Identifying essential long non-coding RNAs in cancer using CRISPRi-based dropout screens

Summary Long non-coding RNAs (lncRNAs) are emerging as key regulators in the initiation, growth, and progression of cancer. High-throughput CRISPR-based techniques systematically assess the function of genes or regulatory elements present in the human genome. Here, we present a protocol for identifying essential lncRNAs in cancer using CRISPRi-based dropout screens. We describe steps to select target sites, design guide RNAs, and generate CRISPRi cell lines. We then detail the execution and analysis of CRISPRi-based dropout screens.


Target identification
This section describes the capture of transcriptional changes associated to specific biological contexts of interest by performing Global run-on sequencing (GRO-Seq).We generally apply GRO-seq to monitor the transcriptional regulation of lncRNAs in response to the disruption of specific transcription factors in a cellular context of interest, and subsequently assess the function of identified deregulated transcripts by CRISPRi-based dropout screens.While this method can be used to identify new essential molecules related to specific transcriptional programs, it can also be employed to uncover key transcriptional changes following specific treatments (e.g., chemotherapy) or to characterize specific cellular states (e.g., chemo-sensitive vs. chemo-resistant cells).Alternatively, GRO-seq can be used to identify all expressed lncRNAs in a cellular context of interest.

CRISPRi screen
Here, we describe the design of gRNAs to target the transcription start sites (TSSs) of selected lncRNAs, using the GRO-Seq data generated in the first part of this protocol.In addition, control gRNAs, which are crucial for downstream analyses, must be included.Among the various controls, we systematically include gRNAs targeting essential (coding) genes (positive controls) that have been previously described and that are expressed in the selected cell system.We also include non-targeting scrambled gRNAs (negative control group 1) that do not align to any genomic sequence in the used model system.A group of gRNAs targeting non-essential 1 or non-expressed genes (negative control group 2) can also be added.Control gRNA sequences are provided in Table S1.
Prior to performing a CRISPRi-based dropout screen, plasmid vectors expressing both KRAB-dCas9 and the gRNAs must be selected.Special attention should be given to ensure complementary selection markers (fluorescent proteins and/or resistance genes).In addition, if a doxycycline (dox) inducible system is preferred, the amount of dox required to adequately control the desired expression in your cell system of interest should be thoroughly tested.
Dropout screens will generally be analyzed over several time points (e.g., day 0, 10 and 20).The ideal harvesting time points may vary between different cell models, which are likely to present distinct doubling times and sensitivities.Therefore, we recommend harvesting multiple time points over a period of 2-3 weeks to ensure the capture of dropout hits.

MATERIALS AND EQUIPMENT STEP-BY-STEP METHOD DETAILS
Target identification Selection of cell system and nuclei isolation An adequate cell system or cell lines of interest should be initially selected.Any cell line that can be expanded in vitro is in principle suited for this protocol.Our laboratory mainly focuses on colorectal cancer and we, therefore, established this procedure using a panel of CRC cell lines (e.g., Colo320HSR, LS180, HT55 and LS411N).Each cell line is then profiled using GRO-seq.To capture nascent transcripts, millions of cell nuclei are isolated and transcription is reinitiated in presence of an analog of UTP (Br-UTP).Nascent transcripts with incorporated Br-UTP can be captured with an agarose-conjugated anti-BrdU antibody (Santa Cruz, IIB5).This section describes how to isolate cell nuclei, perform and analyze GRO-seq.Steps related to the GRO-Seq protocol are illustrated in Figure 2A.
CRITICAL: RNA is sensitive to temperature and enzymatic digestion by RNases, which are normally present on skin, surfaces, etc. Keep all surfaces and equipment clean while performing this protocol.We also recommend to wear gloves and to keep samples on ice whenever indicated in the protocol.Only use RNase-free materials and chemicals.
Note: Here we describe the procedure from the day cells are harvested for nuclei isolation.We do not consider seeding density, expansion time and possible treatments.However, if you have different treatments or conditions in your experimental setup, these factors should be carefully considered in the overall planning.For cancer cell lines, we generally harvest 20 3 10 6 cells per replicate and combine two run-on reactions (each with 5 3 10 6 nuclei) in order to obtain sufficient material for the preparation of a sequencing library.However, if the transcriptional output of your cells of interest is low, you may have to increase the number of pooled run-on reactions per replicate.

Day 1
1. Prepare the swelling, lysis, and freezing buffers and pre-cool them on ice.2. Cool down a centrifuge to 4 C and pre-warm the trypsin.3. Trypsinize the cells until they become rounded and floating.
a. Neutralize the trypsin with 2 volumes of 10% FBS-supplemented medium and transfer each sample to a 50 mL tube.Make sure that cells are properly singularized.b.Centrifuge samples for 5 min at 300 3 g, and discard the supernatants.c.From here on, keep the pellets on ice unless indicated otherwise.Note: For each replicate, we recommend to perform two run-on reactions (each with 5 3 10 6 nuclei).If your harvest did not yield enough nuclei for two run-on reactions, you can repeat the previous steps or control the efficacy of your nuclei isolation by taking along cell lines with high nuclei isolation yield such as HCT116 and HEK293.

Timing: 1 h
Here we describe the use of global run-on sequencing to identify expressed or regulated lncRNAs in a cell line of interest.Compared to conventional RNA-Seq, GRO-Seq captures nascent transcripts, which improves the detection of less stable and lowly transcribed RNAs (including non-polyA RNAs). 9,10In addition, this method allows the precise mapping of transcription start sites (TSSs).This feature is particularly important since current CRISPRi tools achieve best transcriptional regulation when gRNAs are targeted to these regions.
8. Preheat a thermo-mixer to 30 C. 9. Prepare the reaction buffer according to      Note: You can prepare buffers before performing this protocol.However, Tween20-containing buffers should be kept in the dark, and RNase-inhibitors as well as DTT should be freshly added before use.Note: It is recommended to run a small amount of each RNA sample (0.8 mL) on a Bioanalyzer instrument (Agilent) to visually inspect the quality of the run-on, before proceeding to the sequencing library preparation.An example of two run-on reactions analyzed with a Bioanalyzer (Eukaryote Total RNA, RNA 6000 Pico Chip) are shown in Figures 2C and 2D.To avoid RNA degradation, we recommend proceeding immediately to the 1 st strand synthesis step.
52. Proceed with the preparation of the sequencing library according to the manufacturer's manual.

Note:
We generated the libraries for sequencing using the TruSeq Stranded mRNA Library Prep Kit (Illumina).Using a stranded sequencing kit is essential to determine strand-specific transcriptional activities.Moreover, we quantified libraries with the NEBNext Library Quant Kit (New England Biolabs).Barcoded samples were pooled equimolarly and sequenced on the Illumina HiSeq4000 platform (single-end 50 bp reads).

Timing: 2-3 h per sample
In this section, we process the GRO-Seq data to identify expressed or differentially regulated genes.Troubleshooting 2 & Troubleshooting 3.

CRITICAL: The analysis was executed and tested on an Ubuntu (version 20.04) virtual environment via Windows Subsystem for Linux (WSL, version 1).
Before you start: Download the output files from the GRO-Seq and install python.Using a package manager like anaconda with useful pre-installed packages (including jupyter notebook) is highly recommended.Create a new directory called ''CRISPRi'' for all the data ($DIR from now on).Download the ''Cleanup and guide picking.html'',''Ordered csv.html'', and ''Merge files.html''files (Supplementary files) and place them in the CRISPRi folder.
53.In order to identify low quality regions in the generated reads, install and run the tool FastQC.
Note: In general, reads classified with a green flag are high-quality reads and will be used for the GRO-seq analysis.In some cases, the reads classified with yellow flags require further attention.In particular, rRNA contamination and other quality issues should be analyzed, and if the quality of the reads is too low, re-sequencing is necessary.
54.To remove low-quality sequences identified in the previous step, download and install the tool Trimmomatic. 5te: The command line code used in this step will depend on the issue identified by the FASTQC tool.Therefore, we recommend to carefully read the documentation of Trimmomatic and correct the reads accordingly.
55.To map high-quality reads to the human reference genome with the STAR aligner tool, 2 download necessary files to generate the reference genome: Note: The NRSA_guide.txt file includes information about the download and installation of the NRSA tool.The NRSA perl scripts Pause_PROseq.plwill be used for calculating the differential expression of expressed genes.The NRSA contains processed reference files from various organisms.In this protocol, we are using the processed version of the GRC38 reference files which are the same files used for mapping the GRO-seq reads.Ensure that the version of the genome used by NRSA is consistent with the one to which the reads were aligned to in the previous step.
59. To generate output files, execute the following steps sequentially.

Note:
In this example, the mapped reads from two different conditions were named 'condi-tion1.bam'and 'condition2.bam'.In the case of multiple BAM files corresponding to different replicates, all BAM files should be added and separated by a space (e.g.-in1 condition1_ rep1.bamcondition1_rep2.bam-in2 conditions2_rep1.bamcondition2_rep2.bam).
60. Obtain the output files gb_change.txt,which contain transcriptional changes of genes .
Note: These output files can be imported into R, Python, and Microsoft Excel for further analysis.Depending on your experimental design, different cutoffs such as minimum read coverages or specific fold changes (control vs treatment) can be used to define a list of candidate lncRNAs.

Visualizing expression data and defining target regions
Timing: 10-20 min per file Here, we briefly outline the steps to visualize custom expression data on the UCSC genome browser.We further detail how to define and target genomic regions of interest.
Note: Hereafter, we only briefly describe the generation of custom expression data tracks on the UCSC genome browser.Detailed documentation can be found in the User's Guide.Alternatively, the protocol using bigWig files can be found here.
61. Remove any ''track'' or ''browser'' lines from your .wigfiles output obtained in the previous step.62. Generate bigWig files from your .wigfiles using the binary utilities fetchChromSizes and wigTo-BigWig from the UCSC utilities.63.Host the bigWig files on a http(s) or ftp accessible location (e.g., github).64.Generate the tracks with the UCSC protocol for bigWig files linked above.
Note: Here, we describe how to select genomic regions of interest for efficient silencing with the CRISPRi system.Note: In Figure 3, the displayed transcript is encoded on the sense strand (red).The TSS can be defined by locating the beginning of the read coverage for your gene of interest.While annotations can help to pinpoint a region, significant variations can be observed between previously annotated and observed TSS in your specific cell system.When annotations and GROseq reads are not perfectly aligned, we recommend to use your sequencing reads to determine the TSS.In addition, most TSSs will also display some transcription (initiation peak) on the opposite strand (here: blue), which is helpful to pinpoint the right region.
67.Following the ''Zoom in'' on the region of interest, the DNA sequence is retrieved by clicking on ''View'' and ''DNA'' (highlighted).Note: Following these steps, you will have a CRISPRi folder filled with .txtfiles of all candidate TSS sequences (À50 nts to +350 nts = 400 nt total), named by their ENSEMBL Gene-IDs.

Download a human genome file by opening up a new terminal window and running:
71.Then, to install CCTop, run: 72.The installation of BLAT is slightly complex and uses a C compiler to make and make install the packages.Note: Make sure to substitute all the correct file paths and to adapt the ''preseq'' and ''postseq'' variables according to the homology arm sequences present in your selected gRNA vector.

Stable genomic integration of KRAB-dCas9
Timing: 5 days Here we describe the transduction of cell lines to stably integrate an inducible CRISPRi system.In this protocol, we use an inducible KRAB-dCas9 system or the TRE-KRAB-dCas9-IRES-BFP plasmid (Addgene #85449).This construct allows for dox-dependent expression of KRAB-dCas9 and BFP, which enables a more controlled setting as compared to constitutively active promoters.The doxdependent induction of KRAB-dCas9 and BFP requires the co-expression of the reverse tetracycline transcriptional activator (rtTA).Stable integration and constitutive expression of rtTA is achieved using pHR-EF1Alpha-Tet-on 3G (Plasmid #118592).Cell populations that properly induce KRAB-dCas9/BFP (cells co-transduced with KRAB-dCas9/BFP and rtTA) are enriched by FACS sorting.
For an overview of all preparatory steps, see Figure 6A.
CRITICAL: This protocol assumes a basic understanding of the production of lentiviral particles (and associated biosafety procedures) and 2D cell culturing.A general protocol to generate lentiviral particles can be obtained here.

Note:
The amount of cells depends on the growth rate and size of your cells.
91. Thaw the pre-generated viruses (filtered 0.45 mm) OR harvest fresh lentiviruses containing KRAB-dCas9/BFP and rtTA.a. Add 1 mL of virus to the cell suspension and polybrene at a final concentration of 4 mg/mL.b.Centrifuge the plate for 30 min at 500 3 g (32 C). c. Place the plate in a cell culture incubator for 14-18 h at 37 C.
Note: In our experience, using 1mL of lentiviral supernatant (not concentrated) is sufficient to obtain a transduction efficiency that ranges between 1% and 10%.However, concentrating lentiviral supernatants may be necessary to efficiently co-transduce certain cell types.

Day 2
Timing: 30 min 92.Refresh the medium.Note: In this protocol, we use 1 mg/mL of dox in the culture medium to induce the CRISPRi system.However, different cell lines may show different sensitivity to doxycycline, which may require titration experiments to establish ideal experimental conditions.

Day 7
Timing: 2-3 h 96.Trypsinize and collect cells into a 15 mL tube.a. Spin down for 5 min at 500 3 g.b.Remove the supernatant and resuspend the cell pellet in 1-2 mL of medium (depending on pellet size).c.Prepare a 15 mL collection tube for sorting with 2 mL of medium.
Note: In this protocol, we performed cell sorting using the SH800 Cell Sorter (Sony).Settings may differ if other cell lines and other FACS machines are used.An untreated control sample (no dox) is essential to properly gate BFP+ cells.97.Determine the gating.An exemplary gating strategy is depicted in Figures 5B-5F.Note: Depending on the transduction efficiency and proliferation rate of your cell line, dox-induction and sorting steps may be performed with an alternative or better suited time-schedule.Alternatively, single cell (BFP+) clone could be sorted and expanded.However, molecular features of single cell clones may not properly represent the complexity or heterogeneity of the initial population.5F).
Note: We recommend testing the functionality of your CRISPRi cell line, by targeting an essential gene.This allows to validate the transcriptional repression of a target gene and assess potential cellular phenotypic changes.

Setting up and performing a CRISPRi dropout screen
In this section we discuss how to perform a CRISPRi-based dropout screen with adherent cells.An overview of the CRISPRi-dropout screen is shown in Figure 6.
The gRNA library is first cloned into a suited backbone vector, which allows for the generation of lentiviral particles.Then, cells containing the inducible KRAB-dCas9-BFP are transduced with lentiviral particles containing the gRNAs.This should be performed at a low MOI to minimize the presence of multiple integrations per cell.This section describes the titration to obtain a low MOI.
CRITICAL: This section assumes a basic understanding of cloning.We followed the protocol by Wang et al. 12 to clone our gRNA library into a vector backbone and do not further describe these steps here nor are these considered in the estimated timing.We assume a general understanding of lentiviral particles production (and associated safety procedures) and 2D cell culturing.Troubleshooting 4.
Note: We used the pDECKO-mCherry backbone (Addgene #78534) to clone our gRNA library, a vector that contains the mCherry fluorescent protein as well as a puromycin resistance gene for selection.We use the live cell labeling with mCherry to evaluate the transduction efficiency by flow cytometry and to adjust the conditions (e.g., cell density and amount of virus) to obtain a low MOI.Then, we select transduced cells by adding puromycin.Selection of cancer cell lines is usually completed within 2 weeks.We highly recommend using both the antibiotic-based selection as well as a fluorescent marker to easily control and complete these steps.Prior to transduce CRISPRi cell lines, the pDECKO-gRNA library DNA preparation can be PCR amplified (see section ''Preparing sequencing libraries'') and sequenced to validate the presence of each gRNA (optional).

Day 1
Timing: 1 h Note: Number of cells may vary depending on the features (e.g., size and proliferation) of your cells of interest.In general, 70%-80% seeding confluency will yield robust and reproducible results.We usually generate a big volume of viral particles containing the gRNA library to avoid multiple titrations and minimize possible batch effects.We filter, aliquot and store the virus suspension at À80 C. Lentiviral particles can be stored several months without significant impacts on the transduction efficiency.
103.Add different amounts of virus per flask, e.g., 100, 400 and 800 mL of virus.104.Add polybrene at a concentration of mg/mL to each flask.105.Place the flasks for 14-18 h at 37 C in a cell culture incubator.

Day 3
Timing: 2-3 h 107.Harvest and collect cells in tubes suited for flow cytometry analysis.

Note:
We acquire the samples on the SH800 Cell Sorter (Sony) to assess transduction efficiencies.Settings may differ if other cell lines, plasmids and FACS machines are used.
108.Setup the gating strategy as mentioned in step 97.
a. Visualize mCherry+ cells by using the 561 nm laser.An untransduced control sample is required to properly gate mCherry+ cells.b.Examples of transduction efficiencies following the incubation of CRISPRi-engineered cells with an increasing number of lentiviral particles harboring the gRNA library (mCherry+) is shown in Figure 7. Troubleshooting 7.

Transduction and selection of the gRNA library
Here we detail the steps leading to the transduction and selection of cells harboring the gRNA library.
The library is transduced with low MOI (<0.5) and cells are selected with the appropriate marker (e.g., Puromycin or Blasticidin).The screen is usually initiated with a gRNA coverage of 5003 (e.g., 1,000 gRNA library requires >500,000 cells) for each replicate and in presence or absence of dox.We recommend harvesting several time points to follow dropouts over time.
CRITICAL: To ensure a good representation of each gRNA, we usually transduce the cells of interest with a 2003 coverage.Screens can also be initiated with a coverage of at least 2003.However, we do recommend using a coverage of 5003 or more if possible.

Note:
The required number of flasks should be determined based on the size of your library and the titration of your lentiviral suspension.We usually aim to cover every gRNA 200 times with a transduction efficiency varying between 1% to 10%.
110.Thaw pre-generated virus.111.Add an adequate amount of virus per flask to obtain a transduction efficiency ranging between 1% to 10% based on previous titration (e.g., 800 mL would achieve $5% mCherry+ cells in Figure 7).112.Add polybrene at a concentration of 4 mg/mL to each flask.113.Place the flasks for 14-18 h at 37 C in a cell culture incubator.

Day 6
Timing: 1 h 115.Maintain and expand your cells at an appropriate confluency.
a. Transfer cells into a bigger vessel when they reach confluency.

Day 7
Timing: 1 h 116.Start the selection of transduced cells by adding puromycin to the culture medium.
Note: For CRC cell lines, we used concentrations ranging between 1 and 5 mg/mL of puromycin, depending on their sensitivity.Alternatively, determine the amount of puromycin needed to kill untransduced cells by performing a dose-response experiment.Selection of cancer cell lines is usually completed within 2 weeks.Here we describe the conditions and steps required to initiate and complete a CRISPRi dropout screen.
CRITICAL: We initiate our CRISPRi screens with >5003 coverage and recommend to use at least 2003 for each replicate.We also advise to freeze a stock of cells containing your gRNA library that contains at least a 2003 coverage per vial.However, attention should be paid during the thawing step, as important cell death may impact the complexity and therefore skew the representation of gRNAs.

Day 1
Timing: 1-2 h 120.Harvest all the library-transduced cells and collect them in a 50 mL falcon.
a. Count the concentration of cells.
Note: For better accuracy, ensure that trypsinized cells are fully singularized before counting.Note: Depending on the size of your library or properties of your cell line, it might be necessary to seed cells in multiple plates per replicate.We recommend performing 3 replicates in parallel with and without dox to obtain a robust statistical analysis of the screen.

Timing: 2 days
To prepare sequencing libraries, we PCR-amplify gRNAs (PCR1) from gDNA obtained from collected samples.In a second amplification step (PCR2), adapters and indices are added to perform standard Illumina multiplex sequencing.PCR primers used to generate sequencing libraries are shown in Table S2.STAR Protocols 4, 102588, December 15, 2023 Protocol Note: We do not describe the steps required for the library quantification.However, a detailed protocol can be found online at the manufacturer's page: NEBNextâ Library Quant Kit for Illuminaâ.
133.Extract the genomic DNA from each cell pellets with Genomic DNA Purification Kit or a similar product.134.PCR-amplify gRNA sequences using PCR1 primers, which are specific to the vector backbone.
Note: To properly amplify the complexity of a library by PCR, each gRNA should be represented by at least 200 genomes (one integration per genome).Considering that the human diploid genome is approximately 6.5 pg, then a library of 10,000 gRNAs requires to PCRamplify $13 mg of gDNA (200 3 10,000 3 6.5pg).
135.PCR-amplify amplicons from PCR1 with PCR2 primers, in order to add standard Illumina adapters.

Note:
The second PCR is also used to add indices to samples, thus allowing multiplex sequencing.For details, see Table S2.
136.Purify amplicons obtained from PCR2 from the gel using the NucleoSpin Gel & PCR Clean-up kit or similar.137.Quantify the concentration of all libraries, using the NEBNextâ Library Quant Kit for Illuminaâ.138.Dilute samples to the desired concentration (usually between 20-50 nM) and pool all libraries (if multiplex sequencing is performed).
Note: To improve accuracy, we recommend an initial dilution of libraries at a higher (23) concentration (40-100nM).Quantification is then repeated on the ''23'' diluted samples before the final dilution.
139.Sequence all samples using next-generation sequencing.
Note: Using single-end 50 bp sequencing reads (SE50) is sufficient to retrieve your gRNA sequences.We also recommend adding 5% phix and to cover each gRNA with at least 200 reads.

Screen analysis
Timing: 30 min to install and prepare, and 30-60 min per 30 million reads (for step 140) Timing: 30-60 min per 30 million reads (for step 147) Here we describe how to analyze the sequencing data obtained from a CRISPRi dropout screen.We use the MAGeCK package to count the reads and DESeq2 package for subsequent analyses.Specific cutoffs and filters are also applied to the data to produce a list of dropout candidates.Troubleshooting 2 & Troubleshooting 3.
CRITICAL: The analysis is executed and tested on an Ubuntu (version 20.04) virtual environment via Windows Subsystem for Linux (WSL, version 1).
Before you start: Download the .fastqfiles generated by the dropout screen, and place them in a convenient directory (''$DIR'' from now on).If you performed paired-end sequencing, only the forward reads (_R1 or similar) are necessary, therefore place the reverse reads (_R2 or similar) in a separate folder in $DIR.Download and install R (The Comprehensive R Archive Network (r-project.org)) on your system.We highly recommend R studio for writing and executing R codes.
Prior to obtaining the read counts for each gRNA, adapter sequences flanking the gRNA insert must be trimmed.The following steps are adapted from the documentation of Trimmomatic.
140.To install Trimmomatic, execute the following commands in a terminal window: Note: In order to run Trimmomatic, Java should be installed on your system.If Java is not present, run: 141. Identify the gRNA flanking sequences by running the following code on one of your .fastqoutput files (''$FILE'') Example output: Note: Adapter sequences before (yellow) and after (green) the gRNA are highlighted for clarity.
142. Count the length of the adapter that prepends the gRNA inserts (in this case, 22 nucleotides).
143. Pre-process the reads by running the following command.Labeling: ''seqID'' is an identifier for the guide; ''gRNAs'' is the guide sequence; ''targetgene'' is the gene targeted by the gRNA.
159.Make the row names for each target gene identical a. Count all the occurrences of identical genes in a new column called 'good_sgRNAs'.b.Filter the data frame to only include genes which have more than 1 'good sgRNA' c.Save it as a .csv in $DIR: 160.This should output a list of all genes for which at least 2 gRNAs drop out with an adjusted p-value of < 0.05 and an LFC < À1.0.

EXPECTED OUTCOMES
The major outcomes of this protocol are lncRNAs that are likely essential for cell proliferation or survival.However, further experiments to validate screening hits and explore their modes of action will still be required.This method provides a fast, high-throughput and efficient way to test hundreds to thousands of putative candidates, making it ideal to identify essential transcripts.

LIMITATIONS
Dropout screens are mostly limited to the identification of transcripts influencing cell proliferation or survival in a 2D culture system.Of note, CRISPRi allows epigenetic silencing of a targeted locus.While the aforementioned strategy maximizes specific targeting, nearby genomic regions may also be influenced and yield ''false'' positive dropouts.Therefore, we recommend validating any hits with an alternative silencing method (e.g., siRNAs or shRNAs).

TROUBLESHOOTING Problem 1
Generation of output matrices: GRO-Seq reads mostly cover ribosomal RNAs.

Potential solution
One critical step in the GRO-seq procedure reside in the capture of nascent transcripts (harboring Br-UTP) with agarose-conjugated anti-BrdU beads.To reduce rRNA contamination, extra washing steps can be added to the proposed procedure.Alternatively, anti-BrdU pull-down can be performed twice by repeating steps 20 to 51.Certain cell lines have low transcriptional output and, thus, changing the selected cell line could also help to reduce the rRNA contamination.write.csv(as.data.frame(candidates),file='$DIR/candidates.csv'')

Figure 1 .
Figure 1.Workflow to assess the function of expressed/selected lncRNAs with CRISPRi dropout screens The schematic illustrates essential steps for each section of the protocol, including 1-Target identification, 2-CRISPRi screen and 3-Screen analysis.

4 .
Resuspend each pellet in 30 mL of swelling buffer.a. Incubate samples 10 min on ice with occasional inverting of the tubes.b.Centrifuge samples for 5 min at 500 3 g (4 C). c. Carefully remove the supernatant without disturbing the pellet.5. Resuspend each pellet in 10 mL of lysis buffer and transfer the suspension into a 15 mL tube.a.Invert the tube 30 times and spin down for 5 min at 500 3 g (4 C). b.Remove the supernatant.6. Repeat step 5.

Figure 2 .
Figure 2. Global run-on sequencing (A) Scheme highlighting key steps required to perform global run-on sequencing.(B) Example of nuclei obtained from the CRC cell line HCT116, following the procedure illustrated in A1. (C) Analysis of two purified run-on reactions (S1 and S2) using a Bioanalyzer RNA 6000 Pico Chip.The size distribution is indicated by a ladder (L).(D) Electropherograms of samples presented in B. Fragment sizes (nucleotides, nt) are plotted against arbitrary fluorescence units (FU).
To minimize the loss of RNA, use low binding 1.5 mL collection tubes for the following steps.24.Pre-cool a tabletop centrifuge to 4 C. 25.Spin down the BSA-blocked beads for 30 s at 500 3 g.a. Remove the supernatant.b.Resuspend each bead pellet in 500 mL of bead binding buffer.26.Centrifuge RNA samples for 30 min at maximum speed (4 C).A small blue pellet should be visible after the centrifugation step.a. Remove the supernatant without disturbing the pellet.27.Add 800 mL of 75% ethanol to each RNA pellet and vortex the samples for 5 s.a. Spin down samples for 5 min at maximum speed (4 C). b.Remove the supernatant and let the pellets dry with an open lid on the bench.c.Remove residual traces of liquid by carefully flicking the tube.
37. Spin down the beads for 30 s at 500 3 g.a. Carefully remove the supernatant without disturbing the pellet.38.Start the washing steps by adding 500 mL of Low-salt buffer per sample, and place tubes on a rotator for 5 min at 20 C. a. Spin down samples for 30 s at 500 3 g and remove the supernatant.b.Repeat step 38.

39. 3 Timing: 2 - 4 h
Add 500 mL of High salt buffer per sample, and place tubes on a rotator for 5 min at 20 C. a. Spin down samples for 30 s at 500 3 g and remove the supernatant.b.Repeat step 39. 40.Add 500 mL of Tris-EDTA-Tween buffer per sample, and place tubes on a rotator for 5 min at 20 C. a. Spin down samples for 30 s at 500 3 g and remove the supernatant.b.Repeat step 40.41.Add DTT to the Elution buffer and keep at 20 C. 42.Prepare low binding tubes for the elution step.43.Add 150 mL of Elution buffer to each sample and incubate on a rotator for 5 min at 20 C. a. Spin down samples for 30 s at 500 3 g.b.Transfer each eluate to a low binding tube.c.Repeat the elution (step 43) two more times and combine eluates of the same sample (total volume $450 mL).44.Pre-chill a tabletop centrifuge to 4 C. 45.Add 500 mL of acid phenol-chloroform to each sample.a. Vortex samples for 30 s. b.Centrifuge samples for 5 min at 12000 3 g (4 C). c. Transfer 400 mL of the aqueous phase (top) to a new low binding collection tube.d.Discard tubes containing the bottom phase.46.Add 400 mL (13 volume) of chloroform.a. Vortex samples for 30 s. b.Centrifuge samples for 5 min at 12000 3 g (4 C). c. Transfer 400 mL of the aqueous phase (top) to a new low binding collection tube.d.Discard tubes containing the bottom phase.47.Add 1 mL of Glyco Blue to each sample.a. Vortex vigorously for 10 s.Note: Here we use less Glycogen as compared to the previous steps to avoid inhibition of the reverse transcriptase.48.Add 1100 mL (2-33 volume) of 100% ethanol per sample.a. Vortex vigorously for 15 s. b.Incubate samples for 14-18 h at À20 C. Day 49.Pre-chill a tabletop centrifuge to 4 C. 50.Spin samples for 30 min at maximum speed (4 C). a. Remove the supernatant.b.Dry pellets and remove residual liquid around the pellet.Note: You should see a small, light blue pellet.If there is no pellet visible, the yield is probably too low to continue.51.Carefully add 6 mL of H 2 O (supplemented with 1 mL of SUPERaseIN per 10 mL) to each pellet and incubate tubes for 2 min on ice.a. Vortex samples and spin down to collect the liquid at the bottom.b.Keep samples on ice.
65. Open the output matrix obtained from the GRO-Seq data, and use the ENSEMBL gene-IDs to visualize the regions on your UCSC track.A visual example is depicted in Figure 3. 66. Zoom in on the TSS region and define a 400 nucleotides (nts) stretch by selecting 50 nts upstream and 350 nts downstream of the transcription initiation signal.An example is shown in Figure 4.

Figure 3 .
Figure 3. Visualization of GRO-seq data with the UCSC genome browser UCSC genome browser screenshot showing lncRNA transcripts (green\blue annotations) and sequencing tracks of 2 cell lines (track 1 and 2).Read coverage on the sense (red) and antisense (blue) strands are shown as separate tracks.The presence of H3K27Ac (transcriptionally active chromatin) is also displayed.Important features to select the region of interest are highlighted; transcription start site peaks (green box), transcript annotations (black box), and repeat sequences (orange box).

Figure 4 .
Figure 4. Defining and retrieving target sequences (A) Zoomed in view showing the selection of a genomic region encompassing 50 nucleotides before and 350 nucleotides after the transcription start site.The ''View'' button is highlighted in green.(B) The ''View'' dropdown menu, with the correct DNA option selected.

Day 3- 5 Timing: 30 min 93 . 6 Timing: 5
Transfer cells into a bigger vessel when the confluency reaches >90%.94.Maintain and expand cells at an appropriate confluence in at least 2 separate cell culture vessels.Enriching cells harboring the inducible CRISPRi systemTiming: Several weeks This section describes the selection of transduced cells by fluorescence activated cell sorting (FACS).Cells co-transduced with TRE-KRAB-dCas9-IRES-BFP plasmid (Addgene #85449) and EF1Alpha-Tet-on 3G (Plasmid #118592) are dox-treated and BFP+ cells are sorted.To gradually enrich cells that properly induce KRAB-dCas9/BFP, multiple FACS sorting cycles are required.In each cycle, cells are dox-treated, FACS sorted (BFP positive cells sorted) and cells expanded (without dox).Enrichment cycles are performed until the vast majority of cells (>95%) can induce the BFP expression in a dox-dependent manner.In addition, leaky cells can be removed by sorting out BFP positive cells in absence of dox.Note: Our experimental setup uses a construct that allows the selection of cells by a fluorescent marker.Other constructs may contain distinct selection markers and may require a different selection approach.Day min95.Add dox to the culture medium (final conc. 1 mg/mL).a.Incubate the cultures at least 24h to induce the BFP expression.b.Keep one cell culture vessel without dox to use as control sample for the FACS.
a. To gate living cells, use the forward (FSC) and backward scatter (BSC, equivalent to side scatter SSC) area (A).b.Select singlets by using an FSC-height and width density plot.c.BFP+ cells are sorted using the 405 nm laser.98. Sort the BFP+ cell population into the prepared collection tube.a.Transfer sorted cells back into an appropriate cell culture vessel depending on the number of sorted cells.and expand the sorted cells accordingly.

Figure 6 .
Figure 6.CRISPRi dropout screen layoutThe cartoon shows key steps required to perform a CRISPRi dropout screen.Following the transduction of a gRNA library in a CRISPRi cell line (KRAB-dcas9 BFP cells), cells are treated or not with dox and collected at various time points (d0, T1 and T2).Sequencing is performed to evaluate the relative abundance of each gRNA across collected samples.

Figure 7 .
Figure 7. Titration of viral particles harboring the gRNA library CRISPRi-engineered cells were either untransduced (ctrl) or transduced with an increasing amount of lentiviral particles (100, 400 and 800 ml).The transduction efficiency (percentage of mCherry+ cells) is displayed for each condition by plotting the mCherry signal (x-axis) against the BrilliantViolet signal (y-axis).

121.
For the ''day 0'' samples, collect in 3 different tubes (3 replicates) a number of cells corresponding to at least 2003 the size of the gRNA library.a. Centrifuge the cells for 5 min at 500 3 g and remove the supernatant.b.Store cell pellets at À20 C. 122.For the screen, seed an adequate number of cells and culture vessels to reach a library coverage of at least 2003 for each replicate.a. Choose ideal culture vessel and cell seeding density to accommodate 4 days of cell expansion without reaching confluency (>95%).

3 Timing: 10 min 124 . 5 Timing: 2 - 3 h
123.Add dox to a final concentration of 1 mg/mL (or use the ideal dox concentration for your cell line as mentioned above) to the 3 induced (+) replicates.a. Maintain the other 3 replicates untreated (-).b.Label the culture vessels appropriately.c.Place them back in a cell culture incubator at 37 C.Day Refresh the medium of all culture vessels.a. Add dox (1 mg/mL) to induced (+) culture vessels to maintain a sufficient expression of KRAB-dCas9.DayNote: To maintain viable conditions for the cells, we replate cultures every 4 days over the course of the screen.125.Trypsinize cells.a. Collect each replicate in separate and labeled 50 mL tubes.b.Count the cells.c.Seed back cells into new culture vessels as performed in step 122.d.Add 1 mg/mL dox to the medium of induced (+) culture vessels.126.Repeat step 124 every 2 days and step 125 every 4 days until the first time point (T1) is reached.Day X: First time point (T1) Timing: 2-4 h 127.To collect timepoint samples, trypsinize cells.a. Collect each replicate in separate and labeled 50 mL tubes.b.Count the cells.128.For each replicate, transfer a number of cells corresponding to at least 2003 the size of the gRNA library to a labeled 15 mL tube.a. Centrifuge cells for 5 min at 500 3 g.b.Remove the supernatant.c.Freeze the pellets at À80 C. 129.If another time point (TX) is harvested, seed back cells into new culture vessels as performed in step 122.130. Repeat step 124 every 2 days and step 125 every 4 days until the next time point (TX) is reached.131.Day Y: Next time point 132.Perform steps 127 to 130.

Table 2
and preheat to 30 C. 10.If you have stored the nuclei at À80 C, thaw them on ice.

Table 1 .
Buffers for nuclei isolation 11. Add 100 mL of pre-warmed reaction buffer(13 volume)to each reaction tube with 100 mL of nuclei and keep tubes at 20 C. a. Cut off the tip of a 200 mL pipette tip (the mixture is very viscous) and pipette up and down each reaction mixture.Change the tip between each sample to avoid contamination.Pipette carefully to avoid air bubbles.b.Incubate samples for 5 min at 30 C. c. Pipette the reaction mixtures up and down (using cutoff tips) after 2-3 min of incubation, while keeping the samples at 30 C. Pipette gently and avoid making air bubbles.12. Add 600 mL (33 volume) of Trizol LS to each reaction tube and mix samples by vortexing.a. Incubate samples for 5 min at 20 C.

Table 2 .
Buffer to perform run-on reactions

Table 3 .
Buffers for bead preparation

Table 4 .
Buffers for RNA purification and capture of nascent transcripts Composition of buffers to purify and capture nascent RNA transcripts.
Here, we generate a genome with lncRNA annotations only.If you would like to analyze coding genes as well, repeat steps 55 and 56 with protein coding annotation files, which will generate a protein coding reference genome.57.Align reads to the reference genome and generate a bam file containing all mapped reads as output.Troubleshooting 1.
The output should consist of 3 files per input candidate: a .bedfilewith gRNA locations along your input sequence, a .fastafilewith all designed gRNA sequences, and a .xlsfilewith extra information, including the CRISPRater score for each gRNA.82.Create new directories for the different output files and move all files to these directories.In addition, the hg38.fafile is required for BLAT and should therefore be moved to the fasta folder as well: If you are running Ubuntu via WSL, add the command -no-browser.88.Now you should have an organized .csvfile in the ''Clean'' folder for every candidate, as well as a BLAT .pslfile in the ''fasta/Blat/'' folder.89.Create a new folder called ''Final'' in the CRISPRi main folder and run the code from ''Cleanup and guide picking.html'' in jupyter notebook.
If you don't have an existing C compiler on your computer, run: 73.Download and install a library that BLAT uses: 74.Go back to $DIR, download the BLAT source code archive, unpack it, and move all the library files to the /lib/ folder within the BLAT folder: 75.Finally, you have to configure some variables to allow BLAT to run from the terminal: wget http://downloads.sourceforge.net/project/libpng/libpng16/older-releases/1.6.2/libpng-1.6.2.tar.gzcp ./libpng-1.6.2/pnglibconf.h./blatSrc/lib/STAR Protocols 4, 102588, December 15, 2023 Protocol 76.Run BLAT from the terminal by typing blat in the terminal.77.Open the ''.profile'' file in your /home/username/ directory in notepad and add the following line at the end to ensure you can always use blat from the terminal, and use bash loops: 78.Create a bowtie index file from the human genome for CCTop.This should output multiple .ebwtfiles.Execute the following code in $DIR: 79.Run CCTop to design CRISPRi gRNAs on one of our candidate sequences (''$FILE.txt''):80.You can also run all the files automatically by running: 81.bowtie-build -r -f hg38.fahuman export PATH=$PATH:$/bin/$MACHTYPE cctop --input $FILE.txt--targetSize 19 --index human for SAMPLE in *.txt; do cctop --input ${SAMPLE} --targetSize 19 --index human; done sudo mv *.fasta ./fastasudo mv *.xls ./xlssudo mv hg38.fa./fastaSTAR Protocols 4, December 15, 2023 Protocol Note: Protocola.Replace $FILE with your filenames b.Replace the numbers following ''-threads'' and ''HEADCROP:'' with your specific number of cores and the length of the adapter determined in step 142, respectively:144.Run all the files automatically by executing a bash loop in the terminal.

Table 5 .
Example of a MAGeCK library file