Epigenetic regulation of DNA replication studied by super resolution microscopy

Table of Contents

Abstract

Zusammenfassung

List of figures

Abbreviations

1. Introduction
1.1 DNA replication is a basic biological process required by all living organisms
1.1.1 Mammalian DNA is organized into a higher-order chromatin structure
1.1.2 Origins of replication and visualization of distinct replication patterns
1.1.3 DNA replication patterns are formed by replication foci
1.1.4 Chromatin structure can be directly influenced by DNA methylation
1.1.5 Histones exhibit a multitude of post-translational modifications that define chromatin structure
1.1.6 Histone modifications are brought about by specific histone modifying enzymes
1.1.7 Disruption of histone modifying enzymes affects replication dynamics
1.2 Breaking the resolution limit with super resolution microscopy
1.2.1 Principles and applications of 3D structured illumination microscopy
1.2.2 Utilization of 3DSIM data for analysis of nuclear genome organization
1.3 Aims of this study

2. Materials
2.1 Cell lines
2.2 Antibodies
2.3 Fluorescent in situ hybridization primers
2.4 EdU and BrdU detection reagents
2.5 Reagents and chemicals
2.6 Buffers
2.7 Somatic and stem cell medium
2.8 Consumables
2.9 Equipment
2.10 Software

3. Methods
3.1 Cell Culturing
3.1.1 C2C12 and C127 propagation
3.1.2 Seeding C2C12 and C127 cells for immunostaining
3.1.3 Freezing and thawing of C2C12 and C127 cells
3.1.4 Embryonic stem cells propagation
3.1.5 Seeding embryonic stem cells for immunostaining
3.1.6 Pre-coating of flasks and coverslips for embryonic stem cells
3.1.7 Freezing and thawing of embryonic stem cells
3.1.8 Embryonic stem cell differentiation
3.2 Combined pulse-chase-pulse and immunostaining protocol
3.2.1 EdU click chemistry and pulse-chase-pulse experiments
3.2.2 Immunostaining for pulse-chase-pulse experiments
3.3 Transfection of embryonic stem cells for live-cell-imaging
3.4 Fluorescence in situ hybridization (FISH) of somatic and embryonic stem cells
3.4.1 DNA FISH probe preparation for major and minor satellites and LINES-1
3.4.2 DNA FISH probe preparation for telomeric repeats
3.4.3 Nick translation for amplified PCR products for DNA FISH probes
3.4.4 DNA FISH probe preparation from nick translated products
3.4.5 DNA FISH probe hybridization protocol
3.4.6 DNA FISH probe post-hybridization protocol
3.4.7 FISH treatment combined with EdU pulse
3.4.8 Xist RNA FISH probe preparation
3.4.9 Pre-treatment and fixation of cells for Xist RNA FISH
3.4.10 Xist RNA FISH probe hybridization and detection
3.5 3D-SIM to wide-field deconvolution protocol
3.6 Microscopy and image acquisition

4. Results
4.1 DNA replication patterns in somatic mouse cells
4.1.1 DNA replication patterns in C2C12 cells
4.1.2 DNA replication patterns in C127 cells
4.2 Fluorescent in situ hybridization (FISH) of repetitive sequences in C2C12 cells
4.3 Histone modification patterns in C2C12 cells
4.4 Quantification of replication foci in C2C12 cells
4.4.1 Effect of laser intensity on count of RF
4.4.2 Comparison of RF between 3D-SIM and WFD data
4.5 Analysis of EZH2 inhibitor effect on replication and proliferation of C2C12 cells
4.6 DNA replication patterns in mouse embryonic stem cells
4.6.1 DNA replication patterns in HI5 female mESCs
4.6.2 Live-cell S-phase progression in HI5 female mESCs
4.6.3 FISH hybridization of major satellite repeats in HI5 mESCs
4.7 H3K27me3 distribution in HI5 mESCs and the effect of EZH2 inhibitor
4.8 Differentiation of HI5 mESCs into EpiLCs
4.8.1 DNA replication patterns in HI5 EpiLCs
4.8.2 Inactive X chromosome replication analysis in EpiLCs
4.9 Hybridization of Xist RNA probe in somatic, stem, and EpiLCs
4.10 Effect of EZH2 inhibition during differentiation of HI5 mESCs into epiblasts
4.11 Comparison of DNA replication patterns in undifferentiated and differentiated Suv39H1/2 knockout male mESCs

5. Discussion
5.1 Replication patterns in C2C12 and C127 cells with 3D-SIM
5.2 Combination of fluorescent in situ hybridization of specific sequences and 3D-SIM
5.3 RF counts in C2C12 cells with TANGO and 3D-SIM
5.4 Differences in replication dynamics in somatic and mESCs
5.5 Effect of interference with histone modifications on S-phase progression

References A

Acknowledgements H

Abstract

DNA replication is a fundamental biological process responsible for accurate duplication of genetic information necessary for its faithful inheritance to the two daughter cells. Despite much effort, the underlying mechanisms controlling this process are not fully understood. In order to accommodate very large and complex genomes, replication dynamics in eukaryotes evolved to become controlled by major epigenetic mechanisms. Moreover, the spatio-temporal organization of S-phase progression changes throughout cell differentiation and development. The study of genome duplication has been largely hindered by the lack of appropriate monitoring techniques, and any comprehensive understanding ultimately requires quantitative approach.

In this master’s thesis, we analyzed replication patterns in mouse somatic and embryonic stem cells (mESCs) with newly developed three-dimensional structured illumination microscopy (3D-SIM) to register the progression of S-phase in more detail than previously described. We successfully established an automated workflow to produce reliable and reproducible replication foci (RF) counts in C2C12 cells from 3DSIM data and TANGO (Tools for Analysis of Nuclear Genome Organization). Such an approach has not been described before, and could be used to evaluate further cell types and schemes.

Additionally, we observed significant differences in replication timing and progression between somatic (C2C12, C127) and mESCs (HI5). In this report we show that in mESCs S-phase lasts significantly longer (15 h), with a ‘leaky’ chromocenter replication profile compared to somatic cells. Furthermore, differentiated HI5 female mESCs into epiblast-like cells (EpiLCs) exhibit inactive X chromosome and differential replication timing of Xi within two distinct EpiLC populations, and a much shorter S-phase (10 h).

As a final aim of this work, we interfered with specific histone modifications with inhibitors and knockout cell lines. Inhibition of EZH2 methyltransferase resulted in global reduction of H3k27me3 levels in both somatic and mESCs, however replication dynamics were not affected. In contrast to somatic cells, viability of mESCs in presence of inhibitor was greatly reduced, suggesting a more important role of H3K27me3 in mESCs. Suv39H1/H2 double knockout mESCs had no observable effect on replication dynamics or proliferation. Moreover, differentiation of these cells into EpiLCs resulted in a distinct S-phase progression, with replication resembling HI5 EpiLCs.

Zusammenfassung

Replikation der DNA ist ein grundlegender biologischer Prozess, der für die genaue Duplikation genetischer Informationen verantwortlich ist und somit die zuverlässige Vererbung der DNA an die Tochterzellen gewährleistet.Trotz großer Mühe, wurden die zugrundeliegenden Mechanismen, die diesen Prozess steuern, noch nicht vollständig aufgeklärt. Um sehr große und komplexe Genome erfassen zu können, haben sich Replikationsdynamiken dahingehend evolviert, dass sie durch epigenetische Mechanismen reguliert werden. Dazu kommt, dass die räumliche und zeitliche Koordination des Fortschreitens der S- phase sich während der Entwicklung und Zelldifferenzierung verändert. Die Analyse der Genomreplikation war hauptsächlich dadurch gehindert, dass angemessene Beobachtungstechniken fehlten, und jegliches umfassende Verständnis braucht zuletzt quantitative Verfahren.

In dieser Masterarbeit wurden Replikationsmuster während des Fortschreitens der S-phase mittels der neu entwickelten, dreidimensional strukturierten Beleuchtungsmikroskopie (3D-SIM) in murinen somatischen und embryonalen Stammzellen (ESZ) detaillierter erfasst als zuvor. Wir haben einen automatischen Arbeitsprozess erstellt, der verlässliche und reproduzierbare Anzahlen an Replikationsfoci in C2C12 Zellen mittels Daten von 3DSIM und TANGO (Tools for Analysis of Nuclear Genome Organization) errechnet. Diese Herangehensweise wurde noch nicht beschrieben und kann für die Auswertung weiterer Zelltypen und Vorhaben verwendet werden.

Zusätzlich haben wir signifikante Unterschiede zwischen somatischen (C2C12, C127) und ESZ (HI5) bezüglich Replikationszeitpunkte und dem Fortschreiten der Replikation gefunden. In dieser Arbeit können wir zeigen, dass im Vergleich zu somatischen Zellen, die S-phase in mESZ signifikant länger ist (15h) und "durchlässige" Chromozentren Replikation zeigt. Des Weiteren können wir zeigen, dass zu EpiLC differenzierte weibliche HI5 mESZ ein inaktives X Chromosom besitzen, dass der Replikationzeitpunkt des Xi in zwei Zellpopulationen geteilt werden kann, und über eine stark verkürzte S-phase Dauer verfügt (10h).

Zuletzt haben wir auf spezifische Histonmodifikationen mittels Inhibitoren und Knockout-Zelllinien eingegriffen. Inhibierung der EZH2 Methyltransferase resultierte in globaler Verringerung der H3K27me3 Menge in somatischen und mESZ, jedoch waren Replikationsdynamiken nicht beeinflusst. Im Gegensatz zu somatischen Zellen, war das Überleben von mESZ bei Zugabe von Inhibitoren stark verringert. Suv39H1/2 dko mESZ hatten keinen erkennbaren Effekt auf Replikationsdynamiken.Darüber hinaus, resultierte die Differenzierung von Suv39H1/2 dko mESZ zu EpiLCs in einemdistinkten Fortschreiten der S-phase, mit Änderungen, die denen der HI5 EpilCs ähnelten.

List of figures

Figure 1-1: Basic unit of chromatin organization

Figure 1-2: DNA replication patterns in mammalian cells

Figure 1-3: Time lapse series following a single cell throughout S-phase

Figure 1-4: Activation of eukaryotic DNA replication origins

Figure 1-5: Different classes of chromatin modifications

Figure 1-6: Histone modifying enzymes

Figure 1-7: Schematic of a typical acrocentric mouse chromosome

Figure 1-8: Basic principle of 3D-SIM

Figure 4-1: C2C12 pulse-chase-pulse experiment

Figure 4-2: Detailed progression of late S-phase in C2C12 cells

Figure 4-3: Characterization of DNA replication patterns in C2C12 cells

Figure 4-4: DNA replication patterns in C2C12 cells

Figure 4-5: DNA replication patterns in C127 cells

Figure 4-6: Major satellites in context of replication in C2C12 cells

Figure 4-7: Minor satellites in context of replication in C2C12 cells

Figure 4-8: Telomeric repeats in context of replication in C2C12 cells

Figure 4-9: LINE-1 repetitive sequences in context of replication in C2C12 cells

Figure 4-10: Specific histone modifications in C2C12 cells

Figure 4-11: Effect of laser power on data acquisition

Figure 4-12: Comparison of replication foci counts for 10% and 100% laser intensity during data acquisition

Figure 4-13: Comparison of RF counts for different S-phase stages between 3D-SIM and wide-

field deconvolved data

Figure 4-14: Effect of GSK343 inhibitor on replication C2C12 cells

Figure 4-15: Effect of GSK343 inhibitor on replication timing in C2C12 cells

Figure 4-16: Detailed analysis of EZH2 inhibition effect on replication and proliferation of C2C cells

Figure 4-17: Replication patterns in HI5 female mESCs

Figure 4-18: Pulse-chase-pulse experiment in HI5 female mESCs

Figure 4-19: Dynamics of replication progression in HI5 female mESCs

Figure 4-20: Major satellite FISH probe in context of replication in HI5 mESCs

Figure 4-21: Distribution of H3K27me3 in HI5 mESCs

Figure 4-22: Effect of GSK343 inhibitor on replication in HI5 mESCs

Figure 4-23: Detailed analysis of EZH2 inhibition effect on replication and proliferation of HI cells

Figure 4-24: Distribution of H3K27me3 in epiblast-like cells

Figure 4-25: Comparison of H3K27me3 distribution between undifferentiated HI5 mESCs and EpiLCs

Figure 4-26: Distinct S-phases observed in EpiLC population

Figure 4-27: Xi replication analysis in EpiLCs

Figure 4-28: Hybridization of Xist RNA probe in somatic, stem, and EpiLCs

Figure 4-29: Hybridization of Xist RNA probe on EHZ2 inhibited EpiLCs 70 Figure 4-30: Distribution of H3K9me3 in undifferentiated and differentiated Suv39H1/H2 double knockout cells

Figure 4-31: Replication progression in undifferentiated Suv39H1/H2 knockout cells

Figure 4-32: Replication changes in differentiated Suv39H1/H2 knockout cells

Abbreviations

illustration not visible in this excerpt

1. Introduction

1.1 DNA replication is a basic biological process required by all living organisms

One common feature of all living organisms is the ability to procreate. Cell division is a fundamental process responsible for generation of offspring that relies on a precise way to faithfully distribute genetic information to the new generation. This task is accomplished by DNA replication - a basic biological process that is conserved throughout all kingdoms of life. Prior to division, every cell must ensure that its genome is quickly and accurately copied and evenly distributed to the two daughter cells. Despite nearly identical genomic sequences in each cell type in a multicellular organism, the gene expression profile of each type is unique. To maintain proper development and function of an organism, this unique identity of each cell type must be conserved during the division. It is now recognized that in eukaryotes ‘epigenetic’ mechanisms such as DNA methylation and histone modifications are key contributors to the genomic integrity and faithful transmission of cellular identity from the parent cell to daughter cell. Interestingly, to accommodate very large eukaryotic genomes, replication dynamics evolved to become more controlled by epigenetic mechanisms that affect chromosome structure and nuclear organization, rather than DNA synthesis itself.

1.1.1 Mammalian DNA is organized into a higher-order chromatin structure

Mammalian nucleus is organized into functional domains that respond to specific signaling cues and regulate appropriate cellular pathways. These domains are formed by direct association of DNA and histone proteins, resulting in a higher-order structure commonly known as chromatin. Chromatin is a dynamic structure composed of negatively charged DNA coiled around an octamer of four positively charged histone proteins (two of each H2A, H2B, H3, H4), constituting a basic repeating unit known as the nucleosome (Fig. 1-1) (Kornberg, 1974). Each nucleosome contains approximately 147 bp of DNA, forming an array of repeating units and separated by a short histone-free segment called the linker DNA. Additionally, the fifth histone (H1) can tightly bind the nucleosomes with the linker DNA, helping to stabilize the higher order chromatin fiber (Caterino and Hayes, 2011).

illustration not visible in this excerpt

Figure 1-1: Basic unit of chromatin organization. The nucleosome consists of two copies of each histone protein (H2A, H2B, H3 and H4) forming an octamer around which the DNA (black) is wrapped around resembling ‘beads on a string’. Adjacent nucleosomes are separated by nucleosome-free linker DNA. Flexible histone tails (red) are sites of most post-translational modifications and are involved in interactions of neighboring nucleosomes or nuclear factors. (Modified from Marks et al., 2001)

Chromatin exists in two functional states - euchromatin and heterochromatin. Euchromatin is a less condensed form of the chromatin and includes most gene-rich portion of the genome which is often under active transcription. Heterochromatin, comprising of transcriptionally silent DNA, is further subdivided into two distinct forms - facultative and constitutive heterochromatin. Facultative heterochromatin is conditionally silenced and includes the inactive X chromosome of mammalian female cells, imprinted genes, perinuclear DNA and genes involved in tissue-specific silencing. Constitutive heterochromatin contains permanently silenced DNA including centromeres, telomeres, pericentric DNA and silenced mobile elements (Ma et al., 2005). Despite complex organization, specific genomic regions replicate at different times and begin replication at different initiation sites during DNA replication event (Taylor, 1960).

1.1.2 Origins of replication and visualization of distinct replication patterns

DNA replication occurs during DNA synthesis phase (S-phase) of the cell cycle and begins at specific locations - origins of replication. Unlike prokaryotes, where an entire genome can be duplicated from a single replication origin, eukaryotes exhibit a much larger and more complex genome, requiring coordination of thousands of replication origins. Huberman and Riggs successfully demonstrated that in mammalian cells as many as 30,000-50,000 origins are active at each cell cycle (1966). Interestingly, whereas in bacteria specific genetic sequences dictate the replication start sites, no consensus sequence defining replication origins in metazoans has been identified (Gilbert, 2001). To further complicate this matter, replication must be coordinated with chromatin remodeling and ongoing transcription (Schwaiger and Schübeler, 2006). Incorporation of modified nucleotides into the newly synthesized DNA strand and their visualization with light microscopy revealed distinct replication patterns that change throughout S-phase (Nakamura et al., 1986). These distinct replication patterns are characterized into early, mid and late patterns and correspond to specific sequences that replicate in spatial and temporal positions during S- phase (O’Keefe et al., 1992). Figure 1-2 depicts super-resolution light microscopy images of mammalian somatic cells exhibiting characteristic S-phase patterns. In early S-phase, replication begins at many small, discrete foci throughout the nucleus, and corresponding to euchromatic regions or R-bands with high gene density. At the onset of mid S-phase, the patterns become more defined, with facultative heterochromatin at perinuclear and perinucleolar regions replicating first, followed by an even more defined replication pattern of constitutive heterochromatin actively replicating in late S-phase. Finally, in very late S-phase, replication foci accumulate into fewer but larger structures (O’Keefe et al., 1992; Craig and Bickmore, 1993; Bickmore, 2001).

illustration not visible in this excerpt

Figure 1-2: DNA replication patterns in mammalian cells. 3D-SIM super resolution images of C2C12 somatic cells with characteristic early, mid and late S-phase patterns. Euchromatic regions replicate during early S-phase (Se), followed by facultative heterochromatin (Sm), and finally constitutive heterochromatin replicating during late S-phase (Sl). Replicating DNA is visualized by incorporation of modified nucleotide EdU (5-ethynyl-2´-deoxyuridine) and chemical detection after fixation (Top). Bottom panel showing overlay of DNA staining by DAPI (gray) and EdU replication staining (red). Scale bar 5 µm. (Casas-Delucchi and Cardoso, 2011)

In the early studies, analysis of DNA replication has been limited to static time points and fixed cells. Isolated snapshots of fixed cells in S-phase did not provide enough direct evidence for change of replication patterns in individual cells (Leonhardt et al., 2000). In recent years, however, the progression of S-phase has been extensively studied in the living cells. Replication patterns have been clearly observed with cell lines stably expressing proliferating cell nuclear antigen (PCNA) - a central component of replication machinery fused with green fluorescent protein (GFP) (Fig. 1-3). PCNA is crucial for formation of replication factories and has been widely used as a marker of active sites of DNA replication. Live-cell analysis clearly showed that cells undergo reproducible changes in replication patterns throughout S-phase, serving as a marker for early, mid, and late S-phases (Leonhardt et al, 2000).

illustration not visible in this excerpt

Figure 1-3: Time lapse series following a single cell throughout S-phase. Time-lapse snapshots of live-cell imaging of mouse C2C12 cell stably expressing GFP-PCNA. Clearly identifiable sites of active replication undergo characteristic changes throughout S-phase. (Leonhardt et al., 2000).

1.1.3 DNA replication patterns are formed by replication foci

S-phase replication patterns are established by small 120 nm structures consisting of a series of activated replication start sites within close spatial proximity to each other and commonly known as the replication foci (O’Keefe et al., 1992; Baddeley et al., 2010). At these replication origins, the DNA double helix unwinds to allow the replication machinery to begin DNA synthesis of each strand and progress in a bidirectional manner. It has already been well established that before the actual DNA replication, origin selection occurs in the absence of nuclear membrane, shortly after metaphase, with binding of highly conserved origin recognition complex (ORC) to the DNA replication origin. Regardless of targeting mechanism of ORC, which varies between different eukaryotes, ORC initiates the assembly of pre-replication complex (Pre-RC) (Gilbert, 2001). Phosphorylation of Pre-RC leads to the recruitment of additional initiation factors (Mcm10, Sld3, GINS) required for unwinding of the replication origins and recruitment of replicative DNA polymerases (Gambus et al., 2006; Kanemaki et al., 2006). The result is two open replication forks which replicate in opposite directions (Fig. 1-4). Replication continues until the forks of the neighboring replication origins eventually fuse together.

illustration not visible in this excerpt

Figure 1-4: Activation of eukaryotic DNA replication origins. Pre-replication complexes (Pre-RCs) assemble during G1 phase of the cell cycle at the replication origins. Upon activation of Pre-RCs, replication begins in a bidirectional manner. Activation of different origins occurs throughout S-phase - some during early (1 and 2), and some during mid (3) or late (4) S-phase. (Mechali, 2010)

Although mammalian cells contain a large number of replication origins, not all origins activate at the same time and not all activate at each cell cycle (Mechali, 2010). Selection and activation of origins appear to be highly organized, resulting in defined patterns observed during S-phase. However, neither all the known pre-RC factors, nor the ORC can explain how specific origins are chosen in multicellular eukaryotes. Consensus DNA sequence defining replication origins has not been identified and transcriptional activity does not appear to predict replication origins. This could suggest that accessibility to the DNA, or chromatin structure that corresponds to it, might facilitate replication activity and control replication timing (Gilbert, 2001; Wei et al., 1998; Danis et al., 2004; Lunyak et al., 2002). Randomly inactivated X chromosome in female mammalian cells replicates within a different time frame than its active homolog, emphasizing the fact that genetics alone cannot explain the mechanism of replication timing (Taylor, 1960). Early replication origins correlate with actively transcribed genes, and late origins with non-coding genomic regions (Schwaiger and Schubeler, 2006; Donaldson, 2005). Open chromatin, which permits active transcription, could be a potential control mechanism of replication timing. To further support this idea, Zhou and colleagues demonstrated that replication origins are enriched in open chromatin structures (2005). Major epigenetic mechanisms such as DNA methylation and histone post- translational modifications, which play a crucial role in defining chromatin state, appear to be reasonable candidates to control replication dynamics in mammals (Casas-Delucchi and Cardoso, 2011).

1.1.4 Chromatin structure can be directly influenced by DNA methylation

DNA can be directly modified to influence chromatin structure. DNA methylation refers to the covalent addition of a methyl group to a 5’ carbon of the cytosine pyrimidine ring and occurs primarily at CpG dinucleotides in mammalian genomes. Methylation patterns are established during the development and cell differentiation and then maintained in somatic cells. In fact, somatic cells exhibit methylation of 70-80% of all CpGs (Laurent, 2010). Repetitive sequences are highly enriched in CpG methylation levels and CpG islands at promoter regions are usually non-methylated (Chen and Riggs, 2012). DNA is methylated by a family of DNA methyltransferases which include Dnmt1, Dnmt3a and Dnmt3b. Dnmt1 is responsible for maintenance of DNA methylation levels (Song et al., 2011) and Dnmt3a and Dnmt3b are known as de-novo methyltransferases, responsible for establishing new methylation patterns during development (Rottach et al., 2009). DNA methylation primarily acts as a silencing mark, influencing gene expression by affecting interactions of transcription factors and chromatin proteins (Razin and Cedar, 1991). This epigenetic mechanism is essential for normal development and plays an important role in several fundamental processes such as X chromosome inactivation, suppression of retrotransposon elements and genomic imprinting (Li et al., 1992; Okano et al., 1999; Bird, 2002). CpGs in constitutive heterochromatin, as well as the promoter regions of inactive X chromosome and the imprinted genes, are characterized by high levels of DNA methylation (Ma et al., 2005). During S-phase, these regions show characteristic replication patterns and replicate at different times than non-methylated euchromatin. It is reasonable to assume that methylation patterns could influence replication timing and play a role in replication dynamics (Casas-Delucchi and Cardoso, 2011). However, current evidence suggests that replication timing appears to be independent of DNA methylation levels and a clear role of this modification in replication dynamics is yet to be elucidated (Bickmore and Carothers, 1995; Schübeler et al., 2000).

1.1.5 Histones exhibit a multitude of post-translational modifications that define chromatin structure

Histone proteins are known to be some of the most evolutionary conserved proteins in nature and share a common structural ‘histone fold’ domain that mediates histone-DNA and histone-histone interactions by forming a 4-helix bundle of H2A-H2B and H3-H4 heterodimers (Luger et al., 1997). The N- and C-terminals of the core histones extend through the DNA and provide sites for a variety of post-translational modifications (PTMs). Both the histone tails and internal globular domains can serve as the sites of the PTMs. Majority of modifications result in reversible additions of relatively small and structurally distinct moieties, such as methyl, phosphate and acetyl groups by histone modifying enzymes (Taverna et al., 2007). To date, eight PTMs (Fig. 1-5) and over 70 sites for modifications have been reported. Despite such a large number of potential modification sites, only a small subset of amino acid residues is known to be modified. So far, lysine (K), arginine (R), serine (S), threonine (T), tyrosine (Y), histidine (H) and glutamic acid (E) are known to undergo covalent PMTs (Allis et al., 2006). Moreover, additional complexity arises from methylation of lysines and arginines, which can exist in one of the several forms: mono-, di-, or tri-methylation for lysines, and mono- or di-methylation for arginines. Not all histones have the same modifications, therefore it has been suggested that distinct modifications “act sequentially or in combination to form the ‘histone code’ that is read by other proteins to bring about distinct downstream events” (Strahl and Allis, 2000). Ultimately, histone modifications coordinate gene expression programs by marking different genes for either transcriptional activation or silencing.

illustration not visible in this excerpt

Figure 1-5: Different classes of chromatin modifications. Examples of major histone PTMs, modified residues and associated regulatory functions are shown. (Kouzarides, 2007)

In attempt to understand the functional role of histone modifications and their effect on chromatin and chromatin-based events, two models have emerged. One model describes a more direct role of histone modifications that affect the compaction of DNA around the histone core. For example acetylation of lysine residue can neutralize the normally present positive charge, affecting the binding affinity of the negatively charged DNA and making it more accessible to the transcription machinery. Indeed, transcriptionally active promoters are usually H3 and H4 hyperacetylated - a mark that results in more open chromatin conformation (Berger, 2007). The second model is based around the ‘effector-mediated’ mechanisms. In this paradigm, histone PTMs are read by protein modules (effectors) that facilitate recruitment and assembly of variety of other regulatory proteins (Seet et al., 2006; Ruthenburg et al., 2007). One such example is tri-methylation of lysine 9 on histone 3 (H3K9me3) and its association with Heterochromatin Protein 1 (HP1) that can recognize H3K9me3 by its chromo-domain and form transcriptionally silent heterochromatin (Vakoc et al., 2005). Interestingly, histone lysine methylation has been correlated with both gene activation and repression. H3K4me3 has been shown to highly correlate with active transcription start sites, H3K27me3 is linked to facultative heterochromatin, and H3K9me3 is correlated with constitutive heterochromatin (Guenther et al., 2007; Berger, 2007; Kouzarides, 2007). Furthermore, 'bivalent domains' contain H3K4me3 and H3K27me3 marks that can be simultaneously detected on promoters of select genes during stem cell development (Berger, 2007; Kouzarides, 2007).

1.1.6 Histone modifications are brought about by specific histone modifying enzymes

Histone post-translational modifications modulate chromatin structure and function. Such modification patterns are regulated by the histone modifying enzymes. Three main types of histone modifying enzymes have been described: ‘writers’, ‘editors’ (or ‘erasers’) and ‘readers’ (Fig. 1-6). For instance, active genes in a specific cell lineage are marked with active histone modifications like H4 acetylation and H3K4 mono- di- and tri-methylation, while the same genes are enriched with repressive marks like H3K27 tri-methylation in cell types where they are silenced (Butler et al., 2012). Methylation and acetylation are the most studied histone modifications. Such modifications are enzymatically reversible and are ‘written’ by lysine methyltransferases (KMTs) and lysine acetyltransferases (KATs), and ‘erased’ by lysine demethylases (KDMs) and histone deacetylases (HDACs) (Butler et al., 2012).

illustration not visible in this excerpt

Figure 1-6: Histone modifying enzymes. ‘Writers’ are enzymes that establish reversible modification marks on histone tails. These marks can be removed by enzymes known as ‘editors’. ‘Readers’ of histone modifications recognize modifications and mediate their association with other protein complexes. HMT (histone methyltransferase), HAT (histone acetyltransferase), KDM (lysine-specific histone demethylase), HDAC (histone deacetylase), PPP (serine/threonine protein phosphatase). (Plass et al., Nat Rev. 2013)

The H3K4 mono-, di- and tri-methylation is induced by Trithorax (TrxG) and Trithorax related group of chromatin modifiers which act as transcriptional activators, antagonizing the effect of Polycomb Group (PcG) family of modifiers which induce repressive H3K27 tri-methylation marks (Butler et al., 2012). The Trithorax and Polycomb groups of proteins are also involved in regulation of major pluripotency genes such as OCT4 and Nanog, which are involved in proliferation and differentiation decisions in stem and progenitor cells (Ringrose et al., 2004, Butler et al., 2012). A direct link between TrxG complex subunit WDR5 and regulation of OCT4 expression through H3K4me3 modification has clearly been established (Ang et al., 2011). PcG proteins are also involved in maintaining pluripotent state of embryonic stem cells. Polycomb repressive complexes PRC1 and PRC2 repress a large number of critical transcription factors that are indispensable for stem cell pluripotency (Boyer et al., 2006, Butler et al., 2012). The PRC1 catalyzes monoubiquitylation of histone H2A and PRC2 is responsible for tri-methylation of H3K27 (Yoo and Hennighausen, 2012). Enhancer of zeste 2 (EZH2), a well described methyltransferase component of the PRC2, plays an essential role in the epigenetic maintenance of the H3K27me3 repressive chromatin mark. Abnormal expression of Ezh2 has been associated with various cancers and knockouts influence proliferation and differentiation of stem cells (Wang et al., 2010; Ezhkova et al., 2009). H3K27me3 mark has been shown to be closely associated with over 500 developmental regulator genes in embryonic stem cells (Boyer et al., 2006; Lee et al., 2006). In addition to accumulation of Xist RNA at the early stages of X chromosomal inactivation, H3K27me3 mark is established on the inactivating X chromosome by the PRC2 complex for additional level of repression (Plath et al., 2004).

Unlike H3K27me3, which is found preferentially in gene-rich regions and associated with CpG enriched sequences, H3K9me3 is another a silencing mark that is mainly present on gene poor regions and tandem repeat sequences (Pauler et al., 2009; Kim J. and Kim H., 2012). Both marks appear to be more prevalent in the embryonic stem cells, and become less widespread upon differentiation (Kim J. and Kim H., 2012). These repressive modifications are differentially changed to a more permanent repression by DNA methylation in somatic cells (Meissner et al., 2008). Interestingly, most genomic regions with H3K27me3 mark are protected from DNA methylation, while regions enriched in H3K9me3 become DNA-methylated during differentiation into somatic cells (Kim J. and Kim H., 2012). H3K27me3 is considered a temporary repressive modification of facultative heterochromatin and H3K9me3 is a permanent silencing mark of constitutive heterochromatin. In contrast to H3K27me3, which is generated by a single protein EZH2, H3K9me3 is known to be established by several enzymes: SETDB1 (or ESET), Suv39H1, Suv39H2, EHMT (GLP) and EHMT2 (G9A). Suv39H1 and Suv39H2 establish H3K9me3 in telomeric, pericentromeric, and constitutive heterochromatic regions (Peters et al., 2001). EHMT1 and EHMT2 are responsible for establishing H3K9me3 in euchromatic regions (Tachibana et al., 2002). SETDB1, known as a major regulator of pluripotency and self-renewal, is highly expressed in embryonic stem cells (Yeap et al., 2009). Both EZH2 and EHMT2 (G9A) have been shown to interact with Dnmt3a and Dnmt3b methyltransferases responsible for DNA methylation (Viré et al., 2006; Epsztejn-Litman et al., 2008). Further purification studies showed that isolated multiprotein-complexes contained both histone and DNA methyltransferases, suggesting dual functions of such complexes - DNA and histone methylation (Cedar and Bergman, 2009).

1.1.7 Disruption of histone modifying enzymes affects replication dynamics

Different chromatin domains replicate at different times during S-phase, however direct connection between specific histone PTMs and the firing of replication origins at those domains has not yet been described. Nevertheless, some evidence has shown that independently of DNA methylation, late replicating constitutive heterochromatin replicated earlier upon hyperacetylation, suggesting a clear role of histone modifications in dictating replication timing (Donaldson, 2005).

Moreover, under normal physiological conditions the inactive X chromosome (Xi) exhibits a typical mid S-phase replication pattern in somatic mouse cells. Casas-Delucchi and colleagues successfully demonstrated that inhibition of HDAC, enzyme responsible for histone hypoacetylation, resulted in reduced percentage of cells exhibiting Xi replication pattern. Additionally, they showed that conditional knockout of Ezh2 also influenced Xi replication and had a dramatic effect on proliferation of mouse embryonic fibroblasts (Casas-Delucchi et al., 2011). Together these results indicate that removing enzymes responsible for specific histone modifications or blocking their enzymatic activities can grossly affect replication dynamics and cell proliferation.

H3K9me3-enriched pericentric heterochromatin was shown to replicate almost exclusively in the mid-to-late S-phase. However, despite crucial role of H3K9me3 in heterochromatin organization, disruption of Suv39H1/H2 or G9A methyltransferases in mouse embryonic stem cells did not have an observable effect on replication dynamics (Wu et al., 2005). Given enzymatic redundancy, it is possible that other enzymes compensate for the change in methylation state. It is also possible that other, or a combination of multiple modifications establish spatio-temporal control to dictate replication timing (Wu et al., 2005).

In eukaryotes, centromeres are responsible for proper segregation of sister chromatids during cell division. Centromeres are constitutively heterochromatic and remain condensed during interphase. Two types of repetitive sequences are associated with centromeres in mouse cells - major and minor satellite repeats (Fig. 1-7) (Choo, 1997). Although both repeats exhibit H3K9me3 mark, HP1-α specifically accumulates on major satellites, and majority of centromeric proteins (CENP-A,B,C) are associated only with minor satellites (Guenatri et al., 2004). Differences also exist in higher order heterochromatin organization between these two repeats. Chromatin immunoprecipitation revealed that major satellites contain distinct H3K9me3-rich dinucleosomes. Additionally, both sequences replicate at different times during S-phase. Major satellites replicate in mid S-phase, later followed by the minor satellite repeats (Guenatri et al., 2004).

illustration not visible in this excerpt

Figure 1-7: Schematic of a typical acrocentric mouse chromosome. The location of telomere (black), minor satellite (red), major satellite (green) and long arm of the chromosome (blue) are shown (Top). Localization of major satellites (green) and minor satellites (red) by Fluorescent in situ hybridization (FISH) analysis on metaphase chromosome stained by DAPI (blue) (a, b). (Guenatri et al., 2004)

Investigations in somatic mouse cells lacking both Suv39h1 and Suv39h2 methyltransferases showed severely reduced HP1-α association at pericentric heterochromatin. Guenatri and colleagues revealed that the loss of both H3K9me3 mark and HP1-α at pericentric major satellite repeats in Suv39h1,2 -/- cells lead to defect in chromatid cohesion specific to major satellites. Separation of minor satellite chromatids showed no observable effect. Differential replication timing of these repeats appears to be a “key feature for centromere organization and function” (Guenatri et al., 2004).

Ample of evidence has shown that manipulating chromatin can affect replication dynamics (Casas-Delucchi and Cardoso, 2011). However, histone modifications rarely affect chromatin independently, and are often characterized by strong epigenetic crosstalk. Replication dynamics are complex and involve a multitude of factors, direct roles of which are yet to be elucidated. Recent advances in microscopy, accurate visualization of actively replicating structures and improved live-cell imaging techniques are currently advancing our understanding of replication dynamics.

1.2 Breaking the resolution limit with super resolution microscopy

Despite major advances in electron microscopy, about 80% of all microscopy investigations in life sciences are carried out with conventional lenses and visible light (Hell, 2007). Light microscopy allows for non-invasive, 3-dimensional (3D) imaging of the cell and detection of fluorescently tagged nucleic acids, lipids, and proteins. However, dating back to 1873, Ernst Abbe showed that the smallest resolvable distance between two points using a conventional light microscope can never be smaller than half the wavelength of the imaging light - that is limited by diffraction to about 200 nm (Abbe, 1973). Some improvements can be achieved by using light with a shorter wavelength, although in biological samples this comes at the expense of tissue damage. Only recently scientist were able to push the Abbe’s resolution limit with a variety of new approaches that allow for improved fixed and live cell imaging. One such technique is 3D structured illumination microscopy (3D-SIM).

1.2.1 Principles and applications of 3D structured illumination microscopy

To exceed the classical resolution limit, structured illumination can be used in a wide-field fluorescent microscope. The sample can be illuminated with a series of excitation light patterns causing normally inaccessible information to be encoded into observed image, which can further be mathematically reconstructed to achieve an image with twice the normal resolution (Gustafsson, 2000). In 3D-SIM, a sample is illuminated with a series of striped patterns of high spatial frequency by light passing through a movable optical grating (Fig. 1-8) (Schermelleh et al., 2008; Heintzmann and Cremer, 1999; Gustafsson, 2000). As a result of this interference, high spatial frequency information, which is normally below the diffraction limit, is shifted to lower frequency patterns, generating ‘Moiré fringes’ which contain additional information that was previously unresolved. These ‘Moiré fringes’ appear in the emission distribution and can be transferred to the image plane by the microscope. Application of these patterns in different orientations (3X 60° rotation), and processing all acquired images using mathematical algorithms, generates a super-resolution image. The result is 100-130 nm lateral (XY), and 250-350 nm axial (Z) resolutions, allowing an approximately eightfold smaller volume to be resolved, compared to conventional light microscopy (Schermelleh et al., 2008).

illustration not visible in this excerpt

Figure 1-8: Basic principle of 3D-SIM. When fluorescent sample is illuminated by light passing through a movable optical grating, normally unresolved high-frequency information is shifted to lower spatial frequency patterns, generating Moiré fringes. These patterns lead to acquisition of information below diffraction limit. Each z-section has to be imaged with 5 different phase shifts at different orientations (60° rotation, three times), resulting in 15 images per z-slice. Further mathematical reconstruction of all images generates super-resolution image (Schermelleh et al., Rev. 2010).

Standard dyes and staining protocols can be used in multicolor analysis of cellular structures in three dimensions. For example replication foci could be visualized and analyzed by 3D-SIM (Baddeley et al., 2010). Reconstructed data can be further used for a more precise quantitative analysis. Moreover, in combination with fluorescent in situ hybridization (FISH), 3D-SIM has opened new possibilities to precisely analyze nuclear architecture and chromatin dynamics of specific DNA sequences at the ultrastructural level, particularly during S-phase (Markaki et al. 2010; Markaki et al., 2012; Smeets et al., 2014). Full analysis and interpretation of superresolution data is often limited by processing power and inefficient data processing tools. To further enhance our understanding and analysis of nuclear architecture and get closer to molecular ratios and interactions, large sets of super-resolution data can be analyzed by a variety of recently developed image analysis tools described below.

1.2.2 Utilization of 3DSIM data for analysis of nuclear genome organization

Several programs such as CellProfiler, FISH Finder, and NEMO have already been developed for large scale analysis of image-derived data. CellProfiler only performs 2D analysis, and FISH Finder and NEMO have limited image processing and analysis capabilities (Ollion et al., 2013). Tools for Analysis of Nuclear Genome Organization (TANGO), is a recently developed tool that attempts to overcome limitations of other software. TANGO is an open source image analysis tool dedicated for quantitative analysis of nuclear architecture. It permits biologists without programming skills to perform a complete analysis of 3D fluorescence images: image processing, quantitative analysis, and statistical processing (Ollion et al., 2013). It is made up of ImageJ plugin for batch processing, and R package for statistical analysis. Combination of high-resolution images obtained from 3D-SIM and TANGO is a powerful tool which was recently used to gain a more precise understanding of underlying chromatin architecture of inactive X chromosome (Smeets et al., 2014). Bridging the gap between attaining information and quantifying, processing, and analyzing it is the key to improve our current understanding of many cellular processes.

1.3 Aims of this study

DNA replication is a fundamental biological process, which remains to be fully understood. Although many factors involved in replication exhibit high degree of evolutionary conservation, direct and precise involvement of each factor remains elusive. In recent years, ‘epigenetic’ factors have been shown to play crucial roles in majority of biological processes, particularly during development, however their precise roles in S-phase dynamics and initiation of DNA replication are not well defined. Major epigenetic mechanism, such as histone PTMs, has been recently shown to influence chromatin structure and suggested to play a role in defining spatio-temporal replication dynamics. The direct involvement of specific modification in dictating replication dynamics is currently an area of active research.

The primary aim of this study was to establish a systematic process to register replication patterns in somatic cells using a light microscopy approach, 3D-SIM. The obtained data was further utilized with software solutions to establish an automated segmentation and quantification workflow to measure more accurately than before the number of replication foci (RF) in specific S-phases. Additionally, a comparison between data obtained from 3D-SIM and reconstructed wide-field deconvolved images (pseudo-confocal) of the same cells was analyzed, and the increased RF numbers for individual S-phases were assessed.

The second aim of this study was to utilize well-established molecular tools (modified nucleotides, FISH, antibody staining, and introduction of RFP-PCNA overexpression vectors) to characterize S-phase progression and duration in mouse embryonic stem cells (mESCs), both undifferentiated and differentiated, and compare them to the already existing models in somatic cells.

The third aim was to investigate potential roles of histone PTMs in replication patterns and timing by interfering with specific histone modifications with inhibitors and knockout cell lines. The main goal was to analyze the effect of inhibitors and knockouts on progression of S-phase in somatic and undifferentiated mESCs, as well as their effect during differentiation.

2. Materials

2.1 Cell lines

Table 2-1: Cell lines

illustration not visible in this excerpt

2.2 Antibodies

Table 2-2-1: Primary antibodies

illustration not visible in this excerpt

Table 2-2-2: Secondary antibodies

illustration not visible in this excerpt

2.3 Fluorescent in situ hybridization primers

Table 2-3: FISH primers

illustration not visible in this excerpt

[...]