The EMBL nucleotide sequence data library.

nogalamycin and the duplex sequence dTpA (Brown et al., 1984), where it has been suggested that the sugar group at each end of the drug molecule (Fig. 2) sterically blocks the free movement of the chromophore in and out of the intercalation site. The intercalation process in solution has been extensively studied most often with reference to random-sequence DNA, and rather more rarely with respect to definedsequence polynucleotides. For DNA itself, there are ten distinct dinucleoside intercalation sites, and two for an alternating polynucleotide such as poly(dC-dG). Most solution methods for studying drug intercalation (for example, by measurement of affinity constants), tend to average out differences in binding properties at these different sites. The molecular graphics approaches outlined above, do on the other hand, focus attention on just one site, such as the -CpGsequence. This is likely to be a highaffinity site, especially at the low drug levels relevant to physiological conditions (reviewed in Neidle & Abraham, 1984). It will also be an important goal for the future effort in this field to relate the simulation data at these defined sites to intercalative drug behaviour in solution and in the cell as measured by finedetails methods such as stoppedflow kinetics and DNA footprinting.