CgII cleaves DNA using a mechanism distinct from other ATP-dependent restriction endonucleases

Abstract

The restriction endonuclease CglI from Corynebacterium glutamicum recognizes an asymmetric 5'-GCCGC-3' site and cleaves the DNA 7 and 6/7 nucleotides downstream on the top and bottom DNA strands, respectively, in an NTP-hydrolysis dependent reaction. CglI is composed of two different proteins: an endonuclease (R.CglI) and a DEAD-family helicase-like ATPase (H.CglI). These subunits form a heterotetrameric complex with R2H2 stoichiometry. However, the R2H2·CglI complex has only one nuclease active site sufficient to cut one DNA strand suggesting that two complexes are required to introduce a double strand break. Here, we report studies to evaluate the DNA cleavage mechanism of CglI. Using one- and two-site circular DNA substrates we show that CglI does not require two sites on the same DNA for optimal catalytic activity. However, one-site linear DNA is a poor substrate, supporting a mechanism where CglI complexes must communicate along the one-dimensional DNA contour before cleavage is activated. Based on experimental data, we propose that adenosine triphosphate (ATP) hydrolysis by CglI produces translocation on DNA preferentially in a downstream direction from the target, although upstream translocation is also possible. Our results are consistent with a mechanism of CglI action that is distinct from that of other ATP-dependent restriction-modification enzymes.

Figure 1.

The adenosine triphosphate (ATP)-dependent restriction-modification (RM) enzymes. (A) Mechanisms of DNA cleavage by Type I, ISP and III RM enzymes. Nuclease subunit/domains are labeled ‘N’. For Type I enzymes, each complex has two helicase–nuclease subunits which can translocate away each side of the site, drawing in two DNA loops (only one loop is shown per complex for clarity). Upon collision of two converging translocating enzymes at a non-specific site, the nucleases engage and generate a dsDNA break (2). Additional processing leads to DNA shortening and release of small fragments (54,55). For Type ISP enzymes, translocation is unidirectional. Upon collision at a non-specific site, the nuclease domains are held apart. It is proposed that movement of the collision complex results in multiple nicks that generate the dsDNA break (9,40). For the Type III enzymes, ATP hydrolysis leads to a conformation switch into a DNA sliding state. Long-range communication is bidirectional and driven by thermal energy. Upon collision with a second enzyme complex bound to a target site, the nucleases engage and generate a dsDNA break (18). (B) Domain organization of the R- and H.CglI proteins. R.CglI (NCgl1704) is a putative REase with the phospholipase D (PLD)-superfamily nucleolytic and B3-like DNA binding domain. H.CglI (NCgl1705) is predicted superfamily 2 (SF 2) helicase/ATPase containing uncharacterized Z1-superfamily and C-terminal domains. (C) Proposed model for the R₂H₂·CglI complex. R.CglI is composed of the PLD and B3 domains (colored gray and dark gray). H.CglI contains the DEAD, Z1 and C-terminal domains (colored white and light gray). Figure was made according to (28).

Figure 2.

One-site plasmid (A), two-site plasmid (B) and two-site catenane (C) cleavage by CglI. The two-site plasmid substrate was used to form catenane in the presence of Tn21 resolvase (7). The CglI recognition sequence (5′-GCCGC-3′) is shown as an arrowhead (▸). Possible DNA species formed when CglI is incubated with the plasmid or catenane substrates are presented above the graphs: SC—supercoiled circular DNA, OC—open circle DNA (‘nicked’), FLL—full-length linear DNA, L1+L2—linear DNA cut at both CglI sites, SCcat—SC catenane DNA, OC1—catenane nicked in the large ring only, OC2—catenane nicked in the small ring only, OC3—catenane nicked in both rings, OC4—nicked large plasmid, OC5—nicked small ring, FLL1—full-length linear DNA of the large ring, FLL2—full-length linear DNA of the small ring. Note that final FLL products from SCcat can be generated by different routes, requiring either 2, 3 or 4 consecutive independent cleavage events on a single DNA substrate. Reactions contained 10 nM DNA, 4 mM ATP, 500 nM R₂H₂·CglI and were conducted as described in ‘Materials and Methods’ section. The rate constant values (5.3 ± 0.3 × 10⁻³ s⁻¹ for (A), 1.1 ± 0.1 × 10⁻² s⁻¹ for (B), 1.2 ± 0.1 × 10⁻² s⁻¹ for (C)) were obtained by fitting a single exponential to the time courses of supercoiled form depletion. Points are averages with error bars as standard deviation for at least three repeat reactions.

Figure 3.

(A) Cleavage of the SC, OC and linear one-site plasmid DNA by CglI. Accumulation of the final DNA cleavage products with both DNA strands cleaved at the CglI recognition sequence is shown. The one-site SC DNA was used to form the OC DNA (by cleavage of a single DNA strand with the Cas9–crRNA complex) or the full-length linear DNA (FLL) (both DNA strands were cleaved with the restriction endonuclease NdeI). The CglI recognition sequence (5′-GCCGC-3′) is shown as an arrowhead (▸). The rate constant values (3.6 ± 0.2 × 10⁻³ s⁻¹ for SC, 3.0 ± 0.3 × 10⁻³ s⁻¹ for FLL and 4.2 ± 0.7 × 10⁻³ s⁻¹ for OC) were obtained by fitting a single exponential to the time courses of substrates. (B) Cleavage of the linear one- and two-site DNA by CglI. Accumulation of the final DNA cleavage products with both DNA strands cleaved at one CglI recognition sequence is shown. Numbers indicate distances in bp between the CglI targets and DNA ends. The rate constant values (2.0 ± 0.3 × 10⁻³ s⁻¹ for HtN, 8.9 ± 1.9 × 10⁻⁴ s⁻¹ for HtH and 3.3 ± 0.1 × 10⁻³ s⁻¹ for HtT) were obtained by fitting a single exponential to the time courses of substrates. All reactions contained 10 nM DNA, 4 mM ATP, 500 nM R₂H₂·CglI and were conducted as described in ‘Materials and Methods’ section. Points are averages with error bars as standard deviation for at least three repeat reactions.

Figure 4.

Cleavage of the two-site linear DNAs with different target orientations by CglI. Cleavage of the HtT substrates is presented in (A), the HtH and TtT substrates–in (B). The two-site supercoiled circular plasmids were used to form the linear substrates by cleavage with the restriction endonucleases XhoI and NdeI). The CglI recognition sequence (5′-GCCGC-3′) is shown as an arrowhead (▸),‘S’ represents the substrates and any nicked intermediates, while P1 and P2 represent a double strand break at site P1 or P2, respectively. Numbers indicate distances in base pair between the CglI targets and DNA ends. All reactions contained 10 nM DNA, 4 mM ATP, 500 nM R₂H₂·CglI and were conducted as described in ‘Materials and Methods’ section. The rate constant values (3.3 ± 0.1 × 10⁻³ s⁻¹ for HtT1, 4.4 ± 0.4 × 10⁻³ s⁻¹ for HtT2, 8.9 ± 1.9 × 10⁻⁴ s⁻¹ for HtH and 2.0 ± 0.2 × 10⁻³ for TtT) were obtained by fitting a single exponential to the time courses of substrates. Points are averages with error bars as standard deviation for at least three repeat reactions.

Figure 5.

CglI translocase activity analyzed using the DNA triplex displacement assay. (A) CglI displaces TFO. Scheme of the DNA substrate used in the triplex assay is shown above the graph: DNA is shown as a thick line, the CglI recognition sequence (5′-GCCGC-3′) is shown as an arrowhead (▸), the triplex binding sequence is shown as a varied rectangle, numbers indicate distances in base pair. Used abbreviations: H—H.CglI, R—R.CglI, Hmut—the H.CglI (D158A + E159A) mutant, Rmut—the R.CglI (H105A) mutant. The inset shows zoomed part of the graph to clarify the differences of TFO displacement. (B) TFO displacement by CglI is dependent on the recognition sequence. DNA substrates: N—DNA without the recognition sequence, the T and H DNA fragments contain the recognition sequence in both orientations. All reactions contained 5 nM DNA, 2.5 nM TFO, 4 mM ATP, 200 nM R.CglI, 200 nM H.CglI and were conducted as described in ‘Materials and Methods’ section. Solid lines are single or double exponential fits to the data. Rate constants for TFO displacement are presented in Supplementary Table S4. Points are averages with error bars as standard deviation for at least three repeat reactions.

Figure 6.

TFO displacement by CglI in trans. (A) TFO displacement from the linear DNA (N) without the recognition sequence of CglI. DNA substrate (278 bp in length) is shown above the graph. The triplex binding sequence is shown as a varied rectangle. DNA added in trans (278 bp in length) are the same as used in Figure 5B. (B) TFO displacement from the circular versus linear DNA. DNA substrates (3383 bp in length) are shown above graph. The CglI recognition sequence (5′-GCCGC-3′) is shown as an arrowhead (▸). DNA added in trans (3105 bp in length) were linear (L0 and L1) and circular (p0s and p1s) without and with recognition sequence of CglI (are shown on the right). All reactions contained 1 nM DNA substrate, 0.5 nM TFO, 4 mM ATP, 200 nM R.CglI (H105A), 200 nM H.CglI, 4 nM (A) or 10 nM (B) DNA added in trans and were conducted as described in ‘Materials and Methods’ section. Solid lines are single exponential fits to the data. Rate constants are presented in Supplementary Table S4. Points are averages with error bars as standard deviation for at least three repeat reactions.

Figure 7.

Effect of DNA immobilization on TFO displacement by CglI in trans. TFO displacement from the non-specific DNA in the presence of the specific DNA (in solution or immobilized). N, H and T DNA fragments (278 bp in length) are the same as used in Figure 5B. DNA is shown as a thick line, the CglI recognition sequence (5′-GCCGC-3′) is shown as an arrowhead (▸), the triplex binding sequence is shown as a varied rectangle, biotin is shown as black diamond, magnetic bead is shown as gray circle. All reactions contained 1 nM DNA substrate, 0.5 nM TFO, 4 mM ATP, 200 nM R.CglI, 200 nM H.CglI, 50 nM DNA in trans (in solution or immobilized) and were conducted as described in ‘Materials and Methods’ section. Solid lines are single exponential fits to the data. Rate constants are presented in Supplementary Table S4. Points are averages with error bars as standard deviation for at least three repeat reactions.

Figure 8.

Proposed mechanism of action of CglI on various DNA substrates. DNA is shown as a thick line, the CglI recognition sequence (5′-GCCGC-3′) is shown as an arrowhead (▸). The R₂H₂·CglI complex (white ellipse) containing only a single active site binds the recognition sequence, becomes activated (gray ellipse), starts to translocate on DNA using ATP hydrolysis and leaves the target. For clarity here only one of two possible directions of translocation on DNA is shown. The second CglI complex binds to the same recognition sequence. After collision of the target-bound and translocating CglI complexes (black ellipses) a double-stranded break is introduced 7 and 6/7 nucleotides downstream of the 5′-GCCGC-3′ site on the top and bottom strands, respectively. In the case of the linear one-site substrate two target bound/activated CglI complexes associate in trans introducing a double-stranded break near to the one of the targets (inefficient cleavage). For the linear two-site substrates only a scheme of the HtT fragment cleavage is shown. CglI shows a preferential directionality of translocation on DNA (shown by an arrow) resulting in the predominant DNA cleavage at the second site (downstream to the first).