Mutations in the tumour-suppressor gene BRCA2 cause genomic instability and predisposition to cancer. However, the exact function of BRCA2 in safeguarding genomic integrity in normal cells is unclear. New evidence from Ashok Venkitaraman's group indicates that BRCA2 is required to prevent the breakdown of stalled replication forks and that disruption of this function leads to the chromosomal rearrangements that occur spontaneously in dividing BRCA2-mutant cells.

Although it is not part of the normal replication machinery, BRCA2 localizes in the nuclear foci that are formed when replication is inhibited. So, does BRCA2 function in the cellular response to stalled replication? Venkitaraman and colleagues investigated this by blocking DNA synthesis in mouse cells that were homozygous for a targeted truncation of Brca2Brca2Tr — that produces a non-functional protein. They used two-dimensional gel electrophoresis to visualize Y-arcs — DNA structures that are formed at replication forks — at the ribosomal DNA (rDNA) locus. When replication was inhibited using hydroxyurea (HU), these structures disappeared from Brca2Tr/Tr cells, but persisted in wild-type cells, indicating that Brca2 is required for the stability of stalled replication forks.

In replication-defective bacteria, stalled replication forks can be broken down into linear chromosomal fragments. If a similar process occurred as a result of loss of Brca2 function, this could explain the spontaneous genomic instability that arises as a result of mutations in this gene. To investigate this, the authors used pulsed-field gel electrophoresis (PFGE) to check for the production of DNA fragments that would be predicted to form because of the breakdown of replication forks in the rDNA locus. Chromosomal DNA from this locus is usually too large to be visualized by PFGE, but in Brca2Tr/Tr cells, fragments consistent with breakdown at rDNA replication forks were seen following application of HU. This effect of Brca2 inactivation is not restricted to the rDNA genes, as similar breakdown events were shown to occur at other genomic loci.

Importantly, some DNA breakage was also seen in Brca2Tr/Tr cells that were not treated with HU. This indicates that loss of Brca2 function has a similar effect on the stability of replication forks that stall during normal replication — because of DNA lesions or at natural pause sites — as it does in response to a global arrest of replication.

In response to arrested DNA replication, mammalian cells activate a Chk2-dependent checkpoint mechanism that prevents cells from prematurely undergoing mitosis. Does the role of Brca2 in response to replication arrest lie in the activation of this checkpoint? Venkitaraman and colleagues found that Brca2Tr/Tr cells did not enter mitosis prematurely after treatment with HU, and showed that the activated, phosphorylated form of Chk2 is present in these cells. So, Brca2 functions either downstream of Chk2 or as part of an independent pathway.

The breakdown of replication forks as a result of Brca2 inactivation provides the first example of a link between human disease and defects in the response to arrested DNA replication. It will be interesting to see whether similar mechanisms underlie other conditions in which chromosomal instability is related to cancer predisposition.