A novel zinc-finger nuclease platform with a sequence-specific cleavage module

Abstract

Zinc-finger nucleases (ZFNs) typically consist of three to four zinc fingers (ZFs) and the non-specific DNA-cleavage domain of the restriction endonuclease FokI. In this configuration, the ZFs constitute the binding module and the FokI domain the cleavage module. Whereas new binding modules, e.g. TALE sequences, have been considered as alternatives to ZFs, no efforts have been undertaken so far to replace the catalytic domain of FokI as the cleavage module in ZFNs. Here, we have fused a three ZF array to the restriction endonuclease PvuII to generate an alternative ZFN. While PvuII adds an extra element of specificity when combined with ZFs, ZF-PvuII constructs must be designed such that only PvuII sites with adjacent ZF-binding sites are cleaved. To achieve this, we introduced amino acid substitutions into PvuII that alter K(m) and k(cat) and increase fidelity. The optimized ZF-PvuII fusion constructs cleave DNA at addressed sites with a >1000-fold preference over unaddressed PvuII sites in vitro as well as in cellula. In contrast to the 'analogous' ZF-FokI nucleases, neither excess of enzyme over substrate nor prolonged incubation times induced unaddressed cleavage in vitro. These results present the ZF-PvuII platform as a valid alternative to conventional ZFNs.

Figure 1.

Strategies for developing ZF-PvuII fusion enzymes compared to the ‘classical’ ZFN design. (A) Scheme of a ‘classical’ ZFN (ZF-FokI) with two ZFN monomers, each consisting of a ZF DNA-binding domain fused to the N-terminal end of the unspecific cleavage domain of FokI (2FOK). Upon dimerization of the two ZFN monomers at the DNA target sequence ‘Z–Z’, cleavage within the intervening DNA sequence of random 6 bp will be catalyzed by the interacting FokI-cleavage domains. Recombinant ZF-FokI nucleases were purified by affinity chromatography using their N-terminal Strep-tag. (B) Scheme of a ZF-PvuII homodimer (a homodimer fusion approach), in which the ZF DNA-binding domain (N-1-2-3-C) is fused to the N-terminal end of each subunit of the PvuII homodimer (1PVU). The tripartite DNA target sequence ‘Z–P–Z’ is addressed by this ZF-PvuII fusion enzyme. (C) Scheme of ZF-scPvuII (a monomer fusion approach), in which the ZF-binding module is fused to the N-terminal end of the single-chain variant of PvuII (3KSK). This ZF-scPvuII fusion enzyme addresses the bipartite target sequence ‘P–Z’. Purification of recombinant ZF-PvuII and ZF-scPvuII was performed by tandem affinity purification using the N-terminal Strep- and the C-terminal His-tag of both types of fusion enzymes.

Figure 2.

Kinetic analysis of the ZF-PvuII fusion enzymes cleaving an addressed substrate [black triangle; 250 bp PCR product containing either the tripartite (Z12P12Z) or bipartite (P12Z) target sites] and an unaddressed substrate [open diamond; 450 bp PCR product containing a single PvuII site (P)] in competition in a near equimolar stoichiometry (20 nM addressed substrate/20 nM unaddressed substrate/18 nM enzyme). The gel electrophoretic analysis of samples withdrawn from the incubation mixture at defined time intervals is shown as an insert of the activity versus time profile. (A) Preferential DNA cleavage (addressed versus unaddressed) by ZF-PvuII G46 (Z12P12Z versus P). (B) Preferential cleavage by ZF-scPvuII G46 (P12Z versus P). (C) Preferential DNA cleavage by ZF-PvuII G46/F94 (Z12P12Z versus P). (D) Preferential DNA cleavage by ZF-PvuII G46/A83/F94 (Z12P12Z versus P).

Figure 3.

Addressed cleavage of a 4878 bp plasmid substrate by the ZF-PvuII fusion enzymes using a near equimolar stoichiometry (20 nM plasmid/18 nM ZF-PvuII homodimer). The plasmid contains four unaddressed PvuII sites (P) and an ‘addressed cassette’ region (indicated by an open star) with either the ‘addressed tripartite target sequence’ (4xP-Z12P12Z) or another unaddressed PvuII site (P) as control (4xP-P). A unique HindIII site (H) is also included in the plasmid sequence for further analysis of the addressed DNA cleavage (Figure 4). The gel electrophoretic analyses of the cleavage products obtained after defined times of incubation of the plasmids harboring 4xP-Z12P12Z or 4xP-P with PvuII G46, ZF-PvuII G46, ZF-PvuII G46/F94 or ZF-PvuII G46/A83/F94 are shown, respectively. The supercoiled, open circular and linear forms of the plasmid are denoted by sc, oc and lin. The final cleavage products are indicated by their respective lengths (1828, 1122, 870, 608 and 450 bp); intermediate cleavage products are marked by an asterisks. It should be noted that the ZF-PvuII G46/F94 and the ZF-PvuII G46/A83/F94 fusion proteins cleave the plasmid only at the addressed sequence (Z12P12Z).

Figure 4.

Addressed cleavage of a 4878 bp plasmid substrate by the ZF-PvuII fusion enzymes using an excess of enzyme (20 nM plasmid/100 nM ZF-PvuII homodimer). The gel electrophoretic analyses of the cleavage products of the plasmids 4xP-P and 4xP-Z12P12Z produced by ZF-PvuII G46/F94 or ZF-PvuII G46/A83/F94, respectively, are shown. The supercoiled, open circular and linear forms of the plasmid are denoted by sc, oc and lin. The final cleavage products are indicated by their respective lengths (1828 and 1122 bp); intermediate cleavage products are marked by an asterisks. After 24 h of cleavage by the fusion enzyme, a secondary cleavage with HindIII was performed to characterize the linear cleavage product. Only cleavage at the addressed PvuII site and the HindIII site, results in formation of two cleavage products, 3442 and 1436 bp in length. The occurrence of these two products, which are marked by a bracket and denoted with ‘addressed’, demonstrates that the ZF-PvuII fusion enzyme before cleavage by HindIII had cleaved the plasmid only at the addressed PvuII target site. Further cleavage products that are occurring after the subsequent HindIII cleavage and document previous cleavage events at unaddressed PvuII sites by the ZF-PvuII fusion enzymes, are marked by an arrow.

Figure 5.

Influence of mutations in the FokI-cleavage domain (ZF-FokI EA, ZF-FokI KV, ZF-FokI ELD, ZF-FokI KKR) on the ZF-FokI-cleavage specificity of ‘classical’ ZFNs. Cleavage of a 4878 bp plasmid substrate was performed for 24 h, using a near equimolar stoichiometry of DNA substrate and ZFN dimers [20 nM plasmid/36 nM ZF-FokI monomer(s); top panel] and an excess of enzyme [20 nM plasmid/200 nM ZF-FokI monomer(s); bottom panel], respectively. ZF-FokI variants with either the homodimeric wt FokI-cleavage domain, or the obligate heterodimeric FokI-cleavage domains EA, KV, ELD, KKR, each individually tested, or the obligate heterodimeric FokI-cleavage domain pairs EA + KV and ELD + KKR in combination were investigated. A summary of the gel electrophoretic analyses of the cleavage products obtained after 24 h of incubation of the plasmids 4xP-P (unspecific target), 4xP-Z6Z (specific target) and 4xP-Z (half-site target) with the different ZF-FokI variants is shown. The supercoiled, open circular and linear forms of the plasmid are denoted by sc, oc and lin. Further cleavage products that result from cleavage at unspecific target sites are marked by an asterisks. (The detailed kinetic analyses of DNA cleavage by all ZF-FokI variants are shown in Supplementary Figures S12–S18.)

Figure 6.

Influence of the salt concentration on the addressed cleavage of a 4878 bp plasmid substrate by the obligate heterodimeric ZFN pair ZF-FokI ELD + KKR in combination (20 nM DNA/100 nM ZF-FokI ELD monomer/100 nM ZF-FokI KKR monomer; left panel) and by the ZF-PvuII G46/A83/F94 fusion enzyme (20 nM DNA/100 ZF-PvuII G46/A83/F94 homodimer; right panel), respectively. The salt concentration of 50 mM NaCl was complemented with KCl to give total salt concentrations of 50, 75, 100, 125, 150, 175 and 200 mM salt. The gel electrophoretic analyses of the cleavage products of 4xP-P (unspecific target), 4xP-Z6Z (specific target for ZFN) and 4xP-Z12P12Z (specific target for ZF-PvuII) cleaved by ZF-FokI ELD + KKR or ZF-PvuII G46/A83/F94 after 24 h of incubation are shown. The supercoiled, open circular and linear forms of the plasmid are denoted by sc, oc and lin. Further cleavage products that result from cleavage at unspecific target sites are marked by an asterisks. (The corresponding gel electrophoretic analyses of the products obtained after 1 h of incubation are shown in Supplementary Figure S19.)

Figure 7.

ZF-PvuII G46/A83/F94 mediated DNA cleavage in mammalian cells. (A) Schematic of the ZF-PvuII G46/A83/F94 target sites. The addressed Z12P12Z target site harbors an inverted repeat of ZF-binding sites separated by 12-bp spacer sequences that flank a central PvuII site. The unaddressed target site 21P21 is structured identically but lacks the ZF-binding sites. (B) Expression levels of ZF-PvuII G46/A83/F94 and PvuII G46/A83/F94. Cell lysates of transfected HEK293T cells were probed with antibodies against HA-tag or EGFP. (C and D) Cleavage activity in cellula. Cleavage of target plasmids in transfected HEK293T cells was assessed by detecting nuclease-induced mutations due to imperfect repair of DNA DSBs by NHEJ. PCR fragments encompassing the target site were either subjected to digestion with the mismatch-sensitive T7 endonuclease 1 (C) or PvuII (D).