of work: Chromosomal double-strand breaks (DSBs) pose a significant and immediate danger to genome integrity and cell survival. DSBs can arise through the action of exogenous agents (radiation and chemicals), endogenous causes (e.g., collapsed replication forks) and in the course of developmental programs (Ig recombination and class switching, and meiosis). DSBs are repaired by either nonhomologous end joining or homologous recombination (HR). Both pathways are equally important in most organisms, including mammals (reviewed in Sonoda et al. (2006) DNA Repair 5, 1021)). ? ? HR is common to all forms of life and is a multistep process involving many gene products. In eukaryotes, the biologically most important form of recombination is the exchange of genetic information between homologous chromosomes (homologs) in meiosis (reviewed in Gerton and Hawley (2005) Nature Rev. Genet. 6, 477). Meiotic recombination is the central phenomenon in the genetics of eukaryotes, ensures the proper segregation of chromosomes at the first division of meiosis (prevents non-disjunction) and is the main force shaping the evolution of genomes. In all organisms, homologous recombination is inextricably related to DNA repair and replication. Rad51 and the meiosis-specific Dmc1 protein, both homologues of bacterial RecA (the prototypical homologous recombination protein), have been identified in most eukaryotes including man and mouse, and homozygous Rad51 -/- mouse ES cells are not viable. Thus, the study of RecA homologues should yield insights into not only homologous recombination but also the regulation of gene stability and cell proliferation.? ? In meiosis in all organisms, including mammals (Romanienko and Camerini-Otero (2000) Mol. Cell 6, 975), Spo11, a type II topoisomerase, cleaves the chromosomal DNA at many sites (a couple of hundred or more in mammals) in each and every nucleus (reviewed in Keeney and Neale (2006) Biochem. Soc. Trans. 34, 523). In mammals there are more endogenous DSBs in meiotic nuclei than in any somatic nuclei in the life of an organism, by a factor of at least 100. Hence the entire arsenal of the HR machinery invoked in somatic cells plus additional meiosis-specific proteins are required for the efficient repair of DSBs in meiosis. These include p53, Brca1, Brca2, Rad51, Dmc1, Atm, Atr, DNA-PKcs, Chk2, Nbsl, Mre11, Rpa, Blm, etc.? ? When Spo11 cleaves chromosomes in meiosis to smithereens (but still held in place by their sisters) they are repaired exclusively by the more accurate HR pathway leading to chromosome pairing. The resulting interhomolog associations (crossovers) ensure the orderly segregation of chromosomes so that each gamete receives one (and only one) chromosome of each pair. The central problem in the process of chromosome pairing during meiosis is arguably the search for homology (Voloshin and Camerini-Otero (2004) Mol. Cell 15, 965). There is considerable evidence that Rad51 and Dmc1, observed as foci at the sites of DSBs, are the central players in this search. Cooperation between Dmc1/Rad51 and Hop2/Mnd1 is likely to be crucial in vivo, since without Hop2 and/or Mnd1, in yeast, Arabidopsis and mouse (Petukhova et al. (2003) Dev. Cell 5, 927) recombination proceeds properly up to the point when Dmc1 and Rad51 are loaded onto the ends of DSBs, but further progression is impaired. This finding suggested that these proteins might play a heretofore-unrecognized central role in bringing meiotic chromosomes together. Prompted by this biologically inspired hypothesis we studied the biochemical properties of Hop2 (25Kda) and Mnd1 (23Kda), proteins that do not share any homology with any previously described proteins except each other. That Hop2 and Mnd1 could be co-immunoprecipitated from mouse testes extracts and that the purified mouse Hop2 and Mnd1 proteins form a stable heterodimer indicate that this heterodimer is the biologically important moiety. Although the purified mouse Hop2 protein is very proficient at strand invasion, the first intermediate in meiotic recombination, and in this respect is much more active than either RecA homologue, disappointingly the strand invasion activity of Hop2 is abrogated upon its association with Mnd1. Unexpectedly, but most reassuringly, however, the Hop2/Mnd1 heterodimer, but neither Hop2 nor Mnd1 themselves, physically interacts with Rad51 and Dmc1 and stimulates their activity up to 35-fold (Petukhova et al. (2005) Nat. Struct. Mol. Biol. 12, 449). Thus, it appears that Mnd1 may act as a matchmaker that channels this heterodimer to stimulate the eukaryotic recombinases, Rad51 and Dmc1, rather than directly pair DNAs. 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. ? ? In order to elucidate the mechanism by which Hop2/Mnd1 stimulates Dmc1 we identified several intermediate steps in the homologous pairing reaction promoted by Dmc1. We showed that Hop2/Mnd1 is required for Dmc1 to promote synaptic complex formation, a step previously revealed only for bacterial homologous recombinases. This synaptic alignment is a consequence of the ability of Hop2/Mnd1 to: 1) stabilize Dmc1-ssDNA nucleoprotein complexes, and 2) facilitate the conjoining of DNA molecules through the capture of dsDNA by the Dmc1-ssDNA nucleoprotein filament (Pezza et al., submitted). To our knowledge, this is the first time that the mechanism for an accessory protein for Dmc1 has been delineated and Hop2/Mnd1 is the only accessory homologous recombination protein that acts on these two critical and separate steps in mammalian meiotic recombination. In summary, we propose that in meiotic recombination Hop2/Mnd1 promotes chromosome synapsis by stabilizing Dmc1 and Rad51 (the latter in collaboration with Patrick Sungs lab at Yale (Chi et al., submitted)) on the single-strand DNA tails of the DSBs and by participating directly, presumably both by protein-protein and protein-DNA interactions, in the capture of potential partner chromosomes.? ? Recently we have investigated whether strand invasion products produced by Rad51, Dmc1 or Hop2 have distinguishable biochemical properties such as their ability to assimilate different degrees of heterology or their either inherent stability or susceptibility to be dissociated by other proteins.
