ADP-ribosylation reactions have been studied now for over 30 years. They came to light originally in studies of some bacterial toxins, which turned out to be mono-ADP-ribosyl transferases. Subsequently, endogenous mono-ADP-ribosyltransferases were discovered. Although the substrates of the toxins are always so-called G-proteins, the substrate for a muscle mono-ADP-ribosyl transferase has been shown to be an extra-cellular cell adhesion molecule. A new pathway of NAD catabolism is the recently described cyclic ADP-ribose; this seems to be involved in calcium metabolism. Just under 30 years ago, poly(ADP-ribose) polymerase was discovered. From protein studies as well as from recent molecular biology, some amino acids essential for the enzyme activity or for binding to DNA have been identified. I suggest that poly(ADP-ribose) polymerase has several related functions in maintaining the integrity of the genome in eukaryotic cells. The highly-charged polymer, poly(ADP-ribose), is always made when free, ie naked DNA ends appear, in order to ensure the correct processing of these DNA breaks. In particular, the polymer may act to prevent DNA recombination reactions that would interfere with DNA repair. In addition, the polymer may protect the free DNA end from exonuclease action, and thirdly, it may unravel the chromatin structure. This suggests that this enzyme is not a necessary component of the process of DNA excision repair, but that this enzyme is required for correct and efficient DNA excision repair.(ABSTRACT TRUNCATED AT 250 WORDS)