Induction and repair of zinc-finger nuclease-targeted double-strand breaks in Caenorhabditis elegans somatic cells

Abstract

Zinc-finger nucleases are chimeric proteins consisting of engineered zinc-finger DNA-binding motifs attached to an endonuclease domain. These proteins can induce site-specific DNA double-strand breaks in genomic DNA, which are then substrates for cellular repair mechanisms. Here, we demonstrate that engineered zinc-finger nucleases function effectively in somatic cells of the nematode Caenorhabditis elegans. Although gene-conversion events were indistinguishable from uncut DNA in our assay, nonhomologous end joining resulted in mutations at the target site. A synthetic target on an extrachromosomal array was targeted with a previously characterized nuclease, and an endogenous genomic sequence was targeted with a pair of specifically designed nucleases. In both cases, approximately 20% of the target sites were mutated after induction of the corresponding nucleases. Alterations in the extrachromosomal targets were largely products of end-filling and blunt ligation. By contrast, alterations in the chromosomal target were mostly deletions. We interpret these differences to reflect the abundance of homologous templates present in the extrachromosomal arrays versus the paucity of such templates for repair of chromosomal breaks. In addition, we find evidence for the involvement of error-prone DNA synthesis in both homologous and nonhomologous pathways of repair. DNA ligase IV is required for efficient end joining, particularly of blunt ends. In its absence, a secondary end-joining pathway relies more heavily on microhomologies in producing deletions.

Fig. 1.

Diagram showing a pair of ZFNs bound to DNA. The sequence shown is that of the synthetic QQR target. Each finger is represented by a loop shown contacting 3 bp of DNA; cleavage domains are represented by filled ovals; the N and C termini of the protein are indicated. The expected cut sites on the DNA strands are indicated with triangles.

Fig. 2.

Specific DNA targeting by ZFNs in C. elegans. The locations of PCR products resistant (R) and sensitive (S) to the diagnostic enzymes in both panels are indicated. (A) Alteration of a synthetic, extrachromosomal target by induction of the QQR nuclease. A 1-kb region around the QQR target was amplified by PCR and digested with MluI. L, linear size standards; −, a nematode not heat-shocked; +, plasmid positive control; 1–3, individual heat-shocked nematodes. (B) Alteration of the genomic Nw target with specifically engineered nucleases. A 752-bp genomic segment containing the Nw target was amplified and digested with HindIII. L, linear size standards; −, two nematodes that did not carry the NwA and NwB nucleases; 1,2, two heat-shocked nematodes carrying NwA, NwB.

Fig. 3.

PCR products from the QQR target after ZFN induction in wild-type (A) and lig-4 (B) nematodes. The unmodified product is 207 bp long. The format is as in Fig. 2.

Fig. 4.

Models of repair pathways after ZFN cleavage. The double-strand break has 4-bp 5′ overhangs that can be filled in completely or partially by DNA polymerase and the resulting blunt ends ligated. This process depends on DNA ligase IV. Alternatively, 5′->3′ resection at the ends leaves single-stranded 3′ tails that can be substrates for either template-dependent gene conversion or for a nonhomologous end-joining process leading to deletions. Gene conversion would restore the original sequence, which could then be recut and reprocessed. End joining could occur by removal of the single-stranded tails and blunt joining, relying on DNA ligase IV. An alternative end-joining process, independent of ligase IV, relies on microhomologies; the one illustrated envisions extension of a transient microhomology-based primer–template complex by DNA polymerase (41).