We and another lab here in Houston have encountered a problem using the general Seydoux protocol with in vitro synthesized Cas9, tracrRNA and crRNA. Both of our labs have successfully used this approach because of increased efficiency with the dpy-10(cn64) co-CRISPR strategy, though we both had succeeded with plasmid-based approaches.
The problem is what looks like a toxic or dominant negative reagent. The dpy-10 co-CRISPR tools work great, but we lose Dpy/Rol production when we add in crRNA(s) for the second target. With other targets we have no problems, so clearly this is not a general phenomenon. And we carefully tested with subtraction and add-back experiments (for my group, I did all the injections, and so am highly confident of the efficacy). We have now totaled three targets that cause this problem. Has anybody else had similar issues? Any potential solutions?
Edit: Just to clarify, we don’t think it is the mutation that is dominant lethal, or something like that. They are all structurally unrelated and act in diverse processes> Rather, we wonder whether certain crRNAs could lock up the Cas9 so it is non-functional. A solution would be to try different guides (and hence crRNAs) for the same gene.
Are you adding Cas9 last to the mix or does adding the additional reagents create more pipetting steps with Cas9 involved? Our lab recently found that changing the way we pipetted the reagents gave us the expected number of Rol, whereas before, we were averaging about 1 Rol per 20 animals injected (our plasmid based control injections were fine). Now, we don’t even pipet the Cas9, but mix the other reagents together and add slowly and directly to the Cas9 tube, pipetting up and down only once (vigorously trying to mix the Cas9 with everything seems to not yield any Rol, according to another labmate who’s tried). I’ve repeated this 6 times with similar results: ~80% injected P0 yielding Rol. So it seems likely that pipetting was our issue.
Fascinating! So you would hypothesize that targeting two genes (e.g. yfg-1 and dpy-10) would result in twice as much pipetting, and so a drop in efficiency? And there could be a ton of variability from injection to injection depending on the person who injects, their pipetting vigor, and the order in which each person does it.
Update from our neighbors, with whom we have been working to find a fix. These folks were getting nothing, including Dpy/Rols from the dpy-10(cn64) co-CRISPR (identifying info redacted)
Half the crRNA concentration for our target gene as we are injecting two crRNA targeting the same gene.
Gently pipetting the RNA mix into the Cas9 protein, do not pipette up and down.
Use the injection mix immediately following the 15 minute 37C incubation.
Use the needle base to swirl/mix the injection mix immediately before needle loading. {EDIT: this lab loads needles by putting the back end of the needle in the microfuge tube with the injection mix. That is why this step is here.}
Don’t know if we have the knockouts of interest yet, but will keep you updated."
My take: Not so sure about the last modification. But we did consult with some CRISPR luminaries. One recommended half the crRNA concentration for each crRNA (so, same total crRNA for a two-crRNA mix) in the co-CRISPR mix, plus immediate loading and injection after incubation. And, you can see they also were very gentle about assembling the mix. Not 100% sure what is meant by the last step there, but the rest makes sense. Even if most edits are working (as in our case), perhaps overall efficiency can be boosted by incorporating these details into your workflow?