Our system is a cheating co-culture. C. hutchinsonii spends its own resources to break down cellulose into simple sugars, and E. coli, the cheater, consumes the sugars for free. Traditionally, cheaters become predominant in a multi-species community because they have a higher level of competitive fitness.
The Challenge
Our naive co-culture experiment showed that C. hutchinsonii actually dominates the community. We attributed this to C. hutchinsonii’s efficiency in taking up the sugars it produces, since these sugars need to diffuse into solution to become available to E. coli. The additional metabolic burden on E. coli when it has to produce biofuel would only exacerbate the high proportion of C. hutchinsonii, but since we hoped to create an efficient system for consolidated bioprocessing, we wanted more E. coli to generate more product. Our goal was to create a circuit that would stabilize populations with a larger proportion of E. coli in the co-culture.
The Circuit

RelE, a suicide gene, is constitutively expressed in C. hutchinsonii. At high E. coli levels, LuxI is expressed to produce 3OC6HSL, which diffuses in the solution and activates the antidote gene RelB, rescuing C. hutchinsonii from the RelE. Therefore, C. hutchinsonii relies on a certain concentration of E. coli to survive.

Since E. coli’s dependence on C. hutchinsonii creates a natural one-way metabolic link between the two species, we designed a circuit to create dependency in the other direction - in other words, we incentivized C. hutchinsonii to allow more E. coli to grow. In our system, C. hutchinsonii and E. coli constitutively express RelE (a suicide gene) and LuxI (which produces signal molecule 3OC6HSL) respectively. At high enough E. coli levels, diffused 3OC6HSL activates the antidote gene RelB in C. hutchinsonii, rescuing it from the RelE. Therefore, C. hutchinsonii requires a certain concentration of E. coli in order to survive. To adjust the population ratios, we would tune the strength of the RBS and promoter of RelE, LuxR and LuxI.
Engineered Population Control Systems
Previous approaches to controlling co-cultured population range from growth-controlling genetic circuits based on quorum sensing compounds (You et al. 2004) to complementary auxotrophic amino acid exchange (Shou et al. 2007). As proof-of-concept, these studies are quite elegant and encouraging but may suffer on the large scale due to unstable and difficult genetic modifications in industrially relevant organisms.

In our system, E. coli and C. hutchinsonii must release signals necessary to the other's survival in a pathway orthogonal to their linked metabolisms. Such interactions have been engineered before as artificial predator-prey systems (Balagadde et al. 2008) and through population-controlled regulated killing (Lingchong et al. 2004; You et al. (2004).

Our general circuit design was inspired by examples of these published systems. We chose to use the Lux system, originally isolated from Vibrio fischeri, since it is one of the most widely used AHL signaling systems. We chose RelE/RelB as our toxin-antitoxin system because other systems are more specific to E. coli and less likely to be effective in C. hutchinsonii. While the Lux and RelE/RelB systems have been demonstrated to function in E. coli, the novelty in our work lies in how we combined them and designed the system for expression in C. hutchinsonii.
Failure Modes
Due to time constraints we had to keep our circuit simple, and so there are some modes in which the circuit would produce undesirable behaviour. If the concentration of E. coli begins at or drops below a certain threshold, a positive feedback loop will be started in which C. hutchinsonii begins to die from its constitutively expressed RelE. As a result, E. coli will have less sugar and the concentration will drop even more, leading to more C. hutchinsonii dying and so on. Our circuit requires a high starting concentration of E. coli or exogenous addition of the signalling molecule to prevent C. hutchinsonii from immediately dying.
Balagaddé et al., Mol Syst Biol, 4:187 (2008)

Lingchong et al., Nature, 428:868-871, (2004)

Shou et al., Proc Natl Acad Sci USA, 104: 1877–1882 (2007)

You et al., Nature, 428(6985):868-71 (2004)