Repair of DNA by recombination is a high fidelity molecular system whose purpose is to restore genetic material lost as a result of breakage or other severe forms of damage to DNA molecules. The process relies on utilization of a homologous DNA sequence as a template to direct repair of the damaged molecule. The paradigm for recombinational repair holds that the broken molecule is resected from the end to reveal a protruding single strand which then pairs with the homologous sequence in the undamaged molecule forming a joint molecule with a displaced strand or D-loop. After some additional processing that includes a limited amount of DNA synthesis at the D-loop and that may include resolving a Holliday junction, the broken DNA is spliced back together with lost information restored.

The phase in this process that has held particular fascination for me over the course of my scientific career is the homologous pairing step. How are homologous sequences located within DNA molecules and how is the D-loop formed? The laboratory has worked on aspects of this problem for many years using both traditional biochemical means to purify a D-loop forming activity and molecular genetic means to identify the relevant structural genes. The seminal observation in the field made in Charles Radding's lab at Yale over twenty-five years ago was that the key catalytic component in D-loop formation in E. coli was the RecA protein. More recently after genetic studies and database analysis revealed related proteins to be present in eukaryotes, it was shown that the Rad51 protein was the primary player in performing the reaction in eukarytoes.

My laboratory uses a non-mainstream yeast, Ustilago maydis, for experimental work and, in keeping with convention, we found that its Rad51 is fairly standard in promoting the D-loop reaction. What was unexpected was the finding that there is a second protein, related to Rad51 but extremely divergent structurally, that can also promote D-loop formation. While there is nothing unusual in finding an additional Rad51-related protein in U. maydis, (eukaryotes generally have several Rad51 paralogs), it was extraordinary to find one that promotes formation of D-loops. We know that the proteins interact physically and have been trying to understand the mechanistic details of why and how they are functionally dedicated in their actions. This has led us to examine the basic genetic composition of the recombinational repair system in terms of defining other genes in the pathway and in doing so we came across several remarkable features.

In the mainstream yeast world the pivotal gene for all of all recombination genes is RAD52. Not so in U. maydis. There is a highly conserved Rad52 (and no other apparent homolog) but deleting the structural gene results in essentially no phenotype by any of our DNA repair or recombination assays. The key player controlling recombinational repair in U. maydis seems to be a homolog of BRCA2 the product of one of the major breast cancer predisposition genes in human. It in turn is dependent on a tiny acidic protein called Dss1 that binds to it forming a tight complex. The path we are on is to perform experiments that could help sort out the interplay of these protein and to understand their roles in the molecular mechanisms of D-loop dynamics. Our particular challenge is to draw on the current resources and developing technology to design the best experiments addressing the fundamental issues.