Replication timing program in mammalian systems



1. Replication timing program in mammalian systems

DNA contains the genetic instructions used in the development and functioning of all known living organisms. In eukaryotes, DNA, together with a lot of proteins, is packaged into chromatin to carry complicated epigenetic information and to fit into the tiny nucleus of cells. During each cell division, the genetic and epigenetic information have to be duplicated in order to transmit everything necessary to daughter cells. Therefore DNA replication is the absolute central and fundamental process used by all living organisms as it is the basis for biological inheritance. All genetic information has to be copied faithfully for proper development and for reproduction. Any mistake during DNA replication will have dramatic effect on the function of the genome and therefore is one of the most basic reasons for many diseases: infection diseases, inheritance diseases, cancer, and so on {DePamphilis, 2006 #1267}.

DNA replication has been the focus of many great scientists ever since the double-helix structure was found by Watson and Crick. In their original Nature article, Watson and Crick wrote that "It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material" {Watson, 1953 #1268}. During DNA replication, chromatin is opened up by replication machinery and new DNA strands are synthesized. Nature provides such an elegant window to repackage DNA and therefore regulation of DNA replication, including the process per se and the steps to prepare DNA replication, is critical to maintenance of genetic and epigenetic information. We have gained great understanding of basic DNA replication mechanism, thanks to many studies done in bacteria and single eukaryotic budding yeast systems in the last half a century. For example, we know how DNA double strand is opened up to form a replication bubble and how many replication proteins associate together to start replication at replication fork. But there are still a lot of critical questions unsolved yet. In particular, how DNA replication precedes in a packed chromatin environment is a very basic but important question. Human genome contains 3 billion base pairs (the basic unit of DNA). Replication of this huge genome takes place in usually about 10 hours in S-phase of the cell cycle. Such DNA replication can be regulated at many levels, for example specification of replication origins and control of origin license (in other words, where and how to start DNA replication) and the temporal order of DNA replication (when a particular DNA sequence replicates). And of course we are far from understanding the coordination of DNA replication with other basic processes such as transcription and DNA repair. The question of "when" or the replication timing program is the focus of this dissertation.

S-phase progresses through synchronously firing of clustered origins within chromosomal regions (replication domains). Each domain duplicates at a precise time in S-phase, forming an early �C late replication timing program. Such program is re-established and maintained in very cell division until cells change their fate or cells respond to environmental factors.

In eukaryotic organisms, from yeast to human, replication of genome does not start from all origins synchronously or randomly but at different times during S phase. The replication proceeds via synchronously firing of clustered origins within chromosomal segments with each segment duplicating at a precise time, choreographed to form the same temporal order of replication from cell division to cell division, unless cells change their identity {Hiratani, 2009 #1238}. Such temporal order forms the tightly controlled replication timing program (Figure 1). It has been known for a long time that in general euchromatin and transcriptionally active genomic regions replicate early while heterochromatin and silenced regions replicate late in S-phase. Euchromatic and heterochromatic chromosomal domains are segregated into different compartments in the nucleus. Sub-nuclear positions of sequences reflect, at least in part, their transcriptional potential (Hiratani and others). For example, silent genes are generally less mobile and closer to the nuclear periphery and the nucleolus than active genes. Since euchromatin and heterochromatin usually have such distinct subnuclear localizations, the temporal replication order is closely related to the spatial organization and sometimes called spatio-temporal program.

Given the connection between replication timing and chromatin subnuclear organization, such a replication timing program can be monitored by pulse labeling cells with thymidine analog BrdU at different times during S-phase. During DNA replication BrdU will be incorporated into replicating DNA in place of thymidine which can be further stained with BrdU antibodies. In doing so, it has been revealed that DNA replication takes place at discrete nuclear sites, generally called replication foci, which form some characteristic replication patterns. For example, in Chinese hamster ovary (CHO) cells, 5 replication patterns can be identified (Figure 2). While for mouse C127 cells, we can identify 6 distinct replication patterns (Figure 3). This is largely due to existence of chromocenters, clusters of centromeres and pericentric heterochromatin (and may include temoleres) from several chromosomes, which replicate in the second half of S-phase and have characteristic replication patterns (mainly pattern IV).

CHO cells were synchronized in mitosis by selective detachment and released into medium containing aphidicolin to arrest them at the G1/S border. Aphidicolin allows the synthesis of short primers but prevents their processive elongation, arresting replication forks close to the earliest firing origins. At various intervals after release from the aphidicolin block, cells were pulse-labeled with BrdU and indirect immunolocalization of the sites of replication was performed using an anti-BrdU monoclonal antibody. Type I patterns are observed during the first 10-30 minutes of S-phase. Throughout the remaining first half of S-phase (6 hours), nearly all nuclei display Type II patterns. Since Type I is brief and is likely a precursor to Type II, we refer to these patterns collectively as Type I/II. During the next 2-3 hours, Type III predominates. Type IV, and then Type V appear during the last two hours of S-phase. Unless otherwise noted, we refer to Type III through V collectively as late-replicating patterns.

Mouse C127 cells were pulse labeled with BrdU and stained with anti-BrdU antibodies and DAPI. Shown are de-convolution ima ges of each consecutive labeling pattern (I-VI). DAPI show s intense staining of peri-centromeric heterochromatin (chromocenters) characteristic of mouse cells, as well as the nuclear periphery. Chromocenters associate with nucleoli, visualized as DNA-free areas that are devoid of DAPI or BrdU staining. DNA synthesis begins at many small, discrete foci in the internal, euchromatic region of the nucleus, excluding nucleoli (and associated chromocenters) and the nuclear periphery (pattern I). In pattern II, replication continues throughout the euchromatic region, but is also observed at the nuclear and nucleolar periphery. Pattern III is characterized by fewer foci in the interior and increased replication foci at the nuclear and nucleolar periphery. By mid-S-phase, most euchromatic foci have finished replication and DNA synthesis begins within chromocenters (pattern IV). Thereafter, replication at the nuclear periphery becomes more predominant, coinciding with the replication of a few internal but non-peri-centric domains (pattern V). Finally, small numbers of large foci are observed in both the interior and periphery of the nuclei (Pattern VI). (Adapted from Wu, 2006 JCB).

2. Biological significance of replication timing program

Since it was recognized in 1960s that different regions of the eukaryotic genome replicate at specific times during S phase {Taylor, 1968 #226;Taylor, 1968 #225}, the question regarding the significance of replication timing program has puzzled scientists a lot, mostly due to failure in isolating mutants based on replication timing defects. Several possibilities have been proposed for the potential biological function of differential replication timing {Kim, 2001 #155;Wintersberger, 2000 #154}. First, the programmed replication timing might be the adaptation of cells to limited reservoir of dNTPs, DNA polymerases and other replication proteins. Simultaneous replication of all the regions would otherwise demand for higher level of dNTPs and replication proteins than the cell can provide. Second, the programmed replication timing might allow for setting checkpoints during the S-phase progression. In time of cellular emergency, such as DNA damage and dNTP starvation, intra-S checkpoint would suppress the firing of later origins to ensure the faithful duplication of whole genome sequences. Third and most importantly, accumulating evidence suggests that replication timing may play a role in regulating other chromatin-associated cellular processes, such as gene expression {Goren, 2003 #157;Simon, 1996 #156}.

As of now, the exact functional significance of replication timing is not clear yet. But temporal control of DNA replication is linked to many basic cellular processes, both in cell-cycle control and development. Changes in replication timing accompany key stages of development {Hiratani, 2004 #4;Perry, 2004 #1}. Mammalian X-chromosome inactivation is accompanied by a nearly chromosome-wide switch from early- to late-replication {Maxfield Boumil, 2001 #1976;Avner, 2001 #2018} at early embryogenesis, shortly after Xist RNA expression {Chow, 2009 #1288}. Without the replication timing switch, X-chromosome is not stably inactivated. In general, developmentally regulated genes often replicate late in most cell types and early in cell types in which they are expressed (Hiratani). For example, at the mammalian ?-globin locus, over 200kb of DNA is early-replicating and DNase I sensitive in erythroid cells but late-replicating, heterochromatic, and DNaseI resistant in cultured fibroblasts (Ref 22, 28). Activation of ? -globin in erythroblasts correlates with a switch of the locus to early replication. Fusion of non-globin expressing cell lines with globin expressing lines can cause either activation and early-replication of an inactive globin gene, or extinction and late-replication of an active globin gene (Ref 23).

Replication timing program has been found to be involved in many human diseases, potentially by affecting genome stability. Changes in replication timing have been reported at the loci responsible for some diseases conditions such as the Tourette syndrome, Di George velocardiofacial syndrome, and Prader-Willi syndrome {State, 2003 #1347;D'Antoni, 2004 #1348;Gunaratne, 1995 #1342}, and in a number of blood disorders {Amiel, 2002 #971}. Allelic synchronization (two alleles in one cell replicate at the same time during S-phase) is frequently lost in cells derived from patients with various forms of leukemia, correlating with higher frequency of aneuploidy (Korenstein-Ilan et al., 2002). In non-Hodgkin lymphoma, asynchronous allelic replication is indicative of higher risk of relapse (Amiel et al., 2001). In addition, cells from patients with several inherited human diseases show defects in replication-timing which are likely related to mis-regulation of genes during development {Barbosa, 2000 #2014;Hansen, 2000 #2012;Hellman, 2000 #2015;Reish, 2002 #2130}. Moreover, unscheduled alterations in the replication timing program may delay the process of chromosome condensation, sister chromatid cohesion, and increasing chromosome instability {Loupart, 2000 #1801;Pflumm, 2001 #1875}{Breger, 2005 #1315}. This sequence of events has been shown in a number of tumor cell lines and primary tumor samples (Smith et al., 2001). It is thought that late-replicating regions and the transition points between early- and late-replicating domains are likely more prone to suffer replicative stress. The molecular composition of the boundaries between replication domains could contain fork-pausing or barrier elements to prevent replication forks from the early domains to progress into the late domains. Although the details are not clear, such fork barriers are hot spots for DNA breaks and chromosome rearrangements (reviewed by Rothstein et al., 2000). Interestingly, several cancer and other disease-related genes are located in replication timing switching regions and may therefore be more prone to inactivation (Watanabe et al., 2002). Genome-wide studies also showed that the majority of common fragile sites and rare fragile sites in the whole genome are late-replicating (reviewed by Durkin and Glover, 2007). Late-replicating regions are rich in single-nucleotide polymorphisms (Watanabe et al., 2002) and display an increased mutation rate (Stamatoyannopoulos et al., 2009). These intriguing properties of late-replication domains could be due to the inherent difficulties of replicating the more compact heterochromatic DNA, which could result in an accumulation of ssDNA that is more susceptible to damage. As the list of RT alterations associated to disease continues to grow, it will become very important to ascertain the specific contribution of replication timing changes to these phenotypes.

Replication timing potentially modulates transcription activity / chromatin structure. Since chromatin is assembled at the replication fork, replication timing could potentially dictates chromatin states that in turn will regulate replication timing in the subsequent cell cycle, providing a means of epigenetic inheritance during somatic development. The most compelling, while indirect, evidence that different chromatin structures might be assembled at different times during S-phase is that reporter plasmids injected into early or late S-phase mammalian nuclei assembled into hyper- or hypo-acetylated chromatin, respectively {Zhang, 2002 #2134}. The established transcriptional status is mitotically inherited. This result prompted speculation that replication fork proteins like the replication clamp and multi-functional scaffold protein PCNA might recruit different chromatin modifiers at different times during S-phase to assemble different types of chromatin {McNairn, 2003 #2314}. Unfortunately, although dozens of proteins have been localized to replication forks, only two have shown evidence for temporally-specific recruitment (HDAC2 {Rountree, 2000 #2010} and MBD2-MBD3 {Tatematsu, 2000 #2795} localized specifically to sites of late DNA synthesis and both reports monitored ectopically expressed tagged versions of these proteins. Hence, evidence for endogenous proteins that localize to replication sites at specific times during S-phase has been disappointing. Although it is still appealing to envision a scenario in which replication timing program plays an important role in modulating transcriptional activity, how exactly this is realized is still a mystery. Moreover, since replication is regulated at the level of replicons or a cluster of replicons, a change in replication timing could rapidly transmit a change in chromatin state to many genes simultaneously. Clearly this attractive model merits a definitive test, but it is currently impossible to manipulate replication timing without affecting other properties of a chromatin domain. And as will be summarized in the next section, the relationships are so complex that there is no way to establish a simple cause �C effect relationship.

3. Mechanism of replication timing program

Although there are so many linkage data connecting replication timing program to other biological processes as summarized above, the significance of such a program is not generally recognized. This is at least in part due to the lack of mechanistic insights which renders us unable to manipulate it.

3.1 Potential candidates analyzed

3.1.1 DNA sequence does not determine replication timing.

First of all, the replication timing program is not determined by DNA sequence per se. Loci on the transcriptionally active mammalian X chromosome are early replicating, while the homologous loci on the inactive X in females are late-replicating {Schmidt, 1990 #1284}{Hansen, 1993 #1285}{Boggs, 1994 #11;Gartler, 1999 #29}. In addition, allelic differences often exist in the timing of replication of imprinted gene loci {Kitsberg, 1993 #12;Korenstein-Ilan, 2002 #13}{Kitsberg, 1993 #420}{Greally, 1998 #1286}{Chess, 1994 #1287}. Moreover, recent genome-wide replication timing studies from our and others have revealed obvious replication timing switches during early mammalian development, supporting that DNA sequence does not dictate replication timing.

However, DNA sequences that affect transcription may affect replication timing. For example, the LCR at the human "-globin locus can accelerate replication of extended chromosomal domains at the native locus. But it can also delay replication at ectopic locations, similar to its behavior in non erythroid lineage cells. Another intriguing phenomenon was observed by Thayer and colleagues that some chromosomal translocations led to delayed replication timing of entire chromosomes {Breger, 2005 #1315}. All these are potentially secondary effects caused by the chromatin structure modifications, although the exact mechanism is not clear.

3.1.2. Transcription does not seem to directly affect replication timing

A correlation between transcriptional activity and replication timing has long been observed. In general, transcriptionally active domains (euchromatin) replicate early in S phase and transcriptionally repressive domains (heterochromatin) replicate late {Holmquist, 1987 #158}. Recent genome-wide studies revealed a positive correlation between early replication timing and high transcription probability. Significant changes of replication timing are often coupled to changes in transcriptional activity during X-chromosome inactivation, embryonic stem cell differentiation and lymphocyte differentiation {Ermakova, 1999 #159;Hibbard, 1999 #160;Mostoslavsky, 2001 #161;Boggs, 1994 #162;Hiratani, 2004 #185}. In addition, monoallelic expression is found to be closely associated with asynchronous replication timing {Gimelbrant, 2005 #251}, in such cases, the expressed allele tends to replicate earlier than the silent. It was also found in budding yeast that targeting a silencing Sir protein to an early replication origin could induce late replication together with transcriptional silencing of a neighboring gene {Zappulla, 2002 #1132}. And heritably repressed chromatin usually is late replicating, a property conserved in all eukaryotes {Gilbert, 2002 #16}. The general correlation is well-established but still no direct causal relationship has been established between transcription and replication timing by any experimental data.

In addition, exceptions to the correlation between replication timing and transcriptional activity are also commonly observed. Early-replicating heterochromatin that is transcriptionally silenced exists widely in fission yeast {Kim, 2003 #258}. Genome-wide analysis performed in budding yeast does not reveal a connection between replication timing and transcriptional activity {Raghuraman, 2001 #257}. Even in metazoans, the correlation is not perfect. Early-replicating heterochromatin in metazoans has also been reported {Kim, 2003 #258}. Moreover, several lymphocyte specific genes do not replicate late when their transcription is repressed during B lymphocyte differentiation {Azuara, 2003 #259}. In Drosophila, many late-replicating heterochromatic regions also contain transcriptionally active genes {Schubeler, 2002 #252}. A recent paper from {Prioleau G&D accepted?} showed the presence of a proximal gene promoter that is highly acetylated and transcriptionally active does not convey any replication timing shift.

Recent microarray studies have allowed us to sharpen hypotheses regarding replication timing and transcription {reviewed in Hiratani 09 COGD; Woodfine, 2005 #255;Woodfine, 2004 #256;MacAlpine, 2004 #253;Schubeler, 2002 #252;White, 2004 #254; Hiratani more}. It is now clear that the relationship between replication timing and transcription is similar across cell types and species in metazoan and is confined to a specific class of genes. In Drosophila, mouse, and human, most genes replicate in the first third of S-phase and have an equally high probability of being expressed independent of their replication time within this period; a strong relationship between earlier replication timing and transcription is restricted to the 25% of genes that replicate later during S-phase {Schubeler, 2002 #2068;Hiratani, 2008 #2765}. In mammalian cells, early-replicating genes are enriched for high CpG-density promoters, while late-replicating genes are enriched for low CpG-density promoters {Hiratani, 2008 #2765}, which maintain their repressed state even upon loss of DNA methylation or treatment with TSA {Lande-Diner, 2007 #2783}. More powerful than static correlations is the examination of dynamic changes in replication timing and transcription for genes residing within the timing "switching" domains described above, which can reveal distinctions between the majority of genes whose transcription changes coordinately with replication timing versus those that don"t. For instance, high and low CpG-density promoters, which generally possess strong and weak promoter activity, respectively, showed distinct behaviors upon entering a late-replicating environment, with the latter showing higher tendency toward transcriptional down-regulation {Hiratani, 2008 #2765}. Thus, the occasional strong promoter that finds itself located in a replication timing switching domain may "come along for the ride" but be unaffected by the replication timing change. This is consistent with reports in several systems that strong promoters can overcome heterochromatin silencing ({Hiratani, 2008 #2765} and references therein) as well as the recent observation that tethering a chromosomal region to the nuclear periphery repressed some genes but not others {Reddy, 2008 #2781; Kumaran, 2008 #2779; Finlan, 2008 #2780}. Overall, these results reinforce the notion that a strong association exists between replication time and transcription for a specific class of genes, but that this relationship is not a direct one, probably mediated by some aspect of chromatin or chromosome structure.

3.1.3. Alterations in chromatin structure induce modest changes in replication timing

The effect of chromatin structure on replication timing is also not strict forward. On one hand, it is known that during X-chromosome inactivation, the replication timing switch is always coupled to the appearance of essential histone modifications for silenced chromatin structure - histone H3 K9 methylation and histone H3 K27 methylation {Heard, 2004 #261}. We also showed that active modification Me1K9H3 was largely restricted to early replicating, small punctate domains in the nuclear interior. Me2K9H3 was the predominant MeK9 epitope at the nuclear and nucleolar periphery and colocalized with sites of DNA synthesis primarily in mid-S phase. Me3K9H3 decorated late-replicating pericentric heterochromatin and sites of DAPI-dense intranuclear heterochromatin {Wu, 2005 #846}. Multiple studies also suggest that chromatin modifications directly regulate replication timing. Chemical inhibition of histone deacetylases (HDAC) can partially advance replication timing of several mammalian genes {Kagotani, 2002 #313;Bickmore, 1995 #1455} as well as the Epstein Barr Virus mini-chromosome {Zhou, 2008 #2794}. Consistently, histone hyperaceylation induced by 5-aza-2'-deoxycytidine treatment also correlates with advanced replication at murine pericentric heterochromatin {Takebayashi, 2005 #329;Selig, 1988 #324}. Over-expression of a chromatin remodeling complex NoRC delays replication of rRNA genes {Li, 2004 #2431}. In budding yeast, silent chromatin proteins SIR3 {Stevenson, 1999 #1478} and the HDAC Rpd3 {Vogelauer, 2002 #2360;Aparicio, 2004 #2330} delay the firing of specific origins, while tethering of histone acetyltransferase (HAT) partially advances origin firing {Vogelauer, 2002 #2360}. By using either an erythroid differentiation system, or a RMCE-based experimental system in which the timing of DNA replication can be altered in a controlled manner, it has been demonstrated that the replication timing of human ?-globin locus is tightly associated with the levels of histone acetylation: early replication correlates with hyperacetylation (de-condensed euchromatic chromatic states), while, late replication correlates with hypoacetylation (condensed heterochromatic chromatic states) {Lin, 2003 #314;Schubeler, 2001 #323}. Recently, HAT/HDAC-tethering to the human "-globin origin in mouse cells caused similar modest changes (~20% of S-phase) {Goren, 2008 #2772}. Nonetheless, the effects of any particular modification are relatively minor and in many cases are associated with other changes such as transcription and chromatin localization. Furthermore, little to no effect on replication timing of 20 gene loci were detected in Suv39 h1/2, Dnmt1, G9a, Eed, and Dicer mutant mouse ESCs or embryonic fibroblasts {Wu, 2006 #2627;Jorgensen, 2007 #2757}. In the only whole-genome study, no gene was found to change its replication timing after G9a knockout {Yokochi, 2009}. Even in cases there was detectable changes, the effect was limited to specific chromatin regions such as pericentric heterochromatin. Suv39h1/h2 histone methyltransferase (HMTase) mutation slightly advanced replication timing of pericentric heterochromatin in embryonic fibroblasts, but slightly delayed timing in ESCs {Wu, 2006 #2627;Jorgensen, 2007 #2757} [Table 1]. Whereas for G9a knockout, the replication timing of pericentric heterochromatin was slightly advanced in both cell systems ({Wu, 2005 #1306}{Yokochi, 2009}, and this manuscript). Indeed, the role of specific histone modifications in regulating replication timing has been difficult to fathom, since even origins that fire at the same time within the same segment of DNA have different histone modifications {Prioleau, 2003 #2302;Dazy, 2006 #2792;Norio, 2006 #2651}.

Table 1: Effect of Chromatin Modifiers on Replication Timing

Name Function Effect

HP1 Binding to H3K9 Me3 No change @ TDP

Suv39h1,2 H3K9 Me3 No change @ TDP, KO negative

G9a H3K9 Me2/Me1 KO negative, exp. satellite

MII H3K4 me KO negative

Eed H3K27 me KO negative

Mbd3 nucleosome remodeling/HDAC KO negative

Dnmt1 Maintenance of DNA methylation KO negative

Dnmt3a/3b de novo DNA methylation KO negative

Dicer RNAi, RNase KO negative

3.1.4 Sub-nuclear localization and replication timing

As mentioned earlier, in mammalian cells, the genome is organized into small functional domains {Berezney, 2000 #276}. Each domain, representing a cluster of coordinately activated replicons, replicates at a defined time during the S-phase. In most of the interphase, each domain occupies a distinct position within the nuclei and the spatial positioning is relatively stable throughout the cell cycle. In general, early-replicating domains are found in euchromatic regions in the interior of the nucleus, while late-replicating domains are retained in heterochromatic regions that are found either at the nuclear periphery or near nucleoli {Visser, 1998 #281;Zink, 1999 #280}. It has also been demonstrated that asynchronous replication timing of imprinted loci is consistent with differential sub-nuclear localization {Gribnau, 2003 #342;Kagotani, 2002 #313}. Moreover, relocation of IgH (immunoglobin heavy-chain) locus into nuclear periphery is coupled to delayed replication and repressed transcription during the maturation of B-cell {Zhou, 2002 #336}. A high-resolution study on CFTR and its adjacent genes further demonstrated the correlation between late replication and positioning at nuclear periphery {Zink, 2004 #283}. Therefore the spatial organization is closely related to the replication timing program.

Genome-wide replication timing studies also showed a link between replication timing and subnuclear positioning. When regions switching replication timing during embryonic stem cell differentiation were analyzed, differences in replication timing are associated with differences in subnuclear position for several gene loci {Williams, 2006 #2633;Hiratani, 2008 #2765}. Although a genome-wide survey of such relationship is currently impractical, these results imply the existence of a strong relationship.

However, inconsistency between sub-nuclear localization and replication timing has been reported in specific cell lines. For example, early replication was not found to be associated with more nuclear interior-toward localization at Igf2-H19 locus in a parthenogenetic ES cells {Gribnau, 2003 #342} and RET locus in a neuroblastoma cell line {Cinti, 2004 #344}. We also found late replicating peripherial localized genes (Yokochi, 09 and Hiratani ? ) through genome-wide studies. More importantly, similar to the relationship to transcription, it has been difficult to distinguish the causal relationship between replication timing and spatial positioning, partially because there is no molecular handles on either spatial localization or replication timing.

3.1.5 Cell cycle regulators

A few studies suggested that cell cycle regulation factors may control chromatin replication timing. For example using Xenopus egg extract and Xenopus sperm chromatin as template, Shechter et. al. {Shechter, 2004 #1067} suggested that DNA damage checkpoint proteins ATR and Chk1 could suppress late replicating origins. Similar effect was also shown in mouse embryonic fibroblast cells (MEF) by deleting Chk1 which activates cyclin-dependent kinase 1 (CDK1) �C cyclin A complex and promotes early firing of late origins {Katsuno, 2009 #1263}. However in some somatic cells, it was found that Chk1 could only prevent latent origins within active replicons from firing, while had no effect on global replication order {Maya-Mendoza, 2007 #979}. Further, potentially genome-wide, studies are necessary to clarify the discrepancy.

Nevertheless, since specific chromatin regions could change the time of replication during development (REFs) and there is frequent allelic asynchronous replication, it must be something intrinsic to chromatin that "prepares" chromatin for further readout in S-phase at a specific time. Moreover, since we found that the replication timing program is established during early G1-phase and these cell cycle regulators are not functional at that time, such trans-regulators could only be the potential "readers" of replication timing markers. (making sense??)

3.2. Establishment of replication timing program �C the Timing Decision Point (TDP)

Understanding the role of replication timing in transcriptional control or any other cellular process will require knowledge of what programs replication origins to fire at particular times during S phase. This has been a particularly challenging problem as summarized in the previous section. In fact the problem is more complicated in metazoa, because of difficulties in defining the sequence elements that comprise replication origins and the lack of convenient assays for replication {Gilbert, 2002 #858}. Here I summarize some findings from our lab using a unique in vitro replication system.

3.2.1 TDP (Timing Decision Point) Cell-free Xenopus egg extract in vitro replication system

Extracts from Xenopus eggs provide an ideal cell-free system for studying the regulation of DNA replication. The extract can virtually replicate any DNA substrates, including naked plasmid DNA, chromatin and whole cell nuclei {Madine, 1997 #1015;Wu, 1997 #204}. The extract is providing a rich source of S-phase promoting factors that can overcome cell-cycle checkpoints and drive nuclei staged at any time during G1-phase into S. Such replication initiation in Xenopus egg extracts can recapitulate in vivo events when the nuclei are prepared under appropriate conditions. This provides a unique opportunity to examine the steps during G1-phase that prepare cells for DNA replication. For example, using this system, we have identified one important time point in late G1 phase, origin decision point (ODP), at which the specificity of DHFR (dihydrofolate reductase) replication origins is determined {Wu, 1996 #199}. Most importantly, using knowledge about the requirements for passage through the ODP acquired using the in vitro system, we were able to identify conditions in which cultured cells would enter S-phase without specifying origin sites {Wu, 1998 #879}. the Timing Decision Point (TDP)

Using this Xenopus egg extract system, Gilbert lab discovered that replication timing program in CHO is established at a distinct point in early G1 phase termed Timing Decision Point (TDP) {Dimitrova, 1999 #166}. In summary, with nuclei isolated 1 hour after mitosis, (G1-1hr) initiation takes place within both heterochromatic and euchromatic domains without any preference. However, with nuclei isolated 2 hour or 3 hour (cell line dependent) after mitosis (G1-3hr), replication in vitro proceeds according to the proper temporal order that S-phase cells go through. Throughout this manuscript, we will refer to these two time points as the pre-TDP and post-TDP stages of G1-phase, respectively. The conclusion obtained from global chromosome domains was further confirmed by experiments dealing with specific chromosomal loci including DHFR (early replicating), ?-globin (late replicating), and C3 (repetitive heterochromatic sequence) ({Dimitrova, 1999 #166}{Li, 2001 #189} and unpublished) in CHO cells and extended to mouse C127 cells {Wu, 2006 #840}. In particular, by studying mouse late replicating chromocenters, we were able to characterize the role of many heterochromatin related factors in replication timing determination {Wu, 2005 #1306}. It was found that both the H3K9Me3 modification and the high affinity binding of HP1 to pericentric heterochromatin are in place prior to the TDP, implying that an assembly step downstream of the HP1-H3K9Me3 interaction is necessary to delay chromocenter replication. The late replication at chromocenter is generally preserved when a large pool of HP1 protein is disposed by Me3K9H3 peptide competition At cytological resolution, cells lacking Me3K9H3 by knocking out Suv39 h1/h2 still replicate chromocenters in the second half of S-phase, with slight advance relative to wild-type and Suv39-rescued cells {Wu 2006}. In addition, we found that this TDP happens before the origin decision point, suggesting that replication timing and replication origin specification can be uncoupled and replication timing builds on something more than origins per se.

Interestingly, in budding yeast, the late-replication of telomeric origins is also established during early G1 {Raghuraman, 1997 #356, Heun,}, indicating that a similar replication timing decision point exists in yeast as well. This suggests that an early G1-phase establishment of replication timing may be evolutionarily conserved. However difficulties studying early G1-phase events in yeast have hindered progress in this system.

3.2.2 A working model: coincidence of replication timing and subnuclear spatial organization at TDP

In all of the above studies, it was found that the establishment of replication timing at TDP correlates with a transient, but preferential spatial localization at the chromatin domains (Dimitrova and Gilbert, 1999b; Li et al., 2001). {Gilbert, 2001 #176;Heun, 2001 #519;Harmon, 2005 #75}. It was demonstrated by living cell imaging that the chromatin moves vigorously during early G1-phase, presumably facilitating to create sub-nuclear domains {Thomson, 2004 #1052;Chubb, 2002 #726; Walter, 2003 #2782}. Recently, an inducible targeting of a genetic locus to the nuclear lamina was shown to require a passage through mitosis and took place during the transition from late telophase to early G1-phase {Kumaran, 2008 #2779}, also same time window when replication timing is determined.

Based on these observations, a model for the establishment of replication timing in mammalian cells has been proposed {Gilbert, 2001 #176;Dimitrova, 1999 #166} (see Fig.1). This model suggests that repositioning of sequences in early G1-phase creates sub-nuclear microdomains that establish thresholds for the initiation of replication. These sub-nuclear microdomains recruit or create high local concentration of chromatin proteins /modifications (replication timing determinants) to establish thresholds for replication initiation during S-phase. For example, the heterochromatic microdomain, similar to yeast telomeric chromatin {Stevenson, 1999 #357}, may have a higher threshold, limiting the access of replication initiators and thereby delaying the replication time of associated origins in S-phase.

However, there is also evidence that the maintenance of such repositioning and anchorage at the TDP may not be required once cells progress through the early G1-phase past the TDP, both in budding yeast {Heun, 2001 #1852} and in mammals {Bridger, 2000 #1768}, indicating that once the replication timing is determined, the spatial organization is not necessary for the maintenance of such a temporal program. Therefore, spatial organization could somehow cause replication timing determinants be marked on different regions which stay on duty until the time of replication (Figure 4). This manuscript sets up to look for the potential replication timing determinants and made some interesting findings.

C. Other Related Background

1. Pericentric heterochromatin

Heitz first coined the concepts of euchromatin and heterochromatin on the basis of microscopic observations {Heitz 1928}. Experiments performed in the 80 years since then have largely upheld his view of heterochromatin as genetically inactive. While euchromatin condenses at mitosis and de-condenses during interphase, heterochromatin remains compact state throughout the cell cycle {Maison, 2004 #94}. Heterochromatin is distinguished from euchromatin in DNA sequence, histone modification and replication timing {Hennig, 1999 #980}. In general, heterochromatin is composed of "non-coding", repetitive DNA sequences, which are usually heavily methylated on cytosine of the CpG dinucleotides. In addition, heterochromatin is generally characterized with histone hypoacetylation, histone hypermethylation and late-replication. Such densely packed heterochromatin is usually late replicating and displays a low recombination rate. For a long time, heterochromatin has been thought to be "junk" of the genome. This view has been greatly challenged with ever-increasing understanding on the functional and structural character of heterochromatin. Rather than simply a structural component in the nucleus, heterochromatin has already demonstrated its pivotal roles in transcriptional regulation, cellular differentiation and animal development {Arney, 2004 #1041}. Among several better-studied heterochromatic regions is pericentric heterochromatin and the studies conducted on pericentric heterochromatin have contributed a lot to our current knowledge on heterochromatin.

In mouse, centromeres are composed of two types of repetitive DNA sequences: major (") satellite repeats (234bp unit) and minor satellite repeats (120bp unit) {Sullivan, 2002 #894;Guenatri, 2004 #517;Garagna, 2001 #891}. Major satellite sequences are located pericentric, while minor satellite sequences form the centromere core. They all form heterochromatin structure. Since these regions occupy up to 10% of mouse genome and they are one of the most extensively studied constitutive heterochromatin regions. In interphase, centromeres of different chromosomes associate in clusters to form chromocenters, which could be easily detected cytologically as condensed foci. In chromocenters, pericentric major satellite repeats form the large central core, and the corresponding minor satellites are located in the surrounding domain as several separate entities {Maison, 2004 #94;Kalitsis, 1997 #855;Lehnertz, 2003 #27}. In addition to the typical "epigenetic marks" of constitutive heterochromatin, including DNA hypermethylation, histone hypoacetylation and histone methylation (especially trimethylation at lysine 9 of histone H3), pericentric heterochromatin harbors its own hallmark, non-histone protein - HP1" (heterochromatin-associated protein 1), RNA and the enzymes that catalyze the formation of histone/DNA modification {Maison, 2004 #94}. "H3 K9 methylation-HP1-Suv39h" based network is central to the formation and maintenance of pericentric heterochromatin {Zhang, 2001 #935}. Histone methyltransferase Suv39h catalyzes the addition of methyl groups to lysine 9 of histone H3. Histone H3 K9 methylation provides a platform for HP1 to bind. The recruited HP1 is able to recruit more Suv39h for the propagation of heterochromatin {Bannister, 2001 #943;Lachner, 2001 #24}. These three elements form a "self-enforcing" loop, which ensures the propagation of heterochromatin once an initiating site has been established {Bannister, 2001 #943;Maison, 2004 #94}.

"H3 K9 methylation-HP1-Suv39h" is also implicated in the maintenance of pericentric heterochromatin by directly or indirectly recruiting other enzymes that catalyze histone/DNA modification {Craig, 2005 #959}. For instance, H4 K20 trimethylation and corresponding enzyme Su(var)4-20h1/2 localize to pericentric region and this localization is dependent on normal function of Suv39h genes {Schotta, 2004 #50;Kourmouli, 2004 #996;Kourmouli, 2004 #996}. Histone deacetylase complexes are also recruited to pericentric region through the interaction with Suv39h1/HP1 {Vaute, 2002 #1000;Xin, 2004 #146;David, 2003 #986}. Moreover, several enzymes (and/or associated factors) responsible for DNA methylation, such as DNMT1/DMAP1 and Lash1 physically interact with HP1/Suv39h {Xin, 2004 #146;Yan, 2003 #1011}. Except Su(var) 4-201/2 which is consistently present {Kourmouli, 2004 #996}, most of these enzymes are enriched at pericentric regions in late S-phase when pericentric heterochromatin is replicated {Xin, 2004 #146;Yan, 2003 #1011}. Hence, their role is mostly likely to function, together with histone methyltransferases, to pass on "epigenetic codes" from parental heterochromatin to newly packaged daughter chromatin, thereby ensuring the faithful duplication of chromatic states {McNairn, 2003 #191;Ehrenhofer-Murray, 2004 #1047}. Indeed, defects in these enzymes" function disrupt histone methylation patterns {Xin, 2004 #146;Espada, 2004 #1048}.

It should be noted that pericentric heterochromatin structure is dynamic. On one hand, its basic component, HP1, is highly mobile {Cheutin, 2003 #7;Festenstein, 2003 #760}. On the other hand, its organization undergoes some changes during early embryogenesis and differentiation {Probst, 2008 #1320}. For instance, male and female pronuclei have different pericentric heterochromatin structure during the first cell cycle after fertilization and chromocenters are not formed until the following cell cycles {Probst, 2007 #1322}. Elucidating the dynamics in pericentric heterochromatin (and other heterochromatic regions) formation and maintenance during development would reveal the important functions of heterochromatin.

We know from above Figure 3 that in S-phase, chromocenters mainly replicate at middle/late S-phase (patterns III and IV). Chromocenters were studied in detail by a previous graduate student in Gilbert lab {Wu, 2006}{Wu, 2005 #1306} as a late replicating model. Typical heterochromatin marks have been associated with pericentric regions: histones are generally hypoacetylated and specifically methylated on lysine 9 of histone H3 (H3K9, especially trimethylation H3K9Me3); heterochromatin protein 1 (HP1) is enriched in the region. Centromere chromatin contains some unique histone variants and histone modifications which are distinct from that of both euchromatin and pericentric heterochromatin {Guenatri, 2004 #676; Sullivan, 2004 #766}.

2. RNA interference and heterochromatin formation

How might pericentric regions initiate the formation of heterochromatin structure? the transcription from heterochromatin repeat sequences actually has been found to be part of a critical mechanism for heterochromatin assembly in some species. In S. Pombe, deletion of any gene that is involved in RNAi pathway, such as Dicer (dcr1), RdRp (rdp1), Argonaute (ago1), causes H3 Lys9 methylation and Swi6 (HP1 homologue in Pombe) to delocalize from centromeric repeats {Hall, 2002 #17;Volpe, 2002 #54;Grewal, 2004 #16}. Moroever, a RNAi effector complex called RITS (RNA-induced initiator of transcriptional gene silencing) has recently been identified {Verdel, 2004 #853;Motamedi, 2004 #948;Ekwall, 2004 #946;Noma, 2004 #831}, providing the first direct evidence for RNAi-mediated heterochromatin assembly. RITS contains a chromodomain protein, that binds centromeres and an Argonaute family protein, Ago1 that binds siRNA corresponding to centromeric repeats. RITS physically interacts with another RNAi complex, RDRC which contains RdRP {Motamedi, 2004 #948}. It's hypothesized that RITS complex further recruits Clr4 (histone H3 K9 methyltransferase, homologue of Suv39h in fission yeast) to establish methylation of H3K9 pattern for nucleation of SWI6 (homologue of HP1 in fission yeast) {Grewal, 2007 #34}. The evidence for RNAi mediated heterochromatin assembly is also found in other species. In Drosophila, defect in RNAi machinery abolishes heterochromatin-mediated silencing {Grewal, 2007 #831}. Therefore, the role of RNAi in heterochromatin assembly is likely to be evolutionarily conserved among diverse species. However whether or not such a RNAi-mediated silencing mechanism also exists in mammalian systems has not been well established, except a few cases {Morris, 2008 #897}.

Meanwhile, RNA is also implicated in the maintenance of pericentric heterochromatin structure. RNase A treatment causes the removal of HP1 and Me3K9H3 from pericentric region, suggesting that some "structural RNA" components are essential for maintaining the heterochromatin structure {Maison, 2002 #29;Muchardt, 2002 #103}. However, the identity of these RNAs is still a mystery and whether they are involved in initiating heterochromatin assembly also needs to be clarified in the future. On the other hand, although Dicer mutants in mice exhibit elevated levels of major satellite transcripts {Fukagawa, 2004 #28}{Kanellopoulou, 2005 #64}, this has no consequence on histone or DNA methylation in heterochromatin {Murchison, 2005 #349}. In addition, some of the RITS pathway component RNA dependent RNA polymerase (RdRp) is missing while Dicer is located in cytosol, raising questions on the conservation of RNAi-mediated pathway (chapter 2 and {Lu, 2008 #1137}). Therefore it remains unclear as to how RNAi-mediated mechanism of heterochromatin maintenance could function in mammals. Potential functions of these non-coding satellite RNAs in mice are implicated by us in chapter 2 of this dissertation.

Please be aware that the free essay that you were just reading was not written by us. This essay, and all of the others available to view on the website, were provided to us by students in exchange for services that we offer. This relationship helps our students to get an even better deal while also contributing to the biggest free essay resource in the UK!