Optimization and scale-up of a plasmid DNA production process

A 4-factor, 2-level DOE was set up to study the cell growth and pDNA expression of a green fluorescent protein (GFP) plasmid in One Shot TOP10 E. coli cells. The experimental design included:

• Type of growth medium (LB vs TB).
• Cultivation vessel (shake flask versus spin tubes).
• Nutrient supplementation (feed supplementation vs no feed supplementation).
• Carbon source (glycerol vs glucose).

Each factor was evaluated at two levels to understand their individual and interaction effects on cell growth and pDNA expression.

The transformed E. coli cells were inoculated into the respective media and grown under controlled conditions, with OD₆₀₀ and pDNA concentration measured at specified intervals to assess cell growth and plasmid production. This systematic approach enabled the screening and identification of growth conditions for maximizing cell density, plasmid production, and providing valuable insights for transferring the processes into an automated fermenter platform. A summary of the experimental setup is shown in Table 1. The interactions tested during the experimental work are shown in Table 2. Table 3 details the parameters assigned for each condition tested.

Table 1. Factors and levels for DOE

Table 2. Interactions tested for DOE

Table 3. Parameters tested during the upstream screening study

Selected growth parameters from the DOE screening study were applied to a 1 L benchtop fermenter to assess process feasibility and productivity. TB medium with glycerol as the carbon source was used, with feed supplementation during the run. The feeding strategy was established through theoretical calculation of the carbon demand per dry cell weight with unlimited substrate, then adjusted to the targeted growth rate. The setpoints for the continuous feed with a 250 g/L glycerol fortified yeast extract are shown in Table 4. Feeding commenced when glycerol concentration dropped below 1 g/L or dissolved oxygen (DO) levels indicated carbon source depletion. The fermenter was set up to control temperature at 37°C, DO at 30%, and pH of 7.0 ± 0.3. The DO cascade parameters for the fermentation were set considering both the agitation and aeration strategies, as shown in Table 5, to provide the necessary DO levels demanded by the process. Online sensors monitored and controlled DO, temperature, and pH. Hourly samples were collected for metabolite analysis and plasmid DNA quantification. Harvesting occurred at hour 16 and when OD₆₀₀ exceeded 45 absorbance units (AU). The bioreactor temperature was reduced to 25°C and agitation decreased to prevent vortexing during the harvest procedure.

Table 4. Feeding setpoints during the 1 L fermentation run

Table 5. DO cascade parameters for the 1 L fermentation run

The growth parameters identified during the DOE screening study, and later transferred to a 1L bench-scale fermenter to demonstrate process feasibility, were applied to a Xcellerex™ XDR-50 bioreactor with a 30 L working volume to demonstrate process scalability at pilot scale. Throughout the run, process performance was monitored, including cell growth profile, metabolites, and pDNA productivity. TB with glycerol as the carbon source and feed supplementation were used for the XDR-50 bioreactor run. The feeding strategy was established through theoretical calculation of carbon demand and modified, based on the performance of the 1 L fermentation run. The feeding setpoints are shown in Table 6. Feeding commenced when glycerol concentration dropped below 1 g/L or when DO levels indicated carbon source depletion, similar to the bench-scale run. The XDR-50 bioreactor was set up to control temperature at 37°C, DO at 30%, and a pH of 7.0 ± 0.3. The DO cascade parameters for the XDR-50 run were set by considering both the agitation and aeration strategies, as shown in Table 7 and Table 8, to provide the necessary DO levels demanded by the process. Online sensors monitored and controlled DO, temperature, and pH. Hourly samples were collected for metabolite analysis and pDNA quantification. Harvesting occurred at hour 16 and when OD₆₀₀ exceeded 45 AU. The bioreactor temperature was reduced to 25°C and agitation decreased to prevent vortexing during the harvest procedure.

Table 6. Feeding setpoints during the XDR-50 run

Table 7. DO cascade parameters for the XDR-50 run

Table 8. Agitation DO cascade parameters for the XDR-50 run

Parameter screening DOE

The parameter screening DOE were performed using the following methods:

Incubation conditions: For all steps, shaking incubators were set at 300 rotations per minute (rpm), 25 mm throw, temperature was controlled at 37°C, and 80% relative humidity (RH), with CO₂ control turned off.

Culture inoculation:

1. Prewarmed 20 mL LB broth aseptically transferred to a baffled 125 mL SF for a minimum of 20 min in a 37°C incubator.
2. Thaw glycerol stock vial in a 37°C water bath until a single ice crystal remains (~ 2 min).
3. Immediately transfer the full contents of the vial to the baffled 125 mL SF containing the prewarmed LB Miller broth.
4. Incubate at 37°C, 300 rpm, 0% CO₂, and 80% RH.
5. After 6 h, transfer to N-1 at 30 mL in 125 mL SF at a target OD₆₀₀ of 0.02.
6. After 6 h, transfer to N-0 at a target OD₆₀₀ of 0.2 for each condition, as shown in Table 3.

Sampling: Samples were taken every other hour, starting at 6 h to measure OD₆₀₀, pH, and metabolites.

Feeding: Feeding strategy was calculated based on the following equation

F(t) = (µX_BV_B/S_fY_X/S) * e^µt, where:

• F(t) = Feed rate (L/h)
• µ = Desired specific growth rate (h^-1)
• X_B = Biomass at end of batch phase (g dry cell weight/L culture)
• V_B = Initial volume of culture
• S_f = Substrate limiting concentration (g/L)
• Y_X/S = Yield coefficient of biomass from substrate
• t = Time since start of batch

Biomass at the end of the batch phase (X_B) and yield coefficient of biomass from substrate (Y_X/S) were determined during batch shake-flask studies using the same cell line used in these experiments.

Criteria for harvest: Each condition was maintained until OD₆₀₀ reached >12 or exponential growth stopped. For conditions where exponential growth continued, culture duration was extended.

pDNA extraction: At the end of the incubation period, cells were harvested by centrifugation, and pDNA was extracted using a commercially available pDNA extraction kit. Yield was quantified using a spectrophotometer. Purified plasmid was enzymatically cleaved with restriction enzymes and visualized on agarose gel electrophoresis.

pDNA production at 1 L scale

The 1 L scale pDNA production run was performed using the following methods:

• Bioreactor setup: 1 L stirred-tank bioreactor with online control of temperature, pH, DO, and agitation.
• Inoculum preparation: The seed culture was grown in TB media until OD₆₀₀ reached 2.0 AU, then used to inoculate the bioreactor to an initial OD₆₀₀ of 0.2 AU.
• Inoculation and growth: The bioreactor was inoculated with the seed culture, and growth was monitored by measuring OD₆₀₀, metabolites, and sampling for pDNA extraction at regular intervals.
• Fermentation conditions:
• Temperature: 37°C during the production phase.
• pH: Maintained at 7.0 ± 0.3 using phosphoric acid and ammonium hydroxide.
• DO: Maintained above 30% using a cascade control strategy (agitation and air and oxygen flow rate).
• Agitation: Initial speed at 300 rpm, increased to 450 rpm to maintain DO target.

• Feeding strategy: Feeding was initiated when glycerol concentration fell below 1 g/L or DO trend started to show signs of carbon source depletion. Feeding regimen was decided based on the equations outlined in Parameter screening DOE.
• Sampling: Regular sampling every 1 h to monitor OD₆₀₀, metabolites, and pDNA yield.

Harvesting: Harvest was initiated when OD₆₀₀ exceeded 45 AU. The temperature was adjusted to 25°C and agitation reduced to prevent vortexing during harvest.

pDNA production at 30 L scale using Xcellerex™ XDR-50 bioreactor

The 30 L scale pDNA production run was performed using the following methods:

• Bioreactor setup: Xcellerex™ XDR-50 dual purpose stirred-tank bioreactor (30 L working volume) with online control of temperature, pH, DO, and agitation.
• Inoculum preparation: Seed culture was started from a frozen glycerol stock and grown in a 150 mL baffled flask with 30 mL TB media until OD₆₀₀ reached 2.0 AU. The 30 mL culture was used to inoculate the P1 flask at 500 mL working volume, in a 3000 mL baffled flask. The P1 flask was seeded at OD₆₀₀ of 0.2 AU. The shake flask was grown until OD₆₀₀ reached approximately 6.0 AU then used to inoculate the bioreactor to an initial OD₆₀₀ of 0.2 AU.
• Inoculation and growth: The bioreactor was inoculated with the seed culture, and growth was monitored by measuring OD₆₀₀, metabolites, and sampling for pDNA extraction at regular intervals.
• Fermentation conditions:
• Temperature: 37°C during the production phase.
• pH: Maintained at 7.0 ± 0.3 using phosphoric acid and ammonium hydroxide.
• DO: Maintained above 30% using a cascade control strategy (air and oxygen flow rate).
• Agitation: Speed set to 350 rpm.

• Feeding strategy: Feeding was initiated when glycerol concentration fell below 1 g/L or DO trend started to show signs of carbon source depletion. Feeding regimen was decided based on the equations shown in Parameter screening DOE.
• Sampling: Regular sampling every 1 h to monitor OD₆₀₀, metabolites, and pDNA yield.
• Harvesting: Harvest was initiated when OD₆₀₀ exceeded 45 AU. The temperature was adjusted at 25°C and agitation reduced to prevent vortexing during harvest.

Parameter screening experiments

The analysis of individual effects shows that use of glycerol has the highest positive impact on cell density and pDNA production (Figs 1 and 2). The analysis also shows that the use of nutrient rich media (TB), and the use of shake flasks resulted in a positive impact on both cell density and peak pDNA production (Figs 1 and 2). The analysis of interaction effects on cell growth reveals that the combination of TB with glycerol, feed supplementation with glycerol, and use of shake flasks with media type enhance cell density, as well as pDNA production (Figs 3 and 4).

These results suggest that the nutrient-rich composition of TB, when paired with glycerol's efficient utilization as a carbon source, creates an optimal environment for the proliferation of the selected E. coli strain. Additionally, the interaction between glycerol and feed supplementation also results in high cell densities, indicating that the synergistic effect of an external nutrient feed with glycerol further boosts cell growth. These findings suggest that the optimal growth media that achieve high cell densities and improved pDNA productivity with the One Shot Top10 E.coli strain is TB media with glycerol as a carbon source and supplemented with feed. These results highlight the importance of screening and selecting appropriate media and supplementation strategies in optimizing microbial cultures for maximum cell yield.

Fig 1. Graphs showing the effect of (A) vessel type, (B) media types, (C) carbon source, and (D) feed supplementation on cell growth.

Fig 2. Graphs showing the effect of (A) vessel type, (B) media types, (C) carbon source, and (D) feed supplementation on pDNA Titer.

Fig 3. Graphs showing the interaction effects on cell growth.

Fig 4. Graphs showing the effect of (A) vessel type, (B) media types, (C) carbon source, and (D) feed supplementation on pDNA production.

pDNA production at 1 L scale

The OD₆₀₀ readings indicate a steady increase in cell density over time, suggesting healthy cell growth during the fermentation process (Fig 5A). The pDNA titer measurements showed a steady increase in pDNA concentration, achieving a titer of 116.4 mg/L at 16 h (Fig 5B). In-line pH measurements fluctuated slightly, with control measures maintaining the pH within the desired range (Fig 5D). The off-line pH levels remained relatively stable with minor fluctuations within acceptable ranges (Fig 5C). Lactate concentration initially increased and then plateaued around 4 h, indicating stabilization (Fig 5E). Ammonium concentration rose during the first half of the process, then started to decrease after 8 h of culture. The ammonium concentration spiked upwards at the final analysis point indicating potential late-stage cell stress (Fig 5F). Feeding was initiated at hour 3 due to glycerol concentration being below 1 g/L. It accumulated over the first few hours indicating that feeding may have been started prematurely. Once the glycerol consumption rate caught up with the feeding strategy, growth was controlled at the targeted rate (Fig 5H). Vessel temperature was controlled within the desired range (Fig 5I). Agitation was adjusted from 300 to 450 rpm over time to maintain the DO setpoint, with DO stabilizing after the agitation adjustment to 450 rpm. DO levels showed spikes that were related to the increases in agitation rates and feeding times. The continuous feed pump rate data indicated the nutrient feed rate into the system increased during the run, as expected (Fig 5J). No antifoam was added post-inoculation, suggesting effective foam control, possibly due to the run being executed at half the working volume with the top mixer aligning with the culture level.

Fig 5. 1 L benchtop bioreactor cell growth, pDNA titer, metabolite, and DO trends.

pDNA production at 30 L scale using Xcellerex™ XDR-50 bioreactor

Data analysis demonstrates that the bioreactor supported robust cell growth and effective plasmid production, with the appropriate consumption of carbon and nitrogen sources (Fig 6). The objectives of achieving an OD₆₀₀ greater than 45 AU were successfully met. The growth trends demonstrated slower growth during the initial 4 h of fermentation, followed by an increase in growth to achieve the expected target OD₆₀₀ (Fig 6A). This initial slower growth may be attributed to a lower inoculation density, causing the culture to require longer time to achieve a similar cell density to the benchtop run within the same timeframe. Nevertheless, overall measurements confirm expected cell doubling during the fermentation process.

Metabolite profiles were as expected and aligned with the benchtop bioreactor run. Glycerol concentration decreased steadily during the process (Fig 6C). and feeding was initiated when the glycerol concentration reached 1 g/L. Glycerol concentration continued to remain under 1 g/L for the duration of the run (Fig 6C). These trends indicate that glycerol was utilized as the primary carbon source. The lactate concentration was initially elevated but declined over time suggesting that lactate was produced as a metabolic byproduct subsequently consumed or converted by the cells. Ammonium levels also decreased over time, indicating utilization of ammonium as a nitrogen source by the cells (Fig 6F). Throughout the process, pH was maintained within the setpoint, with notable changes correlating to inoculation, feed initiation, and harvest point (Fig 6D).

Overall, the data demonstrates that the bioreactor run was successful, achieving target cell growth and plasmid production, with appropriate utilization of carbon and nitrogen sources. The fluctuations in pH and DO levels were managed to maintain optimal conditions for the process. Peak plasmid titer reached 146.63 mg/L at 16 h, similar to the benchtop bioreactor run (Fig 6B). Manual manipulation of the DO control was required during this engineering run, with adjustments made to the gas flow rate (SLPM) at various intervals, as shown by numbering (Fig 6G) to mitigate DO spikes. All other process conditions were maintained within study parameters.

Fig 6. Xcellerex™ XDR-50 bioreactor run trends for cell growth, pDNA titer, metabolites and DO.

Conclusions

The results of the DOE study identified TB media, supplemented with feed, and glycerol as a carbon source, as the optimal medium for the production of this plasmid DNA (pDNA) in One Shot Top10 E. coli. This medium was chosen due to its ability to achieve higher cell density, maintain a more stable pH during the run, and increase pDNA yield.

We successfully achieved an OD₆₀₀ of 70 absorbance units (AU) in the benchtop bioreactor, with a peak titer of 116.4 mg/L at 16 h. OD₆₀₀ readings suggest that the run did not reach the stationary phase, indicating that the culture duration could potentially be extended to achieve higher cell density and improve pDNA production. Metabolite readings indicate that the feeding strategy has the potential to be adjusted for future runs to enhance cell growth and further improve pDNA productivity.

The scale-up to the 50 L bioreactor was also successful, with minor adjustments to maintain oxygen transfer and mixing. We successfully exceeded the target of 45 absorbance units at OD₆₀₀. Peak plasmid titer reached in the Xcellerex™ XDR-50 bioreactor was 146.6 mg/L at 16 h, demonstrating a 22% increase in pDNA productivity. The increase in pDNA productivity may be attributed to the feeding regimen and the slower cell growth compared to the 1L run.

CY51386-09DEC25-AN