Maintenance of an intact genome and proper regulation of the genes within are crucial aspects for life. The work of this dissertation has implicated the Drosophila condensin II complex in both processes. Condensin II's ability to reconfigure chromosomes into spatially separated and discrete units is necessary to ensure proper meiotic segregation. When this "individualization" activity fails in a condensin II mutant, chromosomes remain entangled, and either cosegregate or become lost during cell division. This leads to the creation of aneuploid sperm. We have also implicated condensin II as a factor necessary to individualize interphase somatic chromosomes from one another. This is relevant in Drosophila because the association of homologous chromosomes is thought to facilitate gene regulation activity in trans. We speculate that condensin II individualization spatially distances aligned chromosomes from one another and prevents this trans-communication between allelic loci. This is supported first by an increase of homologous chromosome pairing in a condensin II mutant background. Secondly, loss of condensin II leads to elevated production from alleles that are known to depend on pairing for transcriptional activation. These meiotic and interphase condensin II roles support its necessity to Drosophila genome integrity and transcriptional regulation. Given the conservation of condensin from bacteria to humans, it is likely that equivalent or related roles exist in a variety of species.

Specification of a germ cell lineage and appropriate maturation are essential for the transfer of genetic information from one generation to the next. Germ cells form from pluripotent precursor cells that migrate into the gonadal ridge and undergo commitment to either the female or male lineage. In the fetal ovary, germ cells enter meiotic prophase I, then arrest at the diplotene stage; in the testis germ cells do not begin meiosis until puberty. Abnormal differentiation of germ cells can result in malignant transformation. Somatic cells play a key role in modulating the developmental fate of the germ cells. Research into germ cell development during fetal life has almost exclusively focused on studies in rodents, but we, and others, have reported several fundamental differences in the expression of germ cell specific markers in the human compared with the mouse. The studies described in this thesis have investigated germ cell-specific gene expression and the possible impact of the somatic cells during development. This was achieved by studying human fetal gonads obtained during the 1st and 2nd trimesters of pregnancy and through the use of both wild-type and mutant mouse ES cell lines. Studies on germ cells in the human fetal testis have extended the findings of others, and confirmed that germ cell populations at different stages of maturation co-exist in the human fetal testis, a situation that is in contrast to that in rodents. For example expression of M2A and AP2γ was restricted to the OCT4-positive gonocyte population, while VASA and NANOS1 were localised exclusively to the to the OCT4-negative prespermatogonia. DAZL was expressed in both populations. Analysis also revealed that both the gonocyte and prespermatogonial populations proliferate throughout the 2nd trimester. Recent studies have implicated retinoic acid (RA) in the control of meiotic entry in germ cells of the fetal mouse ovary. In this study we demonstrated for the first time that two genes implicated in the action of RA in mouse gonad, STRA8 and NANOS2, are also expressed in a similar sexspecific- manner in the human fetal gonads, and that the RA receptors are present in both somatic and germ cells suggesting that RA may regulate germ cell function in the human as well as the mouse. However, whilst the mesonephros appears to be the primary site of RA synthesis in the mouse our initial studies indicate that in the human the gonad itself may be a more likely site of RA biosynthesis. In the fetal mouse testis, RA is degraded by the enzyme Cyp26b1 present in the somatic cells and germ cells do not enter meiosis, our novel findings suggest that CYP26B1 is more abundant in the human fetal ovary than the testis, suggesting that meiotic entry may be controlled by an alternative signalling pathway in the human. One of the methods that can aid our understanding of somatic cell gene expression in the gonad is in vitro culture. To date, there have been no published reports of the successful in vitro culture of somatic cells from the human fetal testis. In the current study, populations of human somatic cells were dissociated and maintained in vitro and characterised. Analysis demonstrated that cells expressing mRNAs characteristic of Sertoli cells, Leydig cells and peritubular myoid (PTM) cells were present initially, but long-term culture resulted in downregulation in expression of mRNAs specific for Sertoli cells and Leydig cells, suggesting that these cells either failed to survive or underwent alterations to their phenotype. In contrast PTM/fibroblast cells proliferated in vitro and initially maintained androgen receptor expression. These cultures therefore hold promise for studies into the signalling or cell-cell interactions in testicular somatic cells especially those relevant to the PTM population. Several studies have claimed differentiation of putative germ cells from ES cells. In the current study, analysis of mouse ES cell lines has expanded on results showing that ES cells and early germ cells express a number of genes in common. Kit signalling was shown to be important for ES cell survival as they differentiate although expression of Kit was heterogeneous. We also demonstrated that ES cells that did not express Kit displayed a decreased expression of the early germ cell genes Blimp1, Fragilis and Stella, implicating Kit signalling in the control of germ cell-associated gene expression in ES cells. This may be important to future studies optimising germ cell derivation from ES cells. In conclusion, this study has demonstrated important differences in protein expression patterns in germ cells of the human fetal testis compared to the mouse, and has raised questions about whether the proposed mechanism controlling meiotic entry of germ cells in the mouse can be applied to the human. The establishment of a system for culturing human fetal gonadal somatic cells may lead to further understanding of gene expression and development in the human fetal testis, and data suggest that the Kit/Kitl signalling system may influence germ cell gene expression in mouse ES cells.

Chez Saccharomyces cerevisiae, la methylation de la lysine 4 de l'histone H3 (H3K4) est catalysée par le complexe à activité methyltransférase Set1, conservé au cours de l'évolution. Durant la méiose, l'absence de Set1 conduit à un retard de démarrage de la phase S, et à un défaut dans la formation des coupures double-brin de l'ADN (CDBs). Nous avons cherché à mieux caractériser ces deux conséquences phénotypiques liées à l'absence de Set1. Nous montrons que le retard de réplication est lié à la perte de méthylation de H3K4 mais qu'il ne résulte pas d'un défaut d'activité des kinases responsables de l'activation des origines de réplication ou de l'activation des voies canoniques de surveillance moléculaire liées aux dommages de l'ADN. L'importante diminution de fréquence de CDB sur la majorité des points chauds chez le mutant set1∆ a été corrélée à l'absence de la marque de H3K4 triméthylée. Nous avons confirmé le role de la méthylation de H3K4 sur la base de la diminution générale de la fréquence des CDBs observée en absence des différentes sous-unités du complexe associé à Set1 (COMPASS) ou chez un mutant exprimant une histone H3 non-méthylable (H3K4R). Pour tester la relation de causalité entre méthylation et CDBs, différentes sous-unités du COMPASS, telles que Set1 et Spp1, ont été fusionnées avec le domaine de fixation à l'ADN de Gal4 pour les cibler vers des régions non méthylées et dépourvues de CDB. Gal4BD-Spp1 stimule fortement la fréquence des CDBs à certains loci, y compris en contexte mutant H3K4R. Ainsi, le ciblage de Spp1 peut etre suffisant pour recruter et/ou activer la machinerie de CDB. / In Saccharomyces cerevisiae, the methylation of the lysine 4 of histone H3 (H3K4) is catalysed by the evolutionary conserved Set1 methyltransferase complex. During meiosis, the absence of Set1 leads to a delay of S-phase onset and to a defect in the formation of double-strand breaks (DSBs). Our work was intended to give some clues about these two phenotypic consequences of Set1 loss. We show that the replication delay is linked to the absence of H3K4 trimethylation but does not result from a defect of the kinases responsible for the activation of replication origins or the activation of the canonical DNA-damage checkpoints. The severe decrease of DSB levels at the majority of recombination hotspots in set1∆ has been correlated with the specific marking of DSB sites by H3K4 trimethylation at some loci. We have confirmed the role of H3K4 methylation by observing a general decrease in DSB frequency similar to that of set1∆ in mutants lacking various subunits of the Set1- associated complex (COMPASS) or expressing a nonmethylatable histone H3 (H3K4R). To test for a causal relationship between H3K4 methylation and DSB formation, we have fused different proteins of the COMPASS, such as Spp1 or Set1, with the DNA binding domain of Gal4, in order to target them to H3K4-unmethylated and DSB-cold regions. Remarkably, Gal4BD-Spp1 strongly stimulates DSB formation in naturally cold DSB regions, even in the H3K4R mutant context. Thus, the specific tethering of Spp1 to a chromosome site is sufficient to recruit and/or activate the DSB machinery.

PLK1 belongs to the extended family of serine/threonine kinases controlling the cell cycle. It is well known for its role in the control of mitosis and contributes also to the regulation of meiotic division. On a basis of Live Cell Imaging (LCI) experiments we can describe the phenotype of the oocytes with PLK1 inhibited by small molecular inhibitor BI2536. PLK1 inhibition leads to delayed nuclear envelope breakdown (NEBD) and chromatin condensation (CC) and also causes desynchronization of NEBD and CC; in contrast to control oocytes, PLK1 inhibited oocytes break down their nuclear envelope with chromatin almost fully condensed. Also duration of these two early nuclear events is prolonged in oocytes with inhibited PLK1. In contrast to somatic cells, PLK1 inhibition in mouse oocytes does not prevent assembly of spindle with two distinct poles but affects the final spindle volume. Similar to somatic cells, mouse oocytes with PLK1 inhibited from the beginning of the meiotic maturation stay arrested in metaphase I but in the case of mouse oocytes, this block is not dependent on Spindle Assembly Checkpoint (SAC) persisting activity. When mouse oocytes are synchronized on metaphase I/anaphase I transition by proteasome inhibition and then PLK1 kinase activity is inhibited, about 2/3 of the oocytes stay arrested...

La méiose est un processus fondamental qui conditionne la formation de gamètes et assure la stabilité des génomes tout en générant de la diversité par brassage génétique. La régularité méiotique nécessite la formation de crossing-overs (CO) exclusivement entre chromosomes homologues. Cette condition est plus difficile à remplir chez les espèces allopolyploïdes qui présentent plusieurs jeux de chromosomes toujours susceptibles de recombiner ensemble. Bien que la polyploïdie soit omniprésente chez les plantes, on connait peu de choses sur le déroulement de la méiose chez ces espèces. Au cours de ma thèse, je suis intéressée à l’effet de la polyploïdie sur la formation et sur la fréquence de CO en utilisant le colza (Brassica napus, AACC, 2n=38) comme modèle d’étude. J’ai notamment cherché à comprendre : (1) quel est l’effet du niveau de ploïdie sur la fréquence de crossovers, et (2) l’origine des variations de COs observées chez les plantes allohaploïdes (AC) produites à partir de différentes variétés de colza, en utilisant une palette d’approches cytologiques et cytogénétiques. Mes travaux ont permis de montrer que le niveau de ploïdie induit une augmentation de la fréquence de crossovers, et que cette augmentation est plus importante dans un contexte triploïde que tétraploïde. J’ai ainsi montré que la fréquence de CO augmente progressivement du diploïde (1,6 CO/bivalent) vers le tétraploïde (2 CO/bivalents) et quelle est maximale chez le triploïde (2,8 CO/bivalent). En ce qui concerne la deuxième question, j’ai montré que la différence entre les allohaploïdes de colzas apparaît tardivement au cours de la méiose. Elle semble être liée à une capacité différente à former des CO en fonction de la variété utilisée pour produire ces allohaploïdes et non pas à une différence dans la reconnaissance de l’homologie. Un de mes résultats original est que la protéine HEI10, impliquée dans la voie de formation des CO interférents, présente une dynamique différente entre les deux variétés, que ce soit à l’état euploïde (AACC) qu’allohaploïde (AC).Mes résultats conduisent à s’interroger sur la relation entre (i) la régulation du nombre de CO formés entre chromosomes homologues et (ii) la suppression des CO entre chromosomes non homologues chez les espèces allopolyploïdes. / Meiosis is a fundamental process required to produce gametes, ensure genome stability and generate diversity within species by creating new chromosome/allele combinations. For all these outcomes the exclusive formation of crossovers (CO) between homologous chromosomes is required. This condition is more difficult to fulfil in allopolyploid species that have more than two sets of chromosomes still able to recombine together. Although polyploidy has been particularly prevalent in plants, little is known about meiosis in polyploids. During my thesis I have analyzed the effect of polyploidy on CO formation and frequency, using oilseed rape (Brassica napus, AACC, 2n=38) as model. My work aimed to investigate (i) the effect of ploidy level on the rate of meiotic COs and (ii) the causes for the observed difference in CO rate between allohaploid plants (AC) produced from different B. napus varieties. To address these questions, I have combined a series of cytological, immunocytological and cytogenetical analyses.My work first indicates that polyploidization leads to increase CO frequency. I showed that the number of COs progressively increases from the diploid (1,6 CO/bivalent) to the tetraploid (2 CO/bivalent) and is maximal in the triploid (2,8 CO/bivalent). In the second part, I have shown that the difference of meiotic behaviors between B. napus allohaploids appears at a late stage of meiosis. This difference seems to be due to a difference in the propensity to form CO between the two varieties rather than a difference in the stringency of homology recognition. This difference could be related to the difference in the pattern and/or chronology of HEI10 (a key protein involved in the interfering CO pathway) signals along chromosomes during prophase I in both euploids (AACC) and allohaploids (AC).My results thus puts under the spotlight the link that may exist between (i) the regulation of CO rate between homologous chromosomes and (ii) the suppression of COs between non-homologous chromosomes in polyploid species.

In Saccharomyces cerevisiae, Ime4 is a protein that is induced during meiosis and has a primary role in regulating sporulation in starving diploids. One function of Ime4 is methylation of adenosine residues within mRNA transcripts. Recent studies have shown Ime4 to be induced in haploids during the mating response, although its role in mating has not been determined. In this report, I identify the subcellular localization of Ime4 during the mating response through treatment with alpha factor. A plasmid containing IME4-GFP under the control of the medium strength promoter CYC1 was created in order to express the protein in a controlled manner. Lastly, mutational analysis was conducted to determine which regions of the protein were necessary for its nuclear localization, aggregation, and sporulation function.

Pendant la méiose, la division cellulaire qui forme les cellules haploïdes, les chromosomes homologues hérités de chacun des deux parents sont appariés et échangent des segments réciproques appelés crossing-overs (CO). Les CO ne sont pas distribués au hasard dans le génome et leur taux varie le long des chromosomes. Certains des mécanismes responsable ont été décrits chez les mammifères et la levure mais ne sont pas conservés chez les plantes. Les CO sont fortement inhibés dans l'hétérochromatine qui est riche en éléments répétés. Le degré élevé de méthylation d l'ADN qui caractérise les séquences répétées pourrait être un inhibiteur des CO. Cela a été clairement démontré chez le champignon Ascobolus immersus et des études récentes ont montré que la perte de méthylation modifiait la distribution des CO chez Arabidopsis thaliana. Le but de ma thèse a été de décrire plus précisément le rôle de la méthylation de l'ADN dans le contrôle des CO en l'absence de polymorphisme de séquence qui affecte aussi la recombinaison.Pour cela, j'ai mesuré la recombinaison dans différentes plantes dans lesquelles la méthylation de l'ADN a été partiellement ou totalement enlevée grâce à la mutation du gène ddm1. Pour tester l'effet opposé d'un gain de méthylation, j'ai aussi essayé de cibler la methylation de l'ADN à un point chaud de recombinaison connu. Mes résultats montrent que la parte de la méthylation de l'ADN entraîne une augmentation globale de la recombinaison. Paradoxalement, l'heterochromatine qui est normalement très méthylée est moins affectée par la perte de méthylation que le reste du chromosome, probablement car la méthylation de l'ADN a des effets à distance. L'augmentation de CO est accentuée dans les générations successives du mutant ddm1. Cependant, l'effet le plus important est observé dans les hétérozygotes où la moitié du génome seulement est hypométhylée, ce qui suggère un rôle complexe de la méthylation. Finalement, j'ai pu montrer que le polymorphisme affecte la recombinaison surtout dans l'hétérochromatine mais pas dans le sens attendu puisque les plantes homozygotes recombinent moins que les plantes hétérozygotes. / During meiosis, the cellular division that gives rise to haploid cells, homologous chromosomes inherited from each parent are paired and are subjected to reciprocal exchanges of chromosome segments called crossing-overs (COs). COs are not randomly distributed in the genome. Some of the involved mechanisms have recently been described in mammals and yeast bu they are not conserved in plants. Repeat-rich heterochromatin is suppressed for COs. The high level of DNA methylation associated with repeats could be an inhibitor of COs. This was clearly demonstrated in the fungus Ascobolus immersus and recent studies have shown that the loss of DNA methylation also affects COs in Arabidopsis thaliana. The aim of my thesis was to describe more precisely the role of DNA methylation in the control of CO distribution in the absence of any DNA sequence polymorphism which are known to affect recombination. For this purpose, I measured recombination in different plants where DNA methylation has been partially or completely removed thanks to the mutation of the DDM1 gene. To test the opposed effect of a gain of DNA methylation,.I also tried to target DNA methylation at a known recombination hotspot. My results show that the loss of DNA methylation induces a global increase of recombination. Paradoxically, the normally highly methylated heterochromatin is less affected by this loss than the rest of the chromosome, probably because DNA methylation has distal effects. The increased recombination is exacerbated in successive generations of the hypomethylated ddm1 mutants. However, the strongest effect is seen in the heterozygotes where only half of the genome is hypomethylated, suggesting a complex role in the control of CO distribution. Finally, I show that DNA sequence polymorphism affects mainly recombination in the heterochromatin but not in the expected sense, since homozygous plants recombine less than heterozygous.

Meiosis is a specialized division of germ cells in sexually reproducing organisms, which is a fundamental process with key implications for evolution and biodiversity. In two consecutive rounds of cell division, meiosis I and meiosis II, a normal, diploid set of chromosome is halved. From diploid mother cells haploid gametes are generated to create genetic individual cells. This genetic uniqueness is obtained during prophase of meiosis I by essential meiotic processes in meiotic recombination, as double strand break (DSB) formation and repair, formation of crossovers (CO) and holiday junctions (HJs). Checkpoint mechanisms ensure a smooth progress of these events. Despite extensive research key mechanisms are still not understood. Based on an analysis of Massive Parallel Sequencing (MPS) data I could identify 2 genes, Mcmdc2 and Prr19, with high implication in meiotic recombination. In the absence of Mcmdc2 both sexes are infertile and meiocytes arrest at a stage equivalent to mid-­‐pachytene in wt. Investigations of the synaptonemal complex (SC) formation revealed severe defects suggesting a role for MCMDC2 in homology search. Moreover, MCMDC2 does not seem to be essential for DSB repair, as DSB markers of early and mid recombination nodules, like DMC1 and RPA, are decreased in oocytes. Nevertheless, late recombination nodules, which are positive for MutL homolog 1 (MLH1), do not form in both sexes. The absence of the asynapsis surveillance checkpoint mechanism in Hormad2 deficient ovaries with Mcmdc2 mutant background allowed survival of oocytes. This points into the direction that Mcmdc2 knock­out oocytes get eliminated after prophase I due to failed homologous synapsis. Interestingly, MCMDC2 contains a conserved helicase domain, like the MCM protein family members MCM8 and MCM9. I therefore hyphothesize that Mcmdc2 promotes homolgy search.

The spliceosomal U5 snRNP contains several proteins with well characterised functions in splicing, including: Brr2, an ATPase/RNA helicase that disrupts U4/U6 and U2/U6 snRNA base pairing during activation of the spliceosome; Snu114, a GTPase that controls the action of Brr2; and Prp8, the largest and most conserved protein considered to have a central role in the spliceosome, which interacts directly with Snu114 and Brr2. Yeast Cwc21 is one of twelve Bact complex proteins that associate with spliceosomes shortly before the first step of splicing catalysis. Cwc21 interacts directly with Prp8 and Snu114, as does its human orthologue, the SR protein SRm300/SRRM2. Although, Cwc21 is not essential for yeast cell viability, it is required for sporulation. This work aims to identify the function of Cwc21 during meiosis. PP1 is a protein phosphatase required for both steps of splicing. Multiple sequence alignments of Snu114 and Prp8 revealed the presence of putative PP1 binding motifs that are well conserved among different species. This led me to hypothesize that PP1 may interact with Snu114 and/or Prp8 to regulate these or other interacting proteins. By screening intron-containing genes that are expressed in meiosis, I found that Cwc21 is required for splicing HRB1 transcripts. In addition, I show that HRB1 is also required during meiosis. The HRB1 intron contains an unusual branchsite sequence, TACTAATG, which when changed to the consensus branchsite sequence restores sporulation in the absence of Cwc21. Therefore, it is likely that Cwc21 promotes the expression of HRB1 during an early stage of meiosis by stabilising its pre-mRNA in the catalytic centre of the spliceosome. This study demonstrates a novel function for Cwc21 during meiosis. Using yeast two hybrid assay I have identified the interacting regions of Cwc21, PP1 and Brr2 in Snu114. Through biochemical studies I provide evidence for mutually exclusive interaction of Cwc21 and PP1 in the putative PP1 binding motif situated in Snu114 domain ‘IVa’. In the case of yeast Snu114, the PP1 binding motif has a novel sequence ‘YGVQYK’. I also show that the affinity of Cwc21 and PP1 for Snu114 is influenced by the different nucleotide-bound states of Snu114. Furthermore, I show that mutations in Snu114 domain ‘IVa’ restrict Snu114 function during meiosis and affect the MER1 splicing regulatory network. Therefore, Snu114 may play a role in modulating the conformational state of the catalytic spliceosome through its interactions with Cwc21/PP1 in regulating subsets of genes during meiosis. Finally, I show that PP1 is a putative regulator of Prp8.

The organisation of microtubule networks into a bipolar spindle is essential for reliable chromosome segregation during cell division. A pair of centrioles surrounded by pericentriolar material (PCM), define the canonical centrosome that acts as the main microtubule organising centre (MTOC) during mitosis. In mammalian meiosis, centrioles are eliminated early on during oogenesis. Despite the absence of centrosomes, a large number of centrosomal proteins are highly expressed in mouse oocytes. Here, I characterise the localisation and function of centrosomal proteins at a previously undescribed meiotic spindle pole domain (MSPD). An initial protein screen identified a group of pericentriolar satellite proteins that localised to a previously undescribed spindle pole domain throughout meiotic maturation in mouse oocytes, including Pericentriolar material 1 protein (PCM1). This domain was distinct from spindle microtubules and the acentrosomal microtubule organising centres (aMTOCs). Initial characterisation focused on PCM1, the main centriolar satellite scaffold protein in somatic cells. Depletion of PCM1 revealed interdependence with the essential aMTOC component, Pericentrin. In the absence of PCM1, aMTOCs could no longer assemble or maintain their structural integrity. PCM1 degradation and disassembly of aMTOCs disrupted spindle assembly and reduced the total amount of nucleated microtubules throughout meiosis. In the absence of the main microtubule nucleating aMTOCs, oocytes relied on the Ran GTPase activity to form a small bipolar spindle. A similar mechanism was previously reported in human oocytes that lack prominent MTOCs. The extended centrosomal protein screen identified additional components of the MSPD. TACC3, under the regulation of Aurora-A at aMTOCs, drive assembly of the MSPD. This domain was absent in MTOC free human oocytes but a second population of TACC3 (identified in mouse oocytes) localised to the meiotic spindle and K-fibres was essential for maintaining spindle pole integrity. Establishing the Lightsheet Z.1 system for live cell imaging of human oocytes enabled us to observe the dynamic distribution of TACC3 in these oocytes. In the absence of prominent MTOCs and the MSPD, human oocytes likely rely on other spindle assembly factors and motor proteins to organise their spindle. Future work to address if the absence of the MSPD could account (in part) for the observed spindle instability in human oocytes is an exciting outlook.