Portal de Periódicos da CAPES

A Super New Twist on the Initiation of Meiotic Recombination

1997; Cell Press; Volume: 89; Issue: 2 Linguagem: Inglês

10.1016/s0092-8674(00)80194-4

1097-4172

James E. Haber,

Acute Lymphoblastic Leukemia research

In meiosis, recombination between homologous chromosomes occurs at least 1000 times more frequently than it does in mitotic cells. This high level of genetic exchange does not simply provide a facile way to generate diversity in sperm or eggs (or spores) from parents harboring hundreds of heterozygosities. Crossing-over also serves a key role in ensuring that paired, homologous chromosomes will properly segregate at the first meiotic division (for general reviews of meiotic recombination, see17Roeder G.S Proc. Natl. Acad. Sci. USA. 1995; 92: 10450-10456Crossref PubMed Scopus (107) Google Scholar, 9Kleckner N Proc. Natl. Acad. Sci. USA. 1996; 93: 8167-8174Crossref PubMed Scopus (344) Google Scholar). Chromosomes that fail to experience at least one crossover show very elevated rates of nondisjunction, leading to aneuploid offspring. It has also been known for some time that topoisomerase II is necessary to disentangle intertwined recombined chromsomes after recombination; but surprisingly, a novel topoisomerease II is also needed to initiate recombination (2Bergerat A de Massy B Gadelle D Varoutas P.-C Nicolas A Forterre P Nature. 1997; 386: 414-417Crossref PubMed Scopus (657) Google Scholar, 7Keeney S Giroux C.N Kleckner N Cell. 1997; 88: 375-384Abstract Full Text Full Text PDF PubMed Scopus (1174) Google Scholar). Meiotic recombination is not simply the heating-up of a gently simmering process found in mitotic cells. There are many important differences. First, recombination is not stimulated to the same extent in different regions of the genome; there are prominent “hot spots” where recombination is 10–50 times more frequent than at other locations. Second, the proportion of gene conversions that are accompanied by a reciprocal crossover is much higher in meiotic cells. Third, recombination in mitotic cells occurs much more frequently between sister chromatids than between homologous chromosomes, while recombination in meiotic cells is much more frequent between homologs. Finally, the distribution of crossovers along a chromosome is highly regulated, reflected in the phenomenon known as interference, whereby the number of exchange events per chromosome arm is kept within a very narrow range. Lurking around all of these events is a large, poorly understood structure—the synaptonemal complex (SC)—that has been invoked to play a key role in all of these aspects of meiotic recombination and chromosome segregation. Soon after premeiotic DNA synthesis, the sister chromatids become associated with a proteinaceous linear structure known as the axial element. By the time that homologous chromosomes pair and recombine, the two axial elements of each pair of homologous chromosomes have become incorporated into the tripartite structure of the SC. Only a small percentage of the DNA is directly associated with these structures, but we do not know if some sequences are preferentially associated with axial element components or how this might influence where and when recombination occurs. In Saccharomyces cerevisiae, it has been recognized for some time that the predominant stimulus of recombination, both in mitosis and in meiosis, is the formation of double-strand breaks (DSBs). In mitotic cells, recombination initiated by site-specific endonucleases such as HO (which initiates mating-type gene switching) or I-SceI (responsible for intron homing) have received great attention, and much is known about the sequence of molecular steps that occur after the creation of a DSB (5Haber J BioEssays. 1995; 17: 609-620Crossref PubMed Scopus (159) Google Scholar). The ends of a DSB are first resected by one or more 5′-to-3′ exonucleases, producing long 3′-ended single-stranded DNA that can invade an intact homologous donor sequence and initiate new DNA synthesis, leading to the repair of the DSB. A key step in understanding meiotic recombination came with the observation that hot spots of meiotic recombination were also sites of DSBs, which were resected to produce 3′-ended tails (20Sun H Treco D Szostak J.W Cell. 1991; 64: 1155-1161Abstract Full Text PDF PubMed Scopus (416) Google Scholar). Physical analysis of meiotic recombination also produced clear evidence of another expected intermediate in DSB-mediated recombination, the branched DNA structure known as a double Holliday junction (18Schwaca A Kleckner N Cell. 1995; 83: 783-791Abstract Full Text PDF PubMed Scopus (386) Google Scholar). But how do meiotic DSBs arise and what is their relation to the restriction endonuclease-like cleavages made by HO and I-SceI? A search for mutation that would prevent such cleavages and thus yield only nonrecombined (and nondisjoined) chromosomes provided an embarrassment of riches: there are at least ten genes that are needed to induce meiotic recombination. However, none of their predicted protein sequences especially recommended any one of them as “the endonuclease.” An important step in deciphering these events came with the study of a special mutation of the RAD50 gene known as rad50S (1Alani E Padmore R Kleckner N Cell. 1990; 61: 419-436Abstract Full Text PDF PubMed Scopus (466) Google Scholar). While a deletion of RAD50 completely prevents recombination and the creation of DSBs, the rad50S mutation permits the formation of DSBs; however, the breaks remain undegraded and subsequent steps in recombination are absent. The persistence of DSBs in turn made it possible to investigate the nature of the breaks formed at various hot spots. Coupled with chromosome-separating gels, it has been possible to map the prominent DSB sites along entire chromosomes (reviewed by10Lichten M Goldman A.S.H Annu. Rev. Genet. 1995; 29: 423-444Crossref PubMed Scopus (292) Google Scholar). The sites of DSBs at several meiotic hot spots were examined at nucleotide resolution. It became quickly apparent that, unlike the cleavage sites of site-specific endonucleases, hot spots could be cleaved at a number of nearby sites within a 100–200 bp region. What distinguishes hot spots from other DNA seems to be their chromatin structure: nearly all hot spots are found in promoter regions, and the promoters of most transcribed genes are cleaved. The sites of DSB formation are also well-correlated with regions of “open” chromatin, as measured by DNase I and micrococcal nuclease hypersensitive sites. However, it is not at all clear why some genes are much “hotter” than others. Does this correlate with the level of transcription, the presence of particular classes of transcription factors that remodel chromatin, or the preferential association of some regions of the chromosome with the axial elements? The mapping of DSBs at the nucleotide level produced an even larger surprise: in rad50S strains, the 5′ ends of the DSBs were covalently attached to a protein. Moreover, rad50S was not the only mutant that would leave unresected DSBs: mutations in another gene, known both as COM1 or SAE2, produced a rad50S-like phenotype (13McKee A.H.Z Kleckner N Genetics, in press. 1997; Google Scholar, 14Prinz S Amon A Klein F Genetics, in press. 1997; Google Scholar). At least some DNA ends had 2 bp, 5′ overhangs. A short, intense period of speculation about the identity of the covalently attached protein has ended with the recent publication of two papers (2Bergerat A de Massy B Gadelle D Varoutas P.-C Nicolas A Forterre P Nature. 1997; 386: 414-417Crossref PubMed Scopus (657) Google Scholar, 7Keeney S Giroux C.N Kleckner N Cell. 1997; 88: 375-384Abstract Full Text Full Text PDF PubMed Scopus (1174) Google Scholar) that identify it as the protein encoded by SPO11, one of the first meiotic-specific genes shown to be required for the initiation of meiotic recombination (8Klapholz S Waddell C.S Esposito R.E Genetics. 1985; 110: 187-216Crossref PubMed Google Scholar). In a new twist on programmed DNA breaks, SPO11 turns out to be related to a novel family of type II topoisomerases. 7Keeney S Giroux C.N Kleckner N Cell. 1997; 88: 375-384Abstract Full Text Full Text PDF PubMed Scopus (1174) Google Scholar identified Spo11 by purifying it from the covalently attached ends of DSBs made from rad50S cells and obtaining peptide sequence information. Antibodies against Spo11p were shown to immunoprecipitate Spo11 and the expected covalently attached fragments of DNA from one particular hot spot. The efforts of the Nicolas and Forterre labs have added additional weight to the argument, by providing evidence that Spo11 does not simply become fixed to the ends of the DSBs; rather, it is very likely to be the nuclease itself. 2Bergerat A de Massy B Gadelle D Varoutas P.-C Nicolas A Forterre P Nature. 1997; 386: 414-417Crossref PubMed Scopus (657) Google Scholar showed that Spo11 is homologous to a novel family of type II topoisomerases first identified in archaebacteria (where it is called topoisomerase VI). A conserved tyrosine in Spo11 that is postulated to act as the catalytic residue in phosphodiester bond cleavage was mutated to phenylalanine, and the resulting spo11-YF135 was unable to induce meiotic DSBs. Archaebacterial topoisomerase VI is composed of two subunits, but homology to only one of them (SPO11) is convincingly found in the Saccharomyces Genome Database. This may suggest that the function of this enzyme has evolved significantly between archae and fungi. Type II topoisomerases normally create a transient double-strand cleavage, with the protein covalently attached to the 5′ DNA ends, and then rejoin the ends. Usually one can only recover the protein-attached DNA when the topoisomerase is denatured or treated with specific inhibitors; however, the appearance of a similar intermediate can also be genetically regulated. In E. coli, the F-factor CcdB protein interacts with gyrase to produce such breaks, leaving gyrase covalently attached to the 5′ ends of the DNA (3Bernard F Kezdy K.E van Melderen L Steyaert J Wyns L Pato M.L Higgins P.N Couturier M J. Mol. Biol. 1993; 234: 534-541Crossref PubMed Scopus (150) Google Scholar). This state is analogous to the DSBs produced in rad50S and com1/sae2 mutants (Figure 1). This raises an interesting question: is Spo11 capable of carrying out the complete topoisomerase reaction or has it become limited to generating the initial, protein-bound cleavage that is subsequently processed by other proteins into a resected and recombinogenic DSB? The presumed properties of Spo11 protein may provide attractive explanations for the regulation of meiotic DSB formation. Intuitively, it would seem disastrous to have a meiotic cell creating 100 or more DSBs with no sort of constraint. It is difficult to imagine 200 broken chromosome ends all simultaneously trying to locate homologous sequences with which such breaks could be repaired, even in an organism such as budding yeast with very few dispersed repeated sequences. One way to ensure that meiotic DSBs would be less likely to create meiotic mayhem would be to require that a region should be paired with a homologous partner before a DSB is permitted to be made. Spo11 could repeatedly create a protein-bound cleaved intermediate and rejoin the DNA (Figure 1). This cycle might be interrupted to produce a recombinogenic break only when a second, homologous region is in proximity to the enzyme. This would, however, require that Spo11 or some accessory protein be able to distinguish homologous from nonhomologous duplex DNA prior to the creation of a recombinogenic DNA end. However, such prior pairing cannot be a prerequisite for cleavage, for DSBs occur in haploids that are deceived to enter meiosis (where there are no homologous chromosomes) at levels and kinetics similar to what is seen in diploids (4Gilbertson L Stahl F.W Proc. Natl. Acad. Sci. USA. 1994; 91: 11934-11937Crossref PubMed Scopus (37) Google Scholar). One could also imagine ways in which Spo11 might be regulated to account for other unusual features of meiotic DSB formation. For example, genetic evidence suggests that, even at the hottest hot spot, DSBs are rarely created on both sister chromatids of one homolog. This situation is distinctly different from what occurs when the HO endonuclease is expressed in meiotic cells, where both chromatids are frequently cleaved (11Malkova A Ross L Dawson D Hoekstra M Haber J.E Genetics. 1996; 143: 741-754PubMed Google Scholar). Moreover, several labs have shown that a high level of DSBs at one site somehow regulates the frequency with which a nearby DSB appears (10Lichten M Goldman A.S.H Annu. Rev. Genet. 1995; 29: 423-444Crossref PubMed Scopus (292) Google Scholar). Perhaps Spo11 sits on the axial element but only at a few sites along each chormosome arm. From this position, Spo11 might be able to cleave DNA that is tethered near the axial element, thus limiting cleavage to only one of several alternative sites. A more complex scenario would be needed to account for observations that the frequency with which DSBs are created on one chromosome is influenced by the extent of homology it shares with its homolog (9Kleckner N Proc. Natl. Acad. Sci. USA. 1996; 93: 8167-8174Crossref PubMed Scopus (344) Google Scholar, 16Rocco V Nicolas A Genes to Cells. 1996; 1: 645-661Crossref PubMed Scopus (47) Google Scholar). It is possible that Spo11 might also play a role in homolog pairing prior to the formation of DSBs. Studies of transient chromosome associations in meiotic cells suggest that spo11 diploids are more severely impaired in recombination-independent chromosome colocalization than other mutants such as rad50 (9Kleckner N Proc. Natl. Acad. Sci. USA. 1996; 93: 8167-8174Crossref PubMed Scopus (344) Google Scholar). 7Keeney S Giroux C.N Kleckner N Cell. 1997; 88: 375-384Abstract Full Text Full Text PDF PubMed Scopus (1174) Google Scholar suggest that the ability of topoisomerase-like proteins to bind two DNA helices might allow such associations, though it is unclear how homologous regions would be identified and why other mutations that abolish Spo11's cleavage function (e.g., a complete deletion of rad50) have less effect on chromosome colocalization. Alternatively, Spo11 might intertwine homologs that have been colocalized by other proteins. Spo11 does not act alone to create DSBs. At least nine other gene products are required before meiotic DSBs are generated (17Roeder G.S Proc. Natl. Acad. Sci. USA. 1995; 92: 10450-10456Crossref PubMed Scopus (107) Google Scholar, 9Kleckner N Proc. Natl. Acad. Sci. USA. 1996; 93: 8167-8174Crossref PubMed Scopus (344) Google Scholar). Three of these genes (RAD50, XRS2, and MRE11) are also expressed in mitotic cells and clearly play a much more complicated role in the life of chromosomes than simply ensuring the formation of a meiotic DSB. Deletions of these genes have notable effects on homologous recombination and especially on nonhomologous end-joining. Deletion of any one of these genes slows down, but does not prevent, HO-induced recombination, at least in part by reducing 5′-to-3′ exonuclease activity (5Haber J BioEssays. 1995; 17: 609-620Crossref PubMed Scopus (159) Google Scholar). Interestingly, RAD50 and MRE11 share homology with the bacterial endo-/exonuclease SbcCD (19Sharples G.J Leach D.R.F Mol. Microbiol. 1995; 17: 1215-1220Crossref PubMed Scopus (188) Google Scholar). These observations are certainly consistent with the idea that Rad50S protein is unable to remove Spo11 protein from the ends of DSBs and thus they accumulate as unresected ends. Rad50, Mre11, and Xrs2 proteins are also needed to create meiotic DSBs in the first place; but their presence is not sufficient to create favorable cutting conditions. MEI4, REC102, REC104, and MER2, all meiotic-specific genes, are also required for DSB formation. Moreover, deletion of HOP1 or RED1, two components of the axial element, markedly reduce DSB formation (12Mao-Draayer Y Galbraith A.M Pittman D.L Cool M Malone R.E Genetics. 1996; 144: 71-86PubMed Google Scholar). (Caveat lector: one endemic problem in studying yeast meiosis is that a number of mutants have signficantly different phenotypes in different strain backgrounds; hence, Mao-Draayer et al. [1996] find red1 strains to have markedly reduced DSB formation in rad50S while (21Xu L Weiner B.M Kleckner N Genes Dev. 1997; 11: 106-118Crossref PubMed Scopus (184) Google Scholar find normal amounts of DSBs in rad50S but significantly reduced levels in a Rad50+ strain.) In any case, one is left with the impression that Spo11 only acts in the context of an assembled structure that may include both components of the the axial element and the Rad50–Xrs2–Mre11 complex to cleave DNA associated with that structure. This might be set up in such a way that Spo11 can only cleave one of two sister chromatids, for example, and also can explain why some promoter regions are much hotter than others. Our present state of understanding raises a number of interesting questions, which will most likely be answered in the near future. (1) Does Spo11 turn over or does it commit suicide by making a single, irreversible cleavage? (2) With what other proteins does Spo11 interact and where are they located? Rad50 and Mre11 proteins have been shown to interact (6Johzuka K Ogawa H Genetics. 1995; 139: 1521-1532Crossref PubMed Google Scholar), but no interaction with Spo11 has been reported. Perhaps Spo11 is modified or requires association with a meiotic-specific protein before it can join the party. 2Bergerat A de Massy B Gadelle D Varoutas P.-C Nicolas A Forterre P Nature. 1997; 386: 414-417Crossref PubMed Scopus (657) Google Scholar remind us that topoisomerase VI in archaebacteria has two subunits, the other of which has some weak homology to a heat shock protein that is specifically induced in meiosis. Possibly the second subunit is involved in regulating the timing of DSB formation. (3) How are DSBs regulated, both between sister chromatids and along chromosomes? How does cleavage at one site exclude nearby cleavages? (4) How are the DSBs formed in wild-type cells processed? How is Spo11 released? 7Keeney S Giroux C.N Kleckner N Cell. 1997; 88: 375-384Abstract Full Text Full Text PDF PubMed Scopus (1174) Google Scholar point out that the enzyme–DNA bond could be hydrolyzed, leaving a 5′ phosphate that a 5′-to-3′ exonuclease could then degrade, or that a combination of a helicase and an endonuclease could cleave the 5′-ended DNA strand at some distance from the end. In this regard it would be interesting to understand more about how other protein–DNA ends are processed in yeast, for the removal of Spo11 is not without precedent. Trans-kingdom DNA transfer, from E. coli to yeast or from Agrobacterium to yeast, has been well documented in mitotic cells. Especially in the latter case, it is evident that DNA transfer into yeast depends on the VirD2 protein bound to the 5′ end of Ti DNA and that subsequently the originally protein-attached end can participate in homologous and nonhomologous recombination (15Risseeuw E Franke-van Dijk M.E.I Hooykaas P.J.J Mol. Cell. Biol. 1996; 16: 5924-5932PubMed Google Scholar). Moreover, Spo11-blocked DNA ends can apparently be repaired when meiotic cells are returned to mitotic growth conditions before cells undergo meiotic chromosome segregation (1Alani E Padmore R Kleckner N Cell. 1990; 61: 419-436Abstract Full Text PDF PubMed Scopus (466) Google Scholar, 12Mao-Draayer Y Galbraith A.M Pittman D.L Cool M Malone R.E Genetics. 1996; 144: 71-86PubMed Google Scholar). How are these ends removed and what relationship does this process have to the processing of meiotic DSBs? (5) What other properties of meiotic recombination (the high frequency of associated crossing-over and the whole phenomenon of crossover interference) are unique to Spo11-generated DSBs? Would meiotic expression of an HO-induced DSB give rise to meiotic- or mitotic-like recombination? Finally, there is the big question: do the events that have been so well delineated in Saccharomyces provide a model for how meiotic recombination occurs in other organisms? The finding that SPO11 is homologous to Schizosaccharomyces rec12 (2Bergerat A de Massy B Gadelle D Varoutas P.-C Nicolas A Forterre P Nature. 1997; 386: 414-417Crossref PubMed Scopus (657) Google Scholar, 7Keeney S Giroux C.N Kleckner N Cell. 1997; 88: 375-384Abstract Full Text Full Text PDF PubMed Scopus (1174) Google Scholar) offers some hope, despite the fact that intensive searches for DSBs at meiotic hot spots in fission yeast have not yet been successful. The discovery that SPO11 is also homologous to an as-yet-uncharacterized Caenorhabditis sequence will soon enable us to ascend another rung of the evolutionary ladder in learning how this critical process occurs in humans.