Autoinhibition of the Nuclease ARTEMIS Is Mediated by a Physical Interaction between Its Catalytic and C-terminal Domains.

The nuclease ARTEMIS is essential for the development of B and T lymphocytes. It is required for opening DNA hairpins generated during antigen receptor gene assembly from variable (V), diversity (D), and joining (J) subgenic elements (V(D)J recombination). As a member of the non-homologous end-joining pathway, it is also involved in repairing a subset of pathological DNA double strand breaks. Loss of ARTEMIS function therefore results in radiosensitive severe combined immunodeficiency (RS-SCID). The hairpin opening activity is dependent on the DNA-dependent protein kinase catalytic subunit (DNA-PKcs), which can bind to and phosphorylate ARTEMIS. The ARTEMIS C terminus is dispensable for cellular V(D)J recombination and in vitro nuclease assays with C-terminally truncated ARTEMIS showing DNA-PKcs-independent hairpin opening activity. Therefore, it has been postulated that ARTEMIS is regulated via autoinhibition by its C terminus. To obtain evidence for the autoinhibition model, we performed co-immunoprecipitation experiments with combinations of ARTEMIS mutants. We show that an N-terminal fragment comprising the catalytic domain can interact both with itself and with a C-terminal fragment. Amino acid exchanges N456A+S457A+E458Q in the C terminus of full-length ARTEMIS resulted in unmasking of the N terminus and in increased ARTEMIS activity in cellular V(D)J recombination assays. Mutations in ARTEMIS-deficient patients impaired the interaction with the C terminus and also affected protein stability. The interaction between the N- and C-terminal domains was not DNA-PKcs-dependent, and phosphomimetic mutations in the C-terminal domain did not result in unmasking of the catalytic domain. Our experiments provide strong evidence that a physical interaction between the C-terminal and catalytic domains mediates ARTEMIS autoinhibition.

Physical incorporation of a single-stranded oligodeoxynucleotide during targeted repair of a human chromosomal locus.

BACKGROUND: Targeted gene repair is an attractive method to correct point-mutated genes at their natural chromosomal sites, but it is still rather inefficient. As revealed by earlier studies, successful gene correction requires a productive interaction of the repair molecule with the target locus. The work here set out to investigate whether DNA repair, e.g., mismatch repair, or a direct incorporation of the correction molecule follows as the step upon the initial interaction. METHODS: Single-stranded 21mer oligodeoxynucleotides (ODNs) of sense orientation were directed towards point-mutated enhanced green fluorescence protein transgene loci in HEK-293-derived cell clones. First gene repair assays compared ODNs carrying the canonical termini 5'-phosphate and 3'-OH with their respective variants harbouring non-canonical termini (5'-OH, 3'-H). Second, a protocol was established to allow efficient recovery of integrated short biotin-labelled ODNs from the genomes of gene-corrected cells using streptavidin-coated beads in order to test directly whether transfected ODNs become bona fide parts of the target locus DNA. RESULTS: Oligodeoxynucleotides with canonical termini were about 34-fold more efficient than their counterparts carrying non-canonical termini in a phosphorothioate-modified backbone. Furthermore, biotinylated fragments were successfully recovered from genomic DNAs of gene-corrected cells. CONCLUSIONS: The experiment showed that ODNs are incorporated into a mammalian genome. This unravels one early repair step and also sets an unexpected example of genome dynamics possibly relevant to other ODN-based cell techniques.

Targeted chromosomal gene modification in human cells by single-stranded oligodeoxynucleotides in the presence of a DNA double-strand break.

A DNA double-strand break (DSB) cannot be tolerated by a cell and is dealt with by several pathways. Here, it was hypothesized that DSB induction close to a targeted mutation in the genome of a mammalian cell might attract oligodeoxynucleotide (ODN)-directed gene repair. A HEK-293-derived cell line had been engineered harboring a single target locus with open reading frames encoding the living-cell reporter proteins LacZ and EGFP, the latter translationally decoupled by a DNA spacer with a unique I-SceI recognition site for defined DSB induction. To enable expression of a fluorescent LacZ-EGFP fusion protein, single-stranded (ss) ODNs (80 or 96 nucleotides long) spanning the DSB were designed to fuse both reading frames by altering a few base-pair positions, deleting 59 bp or introducing a 10-bp fragment. The ssODNs alone generated few EGFP-positive cells. With I-SceI transiently expressed, more than 0.3% of cells revealed EGFP expression 7 days after transfection, with up to 96% of the loci faithfully corrected, depending on the ssODN used. During these correction events, the ssODN did not become physically incorporated into the chromosome, but served only as information template. Unwanted insertional mutagenesis also occurred. Both observations have important implications for gene therapy.