Rewiring Calcium Signaling for Precise Transcriptional Reprogramming

Abstract

Tools capable of modulating gene expression in living organisms are very useful for interrogating the gene regulatory network and controlling biological processes. The catalytically inactive CRISPR/Cas9 (dCas9), when fused with repressive or activating effectors, functions as a versatile platform to reprogram gene transcription at targeted genomic loci. However, without temporal control, the application of these reprogramming tools will likely cause off-target effects and lack strict reversibility. To overcome this limitation, we report herein the development of a chemical or light-inducible transcriptional reprogramming device that combines photoswitchable genetically encoded calcium actuators with dCas9 to control gene expression. By fusing an engineered Ca²⁺-responsive NFAT fragment with dCas9 and transcriptional coactivators, we harness the power of light to achieve photoinducible transcriptional reprogramming in mammalian cells. This synthetic system (designated CaRROT) can also be used to document calcium-dependent activity in mammals after exposure to ligands or chemicals that would elicit calcium response inside cells.

Diverse cellular activities such as cell growth, cell differentiation, metabolism, and homeostasis are dictated by complex gene networks and active transcriptional regulation.^1,2 To illuminate the gene function during these biological processes, interventional tools that would enable real-time manipulation and perturbation of target gene expression are critically needed. The clustered regularly interspaced short palindromic repeats (CRISPR)-associated-9 nuclease also known as Cas9 derived from Streptococcus pyogenes has emerged recently as a powerful tool for genome engineering.^3-7 Cas9 can be guided by a single RNA (sgRNA) to a genomic target site where it is complementary to the sgRNA and juxtaposed to a protospacer adjacent motif (PAM) sequence: NGG.⁸ The CRISPR-Cas9 genome editing system only requires two major molecules to bind to a target DNA sequence, and thus have a great potential to become an RNA-dependent DNA recognition platform. Lately, the engineered catalytically inactive Cas9 (or dCas9) has been developed as a robust tool for targeted endogenous gene regulation without genetically altering the DNA sequence.^3,9-14 The dCas9-sgRNA complex can induce repression of endogenous genes in bacteria by blocking RNA polymerase or perturbing transcription factor binding.^1,10,15 Also, dCas9 fused to effector domains such as multiple tandem copies of Herpes Simplex Viral Protein 16 (VP64) or p65 activator domain (p65AD) have been used to activate reporter genes or endogenous genes both in E. coli and human cells.^1,3,16 However, without temporal control, the application of these reprogramming tools will likely cause off-target effects and lack strict reversibility.

In order to overcome this limitation, we decided to generate a synthetic chemical or light-sensitive dCas9 nuclear translocation systems. Our design combines genetically encoded photoactivatable Ca²⁺ actuators with an engineered Ca²⁺-responsive transcriptional factor and dCas9-effector fusions (Figure 1). The most effective one is using a photoswitchable Ca²⁺ actuator engineered from CRAC channel (Opto-CRAC)^17-20 to remotely control calcium signals and Ca²⁺-dependent nuclear translocation of engineered dCas9 fusions. We improved the Opto-CRAC system to reduce “leakiness” of Ca²⁺ influx, or dark-state background activity. Ca²⁺ influx induced by Opto-CRAC activates calcineurin, a Ca²⁺-dependent phosphatase, which dephosphorylates nuclear factor of activated T cells (NFAT) and subsequently leads to NFAT nuclear translocation (Figure 1).²¹ Hence, by cotransfection with the Opto-CRAC system, NFAT fragment (residues 1–460) fused with dCas9-VP64 would translocate into the nuclei upon blue light illumination. Nuclear dCas9 was further directed toward its target genes by sgRNA to turn on the reporter or endogenous gene expression. The calcium signals can be generated either with calcium channel agonists or light illumination. We demonstrate herein the use of this chemical- or light-inducible transcriptional reprogramming device (designated as CaRROT for “calcium-responsive transcriptional reprogramming tool”) to modulate gene expression with high precision.

Figure 1.

Design of genetically encoded CaRROT to enable spatiotemporal control of transcriptional reprogramming in mammals. This synthetic device is composed of (i) second-generation Opto-CRAC made of LOV2-SOAR chimeras that could photoactivate ORAI calcium channels on the plasma membrane with tight control over Ca²⁺ signals; and (ii) a calcium-responsive dCas9 fusion construct (e.g., NFAT_1–460-dCas9-VP64). The N-terminal NFAT fragment used in the design lacks the C-terminal DNA binding domain to avoid binding to endogenous NFAT targets. In the dark, CaRROT stays in the cytosol. Upon blue light illumination, CaRROT undergoes light-inducible nuclear translocation due to the cleavage of the phosphate groups (P) by calcineurin to turn on gene expression at targeted loci in the presence of small guide RNAs (sgRNAs). In addition to light, chemicals or ligands that could elicit intracellular calcium mobilization could likewise rewire calcium signaling to achieve inducible transcriptional reprogramming at targeted genomic loci.

RESULTS AND DISCUSSION

Design of Second Generation Opto-CRAC and CaRROT Constructs.

In order to regulate specific gene expression, dCas9 is required to locate in the nuclei and is directed by a sgRNA to the promoter of the targeted genes. In most applications, the nuclear localization of dCas9 is enabled by adding several nuclear localization signals (NLS) at both N and C-termini of dCas9.^3,14 Without the NLS signals, dCas9 largely resides in the cytoplasm given its relatively large size and lack of a strong NLS by itself (Supplementary Figure S1A). To enable light-controllable nuclear translocation to execute its function, we designed and constructed several photosensitive dCas9-VP64 systems based on two strategies: (i) fusion of dCas9-VP64 with light-sensitive NLS; or (ii) design of a synthetic Ca²⁺-dependent nuclear translocation device, thereafter termed as calcium-responsive transcriptional reprogramming tool (or CaRROT) (Figure 2A).

Figure 2.

Design and optimization of CaRROT and second-generation Opto-CRAC constructs to enable tight control of dCas9 nuclear translocation. (A) Design of dCas9-fUsion constructs for inducible nuclear translocation: (i) fusion with light-sensitive NLS signals (BiNLS: V1–V2); or (ii) through Ca²⁺-dependent nuclear translocation (V3–V5). (B) Opto-CRAC designed to photoinduce Ca²⁺ influx by optimizing STIM1-CT fragments, the linker and fusion to LOV2-binder Zdk. (C) Basal fluorescence intensities of GCaMP6s-HeLa cells transfected with indicated Opto-CRAC constructs in the dark. At least 30 cells were analyzed in the assay for each construct. (D) Light-inducible fold-change in the GCaMP6s fluorescence intensity (at 2 min postphotostimulation at 470 nm; 50 μW/cm²) in HeLa cells expressing the indicated second generation Opto-CRAC constructs. Data were shown as mean ± SD (n = 30 cells from three independent experiments). (E) Time course showing light-inducible increase of GCaMP6s signals in HeLa cells expressing Opto-CRAC-B10. Representative images showing GCaMP6s fluorescence before and after light stimulation were presented on the right. Data were showed as mean ± SD (n = 30 cells). (F) Monitoring light-inducible translocation of dCas9-VP64 or dCas9-NFAT_1–460-VP64 from cytosol to nuclei in the same cells expressing the indicated constructs by confocal imaging. (G,H) Time course showing the fold-change of nuclear GFP intensity following blue light stimulation (G) and quantification of signals before and after light illumination for 30 min (H). Data were showed as mean ± SD (n = 9). Scale bar: 5 μm. ****P < 0.0001 compared to the dark group (two-tailed Student’s t-test).

For the first approach, light-sensitive NLS is designed by fusing bipartite NLS peptides²² to LOV2 (AsLOV2) derived from Avena sativa phototropin.²³ Bipartite NLS was introduced to the C-terminal Jα helix of the AsLOV2 domain, while dCas9-VP64 was placed at the N-terminus. In the dark, NLS is caged by LOV2 domain and thus shielded from the nuclear import cargo; therefore, the fusion protein is trapped in the cytoplasm. Upon blue light illumination, photoexcitation creates a covalent adduct between LOV2 residue C450 and the cofactor FMN,²³ allowing the undocking of the Jα helix to expose NLS. The NLS binds to importin, which mediates interactions with the nuclear pore complex, thereby causing the translocation of dCas9-VP64 from cytosol to the nuclei. We used two different mCherry-tagged versions of bipartite NLS (NLS1 and NLS2), which has been shown to induce nuclear translocation after blue light induction.²² But this strategy did not work out as anticipated since we failed to observe light-inducible nuclear entry of dCas9 (Figure 2; constructs V1–V2).

We therefore resorted to the second photoactivatable nuclear translocation approach, which is based on a Ca²⁺-dependent system, consisting of two components: (i) a GFP-tagged fusion protein contains dCas9, VP64 and an N-terminal fragment of NFAT (residues 1–460 without the DNA binding domain to avoid binding to endogenous NFAT targets), in which NFAT_1–460 was fused to either the N- or C-terminus of dCas9, and NLS was inserted in different positions depending on the constructs; (ii) Opto-CRAC, which comprises an ORAI-activating fragment from the cytoplasmic domain of STIM1 (SOAR or CAD) and LOV2 domain to induce Ca²⁺ influx by blue light. In the dark, the SOAR/CAD domain was caged by LOV2 to prevent the activation of ORAI calcium channels. Following blue light exposure, the unwinding of the LOV2-Jα helix promoted the exposure of SOAR/CAD, which subsequently moved toward the plasma membrane to directly engage and activate ORAI1 Ca²⁺ channels.^19,24-28

The prototypical design of Opto-CRAC contains LOV2 and a STIM1 cytosolic fragment (aa 336–486) as we described previously.¹⁹ However, this construct showed appreciable dark activity (construct B1, Figure 2B,C), which might cause constitutive nuclear translocation of NFAT in the dark. To confer tightest control over the CaRROT system, we created a series of second generation Opto-CRAC constructs by (i) varying the length of STIM1-CT fragments (B1–3; Figure 2B); (ii) fusion to Zdark (Zdk) protein,²⁹ a light-dependent LOV2 binder and changing the linkers (B4–9); or (iii) using SOAR domain derived from other species (such as After transfecting Opto-CRAC constructs to a GCaMP6s-stable HeLa cells, we next set out to identify the best performing construct based on two criteria: (i) reduced dark activity; (ii) enhanced dynamic ranges of calcium signal changes in response to light stimulation. Since Zdk binds to LOV2 tightly in the dark but dissociates from LOV2 upon light stimulation,²⁹ we reasoned that Zdk might serve as an additional “lock” to further cage LOV2-SOAR fusion in a quiescent configuration, thus reducing the background activation. Indeed, some of the Zdk constructs showed substantially reduced dark activity (constructs B5, B6, and B8; Figure 2C). However, in some constructs, the addition of Zdk led to narrower dynamic ranges (B8 and B9, Figure 2D) and slower onset of light-inducible Ca²⁺ responses when compared to B1 (Supplementary Figure S1B,C). Ultimately, the chimera made of LOV2 and D. rerio SOAR (DrSOAR; residues 341–442) turned out to be an ideal candidate with negligible dark activation but potent photoinduced calcium influx, with the activation and deactivation half-lives of 18.7 and 24.5 s, respectively (construct B10; Figure 2D,E and Supplementary Figure S1D). We therefore used this construct for the downstream applications.

Chemical and Photoinducible Nuclear Translocation of dCas9.

We next evaluated the dCas9 nuclear translocation of different designs upon blue light illumination by transfecting HeLa cells with mCherry-tagged dCas9-VP64-LOV2-NLS constructs (Figure 2A, constructs V1 and V2) or cotransfecting LOV2-DrSOAR (construct B10) with various CaRROT constructs (Figure 2A, constructs V3–5). In cells expressing mCherry-tagged V1 or V2, dCas9 underwent very low translocation from cytosol to nuclei in response to blue light illumination (Figure 2F, first and second panels), suggesting that either the NLS is not strong enough or not fully exposed to drive nuclear import of dCas9. Contrariwise, for cells transfected with V3 or V4, dCas9 was observed in the nuclei prior to light illumination when cotransfected with Opto-CRAC-B10 (Figure 2F, third and fourth panels). The result indicated that NLS inserted in these constructs have a strong affinity for the import machinery, and thus shuttles the fusion protein into nuclei even in the dark. After removing all the NLS in dCas9, the CaRROT construct V5 remained exclusively in the cytosol in the dark. Upon light stimulation, CaRROT-V5 showed light-inducible translocation into the nuclei of cells cotransfected with Opto-CRAC-B10 (~3.5-fold change; Figure 2F,G). This system also showed no discernible dark-state background activity (Figure 2H), which was concordant with the minimal Ca²⁺ influx “leakiness” of the improved Opto-CRAC system (Figure 2C and 2F).

Next, to confirm that our system could be likewise manipulated by chemicals that could alter intracellular calcium signals, we used thapsigargin (TG), which block the sarcoplasmic/endoplasmic reticulum Ca²⁺ ATPase (SERCA) pump to passively induce calcium store depletion in the ER.³⁰ This process is immediately followed by the activation of STIM1 to open ORAI calcium channels and cause bulky flooding of calcium ions into the cytosol. Time-lapse imaging showed that TG, similar to photoactivated Opto-CRAC, could lead to the nuclear entry of dCas9 (~4-fold changes in nuclear fluorescence signals; Figure 2F,G). Taken together, our results establish that we could use both chemical and light to control the cytosol-to-nucleus shuttling of dCas9-fused transcriptional effector, which is an indispensable step for the fusion protein to execute its function within the nuclei.

Chemical and Photoactivatable Transcriptional Modulation.

Several previous studies have demonstrated that gene activation could be achieved by targeting VP64 effector domain to the transcribed region of a gene^8,9,31 due to its ability to recruit basal transcriptional machinery.³² In our design, the dCas9-sgRNA complex serves as a locus-positioning apparatus to direct VP64 to targeted genomic loci, e.g., promoters of targeted genes (Figure 3A). To rapidly evaluate if our CaRROT system can precisely control gene expression in response to light or chemicals, we used an EGFP reporter assay, in which eight repeats of a guide RNA target sequence situated upstream of a minimal cytomegalovirus (CMV) promoter.³³ Since the activated reporter expresses EGFP, we replaced the GFP module on the CaRROT-V5 construct by BFP to avoid fluorescence overlap. Similar to GFP-tagged version, BFP-tagged CaRROT-V5 showed undetectable dark-state background activity, but underwent nuclear translocation when cotransfected with Opto-CRAC-B10 upon blue light stimulation or TG treatment (Supplementary Figure S2A-C). After addition of TG to induce calcium influx in cells transfected with BFP-tagged CaRROT-V5 and corresponding sgRNA, we observed a pronounced increase in EGFP levels compared with control cells without TG administration per different microscopic fields (Supplementary Figure S3A, second panel). We performed the similar experiment by using Opto-CRAC-V5, rather than TG, to elicit calcium entry, and noticed a similar increase of EGFP expression (Supplementary Figure S3A, last panel). As a control, cells transfected with CaRROT-V5 and sgRNA did not show the significant increase of EGFP signals regardless of the presence of light (Supplementary Figure S3A, first panel).

Figure 3.

Use of CaRROT to chemically or photoinduce EGFP reporter expression. (A) Monitoring the EGFP reporter signals in HeLa cells expressing the indicated set of constructs with the combination of Opto-CRAC-B10 or chemical (TG, 2 μM). Scale bar: 5 μm. (B) Time course showing the changes of GFP signals in the same imaging field under the indicated conditions. (C) Quantification of EGFP reporter intensities before and after light stimulation or TG treatment. Data were showed as mean ± SD (n = 9). **P < 0.01; ***P < 0.001 compared to untreated conditions (two-tailed Student’s t-test).

To better visualize the behaviors in the same transfected cells at real time in response to blue light, we cotransfected different sets of vectors for 24 h into Hela cells, and then recorded the time-lapse imaging with pulsed photostimulation (blue LED at 470 nm; power density of 0–50 μW/mm; 5 s ON, 20 s OFF) or by adding TG to the cells (12 h). The assay was performed with the following three groups: (i) BFP-tagged CaRROT-V5, sgRNA, Opto-CRAC-B10 and EGFP reporter; (ii) BFP-tagged CaRROT-V5, sgRNA, adding TG and EGFP reporter; (iii) BFP-tagged CaRROT-V5, sgRNA and EGFP reporter as a negative control. The addition of TG to CaRROT-V5-contained cells (Figure 3A, middle panels) turned on the EGFP signals statistically higher than cells that only received CaRROT-V5 and the reporter construct (Figure 3A, left panels). The similar trend was observed in the group of cells transfected with CaRROT-V5 and Opto-CRAC-B10, in which the EGFP signals in the same cells were markedly higher than before light stimulation and the negative control (Figure 3A, right panels). The EGFP signals raised up to 3-fold compared to the starting point (Figure 3B).

Each cell probably received a different amount of vectors mixture; therefore, the evaluation of the EGFP intensity of the whole cell population would better reflect the efficiency of the CaRROT system. We then used flow cytometry to quantify the numbers of cells showing chemical or light-inducible activation of the EGFP reporter (Supplementary Figure S3B). Since the magnitude of the difference between cells irradiated with light and incubated in the dark varied on the threshold applied to the EGFP fluorescence intensity, we calculated the number of EGFP-positive cells at different indicated thresholds. The cells transfected with CaRROT-V5, Opto-CRAC-B10, sgRNA, and EGFP reporter and subjected to light illumination showed a statistically higher number of EGFP⁺ cells than either cells shielded in the dark or the negative control group at any thresholds (Supplementary Figure S3C). The similar trend was also observed in the group received CaRROT-V5 and TG, which showed significant enrichment of EGFP⁺ cells compared to negative controls (Supplementary Figure S3C).

To demonstrate that our synthetic system would also allow photoactivation of endogenous genes, we used a set of sgRNAs designed to target the promoter regions of the human achaetescute family bHLH transcription factor 1 (ASCL1) or myogenic differentiation 1 (MYOD1), and then evaluated their light-dependent transcription in HEK293T cells. ASCL1 acts as a pioneer transcription factor to control neuronal differentiation;³⁴ whereas MYOD1 is a key regulator for skeletal muscle differentiation, which is able to induce transdifferentiation of fibroblasts or other cell types into myocytes.³⁵ Provided their fundamental roles in developmental biology,^36,37 light-inducible expression of ASCL1 or MYOD1 will likely be useful for future temporal control of the differentiation of neurons or muscle cells in regenerative medicine. We therefore chose these two genes to test the CaRROT system. Cells transfected with dCas9-NLS-VP64 were used as positive control, which showed light-independent expression of both genes (left bars, Figure 4A,B). For cells transfected with CaRROT-V5 and Opto-CRAC-B10, we observed a significant increase in gene expression upon light illumination for each individual or combined sgRNAs. For MYOD1, the mean levels of expression were enhanced by over 200-fold, which was comparable to those of dCas9-NLS-VP64 expressing cells (Figure 4A). The remarkable light-dependent transcription was also observed when we targeted the ASCL1 locus (Figure 4B). In all cases, gene expression levels in cells maintained in the dark were comparable to cells transfected with the empty vectors (right bar, Figure 4A,B). Notably, the coexpression of our system with two sgRNAs targeted MYOD1 or three sgRNAs targeted ASCL1 did not show a significant difference compared to individual sgRNA transfection, suggesting that the expression of multiple guide RNAs targeted to the same gene does not seem to cause synergistic activation of both endogenous genes ASCL1 and MYOD1 (Figure 4A,B). This observation also indicates that by using our system, one well-designed sgRNA probably would be sufficient to activate the expression of endogenous genes.

Figure 4.

CaRROT-mediated light-inducible activation of endogenous gene expression. Light-induced endogenous gene expression of (A) MYOD1 and (B) ASCL1 in HEK293T cells were measured by qRT-PCR. Cells were transfected with dCas9-NLS-VP64 as positive control (PC), BFP-tagged-CaRROT-V5 construct, Opto-CRAC-B10 and indicated sgRNAs or the empty plasmid (pTriEX-BFP). Cells were subjected to pulsed blue light stimulation (470 nm, 50 μW/cm²). *P < 0.05; ***P < 0.001; ****P < 0.0001 compared to the dark group (two-tailed Student’s t-test).

In summary, we have devised a synthetic transcriptional reprogramming device (CaRROT) that can be tightly controlled by chemicals or light to induce endogenous gene transcription with high precision. Since the system relies on the generation of Ca²⁺ signals to drive nuclear translocation of CaRROT, it can be further extended to record or permanently mark Ca²⁺ dependent activities in neurons or lymphocytes once coupled with a reporter gene (e.g., GFP or luciferase).

METHODS

Cloning and Plasmid Construction.

Opto-CRAC vectors were designed by amplifying Homo sapiens STIM1-CT fragments (residues 336–486, 336–442, 347–448) and Danio rerio STIM1(341–442) using the KOD Hot start DNA polymerase (EMD Millipore, Billerica, MA, USA) and inserted downstream of LOV2_404–546 between the HindIII-XhoI restriction sites to replace Rac1 in the pTriEX-mcherry-PA-Rac1 plasmid (Addgene, #22027). The linker 1 (GSGLEGSGG) or linker 2 (GSGLESG) was introduced to the Opto-CRAC vectors at NotI-XhoI sites. cDNAs encoding Zdk1 and Zdk2 were a gift from Dr. Klaus M Hahn at the University of North Carolina at Chapel Hill. They were amplified and inserted between XhoI-XbaI sites.

To construct dCas9-based nuclear translocation vectors, we introduced sequentially NFAT (1–460) and dCas9, VP64 (derived from Addgene plasmid 22027), and GFP or BFP to AflII/AgeI/HindIII and XhoI sites of pcDNA3.1(+). The NLS oligonucleotides were also inserted during amplification depending on the construct. AsLOV2-based bipartite NLS1 and NLS2 generated from biLINUS 9 and biLINUS 11, respectively, were gifts from Dr. Barbara Di Ventura at the University of Heidelberg, Germany. All the restriction enzymes used in our studies were purchased from New England Biolabs.

The sgRNA targeting EGFP reporter, MYOD1 sgRNA 1 and sgRNA2 were obtained from Addgene (#60719, #64137 and #64138). The sgRNAs targeting ASCL1 were generated by annealed oligo cloning using the BsmBI site of LentiCRISPRv2, (Addgene: #52961). EGFP reporter containing eight copies of a gRNA binding site for light-inducible dCas9 activation was obtained from Addgene (#60718).

Cell Culture and Transfections.

HEK293T and HeLa cells from the American Type Culture Collection (ATCC) were maintained in DMEM medium (Gibco) supplemented with 10% FBS, 100 unit/ml penicillin and 100 μg/mL Streptomycin (Gibco) at 37 °C in a humidified atmosphere under 5% CO₂. For confocal imaging, 2 × 10⁵ cultured cells were seeded on 35 mm glass-bottom dishes 24 h before transfection using Lipofectamine 3000 (Life Technologies) according to the manufacturer’s instructions.

Fluorescence Imaging and Statistical Analysis.

Confocal imaging was performed generally at 24 h after the transfection by using an inverted Nikon Eclipse Ti-E microscope customized with Nikon A1R+ confocal laser sources (405/488/561/640 nm). We used an external blue light (470 nm, tunable intensity of 0–50 μW/mm², ThorLabs Inc., Newton, NJ, USA) for photostimulation.

For measurements of Ca²⁺ influx using the green color calcium indicator GCaMP6s or red indicator jRCaMP1b, we cotransfected 100 ng of each Opto-CRAC construct and 100 ng cytosolic GCaMP6s or jRCaMP1b into HeLa cells using Lipofectamine 3000. 24 h after transfection, a 488 nm laser was used to excite GFP, and a 561 nm laser to excite mCherry at intervals of 1–5 s. mCherry-positive cells were used for statistical analysis.

In order to evaluate the nuclear translocation of dCas9 variants, we transfected single dCas9-VP64-mCherry-AsLOV2-bipartite NLS version 1.0 or Version 2.0 into Hela cells or cotransfected Opto-CRAC with one of the following constructs: CaRROT-V3, V4, and V5 then incubated for 1 day. Photostimulation was used to induce Ca²⁺ influx mediated by Opto-CRAC constructs, and the cells were time-lapse recorded for more than 30 min at intervals of 2 min. Nine mCherry-positive cells (the first two constructs) and both mCherry and BFP/GFP-positive cells (the last three constructs) were selected to calculate the ratio of fluorescence signal between nuclei and the total fluorescence (nuclei plus cytosolic intensities).

For EGFP reporter assay, HeLa cells were seeded onto 4-well glass bottom dishes for confocal imaging and transfected with the different sets of vectors in each well: BFP-tagged CaRROT-V5, sgRNA, Opto-CRAC, and EGFP reporter at ratio 3:1:1:2. 24 h later, fluorescent images were acquired at 37 °C and 5% CO₂ at 40× or 60× magnification. Imaged fields were selected from three different areas in each well, and time-lapse recording lasted for 12 h at the intervals of 5 min under blue light exposure (pulsed blue LED at 470 nm; power density of 0–50 μW/mm²; 5 s ON, 20 s OFF) or adding TG (2 μM). At least nine BFP, GFP and mCherry-positive cells (the group used Opto-CRAC) or BFP and GFP-positive cells (without Opto-CRAC) were chosen for EGFP reporter expression analysis.

Flow Cytometry Analysis.

Hela cells were seeded in two 6-well plates transfected with the combination of vectors in each well: BFP-tagged-CaRROT-V5 (750 ng), sgRNA (250 ng), EGFP reporter (500 ng), Opto-CRAC (250 ng). After 24 h of transfection, one plate was kept in the dark, and another plate was added TG (2 μM) or subjected to blue light irradiation (470 nm, tunable intensity of 0–50 μW/mm) for 1 h, followed by pulsed stimulation (5 s ON for every 20 s) for another 24 h to maintain the constant activation of the light-inducible system. Next, the cells were washed, trypsinized and washed with PBS twice. The levels of fluorescence protein were determined using the LSRII flow cytometer (BD Biosciences). Cells were sampled at a medium flow rate, and 10 000 cells were counted for each condition. FlowJo software (TreeStar) was used to analyze the data (EGFP⁺ in BFP⁺ or BFP⁺mch⁺ cell populations). The experiments were conducted in duplicate.

Endogenous Genes Expression Assay.

HEK293T cells were plated at approximately 5 × 10⁴ cells/well in 12-well plates (Corning Inc., USA) and cultured for 24 h. Each well was transfected with 250 ng, 250 ng and 750 ng of the sgRNA expression plasmid, Opto-CRAC, and BFP-tagged-CaRROT-V5, respectively. As a positive control, plasmid encoding dCas9-NLS-VP64 and sgRNA were transfected at a 3:1 ratio. After 24 h of transfection, samples underwent blue light stimulation or incubated in the dark as described above. On the next day, mRNA was extracted using Qiagen RNeasy spin prep columns and reverse transcription PCR was performed using amfiRivert cDNA Synthesis Platinum Master Mix (genDEPOT). Relative levels of cDNA were detected using amfiSure qGreen Q-PCR Master Mix (genDEPOT) and Mastercycle Real-Time PCR (Eppendorf, USA). The data were normalized to GAPDH levels and cells transfected with an empty plasmid (control) using the ΔΔCt method.