High‐resolution analysis of DNA synthesis start sites and nucleosome architecture at efficient mammalian replication origins

DNA replication origins are poorly characterized genomic regions that are essential to recruit and position the initiation complex to start DNA synthesis. Despite the lack of specific replicator sequences, initiation of replication does not occur at random sites in the mammalian genome. This has lead to the view that DNA accessibility could be a major determinant of mammalian origins. Here, we performed a high‐resolution analysis of nucleosome architecture and initiation sites along several origins of different genomic location and firing efficiencies. We found that mammalian origins are highly variable in nucleosome conformation and initiation patterns. Strikingly, initiation sites at efficient CpG island‐associated origins always occur at positions of high‐nucleosome occupancy. Origin recognition complex (ORC) binding sites, however, occur at adjacent but distinct positions marked by labile nucleosomes. We also found that initiation profiles mirror nucleosome architecture, both at endogenous origins and at a transgene in a heterologous system. Our studies provide a unique insight into the relationship between chromatin structure and initiation sites in the mammalian genome that has direct implications for how the replication programme can be accommodated to diverse epigenetic scenarios.     

DNA synthesis generally initiates outside the simian virus 40 core origin in vitro.

The nucleotide positions at which DNA synthesis initiates in vitro, in the vicinity of the simian virus 40 origin, have been determined. Start sites for DNA synthesis are greatly suppressed over the simian virus 40 core origin. Relatively weak start sites are detected over the 21-bp repeats and T-antigen-binding site I; distal to these regions, stronger start sites are detected. Thus, studies using a model system for eukaryotic DNA replication indicate that DNA synthesis events initiate, in general, outside the core origin.     

Topoisomerase I sites cluster asymmetrically at the ends of the simian virus 40 core origin of replication.

In vivo, topoisomerase I cleavage sites are located predominantly on the strands of simian virus 40 DNA that are the templates for discontinuous synthesis (S.E. Porter and J.J. Champoux, Mol. Cell. Biol. 9:541-550, 1989). This arrangement of sites suggests that topoisomerase I may associate with replication complexes in unique functional orientations at replication forks. We have mapped topoisomerase I cleavage sites in the simian virus 40 origin of replication in vitro under conditions suitable for DNA replication. Numerous sites cluster in the inverted repeat and AT-rich domains at the ends of the core origin and are arranged on the same strands that are cut most frequently in vivo. We propose that cleavage at these sites would allow bidirectional extension of the replication bubble induced by T antigen within the core origin of replication early in the initiation of DNA synthesis. A mutational analysis of the topoisomerase I sites confirms the importance of positions -4 to -1 and +1 in the consensus sequence 5'-A/T-A/G-A/T-T-break-G/A-3'. Surprisingly, more distant nucleotide positions also influence topoisomerase I sites in the inverted repeat and AT-rich domains of the core origin. The effects of distant sequences could be mediated by direct interactions with topoisomerase I or by the conformation of DNA in the core origin.     

Fine mapping of secondary structures of fd phage DNA in the region of the replication origin.

A synthetic heptaribonucleotide, GACCCCC, which is complementary to a unique site on fd bacteriophage DNA, primes DNA synthesis of fd by T4 bacteriophage DNA polymerase. The rate of the GACCCCC-primed DNA synthesis was not uniform as reflected by the appearance of discrete DNA fragments as replication intermediates on an alkaline agarose gel. After 10 minutes of synthesis a significant fraction of the DNA product ran as a single band with a length of about 1960 nucleotides. We have isolated this DNA fragment, hybridized back to unlabeled fd DNA template, and mapped the Taq I restriction fragments by urea polyacrylamide gel electrophoresis. This fine mapping procedure has located two major pause sites at fd nucleotide positions 5575 and 5674. These sites reside in the stem of two very stable hairpin helices near the origin of DNA replication of fd. Models for the functional roles of these two hairpin helices are presented.     

The T-antigen-binding domain of the simian virus 40 core origin of replication.

The simian virus 40 origin of replication contains a 27-base-pair palindrome with the sequence 5'-CA-GAGGC-C-GAGGC-G-GCCTC-G-GCCTC-TG-3'. The four 5'-GAGGC-3'/5'-GCCTC-3' pentanucleotides are known contact sites for simian virus 40 T-antigen binding in vitro. We used oligonucleotide-directed cassette mutagenesis to identify features of this palindrome that are important for the initiation of DNA replication in vivo. Each base pair of a pentanucleotide is crucial for DNA replication. In contrast, sequences adjacent to pentanucleotides have little or no effect on replication. Thus, the pentanucleotide is the basic functional unit, not only for T-antigen binding but also for DNA replication. All four pentanucleotides are indispensable in the initiation process. The spacing of pentanucleotides is crucial because duplication of the single base pair between binding sites has a far greater effect on replication than does substitution of the same base pair. Inversion of any pentanucleotide blocks DNA synthesis. Thus, the pentanucleotide is not a functionally symmetrical unit. We propose that each pentanucleotide positions a monomer of T antigen at the proper distance, rotation, and orientation relative to other T-antigen monomers and to other origin domains and that such positioning leads to subsequent events in replication.     

A case for sliding SeqA tracts at anchored replication forks during Escherichia coli chromosome replication and segregation

SeqA is an Escherichia coli DNA-binding protein that acts at replication origins and controls DNA replication. However, binding is not exclusive to origins. Many fragments containing two or more hemi-methylated GATC sequences bind efficiently. Binding was optimal when two such sequences were closely apposed or up to 31 bases apart on the same face of the DNA helix. Binding studies suggest that neighboring bound proteins contact each other to form a complex with the intervening DNA looped out. There are many potential binding sites distributed around the E.coli chromosome. As replication produces a transient wave of hemi-methylation, tracts of SeqA binding are likely to associate with each fork as replication progresses. The number and positions of green fluorescent protein-SeqA foci seen in living cells suggest that they correspond to these tracts, and that the forks are tethered to planes of cell division. SeqA may help to tether the forks or to organize newly replicated DNA into a structure that aids DNA to segregate away from the replication machinery.     

Initiation of Escherichia coli minichromosome replication at oriC and at protein n'' recognition sites. Two modes for initiating DNA synthesis in vitro.

The start sites for leading and lagging DNA strands were determined in vitro with minichromosomes as templates. Fragments from replication intermediates were analyzed by hybridization to single-stranded probes. Leading strand synthesis in the counterclockwise direction was found to originate in or close to (position 248 to -44) the minimal origin. Complementary lagging strand synthesis started several positions to the left outside of oriC. The results suggest in addition a concerted synthesis of leading and lagging strands following the dnaA directed assembly of initiation proteins at double-stranded oricC DNA (pre-replisome). In addition, DNA synthesis could initiate at protein n' recognition sequences located within and clockwise to the asnA gene. Initiation at n' sites was dependent on protein i activity, whereas leading and lagging strand initiation in the oriC region was not affected by protein i. Our results argue against an involvement of the phi X174-type primosome in the initiation of discontinuous DNA synthesis at oriC. An alternative function is suggested.     

Mapping replication units in animal cells

A general approach for assaying the in vivo direction of replication for any DNA segment has been developed. This technique allows the scanning of genomic regions to detect bidirectional tail-to-tail replication, indicating the presence of a functional origin. By this criterion we identified the approximate positions of two origin sites downstream of the Chinese hamster DHFR gene. Further mapping revealed areas of head-to-head replication, signifying locations of replication termination and thus defining the landmarks of a complete animal cell replicon. Genetic proof for the existence of the DHFR origin was obtained by showing that this region serves as a bidirectional DNA synthesis initiation point following its integration into other sites in the genome by transfection. To show the general applicability of this methodology, we studied the APRT domain. Replication mapping together with the use of deletion mutants allowed the identification of an origin at a far-upstream locus.     

DNA synthesis is initiated at two positions within the origin of replication of plasmid R1162.

DNA synthesis of broad host-range plasmid R1162 is initiated from two positions, flanking a large (40 bp stem, 40 bp loop) inverted repeat. Each start-point is located within a highly conserved, but oppositely oriented, 10 base-pair sequence. Synthesis from the two positions converges within the intervening inverted repeat. An analysis of deletions suggests that both start positions must be present for synthesis. A model describing possible early events in replication of plasmid R1162 is presented.     

Sequence-specific pausing during in vitro DNA replication on double-stranded DNA templates

Sequence-specific pausing occurs during DNA synthesis catalyzed by the bacteriophage T4 DNA polymerase holoenzyme in the presence of the T4 helix destabilizing protein (gene 32 protein). Two of the six strongest pause sites on a double-stranded bacteriophage fd DNA template are in regions where hairpin helices are predicted to form when the DNA is single stranded. However, the other pause sites are in regions that are not obviously involved in secondary structure. The positions of the DNA chain ends produced at one pause site of each type were determined to within +/- 2 nucleotides. At this resolution, a clustering of sites is observed, suggesting that the polymerase holoenzyme may become destabilized when moving along selected regions of the DNA and then pause at one or more of several closely spaced positions. The addition of the T4 gene 41 protein (a DNA helicase that forms part of the T4 primosome) to the above replication system greatly increases the rate of fork movement and eliminates detectable pausing. In contrast, the addition of the T4 dda protein (a second DNA helicase that increases the rate of fork movement to a similar extent) has no affect on replication fork pausing. This difference could either be due to specific protein-protein interactions formed between the polymerase holoenzyme and the 41 protein or to the highly processive movement of the 41 protein along the displaced DNA strand.     

Mechanisms and regulation of DNA replication initiation in eukaryotes

Cellular DNA replication is initiated through the action of multiprotein complexes that recognize replication start sites in the chromosome (termed origins) and facilitate duplex DNA melting within these regions. In a typical cell cycle, initiation occurs only once per origin and each round of replication is tightly coupled to cell division. To avoid aberrant origin firing and re-replication, eukaryotes tightly regulate two events in the initiation process: loading of the replicative helicase, MCM2-7, onto chromatin by the origin recognition complex (ORC), and subsequent activation of the helicase by its incorporation into a complex known as the CMG. Recent work has begun to reveal the details of an orchestrated and sequential exchange of initiation factors on DNA that give rise to a replication-competent complex, the replisome. Here, we review the molecular mechanisms that underpin eukaryotic DNA replication initiation – from selecting replication start sites to replicative helicase loading and activation – and describe how these events are often distinctly regulated across different eukaryotic model organisms.     

New insights into replication origin characteristics in metazoans

We recently reported the identification and characterization of DNA replication origins (Oris) in metazoan cell lines. Here, we describe additional bioinformatic analyses showing that the previously identified GC-rich sequence elements form origin G-rich repeated elements (OGREs) that are present in 67% to 90% of the DNA replication origins from Drosophila to human cells, respectively. Our analyses also show that initiation of DNA synthesis takes place precisely at 160 bp (Drosophila) and 280 bp (mouse) from the OGRE. We also found that in most CpG islands, an OGRE is positioned in opposite orientation on each of the two DNA strands and detected two sites of initiation of DNA synthesis upstream or downstream of each OGRE. Conversely, Oris not associated with CpG islands have a single initiation site. OGRE density along chromosomes correlated with previously published replication timing data. Ori sequences centered on the OGRE are also predicted to have high intrinsic nucleosome occupancy. Finally, OGREs predict G-quadruplex structures at Oris that might be structural elements controlling the choice or activation of replication origins.     

Mapping of the in vivo start site for leading strand DNA synthesis in plasmid R1.

We have previously constructed Escherichia coli strains in which an R1 plasmid is integrated into the origin of chromosome replication, oriC. In such intR1 strains, oriC is inactive and initiation of chromosome replication instead takes place at the integrated R1 origin. Due to the large size of the chromosome, replication intermediates generated at the R1 origin in these strains are considerably more long-lived than those in unintegrated R1 plasmids. We have taken advantage of this and performed primer extensions on total DNA isolated from intR1 strains, and mapped the free 5' DNA ends that were generated as replication intermediates during R1 replication in vivo. The sensitivity of the mapping was considerably improved by the use of a repeated primer extension method (RPE). The free DNA ends were assumed to represent normal in vivo start sites for leading strand DNA synthesis in plasmid R1. The ends were mapped to a short region approximately 380 bp away from the R1 minimal origin, and the positions agreed well with previous in vitro mappings. The same start positions were also utilized in the absence of the DnaA protein, indicating that DnaA is not required for determination of the position at which DNA synthesis starts during initiation of replication at the R1 origin.     

Mechanism of replication of human mitochondrial DNA. Localization of the 5' ends of nascent daughter strands

Human mitochondrial DNA contains two physically separate and distinct origins of DNA replication. The initiation of each strand (heavy and light) occurs at a unique site and elongation proceeds unidirectionally. Animal mitochondrial DNA is novel in that short nascent strands are maintained at one origin (D-loop) in a significant percentage of the molecules. In the case of human mitochondrial DNA, there are three distinct D-loop heavy strands differing in length at the 5' end. We report here the localization of the 5' ends of nascent daughter heavy strands originating from the D-loop region. Analyses of the map positions of 5' ends relative to known restriction endonuclease cleavage sites and 5' end nucleotides indicate that the points of initiation of D-loop synthesis and actual daughter strands are the same. In contrast, the second origin is located two-thirds of the way around the genome where light strand synthesis is presumably initiated on a single-stranded template. Mapping of 5' ends of daughter light strands at this origin relative to known restriction endonuclease cleavage sites reveals two distinct points of initiation separated by 37 nucleotides. This origin is in the same relative genomic position and shows a high degree of DNA sequence homology to that of mouse mitochondrial DNA. In both cases, the DNA region within and immediately flanking the origin of DNA replication contains five tightly clustered tRNA genes. A major portion of the pronounced DNA template secondary structure at this origin includes the known tDNA sequences.     

The structure of the initiation complex at the replication origin, oriC, of Escherichia coli.

Two distinct regions in the replication origin, oriC, of Escherichia coli are separately distorted upon initiation complex formation by the initiator protein DnaA. The AT-rich region in the left part of oriC and the start site region in the right part of oriC. Chemical modification of single-stranded DNA was observed at both regions whereas endonuclease recognition of DNA mini-bulges specifically occurred in the start site region. We show that the helical phasing of binding sites for DnaA protein in oriC is important for origin function. An insertion or deletion of one helical turn between the two rightmost binding sites does not alter the efficiency of replication initiation, whereas all modifications of distance by less or more than one helical turn result in inactivation of oriC. DnaA binding and helical distortions in the AT-rich region as well as in the start site region are not affected in the distance mutants irrespective of their functionality in vivo. We propose a specific compact nucleoprotein structure for the initiation complex.