A method for characterizing Cas9 variants via a one-million target sequence library of self-targeting sgRNAs

Abstract

Detailed target-selectivity information and experiment-based efficacy prediction tools are primarily available for Streptococcus pyogenes Cas9 (SpCas9). One obstacle to develop such tools is the rarity of accurate data. Here, we report a method termed ‘Self-targeting sgRNA Library Screen’ (SLS) for assaying the activity of Cas9 nucleases in bacteria using random target/sgRNA libraries of self-targeting sgRNAs. Exploiting more than a million different sequences, we demonstrate the use of the method with the SpCas9-HF1 variant to analyse its activity and reveal motifs that influence its target-selectivity. We have also developed an algorithm for predicting the activity of SpCas9-HF1 with an accuracy matching those of existing tools. SLS is a facile alternative to the much more expensive and laborious approaches used currently and has the capability of delivering sufficient amount of data for most of the orthologs and variants of SpCas9.

An Analysis of gRNA Sequence Dependent Cleavage Highlights the Importance of Genomic Context on CRISPR-Cas Activity

“…3,33 For example, several reports have highlighted specific motifs or nucleotide preferences of high fidelity variants but mechanistic explanations for this are lacking. 8,20,24,26 Additionally, other contextual factors such as target site accessibility or other unknown chromosomal factors may play a role in Cas9 activity. 7,8,34…”

“…11 In addition to the data collected in mouse and human cell lines in their lab, Doench et al 2014 3 provide data extracted from Shalem et al 2014 44 of gRNA targeting essential genes. Finally, for Tálas et al 2021 24 we combined the “balanced” datasets provided by the authors as a subset of the larger ~1.2 million gRNA library. The “balanced” datasets were provided by the authors to better help differentiate features that drive differences in efficient and inefficient gRNA as the majority of the larger ~1.2 million gRNA library would be deemed efficient.…”

Chronos: a CRISPR cell population dynamics model

“…CRISPR knockout (KO) of a gene is an inherently stochastic process. Some sgRNAs will fail to cut their target (Tálas et al, 2021) . In the event that they succeed and double-stranded break repair results in an insertion or deletion, in-frame mutations occur in about 20% of cases and may leave protein function intact (Michlits et al, 2020) .…”

Genome dependent Cas9/gRNA search time underlies sequence dependent gRNA activity

“…Non-target interactions 16,21 Secondary structure (reduced functional gRNA) 17,18 Target copy number 34 Target accessibility 41 R-loop formation 42 Cas9 variant 8,[24][25][26] Conformational shift [23][24][25][26] PAM sequence 3 Repair mechanism 22 Target copy number 34 Supercoiling 43 Reported to be gRNA sequence dependent Not reported to be gRNA sequence dependent Both gRNA 4 nt PAM sequence Target sequence Target site context G 3,6,9,13,24,25,27 A 25 C 21 C 3,6,13,18 T 3,6,9,13,18 A 27,29 C 3,10,27 G 10 T 3,6 No preference 13 C 10 G 3,10 gRNA:target duplex stability 18 Self-folding minimum free energy 18,23,24,31 Self-folding minimum free energy 11 Self-folding minimum free energy 10 Melting temperature 8 GC % 18,23 Extreme GC 2,3,8 Extreme GC 13 GC % 27 Melting temperature 24 Aggregate sequence properties Fig. 1 Summary of known factors that influence Cas9/gRN...…”

“…In some cases, this may influence the apparent or measured activity of a given gRNA. Finally, several studies have found sequence dependent differences in the activity of a given Cas9/gRNA complex when comparing wild-type Cas9 and higher fidelity variants that are known to cleave the target site more slowly [23][24][25][26] . This could indicate a role for sequence in determining the rate at which Cas9 cleaves the target site.…”