A repair protein (green) that lacks its kinase activity (left) or phosphorylation sites (right) still turns up at the repair site (blue).

ADNA repair protein turns up at the job site even without its tool kit, according to Uematsu et al. (page 219). Its visit is then prolonged by its inefficiency.

Fixing double strand breaks in DNA by nonhomologous end joining (NHEJ) requires a DNA-dependent protein kinase (DNA-PK) to bind to the loose ends of broken DNA and a ligase to do the gluing. This protein machinery must turn up and fix the ends rapidly to minimize the chance that DNA diffusion causes the wrong partners to be glued back together.

Very little is known, however, about the in vivo dynamics of NHEJ. Here, Uematsu et al. describe the dynamics of DNA–PK recruitment in vivo. DNA-PK is composed of two subunits: Ku70/80 and DNA-PKCS. They found that both subunits accumulated at damaged sites within two seconds after targeted DNA breakage. This repair site recruitment was dependent on the Ku70/80 subunit.

DNA-PKCS has a kinase domain and a cluster of phosphorylation sites, both of which are needed for mending DNA. Mutation of either, however, did not impair the enzyme's recruitment speed.

Despite turning up for work as usual, the mutant repair proteins didn't leave as quickly as the wild type. DNA-PKCS is well-known for performing autophosphorylation, and this ability is necessary for DNA repair. Since recruitment of the mutant proteins was normal but repair was deficient, the authors suggest that the main target of DNA-PKCS kinase activity is itself. Autophosphorylation might be the signal that the two loose DNA ends have come together. It most likely triggers the release of DNA-PK and allows access for the ligase to do its gluing.