Application Note - Validation of Hamilton TCD Sensor During Pichia pastoris Fermentation
Improving In-Process Control of VHH Production Using Real-time Total Cell Density Measurement
- Industry: Biotechnology, pharmaceuticals
- Application field: Fermentation process monitoring and control, recombinant protein production
- Hamilton products: Dencytee Arc
Efficient production of recombinant proteins such as VHHs (nanobodies) requires accurate monitoring of cell growth. Traditional off-line methods – OD600, wet cell weight (WCW), and dry cell weight (DCW) – are labor-intensive and provide only intermittent data, limiting effective process control.
The Hamilton Total Cell Density (TCD) sensor enables continuous, real-time monitoring by combining near-infrared transmission and reflection. Transmission dominates at low culture densities, while reflection is more accurate at higher densities. To ensure reliable quantification under varying conditions (e.g., changes in agitation and aeration), custom off-line correlation models were developed and validated for different fermentation phases.
Methods
A Hamilton Dencytee TCD sensor was installed directly in a 7.5 L glass bioreactor for Pichia pastoris fermentation, producing VHH. The culture was controlled using an Eppendorf Bioflo320 system with dissolved oxygen (DO) maintained through cascade control of agitation speed and airflow.
The TCD sensor was connected to the Bioflo320 via a 4–20 mA interface for live trend display and simultaneously linked to a computer via Bluetooth for data handling in ArcAir software. ArcAir was used not only to visualize sensor trends but also to establish and validate custom off-line correlation models, which were critical for accurate online DCW calculation.
Growth phase (glycerol): Cells were inoculated into buffered glycerol complex medium (BMGY). Six off-line data points (OD600/WCW) were collected across the growth curve to establish the first correlation model.
Production phase (methanol): Recombinant VHH expression was induced with methanol, resulting in lower oxygen demand in this Pichia strain (mutS). Six additional correlation points were collected at higher densities, forming a second correlation model specific to this phase.
This dual-model approach, enabled by ArcAir, allowed adaptation of the TCD output to phase-specific fermentation conditions—capturing differences in oxygen transfer, agitation, and carbon source utilization between the glycerol and methanol phases.
Results
The TCD sensor reliably tracked cell growth, with reflection increasing and transmission decreasing as OD600 rose (Figure 1). However, results also showed that reflection signals were influenced by agitation
speed and aeration. During the transition from glycerol to methanol feeding, changes in agitation caused deviations, making a single universal model insufficient (Figures 2 and 3).
Applying phase-specific models significantly improved correlation with off-line DCW measurements (R² up to 0.98). A single universal model could not fully capture the dynamics of Pichia pastoris fermentation because the growth and production phases differ substantially in their operating conditions.
Authors:
Niels Haarman, Anna Kolecka
