NONRANDOM DISTRIBUTION OF DNA DOUBLE-STRAND BREAKS INDUCED BY PARTICLE IRRADIATION

Induction of DNA double-strand breaks (dsbs) in mammalian cells is dep endent on the spatial distribution of energy deposition from the ioniz ing radiation. For high LET particle radiations the primary ionization sites occur in a correlated manner along the track of the particles, while for X-rays these sites are much more randomly distributed throug hout the volume of the cell. It can therefore be expected that the dis tribution of dsbs linearly along the DNA molecule also varies with the type of radiation and the ionization density. Using pulsed-field gel and conventional gel techniques, we measured the size distribution of DNA molecules from irradiated human fibroblasts in the total range of 0.1 kbp-10 Mbp for X-rays and high LET particles (N ions, 97 keV/mu m and Fe ions, 150 keV/mu m). On a mega base pair scale we applied conve ntional pulsed-field gel electrophoresis techniques such as measuremen t of the fraction of DNA released from the well (FAR) and measurement of breakage within a specific Nod restriction fragment (hybridization assay). The induction rate for widely spaced breaks was found to decre ase with LET. However, when the entire distribution of radiation-induc ed fragments was analysed, we detected an excess of fragments with siz es below about 200 kbp for the particles compared with X-irradiation. X-rays are thus more effective than high LET radiations in producing l arge DNA fragments but less effective in the production of smaller fra gments. We determined the total induction rate of dsbs for the three r adiations based on a quantitative analysis of all the measured radiati on-induced fragments and found that the high LET particles were more e fficient than X-rays at inducing dsbs, indicating an increasing total efficiency with LET. Conventional assays that are based only on the me asurement of large fragments are therefore misleading when determining total dsb induction rates of high LET particles. The possible biologi cal significance of this non-randomness for dsb induction is discussed .