Spotlight Rewired Cas9s with Minimal Sequence Constraints Sébastien Levesque, 1,@ Daniel Agudelo, 1,@ and Yannick Doyon 1, * ,@ The genome editing toolkit is ever expanding. Although CRISPR-Cas systems can target virtually any gene, single-nucleotide resolution is yet to be achieved. Walton and colleagues engineered nucleases and base editors compatible with every protospacer adjacent motif (PAM) to achieve high-precision targeting. Their findings revealed the striking plasticity of Cas9. The constant evolutionary arms race be- tween bacteria and their viruses led to the emergence of sophisticated CRISPR- Cas defense systems. Some, such as CRISPR-Cas9, have been repurposed to engineer the genomes of all kingdoms of life. To target a specific sequence, Cas9 nucleases use a 20-nucleotide guide RNA (gRNA) to interrogate DNA for potential complementarity. This direc- tional search process is licensed and constrained by the PAM sequence, which must be found immediately down- stream of the target site (Figure 1)[1]. For the most widely used nuclease, Streptococcus pyogenes Cas9 (SpCas9), target exploration is driven by the canoni- cal NGG PAM (where N is any nucleotide followed by two guanines). In its natural prokaryotic habitat, PAM specificity is crucial to prevent self-targeting by Cas9. However, the PAM requirement restricts the targetable genomic space, so Cas9 orthologs with different PAM specificities, such as SaCas9, NmeCas9, and St1Cas9, can expand the targeting range of CRISPR systems [2]. The PAM-interacting (PI) domain, located near the C terminus of Cas9, recognizes its PAM sequence through protein DNA contacts leading to local unwinding of the target duplex, subsequent R-loop forma- tion, and stable binding (Figure 1)[1,3]. The crystal structure of SpCas9 revealed that arginines 1333 (R1333) and 1335 (R1335) specify the second and third gua- nines in the NGG PAM, which is otherwise nestled in a positively charged groove [3]. This provided a framework for the remod- eling of base-specific interactions and the introduction of compensating non- base-speci fic ones to reprogram PAM specificity. While not trivial, these ap- proaches were largely successful and even yielded variants able to function with minimal NGN and NRNH PAMs (where R is A, or G and H is A, C, or T) [2,4–7]. Walton et al. took an iterative approach to bring this concept a step further and created SpG, targeting an expanded set of NGN PAMs, and SpRY, a near- PAMless SpCas9 variant active on NRN and NYN PAMs (where Y is C or T) (Figure 1). They demonstrate their perti- nence in the context of base editors by creating ‘protective’ genetic variants [8]. Equipped with structural data and their previously developed SpCas9-VRQR (D1135V, G1218R, R1335Q, T1337R) molecular scaffold [4], they set out to fur- ther relax PAM preference from NGA to NGN ( Figure 1). First, they developed HT-PAMDA, an in vitro assay that allows the quanti fication of cleavage or base editing rate on DNA targets harboring dif- ferent PAMs. Briefly, whole-cell lysates containing a nuclease, or a base editor variant, were complexed with in vitro- transcribed single guide RNAs (sgRNAs) and incubated with linearized PAM- containing plasmid substrates. The activity of the variants was assessed by next- generation sequencing, which allowed the high-throughput assessment of PAM specificity. This facilitated comprehensive profiling of the activity and PAM preference of a large pool of SpCas9 variants contain- ing substitutions at five residues selected for their spatial proximity to the PAM. When assessed with HT-PAMDA, the SpG variants harboring the S1136W and E1219Q mutations exhibited comparable activity at all combinations of NGN PAM sites (Figure 1). No substantial first PAM, fourth PAM, or first spacer position prefer- ences were observed. In human HEK293T cells, SpG displayed high editing activity at several endogenous loci flanked by NGN PAMs, corroborating HT-PAMDA findings. Furthermore, the nuclease, cytosine base editing, and adenine base editing activity of SpG compared favorably with previously reported NGN-accepting variants [5,6]. To further relax NGN PAM specificity, SpG was used as a scaffold to engineer a variant with a substitution at the R1333 residue to alter the recognition of the second G nucleotide. While this nearly abolished activity in human cells, two extra mutations (L1111R and A1322R) partially rescued activity at NRN PAMs through the addition of nonspecific con- tacts (Figure 1)[6]. Using SpG harboring the L1111R and A1322R substitutions, the impact of single and combinatorial changes imparted at R1333 and three additional residues was profiled using HT-PAMDA and at endogenous loci. This led to SpRY, a variant capable of targeting NRNNNYN PAMs. SpRY harbors a total of 11 residue substitutions, including R1333P and R1335Q, the key positions that specify the second and third guanines in the NGG PAM for wild-type (WT) Cas9 [3]. Of note, A61R located in the arginine-rich bridge helix near the N terminus of Cas9 is the only substitution not found in the PI domain (Figure 1). Assessment of its nuclease and base editing activity in HEK293T cells con- firmed that SpRY is highly active on NRN PAMs. In addition, SpRY was found to be active at 42% of the NYN PAMs tested, although with half the activity observed for NRN PAMs. While SpRY displays reduced overall activity, with clear guide-dependent Trends in Pharmacological Sciences, Month 2020, Vol. xx, No. xx 1 Trends in Pharmacological Sciences TIPS 1716 No. of Pages 3