Replication fork movement sets chromatin loop size and origin choice in mammalian cells.

Genome stability requires one, and only one, DNA duplication at each S phase. The mechanisms preventing origin firing on newly replicated DNA are well documented, but much less is known about the mechanisms controlling the spacing of initiation events(2,3), namely the completion of DNA replication. Here we show that origin use in Chinese hamster cells depends on both the movement of the replication forks and the organization of chromatin loops. We found that slowing the replication speed triggers the recruitment of latent origins within minutes, allowing the completion of S phase in a timely fashion. When slowly replicating cells are shifted to conditions of fast fork progression, although the decrease in the overall number of active origins occurs within 2 h, the cells still have to go through a complete cell cycle before the efficiency specific to each origin is restored. We observed a strict correlation between replication speed during a given S phase and the size of chromatin loops in the next G1 phase. Furthermore, we found that origins located at or near sites of anchorage of chromatin loops in G1 are activated preferentially in the following S phase. These data suggest a mechanism of origin programming in which replication speed determines the spacing of anchorage regions of chromatin loops, that, in turn, controls the choice of initiation sites.

Integrative recombination between adenovirus type 12 DNA and mammalian DNA in a cell-free system: joining by short sequence homologies.

A cell-free system was developed to investigate the mechanism of how junctions are formed between viral and cellular DNAs during adenoviral DNA integration into the hamster cell genome. Recombination between the segment of adenovirus type 12 (Ad12) DNA, that comprises sequence coordinates 20885-24053, subsequently termed PstI-D fragment and the hamster preinsertion DNA sequence p7 was studied in a cell-free system. The p7 DNA segment had served as viral DNA integration site in the Ad12-induced tumor CLAC1. Nuclear extracts initially from uninfected BHK21 hamster cells were fractionated by a series of chromatographic steps. DNAs of the in vitro generated recombinants were analyzed in detail. In the course of the recombination reaction, the two linear molecules were joined. The reaction took place between two short homologous sequences one of which was always at or very close to a DNA terminus, the second one could be several kilobase pairs remote from a DNA terminus. Apparently, the nucleotide sequence at the terminus of one recombining molecule determined the point of junction by searching for short homologies in the partner molecule. The recombination reaction was not conservative, the sequences in-between the short sequence homologies and one of the short sequence homologies were deleted in the in vitro recombinants. Two main criteria influenced the choice of interacting short sequence homologies: perfect homologies of 8-9 bp were most frequently found, they were preferred over more extended, but less perfect homologies. Comparing different short sequence homologies with similar stabilities, those combinations seemed to be chosen in the reaction which led to a minimal loss of nucleotides in the recombinants. The in vitro activity was found in nuclear extracts from both hamster and human cells. The activity was, hence, available for Ad12 DNA in productively infected human and abortively infected hamster cells. The specific recombination activity was increased in nuclear extracts of hamster cells abortively infected with Ad12. The junction sites in the recombinants, which were generated by the cell-free system, were very similar to junctions between adenoviral and cellular DNAs cloned from Ad12-induced tumor cells and Ad12-transformed cell lines.