Zinc finger protein-dependent and -independent contributions to the in vivo off-target activity of zinc finger nucleases

Abstract

Zinc finger nucleases (ZFNs) facilitate tailor-made genomic modifications in vivo through the creation of targeted double-stranded breaks. They have been employed to modify the genomes of plants and animals, and cell-based therapies utilizing ZFNs are undergoing clinical trials. However, many ZFNs display dose-dependent toxicity presumably due to the generation of undesired double-stranded breaks at off-target sites. To evaluate the parameters influencing the functional specificity of ZFNs, we compared the in vivo activity of ZFN variants targeting the zebrafish kdrl locus, which display both high on-target activity and dose-dependent toxicity. We evaluated their functional specificity by assessing lesion frequency at 141 potential off-target sites using Illumina sequencing. Only a minority of these off-target sites accumulated lesions, where the thermodynamics of zinc finger-DNA recognition appear to be a defining feature of active sites. Surprisingly, we observed that both the specificity of the incorporated zinc fingers and the choice of the engineered nuclease domain could independently influence the fidelity of these ZFNs. The results of this study have implications for the assessment of likely off-target sites within a genome and point to both zinc finger-dependent and -independent characteristics that can be tailored to create ZFNs with greater precision.

Figure 1.

Zinc finger nucleases targeting exon2 of kdrl: (A) A schematic drawing showing the ZFNs bound to the target site. The two ZFN monomers (ZFNL and ZFNR) bind respectively to the 9 bp 5′- and 3′-half sites through the associated ZFPs (fingers indicated by numbered ovals), which position the heterodimeric nuclease domain over the 6 bp spacer between the two ZFP half sites. (B) The DNA-binding specificities of the kdrl ZFPs determined at high stringency (5 mM 3-AT) using the B1H system displayed as a sequence logo (46,47).

Figure 2.

Overview of the off-target analysis for the original kdrl ZFNs. (A) The number of active (red) and inactive (blue) off-target sites is depicted in the graph. The sites are subdivided according to the type of site (homodimeric or heterodimeric), where heterodimeric sites were divided into five different groups based on the spacing between the two half-sites. A total of eight active off-target sites (see text for criteria) were found in normal embryos from 10 pg ZFN dose (Table 1), and 11 additional off-target sites were active either only in deformed embryos from 10 or 20 pg ZFN dose or in one of the two biological replicates, as described in the text. (B) Dose-dependent effects of kdrl-ZFNs on its in vivo activity and precision. The lesion frequency was plotted for the on-target (blue) and eight off-target sites active in the morphologically normal embryos at the 10 pg dose.

Figure 3.

Characteristics of active off-target sites. (A) The distribution of the number of matches to the target site for active (red) and inactive (blue) off-target sites with 5- or 6-bp spacing is shown. (B) Base frequency at each position in the ZFPL and ZFPR binding sites are displayed as a logo for the group of 18-active off-target sites. Guanines at seven positions (red boxes) in the binding sites were absolutely conserved within these sequences, while they are more variable in the inactive sites (Supplementary Figure S5). (C) The distribution of the number of active (red) and inactive (blue) off-target sites as a function of the number of guanines preserved in each recognition sequence. (D) Each base in the binding site of the ZFPL and ZFPR was independently mutated to cytosine, which is not found at any position in either of the ZFP recognition sequences, and its influence on ZFP binding was assayed using B1H-based activity assay (38) at 1 mM 3-AT. This assay will detect only the most important positions for recognition, where a reduction in cell survival (plotted as the –log of surviving colonies) indicates a position important for recognition. All of the absolutely conserved guanines—indicated by an asterisk—are critical for activity.

Figure 4.

The specificity of the ZFP domains influences the precision of ZFNs. (A) Binding site specificities of the new and old ZFPs determined using the B1H system displayed as Sequence logos (46,47). The recognition helix sequences for each finger are displayed where the amino acids that differ in the nZFPs are indicated in red. Red rectangles highlight the positions where information content of the desired base was higher in the improved ZFNs. (B) Comparison of lesion frequencies at the on-target site and the 8 active off-target sites were plotted for oZFN- and nZFN-treated embryos. The color scheme remains same as in Figure 2. Lesion frequency at the off-target sites in nZFN-treated embryos was reduced at all but one off-target site.

Figure 5.

Influence of the type of the engineered nuclease domain (DD/RR or EL/KK) on the precision of the original ZFNs. Injection of 50–100 pg of mRNA for oZFN^ELKK was required to achieve on-target lesion frequency similar to 10 pg of oZFNs^DDRR. The lesion frequencies for normal embryos treated with 10 pg dose of ZFNs^DDRR or 100 pg dose of oZFN^ELKK were plotted for the on-target site and a subset (six of eight) of the active off-target sites for ZFNs^DDRR that were assayed in this experiment.