Single cell measurements in microfluidic chip to determine the dynamics of transcription under induction

Single cell experiments require a system that is capable of collecting signals on the scale of a cell. Merging a microfluidic system with a microscopic setup opens up the possibility to collect data on the single cell level in a controlled manner. A microfluidic chip can both separate and trap cells with the right design. The so called “cell traps” are perfectly suited to observe the organism Saccharomyces cerevisiae over long time periods. This work contributes to the design of traps and their fabrication process. In order to reduce the complexity, two layers of the layout are combined into one. The overall orientation of the chip is reduced in size to fit multiple copies onto one device. In a first drawing, the total number of cell chambers could be increased from one to five, in a second iteration even further to twelve. Furthermore, different designs are used for different purposes. One design enables the simultaneous measurement of up to four yeast strains with the same environment and conditions. Another design allows to image again four chambers, but with different media compositions. Both designs were planned to reveal variations between the strains and conditions used. In addition, the structure of the cell trap was optimized for the fabrication devices, as those could not be generated with the same quality. A big oval shape seemed to solve the problems as it has no sharp angles and a big surface to be washed out. All other shapes of traps could only be produced with the use of a laser direct writer. Thereby, differences in the catch rate could be observed. The L-formed traps had the highest catch rate with 90%. The drawback is the limited number of traps that can be placed in close proximity. Smaller trap designs could catch higher numbers of cells as they were more densely packed. In two collaborations both chip designs could be used successfully. The characteriza- tion of a genetic logic gate based on two aptamers was performed with the same yeast strain under four different conditions: (I) positive control with inducer of gene expres- sion, (II) exposure of cells with neomycin after 2 h of induction, (III) cells were exposed to tetracycline and (IV) cells had both ligands in the media. The increase in GFP signal generated from cells with at least one repressor was stopped after 70 min. The second collaboration was about a light-sensitive potassium channel. Here, different strains with alternative DNA sequences for the GFP-tagged channel have been compared. The chip with four chambers for parallel experiment was best suited to keep the media as well as the light conditions identical. Interestingly, the non-codon optimized strain showed the best signal. The temporal development of the relative signal was in all strains the same. In the microscope images the localization of the channel in the ER membrane was visible. Transcription of mRNA is a stochastic process. So called bursts give rise to the mRNA distribution in the cell. In the case of stimulus dependent transcription, the parameters of the bursting are subject of change. This work investigates the question, which parameters change based on the intensity of activation. The genetic background of the yeast cell line to perform measurements on the tran- scription dynamics consists of two parts: a PP7-based system to tag mRNA directly during synthesis and the GEV transcription factor to respond to β-estradiol. The coat protein of PP7 recognizes a stem loop in the mRNA and strongly binds this motif. Mul- tiple stem loops in repetition lead to an accumulation of PP7-GFP on the mRNA, which is visible as a spot in the nucleus. The duration and intensity are directly related to the transcription dynamics. The GEV transcription factor consists of a Gal4 DNA binding do- main, an estrogen receptor and a viral activation domain. GEV relocates into the nucleus during induction and activates gene expression of Gal1 and Gal10, when β-estradiol is present. As expected, the number of cells that respond to the input rose with higher concentrations of the inducer. Interestingly, the number of responders did not reach a plateau, when induced with 500 nM over 4 h, but continuously increased over time. This finding might be explained by an increased re-initiation rate of polymerases. As GEV gets accumulated in the nucleus, it is more likely to observe an actively transcribing cell. The number of mRNA synthesized is not correlated with the induction dose. Once tran- scription is started, the cell produces a number of mRNA molecules following the same distribution for any β-estradiol concentration. In addition, a GC-rich sequence should give knowledge about the influence of DNA template properties on the elongation rate. As the GC pair forms a stronger bond, it can slow down polymerases as indicated by literature. During the design of a suitable fragment, the melting temperature was chosen to be constant for a window of 14 base pairs. Although the cloning of such a construct was successful, the integration into the yeast genome was not. Furthermore, the plasmid had repeats of another coat protein, MS2, in the 3’ UTR. Such a dual tagging strategy would have simplified the determination of the elongation speed. In conclusion, the optimizations of the chip design reduced the waste and increased the number of usable chips after one fabrication run. All tested traps were able to catch and keep cells for longer time periods. Different designs had different advantages and disadvantages. The chips could be utilized in three projects. The main results were: the neomycin-tetracycline gate responds to both repressors; the light-inducible potassium channel relocates into the ER membrane and the number of responding cells increases with higher concentration of β-estradiol