American Society for Cell Biology
Poster: Sunday, December 2, 2007, program 107, board B39
Presented 12:00n-1:30p by Tim Andersen and Tim Otter..
Thank you for visiting our poster sessions.
Please don't hesitate to contact us if we can answer more of your questions.
Stable Oscillations via Quorum Sensing in Virtual Bacteria
Poster image (reduced size suitable for page printing)
To analyze gene regulatory and signaling mechanisms that produce synchronized rhythms in bacteria, we have designed and built “middle-out” in silico models where molecules, genes, gene regulatory networks, and pathways are represented in a cell-based computational architecture that grows a population of virtual bacteria capable of signaling one another by secreted molecules. First we devised an in silico model of a basic 3-element “repressilator” after Elowitz & Leibler (Nature 403:335), and tested several of their predictions. Although it is possible to monitor any molecule in silico without fluorescent reporters (GFP), we included a reporter gene to mimic the in vivo configuration. As predicted, repressilators with an odd number of genes (3, 5, 7, or 117) oscillated, but those with an even number (4, 6) did not. When we added low or high levels of noise to gene promotion and applied algorithms for cell growth and division, oscillations became asynchronous and varied in amplitude and period, consistent with in vivo studies. Implementing a bidirectional oscillator, where each gene represses the next gene but also promotes the previous one, imparted regularity and stability, and made the network more noise resistant. Finally, we created a computational model of a molecular clock where the cells are synchronized by Lux-R type quorum sensing (Garcia-Ojalvo et al., PNAS 101:10955), and tested its stability of synchronization by 3 methods: 1) temporary knockouts of one repressor gene in each cell; 2) adding noise to gene promotion; and 3) varying decay rates of repressilator proteins. In all three cases, cells became highly synchronized after quorum sensing was activated. These results demonstrate that computational models can simulate in vivo rhythmic behavior, test predictions of mathematical models or wet-bench studies, and augment the hypothesis generation and validation cycle of research on natural and engineered gene networks.
CDR is presenting three other posters as ASCB: