A Rapid, Small-Scale Method for Improving Fermentation Medium Performance

摘要

Cell biomass and chemicals (e.g. bioactive compounds) can be produced by fermentation. Optimising a fermentation system involves optimizing many variables such as determining the effect of inoculum quality and media components, and selecting the most appropriate fermenter design and operating conditions (such as agitation aeration and fermentation mode). Identifying the optimal media is very important because it can significantly affect product concentration, yield and productivity. However, the media contains many components so many trials need to be done, which makes the process laborious, expensive, open-ended, and often time-consuming. The data generated from the many trials can be difficult to analyse. This study developed a rapid, inexpensive small-scale technique to identify how media components affected the growth of Streptomyces hygroscopicus and its production of a secondary metabolite, the anti-tumour agent rapamycin. A method was developed using microtitre plates to screen the effect of three concentrations of nine media components on cell growth and rapamycin production using the Box-Behnken experimental design. Firstly, the methodology for microtitre plates was developed, which involved characterizing the physical parameters of a fermentation system, identifying the incubation time to minimize evaporation, modifying the assay method to deal with the small sample volumes, and developing an alternative method to determinate the rapamycin concentration that was cheaper than the HPLC method. Data from shake flasks trials (the normal screening method) were used to validate the microtitre method and to assess the latter's usefulness in predicting scale-up effects.Six media components - sodium chloride (NaCl), di-potassium orthophosphate (K2HPO4), l-aspartic acid, l-arginine, l-histidine and salt (formula 1) solution - significantly affected culture growth and/or rapamycin concentration. The regression tree method was used to indicate the importance and critical concentration range of each factor. The Pearson's product-moment value indicated a good correlation between data from microtitre plates and shake flasks (cell growth: r=0.75 p=0.016 n=8; rapamycin concentration r=0.92 p=0.08 n=6). The speed of the microtitre plate and shake methods were compared by assessing the total cycle time and the time required for various stages in the method. Performance of each method was assessed as cost of media and equipment. Using microtitre plates to screen and optimise media in terms of biomass and secondary metabolite production is faster and cheaper than using shake flasks. Labour efficiency for the numerous, repetitive, small-scale experiments was substantially increased. Trials could be run without well-to-well cross contamination. The regression tree statistics methodology successfully showed the effect of input variables on target variables and identified effective medium component concentrations and any interactions. It is recommended that the microtitre plate procedure developed in this research may be applied to any study investigating the optimum media composition for growing other Streptomyces spp. strains, in screening studies when searching for new bioactive molecules, or for characterizing natural or recombinant/mutated micro-organisms.