Replication Origins Exposed on the Surface of a Replication Domain by Transcription Elongation are Preferentially Fired for DNA Replication

Abstract

In mammals, DNA replication is initiated at discrete loci known as replication origins. Among totally ∼250,000 potential replication origins, only a small subset is activated and used for complete replication of the whole genome during any S phase, but the selection mechanism remains elusive. Using Stochastic Optical Reconstruction Microscopy (STORM) super-resolution imaging, we mapped the spatial distribution of replication origins relative to individual replication domains (RDs). Metabolic pause labeling revealed that DNA replication mainly initiated at the spatial boundary of replication domains. Further, we choose two adjacent RDs and used fluorescence in situ hybridization (FISH) to fluorescently label the RDs and replication origins of them. Interestingly, we found that predominant origins of the early S phase replication domain (ESRD) which were actually inside of the RD became exposed on the surface of the RD from middle G1 phase to the beginning of S phase (G1/S phase), not the case with dormant origins of ESRD or predominant origins of the late S phase replication domain (LSRD). To understand what drives ESRD predominant origins out, we inhibited transcription elongation or knocked down CTCF or cohesin and got spatial redistribution of predominant origins disappeared. To know why origins who were located outside could be fired, we labeled proliferating cell nuclear antigen (PCNA), minichromosome maintenance complex component 2 (MCM2) and CCCTC-binding factor (CTCF) in G1 phase and found PCNA clusters distributed outside of replication domains while MCM2 or CTCF distributed randomly. Taken together, our STORM images reveal that DNA replication initiates at origins exposed on the surface of replication domains where replication proteins distribute. Replication Origins selection may be a result of their spatial redistribution mediated by cooperation of transcription elongation and CTCF-cohesin formed loops.