Winner-takes-all resource competition redirects cascading cell fate transitions

Abstract

Failure of modularity remains a significant challenge for assembling synthetic gene circuits with tested modules as they often do not function as expected. Competition over shared limited gene expression resources is a crucial underlying reason. It was reported that resource competition makes two seemingly separate genes connect in a graded linear manner. Here we unveil nonlinear resource competition within synthetic gene circuits. We first build a synthetic cascading bistable switches (Syn-CBS) circuit in a single strain with two coupled self-activation modules to achieve two successive cell fate transitions. Interestingly, we find that the in vivo transition path was redirected as the activation of one switch always prevails against the other, contrary to the theoretically expected coactivation. This qualitatively different type of resource competition between the two modules follows a 'winner-takes-all' rule, where the winner is determined by the relative connection strength between the modules. To decouple the resource competition, we construct a two-strain circuit, which achieves successive activation and stable coactivation of the two switches. These results illustrate that a highly nonlinear hidden interaction between the circuit modules due to resource competition may cause counterintuitive consequences on circuit functions, which can be controlled with a division of labor strategy.

Fig. 1. Conceptual design of the synthetic cascading bistable switches (Syn-CBS) circuit.

a Diagram of the Syn-CBS circuit, in which two self-activation modules mutually activate each other. The araC self-activation in Module 1 (M1), regulated by L-ara, is designed to achieve one bistable switch. The luxR self-activation in Module 2 (M2), regulated by C6, is designed to achieve another bistable switch. b, c Phase plane analysis shows the two different expected cell fate transition paths depending on the strength of the links between the two switches. b A weak M1-to-M2 link and a strong M2-to-M1 link lead to a cell fate transition from a RFP-low/GFP-low state (black circle), to a RFP-low/GFP-high state (green circle), and then to a RFP-high/GFP-high state (yellow circle). c A strong M1-to-M2 link and a weak M2-to-M1 link lead to a cell fate transition from a RFP-low/GFP-low state (black circle), to a RFP-high/GFP-low state (red circle), and then to a RFP-high/GFP-high state (yellow circle). The nullclines of M1 and M2 are shown in green and red, respectively. The vector field of the system is represented by small arrows, where the color is proportional to the field strength. The three cell fates are indicated by filled circles at the intersections of the two nullclines.

Fig. 2. Resource competition deviates the cell fate transitions in the one-strain Syn-CBS circuit.

a The normalized steady-state signal intensity of average RFP vs. GFP measured by a plate reader shows a two-phase piecewise linear relationship. Data displayed as mean ± SD (n = 3 biological independent samples). b Flow cytometry data show cell state transitions in one-strain Syn-CBS circuit with increasing level of inducer L-ara (D_L-ara). In total, 10,000 events were recorded for each sample. Data shown from one representative of four independent biological replicates. c Diagram of the perturbed state transitions by resource competition. Dash line: expected path. Solid line: perturbed path. d Diagram of the revised model by including resource competition. e Phase plane diagrams. The nullclines of M1 and M2 are shown in green and red, respectively. The vector field of the system is represented by small arrows, where the color is proportional to the field strength. The three cell fates are indicated by filled circles (black, red, and green) at the intersections of the two nullclines. f Calculated potential landscape. Circuit CT61 was used here. Source data are provided as a Source Data file.

Fig. 3. Resource competition between two separate bistable switches.

a Diagram of the two separate bistable switches (Syn-SBS). b Flow cytometry data show cell state transitions with an increasing level of inducer C6 (D_C6) and a fixed dose of L-ara (D_L-ara = 9.5 × 10⁻⁴%). In total, 10,000 events were recorded for each sample. The two inducers were both added at 0 h. Data from one representative of three independent biological replicates. c Cell fates in the space of two inducers L-ara and C6 at addition time 0 h. d Simulated stochastic trajectories highlighted on the phase plane diagram. The nullclines of M1 and M2 are shown in green and red, respectively, while separatrices are shown in pink. The vector field of the system is represented by small arrows, where the color is proportional to the field strength. The three cell fates (red, green, and yellow circles) are found at the intersections of the two nullclines. Two representative single-cell stochastic trajectories (yellow and red highlights) show the evolution of the system from the same initial condition (purple circle, D_L-ara = 0%, and D_C6 = 0 M) to two different states with the same induction (D_L-ara = 9.5 × 10⁻⁴% and D_C6 = 5 × 10⁻⁸ M). e Calculated potential landscape. Circuit IC15 was used here. Source data are provided as a Source Data file.

Fig. 4. Relative strength of module connections determines the winner of resource competition.

a Diagram of the hybrid Syn-CBS circuit with a tetR module for fine-tuning the connection between two bistable switch modules. A hybrid promoter Para/tet is used for controlling the production of C6 to tune the M1-to-M2 connection. b Flow cytometry data showed cell state transitions with various doses of inducer aTc (D_aTc) and a fixed dose of L-ara (D_L-ara). In total, 10,000 events were recorded for each sample. Data from one representative of five independent biological replicates. Circuit CT81 was used here. Source data are provided as a Source Data file.

Fig. 5. Minimize resource competition through a division of labor using microbial consortia.

a Diagram of two-strain Syn-CBS circuits without a tetR module. b Flow cytometry data show the expected stepwise cell state transitions by increasing the dose of inducer L-ara (D_L-ara). In total, 10,000 events were recorded for each sample. Data from one representative of three independent biological replicates. Circuits CT66 and CT67 were used here. Source data are provided as a Source Data file.