Article Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
pubs.acs.org/JACS
Bidirectional Degradation of DNA Cleavage Products Catalyzed by CRISPR/Cas9 Anthony A. Stephenson,†,‡,∥ Austin T. Raper,†,‡,∥ and Zucai Suo*,†,‡,§ †
Department of Chemistry and Biochemistry, ‡Ohio State Biochemistry Program, and §Comprehensive Cancer Center, The Ohio State University, Columbus, Ohio 43210, United States S Supporting Information *
ABSTRACT: Since the initial characterization of Streptococcus pyogenes CRISPR/Cas9 as a powerful gene-editing tool, it has been widely accepted that Cas9 generates blunt-ended DNA products by concerted cleavage of the target (tDNA) and non-target (ntDNA) strands three nucleotides away from the protospacer adjacent motif (PAM) by HNH and RuvC nuclease active sites, respectively. Following initial DNA cleavage, RuvC catalyzes 3′→5′ degradation of the ntDNA resulting in DNA products of various lengths. Here, we found that Cas9 selects multiple sites for initial ntDNA cleavage and preferentially generates staggered-ended DNA products containing single-nucleotide 5′-overhangs. We also quantitatively evaluated 3′→5′ post-cleavage trimming (PCT) activity of RuvC to find that ntDNA degradation continues up to the −10 position on the PAM distal DNA product and is kinetically significant when compared to extremely slow DNA product release. We also discovered a previously unidentified 5′→3′ PCT activity of RuvC which can shorten the PAM proximal ntDNA product by precisely one nucleotide with a comparable rate as the 3′→5′ PCT activity. Taken together, our results demonstrate that RuvCcatalyzed PCT ultimately generates DNA fragments with heterogeneous ends following initial DNA cleavage including a PAM proximal fragment with a blunt end and a PAM distal fragment with a staggered-end, 3′-recessed on the ntDNA strand. These kinetic and biochemical findings underline the importance of temporal control of Cas9 during gene-editing experiments and help explain the patterns of nucleotide insertions at sites of Cas9-catalyzed gene modification in vivo.
■
INTRODUCTION The field of genetic engineering was revolutionized by the discovery of clustered regularly interspaced short palindromic repeats (CRISPRs) and CRISPR-associated (Cas) proteins which can mediate RNA-guided targeting and cleavage of double-stranded DNA.1−9 The CRISPR/Cas system of Streptococcus pyogenes is most widely utilized for gene-editing and consists of the Cas9 endonuclease and a synthetic single-guide RNA (sgRNA).2−8 DNA targeting by sgRNA-bound Cas9 is achieved through recognition of a protospacer adjacent motif (PAM) on the non-target DNA strand (ntDNA) followed by heteroduplex formation between the sgRNA and the target DNA strand (tDNA) (Figure 1).10−12 Subsequently, consecutive activation of the HNH and RuvC nuclease domains through coupled conformational dynamics leads to rapid cleavage of tDNA and ntDNA at the HNH and RuvC active sites, respectively.13−15 Based on DNA sequencing of Cas9 cleavage products, it was previously concluded that Cas9 generates bluntends at a position three base pairs upstream of the PAM.3 However, molecular dynamics simulations suggest that RuvC may actually generate staggered DNA ends with singlenucleotide 5′-overhangs.16 While it is widely recognized that, following initial DNA cleavage by Cas9, the ntDNA strand is further processed by the 3′→5′ post-cleavage trimming (PCT) © XXXX American Chemical Society
activity of RuvC on the PAM distal DNA fragment, it is unknown if this activity can occur before Cas9 dissociates from its initial DNA cleavage products.3,17−19 Gene-editing with Cas9 is not always successful and can have unintended consequences which represent significant obstacles to continued progress as this technology moves toward clinical applications (i.e., gene therapy). As a result, comprehensive characterization of Cas9 DNA cleavage and trimming activities is essential to uncovering functional limitations and experimental considerations for this biotechnology. Here we sought to explore the selection of the initial DNA cleavage site(s) by RuvC and determine if the PCT activity is kinetically significant within a single Cas9·sgRNA and DNA binding event. We found that RuvC can select up to three positions along the ntDNA strand for initial cleavage but prefers to generate DNA products with single-nucleotide 5′-overhangs, rather than blunt-ends. Surprisingly, we found that, in addition to the known 3′→5′ PCT activity, Cas9 also possesses 5′→3′ PCT activity. We determined the kinetics of RuvC-catalyzed 3′→5′ and 5′→3′ PCT and found them to be significant within a single Cas9·sgRNA and DNA binding event (i.e., extensive trimming Received: December 10, 2017
A
DOI: 10.1021/jacs.7b13050 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Article
Journal of the American Chemical Society
(HDR) pathways, which mediate gene knockout and gene knockin, respectively.5,20,21
■
RESULTS AND DISCUSSION RuvC Selects Multiple Sites for Initial DNA Cleavage. To capture the full spectrum of Cas9-catalyzed events including initial DNA cleavage and 3′→5′ PCT, we performed singleturnover assays with time points (0.003−7200 s) spanning 6 orders of magnitude (Figures 2A and S1A). Briefly, Cas9 (100 nM) and sgRNA (300 nM) were pre-incubated for 30 min at 20 °C to form the binary complex (Cas9·sgRNA) before addition of the DNA substrate (10 nM, Figure 1), which was 5′-radiolabeled on ntDNA. Following a 1 h incubation in the absence of MgCl2 to form the ternary complex (Cas9·sgRNA·DNA), the solution was rapidly mixed with MgCl2 (6 mM), in the presence (Figure 2A) or absence (Figure S1A) of unlabeled DNA substrate (1000 nM), to initiate DNA cleavage at 37 °C for various amounts of time before quenching with EDTA (0.37 M). Close inspection of ntDNA cleavage by RuvC (Figures 2A and S1A) revealed appearance of at least two distinct DNA products (27-mer and 26-mer, Figures 2A and S1A) even at the earliest time point of acquisition (3 ms). To a lesser extent, a third DNA product (25mer, Figures 2A and S1A) was also faintly observed in early time points (
