Specification and Simulation of Synthetic Multicelled Behaviors Seunghee S. Jang, Kevin T. Oishi, Robert G. Egbert, and Eric Klavins* Department of Electrical Engineering, University of Washington, Seattle, Washington 98195, United States * S Supporting Information ABSTRACT: Recent advances in the design and construction of synthetic multicelled systems in E. coli and S. cerevisiae suggest that it may be possible to implement sophisticated distributed algorithms with these relatively simple organisms. However, existing design frameworks for synthetic biology do not account for the unique morphologies of growing microcolonies, the interaction of gene circuits with the spatial diffusion of molecular signals, or the relationship between multicelled systems and parallel algorithms. Here, we introduce a framework for the specification and simulation of multicelled behaviors that combines a simple simulation of microcolony growth and molecular signaling with a new specification language called gro. The framework allows the researcher to explore the collective behaviors induced by high level descriptions of individual cell behaviors. We describe example specifications of previously published systems and introduce two novel specifications: microcolony edge detection and programmed microcolony morphogenesis. Finally, we illustrate through example how specifications written in gro can be refined to include increasing levels of detail about their bimolecular implementations. KEYWORDS: multicelled behavior, pattern formation, bacterial growth, specification T he ability to engineer complex, synthetic, multicelled systems could revolutionize tissue engineering, biomass production, biosensing, and biodetection. Recent advances in this area involve combining small genetic networks in E. coli or S. cerevisiae with cell-to-cell signaling to produce synthetic coupled oscillators, 1 multicell logic, 2,3 pattern formation, 4-6 and population control. 7 Although these examples are simple compared to the multicelled behaviors found in nature, they do begin to explore some of the basic mechanisms that synthetic biologists must harness to make further progress: environ- mental response, 8 signaling, 9 genetic network design, 10 and control of growth and apoptosis. 7 To advance the field, more work is needed in understanding and repurposing basic molecular and cellular mechanisms, and the algorithmic foundations of multicelled systems need to be developed to guide how new mechanisms can be used, in what kinds of algorithms, and with what limitations. Design tools developed for synthetic biology are generally focused on the behavior of single genetic circuits. They model systems using chemical reactions, enzyme kinetics, and gene network models. 11-13 The interactions induced by signaling and their effects on geometry have been examined in, for example, spatial simulations of cell signaling, 14 3D cell growth, 15 cell networks, 16 and morphogenesis, 17,18 which combine reaction diffusion models 19 with geometrical models of cell growth and division. Typically such simulations are geared toward modeling existing systems and not toward designing new ones. Separately, research in computer networks, distributed systems, 20 multirobot systems, 21 and stochastic self- organization 22 has become quite advanced. In particular, the theoretical foundations for distributed systems 20 and parallel algorithms 23 have enabled new algorithms and also described the inherent limitations 24,25 of distributed computation. Most of the approaches in the distributed systems literature, however, rely on various assumptions that do not hold in biological, multicelled systems: that nodes have unique identifiers, that they can send and receive arbitrary kinds of data, that they can store data in buffers, and so on. Even work in silico on pattern formation that uses diffusing signals 26 is difficult to imagine being implemented biologically. Here we introduce a new tool, called gro, intended to assist in the specification, design, and exploration of ideas for multicelled behaviors in a 2D environment where the spatial effects of cell growth, cell crowding, and signal diffusion dominate. gro is an open source software package that combines a distributed systems and parallel computing approach with the simulation of up to a few thousand bacterial cells growing in a 2D environment. The simulation component is focused on E. coli-like bacterial microcolonies growing in a single layer as would be viewed with a fluorescence micro- scope. 27 Although only around 10 generations can be grown in this setting before crowding overwhelms the system, we propose that controlling what happens in the initial stages of microcolony formation is an important engineering challenge that could lead to mechanisms for producing synthetic multicelled systems. Our lab and several others are currently Special Issue: Bio-Design Automation Received: April 16, 2012 Published: July 23, 2012 Research Article pubs.acs.org/synthbio © 2012 American Chemical Society 365 dx.doi.org/10.1021/sb300034m | ACS Synth. Biol. 2012, 1, 365-374