Engineering cell sensing and responses using a GPCR-coupled CRISPR-Cas system

Abstract

G-protein-coupled receptors (GPCRs) are the largest and most diverse group of membrane receptors in eukaryotes and detect a wide array of cues in the human body. Here we describe a molecular device that couples CRISPR-dCas9 genome regulation to diverse natural and synthetic extracellular signals via GPCRs. We generate alternative architectures for fusing CRISPR to GPCRs utilizing the previously reported design, Tango, and our design, ChaCha. Mathematical modeling suggests that for the CRISPR ChaCha design, multiple dCas9 molecules can be released across the lifetime of a GPCR. The CRISPR ChaCha is dose-dependent, reversible, and can activate multiple endogenous genes simultaneously in response to extracellular ligands. We adopt the design to diverse GPCRs that sense a broad spectrum of ligands, including synthetic compounds, chemokines, mitogens, fatty acids, and hormones. This toolkit of CRISPR-coupled GPCRs provides a modular platform for rewiring diverse ligand sensing to targeted genome regulation for engineering cellular functions.

Fig. 1

The GPCR ChaCha design outperforms the Tango design. a Two design schemes of coupling CRISPR-Cas9 function to the activity of GPCR. Left: the Tango design fused the effector protein (dCas9-VPR as shown) to the C-terminus of GPCR via a V2 tail sequence and a TEV cleavage sequence (TCS). An adaptor protein, Beta-Arrestin 2 (ARRB2), is fused to the Tobacco Etch Virus protease (TEVp). Right: in the ChaCha design, we fused the effector to the adaptor protein via the TCS, and fused the TEVp to the receptor via the V2 tail. Upon ligand binding to the receptor, both systems recruit the adaptor to the V2 tail, and the protease specifically cleaves at the TCS, releasing the effector protein that translocates into the nucleus for sgRNA-directed gene regulation. b Comparison of the performance of Tango (orange) and ChaCha (blue) designs using the synthetically evolved GPCR, hM3D in HEK293T cells with or without 20 μM clozapine-N-oxide (CNO), the ligand for hM3D (see Methods section for detailed experimental procedure). The free dCas9-VPR (gray) with and without a targeting sgRNA (sgTet) are used as positive and negative controls. The data were normalized to the free dCas9-VPR without a targeting sgRNA. The data represent two independent experiments with 4–8 technical replicates, and the bars represent the mean. c Estimation of the number of dCas9 molecules (n) released per receptor in the ChaCha design. Left, schematic of a stable cell line containing hM3D-CRISPR ChaCha with Doxycycline (Dox), inducing the expression of the ARRB2-TCS-dCas9-VPR and CNO inducing the receptor activity. Middle, GFP activation levels after 3 days of Dox induction and CNO treatment. Right, GFP activation levels based on the rate-model fit at steady state (r = 0.95); n is a measure of dCas9 molecules released per receptor (see Methods section for the detailed model)

Fig. 2

Characterization of kinetics and dose response of the CRISPR ChaCha system. a Time-lapse imaging of stable HEK293T cells containing hM3D-CRISPR ChaCha (see Supplementary Fig. 1 for plasmid architecture) over 48 h with or without CNO treatment. Scale bar, 50 μm. b The dose–response curve of hM3D-CRISPR ChaCha in HEK293T cells after 1-day treatment of different CNO concentrations. EC₅₀, the effective ligand concentration to achieve half-maximal GFP induction, and is shown as mean ± standard deviation of three technical replicates. The biological replicate experimental data is shown in Supplementary Fig. 5. c Reversibility of the hM3D-CRISPR ChaCha system. 10 μM CNO was added to cells for 1 day (shaded magenta area) and removed after day 1 (magenta curve). As a positive control, cells were grown with CNO for 7 days (black). Each data point represents the mean of GFP fluorescence of six technical replicates from two independent experiments, and the error bars represent the standard error of the mean. See Methods section for detailed experimental procedure

Fig. 3

The ChaCha system enables activation of individual and of multiple genes. a Induction of endogenous IL2 and IFN-γ expression and secretion using the hM3D-CRISPR ChaCha system measured by ELISA. Gray, activation using free dCas9-VPR with a targeting (sgIL2 or sjpgNG) or non-targeting sgRNA (sgGAL4). Blue, activation of IL2 or IFN-γ using a targeting sgRNA (sgIL2 for IL2, sJPEGNG for IFN-γ) with or without 10 μM CNO; orange, activation using a non-targeting sgRNA with or without 10 μM CNO. The dotted line represents the detection limit of ELISA (4 pg/mL). b Activation of HBB and HBG using the hM3D-CRISPR ChaCha system measured by quantitative PCR (qPCR) with or without 10 μM CNO. c Simultaneous activation of two endogenous genes using the hM3D-CRISPR ChaCha system. Left, the plasmids of ChaCha system used for multiplexed gene regulation. Middle, a schematic overview of the ChaCha SIMO (single input, multiple output) system. Right, ELISA measure of both endogenous IL2 (orange) and IFN-γ (blue) expression and secretion using the ChaCha SIMO. For ELISA measurements, bars are the mean of three independent experiments with 7–9 technical replicates. For qPCR measurements, bars are the mean of three independent experiments that contained three technical replicates, which are then measured in technical qPCR duplicate. See Methods section for detailed experimental procedure

Fig. 4

Expanding the ChaCha design to other synthetic and natural GPCRs. a The phylogenetic tree of Class A GPCRs. Synthetic (hM3D, KORD) and natural (LPAR, CXCR4, NMBR, ADRB2, AVPR, TRHR) GPCRs tested here are indicated. b The measured performance of diverse GPCRs with the ChaCha architecture for their respective ligands using HEK293T-GFP reporter cell line (see Supplementary Table 4 for ligand concentrations). The data were normalized to the free dCas9-VPR without a targeting sgRNA (for KORD, LPAR, CXCR4, NMBR) or with a non-targeting sgRNA (for ADRB2, AVPR, TRHR). The data represent two independent experiments with 2–3 technical replicates, and the bars represent the mean. For CXCR4, NMBR, and LPAR, we chose the best performing version characterized from a few variants as shown in Supplementary Fig. 9. See Methods section for the experimental procedure. c The dose–response curve of NMBR-CRISPR ChaCha in HEK293T cells after 1-day treatment of different CNO concentrations. EC₅₀, the effective ligand concentration to achieve half-maximal GFP induction, and is shown as mean ± s.d. of three technical replicates. The replicate experimental data is shown in Supplementary Fig. 10. d Induction of endogenous IFN-γ by NMBR-CRISPR ChaCha in HEK293T cells after 2 days of 0.5 μM NMB treatment. sgGAL4, non-targeting sgRNA; sgIFNG, IFNG-targeting sgRNA.+/− indicates with or without Neuromedin B. The fold of activation displayed on top of bars compares+/− treatment conditions. The bars represent the mean, and the data represent two independent experiments with 3 technical replicates