Pitfalls in optical on-line monitoring for high-throughput screening of microbial systems

New high-throughput screening systems for microbial systems, e.g. the BioLector technology, are simple to handle and offer various options of optical online measurements. The parallelization and small scale in microtiter plates allow economical high throughput and, hence, to screen many parameters in reasonable time. Fluorescent proteins as fluorescent tags made the tracking of cellular proteins in-vivo a routine task. All these tools significantly contribute to the understanding of bioprocesses. But, there are some pitfalls which might mislead the user of such techniques.

The here reported phenomena refer to typical experiments in biotechnological labs. For this reason these aspects are highlighted in this work to make operators of such valuable techniques as the BioLector aware for potential pitfalls and resulting misinterpretations. With the right approach it is possible to minimize existing problems and deal with them.

In this work the bacterium E. coli and the yeast K. lactis expressing the recombinant fluorescent proteins GFP, YFP, FbFP and mCherry were investigated. Cultivations were performed applying special microtiter plates with optodes for dissolved oxygen tension (DOT) and pH measurement in the BioLector system. In this way, microbial growth, protein formation, DOT and pH were monitored on-line via optical signals. During these studies it became obvious that fluorescent proteins can interfere with the optical signals leading to incorrect results. In this work these effects are characterized in detail and possibilities are presented how such adverse effects can be corrected or minimized by mathematical procedures or modification of the measuring method. Additionally, it is shown that morphological changes of cells can affect the biomass on-line monitoring via scattered light. Conclusions: The here reported phenomena refer to typical experiments in biotechnological labs. For this reason these aspects are highlighted in this work to make operators of such valuable techniques as the BioLector aware for potential pitfalls and resulting misinterpretations. With the right approach it is possible to minimize existing problems and deal with them.