In vivo combinatorial knockout screens using CRISPR-Cpf1

Genetic interactions have long been challenging to investigate systematically. Recently, CRISPR-Cas9 based approaches have been described that can dissect genetic interactions in a high-throughput manner. However, these existing approaches have important disadvantages that limit their flexibility and scalability, hindering their application to in vivo screens. Here, we develop MCAP (massively-parallel CRISPR-Cpf1 crRNA array profiling), an approach for combinatorial interrogation of double knockouts in vivo. This protocol accompanies Chow et al. Nature Methods, “In vivo profiling of metastatic double knockouts through CRISPR-Cpf1 screens”. higher-order genetic

(also known as Cas12a) has notable advantages over Cas9, as Cpf1 can autonomously process a single array containing multiple crRNAs and subsequently induce multi-site mutagenesis 10-13. In the associated publication, we describe an approach for combinatorial knockout screens using massivelyparallel CRISPR-Cpf1 crRNA array profiling (MCAP). Compared to prior Cas9-based combinatorial screens, MCAP is unparalleled in its simplicity of oligo design, library cloning, and downstream analysis, thereby facilitating its application to the study of complex processes in vivo. " Figure   1":http://www.nature.com/protocolexchange/system/uploads/7575/original/protocolExchangemetastasis. We anticipate the MCAP system to be of use to the scientific community for the study of two-gene or even higher-order genetic interactions.

4.
Anneal the forward and reverse oligos with the following reaction: 1. Identify crRNA spacer sequences as described above for all genes to be represented in the library. Include control crRNA spacers that are non-targeting, or targeting a known control locus (i.e. noncoding regions or gene deserts). As a rule of thumb, the control crRNAs should ideally comprise at least 10% of the total library size in order to accurately model the null distribution. Since the cutting efficiencies of different crRNAs is quite variable, it is also essential to design multiple independent crRNAs targeting each gene.

2.
Here, we describe barcoded MCAP library design for double knockout studies, but we note that triple knockout studies are also feasible using this approach. Design MCAP library oligos according to the schema detailed below, where the "core sequence" is defined as Spacer1-DR-Spacer2-DR, and "revComp" denotes the reverse complement. For each possible gene pair, randomly assign the constituent genes to either the first or second crRNA position. In our experience, the position of the crRNA within the array does not significantly influence Cpf1 cutting efficiency.

5.
Run the PCR product on a gel. The product should be 210 bp. Purify with the QIAquick Gel Extraction Kit.

6.
Clone the purified PCR product into the appropriate vector by Gibson assembly according to manufacturer instructions. Use 330 ng of the digested vector and 50 ng of the PCR-amplified library pool. 14.

15.
Merge the raw paired-end fastq read files to single fastq files by PEAR with the settings -y 8G -j 8 -v 3.

16.
Filter and demultiplex the merged fastq files using Cutadapt, using two different sets of adapters for extraction of crRNA array sequences or the 10mer barcode. For the crRNA array, use the following settings: cutadapt --discard-untrimmed -g tcttGTGGAAAGGACGAAACACCg, followed by cutadapt --discard-untrimmed -a TGTAGATTTTTTT.
For the 10mer degenerate barcodes, use the following Cutadapt settings: cutadapt -discard-untrimmed -a AAGCTTGGCGTGGATC, followed by cutadapt --discarduntrimmed -g TACTAAGTGTAGATTTTTTT. 18. Quantify the resultant sequences to a reference of all possible 10mer sequences. Tabulate all reads that successfully mapped to both the MCAP library and contained a valid barcode.

1.
Grow HEK293FT cells to 80% confluency in 15-cm tissue culture dishes. One hour prior to transfection, change media to Opti-MEM.

6.
Once the lentiviral titer has been determined, infect the target cells at an MOI of 0.2, aiming for at least 1000x library coverage. As an example, for a library size of 10,000 arrays, infecting 1x108 cells at MOI of 0.2 will lead to ~2000x coverage.

7.
The following day, add 3 μg/mL of puromycin. Continue culturing the cells in puromycincontaining media for 7 days to ensure complete selection.

3.
As the forward PCR primers used to readout crRNA array representation were designed to have a variety of barcodes to facilitate multiplexed sequencing, demultiplex these filtered reads with the following settings: cutadapt -g file:fbc.fasta --no-trim, where fbc.fasta contains the 12 possible barcode sequences within the forward primers noted in Table 1.

5.
Remove the 5' DR sequence as follows: cutadapt --discard-untrimmed -e 0.1 -g TAATTTCTACTAAGTGTAGAT -m 80. 7. For data processing on the level of barcoded-crRNAs, we utilized the same trimmed fastq files as above, but instead used the barcoded-crRNA plasmid library (described above) as the reference Bowtie index.

8.
Tabulate the number of reads that had mapped to each crRNA array within the library using the Bowtie mapping output. Normalize the number of reads in each sample by converting raw crRNA array counts to reads per million (rpm).

9.
To analyze the data at the level of barcoded-crRNAs, convert the counts per barcoded-crRNA in each sample to percentages of total reads.

10.
To identify crRNA arrays that are enriched in individual samples, utilize the control crRNA arrays to model the empirical null distribution. Enriched crRNA arrays can be defined as those exceeding the abundance of the "top" control crRNA.

1.
Having identified potential genetic interactions from the primary tumors and/or metastases, validation experiments using individual crRNA arrays can be performed.
Use the oligo-cloning protocol described above to clone in the specific crRNA arrays of interest.

2.
The in vivo tumor model can be performed similarly. We inject 1.8x106 KPD cells per flank, as the cells all contain the same crRNA array and do not necessitate larger cell numbers to accommodate library diversity.

3.
Proceed with tumor growth measurements and tissue dissection procedures as done previously.

T7E1 assays
1. 7 days following the initial lentiviral infection, harvest the puromycin-resistant cells and isolate genomic DNA using the QIAamp DNA Mini Kit.

2.
Design PCR primers to amplify target loci from genomic DNA around cutting site.
Input the crRNA spacer sequence into NCBI Blat, then click "View->Get DNA." Retrieve the 600 bp of DNA sequence upstream and downstream of the crRNA cut site within the genome.