In addition to the selection of new recombinant clones, media optimization is another key process step in early bioprocess development. Commonly, culture media are optimized in shake flasks or fermentors. In these culture vessels volumes from 10 mL–1.0 L are necessary to run fermentations. The media preparation is normally performed manually, often causing pipetting or calculation errors. Due to the scale and limited throughput seen with these bioreactors, researchers are faced with a number of limitations. The huge work effort and significant consumption of the often expensive media components also limit the use of shake flasks and fermentors.
The acquisition of bioprocess information, however, still requires a lot of sampling to perform offline analysis of biomass and protein concentration, as well as byproduct concentrations. Consequently, many bioprocess research labs are attempting to reduce this effort and improve development times by using small-scale bioreactors.
Small-scale stirred tank bioreactors, bubble columns, and microplate-based bioreactors are already being used on a broad scale. Two huge advantages of using microplate-based bioreactors are their intrinsic high-throughput capacity and ease of automation. Automated media preparation in microplates is the next logical step.
RoboLector S Platform
The expansion of m2p-labs’ RoboLector™ platform to include a simple and cost-effective pipetting robot was recently announced. The development of the RoboLector S (small and smart) was conducted with the following improvements in mind: ease of use, compatibility with microplates, inclusion of up to 26 media components, cost-effectiveness, size that fits into standard clean benches, and DoE implementation.
The RoboLector S (Figure 1) is now available in a smaller footprint that fits into standard clean benches and allows for easy integration into existing cabinets. The robot allows for the allocation of up to 37 different sterile stock solutions and water on the worktable, though only 26 of them can be actively implemented into the preparation plan.
The worktable (Figure 2) includes a disposable tip rack, tip drop station, and stock solution vessels. The standard target microtiter plate is a 48-well Flowerplate® or a 48-round-well CellCulture Plate. The layout can be aligned to other microplate formats as well.
The pipetting volume ranges from 10 to 900 µL, whereas volumes higher than 900 µL are pipetted in two or more steps. Initially, water is prefilled into the wells to avoid precipitation of any media component during media preparation. After water, all other media components are pipetted into the wells from their stock solutions.
After media preparation and inoculation of the individual wells, the target microplate can be placed in a normal shaking incubator or can even be used in the BioLector® system.
The BioLector system is able to monitor all relevant fermentation parameters online without interrupting shaking. Biomass concentration, pH and DO values, and three fluorescence signals can all be measured in parallel. Here, the combination of RoboLector S and BioLector facilitates the gathering of insights from the bioprocess due to online monitoring of the process parameters instead of using just offline analysis of final samples.
The RoboLector S is supported by RoboLector Agent software. This software includes a media configuration window, which contains the medium library and the media components list. In the medium library, all media components used in the laboratory can be entered with their lot number and solubility.
In the media component list, the solutions are assigned to the vessels used on the worktable. Water and 26 stock solutions can be placed on the worktable. Feedback functions are implemented in the media component list to prevent errors. Solubility is also checked automatically. Text fields for additional information are available for documentation. An overview of the stock solutions on the worktable and the required volumes and concentrations can be printed.
The second window contains the configuration of the preparation process. On one side of the window the desired overall volume for each well is displayed; on the other side, the corresponding pipetting matrix is shown. The matrix connects a specific well in the target plate with the defined amount of stock solutions in the form of a table. The entries can be made either as volume or concentration.
If concentrations are entered, the software will automatically calculate the necessary volume of stock solution. Several feedback functions are implemented in the matrix to avoid logical errors. In addition, a fill-with-water function fills up the overall well volume with sterile water. Experiment data files can easily be imported or exported.
Design of Experiments (DoE)
Interest in DoE has been increasing over the past few years, and the concept is now broadly used. In fermentations, DoE is being used to assess medium composition or other process parameters that influence the formation of biomass or product.
Using DoE, users get an impression of how strong influences are and if there are interactions between single factors. Also an initial impression of space-time-yield or productivity can be drawn out of these experiments. Furthermore, fermentation parameters can be optimized efficiently based on this knowledge.
Using DoE, bioprocess behavior under different conditions can be simulated. DoE software provides graphical interpretation of the investigated results and helps to predict system behavior under specific conditions. Because of the wide variety of influencing factors, it is very difficult to apply DoE plans manually. Special plans for manual pipetting of microplates are available, but they are always a compromise. Designs for manual use are limited in their variability.
As a result, the application of DoE plans in the RoboLector S was of critical importance. RoboLector Agent software is able to import whole DoE plans. Researchers have to first design their DoE plan with separate DoE software such as MODDE (Umetrics, Sweden) or others. Then, users can import the DoE plan in the RoboLector Agent software and perform the pipetting with the RoboLector S.
After the fermentation run the user can analyze the outcome of the fermentations again in the DoE software and generate response surface models. If necessary, a new DoE plan can be conducted with the RoboLector S. Again here, online monitoring of the fermentations can be beneficial for the DoE model, because the more information available, the more precise the model can be.
The application of the RoboLector S is demonstrated in an experiment with Pseudomonas putida. Here, 48 parallel fermentations were performed with variations in the available nitrogen and phosphate concentration in the culture medium. Figure 3 shows the outcome of the experiment in a 3-D surface plot. The maximum biomass concentration was reached at 100% nitrogen and 50% phosphate concentration. The same considerations can be made for product formation and with other media components.
The new RoboLector S for media preparation is a compact liquid-handling system that can debottleneck current limitations in media optimization. As a result of the automated pipetting and the use of microplates, the study of multifactorial experiment layouts is no longer a limitation for researchers. The additional application of DoE ensures the statistical proof of these experiments. Online monitoring of all relevant fermentation parameters can provide much deeper bioprocess understanding with less effort. The implementation of these new tools opens up new ways of experimentation and at the same time contributes significantly to the QbD and PAT initiatives.