Multistable and dynamic CRISPRi-based synthetic circuits

Abstract

Gene expression control based on CRISPRi (clustered regularly interspaced short palindromic repeats interference) has emerged as a powerful tool for creating synthetic gene circuits, both in prokaryotes and in eukaryotes; yet, its lack of cooperativity has been pointed out as a potential obstacle for dynamic or multistable synthetic circuit construction. Here we use CRISPRi to build a synthetic oscillator ("CRISPRlator"), bistable network (toggle switch) and stripe pattern-forming incoherent feed-forward loop (IFFL). Our circuit designs, conceived to feature high predictability and orthogonality, as well as low metabolic burden and context-dependency, allow us to achieve robust circuit behaviors in Escherichia coli populations. Mathematical modeling suggests that unspecific binding in CRISPRi is essential to establish multistability. Our work demonstrates the wide applicability of CRISPRi in synthetic circuits and paves the way for future efforts towards engineering more complex synthetic networks, boosted by the advantages of CRISPR technology.

Fig. 1. Assessment of CRISPRi repression through a 2-node NOT logic circuit.

a Schematic representation of the circuit architecture, where the first node (N1) is induced by Ara and in turn produces a sgRNA (sg) that represses the second node (N2). b Details of the circuit design. In addition, AraC, dCas9, and Csy4 are constitutively expressed. Bent arrows: promoters; squares: sgRNA binding sites; rectangles: sgRNAs; crosses: csy4 recognition sites; semicircles: RBSs; pointed rectangles: reporter genes with degradation tags; and T-s: transcriptional terminators. c Maximum fold-repression of N2 achieved by four different sgRNAs, calculated as the GFP fluorescence of a fully-induced (0.2% Ara) circuit relative to a control lacking the sgRNA. Truncated variants of sgRNA-1 and -4 lacking the four 5′ nucleotides (t4) were also assessed. Data correspond to seven biological replicates. Boxes represent quartiles Q1 (box lower limit), Q2 (i.e., median, internal line) and Q3 (box upper limit); whiskers comprise data points that are no more than 1.5× IQR (inter-quartile region, i.e., the length of the box) away from the box. d Dose-response curves for the NOT circuits operating through sgRNA-1 to -4 (including t4 variants of sgRNA-1 and -4). The response (GFP fluorescence) of all four circuits can be modulated by adjusting the input dose (Ara). The controls lacking the sgRNAs (C) showcase some level of P_BAD leakiness in the NOT circuits that results in weak repression even in the absence of Ara (0). Mean and s.d. of three biological replicates. Source data are provided as a Source Data file.

Fig. 2. The CRISPRi-based toggle switch (TS).

a Topology of the bistable TS and the monostable controls. b Molecular implementation. Symbols as described in Fig. 1. c Behavior of TS and control circuits over time in response to two induction schemes: Ara-nothing-AHL-nothing (left) or AHL-nothing-Ara-nothing (right). The percentage of green cells was determined by flow cytometry for four biological replicates (shown are the mean and the s.d.). d Histograms showing representative distributions of the TS cells according to their green fluorescence for the four timepoints indicated in panel c: (i) after Ara induction, (ii) after Ara followed by no inducer, (iii) after Ara-nothing-AHL, and (iv) after Ara-nothing-AHL-nothing. Black: cells below threshold for ‘HIGH’ state (gray line); green: cells in ‘HIGH’ state. Source data are provided as a Source Data file.

Fig. 3. Proposed origin of bistability in the CRISPRi TS.

a Binding configurations of two dCas9/sgRNA variants (colors). We assume a specific binding site for one variant (box) and unspecific binding during screening of PAM sequences for both variants. Arrows: binding/unbinding reactions. b Bifurcation diagram for the mathematical model with specific and unspecific dCas9/sgRNA binding (biologically constrained parameters; see Supplementary Fig. 12 and Supplementary Table 4). Solid (dashed) lines: stable (unstable) steady states; red stars: tangent bifurcations; gray: bistability region. Note that sgRNA-1 (abbreviated sg1) and sgRNA-2 (sg2) in the model represent two generic sgRNAs, and not necessarily the specific sequences labeled as sgRNA-1 and sgRNA-2 in the experimental parts of this work.

Fig. 4. Stripe-forming synthetic IFFL based on CRISPRi.

a I2 topology of the circuit design. b Molecular implementation of the synthetic CRISPRi IFFL. Symbols as described in Fig. 1. c Behavior of the network (output) as a function of inducer concentration (input). Fluorescence of the three nodes was recorded and plotted together. Mean and s.d. of three biological replicates. d Spatial pattern formation by the synthetic CRISPRi circuit. Bacteria carrying the synthetic network were plated homogeneously on agar and a paper disk soaked with Ara was placed in the center of the plate (white), creating a radial Ara gradient by diffusion. The initially identical isogenic population developed a 3-domain pattern in which each domain is characterized by differentially expressed genes. Orange, mKO2; green, sfGFP; blue, mKate2. Note that false colors were chosen for mKO2 (emission λ 565 nm) and mKate2 (emission λ 633 nm) to highlight differences between the two. e Topology of the two-IFFL system. f Double-stripe phenotype displayed by bacteria carrying the two parallel but independent CRISPRi stripe networks. Mean and s.d. of three biological replicates. g Topologies of an orthogonal double-inverter and an IFFL circuit. h Bacteria carrying the two CRISPRi circuits shown in g display both functions. Mean and s.d. of three biological replicates. Source data are provided as a Source Data file.

Fig. 5. The CRISPRlator, a CRISPRi-based oscillator.

a Topology of the ring oscillator. b Molecular implementation of the CRISPRlator. Symbols as described in Fig. 1. c Montage showing the oscillations of the three fluorescent reporters over time. Bacteria were grown in continuous exponential phase in a microfluidic device for 3 days and imaged every 10 min. Microscopy images (as those enlarged in the zoom-ins) are displayed together in a timeline montage (kymograph). d Quantification of the population-level fluorescence of ~110 cells over time. Oscillations display a period of 10–12 h. Source data are provided as a Source Data file.