical Society and American Institute of Chemical Engineers Published on Web 01/28/2006