of work: In order to dissect the biochemical steps involved in genetic recombination we have chosen to focus on a key early step(s): homologous pairing and strand exchange between homologous parental DNAs. A fundamental problem in homologous recombination is how the search for homology between the two DNAs is carried out. In all current models a homologous recombination protein, such as the prototypical E. coli RecA protein, loads onto a single-strand DNA generated from one duplex DNA and scans another duplex to form a synaptic (pairing) complex. Eventually, DNA strands are exchanged and a new heteroduplex is formed. While homologous pairing and strand exchange are the earliest contacts between two parental DNAs mediated by RecA and its eukaryotic homologues, Rad51 and Dmc1, homologous recombination is initiated at DNA double-strand breaks (DSBs). The protein that catalyzes DSB formation in meiosis in the budding yeast, Saccharomyces cerevisae, is the product of the SPO11 gene. Surprisingly, Spo11 homologues are dispensable for synapsis in C. elegans and D. melanogaster yet required for meiotic recombination. We have generated a SPO11 mouse knock-out to investigate the biological function of this gene in mammals. Disruption of mouse SPO11 results in infertility. Spermatocytes arrest prior to pachytene with little or no homologous synapsis and undergo apoptosis. Recently, we have been conducting DNA microarray experiments to determine those meiotically expressed genes whose expression is modified by a DSB. In young mice, before degenerative changes have set in as a result of the apoptosis seen in the knockouts, there are only a few dozen genes that are differentially expressed in Spo11 -/- compared to wild type. Among the genes most affected are the Hop2 and Mnd1 genes. These are homologues of genes that affect homologous pairing in yeast. We have generated a knockout of the Hop2 gene in the mouse. Its meiotic phenotype shows a profound meiotic arrest that is unlike any seen previously. Unlike most knockouts with a meiotic phenotype these mice show no synapsis of any kind. That is, whereas most knockouts, show some willy-nilly non-homologous synapsis, spermatocytes from these mice are arrested without almost any synapsis. The chromosomes are somewhat compacted and appear normally decorated with both Rad51 and Dmc1, as if they are on the cusp of synapsis but fail to proceed forward. This finding suggests that the Hop2 protein might play a heretofore-unrecognized central role in bringing meiotic chromosomes together. Prompted by this biologically inspired hypothesis we have purified and studied the biochemical properties of the 25 Kda Hop2 protein. Most gratifyingly, we have been able to show that Hop2 protein can promote the strand invasion (D-loop formation) and strand exchange reactions characteristic of bona fide homologous recombination proteins. Remarkably, with respect to the very important initiating strand invasion activity it is more active than either RecA homologue; Dmc1, mediates strand invasion poorly and Rad51, mediates strand invasion only in the presence of the Rad54 protein. Finally, we have been able to show that Hop2 can act as an accessory protein for the strand exchange carried out by Rad51. Thus, Hop2 is not only a true novel recombinase but also may be the crucial and central protein in initiating interhomolog contacts in meiotic HR. Deletion of the MND1 gene in yeast results in a very similar phenotype to that observed in a hop2 mutant. Furthermore, the two proteins could be immunoprecipitated from yeast meiotic extracts. We have been able to confirm that the two proteins could be coimmunoprecipitated from mouse testes extracts and that, in addition, purified mouse Hop2 and Mnd1 proteins interact in vitro and form a stable heterodimer. Recently, we have shown that this heterodimer, but not the individual proteins by themselves, greatly ssstimulates Rad51- and Dmc1-promoted strand invasion, in the latter case by 35-fold so that almost 70% of the supercoiled substrate is converted to D-loops. Furthermore, we can demonstrate with purified proteins that both Rad51 and Dmc1 interact with the Hop2/Mnd1 heterodimer. Thus, it appears that Mnd1 may act as a chaperone that ensures that the Hop2 pairing activity is only elicited in the presence of the eukaryotic RecA-homologous recombinases, Rad51 and Dmc1. We have been able to prepare mutant Hop2 and Mnd1 proteins that have enabled us to dissect the mechanism by which these proteins stimulate the strand invasion activity of Dmc1. The results indicate that neither the intrinsic strand invasion activity of Hop2 nor the Hop2/Mnd1 DNA binding activity is required for the stimulation of Dmc1 by the Hop2/Mnd1 heterodimer. Furthermore, the Hop2/Mnd1 heterodimer stimulates the binding of Dmc1 to single-strand DNA by 30-fold. Furthermore, the mammalian Hop2/Mnd1 complex facilitates the capture of dsDNA by a ssDNA-Dmc1 nucleoprotein filament. In addition, the heterodimer stabilizes DNA-Dmc1 filaments. Finally, the phenotype of an Mnd1 knockout mouse strongly suggests that, as has been shown in yeast, there are at least two major pathways for the meiotic repair of DSBs in mice and that Hop2 is an essential player in both of these pathways.? ? In the past few years it has become increasingly clear that recombination events are distributed non-randomly in the genomes of eukaryotes. We have used recent developments in computational methods and the availability of extensive human genetic variation data from the HapMap project to generate a high-resolution map of historically stable recombination hotspots of the human genome. Analysis of the recombination rate map revealed the presence of numerous well-defined hotspots of varying strength. We have found a good correlation between the hotspots predicted by us using computational methods and the hotspots previously identified by the analysis of recombination breakpoints in well-studied genomic regions. The overall distribution of hotspots is non-random. There is a correlation of the distribution of hotspots with various genomic and chromosomal features.
Showing the most recent 10 out of 30 publications
