Browsing by Subject "Biosurfactant"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Publication
    Applied molecular bioprocess control using RNA thermometers : exploiting temperature responsive elements for rhamnolipid production
    (2022) Noll, Philipp; Hausmann, Rudolf
    The highest titer reported for heterologous Rhamnolipid (RL) production is 14.9 g/L. However, biomass generation, as a large carbon sink, was a significant drawback in this process with roughly 50 more biomass than product produced. This problem is addressed in this thesis leveraging temperature as control variable and a molecular temperature sensor, an RNA thermometer (RNAT). RNAT generally refers to secondary loop structures, in the 5’ untranslated region of the mRNA, that form at certain temperatures and therefore regulate translation in dependence of temperature. The ROSE (repression of heat shock gene expression) RNAT evaluated in the first original research article in the heterologous system P. putida KT2440 pSynpro8oT_rhlAB originates from P. aeruginosa. The ROSE element regulates, in dependence of ambient temperature, the translation of rhlA and via a polar effect also the translation of rhlB therefore indirectly RL synthesis. It was found that in the ROSE RNAT-controlled system, the RL production rate was 60% higher at cultivations of 37°C than at 30°C. However, besides the regulatory effect of the RNAT, as revealed by control experiments, multiple unspecific metabolic effects may be equally responsible for the increase in production rate. After screening for even more efficient regulatory structures, a fourU RNAT was identified. Natively, this fourU RNAT regulates the expression of the heat shock gene agsA of Salmonella enterica and its regulatory capability can easily be modified by site-directed mutagenesis. The experimental data collected in the second original research article confirms the functionality of the fourU RNAT in the heterologous RL production system. The data suggested improved regulatory capabilities of the fourU RNAT compared to the ROSE element and a major effect of temperature on RL production rates and yields. The average RL production rate increased by a factor of 11 between 25°C and 38°C. Control experiments confirmed that a major part of this increase originates from the regulatory effect of the fourU RNAT rather than from an unspecific metabolic effect. With this system YP/X values well above 1 (about 1.4 gRL/gBM) could be achieved mitigating the problem of high biomass formation compared to product synthesis. Also, YP/S values of about 0.2 gRL/gGlc at elevated temperatures of 37-38°C were reached in shake flasks. The system was subsequently tested in a proof-of-concept bioreactor process involving a temperature switch. With this simple batch experiment and a temperature switch from 25°C to 38°C not only a partial decoupling of biomass formation from product synthesis was achieved but also an around 25% higher average specific rhamnolipid production rate reached compared to the so far best performing heterologous RL production process reported in literature (average specific production rate: 24 mg/(g h) vs. 32 mg/(g h)). However, to achieve higher titers while reducing side product formation a suitable feeding strategy and more complex temperature profiles may be required. Temperature variations in turn cause several metabolic changes, many of which are complex and interdependent. Models that describe biological processes as a function of temperature are thus essential for improved process understanding. The goal of the peer reviewed review article “Modeling and Exploiting Microbial Temperature Response”, shown in this thesis, was to present an overview of various temperature models, aid comprehension of model intent and to facilitate selection and application. Since not all metabolic interdependencies and mechanisms during temperature variation are known for the reasonable connection of input-output relationships, a suitable modeling approach seemed to be neural networks. Neural networks as black box models do not require mechanistic a priori knowledge but representative historic datasets. To collect training data, different temperature profiles or constant temperatures for a bioreactor process with P. putida KT2440 pSynpro8oT_rhlAB were applied and concentration curves for biomass, glucose and RL recorded. Subsequently, the data was fed into the neural network to compute RL titer as output. An exponential temperature profile yielded at the highest RL value of approx. 9 g (around 13 g/L) less biomass (around 12 g/L) than product. These values were reached after only 30 h consuming just 45 g of glucose. Hence, at this timepoint 36 weight-% of the consumed glucose could be assigned to mono-RL (YP/S = 0.19 gRL/gGlc) and biomass (YX/S = 0.17 gBM/gGlc. The so far best performing heterologous RL production process, yielded 23.2 g (14.9 g/L) mono-RL from >250 g of consumed glucose (YP/S = 0.10 gRL/gGlc) in >70 h using the same strain and medium but a constant temperature of 30°C.