Engineered CRISPRa enables programmable eukaryote-like gene activation in bacteria

Abstract

Transcriptional regulation by nuclease-deficient CRISPR/Cas is a popular and valuable tool for routine control of gene expression. CRISPR interference in bacteria can be reliably achieved with high efficiencies. Yet, options for CRISPR activation (CRISPRa) remained limited in flexibility and activity because they relied on σ⁷⁰ promoters. Here we report a eukaryote-like bacterial CRISPRa system based on σ⁵⁴-dependent promoters, which supports long distance, and hence multi-input regulation with high dynamic ranges. Our CRISPRa device can activate σ⁵⁴-dependent promoters with biotechnology relevance in non-model bacteria. It also supports orthogonal gene regulation on multiple levels. Combining our CRISPRa with dxCas9 further expands flexibility in DNA targeting, and boosts dynamic ranges into regimes that enable construction of cascaded CRISPRa circuits. Application-wise, we construct a reusable scanning platform for readily optimizing metabolic pathways without library reconstructions. This eukaryote-like CRISPRa system is therefore a powerful and versatile synthetic biology tool for diverse research and industrial applications.

Fig. 1

Design and function of eukaryote-like CRISPR activation in E. coli. a The circuit design and structure of the CRISPRa device. The circuit includes a sgRNA generator driven by P_lux2 (light blue block, inducible by AHL), a dCas9 generator driven by P_tet (blue block, inducible by aTc), an activator generator driven by P_rhaB (red block, inducible by rhamnose), and a sfGFP reporter driven by P_pspA with wild-type or heterologous UAS. b The activation mechanism of a eukaryote-like long distance regulation in our CRISPRa design, based on the IHF-dependent DNA loop structure. c, d Test of necessity of the three components in CRISPRa. c Combinations of components were achieved via presence (+) or absence (ø) of genetic part in the strain. The mismatch sgRNA had a random sequence as its spacer (sgRNA-LEA3). All strains were cultured with AHL (1.6 µM), rhamnose (0.4 mm) and aTc (2.5 ng mL⁻¹). Statistical difference was determined by a two-tailed Welch’s t-test: (sgRNA cognate/mismatch) p = 0.0002, t = 74.98. d All genetic components for CRISPRa were present in the strain and combinations were achieved via presence (+) or absence (−) of inducers. Inducer concentrations: AHL (1.6 µM), rhamnose (0.4 mm) and aTc (2.5 ng mL⁻¹). Statistical difference was determined by a two-tailed Welch’s t test: (+++/−−−), p = 0.0014, t = 26.82, (+++/−+−), p = 0.0015, t = 25.98. e A PAM inserted P_pspA with wild-type UAS (referred as mut) or heterologous UAS (referred as G6) was tested against different spacers for activation in a pspF knocked out strain E. coli MC1061ΔpspF. Concentrations of inducers: aTc (2.5 ng mL⁻¹), AHL (1.6 µM). Error bars, s.d. (n = 3); a.u., arbitrary units; p value summary: ****p value < 0.0001, 0.0001 < ***p value < 0.001, 0.001 < **p value < 0.01, 0.01 < *p value < 0.05, p value ≥ 0.05: n.s. Source data are provided as a Source Data file

Fig. 2

Optimization on synthetic UAS sites and sgRNA-dependent dynamic ranges. a Spatial relationship between the activator and the core P_pspA with an artificial UAS LEB2 (Supplementary Data 1). Activator was driven by the constitutive promoter BBa_J23106. Concentrations of inducers used in the ON state: 0.4 µM AHL (sgRNA) and aTc 2.5 ng mL⁻¹ (dCas9). No inducers were added in the OFF state. b sgRNA-dependent dynamic range optimization by “Buffer Terminators” which were previously characterized. c sgRNA-dependent dynamic range optimization by mismatches on the sgRNA scaffold. d sgRNA-dependent dynamic range optimization by moving RNA aptamer to different positions. The first three designs contained a wild-type U-A pair at the “+5” site. The circuits for b–d all employed P_pspA with two synthetic UAS LEA2 (Supplementary Data 1) and the activator was driven by the constitutive promoter BBa_J23106. Concentrations of inducers used: 0.4 µM AHL (sgRNA) and aTc 2.5 ng mL⁻¹ (dCas9). Error bars, s.d. (n = 3); a.u., arbitrary units. Source data are provided as a Source Data file

Fig. 3

Functional expansion with different σ⁵⁴-dependent promoters and bEBPs. a Activation of P_pspA, P_hrpL, P_nifH, P_nifJ in E. coli. Inducer concentrations used: For P_pspA and P_hrpL activation, 1.6 µM AHL (sgRNA), 2.5 ng mL⁻¹ aTc (dCas9), 0.4 mm rhamnose (activator). For P_nifH and P_nifJ activation, 1.6 µM AHL (sgRNA), 2.5 ng mL⁻¹ aTc (dCas9), 3.2 mm rhamnose (activator). Statistical difference was determined by a two-tailed Welch’s t test: P_pspA, p = 0.0011, t = 30.69; P_hrpL, p = 0.0006, t = 38.34; P_nifH (H1), p = 0.0134, t = 8.093; P_nifJ (H7), p = 0.0536, t = 4.053; P_nifJ (H9), p = 0.0242, t = 6.137. The data of P_nifH and P_nifJ are identical to those shown in Fig. 4c. b P_nifH and P_nifJ activation in K. oxytoca. Inducer concentrations: 2.5 ng mL⁻¹ aTc (dCas9), 3.2 µM AHL (sgRNA). Statistical difference was determined by a two-tailed t test: P_nifH (H1), p < 0.0001, t = 18.85; P_nifJ (H7), p < 0.0001, t = 20.73; P_nifJ (H9), p = 0.0165, t = 3.976. c The dynamic ranges of hybrid σ⁵⁴-dependent promoter activation. The asterisk represents the P_flhDp2 was a predicted σ⁵⁴-dependent promoter. Activator expression was driven by the constitutive promoter BBa_J23106. Inducer concentrations used: 1.6 µM AHL (sgRNA), 2.5 ng mL⁻¹ aTc (dCas9). d Orthogonal activators in E. coli TOP10. Left: Truncated NorRs were fused with a λN22plus adaptor. Right: four activators without the HTH domain truncations were fused with a λN22plus adaptor and tested. All engineered activators were driven by the promoter BBa_J23106. P_pspA-20 was used for the reporter, which could be activated by sgRNA-LEB2. Inducer concentrations used: 0.025 µM AHL (sgRNA), 2.5 ng mL⁻¹ aTc (dCas9) for ON states. Statistical difference was determined by a two-tailed t test: NorR (179-504), p < 0.0001, t = 27.54; NorR (188–504), p = 0.0007, t = 9.456; WtsA, p < 0.0001, t = 24.42. Error bars, s.d. (n = 3); a.u., arbitrary units; p value summary: ****p value < 0.0001, 0.0001 < ***p value < 0.001, 0.001 < **p value < 0.01, 0.01 < *p value < 0.05, p value ≥ 0.05: n.s. Source data are provided as a Source Data file

Fig. 4

Orthogonality tests of the CRISPRa system. a Orthogonality owing to UAS-sgRNA specificity. The spacer length for LEA site is 20 bp, and the spacer length for LEB site is 17 bp. Activator expression was driven by the constitutive promoter BBa_J23106. Inducer concentrations used: 1.6 µM AHL (sgRNA), 2.5 ng mL⁻¹ aTc (dCas9). b Orthogonality between different RNA aptamer-adaptor pairs. Activators expressions were driven by the promoter BBa_J23106. Inducer concentrations used: 0.08 mm arabinose (sgRNA), 2.5 ng mL⁻¹ aTc (dCas9). c Orthogonal UAS for P_nifH and P_nifJ. Nine selected UAS between the two promoters were tested in activating P_nifH and P_nifJ, respectively (bottom left). Inducer concentrations used: 1.6 µM AHL (sgRNA), 2.5 ng mL⁻¹ aTc (dCas9), 3.2 mm rhamnose (activator). Error bars, s.d. (n = 3); a.u., arbitrary units. Source data are provided as a Source Data file

Fig. 5

CRISPRa using dxCas9 and non-canonical PAM and its lowered background expression. a Three different NGT PAM were tested by CRISPRa with dCas9 or dxCas9. The mutated PAM was introduced into the P_pspA-LEA3B3, which was targeted by sgRNA-LEA3 in this experiment. Activator expression was driven by the constitutive promoter BBa_J23106. Inducer concentrations used: 1.6 µM AHL (sgRNA), 1.25 ng mL⁻¹ aTc (dCas9/dxCas9). b Two spacers with TGT PAM (A1, A2) within the enhancer region of wild-type P_pspA were selected as CRISPRa target UAS. Inducer concentrations used: 1.6 µM AHL (sgRNA), 2.5 ng mL⁻¹ aTc (dCas9/dxCas9), 0.4 mm rhamnose (activator). c The dynamic range optimization effects of dxCas9 on a synthetic promoter P_pspA-LEA1B1. Activator expression was driven by the constitutive promoter BBa_J23106. Inducer concentrations: 1.6 µM AHL (sgRNA), 1.25 ng mL⁻¹ aTc (dCas9/dxCas9). These data are also shown in Supplementary Fig. 12a. d Two-layer CRISPRa using dCas9 or dxCas9. P_pspA-LEA3B3 was used as the second promoter for the transcription of the sgRNA-LEB1, which activates the third promoter P_pspA-LEA1B1. e A positive feedback loop was constructed using dxCas9 and was compared to a basic activation without feedback. Inducer concentrations used for d and e: 0, 0.006, 0.025, 0.100, 0.400 µM AHL (sgRNA), 1.25 ng mL⁻¹ aTc (dCas9/dxCas9), 0.2 mm rhamnose (activator). The data of basic activation by CRISPRa is also shown in Supplementary Fig. 14. Error bars, s.d. (n = 3); a.u., arbitrary units. Source data are provided as a Source Data file

Fig. 6

An expression profiles scanning tool for metabolic pathway. a Three genes from the violacein production pathway (vioA/vioD/vioC) were chosen as target genes for expression tuning through CRISPRa. Each sgRNA that targeted its cognate promoter could be driven by one of the three constitutive promoters with different strengths and the resulting sgRNA generators were mixed to give a multi-sgRNA generator library. b The refactored violacein pathway was co-transformed with the library and cultured on agar plates for 16 hr with 1 µM rhamnose for activator induction and 0.63 ng mL⁻¹ aTc for dxCas9 induction (top middle). For negative control (top left), an empty vector was used in place of the sgRNA generator library. For positive control (top right), the strong constitutive promoter BBa_J23100 was used to drive expression of all sgRNA. Colonies with visibly different color intensities were marked by triangles according to their levels of purple color: dark purple (magenta triangles), weak purple (yellow triangles) and white (green triangles). Sequencing results revealed the transcription patterns of sgRNAs in each colony. The number of purple bricks correlates with strengths of the constitutive promoters. Images of the same plates are showed in Supplementary Fig. 15. c Pathway optimization by combining CRISPRa and CRISPRi. A new library was constructed by mixing CRISPRi sgRNA (they targeted the coding regions of the three genes and have no RNA aptamers) under a strong promoter (BBa_J23100) and CRISPRa sgRNA under BBa_J23100 or a weak promoter (BBa_J23114). The library was used for profile scanning and results were obtained after 20 hr of culture on agar plates. The images of the same plates are showed in Supplementary Fig. 16. d sgRNA transcription profiles were pooled and analyzed. For each regulated gene, the type of sgRNA and their promoter strengths were plotted against the categorized bins of violacein production. Each dot represents a connection that maps the violacein production strength to the sgRNA promoter strength for a given target gene (vioA/vioD/vioC). Source data are provided as a Source Data file

Fig. 7

The reusability of the multi-gRNA generator library, and the stability of expression profiles. a The multi-gRNA generator library (with sgRNA for CRISPRa and CRISPRi) we used for violacein pathway could be used on other target circuits. The CRISPRa device can project the transcription profiles of each multi-gRNA generators to target circuits and generate a variety of gene expression pattern. Three different fluorescent protein genes were used as reporters in this experiment. The genes sfGFP, mCherry and mTagBFP2 were controlled by the same artificial σ⁵⁴ promoters for vioA, vioD, vioC respectively. Each gene may be activated by CRISPRa to different extents: (1) No activation (the sgRNA designed for CRISPRi on violacein genes was effectively a mismatch sgRNA), (2) moderate activation, and (3) strong activation. b Absolute fluorescence/A₆₀₀ values of the three reporters detected from each strain. Left: values from BFP channel; Middle: GFP; Right, RFP. For high-induction condition, 1.25 ng mL⁻¹ aTc (dxCas9), 0.4 mm rhamnose (activator) were used. For low induction condition, 1.25 ng mL⁻¹ aTc, 0.1 mm rhamnose were used. 1.25 ng mL⁻¹ aTc, 0 mm rhamnose were used for OFF state. The PC strain was a positive control strain that carried a multi-gRNA generator with strong promoter BBa_J23100 for all sgRNA for CRISPRa. c PC strain-normalized fluorescence/A₆₀₀ values, which shows the relative proportions of the three fluorescent proteins in each strain under different induction conditions. Legend for colored circles beneath the data bars: dark color, the sgRNA transcription is driven by strong promoter; light color, driven by weak promoter; white color, mismatch sgRNA. Information for colored circles come from the sequencing results of each strain. d Serial transfers (100-fold every 6 hr) and induced growth up to 24 hr. 1.25 ng mL⁻¹ aTc (dxCas9), 0.4 mm rhamnose (activator) were used in this test. The bar chart shows PC strain-normalized fluorescence/A₆₀₀ values. Error bars, s.d. (n = 3); a.u., arbitrary units. Source data are provided as a Source Data file