Continuous biopropane production would be economical choice for large-scale processes as the gas product can be easily separated. Therefore, the product would not disturb increased cell growth levels which are needed for continuous processes. Product's purification and liquefaction could be also added next to a gas stream outline. Continuous manufacturing needs smaller equipment as the volumetric productivity is greater than for the regular batch processes. First step for scaling up with a production strain is to test it in a chemostat.
Because of the time limit, our propane producing E.coli strain wasn't competely constructed. We decided to try small scale and chemostat production with E.coli BL21 (DE3 ΔyjgB ΔyqhD, pET-TPC4 + pCDF-cAD + pACYC-Fdx-Fpr) made by Pauli Kallio's research group which differs from our own strain by two enzymes at the beginning of the reaction pathway. Thus, experiments with Kallio's strain largely revealed how our own strain would have behaved. More information available at the Project page.
Before trying to grow new cells in a reactor, the strain had to be tested with vial-scale cultivations. In this way we could gather information about the induction, cell growth rates in TB-media, and some estimates of how much time is needed for propane formation after adding IPTG. At the same time we could practise the usage of gas gromatography/mass spectrometry and make proper standards for propane detection. Basically, terrific broth (TB) containing induced bacteria would be added into 22 ml GC-vials. The cells are incubated different lenghts of time and propane accumulates into the head-space of the vial where the injection and analysis would happen. The protocol contains more detailed information for the preparation of samples.
Two samples were incubated for 17 h with 50 RPM shaking in 22 ml vials. Propane concentrations were 111 μg/l and 88 μg/l where the litres were as in the volume of cultivation media. Furthermore, the growth rate of the strain was significant as the OD600 (1 ml cuvettes) increased from 0.1425 to 0.5212 for 1h 10 min time scale. The rate was assumed to be much lower because TB-media contained four different antibiotics which were needed to keep plasmids inside of bacteria. Details about standards, peaks etc. for GC/MS can be seen in this file. The GC/MS analysis and calculcations were done by the personnel of VTT, which is mentioned in the Attributions page.
0.5 l chemostat experiment with 44 ml/h flow rate was successful. With 1.0 l/h aeration, the propane content of the reactor's gas phase was determined to be 4.5 μg/l and 22.7 μg/l with GC/MS 145 h after starting the experiment. The results were calculated when 0.1-0.5 μg of propane was gathered from reactor's gas exhaust into a 22ml gas chromatography vial. However, these values were below GC-standard line so the qualitative accuracy may have decreased. Propane was still identified to be present in gas. The GC-samples were also gathered when a third steady state was running which had much lower cell density than second steady state had. Nevertheless, the result suggests that propane could be produced industrially with continuous production and scale-up experiments should be done.
The batch phase took six hours, and soon after starting the process the amount of dissolved oxygen decreased to zero. Therefore, oxygen became the limiting factor of cell growth. Continuous phase was started and the first steady state was reached 39 h after starting to feed fresh media into the reactor. Overall, the steady state lasted for 15 hours. The density of cell population was calculated to be 9.85 g/l at this point, which can be assessed relatively high. Interestingly, our bacteria acted as a facultative anaerobe. Saturated oxygen level increased from zero for eight hours during the exponental phase of growth. Therefore, cells didn't require oxygen for growing at that time.
IPTG induction started with 2mM concentration and the population growth started to react to changed process conditions. New steady state was achieved 40 h later and the population density became 9.28 g/l. Thus, there was a 0.57 g/l difference of the biomass between the normal growth and the growth with propane production. For some reason, the growth started to decrease again and the second steady state didn't continue, when the feed stock was changed to a bigger container containing the same terrific broth because the old one was running out.
Propane samples weren't gathered from the second steady state during that day because of the time limit, so we had to wait over the weekend until the third steady state had settled. At this point the cell density was 2,625 g/l. Reactor's glucose consentration decreased from original 20 g/l to zero during the exponential grow phase as the cells used it as a carbon and energy source. However, it was hard to recognize whether glucose or oxygen limited more the growth. During the first steady state, glucose concentration increased when the need for the anabolism of biomass was lower. Glucose analyses were done with liquid chromatography which data is gathered here.
Before starting to feed fresh media, the end of batch phase was determined from decreased pH-level when the cells produced acids as a response to nutrient deficiency. TB-media used for cultivation contained phosphates which had buffer capacity but it wasn't enough. pH-control was assembled in order to maintain pH-values above 5.0.
More information about the experiment available at the Process data sheet and the Lab book. Details about standards, peaks etc. for GC/MS can be seen in this file. The GC/MS analysis and calculcations were done by the personnel of VTT. In the case of continuous production, the 22 ml GC-vials were just filled with gas without media.