UC Berkeley

A diverse array of molecular functions has evolved in biological

systems. The ability to reorganize and utilize these remarkable capabilities in order to design pathways for in vivo chemical synthesis requires expanding on foundational biochemical principles in order to better understand the function of entire pathways, thereby allowing for de novo design. Furthermore, in order to integrate these synthetic pathways effectively into organisms such that they are capable of working at the desired high flux, it is necessary to understand the principles affecting pathway flux and pathway interactions with the native metabolism.

To this end, we have engineered Escherichia coli to synthesize n-butanol from acetyl-CoA. In design of a high flux de novo pathway, we have demonstrated that using the enzymatic reaction mechanism of the reduction of the enoyl-CoA intermediate as a kinetic control element can drive flux to the final product. This is in constrast to high yield pathways that use a thermodynamic control element to drive the pathway to completion, generally decarboxylation or sequestration of an insoluble product. These methods have the disadvantages of using ATP and limiting the range of target molecules, respectively. Using the enoyl-CoA reduction to drive pathway flux, we have constructed a chimeric pathway capable of producing titers of n-butanol competitive with native fermentative pathways in E. coli (4650 ± 720 mg/L).

The enoyl-CoA reductase, tdTer, plays a key role in the production of high titers of n-butanol. We characterized this enzyme structurally and biochemically to examine its potential for use with other substrates. The crystal structure of tdTer was determined to 2.00 Å resolution and shows the enzyme is highly similar to members of the FabV family of enoyl-CoA reductases. Biochemical studies show that similar to other enoyl-CoA (ACP) reductases, the enzyme utilizes an ordered bi-bi reaction mechanism initiated by binding of the NADH redox cofactor. Analysis of the activity and inhibition of the C₄, C₆, and C₁₂ substrates and products with a variety of binding loop mutants suggests the region is important in discrimination of chain length and points to the major portal as a target for modifying chain length specificity.

The introduction of tdTer as the enoyl-CoA reductase was shown to be a key facor in the high yield of n-butanol from our synthetic pathway. In order to study the flux through this step, we implemented a protein scaffold system to colocalize enzymes in the pathway via fusing the enzymes of interest to ligand binding domains. We colocalized tdTer with the enzyme catalyzing the previous step in the biosynthetic pathway and separately with enzyme catalyzing the next step in the pathway. Initial results suggest that the colocalization due to scaffolding can increase n-butanol production, and that the fusion of an affinity tag with the AdhE2 enzyme is capable of increasing specificity of the enzyme. Unfortunately, the deleterious effects of introducing the scaffolding system and the lack of reproducibility in the results prevents the use of the system to effectively probe flux through the n-butanol synthetic pathway.

In order to expand the ability of the pathway to probe the metabolism of E. coli, the pathway was modified to draw on the malonyl-CoA pool in addition to acetyl-CoA. To do so, the first enzyme was replaced with NphT7, which condenses one malonyl-CoA and one acetyl-CoA to one acetoacetyl CoA. The introduction of NphT7 creates a significant bottleneck primarily related to malonyl-CoA availability. Increasing the intracellular malonyl-CoA concentration through genetic engineering and heterologous expression was capable of increasing n-butanol production, but not to the levels seen in the acetyl-CoA dependent pathway. We would like to use the modified pathway to identify changes to NphT7 or the host metabolism capable of increasing flux through malonyl-CoA. In order to identify desired mutations to nphT7 and/or the E. coli genome, a strain was developed in which growth rate was correlated with n-butanol production. Use of this strain allowed for the selection of mutated strains with increased n-butanol production; however, low n-butanol yield under selecting conditions and contamination of high yield strains have thus far prevented the idenification of high yield strains. Additional method development allowing for new types of library design is likely necessary in order to identify mutants resulting in high yield malonyl-CoA based pathways.