| In this thesis, metabolic flux analysis was used to identify potentially yield-enhancing metabolic engineering strategies. Such flux analyses were conducted for growth on the two most routinely employed carbon/energy sources in plasmid production culture media: glucose (for fast growth kinetics) or glycerol (for acetate minimization).;While the three aforementioned factors were found to improve plasmid production irrespective of the carbon source, additional metabolic factors were identified that pertained individually to growth on either glucose or glycerol. For glucose, the elimination of pyruvate kinase (Pyk) activity was predicted to be favorable. In fact, the largest theoretical gain in plasmid yield stands to result when Pyk is abolished if each of the strain-intrinsic factors is considered independently and plasmid selection remains based on marker production. Finally, specific to glycerol, increasing its carbon uptake and growth rate by engineering glycerol kinase to be insensitive to fructose-1,6-bisphosphate-mediated feedback inhibition (GlpKi) were shown to theoretically improve plasmid production.;Whether antibiotic resistance was used for plasmid selection or not, theoretical plasmid production in the GlpKi-deregulation/glycerol case was the highest of all cases considered, whereas the non-engineered/glucose case was the lowest. Intermediate were the non-engineered/glycerol and Pyk-deletion/glucose cases. When selection was by antibiotic resistance, Pyk-deletion/glucose outperformed non-engineered/glycerol, while the reverse was true if production of antibiotic resistance marker protein was unnecessary for selection.;This thesis extensively explored the impact of the Pyk-deletion strategy. A Pyk-deficient strain of E. coli (PB25) and its wild-type counterpart (JM101) were transformed with a pUC-based plasmid that relied on antibiotic resistance for selection. Most therapeutic plasmid backbones depend on antibiotic resistance markers for selection, and nearly all use the temperature-sensitive pUC origin of replication. This origin confers high plasmid copy numbers at 37°C, where growth kinetics are maximal, and even higher copy numbers upon shifting temperature to 42°C.;Modeling results revealed three metabolic factors that could potentially enhance plasmid production regardless of the carbon source: (i) overexpression of transhydrogenase or maximization of its activity for improved NADPH availability, (ii) minimization of acetate production, and (iii) using a plasmid whose backbone contains a selection mechanism that does not rely on antibiotic resistance marker production, as theoretical plasmid yield decreased sharply with increasing marker protein promoter strength.;When grown on glucose at 37°C, plasmid production in PB25 (16.1 mg/g, 9.3 mg/g/h) was found to be significantly better than in the cases where wild-type JM101 was grown on either glucose or glycerol, and production in JM101 was slightly better when grown on glycerol (9.6 mg/g, 4.8 mg/g/h) than glucose (4.1 mg/g, 3.1 mg/g/h). The same trend for the three cases occurred during the first 2.5 culture doublings after plasmid production was increased by a temperature shift from 37 to 42°C (PB25/glucose: 32.0 mg/g, 16.3 mg/g/h; JM101/glycerol: 36.7 mg/g, 14.7 mg/g/h; JM101/glucose: 9.9 mg/g, 6.5 mg/g/h). At later generations, glycerol-grown JM101 became overburdened by plasmid production, which reduced its growth rate and caused plasmid levels to rise in accord with decreased cytoplasmic dilution. As a result, prolonged plasmid production at 42°C was highest in glycerol-grown JM101 (73.3 mg/g, 22.7 mg/g/h), though glucose-grown PB25 still substantially outperformed glucose-grown JM101. In neither of these two latter cases did growth rate decrease due to plasmid-derived metabolic burden.;The trends with the experimental plasmid production results are in full agreement with those predicted by the metabolic flux analyses. Though the absolute values of plasmid production between those observed and those predicted vary, the fold-enhancements are remarkably similar. Thus, the model proved to be a valuable tool for predicting a mutation (i.e., Pyk deletion) and a medium alteration (i.e., glycerol vs. glucose) that resulted in substantially enhanced plasmid production when explored experimentally. (Abstract shortened by UMI.)... |
