DNA replication sites (RS) in synchronized HeLa cells have been studied at the electron microscopic level. Using an improved method for detection following the in vivo incorporation of biotin-16-deoxyuridine triphosphate, discrete RS, or foci are observed throughout the S-phase. In particular, the much larger RS or foci typically observed by fluorescence microscopic approaches in mid- and late-S-phase, are found to be composed of smaller discrete foci that are virtually identical in size to the RS observed in early-S-phase. Pulse-chase experiments demonstrate that the RS of early-S-phase are maintained when chased through S-phase and into the next cell generation. Stereologic analysis demonstrates that the relative number of smaller sized foci present at a given time remains constant from early through mid-S-phase with only a slight decrease in late-S-phase. 3-D reconstruction of serial sections reveals a network-like organization of the RS in early-S-phase and confirms that numerous smaller-sized replication foci comprise the larger RS characteristic of late-S-phase.

5-Bromo-2′-deoxyuridine (BrdU) and 2′-deoxy-5-ethynyluridine (EdU) are widely used as markers of replicated DNA. While BrdU is detected using antibodies, the click reaction typically with fluorescent azido-dyes is used for EdU localisation. We have performed an analysis of ten samples of antibodies against BrdU with respect to their reactivity with EdU. Except for one sample all the others evinced reactivity with EdU. A high level of EdU persists in nuclear DNA even after the reaction of EdU with fluorescent azido-dyes if the common concentration of dye is used. Although a ten-time increase of azido-dye concentration resulted in a decrease of the signal provided by anti-BrdU antibodies, it also resulted in a substantial increase of the non-specific signal. We have shown that this unwanted reactivity is effectively suppressed by non-fluorescent azido molecules. In this respect, we have tested two protocols of the simultaneous localisation of incorporated BrdU and EdU. They differ in the mechanism of the revelation of incorporated BrdU for the reaction with antibodies. The first one was based on the use of hydrochloric acid, the second one on the incubation of samples with copper(I) ions. The use of hydrochloric acid resulted in a significant increase of the non-specific signal. In the case of the second method, no such effect was observed.