Protease titer drop after a peptone lot change
An industrial biotechnology company runs a 50 L stirred-tank bioreactor for alkaline protease production by Bacillus licheniformis. The historical batch reached 1200 U/mL in 24 hours with 18 g/L dry biomass. After switching to a cheaper peptone lot, the same recipe gives 780 U/mL but 22 g/L biomass. During the run, dissolved oxygen falls below 10% for four hours, pH rises uncontrolled to 8.6, and foaming overflows the exhaust filter. The vessel is filled to 70% of its working volume. Downstream is a 30 kDa ultrafiltration step, and the product specification is at least 900 U/mL. The production manager proposes raising aeration from 0.5 to 1.5 vvm, adding antifoam, and shipping the batch after blending it with a retained high-potency lot.
1. Use the numbers to diagnose why the titre fell, and distinguish volumetric from specific productivity.
Read the worked answer
The key comparison is specific, not volumetric. Historically, 1200 U/mL divided by 18 mg/mL biomass gives about 67 U per mg dry biomass. Now 780 divided by 22 gives about 35 U per mg. Volumetric titre fell by 35%, but specific productivity fell by roughly 47%. More biomass is being made with less product per cell, so the cheaper peptone is supporting growth while something is limiting product formation or destroying product.
Four candidate causes fit the data:
- Peptone quality: lower free amino nitrogen or a different amino acid profile can derepress growth but reduce protease synthesis. Peptone lot variability is a classic cause of this pattern.
- Oxygen limitation: DO below 10% for four hours is severe for an obligate aerobe, and protease synthesis is oxygen-demanding. The 70% fill plus foaming leaves too little headspace for oxygen transfer.
- pH: an uncontrolled rise to 8.6 changes both growth and the stability of the secreted protease.
- Proteolysis: a serine protease can degrade itself and other proteins, so a longer or stressed fermentation can reduce measured activity even when protein is produced.
The most likely explanation is peptone lot quality interacting with oxygen limitation and loss of pH control, not a single cause.
2. Design a controlled investigation that would separate these causes before any process change is committed.
Read the worked answer
Freeze a reference sample of the old peptone and characterise both lots: total nitrogen, amino nitrogen, ash, moisture, sodium chloride and endotoxin. Then run small-scale (2 to 5 L) bioreactors in a factorial design rather than changing one factor at a time:
- Factor 1: old versus new peptone.
- Factor 2: pH controlled at 8.0 versus uncontrolled.
- Factor 3: DO maintained at 30% versus allowed to fall below 10%.
Measure biomass, protease activity, residual peptone and protease stability over time. Include a cell-free incubation of active protease in spent medium to test whether the enzyme is being degraded. Measure kLa at 0.5 and 1.5 vvm; increasing vvm may improve oxygen transfer but will worsen foaming and filter wetting, so impeller speed, back pressure and a mechanical foam separator should be compared. Reduce fill volume to 60% to create headspace.
Reasoning: the historical and current runs differ in several variables at once, so only a design that varies them independently can attribute the loss. Only then should the aeration or antifoam change be made.
3. Should the batch be blended with the retained high-potency lot and released? Justify your answer.
Read the worked answer
No. Blending to reach a specification is not a legitimate correction for an out-of-specification fermentation. The batch failed its process parameters: DO below 10%, pH uncontrolled, foaming overflow and a titre below 900 U/mL. Those deviations may also have changed product quality, including proteolytic degradation, aggregation or glycosylation-independent stability, and no data are presented to show the blend is homogeneous or stable.
The quality objections are specific:
- Traceability: a blend creates a new lot whose composition and homogeneity must be validated, with mixing studies and replicate sampling.
- The blend would mask a process failure and prevent root-cause investigation.
- Release testing must be done on the blended material, not calculated from the high-potency component.
- Any change to the manufacturing process requires change control and, for a regulated product, may require validation and notification.
The correct action is to quarantine the batch, investigate the peptone lot and the oxygen and pH deviations, and release only material made under validated conditions that meets specification.