BRCA2 was one of the first genes to be associated with breast cancer. Mutations in that gene and its sister gene, BRCA1, raise a woman's cumulative lifetime risk of getting breast cancer from about 13 percent to more than 60 percent. And that's a conservative estimate - some studies have found a lifetime risk of up to 80 percent.

But despite the obvious deleterious effects of BRCA mutations, not that much is known to date about how normal BRCA protein actually works.

BRCA2 is involved in repairing certain types of DNA damage, including double-stranded breaks, in which both strands of DNA are broken or damaged. "If only one strand is damaged, repair is relatively easy, because the information needed to repair it is contained in the other strand. But when both strands are damaged, this is not possible," said Nikola Pavletich, chairman of the structural biology program at Memorial Sloan-Kettering Cancer Center. At the same time, repair of such double-stranded breaks is critical because if they are not repaired, they can be a prelude to chromosomal rearrangements, which in turn can cause cancer.

So the cell must copy the information from a sister chromosome or, if there has been no chromosomal duplication in preparation for cell division, from the homologous chromosome. The cell searches for such homologous information by coating the single-stranded DNA with a so-called filament protein, which is able to search for template DNA to repair the double-stranded break.

"The filament is amazing because it can search for the homologous chromosome in the soup of the nucleus," Pavletich told BioWorld Today.

One of the reasons information on its function has been slow in coming is that the BRCA2 protein is quite a whopper.

"The average human protein is maybe 500 amino acids long," Pavletich said. "BRCA2 is 3,418 amino acids long. This is part of why even though BRCA2 was identified in 1995, it's been very slow going to figure out what it does."

Despite the technical challenges of crystallizing large proteins, in research published in Science in 2002, Pavletich and his colleagues managed to show the crystal structure of BRCA2 complexed to a mammalian filament protein, DSS1, and single-stranded DNA. That paper showed BRCA2 "looked like other proteins that bind to special types of DNA damage," such as double-stranded breaks, Pavletich said. The structural studies gave the scientists a hunch as to what the function might be, Pavletich said.

In the Feb. 10, 2005, issue of Nature, Pavletich and his colleagues from Sloan-Kettering and Cornell University's Weill Medical College, both in New York, confirmed that hunch biochemically, and elucidated the details of the mechanism by which the BRCA2 protein helps repair double-stranded breaks.

There's A Fungus Among Us

The scientists used a fungus (Ustilago maydis) in their experiments instead of the more popular baker's yeast (Saccharomyces cerevisiae). "In this particular area of DNA binding, the fungus resembles mammalian cells a lot more than yeast. In fact, the yeast does not have a BRCA2 homologue," Pavletich said. While yeast does have a mechanism to repair double-stranded DNA breaks, when the gene for the homologous protein is deleted in mice, it does not affect their ability to repair double-stranded DNA damage. The fungus, on the other hand, does have a smaller and somewhat simpler BRCA2 homologue, named BRH2, for BRCA homologue.

The scientists used the fungus to investigate the interactions between DNA, Brh2, Rad51, which forms the filament that searches for template DNA for repair, and replication protein A, or RPA. That protein also binds to single-stranded DNA.

In essence, what they found was that Brh2 acts as a "catalyst of a catalyst," by initiating filament formation, as Stephen Kowalczykowski phrased it in an accompanying commentary.

Pavletich explained.

"Say you have 10 Rad51s already. The eleventh one binds to both the tenth Rad51 and the DNA - it gets binding points for both, so to speak. But of course, the first Rad51 can only bind to the DNA; it gets less binding points." As a matter of fact, the first Rad51 racks up too few binding points from just the DNA to displace RPA and get the filament formation started, which is where Brh2 comes in. Even with double-stranded damage, there usually is not a smooth break; instead, there is a single-stranded overhang. The Brh2 protein preferentially binds at the junction between single-stranded and double-stranded DNA, where it helps initiate filament formation by Rad51.

The scientists investigated the quantitative requirements for Brh2 and found that it is effective at substoichiometric quantities - that is, at a Brh2:Rad51 ratio of less than 1-to-1. What that means is that "all you need is one Brh2 to go to the single-stranded/ double-stranded junction. You don't need one Brh2 per Rad51," Pavletich said.

Asked about possible practical applications down the line, Pavletich said there is an "unproven idea" that inhibiting the pathway might prove useful in cancer treatment.

And basic research tends to lead to clinical applications, if at all, via a circuitous route: "By understanding things better, we can control them better, and that hopefully leads to better therapies," he said, adding that "although, of course, human cells are a lot more complicated than fungi."