Crystal Structure of Cpf1 in Complex with Guide RNA and Target DNA

Abstract

Cpf1 is an RNA-guided endonuclease of a type V CRISPR-Cas system that has been recently harnessed for genome editing. Here, we report the crystal structure of Acidaminococcus sp. Cpf1 (AsCpf1) in complex with the guide RNA and its target DNA at 2.8 Å resolution. AsCpf1 adopts a bilobed architecture, with the RNA-DNA heteroduplex bound inside the central channel. The structural comparison of AsCpf1 with Cas9, a type II CRISPR-Cas nuclease, reveals both striking similarity and major differences, thereby explaining their distinct functionalities. AsCpf1 contains the RuvC domain and a putative novel nuclease domain, which are responsible for cleaving the non-target and target strands, respectively, and for jointly generating staggered DNA double-strand breaks. AsCpf1 recognizes the 5'-TTTN-3' protospacer adjacent motif by base and shape readout mechanisms. Our findings provide mechanistic insights into RNA-guided DNA cleavage by Cpf1 and establish a framework for rational engineering of the CRISPR-Cpf1 toolbox.

Figure 1. Overall structure of the AsCpf1–crRNA–target DNA complex

(A) Domain organization of AsCpf1. BH, bridge helix. (B) Schematic representation of the crRNA and target DNA. TS, target DNA strand; NTS, non-target DNA strand. (C and D) Cartoon (C) and surface (D) representations of the AsCpf1–crRNA–DNA complex. Molecular graphic images were prepared using CueMol (http://www.cuemol.org). See also Figures S1–S3 and Table S1.

Figure 2. Structure of the crRNA and target DNA

(A) Schematic representation of the AsCpf1 crRNA and the target DNA. The disordered region is surrounded by dashed lines. (B) Structure of the AsCpf1 crRNA and the target DNA. (C) Structure of the crRNA 5′-handle (stereo view). (D–F) Close up view of the U(−1)•U(−16) base pair (D), the reverse Hoogsteen U(−10)•A(−18) base pair (E), and the U(−13)-U(−17)-U(−12) base triple (F). Hydrogen bonds are shown as dashed lines. (G) Binding of the crRNA 5′-handle to the groove between the WED and RuvC domains. (H and I) Recognition of the 3′-end (H) and the 5′-end (I) of the crRNA 5′-handle. Hydrogen bonds are shown as dashed lines.

Figure 3. Schematic of the nucleic acid recognition by Cpf1

AsCpf1 residues that interact with the crRNA and the target DNA via their main chain are shown in parentheses. Water-mediated hydrogen-bonding interactions are omitted for clarity. See also Figure S4.

Figure 4. Recognition of the crRNA–target DNA heteroduplex

(A) Recognition of the crRNA–target DNA heteroduplex by the REC1 and REC2 domains. (B) Recognition of the target DNA strand by the bridge helix and the RuvC domain. Hydrogen bonds are shown as dashed lines. (C) Recognition of the crRNA seed region and the +1 phosphate group (+1P) (stereo view). Hydrogen bonds are shown as dashed lines. (D) Mutational analysis of the nucleic-acid-binding residues. Effects of mutations on the ability to induce indels at two DNMT1 targets were examined (n = 3, error bars show mean ± SEM). (E) Stacking interaction between the 20th base pair in the heteroduplex and Trp382 of the REC2 domain.

Figure 5. Recognition of the 5′-TTTN-3′ PAM

(A) Binding of the PAM duplex to the groove between the WED, REC1 and PI domains. (B) Recognition of the 5′-TTTN-3′ PAM (stereo view). Hydrogen bonds are shown as dashed lines. (C–E) Recognition of the dA(−2):dT(−2*) (C), dA(−3):dT(−3*) (D), and dA(−4):dT(−4*) (E) base pairs. (F) Mutational analysis of the PAM-interacting residues. Effects of mutations on the ability to induce indels at two DNMT1 targets were examined (n = 3, error bars show mean ± SEM). See also Figure S5.

Figure 6. RuvC and Nuc nuclease domains

(A) Structures of the RuvC and Nuc domains. The α helices (red) and β strands (blue) in the RuvC (RNase H fold) and Nuc domains are numbered. Disordered regions are shown as dashed lines. (B) Active site of the RuvC domain. (C) Mutational analysis of key residues in the RuvC and Nuc domains. Effects of mutations on the ability to induce indels at two DNMT1 targets were examined (n = 3, error bars show mean ± SEM). Indel values are normalized against wild-type AsCpf1. (D) Spatial arrangement of the nuclease domains relative to the potential cleavage sites of the target DNA. The catalytic center of the RuvC domain is indicated by a red circle. The REC1 and PI domains are omitted for clarity. A schematic of the crRNA and target DNA is shown above the structure. The DNA strands not contained in the crystal structure are represented in light gray. (E) Interaction between Trp958 and the hydrophobic pocket in the REC2 domain. (F) The AsCpf1 R1226A mutant is a nickase cleaving the non-target DNA strand. The wild type or the R1226A mutant (inactivation of the Nuc domain) of AsCpf1 was incubated with crRNA and the target DNA, which was labeled at the 5′ ends of both strands (DNA 1), or at the 5′ end of either the non-target strand (DNA 2) or the target strand (DNA 3). The cleavage products were analyzed by 10% polyacrylamide TBE-Urea denaturing gel electrophoresis. The SpCas9 D10A mutant (inactivation of the RuvC domain) is a nickase cleaving the target strand, and was used as a control. See also Figure S6.

Figure 7. Comparison between Cas9 and Cpf1

(A and B) Comparison of the domain organizations and overall structures between SpCas9 (PDB ID 4UN3) (A) and AsCpf1 (B). The catalytic centers of the RuvC domain are indicated by a red circle. (C and D) Models of RNA-guided DNA cleavage by Cas9 (C) and Cpf1 (D). (E and F) Comparison of the RuvC domains of SpCas9 (PDB ID 4UN3) (E) and AsCpf1 (F). The secondary structures of the conserved RNase H fold are numbered. See also Figure S7.