CRISPR Gene Editing Tool for MicroRNA Cluster Network Analysis

Summary

This protocol describes a high-throughput clustered regularly interspaced short palindromic repeats (CRISPR) gene editing workflow for microRNA cluster network analysis that allows the rapid generation of a panel of genetically modified cell lines carrying unique miRNA cluster member deletion combinations as large as 35 kb within a single experiment.

Abstract

MicroRNAs (miRNAs) have emerged as important cellular regulators (tumor suppressors, pro-oncogenic factors) of cancer and metastasis. Most published studies focus on a single miRNA when characterizing the role of small RNAs in cancer. However, ~30% of human miRNA genes are organized in clustered units that are often co-expressed, indicating a complex and coordinated system of noncoding RNA regulation. A clearer understating of how clustered miRNA networks function cooperatively to regulate tumor growth, cancer aggressiveness, and drug resistance is required before translating noncoding small RNAs to the clinic.

The use of a high-throughput clustered regularly interspaced short palindromic repeats (CRISPR)-mediated gene editing procedure has been employed to study the oncogenic role of a genomic cluster of seven miRNA genes located within a locus spanning ~35,000 bp in length in the context of prostate cancer. For this approach, human cancer cell lines were infected with a lentivirus vector for doxycycline (DOX)-inducible Cas9 nuclease grown in DOX-containing medium for 48 h. The cells were subsequently co-transfected with synthetic trans-activating CRISPR RNA (tracrRNA) complexed with genomic site-specific CRISPR RNA (crRNA) oligonucleotides to allow the rapid generation of cancer cell lines carrying the entire miRNA cluster deletion and individual or combination miRNA gene cluster deletions within a single experiment.

The advantages of this high-throughput gene editing system are the ability to avoid time-consuming DNA vector subcloning, the flexibility in transfecting cells with unique guide RNA combinations in a 24-well format, and the lower-cost PCR genotyping using crude cell lysates. Studies using this streamlined approach promise to uncover functional redundancies and synergistic/antagonistic interactions between miRNA cluster members, which will aid in characterizing the complex small noncoding RNA networks involved in human disease and better inform future therapeutic design.

Introduction

Better research tools are needed to investigate the contribution of noncoding RNAs in human disease. MiRNA dysregulation is often observed in human disorders such as cancer when comparing the expression profiles of these small noncoding RNAs in the tissues and body fluids (e.g., blood, urine) of cancer patients versus noncancer, healthy individuals, employing microarrays, quantitative real-time PCR (qRT-PCR), and next-generation deep sequencing technologies^1,2. Recent work has characterized a large subset of these miRNAs as tumor suppressor, oncogenic, and metastasis factors that control tumor formation, disease progression, and drug resistance. Experimental overexpression and/or downregulation/loss of miRNAs result in functional and pleiotropic consequences in the cell, reflecting the wide range of cancer-associated activities these noncoding RNAs coordinate - growth, apoptosis, differentiation, chromatin remodeling, angiogenesis, epithelial to mesenchymal (EMT) and MET transitions, metabolism, and immune response³.

MiRNAs are encoded as single genes or reside in genomic clusters, which are transcribed in the nucleus and extensively processed before generating the biologically mature, single-stranded ~22 nucleotide (nt) miRNA species localized in the cytoplasm⁴. These small RNAs exert their effects post-transcriptionally and act as negative gene regulators that bind to messenger RNA (mRNA) targets in a sequence-specific manner to bring the catalytic RNA-induced silencing complex (RISC) to the mRNA site, resulting in mRNA degradation and/or a block in protein translation. MiRNAs are an extremely abundant class of noncoding RNAs in animal systems, and 2,654 mature miRNAs exist in the human genome (miRBase release 22.1)⁵. MiRNAs typically associate with incomplete complementarity to their mRNA targets⁴. Therefore, a single miRNA can regulate tens to hundreds of distinct mRNA targets and functionally impact a large range of biological pathways. To add to the complexity of miRNA-based mechanisms, a single mRNA can be regulated by multiple, distinct miRNAs. It is, thus, challenging to investigate how miRNA dysregulation disrupts the body's homeostatic balance and leads to human malignancy.

The majority of published studies have focused on a single miRNA when characterizing their role in disease events. However, ~30% of human miRNA genes are organized in clustered units (typically ~10 kilobases [kb]) that are often transcribed in the same orientation and co-expressed, indicating a coordinated and complex system of noncoding RNA regulation⁶. The largest polycistronic human miRNA cluster is the 14q32 cluster comprising 54 miRNA precursors. One of the most well-studied clustered miRNA units associated with human cancers is the miR-17-92 polycistronic cluster comprised of miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, and miR-92-1 residing within intron 3 of the noncoding RNA, c13orf25. The miR-17-92 cluster is frequently amplified in hematopoietic malignancies and overexpressed in solid tumors and has established oncogenic roles in promoting cell cycle progression, apoptosis, and angiogenesis⁷. In addition, the tumor suppressive miR-15a and miR-16-1 cluster located within the intron of the noncoding gene Leu2 is often deleted in leukemias and downregulated in a large range of cancers, functioning to block tumor growth by targeting the antiapoptotic gene BCL2 and additional cell cycle progression genes⁸. The miR-888 cluster is elevated in patients with high-grade prostate cancer and consists of seven miRNA genes (miR-892c, -890, -888, -892a, -892b, -891b, and -891a) located on human chromosome Xq27.3^9,10.

The miR-888 cluster maps within the HPCX1 locus (Hereditary Prostate Cancer, X-linked 1) spanning Xq27-2, which was identified by linkage analysis of hereditary prostate cancer family pedigrees^{11,12,13,14,15,29}. Functional characterization of individual miR-888 clustered members using conventional miRNA misexpression tools - miRNA mimics and antisense inhibitors - indicated that these miRNAs play overlapping roles in regulating prostate tumor growth and invasion^9,10. However, these experimental methods do not lend themselves easily to studying how multiple miR-888 clustered members act synergistically or antagonistically in a noncoding RNA network to control tissue homeostasis and cancer progression. This described streamlined protocol using high-throughput CRISPR gene editing technology is modified to molecularly dissect miRNA clusters associated with human cancers (e.g., miR-888 cluster) to bridge this knowledge gap.

Bacterial CRISPR and CRISPR-associated (cas) genes mediate adaptive immunity against bacteriophages¹⁶. The discovery of this ancient procaryotic surveillance system was quickly adapted as an efficient scientific tool to easily target any desired genomic locus and make DNA sequence alterations within a large range of animal systems and cell types both in vitro and in vivo^16,17. This technique holds great promise as an effective method to interrogate miRNA networks in the context of human disease. To this end, this high-throughput CRISPR gene editing protocol to study the miR-888 cluster spanning ~35 kb on human chromosome X in immortalized human prostate cancer cell lines (LNCaP, PC3-ML) is constructed to interrogate how cluster members coordinate cancer progression pathways. This approach can be applied to characterizing any miRNA cluster and allows investigators to rapidly generate human cell lines carrying the entire miRNA cluster deletion and individual and combination miRNA gene cluster deletions within a single experiment.

In this procedure, stable cell lines are established carrying a doxycycline (DOX)-inducible lentiviral expression system that enables the investigator to control Streptococcus pyogenes CRISPR-associated endonuclease Cas9 (csn1) gene expression through the constitutive DOX-inducible promoter TRE3G. The Tet-On 3G bipartite system involves a constitutive human elongation factor 1 alpha (hEF1alpha) promoter to drive the transcription of both the Tet-On 3G gene and the blasticidin resistance gene (Blast^R) as a bicistronic transcript. The Tet-On 3G transactivator protein only binds to the TRE3G promoter in the presence of DOX, resulting in robust Cas9 transcription. In the absence of DOX, there is no or very minimal basal Cas9 expression. Therefore, the investigator can induce high Cas9 protein production in cells grown in media supplemented with DOX during the CRISPR gene editing steps and control for rapid CAS9 protein clearance upon DOX withdrawal.

This protocol also describes the design of synthetic CRISPR RNA (crRNA) oligonucleotides targeting regions flanking the entire miRNA cluster, individual miRNA hairpin (premiRNA) regions, and/or subsets of miRNA genes within the cluster. Each designed crRNA contains a unique 5'-terminal 20 nt guide sequence (complementary to the genomic sequence of interest to be targeted), followed by an invariant 22 nt S. pyogenes repeat sequence (5'-XXXXXXXXXXXXXXXXXXXX-GUUUUAGAGCUAUGCUGUUUUG-3') that enables base-pairing with the universal trans-activating CRISPR RNA (tracrRNA) oligonucleotides¹⁸. Together, the annealed crRNA and tracrRNA (mixed 1:1 ratio) function as the guide RNA for this protocol (Figure 1A). In each experiment, two synthetic guide RNAs are transfected into DOX-induced cells to associate and escort the bacterial Cas9 protein to the genomic DNA sites (5' and 3') flanking the miRNA cluster region targeted for removal (Figure 1B).

A protospacer adjacent motif (PAM) sequence (5'-NGG-3' for wild type S. pyogenes Cas9) must be present in the cell genome and located immediately adjacent to the 20 nt DNA sequence targeted by the guide RNA¹⁷. The PAM sequence serves as a binding signal and positions the catalytic region of the endonuclease Cas9 enzyme on the targeted genomic DNA site, subsequently leading to directed, double-stranded (ds) DNA cleavage located ~3 nt upstream of the PAM. The cell's DNA repair machinery repairs the cleaved DNA ends, which can result in perfect ligation, but often nonhomologous end joining (NHEJ) occurs, causing small insertions or deletions (indels) at the repair site. Since miRNAs are noncoding genes often located within intergenic and intronic regions, these indels carry a low risk of creating unwanted nonsense/missense mutations.

By employing synthetic RNA oligonucleotides (annealed crRNA and tracrRNA, 1:1 molar ratio) encoding for the guide RNA complex in these experiments, this gene knockout strategy avoids time-consuming DNA vector subcloning and allows for huge flexibility in transfecting unique guide RNA combinations to cells in a 24-well format. Preparation of crude cell lysates for PCR genotype screening also avoids expensive and time-consuming DNA purification methods, while allowing for streamlined single colony cell line generation and phenotypic analysis. Indeed, this high-throughput CRISPR gene editing protocol has been used successfully to transfect cultured prostate cancer cell lines (LNCaP, PC3-ML) with 32 unique guide RNA combinations in a single experiment and generate knockout lines carrying deletions for the entire ~35 kb miR-888 cluster region; smaller deletion combinations for miR-888 cluster members belonging to the miR-743 and miR-891a families; as well as deletions for individual miRNA members within the miR-888 cluster. Studies like these will provide a clearer understating of how clustered miRNAs function cooperatively to regulate tumor growth, aggressiveness, and drug resistance before translating miRNAs to the clinic as therapeutic and diagnostic tools.

Protocol

1. Preparation for CRISPR gene editing and guide RNA design to generate miRNA cluster knockout cell lines

1. Create a DNA file containing the complete genomic sequence of the miRNA cluster region of interest (intergenic, intronic) and at least 1 kB of surrounding genomic regions using a DNA sequence annotation software program¹⁹.
   1. Use a feature editing tool in the created DNA file to mark the targeted miRNA cluster locus (intergenic, intronic) and each individual miRNA hairpin sequence belonging to the miRNA cluster, as well as notating other nearby coding and noncoding genes and/or regulatory features^5,20,21,22.
   2. Ensure the miRNA regions targeted for deletion do not inadvertently disrupt nearby protein-coding genes, noncoding genes, or important regulatory DNA elements.
      NOTE: Consider the genomic complexity (e.g., chromosome duplications/fusions) of the cell line used in this protocol.
2. Design synthetic crRNAs targeting 20 nt of genomic DNA sequence flanking the entire miRNA cluster, individual miRNA hairpin (pre-miRNA) regions, and/or subsets of miRNA genes within the cluster using a bioinformatic guide RNA design software tool (e.g.,²³). Ensure the appropriate Cas9 enzyme is notated in the guide RNA design tool (i.e., S. pyogenes Cas9).
   NOTE: The synthesized crRNA will contain this unique 5'-terminal 20 nt guide sequence (complementary to DNA target) followed by an invariant 22 nt S. pyogenes repeat sequence to allow base-pairing with the synthesized universal tracrRNA oligonucleotide. Annealed together, they will function as the guide RNA in this protocol.
   1. Referencing the created DNA file (Step 1.1.), enter ~150 nt of DNA sequence residing immediately upstream (5') or downstream (3') of the miRNA hairpin/cluster locus targeted for deletion into the bioinformatic guide RNA design software tool²³.
      NOTE: The design tool will identify unique 20 nt sequences for crRNA oligonucleotide design that reside immediately adjacent to the PAM sequence (on the nontargeted DNA strand) (e.g., S. pyogenes Cas9 PAM sequence NGG). Provided below is an example of crRNA design targeting a site immediately upstream (5') of the mir-891a gene locus (specifically 5A(891a) discussed in the representative results section and Table 1).
      1. Open the guide RNA design software tool²³.
      2. Enter the name 5A(891a) in the Enter Name window.
      3. In the Enter a DNA sequence window, enter 150 nt of genomic sequence residing immediately upstream (5') of the mir-891a precursor gene: 5'-AAGAAAATATACATGCTGATAGTTACACAGG
         TTGTGAATAGTAAATTAGTATGTCATTTATTT
         TAGGTATTCAATGTTGCATAGTCATATTATA
         ATTACGTAGCTTCTTTGTTTTTTTCTAGGTT
         CCCAAAGAGTCTACAAATGTTGTCT-3'
      4. Check the box marked Enable Specificity Check to exclude off-targeting sites detected computationally by the program. Choose the appropriate organism (i.e., Human [Homo sapiens]).
      5. Specify the correct Cas9 enzyme PAM employed for the studies, in this case, Streptococcus pyogenes Cas9-PAM:NGG, to generate the following default settings: PAM relative to target: After; Repeat Sequence: GUUUUAGAGCUAUGCUGUUUUG; Relative to Guide: 3; Target sequence length: 20; Cut relative to target 5' start: Sense 17 Anti-sense 17.
      6. Click on the Next button to view the results for synthetic crRNA synthesis.
         NOTE: In this example, the suggested crRNA to be synthesized will recognize the DNA target ATACATGCTGATAGTTACAC and will be located adjacent to PAM:AGG.
   2. Reference the DNA file (created in Step 1.1.) to determine the best candidate crRNA 20 nt target sequences that reside closest to the miRNA locus to be deleted and possess the highest specificity for the intended genomic target.
      1. Use computational off-targeting analysis tools to evaluate the candidate crRNA specificity and predict potential off-targeting sites in genomic loci unrelated to the targeted miRNA cluster region of interest (e.g.,^24,25,26,27).
      2. Record potential off-targeting results for later reference when genotyping single-colony CRISPR clonal cell lines by PCR (Step 4.2.6.).
         NOTE: The extent of sequence complementarity shared between the designed crRNA oligonucleotide and the genomic target sequence will dictate how selectively the Cas9 enzyme cleaves the genome at that locus. Since the guide RNA anneals to the genomic target in a 3' to 5' direction, mismatches at the 5' end of the DNA target sequence (the first 8-10 bases) will often block Cas9-mediated cleavage. If the genomic target sequence shares homology with another genomic locus, except within the first eight bases of the DNA sequence, this guide RNA is predicted to have a low chance of off-targeting at this site.
      3. Design four unique crRNAs for each targeted miRNA locus: two designed crRNAs to target complementary DNA sequences immediately 5' of the miRNA hairpin/cluster (5'A, 5'B) and two designed crRNAs to target complementary DNA sequences immediately 3' of the miRNA hairpin/cluster (3'A, 3'B).
         NOTE: When performing the CRISPR transfections (in Section 3), four possible 5' and 3' guide RNA pair combinations (5'A + 3'A; 5'A + 3'B; 5'B + 3'A; 5'B + 3'B) will be tested for each targeted miRNA locus to increase the probability of generating successful miRNA locus deletions in a single experiment.
      4. Use a feature editing tool in the created DNA file (Step 1.1.) to mark the DNA target sequence (with adjacent PAM) for each designed crRNA to be synthesized.
3. Design and synthesize PCR primers (forward, reverse) flanking the targeted miRNA cluster regions for genotyping CRISPR cell lines (and label the primers in the created DNA file).
   1. For genotyping cell lines carrying small genomic deletions for closely clustered or individual miRNAs, design PCR primers (forward, reverse) residing approximately 150 bp on each side of the Cas9 cleavage site/knockout region that will enable detection of both the wildtype (~600 bp) and knockout (~300 bp) DNA fragments in a single PCR reaction via gel electrophoresis.
   2. For genotyping cell lines carrying large genomic deletions for miRNA precursor combinations or the entire miRNA cluster, again design PCR primers (forward, reverse) residing approximately 150 bp on each side of the Cas9 cleavage/knockout site. Note that, if the targeted miRNA cluster deletion spans >4 kb in length, the PCR reaction will likely only detect the smaller (~300 bp) knockout DNA fragment but not the wildtype DNA fragment. Therefore, design additional PCR primers (forward, reverse) that can detect the presence or absence of internal miRNA cluster genes to aid the screening process and validate for homozygous knockout cell lines.

2. Generation of stable lentiviral cell lines carrying the DOX-inducible Cas9 expression cassette

NOTE: Perform all cell culture and virus work in a certified biosafety hood using BSL2 procedures and aseptic technique.

1. Determine the appropriate cell line type for the CRISPR studies and preferred growth media.
   NOTE: This protocol was developed for human prostate cancer cell lines LNCaP (preferred medium: RPMI, 10% fetal bovine serum [FBS]) and PC3-ML (preferred medium: DMEM, 10% FBS).
2. Perform a blasticidin "death" curve to determine the optimal drug concentration to select for lentivirus-infected cells carrying the blasticidin resistance gene cassette (Step 2.3.).
   1. Plate cells into 11 wells of a 24-well plate and grow at 37 °C, 5% CO₂ in the preferred medium until approximately 75% confluent.
   2. Dilute the blasticidin (working stock 1 mg/mL) in the preferred medium with a final concentration ranging from 0, 1, to 10 µg/mL and diluted in 1 µg/mL increments.
   3. Label the wells of the seeded 24-well plate from 1 to 11. Remove the growth medium from the wells and replace with the appropriate blasticidin concentrations.
   4. Change the medium with the appropriate blasticidin concentrations every other day for 7-10 days.
   5. Examine the cells daily during the time course and determine the minimum blasticidin concentration required to kill all the cells within 5-7 days.
      NOTE: A blasticidin "kill" curve will need to be performed for each specific cell line employed for the CRISPR gene editing experiments.
3. Infect the cells with lentivirus carrying the DOX-inducible Cas9 expression cassette and perform blasticidin selection using the optimal concentration determined in Step 2.2.
   1. In three wells of a 6-well plate, seed 5 × 10⁴ cells per well and grow in the preferred medium at 37 °C, 5% CO₂ overnight (ON).
   2. The next day, thaw the lentivirus expression vector for DOX-inducible Cas9 (Lenti-iCas9) on ice. Gently mix (do not vortex virus particles).
      NOTE: Store the lentivirus in aliquots at -80 °C to avoid multiple freeze/thaw cycles, which will decrease viral titers and infection efficiency.
   3. Calculate the volume needed to transduce 5 × 10⁴ cells with virus at a multiplicity of infection of 0.3 (MOI = 0.3) based on the viral titer concentration.
   4. Pipet the appropriate virus volume into 250 µL (per well) of serum-free and antibiotic-free preferred medium supplemented with 0.8 µg/mL hexadimethrine bromide. Mix gently.
      NOTE: Hexadimethrine bromide is a cationic polymer found to increase viral adsorption and infection efficiency.
   5. Remove the medium. Add the 250 µL of prepared transduction medium with Lenti-iCas9 virus (MOI = 0.3) to the cells and incubate ON at 37 °C, 5% CO₂. (Shorten the incubation time to 4-6 h if the cells exhibit virus-related toxic effects.)
   6. Replace the medium with fresh preferred medium and allow the cells to recover for 1-2 days.
   7. Perform blasticidin selection on the infected cells for 10-14 days using the optimal blasticidin concentration derived from the "kill" curve described in Step 2.2.
      1. In parallel, grow the uninfected cells in medium with the optimal blasticidin concentration to be used as a positive control for viral selection.
   8. Change the medium supplemented with the optimal blasticidin concentration every 3 days during the selection time course to remove dead cells.
      NOTE: Only cells that survive blasticidin selection will have integrated the lentiviral vector.
   9. Expand the blasticidin-selected Lenti-iCas9 cell lines.
      NOTE: Record the cell passage number upon each expansion.
      1. Freeze a portion of the Lenti-iCas9 cells in preferred medium supplemented with 10% dimethyl sulfoxide (DMSO) for long-term cryovial storage.
      2. Analyze a portion of the Lenti-iCas9 cells by western blot⁹ to identify the cell line exhibiting the most robust DOX-inducible Cas9 protein levels (see Step 2.3.).
4. Perform a doxycycline concentration curve on newly established Lenti-iCas9 cell lines to determine the optimal conditions for Cas9 protein induction.
   1. Set up four independent 6-well plates for each cell line tested to perform this DOX concentration time course. Plate 5 × 10⁴ cells per well of each 6-well plate and grow at 37 °C, 5% CO₂ in the preferred medium for 24 h.
   2. Dilute DOX (working stock 1 mg/mL) in the preferred medium with a final DOX concentration curve ranging from 0, 50, 100, 150, 250 to 500 ng/mL.
   3. Label the wells of each seeded 6-well plate from 1 to 6. Remove the medium and replace with the appropriate DOX concentrations. Grow plates 1-4 until the following time points: 24 h, 48 h, and 72 h DOX induction; and 120 h post DOX withdrawal (initially 72 h DOX-induced).
   4. Harvest the wells of the plate at each time point using appropriate methods (e.g., trypsinization). Store the cell pellets at -80 °C. Process the cell pellets in RIPA buffer for lysate isolation and Cas9 western blot analysis.
      1. Run 40 µg of cell lysate for each sample of the DOX-induction time course using a denaturing 4%-12% Bis-Tris gel and transfer the proteins to an immunoblot membrane.
      2. Hybridize the immunoblot membrane with a primary antibody against Cas9 to detect the 160 kDa protein. Normalize Cas9 levels using a housekeeping gene such as GAPDH (37 kDa).
   5. Based on the western blot results, determine the optimal DOX concentration to induce Cas9 in the Lenti-iCas9 cell lines.
      NOTE: This time course will also reveal how tightly Cas9 expression is regulated upon 120 h post DOX removal.
   6. Establish single-cell colonies for the Lenti-iCas9 cell lines showing robust Cas9 protein expression to optimize the CRISPR transfections (in Section 3).

3. Performing CRISPR reactions by Cas9 induction and synthetic guide RNA transfection of cells using a high-throughput format

NOTE: A workflow diagram is shown in Figure 1.

1. On paper, map out the crRNA pair combinations that will be transfected into each well of a 24-well plate targeting the entire miRNA cluster, various clustered miRNA gene combinations, as well as individual miRNA cluster members.
   1. On the map, label four wells per targeted miRNA locus. Have each well represent a transfection reaction, with two unique crRNAs positioned 5' and 3' of the desired miRNA knockout genomic locus and the tracrRNA (annealed 1:1 molar ratio).
   2. Ensure that each of the four labeled wells represents a unique crRNA pair for transfection (e.g., 5'A + 3'A; 5'A + 3'B; 5'B + 3'A; 5'B + 3'B).
      NOTE: The design of two 5' crRNAs and two 3' crRNAs flanking each targeted miRNA locus (Section 1) enables the generation of four possible 5' and 3' guide RNA pairs in the transfection reaction.
2. Plate the Lenti-iCas9 cell line at 5 x 10⁴ cells/well of a 24-well plate in the preferred medium containing the optimized DOX concentration (determined in Step 2.4.). Grow the cells for 24-48 h at 37 °C, 5% CO₂ to induce Cas9 protein expression.
3. Transfect the Lenti-iCas9 cells with the prepared 5' and 3' guide RNAs.
   1. Reconstitute the synthesized crRNA and tracrRNA oligonucleotides with nuclease-free water to a stock concentration of 10 µM. Store at -80 °C long term.
   2. Label four 1.5 mL microcentrifuge tubes per targeted miRNA locus based on the experimental map designed in Step 3.1., representing the four possible 5' and 3' crRNA pair combinations (5'A + 3'A; 5'A + 3'B; 5'B + 3'A; 5'B + 3'B).
   3. In each tube, mix a 1:1 molar ratio of tracrRNA and unique crRNAs (5' and 3') together to form the 2 µM guide RNA complex using the following reaction (final volume of 10 µL):
      1. Add 2.5 µL of tracrRNA (10 µM stock), 1.25 µL of the 5' positioned crRNA (10 µM stock), 1.25 µL of the 3' positioned crRNA (10 µM stock), and 5 µL of 10 mM TRIS-HCL pH 7.5 to a 1.5 mL centrifuge tube. Microcentrifuge for 30 s at 16,000 × g to mix. Incubate at room temperature (RT) for 5 min.
   4. To each tube, add 40 µL of reduced serum medium (e.g., Opti-MEM) to the 10 µL of guide RNA complex reaction (50 µL total volume). Mix gently by pipetting up and down.
   5. In a clean 1.5 mL centrifuge tube, mix 2 µL of the transfection reagent²⁸ (see NOTE 3.3.5.1.) and 48 µL of reduced serum medium (e.g., Opti-MEM, final volume of 50 µL) gently by pipetting up and down. Incubate at RT for 5 min.
      1. Prepare a master mix of the transfection reagent by multiplying the reagent (2 µL) and reduced serum medium (e.g., Opti-MEM, 48 µL) amounts by the number of wells to be transfected, plus 10% extra volume to account for slight pipetting inaccuracies.
         NOTE: Select a transfection reagent that is optimized for small RNA and CRISPR RNA oligonucleotide transfections, as well as for the cell line type²⁸.
   6. To each tube containing the 50 µL of guide RNA mix (from Step 3.3.4.), add the 50 µL of the diluted transfection reagent (from Step 3.3.5.). Pipet the mixture slowly up and down once to mix. Incubate at RT for 20 min.
   7. Add 400 µL of antibiotic-free medium supplemented with DOX (optimal concentration) to each tube containing the 100 µL of the guide RNA/transfection mix. Pipet gently to mix.
4. Remove the medium from the DOX-induced Lenti-iCas9 cells (Step 3.2.). Add 500 µL of media/DOX/guide RNA/transfection mix based on the 24-well plate experimental map (Step 3.1.). Incubate the transfected cells for 48 h at 37 °C, 5% CO₂.
5. Replace the medium with the fresh preferred medium without DOX (to clear Cas9 protein from the cells). Allow the cells to recover for another 24-48 h at 37 °C, 5% CO₂.
6. Harvest the transfected cells (trypsinization) and prepare for genotyping and single-cell dilutions.
   NOTE: Cells represent a mixed population containing transfected CRISPR gene-edited and untransfected wild-type cells. This step will determine the efficiency of CRISPR editing before investing time/resources in single-cell colony expansion.
   1. Wash the cells 1x with 1 mL of phosphate-buffered saline (PBS). Incubate the cells in 100 µL of trypsin for 5 min at 37 °C, 5% CO₂ and transfer the cells to a clean 1.5 mL centrifuge tube containing 1 mL of medium.
   2. Invert to mix and microcentrifuge the cells for 5 min at 200 × g at 20 °C. Remove the medium and resuspend the cell pellet in 1 mL of PBS.
   3. Microcentrifuge the washed cells for 5 min at 200 × g at 20 °C. Remove the PBS and resuspend the cell pellet in 150 µL of the preferred medium.
7. Determine the cell number per microliter of the 150 µL of resuspended cells using a hemocytometer or automated cell counting instrument.
8. Separate the 150 µL of resuspended cells into three parts.
   1. Freeze 1/3 of transfected cells (50 µL) in medium plus 10% DMSO (final volume of 150 µL) in cryovials for future expansion/long-term storage.
   2. Transfer 1/3 of transfected cells (50 µL) into a clean 0.2 mL PCR tube for genotyping.
      1. Microcentrifuge the cells for 5 min at 200 × g at 4 °C and remove the medium.
      2. Resuspend the cell pellet in 100 µL of PBS.
      3. Microcentrifuge the cells for 5 min at 200 × g at 4 °C and remove the PBS. Store mixed-population cell pellets long term at -80 °C.
      4. Process the cell pellet for genotyping using the workflow in Section 4.
   3. Prepare the final 1/3 of the cells (50 µL) for dilution and plating in a 96-well plate format to generate single-cell colonies.
      1. Based on the cell count in Step 3.7., calculate the dilutions required to achieve 10, 5, 2, and 1 cell(s) in 100 µL of medium per well of a 96-well plate.
      2. Using a 12-channel multi-pipettor and a sterile reagent reservoir, add 100 µL of diluted cells per well to two rows of the plate (24 wells total) for each dilution (10, 5, 2, or 1 cell per well). Allow 4-6 weeks for the cells to grow to confluency for single-cell colony expansion. Observe the cells weekly under a light microscope and note if the colonies represent single- or multiple-cell colonies.
      3. Collect ~10-15 single-cell colonies for genotyping (Section 4). Identify at least three independent, knockout, single-cell colony lines for each targeted miRNA locus. Retain wild-type single-cell lines as controls.
         NOTE: The screening number will depend on the cell line type and the impact of miRNA knockout phenotype on cell growth/viability.

4. PCR genotyping of CRISPR cell lines using crude cell lysates

1. Harvest individual, single-cell colonies from near-confluent wells of a 96-well plate.
   NOTE: The removal of medium/PBS described in these steps can be performed manually using a vacuum aspirator in the biosafety cabinet, but be careful to avoid cross-contamination. Use a new sterile "yellow" 200 µL pipetman tip for each well of the 96-well plate when aspirating, in which each exchanged yellow tip is firmly placed at the end of a sterile glass pasture pipette attached to the vacuum aspirator.
   1. Wash the wells 1x with 100 µL of PBS. Aspirate the PBS.
   2. Add 50 µL of trypsin and incubate for 2-5 min at 37 °C, 5% CO₂.
   3. Gently pipet up and down and transfer the resuspended cells into a 1.5 mL centrifuge tube containing 1 mL of medium.
   4. Invert to mix and gently pellet the cells in a microcentrifuge for 5 min at 200 × g at 20 °C.
   5. Remove the medium and add 1 mL of PBS. Gently flick the tube to disrupt the pellet and invert several times to mix.
   6. Microcentrifuge for 5 min at 200 × g at 20 °C to pellet the cells.
   7. Remove the PBS and resuspend the pellet in 300 µL of fresh PBS.
   8. Split the 300 µL sample into two parts.
      1. Transfer 200 µL of the cells (2/3 of the pellet) to a well of a 24-well plate containing 1 mL of the preferred medium. Once the genotype is validated, use these cells to expand the CRISPR-mediated knockout cell lines for functional analysis.
      2. Transfer 100 µL of the cells (1/3 of the pellet) to a 0.2 mL PCR tube. Microcentrifuge for 5 min at 3,000 x g at 4 °C. Remove the PBS. Store the pellet at -80 °C.
2. Prepare crude cell lysates for genotyping and sequence analysis using the PCR primers designed in Step 1.3. Include the cell lysates prepared from untransfected cells in these experiments to use as wild-type positive controls.
   1. Resuspend the cell pellet in 4 µL of 5x DNA polymerase buffer (see the Table of Materials, final concentration 1x), 1 µL of Proteinase K (20 ng/mL stock), 1 µL of RNase A (10 ng/mL stock), and nuclease-free water up to a total volume of 20 µL.
   2. Lyse the cells in a thermocycler using a PCR program: 56 °C for 30 min and 96 °C for 5 min. Store the cell lysates at -20 °C.
   3. Perform a PCR reaction using the designed PCR primers (forward, reverse) that flank the guide RNA 5' and 3' guide RNA targeted sites (Table 1).
      NOTE: The DNA polymerase buffer and thermocycler conditions will depend on the Taq DNA polymerase enzyme used for the PCR reaction. This protocol was optimized for a Phusion HF DNA Polymerase.
      1. Set up the PCR reaction mix on ice: 5 µL of 5x DNA polymerase buffer, 0.5 µL of forward PCR primer (10 µM stock), 0.5 µL of reverse PCR primer (10 µM stock), 0.5 µL of 10 mM dNTPs, 0.25 µL of DNA Polymerase, 1-4 µL of cell lysate, and nuclease-free water up to a total volume of 25 µL.
      2. Run the PCR reaction in a thermocycler using the program: 98 °C for 3 min, and 35 cycles of 98 °C for 10 s, 62 °C for 15 s, 72 °C for 15 s, and then 72 °C for 10 min. Hold for 4 °C. See NOTE 4.2.3. above.
   4. Load the PCR products onto a 1% agarose gel for electrophoresis analysis.
   5. Extract DNA from the isolated PCR fragments of the predicted molecular size for the knockout (and wild-type) genotypes. Prepare the samples for DNA sequencing.
   6. Validate that the CRISPR reaction was successful and perform DNA sequencing of the isolated PCR fragments to identify the Cas9 cleavage site and confirm miRNA locus deletion.
      1. Isolate at least three independent knockout cell lines for each miRNA locus targeted by CRISPR. Retain wild-type (and heterozygous) cell lines to use as controls in downstream functional studies.
      2. Perform qRT-PCR or northern blot analysis to confirm that the knockout cell lines do not express mature miRNA for the deleted locus.
         NOTE: Whole-genome sequencing (WGS) is the most definitive method to validate CRISPR gene editing and off-targeting.
      3. Test for CRISPR off-targeting effects by PCR, which were predicted computationally (Step 1.2.3.)^24,25,26,27.
      4. If extensive single-cell colony screening only results in heterozygous genotypes, repeat the guide RNA transfection in the single-cell heterozygous lines.
         NOTE: Inability to obtain homozygous knockout cell lines could indicate lethal effects due to miRNA loss or inherent genomic complexity (e.g., chromosome duplications/fusions) of the parental cell line that require further analysis.

Representative Results

This high-throughput CRISPR deletion protocol was successfully employed using transfection of Cas9-inducible LNCaP and PC3-ML human cancer cell lines with synthetic oligonucleotide guide RNAs targeting the miR-888 cluster, which were studied in the context of prostate cancer. The miR-888 cluster was initially identified in an expression profiling screen as being elevated in prostate cancer patients with high-grade disease compared to low-grade and noncancer patients^9,10. The miR-888 cluster, spanning 35,124 bp, consists of seven miRNAs (miR-892c, -890, -888, -892a, -892b, -891b, and -891a) on human chromosome Xq27.3 (Figure 2A). This locus lies within the HPCX1 (Hereditary Prostate Cancer, X-linked 1, Xq27-28), which is associated with hereditary prostate cancer^{11,12,13,14,15,29}. miR-888, -890, -892a, -892b, and -892c share sequence homology and belong to the mammalian-conserved miR-743 family. Cluster members miR-891a and miR-891b only share sequence homology with one another and belong to the primate-specific miR-891 family. The genomic organization of the miR-888 cluster is conserved across primates, implying functional conservation of an important signaling network³⁰.

Previous work using experimental overexpression (miRNA mimics) and inhibition (antisense oligonucleotides) studies to evaluate individual miR-888 cluster member activity revealed an overlapping role for all the miR-888 cluster members to promote prostate cancer cell invasiveness using Matrigel Boyden chamber assays^9,10. Most notably, miR-888 and miR-891a functioned similarly to induce prostate cell growth (WST-1 assays), accelerate prostate tumor load in mouse xenografts, and regulate common targets (e.g., TIMP-2)^9,10. However, this work suggested a more complex interaction amongst the miR-888 cluster members. For example, miR-888 and miR-891a showed synergistic effects in promoting PC3-N prostate cell proliferation, whereas both miR-890 and miR-892c showed inhibitory growth effects using WST-1 assays⁹. Therefore, this high-throughput CRISPR protocol was employed to genetically interrogate the miR-888 cluster network in the context of prostate cancer.

The described workflow for CRISPR Cas9 gene editing (Figure 1) allowed the testing of 32 unique CRISPR guide RNA transfection reactions in a single experiment using a 24-well format. This protocol resulted in the successful identification of a panel of human prostate cancer knockout cell lines carrying deletions for the entire miR-888 cluster, specific miRNA cluster member combinations, and individual miRNA members. The generation of stable inducible Cas9 cell lines was first established. Briefly, PC3-ML and LNCaP human prostate cancer cells were infected with a DOX-inducible Streptococcus pyogenes Cas9 lentivirus vector (Lenti-iCas9; EditR Inducible Lentiviral hEF1aBlastCas9 Nuclease Particles; MOI 0.3) for 6 h (LNCaP) or overnight (PC3-ML), depending on inherent cell line susceptibility for virus toxicity.

The infected cell lines subsequently underwent blasticidin selection (7.5 µg/mL) and single-cell colony expansion. Western blot studies were used to select the single-colony PC3-ML and LNCaP Lenti-iCas9 cell lines exhibiting the most robust Cas9 expression based on a DOX-induction time course (Figure 2A). This lentiviral Cas9 inducible system was tightly controlled, and, upon DOX withdrawal for 120 h, most of the Cas9 protein was cleared from the cells (Figure 2A). A DOX concentration curve determined that 250 ng/mL DOX was the optimal dose to induce robust Cas9 protein expression in these Lenti-iCas9 prostate cell lines (Figure 2B). Using an online crRNA design tool²³, unique crRNAs (two 5' and two 3' crRNAs flanking each planned miRNA knockout locus to allow for four possible 5' and 3' guide RNA combinations) were designed that targeted the entire ~35 kb miR-888 cluster region; smaller cluster combinations within the miR-743 family or the miR-891 family; as well as individual miRNA deletions (miR-888, -891a) (Figure 3A). When performing the CRISPR experiments, human prostate cancer Lenti-iCas9 cell lines were grown in medium supplemented with 250 ng/mL DOX for 48 h to induce Cas9 expression.

The cells were then transfected in a 24-well plate format with the synthetic universal tracrRNA complexed (1:1 molar ratio) with a unique set of 5' and 3' genomic site-specific crRNAs (Table 2). DOX was removed from the cell culture medium 48 h post transfection to clear Cas9 protein from the prostate cancer cells. The wells were harvested 48 h later (96 h post transfection) and one-third of each harvested cell pellet was prepared for PCR genotyping using crude cell lysates (Table 3). Gel electrophoresis analysis of the PCR reactions indicated that the predicted DNA fragment size representing the knockout genotype was amplified for each CRISPR transfection (Figure 3B). DNA sequencing of these isolated knockout PCR fragments confirmed that the transfected 5' and 3' guide RNAs directed Cas9 cleavage/ligation ~3 nt upstream of the PAM sites and validated genomic loss of the targeted miRNA locus (representative example shown in Figure 4C). Since the miR-888 cluster resides in an intergenic region of chromosome X (far from any protein-coding genes), it was not anticipated that knockout cell lines, which often possessed INDELs at the Cas9 cleavage site due to NHEJ, would cause downstream artifactual effects.

Although the CRISPR transfection reactions were successful (Figure 3B), the transfected cells represented a mixed-cell population of transfected (knockout) and untransfected (wild-type) cells. This population also represented a mixture of cells carrying homozygous and heterozygous deletions for the targeted miRNA locus. LNCaP and PC3 prostate cancer cell lines are derived from males possessing abnormal karyotypes that include two X chromosomes^31,32. Therefore, a portion of each cell transfection reaction was serially diluted in a 96-well plate format to obtain single-cell colony lines. These colonies were screened by PCR to select for homozygous cell lines that lacked all copies of the targeted miRNA locus.

This procedure was important to follow in a timely manner since past data have indicated a role for the miR-888 cluster in promoting tumor cell growth and disease aggressiveness. Cancer cells carrying CRISPR deletions for the miR-888 cluster members were predicted to grow slower than wild-type cells. A valid concern was that, over time, the guide RNA-transfected wells, containing a mixed population of miRNA knockout and wild-type cells, would result in wild-type cells rapidly outcompeting the compromised mutant cells in culture. This was noted to occur and was especially pronounced in experiments using the highly aggressive PC3-ML cell lines. Thus, steps involving serial dilutions to generate single-cell colony lines or cell preparation for cryostorage were performed within 48-72 h after performing the CRISPR transfection experiments to preserve the knockout cells in the mixed population samples.

The most laborious part of the CRISPR workflow for the miR-888 cluster was the generation of homozygous, single-cell colony miRNA deletions. One challenge in this process was that the cultured LNCaP and PC3-ML cells typically did not thrive well when plated at single-cell densities. In addition, cell growth appeared to be phenotypically compromised upon combinatorial loss of miR-888 cluster members (and correlated with aggressiveness of the cell line). Therefore, for some of the transfection reactions targeting certain miRNA locus combinations, additional cell dilutions (using a 96-well format) were required to generate enough single-cell colony lines for genotyping. When screening the genotypes of the single-cell colonies by electrophoresis gel imaging, it was helpful to compare the band intensity of the wildtype compared to the knockout PCR fragments and determine if the cell line was heterozygous (equal band intensity) or represented a multi-cell colony (unequal band intensity), which would benefit from additional rounds of single-cell dilutions. If cell colonies were noted to possess PCR fragments of abnormal size and inconsistent with wild-type or knockout genotypes, these lines were eliminated from the study.

DNA sequencing of isolated PCR fragments consistent for the wild-type and knockout genotypes were revalidated for each established single-cell colony and confirmed using independent methods (i.e., qRT-PCR, WGS). In terms of the typical efficiency of generating homozygous null cell lines in these experiments, this varied with the targeted miRNA locus and the cell type used. For example, the efficiency in obtaining single-colony mir-891a knockout lines using LNCaP was 30% but dropped to ~7% in PC3-ML cells. This difference could be due to inherent cell type differences (e.g., PC3-ML cells possess higher division rates/metastatic potential than LNCaP cells). CRISPR efficiency differences could also reflect the functional impact of miRNA loss (e.g., miR-891a promotes cell growth and aggressiveness^9,10) and, thus, knockout phenotypes could be magnified in more aggressive cell lines such as PC3-ML. Obtaining single-colony lines for larger miRNA locus deletions of the miR-888 cluster exhibited reduced efficiencies. We were unable to determine if this was due to knocking out larger regions of the DNA or underscored the essential and overlapping functional roles of miR-888 members in prostate cancer cells.

Detailed description and phenotypic analysis of the single-cell colony-expanded knockout miR-888 cluster cell lines generated from this workflow are ongoing and will be published elsewhere. As a representative result of miRNA deletions generated in these CRISPR gene editing studies, the findings for LNCaP prostate cancer cells carrying an individual mir-891a deletion are shown. Single-cell colonies were genotyped by PCR and sequence verified (Figure 4A-C). Cell lines exhibiting abnormally sized PCR fragments were not analyzed (e.g., clone L14 H7 shown in Figure 4D). Once knockout and wild-type single-cell colonies were obtained for the mir-891a locus, functional studies were performed. Past work has shown that overexpression of miR-891a (miRNA mimic lentiviral vector) induces prostate cell growth⁹ and, therefore, it was predicted that miR-891a loss would show reciprocal effects. WST-1 proliferation assays comparing mir-891a knockout and wild-type cells confirmed this hypothesis, and miR-891a loss resulted in slowed growth (Figure 4D,E).

Figure 1: CRISPR gene editing workflow to generate miRNA cluster deletions. (A) The guide RNA consists of the annealed synthetic crRNA and tracrRNA oligonucleotides (mixed 1:1 molar ratio). The crRNA oligonucleotide is designed to contain a unique 5'-terminal 20 nt guide sequence (yellow) complementary to the targeted genomic DNA site and is followed by an invariant 22 nt Streptococcus pyogenes repeat sequence (green) that base-pair with the tracrRNA oligonucleotide. Red arrows indicate Cas9 enzyme double-stranded DNA cleavage sites typically located 3 nt upstream of the PAM. (B) This experimental workflow depicts a single well of a 24-well plate. Cells infected with a DOX-inducible Cas9 lentivirus vector (Lenti-iCas9, MOI 0.3) are grown in medium with DOX for 48 h to induce Cas9 protein expression. Two synthetic guide RNAs (mixed 1:1 molar ratio of crRNA::tracrRNA) are transfected into Cas9-induced cells. The guide RNAs escort Cas9 to the genomic DNA sites (5' and 3') flanking the miRNA cluster region targeted for removal. Cells are prepared 48 h post transfection for dilution using a 96-well format to generate single-cell colonies for PCR genotyping and downstream phenotypic analyses. Abbreviations: CRISPR = clustered regularly interspaced short palindromic repeats; miRNA = microRNA; crRNA = CRISPR RNA; tracrRNA = trans-activating CRISPR RNA; nt = nucleotide; Cas9 = CRISPR-associated endonuclease; PAM = protospacer adjacent motif; DOX = doxycycline; MOI = multiplicity of infection. Please click here to view a larger version of this figure.

Figure 2: Cancer cell lines carrying a doxycycline-inducible Cas9 lentiviral cassette exhibiting tight Cas9 protein regulation. (A) Western blot analysis showed that human prostate cancer PC3-ML cell lines infected with the Lenti-iCas9 vector (EditR Inducible Lentiviral hEF1aBlastCas9 Nuclease Particles, MOI 0.3) and grown for 48 h in medium containing 250 ng/mL DOX expressed optimal Cas9 protein levels. Cas9 expression was normalized to GAPDH expression. (B) PC3-ML cells grown in medium with 250 ng/mL DOX for 48 h and 120 h (+DOX) showed robust Cas9 protein expression. DOX removal from the media (-DOX) for 120 h (120 h post) resulted in Cas9 protein clearance, as measured by western blot analysis. Similar results were obtained when generating the LNCaP Lenti-iCas9 cell line (not shown). Abbreviations: CRISPR = clustered regularly interspaced short palindromic repeats; Cas9 = CRISPR-associated endonuclease; DOX = doxycycline; MOI = multiplicity of infection; GAPDH = glyceraldehyde 3-phosphate dehydrogenase. Please click here to view a larger version of this figure.

Figure 3: High-throughput CRISPR experimental design to generate multiple miR-888 cluster deletion combinations in a single experiment. (A) Diagram of the miR-888 cluster located on human chromosome Xq27.3, indicating the mammalian-conserved miR-743 family members mir-892c, -890, -888, -892a, and -892b (blue boxes) and the primate-specific miR-891 family members miR-891a and -891b (maroon boxes). Unique synthetic crRNA oligonucleotides (green arrows) were used to generate the miR-888 cluster knockout combinations. Forward and reverse PCR primers (small red rectangles) were designed to genotype cells and screen for knockouts. (B) This representative result shows 25 CRISPR reactions in transfected PC3-ML cells (Lenti-iCas9) that targeted a range of miR-888 cluster knockout combinations in a single experiment. Each transfected well of a 24-well plate contained two unique gRNAs flanking the 5' and 3' sides of the targeted miRNA genomic locus (Δ) (as listed in Table 2). Crude cell lysates were prepared for each CRISPR reaction and PCR genotyped. Isolated PCR fragments migrating at the predicted knockout sizes by gel electrophoresis were DNA-sequenced and validated for targeted Cas9 cleavage and miRNA locus removal. Predicted WT PCR fragment bp sizes are indicated in parentheses, and the desired CRISPR KO PCR fragment sizes are shown at the bottom of the gel. Abbreviations: CRISPR = clustered regularly interspaced short palindromic repeats; Cas9 = CRISPR-associated endonuclease; gRNA = guide RNA; DOX = doxycycline; bp = base pair; WT = wild-type; KO = knockout. Please click here to view a larger version of this figure.

Figure 4: Prostate cancer cells deleted for mir-891a using CRISPR gene editing exhibiting suppressed proliferation. (A) Design of PCR primers (purple) and the two 5' and two 3' guide RNAs (orange) targeting genomic sequences flanking the mir-891a gene (green). (B) PCR genotyping of transfected LNCaP cells with indicated 5' and 3' guide RNAs confirmed that Δmir-891a deletions were obtained for all four guide RNA combinations. Cell lysates were made from mixed-cell populations 48 h post transfection. (C) PCR fragments were sequenced and validated. A representative example shows the DNA sequence of transfected cells treated with 5A(891a) and 3B(891a) guide RNAs, resulting in the predicted Cas9 cleavage site (3 nt upstream of the PAM site) and CRISPR Δmir-891a deletion. (D) Single-cell colonies were obtained and PCR-genotyped. Clone L14 G9 was sequence-validated as a KO and clone L14 G9 as a WT for the mir-891a gene. (E) miR-891a was previously shown to induce prostate cancer cell growth and, therefore, miR-891a loss was predicted to have the reciprocal phenotype. WST-1 assays of prostate cancer cells deleted for mir-891a (L14 G9 [KO], blue) showed suppressed proliferation compared to WT cells (L14 G9 [WT], red). Abbreviations: CRISPR = clustered regularly interspaced short palindromic repeats; Cas9 = CRISPR-associated endonuclease; WT = wild-type; KO = knockout; WST-1 = water-soluble tetrazolium salt-1. Please click here to view a larger version of this figure.

PCR reaction is set up on ice.
Component	25 µL Reaction	Final Concentration
5x DNA polymerase Buffer	5 µL	1x
Forward PCR Primer (10 µM)	0.5 µL	0.2 µM
Reverse PCR Primer (10 µM)	0.5 µL	0.2 µM
10 mM dNTPs	0.5 µL	200 µM
DNA Polymerase (2 U/µL)	0.25 µL	0.125 U/ 25 µL
Crude cell lysate	1 - 4 µL	variable
Nuclease-free water	up to 25 µL
Thermocycler conditions for PCR reaction:
Step	Temperature	Time
Initial denaturation	98 °C	3 min
	98 °C	10 s
35 cycles	62 °C	15 s
	72 °C	15 s
Final extension	72 °C	10 min
Hold	4 °C

Table 1: PCR reaction conditions for crude lysate genotyping. Abbreviation: dNTP = deoxynucleotides.

CRISPR-targeted Deletion Site	Guide RNA Combination	Wild-type (bp)	Knockout (bp)	PCR Primer Pairs
∆Full Cluster	5A(891a)+3A(892c) 5B(891a)+3A(892c) 5A(891a)+3B(892c) 5B(891a)+3B(892c)	35,664	389 496 378 485	891a FWD 892c REV
∆892c/890/888	5A’(888)+3A(892c) 5B’(888)+3A(892c) 5A’(888)+3B(892c) 5B’(888)+3B(892c)	2,739	478 487 467 476	888 FWD 892c REV
∆892c/890	3A’(888)+3A(892c) 3B’(888)+3A(892c) 3A’(888)+3B(892c) 3B’(888)+3B(892c)	2,739	710 754 699 743	888 FWD 892c REV
∆892b/892a	5A’(888)+5A(892b) 5B’(888)+5A(892b) 5A’(888)+5B(892b) 5B’(888)+5B(892b)	2,938	554 547 573 566	892b FWD 888 REV
∆892b/892a/888	5A(892b )+3A’(888) 5A(892b )+3B’(888) 5B(892b )+3A’(888)	2,938	322 278 341	892b FWD 888 REV
∆743 Family	5A(892b)+3A(892c) 5B(892b)+3A(892c) 5A(892b)+3B(892c) 5B(892b)+3B(892c)	5,015	370 389 359 378	892b FWD 892c REV
∆891 Family	5A(891a)+3A(891b) 5B(891a)+3A(891b )	27,294	330 270	891a FWD 891b REV
∆891a	5A(891a)+3A(891a ) 5A(891a)+3B(891a ) 5B(891a)+3A(891a ) 5B(891a)+3B(891a )	554	333 273 454 394	891a FWD 891a REV

Table 2: CRISPR guide RNAs used to generate miR-888 cluster knockout combinations. Abbreviations: CRISPR = clustered regularly interspaced short palindromic repeats; bp = base pair; FWD = forward; REV = reverse.

Primer	Sequence	Tm (°C)
888 FWD	AGGCATCCTCAAGATAAGTTAAG	52.2
888 REV	CTGGATGATGGCCAAGACAAG	55.9
891a FWD	TGGGAAGTCGGAATTCTTACAA	53.9
891a REV	TTCATCTTGCTGCTACCTGTC	54.8
891b REV	TCATCTTGCTGCTATACGTCCT	55.6
892b FWD	GCTCCTGTCAATCATCTAGGTA	53.5
892b REV	CTCTATGTTTACCCAGTGATTGC	53.5
892c FWD	TGTGAGCACCTACAGGTTAG	53.9
892c REV	CACTCTCACTTGGCTTCATG	53.5

Table 3: PCR primers used for genotyping. Abbreviations: FWD = forward; REV = reverse; Tm = melting temperature.

Discussion

This CRISPR gene editing procedure allows the investigator to quickly generate an entire panel of cell lines carrying unique miRNA cluster deletion combinations. The transfection of synthetic guide RNAs composed of 5' and 3' genomic site-specific crRNAs annealed with synthetic tracrRNA (1:1 molar ratio) in this protocol avoids time-consuming plasmid vector subcloning and allows for a more flexible and high-throughput experimental design using a 24-well format. The generation of cell lines carrying a DOX-inducible Cas9 lentiviral cassette enables tightly DOX-regulated and robust Cas9 expression during the cell transfection of guide RNAs and rapid clearance of this bacterial protein upon DOX withdrawal from the culture medium in subsequent steps. This procedure also relies on crude cell lysates for PCR genotyping, which requires minimal cell material for analysis and avoids the use of costly and labor-intensive silica column DNA purification methods or the need to measure nucleic acid concentrations using a spectrophotometer, further streamlining the methods for the investigator.

Indeed, the representative results shared here showed successful adaptation of this method to transfect human prostate cancer cell lines with 32 unique guide RNA combinations in a single experiment and generate multiple single-cell colony knockout lines carrying deletions for the entire ~35 kb miR-888 cluster region; smaller deletion combinations for miR-888 cluster members belonging to the miR-743 and miR-891a families; as well as deletions for individual miRNA members within the miR-888 cluster. This workflow generated a valuable cell line resource to begin to molecularly dissect the miR-888 cluster network in the context of human prostate disease and determine how miR-888 cluster members functionally overlap or interact synergistically/antagonistically with one another to regulate prostate tumor growth, cancer aggressiveness, and drug resistance (to be published elsewhere). These studies will be crucial as this research moves toward diagnostic and therapeutic applications for treating patients with advanced and metastatic forms of prostate cancer.

There are several critical steps in this high-throughput CRISPR gene editing protocol that investigators should carefully consider when adapting this method to study additional miRNA cluster networks in other disease contexts. First, the most important step in this process is the careful design of the guide RNAs that flank the miRNA cluster for CRISPR-medicated removal, as well as additional guide RNAs to delete individual miRNAs or combinations of miRNA cluster members. The investigator should place careful consideration to not disrupt the surrounding genes or regulatory elements, especially if the miRNA cluster resides within a gene intron or is situated antisense to another gene. There are CRISPR off-targeting prediction resources^24,25,26,27 available that can aid the researcher when choosing the most discriminating crRNA oligonucleotides to synthesize. This workflow describes the design of two 5' and two 3' crRNAs flanking each targeted miRNA locus that allow four unique guide RNA combinations to be tested in the CRISPR transfection step and increase the chances of generating the desired knockout cell line. However, the investigator only needs one of these guide RNA combinations to work successfully to disrupt a particular miRNA gene locus.

Second, the investigator should decide the most appropriate method for delivering Cas9 to the cells for CRISPR gene editing. This procedure used the DOX-inducible Cas9 lentiviral expression system, but alternative approaches for Cas9 protein production (e.g., Cas9 recombinant protein, Cas9 mRNA delivery, or the use of conventional Cas9 expression plasmid vector systems) can easily be adapted for this protocol. For example, experimental considerations could dictate the use of Cas9 transient expression versus generating stable lentiviral expression cell lines, the promoter used to drive Cas9 (ubiquitous, cell type-specific, inducible), and the incorporation of positive/negative selection systems to ensure Cas9 expression vector delivery to the cells of interest (i.e., antibiotic resistance).

Third, it is important to quickly genotype mixed-population transfected cells within 48-72 h after guide RNA delivery in the event that loss of the miRNA cluster may negatively impact the viability/growth of the cells. For instance, when deleting tumor-promoting factors such as miR-891a in prostate cancer cell lines, it was noted that these mir-891a knockout cells grew slower than untreated cells and, subsequently, the wild-type cells quickly overtook the cell population. Finally, it is crucial to integrate additional validation methods to ensure miRNA locus removal during the genotypic screening process, especially for larger genomic deletions. These measures include using internal PCR genotyping controls that are only detected if the wild-type allele is present and performing qRT-PCR analysis on single-cell colonies to ensure that knockout cell lines are not expressing the targeted miRNAs. All these considerations will ensure successful outcomes for the investigator.

There were some obvious limitations encountered when incorporating this CRISPR technique into the laboratory. For instance, despite the high-throughput nature of the guide RNA transfection steps described in this protocol to quickly generate multiple miRNA cluster deletion combinations, the rate-limiting step of this procedure was the genotype screening and expansion of single-colony cell lines for each CRISPR-mediated miRNA deletion. The incorporation of positive selection strategies into this protocol would be an advantage; however, it would still take ~3-4 weeks to expand a line from a single-cell colony grown in a 96-well format. Indeed, the collection of at least three independent single-colony cell lines per miRNA locus knockout plus wild-type cells will help to ensure that the phenotypes assayed are not due to artifact/off-targeting effects, but this again adds to the time-consuming nature of this procedure. The investigator may also need to perform multiple rounds of CRISPR gene editing to obtain homozygous null cell lines. This is especially pronounced in studies using established immortalized cancer cell lines that often possess a complex genomic landscape of chromosomal rearrangements and abnormal karyotypes.

For example, this representative study used LNCaP and PC3-derived human prostate cancer cell lines to generate deletions for the miR-888 cluster, mapping onto the X chromosome. It was important to recognize that these male-derived prostate cancer cell lines have abnormal karyotypes and possessed two X chromosomes (and PC3 cells also have partial X chromosome translocations on other autosomes)^31,32. Despite these conditions, CRISPR gene editing was robust enough to generate homozygous knockouts using these prostate cell lines. However, there may be cases when an investigator's cell line with an abnormal karyotype (e.g., insertions, deletions, inversions, duplications) could harbor "invisible mutations" that will obscure the detection of second (or more) copies of the targeted gene locus due to missing elements that are required for successful detection using the designed PCR primers/guide RNAs. This underscores the importance of verifying the homozygous null CRISPR cell lines using independent methods (i.e., qRT-PCR, northern blot, WGS) before performing functional assays. Lastly, this protocol is designed to only knock out miRNA cluster combinations that map adjacent to one another. If an investigator wishes to knock out miRNAs that are separated between other miRNA genes, then multiple rounds of CRISPR transfections will be needed, requiring careful genotyping and validation.

Despite these challenges, the ability to quickly generate a complete panel of miRNA deletion combinations for any given miRNA cluster using this described protocol is a huge advantage over the existing methods in the miRNA research toolbox. Overexpression studies using miRNA mimics to characterize noncoding RNA function can lead to unintentional artifacts or cell toxicities. Likewise, the use of antisense antimir oligonucleotide inhibitors to determine loss-of-function phenotypes can be difficult since antimirs usually result in a knockdown and not a complete elimination of miRNA activity. Attempting to use combinations of miRNA mimics or miRNA antimirs to elucidate miRNA cluster network signaling has proven laborious and difficult to interpret. Therefore, the incorporation of CRISPR gene editing technology toward interrogating how miRNA members interact within a given noncoding genomic cluster will enable researchers to gain a clearer understanding of how noncoding RNAs influence disease signaling pathways and promises to have huge implications in the diagnostic and drug discovery fields.

Materials

Name	Company	Catalog Number	Comments
0.2 mL PCR tubes, flat cap	Fisher	14-230-225	Plasticware
1.5 mL Microcentrifuge Tubes	Seal-Rite	1615-5500	Plasticware
24-well tissue culture plate	Corning Costar	09761146	Plasticware
5x Phusion HF Buffer	Thermo Scientific	F-518L	PCR reagent, genotyping
6-well tissue culture plate	Fisher	FB012927	Plasticware
96-well tissue culture plate	Falcon	08-772-2C	Plasticware
Anti-CRISPR-Cas9 antibody [7A9-3A3], Mouse monoclonal	AbCam	ab191468	Western blot reagent; 160 kDa; dilution 1:200
Antibiotic-Antimycotic (100x)	Gibco	15240062	Tisuue culture reagent
BLAST nucleotide search engine	National Center for Biotechnology Information	NA	Freeware; website: blast.ncbi.nlm.nih.gov
Blasticidin S, Hydrochloride, Streptomyces griseochromogenes	Calbiochem	CAS 589205	CRISPR reagent; Chemical, Working stock = 1 mg/mL in water
Countess 3 Automated Cell Counter	Invitrogen	A50298	Equipment; Cell counter
Cryogenic tubes	Thermo Scientific	50001012	Plasticware
Dharmacon CRISPR Design Tool	Horizon Discovery Ltd.	NA	Freeware; website: horizondiscovery.com/en/ordering-and-calculation-tools/crispr-dna-region-designer
DharmaFECT siRNA Transfection Reagent #2	Dharmacon, Inc.	T-2002-02	Transfection reagent; LNCaP and PC3-ML cell lines
Dimethyl sulfoxide (DMSO)	Fisher	BP231100	Tisuue culture reagent
DMEM Medium with L-Glutamine, 4.5g/L Glucose and Sodium Pyruvate	Corning	MT10013CV	Tisuue culture reagent; Media for PC3-ML cells
Doxycycline Hydrochloride, Ready Made Solution	Sigma-Aldrich	D3072-1ML	CRISPR reagent; Chemical, Working stock = 1 mg/mL stock in water
DPBS, no calcium, no magnesium	Gibco	14190144	Tisuue culture reagent
Edit-R CRISPR-Cas9 Synthetic tracrRNA, 20 nmol, designed	Dharmacon, Inc.	U-002005-20	CRISPR reagent; Universal tracrRNA oligonucleotides
Edit-R Modified Synthetic crRNA, desalted/deprotected, 2 nmol	Dharmacon, Inc.	crRNA-460XXX	CRISPR reagent; Designed crRNA oligonucleotides
EditR Inducible Lentiviral hEF1aBlastCas9 Nuclease Particles, 50 μL, 107 TU/mL	Dharmacon, Inc.	VCAS11227	CRISPR reagent; Lenti-iCas9; Doxycycline-inducible lentiviral Streptococcus pyogenes Cas9 vector system
Ensembl genomic viewer	Ensembl	NA	Freeware; website: [ensembl.org] Use: Genome browser to identify a nucleotide seuqnces containing the miRNA cluster and surrounding gene/regulatory sequences.
GAPDH Antibody (FL-335), rabbit polyclonal	Santa Cruz Biotechnology	sc25778	Western blot reagent; 37 kDa; dilution 1:500
Gel/PCR DNA Fragment Extraction Kit	IBI Scientific	IB47010	Nucleic acid gel electrophoresis reagent
Gibco Fetal Bovine Serum, certified	Gibco	16000044	Tisuue culture reagent
Hexadimethrine bromide	MilliporeSigma	H9268	CRISPR reagent; Chemical, Working stock = 0.8 mg/mL
Immobilon-FL PVDF Membrane	MilliporeSigma	IPFL10100	Western blot reagent; nitrocellulose
Intercept (TBS) Blocking Buffer	LI-COR	927-60001	Western blot reagent
IRDye 680RD Goat anti-Mouse IgG Secondary Antibody	LI-COR	926-68070	Western blot reagent
IRDye 800CW Goat anti-Rabbit IgG Secondary Antibody	LI-COR	926-32211	Western blot reagent
Microcentrifuge	Eppendorf	5425 R	Plasticware
miRBase microRNA viewer	miRBase	NA	Freeware; website: [mirbase.org] Use: Borowser lists all annotated miRNA hairpins, mature miRNA sequences and associated clustered miRNAs mapping within 10 kb.
NuPAGE 4 to 12%, Bis-Tris, 1.0 mm, Mini Protein Gel, 10-well	Invitrogen	NP0321BOX	Western blot reagent
Odyssey CLx Imaging System	LI-COR	CLx	Equipment; Western blot imaging
Opti-MEM Reduced Serum Medium	Gibco	31985062	Transfection reagent
Owl D2 Wide-Gel Electrophoresis System	Owl	D2-BP	Equipment; Nucleic acid gel electrophoresis system
PCR Primers, designed (single-stranded DNA oligonucleotides)	Integrated DNA Technology	NA	PCR reagent, genotyping
Phusion High-Fidelity DNA Polymerase (2 U/µL)	Thermo Scientific	F530S	PCR reagent, genotyping
RIPA Lysis Buffer 10x	MilliporeSigma	20-188	Western blot reagent
Proteinase K Solution (20 mg/mL), RNA grade	Invitrogen	25530049	PCR reagent, genotyping
RNase A, DNase and protease-free (10 mg/mL)	Thermo Scientific	EN0531	PCR reagent, genotyping
RPMI 1640 Medium	Gibco	MT10041CV	Tisuue culture reagent; Media for LNCaP cells
SnapGene Viewer	Snap Gene	NA	Freeware; website: [snapgene.com] Use: DNA sequence annotation software program to create a DNA file for gRNA design and which  highlights the miRNA cluster locus (intergenic, intronic), each individual miRNA hairpin sequence belonging to the miRNA cluster, and other nearby coding and non-coding genes and/or regulatory features
TrypLE Select Enzyme (1x), no phenol red	Gibco	12563029	Tisuue culture reagent; Recombinant trypsin
UltraPure Agarose	Invitrogen	16500100	Nucleic acid gel electrophoresis reagent
Veriti 96-Well Fast Thermal Cycler	ThermoFisher Scientific	4375305	Equipment
XCell SureLock Mini-Cell Electrophoresis System	Invitrogen	EI0001	Equipment; Protein gel electrophoresis system