A Cre-Dependent CRISPR/dCas9 System for Gene Expression Regulation in Neurons

Abstract Site-specific genetic and epigenetic targeting of distinct cell populations is a central goal in molecular neuroscience and is crucial to understand the gene regulatory mechanisms that underlie complex phenotypes and behaviors. While recent technological advances have enabled unprecedented control over gene expression, many of these approaches are focused on selected model organisms and/or require labor-intensive customization for different applications. The simplicity and modularity of clustered regularly interspaced short palindromic repeats (CRISPR)-based systems have transformed genome editing and expanded the gene regulatory toolbox. However, there are few available tools for cell-selective CRISPR regulation in neurons. We designed, validated, and optimized CRISPR activation (CRISPRa) and CRISPR interference (CRISPRi) systems for Cre recombinase-dependent gene regulation. Unexpectedly, CRISPRa systems based on a traditional double-floxed inverted open reading frame (DIO) strategy exhibited leaky target gene induction even without Cre. Therefore, we developed an intron-containing Cre-dependent CRISPRa system (SVI-DIO-dCas9-VPR) that alleviated leaky gene induction and outperformed the traditional DIO system at endogenous genes in HEK293T cells and rat primary neuron cultures. Using gene-specific CRISPR sgRNAs, we demonstrate that SVI-DIO-dCas9-VPR can activate numerous rat or human genes (GRM2, Tent5b, Fos, Sstr2, and Gadd45b) in a Cre-specific manner. To illustrate the versatility of this tool, we created a parallel CRISPRi construct that successfully inhibited expression from a luciferase reporter in HEK293T cells only in the presence of Cre. These results provide a robust framework for Cre-dependent CRISPR-dCas9 approaches across different model systems, and enable cell-specific targeting when combined with common Cre driver lines or Cre delivery via viral vectors.


Introduction
Genetic and epigenetic targeting are fundamental strategies to study gene regulation and function, and also provide novel therapeutic avenues for genetic diseases. In recent years, clustered regularly interspaced short palindromic repeats (CRISPR) systems have revolutionized the field as tools for site-specific DNA and RNA editing (Jinek et al., 2012;Cong et al., 2013;Savell and Day, 2017;Pickar-Oliver and Gersbach, 2019). In these systems, a Cas nuclease is directed to target DNA by an engineered CRISPR single guide RNA (sgRNA), resulting in nucleasemediated cleavage of the targeted nucleic acid sequence. In addition to CRISPR-Cas9 systems, a number of new approaches have been developed based on a mutated, catalytically dead Cas9 (dCas9). CRISPR-dCas9 strategies target specific genomic loci without causing double-strand breaks, and can instead be fused to an array of different effector proteins (Savell and Day, 2017). dCas9 systems are commonly used to induce transcriptional activation (CRISPRa) or interference (CRISPRi), epigenetic modifications, chromatin looping, tagging, or can be used as an anchor system to deliver non-coding RNAs (CRISPR-Display; Hilton et al., 2015;Konermann et al., 2015;Shechner et al., 2015;Yeo et al., 2018;Chen et al., 2019;Savell et al., 2019a;Carullo et al., 2020).
DNA recombinases are commonly used to enable inversion, deletion, or integration of transgenes in a cell-specific manner. The Cre/Lox system is a commonly used recombination approach in which the Cre recombinase recognizes specific 34-bp palindromic Lox sites within a DNA sequence (Anton and Graham, 1995;Gibb et al., 2010). Cre-mediated inversion or deletion occurs when flanking Lox sites are oriented in opposing or parallel orientations, respectively. Importantly, Cre recombinase expression driven under cell type-specific promoters enables targeted gene manipulation of these cell populations (Madisen et al., 2012;Daigle et al., 2018). Due in part to the ease and versatility of this system, hundreds of different Cre-driver transgenic lines have been generated for cell-targeted approaches (Gong et al., 2007;Madisen et al., 2012;Daigle et al., 2018). Likewise, Cre-dependent genetic FLEX switches (flip-excision; also known as double-inverted open reading frame, or DIO switches) invert DNA sequences to enable gene activation or silencing (Schnütgen et al., 2003). This approach has widely been used in neuroscience, enabling cell type-specific optogenetic and chemogenetic activation, overexpression or knock-down of transgenes, and perturbation of endogenous genes (Kühn et al., 1995;Tsien et al., 1996;Gong et al., 2007;Atasoy et al., 2008;Madisen et al., 2012;Erwin et al., 2020).
While CRISPR/dCas9 approaches have enabled targeted induction of transcriptional or epigenetic states at selected genes, inducible cell type-specific CRISPR tools based on these platforms remain limited (Kumar et al., 2018;Bäck et al., 2019). For example, most currently available transgenic CRISPR-dCas9 lines are limited to the mouse (Zhou et al., 2018;Gemberling et al., 2021), and existing rat transgenic lines rely on Cre-dependent CRISPR guide RNA constructs to achieve cell type specificity (Bäck et al., 2019). These approaches are limited to the specific transgenic model system, and require costly crosses with Cre lines and time-consuming animal colony management. Similar approaches based on viral delivery of CRISPR machinery have so far been limited to constitutive transgene expression (Zheng et al., 2018;Savell et al., 2019a), or only enable genome editing via the introduction of double-strand breaks (Kumar et al., 2018). Here, we developed novel Cre-dependent CRISPRa and CRISPRi systems that incorporate a synthetic intron into the dCas9 transgene. Creating an intron-containing dCas9 construct offers two advantages: (1) introns can increase transgene expression compared with intron-lacking counterparts (Shaul, 2017); and (2) introns provide a non-coding locus for insertion sites of Lox sequences without disturbing the integrity of the coding protein. By separating the dCas9 transgene into two segments with a short SV40 intron (SVI), we created CRISPRa and CRISPRi systems in which the first segment of dCas9 is double-floxed and inverted. Importantly, this SV40 intron is only 97 bp, and therefore does not substantially affect construct size or compatibility with viral vector delivery. This approach prevents leaky expression of functional CRISPR constructs by requiring Cre-induced inversion to orient both dCas9 segments into an open reading frame. Furthermore, the construct is designed for easy rearrangement and replacement of promoters and effector fusions and can be applied for various CRISPR/dCas9 systems.

Cultured neuron experiments
Primary rat neuronal cultures were generated from embryonic day 18 rat striatal tissue as described previously (Savell et al., 2019a). Briefly, cell culture wells were coated overnight at 37°C with poly-L-lysine (0.05 mg/ml for culture wells supplemented with up to 0.05 mg/ml laminin) and rinsed with diH 2 O. Dissected tissues were incubated with papain for 25 min at 37°C. After rinsing in HBSS, a single-cell suspension of the tissue was re-suspended in Neurobasal media (Invitrogen) by trituration through a series of large to small fire-polished Pasteur pipets. Primary neuronal cells were passed through a 100 mM cell strainer, spun and re-suspended in fresh media. Cells were then counted and plated to a density of 125,000 cells per well on 24-well culture plate with or without glass coverslips (60,000 cells/cm). Cells were grown in Neurobasal media plus B-27 and L-glutamine supplement (complete neurobasal media) for 11 days in vitro (DIV) in a humidified CO 2 (5%) incubator at 37°C.
For virus experiments, cells were transduced with lentiviruses on DIV4 or DIV5. All viruses had a minimum titer of 1 Â 10 9 GC/ml, with a target multiplicity of infection (MOIs) of at least 1000. After an 8-to 16-h incubation period, virus-containing media was replaced with conditioned media to minimize toxicity. A regular half-media change followed on DIV8. On DIV11, transduced cells were imaged and virus expression was verified before RNA extraction. EGFP and mCherry expression was also used to visualize successful transduction using a Nikon TiS inverted epifluorescence microscope.

RNA extraction and RT-qPCR
Total RNA was extracted (RNAeasy kit, QIAGEN) with DNase treatment (RNase free DNase, QIAGEN), and reverse-transcribed (iScript cDNA Synthesis kit, Bio-Rad). cDNA was subject to qPCR for genes of interest, as described previously (Savell et al., 2016). A list of PCR primer sequences is provided in Table 1.

CRISPR-dCas9 construct design
To achieve transcriptional activation or inactivation, lentivirus-compatible plasmids were engineered to express dCas9 fused to either VPR or KRAB-MeCP2, based on existing published plasmids [Addgene plasmid #114196 (Savell et al., 2019a); Addgene plasmid #155365 ]. A Cre-dependent DIO version of the dCas9-VPR construct was generated by insertion of LoxP and Lox2272 sequences flanking the dCas9-VPR cassette. dCas9-VPR was PCR-amplified to insert additional restriction sites (KpnI, BmtI, and BspDI) to allow for LoxP and Lox2272 insertion and was subsequently inserted in reverse orientation to create an intermediate construct via sequential digest and ligation (AgeI, EcoRI). LoxP and Lox2272 sites were amplified from a DIO construct (Addgene plasmid #113685; Don et al., 2017) to create restriction sites (KpnI and BmtI around one set; BspDI and KpnI around another set) and were inserted into the intermediate construct via sequential restriction digest and ligation (KpnI, BmtI, BspDI, and EcoRI). Intron-containing dCas9-VPR was constructed by the insertion of a gBlock containing the SV40 intron into the original dCas9-VPR plasmid via Gibson assembly (Gibson Assembly kit, New England BioLabs). The SVI-FLEX construct was built via Gibson assembly of the original dCas9-VPR backbone and two gBlocks encoding the SV40 intron sequence, LoxP and Lox2272 sites (sequences based on DIO construct described above) and dCas9-part1 in inverted orientation. The intron-containing CRISPRi construct (SVI-dCas9-KRAB-MeCP2) was built via restriction digest and Gibson assembly (SfiI, EcoRI, and Xhol, Gibson Assembly kit, New England BioLabs) of the KRAB-MeCP2 and SVI-dCas9-VPR constructs. A Cre-encoding construct (Addgene plasmid #49056; Kaspar et al., 2002) was used to amplify and insert a Cre transgene into a lentivirus compatible backbone that contained the hSYN promoter and expressed mCherry for visualization via Gibson assembly. To create an additional Cre construct, mCherry was replaced with GFP via sequential digest and ligation (EcoRI and XhoI). dCas9-VPR-expressing constructs were co-transduced with sgRNA-containing constructs. Gene-specific sgRNAs were designed using an online sgRNA tool, provided by the Zhang Lab at MIT (crispr. mit.edu) and inserted in a previously described lentivirus compatible sgRNA scaffold construct (Addgene plasmid #114199; Savell et al., 2019b). To ensure specificity, all CRISPR crRNA sequences were analyzed with the National Center for Biotechnology Information's (NCBI) Basic Local Alignment Search Tool (BLAST) and Cas-OFFinder (http:// www.rgenome.net/cas-offinder/). sgRNAs were designed to target GRM2, Tent5b, Sstr2, Gadd45b, and Fos, respectively. A list of the target sequences is provided in Table 1. crRNA sequences were annealed and ligated into the sgRNA scaffold using the BbsI or BsmBI cut site. Plasmids were sequence-verified with Sanger sequencing; final crRNA insertion was verified using PCR. Lentivirus-compatible SYN-SVI-DIO-dCas9-VPR and SYN-SVI-DIO-dCas9-KRAB-MeCP2 plasmids are available on Addgene (plasmids #164576 and #170378).

Lentivirus production
Viruses were produced in a sterile environment subject to BSL-2 safety by transfecting HEK293T cells with specified CRISPR-dCas9 plasmids, the psPAX2 packaging plasmid, and the pCMV-VSV-G envelope plasmid (Addgene plasmids #12260 and #8454) with FuGene HD (Promega) for 40-48 h as previously described (Savell et al., 2019b). Viruses were purified using filter (0.45 mm) and ultracentrifugation (25,000 rpm, 1 h 45 min) steps. Viral titer was determined using a qPCR Lentivirus Titration kit (Lenti-X, qRT-PCR Titration kit, Takara). For smaller scale virus preparation, each sgRNA plasmid was transfected in a 12-well culture plate as described above. After 40-48 h, lentiviruses were concentrated with Lenti-X concentrator (Takara), resuspended in sterile PBS, and used immediately. Viruses were stored in sterile PBS at À80°C in single-use aliquots.

Luciferase assay
Bidirectional regulation by SVI-DIO CRISPRa and CRISPRi machinery was examined using a previously described Fos luciferase reporter plasmid . A total of 80,000 HEK293T cells were plated in 500 ml HEK Media. After cells reached 40-50% confluence, 500 ng total plasmid DNA was transfected with 1.5 ml FuGene HD (Promega) as follows: 50 ng of luciferase plasmid, 450 ng in 1:2 molar ratio of total sgRNA: CRISPRa or CRISPRi plasmid. A luciferase glow assay was performed according to manufacturer's instructions 40 h following transfection (Thermo Scientific Pierce Firefly Glow assay; Thermo Scientific 16177). Cells were lysed in 100 ml 1Â luciferase cell lysis buffer while shaking at low speed and protected from light for 30 min; 20 ml of cell lysate was then added to an opaque 96-well microplate (Corning 353296) and combined with 50 ml 1Â D-luciferin working solution supplemented with 1Â firefly signal enhancer (Thermo Scientific Pierce Firefly Signal Enhancer; Thermo Scientific 16180). Following a 10-min dark incubation period to allow for signal stabilization, luminescence was recorded using a Synergy 2 Multi-Detection Microplate Reader (BioTek). Luminescence in dCas9-KRAB-MeCP2 and dCas9-VPR experiments was recorded with a read height of 1 mm, 1-s integration time, and 135or 100-ms delay, respectively. Representative images of luciferase reporter activity assays were captured using an Azure c600 imager (Azure Biosystems).

Results
Creation and validation of a traditional FLEX-CRISPRa system AAV-driven FLEX systems are among the most common Cre/Lox approaches used in neurons, enabling inducible expression by combination with Cre-driver animal models. While this strategy has proven to be a versatile approach for cell-specific targeting, few tools have been described for Cre-dependent CRISPR manipulation. Given the flexibility of CRISPRa and CRISPRi systems for transcriptional manipulation at endogenous loci, development of Cre-dependent platforms for these systems are of interest. To create a Cre-inducible system for transcriptional activation, we adapted a CRISPRa tool recently validated in neurons (Savell et al., 2019a) using a common DIO strategy. In this DIO system, an inverted dCas9-VPR cassette was flanked by LoxP and Lox2272 sites (Fig. 1A), enabling Cre-dependent inversion of dCas9-VPR into the correct orientation followed by the excision of one antiparallel Lox site to prevent further inversion events. To validate this system in dividing cells, HEK293T cells were co-transfected with either a constitutively active dCas9-VPR construct or the DIO-dCas9-VPR construct, in tandem with sgRNA plasmids targeting either the human GRM2 promoter, or a non-targeting control (sgRNA for the bacterial lacZ gene). The DIO-dCas9-VPR construct was tested with or without transfection of a Cre-2A-mCherry plasmid that was driven under the human synapsin (hSYN) promoter (Fig. 1B). Constitutive activation of the endogenous GRM2 gene yielded a 75-fold increase of GRM2 mRNA compared with a non-targeting lacZ control sgRNA. While the Cre-dependent DIO-dCas9-VPR system was able to induce GRM2 mRNA transcription by 112-fold, we also observed a 49-fold increase in the absence of Cre. Next, to examine gene induction in neurons, we transduced rat primary striatal neuron cultures with lentiviruses expressing sgRNAs, CRISPR machinery, and Cre recombinase (Fig. 1C). CRISPR sgRNAs targeting the Tent5b (also known as Fam46b) promoter resulted in strong upregulation of Tent5b mRNA with both the constitutive and Cre-dependent constructs. However, in neurons the DIO system also exhibited leaky expression, indicated by an 8-fold increase of Tent5b mRNA in the absence of Cre. Tent5b is a highly inducible gene which could be more susceptible to baseline/leaky induction compared with other genes. Therefore, we tested the DIO-dCas9-VPR system at two other genes including a gene encoding the neuropeptide receptor Sstr2 and DNA-damage inducible gene Gadd45b (Fig. 1D). Although these genes are less inducible than Tent5b (with induction rates of 8-to 19-fold compared with lacZ controls), we still observed non-specific gene induction in the absence of Cre (2-to 5-fold induction without Cre). While we chose efficient gRNAs validated in prior studies (Savell et al., 2019aCarullo et al., 2020), it is possible that the differences in inducibility across gene targets are driven by the respective gRNA properties rather than inherent inducibility of the target genes. Alternately, this difference may correspond to variation in gene inducibility, driven by baseline expression, gene length, or other features of transcript regulation. Nevertheless, the induction of target genes in the absence of Cre recombinase is likely because of leaky expression of the DIO-dCas9-VPR transgene, since no transgene inversion could be detected in PCR verification using genomic DNA isolated from cells lacking Cre (Fig. 1E). These results are consistent with recent evidence that DIO transgenes can be expressed in the inverted orientation at low levels in the absence of Cre recombinase (Fischer et al., 2019).

Intron insertion into dCas9 increases expression and enables creation of a split-dCas9 FLEX system
Given that an inverted dCas9-VPR cassette still resulted in target gene induction in the absence of Cre, we next aimed to design a system that would prohibit expression of the full length dCas9-VPR fusion protein without Cre-mediated recombination. To achieve this goal, we designed an intron-containing FLEX system. First, we inserted a small SV40 intron (SVI) into the dCas9 sequence to create a constitutively active, intron-containing construct (SVI-dCas9-VPR) in which the dCas9 cassette is divided into two segments by the SV40 intron ( Fig. 2A). We created two versions (1.0 and 2.0) of this SVI-dCas9-VPR construct to determine optimal intron positioning and maximal splicing efficiency. While the intron was only slightly shifted between versions 1.0 and 2.0, SVI-dCas9-VPR 2.0 created a frameshift immediately after the intron, resulting in a premature stop codon within ,100 bases if the product was not spliced (Fig. 2B). PCR amplification from cDNA with primers spanning this intron sequence following HEK293T plasmid transfection revealed that both constructs produced detectable dCas9 expression. However, splicing efficiency of the SVI-dCas9-VPR 2.0 version was higher as compared with SVI-dCas9-VPR 1.0 (Fig. 2C,D). This variability is not surprising given the known effects of intron placement on splicing efficiency (Shaul, 2017). While we still detected low levels of unspliced mRNA using the SVI-dCas9-VPR 2.0 construct, this is potentially driven by nascent RNA that has not yet undergone splicing rather that intron retention (which would result in a premature stop codon).
Using the more efficient SVI-dCas9-VPR 2.0, we next generated a split-dCas9 DIO cassette by inverting and flanking the first segment of dCas9 with LoxP and Lox2272 sites to prevent leaky transgene expression (Fig.  2E). With only part of the dCas9 cassette inverted, Credependent inversion and recombination is required to yield a full-length functional dCas9-VPR fusion protein.
Using the same intron-spanning primer set as in previous validations, we compared PCR products amplified from the original constitutive dCas9-VPR, the intron-containing constitutive VPR (SVI-dCas9-VPR 2.0), and the new Credependent SVI-dCas9-VPR (SVI-DIO-dCas9-VPR) plasmids. As predicted, the SVI-dCas9-VPR 2.0 plasmid yielded a longer product compared with the original dCas9-VPR with no intron (Fig. 2F), because of the 97 bp intron in the SVI-dCas9-VPR 2.0 plasmid. However, because the SVI-DIO-dCas9-VPR contains an inverted dCas9 segment, the validation primers target the same strand and yield no PCR product. These results further validate the absence of non-specific recombination following plasmid transfection.
We next transfected HEK293T cells with each of these three constructs to compare recombination, expression, and splicing efficiency of the SVI-DIO-dCas9-VPR construct. The SVI-DIO-dCas9-VPR groups also received a Cre-EGFP plasmid that was driven under the hSYN promoter to initiate recombination. PCR amplification of cDNA generated from the transfected HEK293T experiments with intron-spanning primers revealed strong signals for the short, spliced PCR product for all three groups (Fig. 2G). This indicates that all three constructs were expressed in HEK293T cells and that both SVI-dCas9-VPR 2.0 and SVI-DIO-dCas9-VPR transgenes  generated mRNA transcripts with efficient splicing and intron exclusion. We further observed that, in HEK293T cells, SVI-dCas9-VPR and SVI-DIO-dCas9-VPR mRNA was expressed at levels 4-fold and 2-fold higher (respectively) than observed with the constitutive dCas9-VPR plasmid (Fig. 2H). Together, these experiments demonstrate that this newly developed intron-containing FLEX system is capable of efficient splicing and transcription and that Cre-dependent recombination is required for inversion and expression of the dCas9-VPR transgene.
Similarly, targeting the highly inducible Tent5b gene in cultured striatal neurons using lentiviral transgene delivery resulted in strong upregulation of Tent5b mRNA with the Cre-dependent constructs (41-fold), but not in the absence of Cre (2-fold; Fig. 3C). These patterns of Cre-dependent induction without leaky effects were also observed for both of the moderately inducible target genes tested in Figure 1 (Sstr2 and Gadd45b; Fig. 3D). PCR amplification of genomic DNA that was extracted from transduced primary neurons further validated that recombination required expression of Cre (Fig. 3E). Together, these data demonstrate that this novel SVI-  DIO-dCas9-VPR construct mitigates the leaky gene induction seen in conventional DIO systems for CRISPRa approaches while maintaining the capacity for robust and specific gene induction.

Cre-dependent CRISPRa and CRISPRi constructs modulate expression from a luciferase reporter
To further validate functionality of our SVI-DIO-dCas9-VPR construct, we tested this system at a luciferase reporter plasmid driven by the rat Fos promoter (Fig. 4A,B; Duke et al., 2020). Luciferase assays are a common luminescence reporter system used to validate effects of an effector molecule on gene expression. Luciferase, when paired with its substrate D-luciferin and in the presence of ATP, O 2 , and Mg 21 , produces bioluminescence that can provide insight into the direct transcriptional activity at a regulatory element driving luciferase expression. HEK293T cells were co-transfected with plasmids expressing the Fos luciferase plasmid, constitutive or Cre-inducible CRISPRa plasmids, and plasmids expressing sgRNAs targeting either the rat Fos promoter or a non-targeting lacZ control. Additionally, the SVI-DIO-dCas9-VPR construct was transfected with a Cre-EGFP plasmid. Notably, the constitutive dCas9-VPR constructs with and without the SV40 intron increased luminescence 26-fold and 25-fold change, respectively, when compared with lacZ. Likewise, the SVI-DIO-dCas9-VPR increased luminescence in a similar range (18-fold), but only in the presence of Cre recombinase (Fig. 4C). Consistent with our previous results in neurons and HEK293T cells (Fig. 3), this Cre-dependent CRISPRa construct exhibited no leaky induction in the absence of Cre, confirming the utility of this approach for Cre-specific expression and gene induction.
While gene overexpression is useful for gain-of-function experiments, CRISPR/dCas9 approaches benefit from the ability to use the same sgRNA for both gene activation and repression, based on the identity of the protein fused to dCas9. However, given the possibility of DIO transgene expression in reverse orientation (Fischer et al., 2019), leaky expression is likely not limited to dCas9-VPR based systems but could also occur in other CRISPR-DIO based systems. In order to generate Cre-dependent system for CRISPRi, we created a parallel Cre-dependent construct in which dCas9 is fused to a KRAB-MeCP2 repressor  Figure 3. Validation of the SVI-DIO-dCas9-VPR CRISPRa system. A, Illustration of CRISPR construct designs for the intron-containing SVI-dCas9-VPR (top) and a Cre-dependent double-floxed inverted open reading frame (DIO) SVI-dCas9-VPR (bottom). B, In transfected HEK293T cells the constitutive VPR construct with and without the SV40 intron induced the endogenous target gene GRM2 by 55-fold and 138-fold compared with the non-targeting lacZ control (Welch's t test for constitutive t (2) = 3.846, p = 0.0614, and constitutive with intron t (2) = 6.322, p = 0.0241). SVI-DIO-dCas9-VPR significantly induced transcription of GRM2 with but not without the Cre construct (n = 3 per group, two-way ANOVA F (1,8) = 179.5, p , 0.0001). C, Lentiviral expression of SVI-DIO-dCas9-VPR and Cre in striatal neurons increased mRNA for the target gene Tent5b (n = 10 per group, two-way ANOVA F (1,36) = 34.97, p , 0.0001). D, Additional target genes tested in neurons demonstrated specific induction only in presence of Cre recombinase (two-way ANOVA with n = 3 per group for Gadd45b F (1,48) = 4.578, p = 0.0375, and for Sstr2 F (1,48) = 6.615, p = 0.0133). E, Illustration of primer and PCR product position for recombination validation experiments (left). qPCR on transduced neuronal DNA revealed high levels of the inverted dCas9-VPR cassette (p3, middle, n = 2 per group, Kruskal-Wallis test F (2,4) = 3.714 p = 0.2) with and without Cre. Inversion into the correct orientation only occurred in the presence of Cre (p4, right, n = 2 per group, Kruskal-Wallis test F (2,4) = 4.57 p = 0.0667). SVI-DIO-dCas9-VPR groups with and without Cre are compared with a non-transduced control (NTC). Data expressed as mean 6 SEM.
domain that results in robust gene silencing when targeted to gene promoters (Yeo et al., 2018;Duke et al., 2020;Fig. 4D). Following similar transfection and luciferase assay protocols used for our CRISPRa system, we found that constitutive KRAB-MeCP2 both with and without the SV40 intron effectively inhibited luminescence from the Fos-luciferase reporter in HEK293T cells (Fig. 4E). Likewise, the SVI-DIO-dCas9-KRAB-MeCP2 inhibited luciferase activity only in the presence of Cre recombinase. Together, these results demonstrate the capacity of our SVI-DIO CRISPR constructs to target and modulate gene expression at specific transcriptional regulatory regions in a Cre-dependent manner. Likewise, because luciferase reporter assays require successful translation of luciferase protein, these experiments also demonstrate the capability of this system to bidirectionally regulate protein levels in addition to mRNA of targeted genes. Furthermore, adaptation of this SVI-DIO system for multiple CRISPR approaches, such as CRISPRi, highlights the potential and versatility of this tool for cell type-specific and gene-specific regulation in future experiments.

Discussion
The mammalian brain consists of heterogeneous cell populations with distinct characteristics and functions. Technologies to study gene function in specific tissues, brain regions, or cell populations are therefore necessary to understand their role in behavior and disease. For example, cell type-selective promoters have been used for specific targeting of excitatory neurons (Camk2a; Mima et al., 2001), astrocytes (Gfap; Morelli et al., 1999), and cell populations that express specific enzymes or receptors such as tyrosine hydroxylase (Th; Savitt et al., 2005), dopamine or serotonin transporters (Slc6a3, Slc6a4; Zhuang et al., 2005;Bäck et al., 2019), or dopamine receptors (Drd1, Drd2;Lobo et al., 2006;Zhang et al., 2006). In this study we developed and optimized Cre-dependent CRISPRa and CRISPRi systems that enable gene and cell type-specific transcriptional activation and inactivation, respectively. In cultured dividing and non-diving cell lines, we demonstrate that this SVI-DIO-dCas9 based CRISPRa system can activate highly inducible endogenous genes in HEK293T cells (GRM2 gene) and striatal neurons (Tent5b), without undesired gene induction in   the absence of Cre recombinase (Figs. 3, 4). This system also performed well when targeted to moderately inducible genes such as Sstr2 and Gadd45b in neurons (Fig. 3), demonstrating a wide range of effect size and application possibilities. Furthermore, successful activation and repression of a Fos luciferase reporter in HEK293T cells using SVI-DIO CRISPR constructs demonstrates the versatility of this machinery to provide robust bidirectional control over gene expression and protein levels.
Using a classic DIO strategy in which the entire dCas9-VPR cassette was double-floxed and inverted, we observed leaky target gene induction in the absence of Cre (Fig. 1). While it is unclear what drives this leaky expression, it is conceivable and in line with recent findings that DIO transgenes can be expressed in the inverted orientation in the absence of Cre (Fischer et al., 2019). To avoid this leaky expression, we generated a novel SVI-DIO-dCas9 system in which dCas9 was broken up into two segments by an SV40 intron. This intron provided a location for insertion of Lox sites to invert and double-flox the first dCas9 fragment while leaving the second fragment undisturbed. Without Cre-mediated recombination, the dCas9 segments remain in opposing orientations and are thus unable to yield functional dCas9 protein. Consistent with observations that introns can facilitate nuclear mRNA export and translation processes and potentially increase functional protein levels (Shaul, 2017), we found that intron insertion increased dCas9-VPR mRNA expression following plasmid transfection (Fig. 2H). Critically, because of the small size of this intron (97 bp), it is unlikely to substantially limit viral packaging and could easily be applied to increase transgene expression in other constructs.
One of the first cell type-specific CRISPR systems used in neurons was based on Cre-dependent expression of CRISPR sgRNAs. Bäck et al., developed a Lox-stop-Lox based sgRNA construct and used it for CRISPR-Cas9 mediated gene knock-out (Bäck et al., 2019). This study tested this approach in transgenic rats that expressed Cre recombinase under the rat dopamine transporter (Slc6a3) promoter to specifically target and knock-out the Th gene in dopaminergic neurons. While this approach allowed for cell type-specific expression of sgRNA constructs, a concern with similar systems is the untargeted overexpression of Cas9 or dCas9 fusion proteins, which could potentially cause unintended and non-specific effects on gene expression.
Our system extends this previous work in two ways. First, the SVI-DIO-dCas9 approach alleviates some of these concerns, as the dCas9 construct itself is Cre dependent, and therefore, functional fusion proteins are not expressed without Cre recombinase. Second, our SVI-DIO-dCas9 system is compatible with traditional validated sgRNA constructs, which can be multiplexed for effect size titration at a single gene (Savell et al., 2019a). Additionally, this approach is also compatible with more complex sgRNA arrays that target entire gene programs . Thus, this study adds another approach for gene regulation to the CRISPR toolbox, while also outlining a novel intron/DIO strategy to avoid leaky Cre-independent transgene expression. Compared with traditional Cre-dependent overexpression vectors, Credependent CRISPR strategies provides several advantages as they enable targeting of one or multiple endogenous genomic loci and titration of effect size for more physiological expression levels.
Prior work has incorporated temporal specificity into Cre/Lox systems, either via use of optically or chemically inducible proteins. For example, selective estrogen receptor modulator (SERM) inducible systems have been generated by fusion of the ligand binding domain of the ER to Cre (Cre-ER). In this approach, a mutated version of the mouse ER that binds tamoxifen but not estrogen was used to create a Tamoxifen-inducible Cre system (Danielian et al., 1993;Metzger and Chambon, 2001). Activity of Cre recombinase can therefore be activated or inactivated on Tamoxifen injections, increasing temporal control of gene editing events.
In addition to cell type-specificity, selective promoters can be used to drive expression in response to stimulation or experience. Bacterial artificial chromosome (BAC) transgenic animals were generated to express GFP or channelrhodopsin (ChR2) under an immediate early gene promoter such as Fos or Arc (Guenthner et al., 2013). Using activity-induced approaches in combination with our SVI-DIO-dCas9 system would allow for experiencedependent genetic and epigenetic manipulations.
Cre/Lox systems are also among the most common approaches for intersectional circuit-specific manipulations in neuroscience. For example, in recent work, circuit-specific CRISPR genome editing was achieved by injection of a Cre-dependent Cas9 viral vector in the nucleus accumbens (NAc) and injection of a monosynaptic rabies virus expressing a sgRNA targeting the Fosb gene in the ventral hippocampus (vHPC; Eagle et al., 2020). While the sgRNA virus successfully spread through all hippocampal projections, significant knock-out of Fosb occurred only in vHPC neurons that projected to the NAc. In future studies, our SVI-DIO-dCas9 system could similarly be used to gain a better understanding of projections between specific cell types in the brain. The intron-containing effector construct could be delivered to one brain region and a Cre-expressing construct to the second brain region through a monosynaptic rabies virus or retrograde AAV serotype. Genetic or epigenetic editing would then only occur in cells with projections between the two targeted brain regions without leaky expression or side effects related to effector protein overexpression in non-targeted cells.
While the use of our SVI-DIO-dCas9 system in Cre driver animal models is an expected application, this approach does not require transgenic organisms. In addition to sgRNA and SVI-DIO-dCas9-VPR or SVI-dCas9-KRAB-MeCP2 delivery, a separate construct expressing Cre can be delivered as necessary to target brain tissues. Additionally, this intron-containing dCas9 provides a basic framework that can be customized and combined with a number of effector proteins to introduce a variety of genetic or epigenetic modifications. The system can easily be customized for expression in various cell types and brain regions and adjusted for inducible efforts and enhanced temporal specificity.