The role of amino acids in the amplification and quality of DNA vectors for industrial applications.

In this study, we have demonstrated that the type and feeding regimen of amino acids have a significant impact on the quality as well as the quantity of DNA vectors produced. Nutrient pool and factorial design experiments were carried out in order to identify the amino acids involved in increased biomass and induction of plasmid amplification. Leucine, glycine, and histidine were responsible for increased biomass and leucine starvation in the presence of histidine was implicated in plasmid amplification. Supercoiling of the plasmid was optimized using a dual feeding strategy. As a result of this, a fed-batch fermentation strategy for the production of a 6.9 kb plasmid, pSVß, in Escherichia coli DH5α was developed. In batch fermentation, a maximum plasmid yield of 39.4 mg/L equivalent to 11.3 mg/g dry cell weight (DCW) was achieved with casein hydrolysate limitation. About 90% of plasmid was in the supercoiled (SC) form after 31 hr of fermentation but only remained so for a short period, leading to a very brief window for harvesting cells at scale. Subsequently, a fed-batch fermentation using a dual feeding strategy was employed. A mean maximum plasmid yield of 44 mg/L equivalent to 9.1 mg plasmid/g DCW was achieved. After 25 hr, 90% of plasmid was in the SC form and remained at this level for the remaining 10 hr of the fermentation, allowing adequate time for the harvesting of cells without the loss of supercoiling of product. This study emphasized that optimizing fermentation strategy and identifying the essential nutrients are beneficial for bioprocessing of plasmid DNA for therapeutic applications.

Large-scale plasmid DNA processing: evidence that cell harvesting and storage methods affect yield of supercoiled plasmid DNA.

The effect of bacterial-cell centrifugation and handling on the initial stages of plasmid processing was investigated. Escherichia coli cells containing either a 6 or 20 kb plasmid were grown in 75- and 450-litre bioreactors, and the process yield of the early recovery stages was characterized in terms of SC pDNA (supercoiled plasmid DNA) recovered. In all cases, the cells were totally recovered using either a continuous-feed, intermittent-solids-discharge, disc-stack centrifuge or a continuous-feed, batch-discharge, solid-bowl centrifuge. The cells were then either processed immediately or stored frozen. The centrifugation method considerably affected the yield of SC pDNA, and there was evidence that the intermittent discharge of cells from a centrifuge operating at high speed led to a sediment containing lysed cells and degraded pDNA. This led to estimated plasmid yield losses of up to 40% as compared with cells recovered from laboratory or solid-bowl centrifuges, where there is evidently no cell stress on discharge. By inference, the cell stress on feed to either of the continuous centrifuges studied was not implicated in product loss. Freezing of the recovered cells gives a convenient hold stage prior to further processing. In all cases, this extra freeze-thaw stage led to loss of SC pDNA, and this was in addition to the loss attributed to cell lysis during centrifugation discharge. Only average yields can be gained from pilot plant-scale studies; separate laboratory-based experiments indicated that this loss of SC pDNA is determined by the time and temperature for which the resuspended cells are held.

Use of operating windows in the assessment of integrated robotic systems for the measurement of bioprocess kinetics.

This study examines the utility of an automated liquid handling robot integrated with a microwell plate reader to enable the rapid acquisition of bioprocess kinetic data. The relationship between the key parameters for liquid handling accuracy and precision and the sample detection period has been characterized for typical low-viscosity (<2.0 mPa x s) aqueous and organic phases and for a high-viscosity aqueous phase (60 mPa x s), all exhibiting Newtonian rheology. The use of a simple graphical method enables the suitability of a given automation platform to be assessed once the user has determined the minimum sample detection period and the minimum accurate and precise dispense volume. This provides for a reduction in the duration of any experiment by maximizing well usage within each microwell plate. The suitability of employing an integrated automation platform to gather kinetic data for systems typical of those encountered in bioprocessing is analyzed via a series of case studies. Application to alkaline cell lysis, where disruption is complete within 120 s, showed that the range of available dispense volumes and the number of wells that can be utilized is limited. In contrast, analysis of a system exhibiting slow process kinetics, the fermentation of Escherichia coli TOP10 pQR239 in microwell plates, demonstrated that, for a typical sample detection period of 30 min, the only restrictions on the degree of well utilization are the liquid handling accuracy and precision and the volume capacity of the liquid handling robot. Finally, liquid-liquid extraction, an example of a kinetically independent operation, was also examined. In this case, only a single equilibrium measurement is required, which means that the only restrictions to the utilization of the integrated devices are the liquid handling accuracy and precision. Integrated automation platforms represent a powerful process development tool over traditional experimental methods used for bioprocess development. Smaller volumes of reagent and sample can be used to achieve greater throughput, while high levels of reproducibility and sensitivity are maintained.

Effects of fermentation strategy on the characteristics of plasmid DNA production.

The synthesis of supercoiled plasmid DNA (SC-pDNA) for therapeutic use will involve large-scale production in bioreactors. The success of these fermentations will be dependent on the interactions between the host organism, the recombinant plasmid vector and the growth environment. In the present study, the recombinant host, Escherichia coli DH5 alpha bearing the recombinant plasmid pSV beta, was grown in shake flasks, batch and exponentially fed-batch bioreactors. Specific and volumetric pDNA yields were increased 8- and 25-fold respectively using exponentially fed-batch cultures in comparison with shake-flask cultures. The percentage of SC-pDNA as a proportion of total plasmid DNA decreased over time in batch cultures, but remained relatively constant during fed-batch cultures. The relative merits of different modes of fermentation and their effects on the quality of alkaline lysate extracts of pDNA with respect to genomic contamination and the percentage of SC-pDNA are discussed.