Two-Vector CRISPR System Is Better Approach for Knockout Screens

Since our introduction of the All-in-One CRISPR-Cas9 lentiviral system in the Spring, we have done a bit more work and moved forward with a two-vector system for our pooled libraries of sgRNA expression constructs. If the All-in-One CRISPR-Cas9 vector works well—which it does—why, you might wonder, would we develop and recommend a two-vector system for pooled CRISPR libraries? It turns out there are two reasons.

Cas9 Too Big for Lentiviral Vectors

Due to their efficiency and ease-of-use, we utilize lentiviral vectors almost exclusively for our RNAi and CRISPR libraries and constructs. While they make it easy to stably introduce shRNA, sgRNA, and even cDNA into virtually any mammalian system, one crucial limitation of the system is that the size of the provirus (the DNA that produces the RNA packaged in a viral particle) is limited. A provirus longer than 8kb does not package well.

The CRISPR system, of course, disrupts genes by targeting the Cas9 DNAs to a specific genomic site using small guide RNA (sgRNA). As a result, both Cas9 protein must be expressed with the sgRNA in a cell to knockout a gene. However, the Cas9 gene weighs in at 4101 bp, so it requires a bit more than half the maximum length of the provirus. Squeezing this gene with a puromycin selection marker, the sgRNA expression cassette, and a few other elements, into the lentivector creates a provirus that's about 8.2kb and packages very poorly. The result is that the one-vector Cas9-sgRNA yields low titers of VSV-g pseudotyped viral particles.

When transducing a small number of cells with the same construct to knock out a single target gene, there is no need for high viral titers. However, when transducing many thousands of constructs into large populations of cells, as is required for loss of function screens with pooled sgRNA libraries (run similar to shRNA screens), then high titers are a necessity.

To generate robust results in a loss-of-function genetic screen, the full library of pooled sgRNA constructs to be screened must be represented in a cell population a couple hundred fold. For libraries with tens of thousands of sgRNA constructs, millions of viral particles are needed for each screen. With a two-vector system, cells can be transduced with the one plasmid construct to produce the Cas9 protein, then grown to the point where they can transduced with the complex sgRNA library to carry out the CRISPR screen.

Pre-Transducing Cells with Cas9 Reduces Screening Noise

The explanation above, however, makes it seem as though the two-vector CRISPR system is a compromise due to the size limitation of the lentiviral system. It isn't though. The two vector approach is better for CRISPR screens because it reduces noise.

The knockout rate of a gene with CRISPR depends on the level of Cas9 in the cell. More Cas9 gives faster knockouts. When cells are transduced with a Cas9-expressing lentiviral vector, the level of Cas9 expression varies depending on where the lentivirus integrates. As a result, cell-to-cell levels of Cas9 vary greatly across a freshly transduced population. Because of this variation, it is actually not a good idea to transduce cells with a library of sgRNAs carrying the Cas9 gene on the same vector. This will produce a large range of Cas9 levels across the population of cells since each cell that picks of an sgRNA will express different levels of Cas9. The result increases variance in the screen.

On the other hand, by transducing the Cas9 construct first, into a small population of cells, then selecting and growing these cells for consistent high levels of Cas9 expression, a population is generated that will produce gene knockouts more quickly and reliably when transduced with the sgRNA library. The "stage is set" in a way for the main event—the screen.

We'll show some more data on this second point about the knockout rate and Cas9 levels in the next blog.

2-Vector vs. 1-Vector CRISPR Titers

Please email with any comments.

Also in Cellecta Blog & News

Perturb-Seq Screening: Cell-by-Cell Analysis of Gene Perturbations Induced by Pooled CRISPR sgRNA Libraries

Read More
Gene Expression Profiling of Single-Cell Samples: DriverMap Targeted Expression Profiling vs SMART Technology

Single-cell expression analysis provides insights about gene expression and cell heterogeneity at the single-cell level. It enables the elucidation of intracellular gene regulatory networks and intracellular pathways that would otherwise be masked in bulk analysis (Massaia et al., 2018). The DriverMap™ Targeted Gene Expression Profiling (TXP) assay combines highly multiplexed RT-PCR amplification with the depth and precision of Next-Generation Sequencing (NGS) to quantitatively measure gene expression of up to 19,000 target genes in a single assay–even down to the single-cell level.
Read More
Comparing DNA vs. RNA Samples for Immune Repertoire Profiling

Adaptive immunity relies on B and T cells that recognize foreign antigens via hypervariable B cell and T cell receptors (BCRs and TCRs). Diversity among B cell and T cell receptors is primarily produced by V(D)J recombination, which involves the shuffling and joining of the variable (V), diversity (D), joining (J), and constant region (C) gene segments. This results in a diverse repertoire called the adaptive immune repertoire (AIR) that comprises multiple individual clonotypes (sequence) for particular receptor chains.
Read More