Mammalian DNA Replication Timing

Abstract

Immediately following the discovery of the structure of DNA and the semi-conservative replication of the parental DNA sequence into two new DNA strands, it became apparent that DNA replication is organized in a temporal and spatial fashion during the S phase of the cell cycle, correlated with the large-scale organization of chromatin in the nucleus. After many decades of limited progress, technological advances in genomics, genome engineering, and imaging have finally positioned the field to tackle mechanisms underpinning the temporal and spatial regulation of DNA replication and the causal relationships between DNA replication and other features of large-scale chromosome structure and function. In this review, we discuss these major recent discoveries as well as expectations for the coming decade.

Figure 1.

Measuring and interpreting replication timing (RT) profiles. (A) DNA RT profiles of chromosome 1 in mouse embryonic stem cells (mESCs). Profiles were generated using three different methods: (Top) E/L Repli-seq displays the population-averaged log₂ ratio of DNA synthesized in early S phase to DNA synthesized in late S phase. One example each of a constant timing region ([CTR], red) and a timing transition region ([TTR], blue) are shown. (Middle) High-resolution Repli-seq displays a heat map (reads per million per 50 kb bin for each of 16 temporal intervals) of the sequences replicated in each of many S phase temporal fractions, in this case 16 fractions. (Data in A is adapted from Zhao et al. 2020 and available at data.4dnucleome.org). (Bottom) Single-cell Repli-seq displays the copy number (red = 2; gray = 1) of sequences per chromosome in each of many (in this case 71) individual cells, ranked from top to bottom in order of their total cellular DNA content. (Bottom panel adapted from Dileep and Gilbert 2018 with permission from the authors who are also the copyright holders.) (B) Replication domains and initiation zones. (Top) Schematic of how initiation events appear as they can be observed on single DNA fibers. There is usually one, occasionally more, initiation site(s) per locally earliest replicating peak. Different cells initiate at different sites throughout initiation zones that elongate(s) with time to give a population-averaged RT profile. (Middle) An ∼0.5-Mb replication domain (red lines) on mouse chromosome 16 defined by E/L Repli-seq as a region of coordinated RT change from early to late during differentiation of mESCs (ESC, black) to neural precursor cells ([NPCs], red). Boundaries of the domain are determined by computing the point of significant slope change. (Bottom) High-resolution Repli-seq of this same region in mESCs reveals a single initiation zone (blue lines) within which this domain initiates. Because E/L Repli-seq sums the reds of the first and second half of S phase, the log₂ ratio peak is relatively flat and extends until the point at which cells in the second half of S phase begin to replicate the region.

Figure 2.

Organization of replicons. (A) Chinese hamster ovary cells were labeled for 10 min with BrdU, fixed and stained with anti-BrdU antibodies to reveal the spatial patterns of DNA synthesis in early, middle, and late S phase. (Images in A courtesy of J. Lu.) Scale bar, 5 μM. (B) Normal rat kidney cells were labeled live with ATTO 633-dUTP and then chased for several generations. Shown is a comparison of a high magnification (scale bar, 500 nm) confocal image of a cluster of replication foci to the super-resolution (2D-STORM) image of the same replication foci, demonstrating that each replication focus consists of a cluster of labeled sites. (Panel B reprinted from Xiang et al. 2018 courtesy W. Xiang © 2018 in conjunction with a Creative Commons License [Attribution 4.0 International] www.ncbi.nlm.nih.gov/pmc/articles/PMC5987722.) (C) The organization of replicons in these foci remains unknown. Sister forks could be replicated independently (left), by a common replisome or clustered replisomes (middle) or multiple replicons, and their sister forks could be replicated by a common replisome or cluster of replisomes (right). (Panel C courtesy of C. Marchal.)

Figure 3.

Nuclear organization and chromatin architecture. (A) Replication timing (RT), displayed both in high-resolution Repli-seq and E/L Repli-seq, highly correlates with Hi-C-derived first principal component (PC1), the distance of the locus from the nuclear speckle protein SON and the nuclear lamina protein LaminB measured via TSA-seq, and with the contact frequency of chromatin with the nucleolus measured via DamID, an alternative to ChIP, which uses Escherichia coli adenine methyltransferase (Dam) fused to a protein of interest to methylate adenines in DNA sequences that come to close proximity to the protein of interest. Information on the distance and contact frequency of chromatin loci from subnuclear landmarks is combined with Hi-C and histone modification data to stratify the genome into 10 SPIN states. A SPIN state is composed of genomic loci sharing unique combinations of histone modifications, compartmentalization, and association with nuclear landmarks. The data displayed originate from a 15 Mb window in chromosome 11 of human bone marrow lymphoblast K562 cells, mapped using the Nucleome Browser (vis.nucleome.org). (B) Spearman correlation of five subcompartments in Rao et al. (2014) to the nuclear lamina, the nucleolus, and early RT. (Data in B adapted from Rao et al. 2014.) (C) Spearman correlation of the 10 SPIN states in Wang et al. (2020b) to six fractions of the S phase and G2. (Data in C plotted using data adapted from Wang et al. 2020b.)

Figure 4.

Model of early replication control element (ERCE) interactions in the Dppa2/4 domain. (A) ERCEs interact to influence the 3D architecture of topologically associated domains (TADs), the interaction of TADs with other domains (compartmentalization), transcription, and early replication timing (RT) (Sima et al. 2019). ERCEs resemble enhancers, being occupied by H3K27ac, the p300 acetyltransferase, and the major pluripotency transcription factors (TFs) Oct4, Sox2, and Nanog (OSN). In this working model, lineage-specific TFs such as OSN promote histone acetylation, recruiting acetylation readers Brd2/4 and promoting chromatin interactions (Kim 2009; Wu et al. 2015, 2018) to form a 3D hub highly enriched for Brd2/4. The replication initiation protein Treslin interacts with Brd2/4, linking OSN and histone acetylation to initiation of replication (Sansam et al. 2018). Meanwhile, Rif1 coats late replicating regions to prevent them from replicating early during S phase. (B) Deletion of all three ERCEs in this domain causes the domain to switch from early to late replicating, eliminates all intradomain transcription, changes compartments (A to B), and shifts the domain toward the nuclear lamina (NL). (C) Rif1 knockout allows for late replicating regions near the NL to become accessible to replication factors, thus diverting replication resources toward these late replicating regions, resulting in highly stochastic RT, redistribution of chromatin marks, and alterations in chromatin compartments (Panels A and B adapted from Sima et al. 2019 courtesy of Creative Commons Public License.)

Figure 5.

Asynchronous replication and autosomal RNAs (ASARs) are necessary to ensure that whole chromosomes are replicated in a timely fashion. (A) Delayed mitotic condensation and spontaneous damage. Mitotic cells containing an uncondensed i(3q), from human rhabdomyosarcoma cells RH30. The i(3q) was identified by FISH with a centromeric probe (red; *). DNA was stained with DAPI (blue), and arrows mark sites of spontaneous damage. (B) Monoallelic expression of multiple ASARs from a hypothetical chromosome pair. The expressed (Active; colored clocks) alleles of different ASARs result in “clouds” of RNA that are retained within the chromosome territories from which they are transcribed. ASAR RNA clouds are speculated to regulate early replication control element (ERCE) activity. The nonexpressed ASARs (Silent; white clocks) are inactive. (Panels A and B courtesy of M. Thayer.)