Evidence for sequential and increasing activation of replication origins along replication timing gradients in the human genome

Genome-wide replication timing studies have suggested that mammalian chromosomes consist of megabase-scale domains of coordinated origin firing separated by large originless transition regions. Here, we report a quantitative genome-wide analysis of DNA replication kinetics in several human cell types that contradicts this view. DNA combing in HeLa cells sorted into four temporal compartments of S phase shows that replication origins are spaced at 40 kb intervals and fire as small clusters whose synchrony increases during S phase and that replication fork velocity (mean 0.7 kb/min, maximum 2.0 kb/min) remains constant and narrowly distributed through S phase. However, multi-scale analysis of a genome-wide replication timing profile shows a broad distribution of replication timing gradients with practically no regions larger than 100 kb replicating at less than 2 kb/min. Therefore, HeLa cells lack large regions of unidirectional fork progression. Temporal transition regions are replicated by sequential activation of origins at a rate that increases during S phase and replication timing gradients are set by the delay and the spacing between successive origin firings rather than by the velocity of single forks. Activation of internal origins in a specific temporal transition region is directly demonstrated by DNA combing of the IGH locus in HeLa cells. Analysis of published origin maps in HeLa cells and published replication timing and DNA combing data in several other cell types corroborate these findings, with the interesting exception of embryonic stem cells where regions of unidirectional fork progression seem more abundant. These results can be explained if origins fire independently of each other but under the control of long-range chromatin structure, or if replication forks progressing from early origins stimulate initiation in nearby unreplicated DNA. These findings shed a new light on the replication timing program of mammalian genomes and provide a general model for their replication kinetics.     

Regulation of the replication of the murine immunoglobulin heavy chain gene locus: evaluation of the role of the 3' regulatory region.

DNA replication in mammalian cells is a precisely controlled physical and temporal process, likely involving cis-acting elements that control the region(s) from which replication initiates. In B cells, previous studies showed replication timing to be early throughout the immunoglobulin heavy chain (Igh) locus. The implication from replication timing studies in the B-cell line MPC11 was that early replication of the Igh locus was regulated by sequences downstream of the C alpha gene. A potential candidate for these replication control sequences was the 3' regulatory region of the Igh locus. Our results demonstrate, however, that the Igh locus maintains early replication in a B-cell line in which the 3' regulatory region has been deleted from one allele, thus indicating that replication timing of the locus is independent of this region. In non-B cells (murine erythroleukemia cells [MEL]), previous studies of segments within the mouse Igh locus demonstrated that DNA replication likely initiated downstream of the Igh gene cluster. Here we use recently cloned DNA to demonstrate that segments located sequentially downstream of the Igh 3' regulatory region continue to replicate progressively earlier in S phase in MEL. Furthermore, analysis by two-dimensional gel electrophoresis indicates that replication forks proceed exclusively in the 3'-to-5' direction through the region 3' of the Igh locus. Extrapolation from these data predicts that initiation of DNA replication occurs in MEL at one or more sites within a 90-kb interval located between 40 and 130 kb downstream of the 3' regulatory region.     

Long-distance control of origin choice and replication timing in the human beta-globin locus are independent of the locus control region

DNA replication in the human beta-globin locus is subject to long-distance regulation. In murine and human erythroid cells, the human locus replicates in early S phase from a bidirectional origin located near the beta-globin gene. This Hispanic thalassemia deletion removes regulatory sequences located over 52 kb from the origin, resulting in replication of the locus from a different origin, a shift in replication timing to late S phase, adoption of a closed chromatin conformation, and silencing of globin gene expression in murine erythroid cells. The sequences deleted include nuclease-hypersensitive sites 2 to 5 (5'HS2-5) of the locus control region (LCR) plus an additional 27-kb upstream region. We tested a targeted deletion of 5'HS2-5 in the normal chromosomal context of the human beta-globin locus to determine the role of these elements in replication origin choice and replication timing. We demonstrate that the 5'HS2-5-deleted locus initiates replication at the appropriate origin and with normal timing in murine erythroid cells, and therefore we conclude that 5'HS2-5 in the classically defined LCR do not control replication in the human beta-globin locus. Recent studies also show that targeted deletion of 5'HS2-5 results in a locus that lacks globin gene expression yet retains an open chromatin conformation. Thus, the replication timing of the locus is closely correlated with nuclease sensitivity but not globin gene expression.     

Replication and subnuclear location dynamics of the immunoglobulin heavy-chain locus in B-lineage cells

The murine immunoglobulin heavy-chain (Igh) locus provides an important model for understanding the replication of tissue-specific gene loci in mammalian cells. We have observed two DNA replication programs with dramatically different temporal replication patterns for the Igh locus in B-lineage cells. In pro- and pre-B-cell lines and in ex vivo-expanded pro-B cells, the entire locus is replicated early in S phase. In three cell lines that exhibit the early-replication pattern, we found that replication forks progress in both directions through the constant-region genes, which is consistent with the activation of multiple initiation sites. In contrast, in plasma cell lines, replication of the Igh locus occurs through a triphasic pattern similar to that previously detected in MEL cells. Sequences downstream of the Igh-C alpha gene replicate early in S, while heavy-chain variable (Vh) gene sequences replicate late in S. An approximately 500-kb transition region connecting sequences that replicate early and late is replicated progressively later in S. The formation of the transition region in different cell lines is independent of the sequences encompassed. In B-cell lines that exhibit a triphasic-replication pattern, replication forks progress in one direction through the examined constant-region genes. Timing data and the direction of replication fork movement indicate that replication of the transition region occurs by a single replication fork, as previously described for MEL cells. Associated with the contrasting replication programs are differences in the subnuclear locations of Igh loci. When the entire locus is replicated early in S, the Igh locus is located away from the nuclear periphery, but when Vh gene sequences replicate late and there is a temporal-transition region, the entire Igh locus is located near the nuclear periphery.     

Regulation of replication at the R/G chromosomal band boundary and pericentromeric heterochromatin of mammalian cells

Mammalian chromosomes consist of multiple replicons; however, in contrast to yeast, the details of this replication process (origin firing, fork progression and termination) relative to specific chromosomal domains remain unclear. Using direct visualization of DNA fibers, here we show that the rate of replication fork movement typically decreases in the early-mid S phase when the replication fork proceeds through the R/G chromosomal band boundary and pericentromeric heterochromatin. To support this, fluorescence in situ hybridization (FISH)-based replication profiles at the human 1q31.1 (R-band)-32.1 (G-band) regions revealed that replication timing switched around at the putative R/G chromosomal band boundary predicted by marked changes in GC content at the sequence level. Thus, the slowdown of replication fork movement is thought to be the general property of the band boundaries separating the functionally different chromosomal domains. By simultaneous visualization of replication fork movement and pericentromeric heterochromatin sequences on DNA fibers, we observed that this region is duplicated by many replication forks, some of which proceed unidirectionally, that originate from clustered replication origins. We showed that histone hyperacetylation is tightly associated with changes in the replication timing of pericentromeric heterochromatin induced by 5-aza-2'-deoxycytidine treatment. These results suggest that, similar to the yeast system, histone modification is involved in controlling the timing of origin firing in mammals.     

Replication forks reverse at high frequency upon replication stress in Physarum polycephalum

The addition of hydroxyurea after the onset of S phase allows replication to start and permits the successive detecting of replication-dependent joint DNA molecules and chicken foot structures in the synchronous nuclei of Physarum polycephalum. We find evidence for a very high frequency of reversed replication forks upon replication stress. The formation of these reversed forks is dependent on the presence of joint DNA molecules, the impediment of the replication fork progression by hydroxyurea, and likely on the propensity of some replication origins to reinitiate replication to counteract the action of this compound. As hydroxyurea treatment enables us to successively detect the appearance of joint DNA molecules and then of reversed replication forks, we propose that chicken foot structures are formed both from the regression of hydroxyurea-frozen joint DNA molecules and from hydroxyurea-stalled replication forks. These experiments underscore the transient nature of replication fork regression, which becomes detectable due to the hydroxyurea-induced slowing down of replication fork progression.     

Transitions in replication timing in a 340 kb region of human chromosomal R-Band 1p36.1

DNA replication is initiated within a few chromosomal bands as normal human fibroblasts enter the S phase. In the present study, we determined the timing of replication of sequences along a 340 kb region in one of these bands, 1p36.13, an R band on chromosome 1. Within this region, we identified a segment of DNA (approximately 140 kb) that is replicated in the first hour of the S phase and is flanked by segments replicated 1-2 h later. Using a quantitative PCR-based assay to measure sequence abundance in size-fractionated (900-1,700 nt) nascent DNA, we mapped two functional origins of replication separated by 54 kb and firing 1 h apart. One origin was found to be functional during the first hour of S and was located within a CpG island associated with a predicted gene of unknown function (Genscan NT_004610.2). The second origin was activated in the second hour of S and was mapped to a CpG island near the promoter of the aldehyde dehydrogenase 4A1 (ALDH4A1) gene. At the opposite end of the early replicating segment, a more gradual change in replication timing was observed within the span of approximately 100 kb. These data suggest that DNA replication in adjacent segments of band 1p36.13 is organized differently, perhaps in terms of replicon number and length, or rate of fork progression. In the transition areas that mark the boundaries between different temporal domains, the replication forks initiated in the early replicated region are likely to pause or delay progression before replication of the 340 kb contig is completed.     

Organization of DNA replication in Physarum polycephalum. Attachment of origins of replicons and replication forks to the nuclear matrix.

We have investigated the attachment of the DNA to the nuclear matrix during the division cycle of the plasmodial slime mold Physarum polycephalum. The DNA of plasmodia was pulse labelled at different times during the S phase and the label distribution was studied by graded DNase digestion of the matrix-DNA complexes prepared from nuclei isolated by extraction with 2 M NaCl. Pulse labelled DNA was preferentially recovered from the matrix bound residual DNA at any time of the S phase. Label incorporated at the onset of the S phase remained preferentially associated with the matrix during the G2 phase and the subsequent S phase. The occurrence of the pulse label in the matrix associated DNA regions was transiently elevated at the onset of the subsequent S phase. Label incorporated at the end of the S phase was located at DNA regions which, in the G2 phase, were preferentially released from the matrix by DNase treatment. From the results and previously reported data on the distribution of attachment sites it can be concluded that origins of replicons or DNA sites very close to them are attached to the matrix during the entire nuclear cycle. The data further indicate that initiations of DNA replication occur at the same origins in successive S phases. Replicating DNA is bound to the matrix, in addition, by the replication fork or a region close to it. This binding is loosened after completion of the replication.     

Localization of a bidirectional DNA replication origin in the native locus and in episomally amplified murine adenosine deaminase loci.

Gene amplification is frequently mediated by the initial production of acentric, autonomously replicating extrachromosomal elements. The 4,000 extrachromosomal copies of the mouse adenosine deaminase (ADA) amplicon in B-1/50 cells initiate their replication remarkably synchronously in early S phase and at approximately the same time as the single-copy chromosomal locus from which they were derived. The abundance of ADA sequences and favorable replication timing characteristics in this system led us to determine whether DNA replication initiates in ADA episomes within a preferred region and whether this region is the same as that used at the corresponding chromosomal locus prior to amplification. This study reports the detection and localization of a discrete set of DNA fragments in the ADA amplicon which label soon after release of synchronized B-1/50 cells into S phase. A switch in template strand complementarity of Okazaki fragments, indicative of the initiation of bidirectional DNA replication, was found to lie within the same region. This putative replication origin is located approximately 28.5 kbp upstream of the 5' end of the ADA gene. The same region initiated DNA replication in the single-copy ADA locus of the parental cells. These analyses provide the first evidence that the replication of episomal intermediates involved in gene amplification initiates within a preferred region and that the same region is used to initiate DNA synthesis within the native locus.     

Two alleles of a developmentally regulated alpha-tubulin locus in Physarum polycephalum replicate on different schedules.

The replication timing of a pair of natural alleles was compared at two alpha-tubulin loci of the Physarum plasmodium. Taking advantage of the naturally synchronous cell cycle of nuclei within the syncytial plasmodium, we analyzed the replication schedule of specific DNA fragments to a resolution of 10-min intervals within a 3-h S phase. At this level of resolution, differences in replication timing between polymorphic alleles at the same locus can be detected in a heterozygote. Specifically, the 3' region of the altA1 allele completes replication at between 20 and 40 min of S phase. The same region of the altA2 allele completes replication at between 40 and 80 min of S phase. In contrast, both alleles at the altB locus replicate concurrently within the first 10 to 15 min of S phase. Previous studies showed that both altA and altB are expressed in the plasmodium, their message levels peaking at mitosis, just minutes before the onset of S phase. However, altB message is detected at substantially higher levels than altA message on Northern (RNA) blots. The temporal windows over which the altA alleles each replicate are very broad in comparison with the levels of mitotic synchrony and altB replication synchrony in a single plasmodium. The allele-specific replication schedule of the altA locus demonstrates that the temporal organization of replicons is not strictly conserved between homologous chromosomes.     

Checkpoint responses to replication fork barriers

The fidelity of DNA replication is of paramount importance to the maintenance of genome integrity. When an active replication fork is perturbed, multiple cellular pathways are recruited to stabilize the replication apparatus and to help to bypass or correct the causative problem. However, if the problem is not corrected, the fork may collapse, exposing free DNA ends to potentially inappropriate processing. In prokaryotes, replication fork collapse promotes the activity of recombination proteins to restore a replication fork. Recent work has demonstrated that recombination is also intimately linked to replication in eukaryotic cells, and that recombination proteins are recruited to collapsed, but not stalled, replication forks. In this review we discuss the different types of potential replication fork barriers (RFB) and how these distinct RFBs can result in different DNA structures at the stalled replication fork. The DNA structure checkpoints which act within S phase respond to different RFBs in different ways and we thus discuss the processes that are controlled by the DNA replication checkpoints, paying particular attention to the function of the intra-S phase checkpoint that stabilises the stalled fork.     

rDNA enhancer affects replication initiation and mitotic recombination: Fob1 mediates nucleolytic processing independently of replication

To investigate the influence of the ribosomal DNA enhancer on initiation of replication and recombination at the ribosomal array, we used yeast S. cerevisiae strains with adjacent, tagged rRNA genes. We found that the enhancer is an absolute requirement for replication fork barrier function, while it only modulates initiation of replication. Moreover, the formation of monomeric extrachromosomal ribosomal circles depends on this element. Our data indicate that DNA double-strand breaks occur at specific sites in the parental leading arm of replication forks stalled at the replication fork barrier. Additionally, nicks upstream of the replication fork barrier were visualized by nucleotide-resolution mapping. They coincide with essential sequences of the mitotic hyperrecombination site HOT1, which previously has been determined at ectopic sites. Interestingly, these nicks are strictly dependent on the replication fork blocking-protein (Fob1), but are replication independent, suggesting that intrachromosomal ribosomal DNA recombination may occur outside of S phase.     

DNA replication during a serum-induced S phase in primate CV-1 cells

We have investigated components of DNA replication in a serum-induced S phase of primate CV-1 cells. Using DNA fiber autoradiography, we found a relative decrease in the frequency of initiation events in mid-S compared with early and late S phase. The other components of DNA replication measured by autoradiography—synchrony of initiation events, size of replication units, incidence of bidirectional replication, and the rate of replication fork movement—remained constant through S phase. When fork movement was measured by density gradient analysis of BUdR- and [³H]-thymidine-substituted DNA, it was also found to remain constant. These results show that most components of DNA replication are invariable through a serum-induced S phase. The changes in initiation frequency support the view that it may be critical in the regulation of ongoing replication.     

Oligomerization of DNA replication regulatory protein RADX is essential to maintain replication fork stability

Genome integrity requires complete and accurate DNA replication once per cell division cycle. Replication stress poses obstacles to this process that must be overcome to prevent replication fork collapse. An important regulator of replication fork stability is the RAD51 protein, which promotes replication fork reversal and protects nascent DNA strands from nuclease-mediated degradation. Many regulatory proteins control these RAD51 activities, including RADX, which binds both ssDNA and RAD51 at replication forks to ensure that fork reversal is confined to stalled forks. Many ssDNA-binding proteins function as hetero- or homo-oligomers. In this study, we addressed whether this is also the case for RADX. Using biochemical and genetic approaches, we found that RADX acts as a homo-oligomer to control replication fork stability. RADX oligomerizes using at least two different interaction surfaces, including one mapped to a C-terminal region. We demonstrate that mutations in this region prevent oligomerization and prevent RADX function in cells, and that addition of a heterologous dimerization domain to the oligomerization mutants restored their ability to regulate replication. Taken together, our results demonstrate that like many ssDNA-binding proteins, oligomerization is essential for RADX-mediated regulation of genome stability.     

Visualization of altered replication dynamics after DNA damage in human cells

Eukaryotic cells respond to DNA damage within the S phase by activating an intra-S checkpoint: a response that includes reducing the rate of DNA synthesis. In yeast cells this can occur via checkpoint-dependent inhibition of origin firing and stabilization of ongoing forks, together with a checkpoint-independent slowing of fork movement. In higher eukaryotes, however, the mechanism by which DNA synthesis is reduced is less clear. We have developed strategies based on DNA fiber labeling that allow the quantitative assessment of rates of replication fork movement, origin firing, and fork stalling throughout the genome by examining large numbers of individually labeled replication forks. We show that exposing S phase cells to ionizing radiation induces a transient block to origin firing but does not affect fork rate or fork stalling. Alkylation damage by methyl methane sulfonate causes a slowing of fork movement and a high rate of fork stalling, in addition to inducing a block to new origin firing. Nucleotide depletion by hydroxyurea also reduces replication fork rate and increases stalling; moreover, in contrast to a recent report, we show that hydroxyurea induces a strong block to new origin firing. The DNA fiber labeling strategy provides a powerful new approach to analyze the dynamics of DNA replication in a perturbed S phase.     

Mechanisms of polar arrest of a replication fork

A DNA replication terminator sequence blocks an approaching replication fork when the moving replisome approaches from just one direction. The mechanism underlying polar arrest has been debated for years, but recent work has helped to reveal how a replication fork is blocked in Escherichia coli . Early work suggested that asymmetric interaction between terminator protein and terminator DNA contributes to polar fork arrest. A later study demonstrated that if the terminator DNA is partially unwound, the resulting melted DNA could bind tightly to the terminator protein, suggesting a mechanism for polar arrest that involves a locked complex. However, recent evidence suggests that the terminator protein–DNA contacts are not sufficient for polar arrest in vivo . Furthermore, polar arrest of a replication fork still occurs in the absence of a locked complex between the terminator protein and DNA. In E. coli and Bacillus subtilis , the bound terminator protein makes protein–protein contacts with the replication fork helicase, and these contacts are critical in blocking progression of the advancing fork. Thus, we propose that interactions between the replication fork helicase and terminator protein are the primary mechanism for polar fork arrest in bacteria, and that this primary mechanism is modulated by asymmetric contacts between the terminator protein and its cognate DNA sequence. In yeast, terminator sequences are present in rDNA non-transcribed spacers and a region immediately preceding the mating type switch locus Mat1, and the mechanism of polar arrest at these regions is beginning to be elucidated.