WEISS LECTURE
Carbohydrate radicals: from ethylene glycol to DNA strand breakage

Abstract

Radiation-induced DNA strand breakage results from the reactions of radicals formed at the sugar moiety of DNA. In order to elucidate the mechanism of this reaction investigations were first performed on low molecular weight model systems. Results from studies on deoxygenated aqueous solutions of ethylene glycol, 2-deoxy-d-ribose and other carbohydrates and, more relevantly, of d-ribose-5-phosphate have shown that substituents can be eliminated from the β-position of the radical site either proton and base-assisted (as in the case of the OH substituent), or spontaneously (as in the case of the phosphate substituent). In DNA the C(4’) radical undergoes strand breakage via this type of reaction. In the presence of oxygen the carbon-centred radicals are rapidly converted into the corresponding peroxyl radicals. Again, low molecular weights models have been investigated to elucidate the key reactions. A typical reaction of DNA peroxyl radicals is the fragmentation of the C(4’)-C(S’) bond, a reaction not observed in the absence of oxygen. Although OH radicals may be the important direct precursors of the sugar radicals of DNA, results obtained with poly(U) indicate that base radicals may well be of even greater importance. The base radicals, formed by addition of the water radicals (H and OH) to the bases would in their turn attack the sugar moiety to produce sugar radicals which then give rise to strand breakage and base release. For a better understanding of strand break formation it is therefore necessary to investigate in more detail the reactions of the base radicals. For a start, the radiolysis of uracil in oxygenated solutions has been reinvestigated, and it has been shown that the major peroxyl radical in this system undergoes base-catalysed elimination of , a reaction that involves the proton at N(l). In the nucleic acids the pyrimidines are bound at N(l) to the sugar moiety and this type of reaction can no longer occur. Therefore, with respect to the nucleic acids, pyrimidines are good models only in acid solutions where the elimination reaction is too slow to compete with the bimolecular reactions of the peroxyl radicals. Moreover, the long lifetime of the radical sites on the nucleic acid strand may allow reactions to occur which are kinetically of first order, and which cannot be studied in model systems at ordinary dose rates. It is therefore suggested to extend model system studies to low dose rates and to oligonucleo-tides. Such studies might eventually reveal the key reactions in radical-induced DNA degradation.