DNA replication fork progression rate and temporal organization of S phase in normal epidermis and in basal cell carcinoma

The double-pulse labeling technique for DNA fiber autoradiography was applied to epidermal cells from normal human skin and from human basal cell carcinoma (BCC). We aimed to measure the size and replication rate of the replication unit (RU) for both types of cell and to account, from these results, for our previous observation of a near doubling of S-phase duration in BCC, compared with normal skin. The mean RU size was 76 +/- 4 micron in BCC, not significantly different from the 68 +/- 6 micron value found in normal skin, so the mean of those two values (i.e., 72 micron), was used in further calculations. The rate of replication fork progression was 0.59 +/- 0.005 micron/min in the normal epidermis and 0.33 +/- 0.03 micron/min in BCC, corresponding to a replication time of the average RU equal to 61 min and 109 min, respectively. Thus, with an unchanged RU size in BCC, the observed 1.8-fold decrease in the rate of fork progression in the tumor can account entirely for our previous observation of a 1.8-fold increase in S-phase duration in this tumor, without requiring the assumption of any change in the temporal organization of DNA synthesis in the malignant cells. Considering S phase as an ordered process in which a major part, if not all, of the genome replicates at genetically determined times, we suggest that the clusters of replication units are, in turn, organized into temporally defined "sets". These sets are composed of all the clusters (whatever their chromosomal location) that are programmed to initiate replication during the same fraction of the S period. This hypothesis implies that DNA synthesis in a given set is triggered by some event coupled to progression of replication in the immediately preceding set. Based on a S-phase duration of 10.2 hours in normal skin and of 19.2 hours in BCC (our previous data), and assuming perfect synchrony and homogeneity of the clusters within each set and of each cluster's constitutive RUs, the minimum number of sequentially replicating sets, in both instances, can be estimated as roughly equal to 10.