CRISPR is changing how we do biology (Addgene has a great write-up if you need a technical refresher). If you goal is screening for putative function, introducing DNA into the genome, or even inducing gene expression, CRISPR has a solution for you. In our work we’ve been thinking about how to integrate multiple CRISPR guides into a single location in the genome and have all of them expressed. This has been lagging a little behind the one-target / one -guide work, but there are a couple of options out there to get this done.

  1. I’ll make this option zero, as it doesn’t really apply, but you could use single-guide integration strategies to put many guides randomly into the genome, and screen for cells or animals where you’ve successfully integrated one of each guide. This is a lot of work.
  2. Hammerhead ribozymes, demonstrated in this paper and in other works. It was a little unclear to me how well this would work when the number of guides to excise out rises to 5-10.
  3. Repeated pairs of promoters and guides, as seen in this paper. They built this out for 7 guides, but we wanted to be able to do a few more than that.
  4. A tRNA excision system, which was just detailed in a bioRxiv paper from the Bullock lab from the MRC. This is the paper I want to look at now.

Port et. al.

Say you want to drop out a piece of DNA like in the above picture. Or maybe you want to activate a whole bunch of genes using CRISPRa. Either way you need to get a series of guides into the genome of interest before they successfully join up with Cas9 to carry out your mission and target the gene or region of interest. You’re also generally constrained by the fact that you need the whole system to fit into a single vector, which for lentiviral systems mean you have a couple of kilobases to work with. Lastly, it would also be nice to have a single well characterized promoter drive expression of all of the guides. This way there wouldn’t be competition between guides for the attention of the polymerase. Port et. al. proposes a solution using a built in system to excise RNA between individual guides, the tRNA system. The high level view is:

The authors has based this system on work recently published in PNAS in rice. The idea is that individual guides are arranged between tRNA sequences, which are naturally spliced out by native machinery. This means that you’re left with individual guide sequences floating around, which can complex with Cas9 and go onto editing the genome. In initial experiments it seemed as if a single tRNA+gDNA worked better than the gDNA alone, though this was only sampled in a few experiments. Using a tiled series of guides, the authors show that you can use such a system to generate numerous edits with varying efficiency:

Targeting four different genes that regulate four different phenotypes, they authors were able to show that you could achieve combinations of phenotypes using this approach. It’s unclear to me if the trailing guide (with 31% efficiency) is the result of poor targeting by that guide, or if the last guide in an array is the least likely to be cleaved. The authors also go on to show that this system can be used with cpf1, an alternate CRISPR genome editing protein. Again, a really cool concept, and something we’ll try out in the near future.

The citation for their paper (on BioRxiv) is something like this:

“Expansion of the CRISPR toolbox in an animal with tRNA-flanked Cas9 and Cpf1 gRNAs”. Fillip Port, Simon L Bullock. bioRxiv 046417; doi: http://dx.doi.org/10.1101/046417