Common fragile sites (CFSs) are chromosome regions prone to breakage upon replication stress known to drive chromosome rearrangements during oncogenesis. Most CFSs nest in large expressed genes, suggesting that transcription could elicit their instability; however, the underlying mechanisms remain elusive. Genome-wide replication timing analyses here show that stress-induced delayed/under-replication is the hallmark of CFSs. Extensive genomewide analyses of nascent transcripts, replication origin positioning and fork directionality reveal that 80% of CFSs nest in large transcribed domains poor in initiation events, replicated by long-travelling forks. Forks that travel long in late S phase explains CFS replication features, whereas formation of sequence-dependent fork barriers or head-on transcriptionreplication conflicts do not. We further show that transcription inhibition during S phase, which suppresses transcription-replication encounters and prevents origin resetting, could not rescue CFS stability. Altogether, our results show that transcription-dependent suppression of initiation events delays replication of large gene bodies, committing them to instability.

Replication stress is a primary threat to genome stability and has been implicated in tumorigenesis 1,2 . Common fragile sites (CFSs) are loci hypersensitive to replication stress 3 and are hotspots for chromosomal rearrangements in cancers 4 . CFSs replicate late in S-phase 3 , are cell-type dependent 4-6 and nest within very large genes 4,[7][8][9] . The mechanisms responsible for CFS instability are still discussed, notably the relative impact of transcription-replication conflicts 7,8,10 versus their low density in replication initiation events 5,6 . Here we address the relationships between transcription, replication, gene size and instability by manipulating the transcription of three endogenous large genes, two in chicken and one in human cells.Remarkably, moderate transcription destabilises large genes whereas high transcription levels alleviate their instability. Replication dynamics analyses showed that transcription quantitatively shapes the replication program of large genes, setting both their initiation profile and their replication timing as well as regulating internal fork velocity. Noticeably, high transcription levels advance the replication time of large genes from late to mid S-phase, which most likely gives cells more time to complete replication before mitotic entry.Transcription can therefore contribute to maintaining the integrity of some difficult-toreplicate loci, challenging the dominant view that it is exclusively a threat to genome stability.It is largely agreed that CFSs tend to remain incompletely replicated until mitosis upon replication stress. Incompletely replicated regions are processed by specific endonucleases promoting mitotic DNA synthesis and sister chromatid separation, eventually at the cost of chromosomal rearrangements [11][12][13][14][15] . Two main mechanisms have been suggested to explain this delayed replication completion. One postulates that secondary DNA structures 10 or transcription-dependent replication barriers, notably R-loops 7,8,10 , lead to fork stalling and collapse. The other proposes that replication of the core of the CFSs by long-travelling forks due to their paucity in initiation events is specifically delayed upon fork slowing 5,6 . Here we .

Replication stress is a primary threat to genome stability and has been implicated in tumorigenesis1, 2. Common fragile sites (CFSs) are loci hypersensitive to replication stress3 and are hotspots for chromosomal rearrangements in cancers4. CFSs replicate late in S-phase3, are cell-type dependent4–6 and nest within very large genes4, 7–9. The mechanisms responsible for CFS instability are still discussed, notably the relative impact of transcription-replication conflicts7, 8, 10versus their low density in replication initiation events5, 6. Here we address the relationships between transcription, replication, gene size and instability by manipulating the transcription of three endogenous large genes, two in chicken and one in human cells. Remarkably, moderate transcription destabilises large genes whereas high transcription levels alleviate their instability. Replication dynamics analyses showed that transcription quantitatively shapes the replication program of large genes, setting both their initiation profile and their replication timing as well as regulating internal fork velocity. Noticeably, high transcription levels advance the replication time of large genes from late to mid S-phase, which most likely gives cells more time to complete replication before mitotic entry. Transcription can therefore contribute to maintaining the integrity of some difficult-to-replicate loci, challenging the dominant view that it is exclusively a threat to genome stability.