RESEARCH Open Access Designing and engineering evolutionary robust genetic circuits Sean C Sleight*, Bryan A Bartley, Jane A Lieviant, Herbert M Sauro Abstract Background: One problem with engineered genetic circuits in synthetic microbes is their stability over evolutionary time in the absence of selective pressure. Since design of a selective environment for maintaining function of a circuit will be unique to every circuit, general design principles are needed for engineering evolutionary robust circuits that permit the long-term study or applied use of synthetic circuits. Results: We first measured the stability of two BioBrick-assembled genetic circuits propagated in Escherichia coli over multiple generations and the mutations that caused their loss-of-function. The first circuit, T9002, loses function in less than 20 generations and the mutation that repeatedly causes its loss-of-function is a deletion between two homologous transcriptional terminators. To measure the effect between transcriptional terminator homology levels and evolutionary stability, we re-engineered six versions of T9002 with a different transcriptional terminator at the end of the circuit. When there is no homology between terminators, the evolutionary half-life of this circuit is significantly improved over 2-fold and is independent of the expression level. Removing homology between terminators and decreasing expression level 4-fold increases the evolutionary half-life over 17-fold. The second circuit, I7101, loses function in less than 50 generations due to a deletion between repeated operator sequences in the promoter. This circuit was re-engineered with different promoters from a promoter library and using a kanamycin resistance gene (kanR) within the circuit to put a selective pressure on the promoter. The evolutionary stability dynamics and loss-of-function mutations in all these circuits are described. We also found that on average, evolutionary half-life exponentially decreases with increasing expression levels. Conclusions: A wide variety of loss-of-function mutations are observed in BioBrick-assembled genetic circuits including point mutations, small insertions and deletions, large deletions, and insertion sequence (IS) element insertions that often occur in the scar sequence between parts. Promoter mutations are selected for more than any other biological part. Genetic circuits can be re-engineered to be more evolutionary robust with a few simple design principles: high expression of genetic circuits comes with the cost of low evolutionary stability, avoid repeated sequences, and the use of inducible promoters increases stability. Inclusion of an antibiotic resistance gene within the circuit does not ensure evolutionary stability. Background Synthetic biology is the design and engineering of new biological functions and systems that do not occur in nature. This relatively new field has provided insight into the mechanisms of natural gene networks [1,2] and engineered multicellular pattern formations [3], bacterial photography [4], tumor-targeting bacteria [5], feed-for- ward network based concentration sensors [6], robust and tunable oscillators [7], and genetic networks that count [8]. On the genome level, entire metabolic path- ways have been engineered to overproduce an anti- malaria compound [9], biofuels from plant biomass [10,11] lycopene through automated genome engineer- ing and accelerated evolution [12], and a synthetic chro- mosome [13] transplanted into a host bacterium [14]. Despite recent efforts of engineering at the genome level, most synthetic biology constructs are engineered at the level of genetic circuits encoded on plasmids. Genetic circuits are built bottom-up from biological parts. A biological part is a DNA sequence that encodes a basic biological function [15]. Examples of parts * Correspondence: sleight@u.washington.edu Department of Bioengineering, University of Washington, Seattle, WA 98195, USA Sleight et al. Journal of Biological Engineering 2010, 4:12 http://www.jbioleng.org/content/4/1/12 �� 2010 Sleight et al licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

include promoters, ribosome binding sites (RBS), protein or RNA coding regions, and transcriptional terminators. Biological engineers can assemble individual parts or combination of parts together using a BioBrick assembly standard for physical composition [16] (described in [17]). Parts that conform to this BioBrick assembly stan- dard are BioBrick standard biological parts, or BioBricks. Standard Assembly involves digestion of two BioBricks encoded on plasmids with different restriction enzymes that leave compatible sticky ends which can be ligated together into a new BioBrick. This assembly method effectively replaces the restriction sites between the assembled parts with a ���scar��� sequence, allowing for the new BioBrick to be later combined with other BioBricks. Alternative assembly strategies have recently been pro- posed [18,19] to improve upon the original assembly standard. The MIT Registry of Standard Biological Parts (called ���The Registry��� from here on) maintains over 3000 BioBricks encoded on plasmids that are available to researchers with a wide variety of different functions, from bacterial photography, to quorum sensing to odor production and sensing. BioBricks are widely available for the design of more complex systems, but in general are not well-character- ized [15,17]. The most well-characterized part to date is a cell-cell communication receiver device [17], which was provided with a published prototype ���biological part datasheet��� containing information engineers would need to use it in their own designs. One of the figures in this datasheet describes the reliability of this circuit over evolutionary time. Connecting the receiver device to a GFP-reporter device causes this circuit to repeatedly lose function in less than 100 generations due to a dele- tion mutation between transcriptional terminators that are repeated in both the receiver and reporter devices. Another example of genetic circuits losing function over evolutionary time is illustrated by studies of microche- mostat-evolved strains containing a cell density regula- tion circuit that loses function in less than 100 hours [20,21]. The evolutionary stability of whole circuits is therefore an emergent property of the context of its bio- logical parts. Evolutionary stability is a problem in genetic circuits if there is no selective pressure to maintain function of the circuit. The current belief is that this loss-of-function occurs because any cell in the population that acquires a mutation in the genetic circuit often has a growth advan- tage and can outcompete the cells in the population with all functional plasmids. As the cells divide, any cell with a larger percentage of mutant plasmids will eventually domi- nate the population until only cells with mutant plasmids remain. A simulation study predicted that the time for a non-functional mutant of a synthetic microbe to become the majority of the population is a function of the growth rate difference between the mutant and functional cells, circuit size, circuit architecture, and mutation rate [22]. Non-functional mutants often have a growth advantage because a mutation that inactivates a genetic circuit can reduce its metabolic load. The magnitude of metabolic load caused by expression and replication of foreign genes is dependent on many factors such as plasmid size, plas- mid copy number, the foreign gene being expressed, anti- biotic resistance gene, metabolic state of the cell, growth media, and amount of dissolved oxygen in the media [23]. Dekel and Alon [24] directly measured the cost associated with expression and maintenance of Lac proteins when they provided no fitness benefit and found mutations that alleviated this cost in the non-selective environment. There are also examples of chromosomal genes that have lost function over evolutionary time when not under selec- tion [25,26] and so encoding synthetic circuits into the chromosome will only delay this problem. The evolutionary stability of genetic circuits within synthetic microbes will be an increasingly significant issue as these circuits become more complex and need to be functional over longer periods of time. The ability to engineer evolutionary robust genetic circuits will be important for applied uses of synthetic microbes that perform long-term functions in the environment and possibly in the human body. This ability will also be important for the study of genetic circuits in microche- mostats and microfluidic devices over multiple genera- tions. Ideally, a selective regime should be used to maintain circuit function over evolutionary time. How- ever, design of a selective regime for synthetic microbes is unique to the genetic circuit of interest, and design for maintaining function of a particular circuit is often difficult. Therefore, general design principles are needed for engineering evolutionary robust circuits that will maximize stability over time. As a first step towards this goal, this study aimed to understand the loss-of-function mutations that occur in two genetic circuits over evolutionary time and their evolutionary stability dynamics. Next, we re-engineered these circuits in various ways to determine the predict- ability of mutations in replicate evolved populations and whether we could make these circuits more evolutionary robust. The results from these experiments allowed us to observe the mutations in several diverse circuits, determine their evolutionary stability dynamics, and develop simple design principles for engineering evolu- tionary robust circuits. Results Loss-of-function mutations and evolutionary stability dynamics in two genetic circuits We first measured the evolutionary stability dynamics of two genetic circuits propagated in Escherichia coli Sleight et al. Journal of Biological Engineering 2010, 4:12 http://www.jbioleng.org/content/4/1/12 Page 2 of 20

MG1655 in order to determine the loss-of-function mutations that cause their instability and which circuit is the most robust over evolutionary time. High-copy plasmids were used instead of low or medium-copy plasmids to maximize selective pressure so that evolu- tion would occur more rapidly since replication and expression of genetic circuits encoded on high-copy plasmids will increase metabolic load and lower fitness. Cells with a low metabolic load (e.g., cells with mutant plasmids) have greater fitness than cells with a higher metabolic load (e.g., cells with functional plasmids) (unpublished results). Therefore, we expect that mutants will be able to rapidly outcompete functional cells that have a high expression level. However, other factors besides expression level will play a role in this evolution- ary process such as mutation rate and the metabolic load associated with plasmid replication. The two circuits we used to measure the evolutionary stability dynamics and determine the loss-of-function mutations were T9002 (Figure 1a) and I7101 (Figure 2a). T9002 is the Lux receiver circuit previously described [17] and expresses luxR that activates GFP expres- sion when the inducer AHL is added to the media (see Figure 1 legend for details). I7101 has a lacI-regulated promoter and expresses GFP only when the inducer IPTG is added to the media since lacI is constitutively overexpressed from the chromosome in this particular strain (Escherichia coli MG1655 Z1). The evolutionary stability dynamics were measured by serial propagation with a dilution factor that allows for about 10 generations per day. Figure 1b shows the evolutionary stability dynamics of the T9002 circuit propagated in high input (with AHL) and low input (without AHL) conditions. From different timepoints in the experiment, the low and high input populations were induced with AHL to measure their normalized expression (here measured by fluorescence divided by cell density) over time. The low input evolved populations slowly lose their function to about 50% of the maximum after 300 generations. The evolved popu- lations in high input conditions rapidly lose their func- tion in less than 30 generations (the dynamics of this evolutionary stability are described below in Figure 3). No functional clones were observed after 30 generations as determined by measurement of individual colonies. The mutation that repeatedly causes loss-of-function in the high input evolved populations is a deletion between two homologous transcriptional terminators (Figure 1c), the same mutation described in [17]. This mutation evi- dently occurs at such a high rate that mutants quickly take over the population. In fact, Canton et al (2008) [17] were unable to isolate a population derived from a single isolate that did not already carry the deletion. The mutant plasmid was transformed back into the progenitor and was shown not to fluoresce after induc- tion with AHL. In this initial study we also tested the evolutionary stability of a BioBrick engineered ver- sion of the repressilator circuit [1]. We could not mea- sure its function over time due to unstable GFP expression at the population level, but found that the circuit repeatedly had a deletion between homologous tetR promoters. Figure 2b shows the evolutionary dynamics of the I7101 circuit propagated in high (with IPTG) and low input (without IPTG) conditions. The evolved popula- tions in low input conditions lose about 70% of their function over 300 generations. The high input evolved populations lose about half their function in 30 genera- tions and nearly all function after 300 generations. For this circuit, the loss-of-function is repeatedly a deletion between two homologous operator sequences in the promoter (Figure 2c). The mutant plasmid was trans- formed back into the progenitor and was shown not to fluoresce when induced with IPTG. This initial study suggests that the use of repeated parts in synthetic circuits should be avoided due to the high mutation rate. Also, there is a high metabolic load associated with the expression of genetic circuits on high-copy plasmids since keeping these circuits off substantially improves evolutionary stability. Evolution experiment with re-engineered circuits Based on the results of the previous experiments, we re-engineered the T9002 and I7101 circuits to test var- ious predictions of evolutionary stability and mutational predictability. For the T9002 circuit, the loss-of-function mutation was repeatedly a deletion between two homo- logous transcriptional terminators. Mutations and genetic rearrangements can occur due to misalignment of homologous sequences during replication (termed ���replication slippage���) [27]. Deletion mutations between repeated sequences are known to be dependent upon repeat length, proximity, and homology level [28]. These deletions are recA-independent if the repeat length is less than 200 bp [27,29], as is the case with the repeated terminators in the T9002 circuit. Thus, we re-engi- neered the last terminator of T9002 with various termi- nators available in the Registry to measure the effect of terminator homology level and orientation with evolu- tionary stability. We predicted that we could increase evolutionary robustness by decreasing the mutation rate of this deletion. Furthermore, although there have been several studies on recombination between repeated sequences, this phenomenon has not been studied in the context of synthetic biology using genetic circuits constructed from BioBricks. For instance, we do not know the effect of using various BioBrick terminators with different homology levels in the same circuit. The Sleight et al. Journal of Biological Engineering 2010, 4:12 http://www.jbioleng.org/content/4/1/12 Page 3 of 20