Synthetic Bioengineering
The ambitious idea of engineering cells that will function as miniature factories has given rise to new fields of research, systems and synthetic biology. We combine multi-scale, stochastic simulations with experimental genetic engineering techniques to synthesize novel biological functions and phenotypes.
Synthetic biology is a forward engineering discipline, with the objective of designing novel biological circuits, devices and systems. As in any engineering discipline, modeling plays a key role in rationalizing the design of complex synthetic phenotypes. In our modeling formalism, we include all the biomolecular interactions involved in transciption, translation, regulation and induction. This results in very large networks of reactions (hundreds to thousands) that span multiple time scales. The need then emerges for development of sophisticated algorithms that simulate networks of reactions away from the thermodynamic limit.
Our theoretical method approximates fast and continuously occurring reactions as a continuous Markov process, but maintains slow or discontinuously occurring reactions in its original form, which is a jump Markov process. Because these two processes are highly coupled, their solution must proceed simultaneously. We describe the effects of the fast reactions using a chemical Langevin equation while the times of the slow reaction events are described by a system of differential Jump equations, a type of stochastic differential equation (SDE) which we have recently derived. To our knowledge, this is the first successful attempt to develop a truly multi-scale algorithm that accurately and efficiently combines stochastic-discrete with stochastic-continuous models. We have made all our algorithms available on the web with open access licenses for the community to use with the Synthetic Biology Software Suite, readily available at http://synbioss.sourceforge.net.
We can now efficiently simulate and predict the dynamics of realistic reaction networks containing many thousands of reactions and chemical species and with widely disparate timescales, such as those found in protein/gene networks. We have used these techniques in gene regulatory network discovery and design of novel gene regulatory networks. The National Science Foundation currently supports this work.
Bio-logical AND gate
Recently, we have moved our predictions in the laboratory where we are experimenting with optimized gene circuits, such as bistable switches and bio-logical AND gates. Using an integrated theoretical-experimental approach we engineered a synthetic gene regulatory circuit with robust AND gate switch functionality in E.coli. The single promoter, hybrid system comprising of two tet (T) and one lac (L) operator sites in and around PL (l-phage) promoter permitted construction of three different motifs: TLT, TTL and LTT. By manipulating regulatory topology, we successfully generated AND gate functionality of varying induction thresholds from a single biological circuit. Our detailed statistical thermodynamic and stochastic kinetic models furnished new elemental insight regarding effective lacO position on the promoter thus facilitating high fidelity biological AND gates. We are now experimenting with other biological networks, such as toggle switches and OR gates.
Comprehensive overview of the synthetic hybrid biological AND gates. Architecture (LTT: 1a, TLT: 1b, TTL: 1c), experimental outcome (LTT: 1d, TLT: 1e, TTL: 1f) and corresponding model predictions (LTT: 1g, TLT: 1h TTL: 1i) of the synthetic AND logic circuit. 1a-c, the promoter topology is varied by successively shifting lacO position from upstream of -35 hexamer to downstream of -10 hexamer respectively. The lacO and tetO represent the operator sites at which LacI and TetR proteins bind. 1d-f, Surface plots representing the mean of gfp fluorescence influenced by the grid of inducers concentrations (aTc, 0-200ng/ml; IPTG, 0-2mM) demonstrate excellent agreement between the experimental results and deterministic model predictions. Presence of both inducers controlled gfp expression in a dose-dependant manner saturating at high concentrations. Notably induction thresholds in the presence of aTc and IPTG absent appeared to be a function of lacO location. The TTL promoter showed the best transcriptional regulation and AND gate phenotype of the 3 designs.
