Using the dCas9-KRAB system to repress gene expression in hiPSC-derived NGN2 neurons

Summary We describe a CRISPR inhibition (CRISPRi) protocol to repress endogenous gene expression (e.g., ATP6V1A) in human induced pluripotent stem cell-derived NGN2-induced glutamatergic neurons. CRISPRi enables efficient and precise gene repression of one or multiple target genes via delivering gRNA(s) to direct a dCas9-KRAB fusion protein to the gene(s) of interest. This protocol can also be adapted for gene activation and high-throughput gene manipulation, allowing assessment of the transcriptomic and phenotypic impact of candidate gene(s) associated with neurodevelopment or brain disease. For complete details on the use and execution of this protocol, please refer to Ho et al. (2017) and Wang et al. (2021).


SUMMARY
We describe a CRISPR inhibition (CRISPRi) protocol to repress endogenous gene expression (e.g., ATP6V1A) in human induced pluripotent stem cellderived NGN2-induced glutamatergic neurons. CRISPRi enables efficient and precise gene repression of one or multiple target genes via delivering gRNA(s) to direct a dCas9-KRAB fusion protein to the gene(s) of interest. This protocol can also be adapted for gene activation and high-throughput gene manipulation, allowing assessment of the transcriptomic and phenotypic impact of candidate gene(s) associated with neurodevelopment or brain disease. For complete details on the use and execution of this protocol, please refer to Ho et al. (2017) and Wang et al. (2021).

Timing: 4-5 weeks
This protocol uses lentiviral delivery to transduce the stable neural progenitor cells (NPCs) lines expressing the CRISPR-repressor system from human induced pluripotent stem cells (hiPSCs) with pLV-TetO-hNGN2-eGFP-Neo, rtTA, and gRNAs. Human NGN2 (hNGN2) can rapidly yield excitatory glutamatergic neurons from NPCs within 21 days. A CRISPR-repressor complex, paired with designed specific gRNA, can repress the expression of the gene of interest. For instance, we utilized CRISPR inhibition (CRISPRi) to reduce endogenous ATP6V1A, an example gene in NGN2 neurons (Wang et al., 2021). Researchers can manipulate their gene(s) of interest using a similar approach.
Each vial contains 1 3 10 6 viable NPCs. f. Once thawed, grow NPCs in NPC medium containing 1 mg/mL puromycin.  a. Select 'Gene repression' for step 1 to specify the type of manipulation on the genome. b. Select 'Human(GRCh37/h19)' for step 2 to specify the organism. c. Click 'Submit step 2, Next Step'. d. Type the gene name for step 3 to specify the gene of interest. e. Click 'CRISPR-ERA search'. f. Select 'Using U6 promoter', as we will clone gRNAs into lentiGuide-Hygro-mTagBFP2 (Addgene, #99374), which contains a U6 promoter. g. Select top 6 sgRNAs with highest E+S score that are also well-located within 1 kb upstream from Transcription Start Site (TSS). i. If the 'Distance to TSS' values of chosen sgRNAs are close (% 20 difference) then choose another sgRNA that has the next best E+S score but a greater difference in 'Distance to TSS' value to other sgRNAs chosen. h. Example of ATP6V1A sgRNA selection to repress gene expression: Design 6 sgRNAs to target the promoter region for ATP6V1A knockdown (KD) (Figure 1). 3. Create a manifest file containing information with target gene name, species, distance to TSS, E (efficacy) + S (Specificity) scores, sgRNA sequence, and CRISPR application. a. Note that the sgRNA sequences should be sure to include the oligo overhangs. Importantly, we add a 'G' as the initiation nucleotide, making the transcription more efficient. b. Our lab usually adds a 'G' when ordering oligos. On the 5' end of the Forward Oligo, there will be a 'CACCG' sequence, and on the 5' end of the Reverse Oligo, there will be an 'AAAC' sequence. The overhangs are required for the oligos to be complementary to the linearized plasmid backbone during ligation, which will allow for successful cloning. Thus, the oligo orders should be '5'-CACCGN...N-3' and '5'-AAACN...NC-3'. c. The ordered sgRNAs will arrive as two single-stranded guide RNA per gRNA. An annealing step must be performed prior to cloning. 4. (Optional. This step is additional verification for efficiency and specificity of designed sgRNAs.) Off-target sgRNA comparison prevents unwanted gene manipulation. a. To check for off-target interactions for your generated sgRNAs you can use various tools, we used http://crispor.tefor.net/. b. For 'step 1' copy and paste the binding element sequence for the sgRNA you wish to verify.
i. The sequence must be at least 20 base pairs long.  i. We used the default selection '20bp-NGG-Sp Cas9, SpCas9HF1, eSpCas9 1.1'. e. Select 'Submit', and wait about 10-15 s to get the results. f. The two specificity scores, MIT and CFD, can then be used to gauge the likelihood of your sgRNA having off-target genome effects. i. The scores range from 0-100, with the higher scores indicating a lower off-target effect on the genome of interest. ii. sgRNAs should have a specificity score greater than 50 to be utilized for cloning and later transfection.
Note: The U6 RNA Pol III promoter transcription starts at the +1 position (23 nucleotides after the TATA box), with G as the preferred initiation nucleotide for RNA Pol III promoters. A Guanine nucleotide must be included on one of the 5' ends of the sgRNA, if it is not already included, to ensure that there is complete and efficient transcription (Kim et al., 2020).
Note: Check whether sgRNAs target the sharing motifs, which may potentially influence the expression of neighboring genes.

Cloning custom sgRNAs into lentiviral vector of lentiGuide-Hygro-mTagBFP2
Timing: 1 week 5. Anneal DNA Oligos a. Order 25 nmole DNA oligos. b. Reconstitute the dehydrated oligos with UltraPure DNase/RNase-free distilled water to a 100 mM stock. Store at -20 C. i. Alternatively, order the oligos in a reconstituted form to save time if working with a large quantity of sgRNA. c. Set up the oligo phosphorylation and annealing for each forward (FWD) and reverse (REV) oligos. Oligos can be ordered phosphorylated.
d. Incubate the reaction mixture in a thermocycler using the following parameters ( Figure 2). Following this step, the oligos will be annealed and phosphorylated. e. Combine 3 mL of each annealed oligos to create a pool of all annealed oligos. Once pooled, dilute the annealed oligos at a 1:10 dilution with UltraPure DNase/RNase-free distilled water.
Pause point: Store both the un-pooled annealed oligos and the pooled/diluted annealed oligos at -20 C until future use. 6. Golden Gate Assembly allows for simultaneous digestion and ligation. a. Prepare the backbone of lentiGuide-Hygro-mTagBFP2 with a concentration of 25 ng/mL. Check for the BsmBI recognition sequence prior to running the assembly. b. Clone the annealed oligos (from step 5) into the lentiGuide-Hygro-mTagBFP2 plasmid via the BsmBI restriction sites (Figure 3). c. Complete the following reaction in a PCR tube. d. Incubate the reaction mixture in a thermal cycler using the following parameters. 15 cycles for Stage 1 (step 1: 5 min at 37 C; step 2: 5 min at 23 C) (Figure 4).  e. Heat-shock each transformation tube by placing the bottom 1/2 to 2/3 of the tube into a 42 C water bath for 30-60 s (45 sec is usually ideal). f. Put the tubes back on ice for 2 min. g. Add 250-1,000 mL LB or SOC medium (without antibiotic) to the bacteria and grow in 37 C shaking incubator for 45 min. h. Plate some or all of the transformation onto a 10-cm LB agar plate containing the appropriate antibiotic (ampicillin). i. Incubate plates at 37 C overnight (18-24 h). j. Next morning, check for colonies (see troubleshooting 1).
Pause point: Bacterial colony plates may be stored at 4 C for weeks. However, we recommend the verification of recombinant colonies by DNA sequencing soon.   Pause point: The sequence-verified plasmid DNA can be stored at -20 C until future use.
Note: In the case of ATP6V1A gRNAs cloning, we notice that Sanger sequencing verified a high efficiency cloning into the plasmid backbone.

Lentivirus Production
Timing: 4 days This step uses a PEI (Polysciences, #23966-2)-based 293T cell transfection method to produce and harvest lentivirus for downstream 2D cell culture infection protocols.
CRITICAL: NIH guidelines require the maintenance of a Biological Safety Level 2 (BSL-2) facility for work involving lentivirus. BSL-2 labs require: (1) all personnel to wear lab coats and gloves, (2) all procedures that may produce infectious aerosols or splashes to be performed within a biological safety cabinet (BSC), (3) an autoclave, or alternative method of decontamination, to be available, (4) for the laboratory doors to be self-closing, and (5) for sink and eyewash stations to be readily available. For more info, please refer to this link: https://www.cdc.gov/training/quicklearns/biosafety/).

CRITICAL:
Steps of lentivirus collection and concentration should be performed in a BSC. a. Perform transfection when 293T cells are at approximately 80% confluency on a 15-cm plate. b. Thaw or prewarm PEI and Opti-MEM on a 37 C water or bead bath. c. In a 15 mL conical tube, combine PEI and DNA mixture in a 1:1 ratio.
i. Drop PEI mixture into DNA mixture slowly to avoid precipitating the DNA (can be seen as white wisps). ii. Gently agitate once between drops to minimize precipitated DNA.
iii. Gently swirl 5 times after adding all the PEI. iv. Invert 5 times and gently vortex for 10 sec. d. Incubate the mixture for 15 min at room temperature (20 C-25 C) (the solution will become translucent). e. Drop 700 mL of the PEI/DNA mixture onto a 15-cm plate with 80% confluent 293T cells and put the plate into the incubator. f. After 6-8 h, perform a full media change (15 mL). g. After 48 h, harvest the first batch of viral media into a 50 mL conical tube or another sterile container. h. Add 15 mL of media to the plate and return the plate to the incubator. i. After 24 h, harvest the second batch of viral media into a 50 mL conical tube or another sterile container. j. Discard the plate in an appropriate manner. Treat the 293T cell plates with a 10% bleach solution for 10 min and aspirate the solution before discarding plates into the biohazard bin. k. Filter collected viral media from both batches by passing through a 0.22 mm pore size filter into a 50 mL conical tube or another sterile container.
Note: VSVG is an envelope protein, acting as an empty backbone for lentiviral production. MDL and Rev are the 3 rd generation lentiviral packaging plasmids that are required for generating functional lentiviral particles.
Pause point: Store filtered viral media on ice in a bucket at 4 C for less than 3 days.
CRITICAL: Under microscope, 293T cells should give very high transfection efficiency (R97%) for all gRNAs. This is a necessity for high-titer virus production (see troubleshooting 2).

Lentivirus Concentration
Timing: 4 h This step produces concentrated lentivirus in tissue culture media for use in 2D cell culture infection protocols.
Note: It is recommended to concentrate lentivirus on the day of the second viral media harvest to maximize virus yield. However, the viral media can also be stored for less than 3 days at 4 C for concentrating. a. Secure the centrifuge rotor into the ultracentrifuge (Beckman Coulter L-100XP Ultracentrifuge, or Optima XE-90 with 32 SW-Ti rotor) and pre-cool to 4 C. b. In the BSC, using sterile forceps, insert empty 38.5-mL polypropylene centrifuge tubes into the metal centrifuge tubes for the rotor being utilized. c. Cap the metal centrifuge tubes and pre-chill in 4 C for 15 min. d. Retrieve tubes and add 25-38 mL of filtered viral media (see above, Lentivirus Production, step 9-k) into polypropylene centrifuge tubes. e. Cap metal centrifuge tubes and carefully bring to ultracentrifuge. h. Carefully open the metal centrifuge tubes and use sterile forceps to remove the inserted polypropylene centrifuge tubes. i. Remove supernatant media by gently tilting the polypropylene centrifuge tubes and aspirating close to the top. j. Gently resuspend virus pellet (almost invisible) with DMEM or medium of choice.
i. Avoid pipetting up and down during resuspension more than 10 times.
ii. Volume depends on how much you want to concentrate the virus.
iii. Resuspension with 500 mL should yield virus on the order of 10 6 IU/mL. k. Pipette the resuspended virus into a sterile 15 mL conical tube or 1.5 mL microcentrifuge tube. l. Centrifuge at 3,000 3 g for 5 min to pellet any residual cell debris. Aliquot supernatant into 0.5 mL microcentrifuge tube.
Pause point: Store at -80 C until future use.
CRITICAL: Minimize freeze-thaw cycles: a single freeze-thaw cycle can reduce infection efficiency by 50%.

Lentivirus Titration
Timing: 2 h This step uses the abm's qPCR lentivirus titration kit (Cat#: LV900) to assess the amount of virus particles in the concentrated lentiviral preparations. These steps follow those outlined in the manufacturer's protocol. a. In a 0.5 mL microcentrifuge tube, add 2 mL of lentivirus to 18 mL of PBS (1:10 dilution). b. In a 0.5 mL microcentrifuge tube, add 2 mL of the diluted virus (from step 11-a) to 18 mL of Virus Lysis Buffer (1:10 dilution). i. Gently mix by swirling pipette tip around or tapping tube. c. Incubate mixture at room temperature (20 C-25 C) for 3 min. This solution is now referred to as the 'viral lysate'.
Note: The Ct value of the viral lysate will be used to determine the titer of the concentrated lentivirus preparation. The viral sample has since been diluted 1:10 twice; therefore, these dilution factors need to be considered in the titration calculation. d. Set-up qPCR reactions in triplicate, except negative control (NTC).
e. Program the real-time qPCR instrument with the following cycle protocol ( Figure 6). Note: Because the viral lysate had been diluted 1:10 twice, a total dilution factor of 1003 was factored into the titer calculation. Adjust the dilution factor as needed.

Lentiviral spinfection
Timing: 6 h The time of spinfection and the G-force may increase the transduction efficiency.
Pause point: D21 NGN2 neurons can be fixed with 4% PFA and stored at 4 C for future immunofluorescence. TRIzol/RIPA-lysed samples can be stored at -20 C for future biochemical assays.
Note: Dox induction should yield a high percentage of GFP-positive cells.
Note: To measure the neuronal electric activity, we strongly recommend NGN2 neurons should be co-cultured with human fetal astrocytes to enhance neuronal maturation. On day 3, astrocytes are split as 17,000 cells/well in a Matrigel-coated 48W MEA plate (Axion Biosystems, M768-tMEA-48W) and maintained as above. On day 7, 70,000 NGN2 neurons are detached, spun down, and seeded on the astrocytes culture. We can detect the population-wide neuronal activity, including frequency and coordination of network firing, and the amplitude of voltage-gated potassium (K + ) current and sodium (Na + ) current using an Axion multi-electrode array (MEA) or whole-cell patch-clamp recordings respectively. We can keep the neuronal culture up to 35 days for recording if necessary.
CRITICAL: If there is a need to seed cells on to a microplate, such as a CytoView 48-well MEA plate, for neuronal activity measurement, then we highly recommend that you split NGN2 neurons using Accutase before differentiation day 7. Dissociating neurons at later stages of neuronal differentiation may lead to increased cell death and potential loss of cultures.

Validation of CRISPRi-induced gene repression by qPCR and western blot
Timing: 2 days Antibiotics Resistance gene in plasmids Selection

OPEN ACCESS
The validation step is a critical prerequisite before starting a functional study. There is a necessity to redesign and test new gRNAs if CRISPRi fails to significantly repress gene mRNA and protein expression (see troubleshooting 7). Reduced gene expression may vary greatly according to genes, gRNAs, transduced tissue types, and batches. We expect a consistent and significant reduction of gene transcription and protein translation. A satisfying level of CRISPRi can sufficiently knockdown the genes of interest to yield a phenotype, even though a small reduction may be achieved.  tail (13,Thermo Fisher Scientific,#78440). Alternatively, we can make our own RIPA lysis solution (see materials and equipment). b. Samples are sonicated for 1 min then centrifuged at 15,871 3 g for 10 min. c. The supernatant is collected, and the total protein concentration is determined by using Quick Start Bradford Protein Assay (Bio-Rad, #5000201) following the manufacturer's instructions. d. Use an antibody (ATP6V1A rabbit antibody: ab199326, 1:1,000 dilution) to obtain the protein expression of the target protein. b-Actin (mouse monoclonal antibody, ab8227, 1:1,000 dilution) is used as a loading control. e. Secondary antibodies used include IRDye 800CW Donkey anti-Mouse IgG (1:15,000, LI-COR, #925-32212) and IRDye 680RD Donkey anti-Rabbit IgG (1:15,000, LI-COR, #925-68073). f. Capture and quantify images using the Odyssey Imaging Systems (LI-COR Inc.). Alternatively, we can apply a conventional western blot using enhanced chemiluminescent (ECL, GE, #RPN2236).
Note: Antibody for the detection of novel target proteins is not always available.

EXPECTED OUTCOMES
The lentiGuide viral delivery has a robust transduction result of >90% BFP-positive cells after 48 h post-infection ( Figure 8A). The BFP signal should remain constant throughout neuronal maturation. On differentiation day 21, NPC-induced glutamatergic neurons show neuronal cell morphology with multifold neurite extensions ( Figure 8B). After Dox withdrawal the NGN2-GFP signal may gradually decrease. Cells are immunostained with MAP2 and TUJ1 ( Figure 8C), and images are captured using a confocal microscope (LSM 780, Zeiss) with a 633 objective lens. We expect to generate a highly homogeneous neuronal population.

LIMITATIONS
CRISPRi provides a powerful tool to decipher the impact of down-regulation of gene expression in development and disease. One must carefully design knockdown experiments to distinguish the impact of CRISPRi on neuronal patterning, maturation, and function; towards this, the dCas9-KRAB effector and/or gRNAs can be delivered or induced to express either early or late in experimental design. We find that a minimum of seven days is necessary to ensure robust knockdown.
The approach depends on a successful search for effective gRNAs. It is critical to validate each gRNA in every donor and cell type, as the magnitude of CRISPRi achieved varies significantly according to ll OPEN ACCESS cellular context. On occasion, we have failed to detect any gRNA that significantly impacts the expression of a gene of interest (Ho et al., 2017). For single-cell experiments, particularly pooled CRISPR screening experiments (Tian et al., 2019), it may be necessary to ensure a comparable degree of knockdown between all neurons in the population, which is likely to be best achieved by generating clonal CRISPRi hiPSCs that express dCas9-KRAB equally in all cells.
It is critical to rule out potential cis-and trans-off-target effects, which may arise either from homologous sequence or proximity in the linear (Schrode et al., 2019) or 3D genome. For this reason, CRISPRi may not be ideal if: (a), the targeting sequence is a bidirectional promoter, or (b), the transcripts for the gene of interest overlap or are very close to the transcripts of one or more different genes.
In theory, the KRAB domain transcriptionally represses all RNA isoforms of a gene of interest. For this, CasRx approaches that bind RNA targets may be a more suitable approach for the functional evaluation of the impact of down-regulation of specific isoforms or RNA splicing.

TROUBLESHOOTING Problem 1
No colonies grow on agar plates after transformation (step 7-j).

Potential solution
The quantity and quality of DNA is not sufficient for plasmid transformation.
Solution 1: Increase the amount of DNA when mixed with NEB 10-beta competent cells.
Solution 2: There might be a need to re-do the Golden Gate Assembly. Pay more attention when pipetting small volumes.
Solution 3: Check the sgRNA design to make sure gRNAs can be ligated into the BsmBI-digested lentiGuide-Hygro-mTagBFP2. Problem 2 BFP fluorescence is low at the 36h point after initial viral transfection of 293T cells (step 9-CRITICAL).

Potential solution
We have found that replacing PEI with PEIMax (Polysciences #24765; 1 MG/ML pH 7.0 stock) at a ratio of 1 mg DNA to 1 mL PEIMax can significantly increase the transfection efficiency during lentiviral production. To optimize this protocol with new DNA and PEIMax stocks, you will need to test transfection efficiency ranging from 1:1 to 1:10 and pick the concentration with the highest efficiency.
Problem 3 Lentiviral titer is low after virus concentration by ultracentrifuge (step 11-f).

Potential solution
We found this problem to be most likely associated with the low transfection of 293T cells. It is better to optimize the protocol and plasmid concentrations to improve the transfection efficiency.
Problem 4 gRNA BFP signal cannot be detected in transduced NPCs/neurons using fluorescent microscopy (step 17-Note).

Potential solution
We should exclude the possibility of inefficient gRNA transduction. Transduction efficiency may increase by the infection of suspension cells after seeding.
BFP signal usually is faint in the days immediately following transduction and may be difficult to determine by eye. The strength of the signal should increase until it peaks around 3-4 days after transduction. We advise keeping a control well that has the gRNA vector but has not been transduced; this control will act as a comparison to the cultures with and without BFP signal.

Problem 5
Cultures perish during selection with multiple antibiotics (step 18-b).

Potential solution
Antibiotic selection may be staggered (i.e., G418 then hygro selection) to aid culture viability. Typically, problems with culture viability during antibiotic selection arise from low viral transduction rates. Check GFP and BFP signals before selection with antibiotics to ensure viral transduction is successful.

Problem 6
During neuronal differentiation, NGN2 neuronal cultures clump together, risking detachment and interfering with phenotyping of mature cultures (step 21).

Potential solution
This is typically an issue with Matrigel coating. A strong recommendation is to optimize Matrigel concentration prior to start an extensive experiment. Sequential coating with polyethylenimine (PEI, branched) and Matrigel coating promotes firm attachment. This coating method is well suited for imaging purpose.
Clumping may be reduced during seeding by gently pipetting NPC pellets after spinning using a P1000 tip to break apart cell clusters and aid resuspension of cells as a monolayer.

Potential solution
Insufficient gene repression may occur for several reasons. We need to: Increase the viral amount of gRNA, if the problem is due to a low lentiviral transduction. Check whether low expression of specific genes may be causing apoptotic cell death. The cells with low gene expression will be washed away when changing the media. Measure dCas9 expression levels in parental NPCs or NGN2 neurons. Combine multiple gRNAs that may lead to significant repression. If the problem continues, design and test more gRNAs.
More information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Aiqun Li (aiqun.li@mssm.edu).

Materials availability
This study did not generate any unique reagents.

Data and code availability
The accession number for ATP6V1A knockdown RNA sequencing data reported in this paper is Gene expression omnibus (GEO): GSE128367.