It's important for the life of a cell that its chromosomes remain intact. DNA breakage can occur following radiation exposure, chemical damage, or viral infection. In some cases, breakage repair results in DNA pieces attached to the wrong chromosome. This phenomenon, known as a translocation, can cause certain cancers, such as acute myeloid leukemia.
In the latest issue of Molecular Cell, Elliott et al. investigated how mammalian cells rejoin DNA strands during translocation. Elliott et al. were especially interested in how translocations occur around the gene MLL, which is frequently the site of translocations in people with acute myeloid leukemia. In addition, they wanted to know in whether short, highly related DNA repeat regions known as Alu sequences , which appear to form hotspots of DNA breakage and rejoining events in tumors, could influence the mechanism of DNA repair.
Translocation breaks can be repaired either with loose ends, carrying possible insertions or deletions, simply joined together (this is called NHEJ, for non-homologous end joining), or via a mechanism called SSA (for single-strand annealing), which uses a match in sequences near the breakpoints as a template to "edit" the new junction and remove one copy of the match. (These mechanisms plus a third mechanism, homologous recombination, are also seen in chromosomal reattachments, when a fragment is re-joined to its original chromosome. In this more normal situation, junction editing makes sense, as extra DNA can be as damaging in the wrong context as a stray semicolon in computer code.)
To study the events surrounding translocation, the authors designed a test (see the picture) where genes encoding resistance to two different drugs (shown as a green region and a red region) were broken in two, and the halves were swapped, joined by a controlled breakage point called a restriction site and an Alu repeat (the blue box). Following DNA strand breakage, about 1 cell in 100,000 showed a chromosomal translocation in which the drug resistance genes were reconstituted. These rare cells were now drug resistant. The neomycin resistance gene (green) had additionally been designed with splicing donor and acceptor sites, a trick which allowed expression no matter how the translocation was joined. But for a cell to achieve puromycin resistance (red), a further constraint was that overlapping sequences (shown as complementary triangles) had to be edited via SSA. Thus, by evaluating the puromycin resistance of neomycin resistant cells, they could see whether the translocation occured by NHEJ or by SSA. (Of course the cells could elect to make one translocated juction by SSA and the other by NHEJ. )
Using this system, they found by far the most common translocation event (80% of all events) utilized SSA on both chromosomes. But they got a suprise when they tinkered with the system. The intervening Alu sequences (the blue box) had been identical to one another in the first run, but when they replaced one Alu sequence with a 25% divergent Alu sequence, neomycin resistance gene (green) translocations now occured almost entirely via NHEJ. The change to NHEJ was not the result of disabling the SSA machinery, as the translocations creating puromycin resistance continued to favor SSA. The total frequency of translocation was not affected, even though a different repair mechanism was used. But this was not completely random NHEJ. The weakly homologous Alu sequences somehow biased the joining event, because deletions associated with the NHEJ often occured within short stretches of identity between the Alu repeats!
The most important observation from this experiment is that translocations are strongly affected by the degree of sequence homology near the broken ends to be joined. If, as in the first experiments, the sequences match closely, an SSA mechanisn will edit the junction so that an overlap is removed. But translocations occuring near moderately related Alu repeats cannot evoke SSA, and instead favor NHEJ joining, with however a bias to the joining. The exact role of Alu repeats in biasing the joining remains a bit mysterious. The authors believe that the NHEJ is sensitive somehow to microhomologies within divergent Alu sequences in their system, and that this sensitivity may be significant in clinical leukemias as well.