Temperature-dependent dynamic control of the TCA cycle increases volumetric productivity of itaconic acid production by Escherichia coli

Abstract

Based on the recently constructed Escherichia coli itaconic acid production strain ita23, we aimed to improve the productivity by applying a two-stage process strategy with decoupled production of biomass and itaconic acid. We constructed a strain ita32 (MG1655 ΔaceA Δpta ΔpykF ΔpykA pCadCs), which, in contrast to ita23, has an active tricarboxylic acid (TCA) cycle and a fast growth rate of 0.52 hr^-1 at 37°C, thus representing an ideal phenotype for the first stage, the growth phase. Subsequently we implemented a synthetic genetic control allowing the downregulation of the TCA cycle and thus the switch from growth to itaconic acid production in the second stage. The promoter of the isocitrate dehydrogenase was replaced by the Lambda promoter (p_R ) and its expression was controlled by the temperature-sensitive repressor CI857 which is active at lower temperatures (30°C). With glucose as substrate, the respective strain ita36A grew with a fast growth rate at 37°C and switched to production of itaconic acid at 28°C. To study the impact of the process strategy on productivity, we performed one-stage and two-stage bioreactor cultivations. The two-stage process enabled fast formation of biomass resulting in improved peak productivity of 0.86 g/L/hr (+48%) and volumetric productivity of 0.39 g/L/hr (+22%) in comparison to the one-stage process. With our dynamic production strain, we also resolved the glutamate auxotrophy of ita23 and increased the itaconic acid titer to 47 g/L. The temperature-dependent activation of gene expression by the Lambda promoters (p_R /p_L ) has been frequently used to improve protein or, in a few cases, metabolite production in two-stage processes. Here we demonstrate that the system can be as well used in the opposite direction to selectively knock-down an essential gene (icd) in E. coli to design a two-stage process for improved volumetric productivity. The control by temperature avoids expensive inducers and has the potential to be generally used to improve cell factory performance.

Keywords: Escherichia coli; dynamic metabolic control; genetic switch; itaconic acid production; two-stage process; volumetric productivity.

Figure 1

Schematic representation of the metabolic networks of strains ita23, ita32, and ita36A. Green bold line: upregulation of gene; red crosses: gene deletions, or (in case of icd in ita23) strong downregulations. Orange point (ita36A): temperature‐dependent regulation

Figure 2

Dynamic control of icd expression. (a) The repressor CI857 is inactive at 37°C allowing expression of icd for an active TCA cycle and biomass formation. (b) At 30°C, the repressor becomes active, binds to the promoters, (p_R/p_L) and, therefore, blocks transcription of icd and thus the TCA cycle redirecting flux to itaconic acid

Figure 3

Top: Specific activities of the isocitrate dehydrogenase (ICD). The activity was normalized to the total protein content in the sample. Bottom: Relative expression of icd normalized to the wild type at 37°C. Error bars represents the standard deviation of two independent biological samples (for the activity) and three independent biological samples (for the expression level). The expression level of ita36A at 28°C was not determined

Figure 4

One‐stage and two‐stage bioreactor cultivations of ita36A at 28°C. Both processes differed with respect to their preculture conditions: the one‐stage preculture was conducted at 28°C (optimal for production), the two‐stage preculture at 37°C (optimal for growth). This enabled fast growth of ita36A at the initial stage of the two‐stage process (0–12 hr). The point where the transition from growth to production phase occurred is marked by the red dashed line. The addition of feed solution is marked by a sharp increase in glucose concentrations. The one‐stage process was conducted as triplicate, the two‐stage process as duplicate. Error bars represents the standard deviation of the independent biological samples. Glutamate was not produced in the two‐stage cultivation

Figure 5

Dependence of the volumetric productivity of Ita36A on the cultivation time and process design. Error bars represent the standard deviation of the independent biological samples