General and specific replication profiles are detected in normal human cells by genome-wide and single-locus...

Mammalian genomes are replicated under a flexible program, with random use of origins and variable fork rates, and many details of the process must be still unraveled. Molecular combing provides a set of direct data regarding the replication profile of eukaryotic cells: fork rates; organization of the replication clusters; proportion of unidirectional forks; and fork dynamics. In this study the replication profiles of different primary and immortalized non-cancer human cells (lymphocytes, lymphoblastoid cells, fibroblasts) were evaluated at the whole-genome level or within reference genomic regions harboring coding genes. It emerged that these different cell types are characterized by specific replication profiles. In primary fibroblasts, a remarkable fraction of the mammalian genome was found to be replicated by unidirectional forks, and interestingly, the proportion of unidirectional forks further increased in the replicating genome along the population divisions. A second difference concerned in the proportion of paused replication forks, again more frequent in primary fibroblasts than in PBL/lymphoblastoid cells. We concluded that these patterns, whose relevance could escape when genomic methods are applied, represent normal replication features. In single-locus analyses, unidirectional and paused replication forks were highly represented in all genomic regions considered with respect to the average estimates referring to the whole-genome. In addition, fork rates were significantly lower than whole-genome estimates. Instead, when considering the specificities of each genomic region investigated (early to late replication, normal or fragile site) no further differentiating features of replication profiles were detected. These data, representing the integration of genome-wide and single-locus analyses, highlight a large heterogeneity of replication profiles among cell types and within the genome, which should be considered for the correct use of replication datasets.