In less than a decade, the CRISPR-Cas system has become an indispensable part of functional genomics screening. Through an ever-expanding CRISPR toolkit, researchers can dissect complex gene-phenotype relationships and delve deeper into disease processes, potentially uncovering novel therapeutic targets along the way.
However, it’s increasingly clear that genes operate in dynamic networks wherein redundant, synergistic, and antagonistic genes collaborate to produce robust phenotypes. Attempts to identify the function of any of the network’s genes may be foiled by its ability to compensate and maintain homeostasis. Therefore mapping the structure of gene networks remains a considerable challenge, one that substantially limits our understanding of clinically relevant processes (such as malignant transformation, development of drug resistance, and cellular plasticity).
To overcome this challenge, researchers from New York University and the New York Genome Center have developed Cas13 RNA Perturb-seq (CaRPool-seq), a new approach to CRISPR screening that leverages both single-cell sequencing and a combinatorial perturbation strategy (in which multiple genes are simultaneously manipulated within each cell). In effect, CaRPool-seq allows researchers to begin unraveling gene networks and the processes they support.
The Challenges of Combinatorial CRISPR Screening
To identify the interplay between multiple genes in a network, researchers need the ability to simultaneously perturb its components, both individually and in various combinations. It is possible to do this with Cas9, however, it’s not practical. Common inefficiencies in gRNA delivery and editing mean that, on average, approximately 20% of cells in a traditional screen will fail to express detectable levels of the appropriate gRNA. This inefficiency further compounds with each additional gRNA that needs to be delivered. Therefore the number of cells in each screen that receive the correct number and combination of gRNAs can be prohibitively low and greatly limits the scale of combinatorial screening.
Many alternatives to Cas9 have been developed, including RfxCas13d (for simplicity, we refer to it as Cas13d). Like Cas9, Cas13d is an RNA-guided nuclease that can be used to target specific genes. However, Cas13d has several unique qualities that make it particularly well-suited for combinatorial screening.
🧬 Synthesis of crRNA Arrays
It is notable that building crRNA arrays—which can approach 300 nucleotides (nt) in length—is far from trivial. Most oligonucleotide synthesis platforms lose accuracy at these longer lengths (with most performing optimally at 150nt or shorter in length). Inaccuracies can be highly problematic in a CRISPR screen as they can lead to inefficient target binding and mistargeting, both of which can produce false positive or false negative results.
Wessels et al. ordered crRNA arrays from Twist Bioscience, whose DNA synthesis platform is uniquely capable of generating pools of oligonucleotides up to 300nt in length with a high degree of accuracy (incurring only 1 error for every 3000 bases).
Unlike Cas9, Cas13d targets single-stranded RNA. Therefore it is typically used for gene knockdown, which reduces the expression of a gene rather than gene knockout, which eliminates a gene altogether. Cas13d also uses a much smaller guide RNA known as a CRISPR RNA (crRNA) to target specific transcripts. Given their smaller size, several distinct crRNAs can be delivered to a cell through a single vector. Once transcribed, this tandem series of crRNAs is known as an array. Cas13d is capable of processing this array and producing mature crRNAs, each ready to guide the nuclease to distinct targets.
Therefore using a crRNA array and Cas13d allows researchers to deliver and effectively edit multiple genes within a cell, all without the compounding inefficiencies that plague Cas9-based approaches. Put another way, Cas13d is ideal for combinatorial screening.
High-Content, Combinatorial CRISPR Screening with CaRPool-seq
Whereas traditional CRISPR screening approaches rely on simple end-points, such as cell death or proliferation, recent advances have made it possible to assess ever more complex phenotypic readouts. Perturb-seq, for example, leverages single-cell RNA sequencing to link gene perturbations to their elicited transcriptomic changes. Such detailed analyses allow researchers to gain a more nuanced understanding of a gene’s function and its potential role within broader gene networks.
Key to performing these analyses is being able to identify which perturbations each cell received—a difficult task when it comes to pooled combinatorial CRISPR screening.
Want to learn more about single-cell CRISPR screening? Watch this webinar to hear about the value of moving beyond simplistic readouts.
In a recent study published in Nature Methods, Wessels et al. described a new screening approach that allows for combinatorial CRISPR screening in a pooled format with single-cell readouts. The approach, known as CaRPool-seq, delivers an array of tandem crRNAs to cells that stably express Cas13d. Appended to the end of this array is a PCR primer binding site and a unique non-targeting barcode sequence. During data collection, sequencing of the PCR amplified barcode indicates which combination of crRNAs each cell received, thus making it possible to perform combinatorial perturbations with single-cell analyses.
CaRPool-seq stands out for its efficiency. In a series of experiments, the authors showed that nearly 70-80% of screened cells expressed a detectable barcode, regardless of the number of crRNAs in the array (0-3 tested). A benchmarking analysis was done to compare CaRPool-seq performance with Cas9-based approaches. In contrast to CaRPool-seq, successful delivery of gRNAs—meaning the right number in the right combination—occurred in only 49.6% of cells, likely reflecting the compounded difficulty of delivering three gRNAs to the same cell.
🤔 Pooled vs Arrayed screening
One of the highest-level divisions between CRISPR screening strategies concerns pooled vs arrayed screens.
In a pooled screen, gRNA libraries are delivered in bulk to a population of target cells, such that each cell in the study may carry a gRNA that differs from its neighbor’s. This approach is often favored for large-scale applications where the number of gRNAs far exceeds the practical limitations of array-based screens. Because the sgRNAs are randomly dispersed throughout the population in a pooled screen, however, researchers need to devise a way to link gRNAs with observed phenotypes.
Arrayed screens, on the other hand, deliver sgRNAs to specific, isolated populations of cells (often separated into different wells on a plate). As it is known which sgRNA each population received, there is no need for sequencing technology to associate guides with their induced phenotypes.
While Cas9-based approaches, such as Perturb-seq, can be used for combinatorial screening, the authors conclude that the efficiency gained through CaRPool-seq makes it an attractive tool for dissecting gene networks. As a proof point, the authors go on to explore the gene networks involved in the malignant differentiation of acute myeloid leukemia cells using CaRPool-seq together with CITE-seq (cellular index of transcriptome and epitopes). This detailed screen revealed several redundant and synergistic gene-gene relationships that contribute to the differentiation process.
Efficiency is critical in any CRISPR screen, particularly those that are large in scale and complexity. Wessels et al. have developed a highly efficient approach to functional CRISPR screening of gene networks. Part of this efficiency is predicated on the accurate synthesis of crRNA arrays. For that, the team turned to Twist Bioscience, whose DNA synthesis platform is uniquely capable of synthesizing long oligonucleotides (up to 300nt in length).
Learn more about how Twist can support CRISPR screening here >>
What did you think?
Like
Dislike
Love
Surprised
Interesting
