The first step of the reaction between DNA and the antitumor drug cisplatin or its clinically inactive isomer transplatin yields monofunctional adducts. Most of the cisplatin monofunctional adducts further react and rather rapidly (t(1/2) smaller than a few hours) to form intrastrand and interstrand cross-links. It is generally accepted that the clinical activity of cisplatin is related to the formation of bifunctional lesions. As concerns transplatin, several studies disagree on the rate of closure of the monofunctional adducts and the nature of the bifunctional lesions. In order to explain these discrepancies, we have prepared several duplexes containing a single monofunctional trans-[Pt(NH3)2(dG)Cl]+ adduct and zero to two monofunctional [Pt(dien)(dG)]2+ adducts at defined positions. In these duplexes, the inert [Pt(dien)(dG)]2+ adducts mimic the presence of transplatin monofunctional adducts. We show that the closure of the transplatin monofunctional adducts is strongly affected by the presence of other adducts and by the length of the duplexes. These findings suggest that the discrepancies in the literature originate from the nature of the platinated samples (molar ratio of bound platinum per nucleotide, length of the DNA fragments). Our general conclusion is that within transplatin-modified DNA, at a low level of platination, the monofunctional adducts evolve slowly (t(1/2) > 24 h) into bifunctional lesions and that these bifunctional lesions are mainly interstrand cross-links. This could explain, at least in part, the clinical inefficiency of transplatin.