First, this work was co-led by postdocs Mat Legut and Zharko Daniloski and aided by several others in our group. (Special shout-out to Xinhe Xue, who spent her PhD rotation helping us launch this project and then joined our lab!) (2/15)
-
-
Prikaži ovu nit
-
First a little background: For gain-of-function screens, we were interested to know whether we could target smaller regions (and hence those containing few NGG PAMs) using the recent variants, xCas9 and Cas9-NG, which only requiring a single G at the target site. (3/15)
Prikaži ovu nit -
So, instead of a typical CRISPR pooled screen with a library of guide RNAs, we also had all 3 different Cas9 enzymes in the same pool (transduced separately). For the multi-CRISPR pooled screen, we added a 6nt barcode after the gRNA to encode enzyme & KO/i/a modality. (4/15)pic.twitter.com/FBbSnXgKJl
Prikaži ovu nit -
Thus, in a single screen, we could capture both target site/PAM preferences and also relative efficiency of each Cas9 variant. (5/15)pic.twitter.com/dDzP3XdJbn
Prikaži ovu nit -
Overall, we found that PAM flexibility of Cas9 mutants comes at a cost of reduced editing efficacy. For example, here are several guides targeting the cell-surface marker CD46. Overall, wild-type Cas9 leads to highest knock-out. (6/15)pic.twitter.com/gUUMzyV4ov
Prikaži ovu nit -
And, in case you’re curious, if you deep sequence the target sites, DNA editing/indels are very well correlated (as expected) with protein loss across all enzymes: (7/15)pic.twitter.com/nkYFQGwtTq
Prikaži ovu nit -
Since CRISPRi and CRISPRa require binding at the target site (but not cutting since they utilize nuclease-null dCas9), we wondered if results might be different for those applications. But they were broadly the same as what we’d seen with nuclease. (8/15)pic.twitter.com/43mTzteoWA
Prikaži ovu nit -
We also detected Cas9-NG activity at unconventional NAD PAMs. Upon testing, we found that some (but not all) NAD target sites work with Cas9-NG. For example, only Cas9-NG can drive CD45 expression at the NAA target site shown in the last row below. (9/15)pic.twitter.com/JoPDgQYyZ3
Prikaži ovu nit -
At this point, you might have noticed that there is also a difference in the PAM-flexible Cas9 variants: Specifically, that Cas9-NG outperforms xCas9 at NGH PAMs. (H= not G) Indeed, this is the case across all of our screens: knock-out, CRISPRi, and CRISPRa. (10/15)pic.twitter.com/iqoAOJX27D
Prikaži ovu nit -
Given this data, we wondered whether the lower activity of xCas9 could be rescued using the Cas9-NG mutations. So, we made a frankenzyme: xCas9-NG. (11/15)pic.twitter.com/14c8JFW89L
Prikaži ovu nit -
Although there was only a partial rescue for nuclease activity (yield an intermediate level of activity between xCas9 and Cas9-NG), we found that — much to our surprise — xCas9-NG yielded a better transcriptional activator than either existing PAM-flexible enzyme. (12/15)pic.twitter.com/dY5ZYwmWV3
Prikaži ovu nit -
What to make of all this? First, multi-CRISPR pooled screens provide a controlled way to benchmark different genome editing tools. (13/15)
Prikaži ovu nit -
Second, none of the three PAM flexible Cas9 mutants were capable of matching the activity of wild-type Cas9 at NGG PAM sites, so these mutations likely incur a fitness cost. (14/15)
Prikaži ovu nit -
Finally, thinking to the future, one thing that is evident is that there is a lot of room for new strategies to design efficient, PAM-flexible — or perhaps even PAM-independent — Cas9 enzymes. (15/15)
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.