Meiotic sister chromatid cohesion and recombination in two filamentous fungi

<p>Homologous recombination and sister chromatid cohesion play important roles in the maintenance of genome integrity and the fidelity of chromosome segregation in mitosis and meiosis. Within the living cell, the integrity of the DNA is threatened by various factors that cause DNA-lesions, of which DNA double-strand breaks (DSBs) are considered particularly deleterious. The causative agents can be of endogenous origin, such as metabolically produced free radicals, and of exogenous origin, such as ultraviolet light and ionizing radiation. The accurate repair of DSBs is important to prevent chromosomal fragmentation, translocations and deletions. Of the sophisticated (networks of) DNA repair pathways that have evolved, homologous recombination, which repairs the DSB by copying information from an intact homologous DNA-template, is considered one of the most accurate. In mitotic G2, the sister chromatid is preferentially used as a template for recombinational repair of DSBs.</p><p>DSBs can also arise as normal intermediates in several DNA repair and recombination pathways, including meiotic recombination. In meiotic recombination (in yeast), the cells actively induce large numbers of DSBs, and channel their search for a homologous template towards a non-sister chromatid of the homologous chromosome. In each pair of homologous chromosomes, at least one DSB is repaired by reciprocal exchange of precisely corresponding segments of non-sister chromatids (crossing over), whereas additional DSBs in the same chromosome pair are repaired by either reciprocal or non-reciprocal exchange. The reciprocal exchanges between non-sister chromatids (visible as chiasmata) are essential for the proper disjunction of homologous chromosomes during the first meiotic division (meiosis I).</p><p>When I started my investigations for this thesis, it was recognized that reciprocal exchanges as such could not direct the proper segregation of homologous chromosomes at meiosis I; some "glue" should keep the chiasmata in place, either by binding to the chiasmata, or by maintaining cohesion between the sister chromatids distal to the chiasmata. Although mutants existed that appeared to be defective in the production of this glue, its nature remained unknown. In this thesis, I have tried to identify components involved in meiotic sister chromatid cohesion and recombination, to analyze the interplay between these two processes in meiosis and to gain insight into their relationship with mitotic DNA repair and recombination.</p><p>In <strong>chapter 1</strong> , I explain the choice of the experimental model systems that I used for the research described in this thesis. All investigations were performed in two filamentous fungi, namely <em>Sordaria macrospora</em> and <em>Aspergillus nidulans</em> . The most important reason for the choice of these two fungi was that mutants were available (or easily obtainable) that were defective in meiotic sister chromatid cohesion and/or recombination, and that it should be feasible to clone the corresponding wild-type genes by means of transformation complementation of the mutant defects. <em>S. macrospora</em> had the additional advantage of a well-developed cytology and <em>A. nidulans</em> had the additional advantages of well-developed molecular genetic tools and the presence of a parasexual cycle in addition to the sexual cycle. This latter feature offers the possibility of analyzing mitotic allelic recombination. Moreover, <em>A. nidulans</em> is one of the two known organism that do not assemble synaptonemal complexes (SCs) during meiotic prophase and that do not display positive crossover interference. The choice for both <em>S. macrospora</em> and <em>A. nidulans</em> would thus make it possible to compare the role(s) of genes involved in meiotic sister chromatid cohesion and/or recombination in a organism with and one without SCs.</p><p>In <strong>chapter 2</strong> , we describe the cloning of the <em>SPO76</em> gene of <em>S. macrospora</em> by transformation complementation of the meiotic defects of the <em>spo76-1</em> (non-null) mutant. <em></em> It was known that this mutant displayed defects in meiotic sister chromatid cohesion, meiotic recombination and mitotic DNA-repair. Furthermore, we analyzed the localization of the Spo76 protein throughout wild-type mitosis and meiosis and performed a detailed analysis of the <em>spo76-1</em> mutant phenotype. We show that Spo76p is chromosome-associated during all stages of mitosis and meiosis, except at metaphase(s) and anaphase(s). During mitosis, Spo76p disappears from the chromosomes at prometaphase. During meiotic prophase I, Spo76p is more abundant than during any other cell cycle stage, and localizes preferentially close to the chromosome axes. Spo76p disappears from the chromosomes at diplotene. In the <em>spo76-1</em> mutant, we observed a transient defect in chromosome organization at (mitotic) prometaphase: the duration of this stage was prolonged and chromosome morphology was abnormal. Strikingly, chromosomal regions with defects in both sister chromatid cohesion and chromosome compaction alternated with regions with apparently normal cohesion and compaction. We speculate that the mitotic prophase to metaphase transition involves forces that tend to disrupt cohesion and that Spo76p promotes the maintenance of a minimum of cohesion in combination with chromosome compaction. Likewise, we observed in meiotic prophase of <em>spo76-1</em> , from late leptotene on, that sister chromatid cohesion and chromosome compaction were coordinately affected on a regional basis. Regions with widely split axial elements alternated with regions with unsplit segments of axial elements. These unsplit segments could form stretches of SC, which contained rare late recombination nodules (late RNs: ultrastructurally recognizable enzyme-complexes involved in the later steps of meiotic recombination). Whereas the number of late RNs was strongly reduced <em>in spo76-1</em> , early RNs (as recognized by immunocytochemical labelling of Rad51 and Dmc1) occurred at only slightly reduced levels in the mutant and persisted longer. This suggest that <em>spo76-1</em> is deficient in some intermediate step of meiotic recombination. The role of Spo76p in the meiotic leptotene/zygotene transition may be related to its role during the mitotic prophase/metaphase transition, in that the leptotene/zygotene transition might also bring along forces that tend to disrupt cohesion, and that Spo76p is also needed for maintenance of cohesion during this meiotic transition. Spo76p may have an additional role in late meiotic prophase because in <em>spo76-1</em> , meiotic sister chromatids separated completely from diplotene on, whereas in wild type this does not occur before anaphase II.</p><p>The predicted protein encoded by the <em>SPO76</em> gene is evolutionary conserved from fungi to man; the highest percentage of amino acid identity (44%) was found with the BIMD protein of <em>A. nidulans.</em> The <em>A. nidulans</em><em>bimD6</em> mutant was previously identified as a conditional lethal mutant with a (lethal) mitotic chromosome segregation defect at high temperature and a (non-lethal) DNA repair deficiency phenotype at low temperature.</p><p>In <strong>chapter 3</strong> , we demonstrate by heterologous complementation that the <em>SPO76</em> gene of <em>S. macrospora</em> can complement both the temperature and the MMS (methyl methane sulphonate) sensitivities of <em>bimD6</em> in <em>A. nidulans</em> , implying direct functional homology between the mitotic roles of two proteins. We also show that, like <em>spo76-1</em> , <em>bimD6</em> mutants do not form sexual spores (ascospores). However, <em>bimD6</em> mutants display, unlike <em>spo76-1</em> , disturbances in premeiotic development and form only few asci. In the few asci that entered meiosis, sister chromatids separate prematurely, but the extent of meiotic sister chromatid separation is less severe in <em>bimD6</em> than in <em>spo76-1</em> . In addition, whereas the mitotic localization of the two proteins was roughly similar, the meiotic localization of the two proteins showed some important differences. Unlike Spo76p in <em>S. macrospora</em> , BIMD in <em>A. nidulans</em> was not more abundant during meiotic prophase than in the mitotic cycle, and did not localize preferentially close to the chromosome axes during the pairing of homologous chromosomes. Moreover, <em>SPO76</em> could not complement the sexual sporulation defects of <em>bimD6</em> in <em>A. nidulans,</em> and <em>vice versa</em> , BIMD could not restore the sexual sporulation defects of the <em>S. macrospora spo76-1</em> mutant. These results indicate that Spo76p and BIMD may differ in a species-specific manner with respect to their meiotic roles. The species-specific aspects of the roles of Spo76p and BIMD6 in meiosis are possibly related to the differences in meiotic chromosome organization between the two fungi (see above; <strong>chapter 1</strong> ): in contrast to <em>S. macrospora</em> , <em>A. nidulans</em> does not form SCs and does not display positive interference of meiotic crossovers.</p><p>We also show that <em>bimD6</em> mutants are hypersensitive to X-rays in addition to their elevated sensitivities to UV and MMS, but only when dividing cells are exposed to these agents. The <em>bimD6</em> mutant thus closely resembles recombination-deficient mutants of <em>A. nidulans</em> , such as <em>uvsC114</em> (the <em>uvsC</em> gene of <em>A. nidulans</em> is homologous to <em>RAD51</em> of <em>Saccharomyces cerevisiae</em> ; <strong>chapter 4</strong> ). We have therefore compared <em>bimD6</em> with <em>uvsC114</em> regarding <em></em> defects in mitotic recombination. When assayed for allelic recombination, <em>bimD6</em> and <em>uvsC114</em> yielded similar results. In both mutants, the absolute frequencies of allelic recombination were strongly reduced, although the distribution of recombinants among the various classes was comparable to wild type. However, when assayed for intrachromosomal conversions, <em>bimD6</em> and <em>uvsC114</em> produced different results. Intrachromosomal conversions between interrupted duplications were significantly reduced in <em>uvsC114</em> , but occurred at wild-type frequencies in <em>bimD6</em> . We propose that BIMD is required for homologous recombination when the template is located on another (sister or non-sister) chromatid, but not when the template is available in close proximity on the same sister chromatid or the same chromatin domain/loop. The repair machinery may thus be obliged to cooperate with cohesion complexes (which are probably located at the borders between chromatin domains/loops) if the homologous template lies outside an intrachromatid loop domain. In contrast, <em>uvsC</em> was required for both types of repair.</p><p>In <strong>chapter 4</strong> , we describe the cloning of the <em>uvsC</em> gene of <em>A. nidulans</em> by transformation complementation of the mitotic repair defects of a <em>uvsC114</em> mutant. Furthermore, we disrupted the entire <em>uvsC</em> gene, and we compared the phenotypic effects of the resulting null mutation with those of <em>uvsC114</em> . The predicted UVSC protein shows 67% amino acid identity with Rad51p of <em>S. cerevisiae</em> and 27% amino acid identity with the RecA protein of <em>Escherichia coli.</em> These proteins are involved in strand invasion and exchange during homologous recombination. We found that in the absence of DNA-damaging agents, the <em>uvsC</em> gene was transcribed at a higher level in the <em>uvsC114</em> mutant than in wild-type. Transcription of <em>uvsC</em> was inducible by MMS in wild-type and <em>uvsC114</em> mutant strains. We compared the mitotic and meiotic phenotypes of mutants carrying the <em>uvsC114</em> point mutation (a deletion of 6 bp in core domain I) with those of the <em>uvsC</em> null mutant. The <em>uvsC</em> null mutant was more sensitive to UV and MMS than <em>uvsC114</em> , indicating that <em>uvsC114</em> is not a null mutation. The sexual developmental phenotypes of the two mutants also differed. In the <em>uvsC</em> null mutant, sexual development was disturbed before the onset of meiosis, whereas in <em>uvsC114</em> it was blocked in meiotic prophase. We observed large, multi-nucleated cells in older cleistothecia of the <em>uvsC</em> null mutant These cells possibly represent degenerated croziers, which might have been blocked at premeiotic S-phase. In <em>S. cerevisiae</em> , disruption of <em>RAD51</em> has no effect on mitotic growth and causes arrest at meiotic prophase I, whereas in mouse, disruption of <em>Rad51</em> results in embryonic lethality. In <em>A. nidulans</em> , disruption of <em>uvsC</em> had no effect on mitotic growth, and caused an arrest at a premeiotic stage of sexual development. Disruption of <em>RAD51</em> -homologous genes thus has different effects in different organisms.</p><p>In <strong>chapter 5</strong> , we describe the isolation and characterization of sexual sporulation mutants of <em>A. nidulans</em> . Vegetative spores were treated with a high dose of UV (1.5 % survival) and surviving colonies were plated on supplemented minimal medium. Colonies that did not display aberrant vegetative growth were visually screened for the appearance of "barren" fruiting bodies (=fruiting bodies without or with only few ascosores). This screen (1250 colonies analyzed) yielded 20 mutants with the desired phenotype. After two successive rounds of backcrosses with wildtype, two mutants yielded no longer progeny with the original mutant phenotype; these mutants were not analyzed further. The remaining 18 mutants were all recessive and were assigned to 15 complementation groups. Based on these numbers, we estimate that, under the growth conditions tested, about 50-100 genes are specifically involved in ascospore formation in <em>A. nidulans</em> . Three mutations were mapped by parasexual analysis: two mutations could be assigned to a specific chromosome, and one was associated with a translocation breakpoint. For all 18 mutants, the contents of the "barren" fruiting bodies were cytologically analyzed. A large proportion of the mutants, namely 11 out of 18, arrested in meiotic prophase I (like <em>uvsC114</em> ; <strong>chapter 4</strong> ) or metaphase I (like <em>bimD6</em> ; <strong>chapter 3</strong> ). It is thus possible that this new collection contains mutants that are specifically affected in meiotic recombination and/or sister chromatid cohesion. We suggest a strategy to clone the corresponding wild-type genes, by selecting for the appearance of prototrophic progeny clones that result from meiotic reassortment of auxotrophic markers.</p><p>In the General Discussion ( <strong>chapter 6)</strong> , we speculate upon the possible links between sister chromatid cohesion and recombination in mitosis and meiosis. Whereas in mitosis centromeric cohesion primarily serves chromosome disjunction, arm cohesion may play additional roles in repair by recombination. Cohesion complexes at the basis of the chromatin domains/loops may function as nucleation sites for the assembly of recombinational repair complexes and assist these complexes in finding a homologous template in a precisely corresponding segment of the undamaged sister chromatid. They may thus be responsible for the observed bias for the sister chromatid as a template for recombinational repair during mitotic G2. In meiosis I, recombination has to be directed towards a non-sister chromatid of the homologous chromosome and, at the same time, arm cohesion has to be maintained and possibly even reinforced to ensure correct reductional chromosome segregation. Consequently, both the recombinational repair machinery and the cohesion complexes function in meiosis in a modified form. In this modified form, cohesion complexes may serve as a basis for axial element formation. DSBs are probably produced concomitantly with axial element assembly. As we proposed for mitotic recombinational repair, we hypothesize that meiotic DSBs are transferred to the basis of chromatin loops/domains, where they are brought into contact with the meiotic cohesion complexes and additional proteins required for recombinational repair. However, we speculate that homology search on the sister chromatid will now be blocked by linear element components so that another template for homologous recombination has to be found. Furthermore, axial element components, in concert with the modified recombination complex, will possibly establish DNA-DNA contacts with a non-sister chromatid of the homologous chromosome. Thus, by providing the basis for linear element formation and assembly of a meiosis-specific recombination complex, cohesion complexes may contribute to the preference for a non-sister chromatid of the homologous chromosome as a template for homologous recombination in meiosis.</p&gt