home - index - site map   -   deutsch
if you don't see a molecular model:

Actinomycin D


To highlight described properties click the boxes . Move the molecule anytime with the mouse - if the view is messed up, clicking the large boxes will restore a convenient view.

The selfcomplementary DNA sequence CGATCGATCG doesn't contain the favourite intercalation site GC for actinomycin D (ActD). Nevertheless a stable, crystalizable complex is formed. At first sight this complex has rather a bizarre structure at the atomic level. At second sight there are some "normal" features. Though the DNA won't pair according to its whole selfcomplementarity, there are Watson-Crick basepairs . Note the gaps between the terminal basepairs and the stacked central pairs. These gaps are partly filled by bases from the other two DNA strands . The whole complex is symmetric, of course . So by swapping the thymines the four strands are linked together. More contacts are between adenines . There are two unpaired bases left at the ends of the strands (at one end the terminal bases could not be modelled from the data).

The overhanging 5'-ends of the DNA strands are the attachment points for the actinomycin molecules. The last CG basepair sets the stage for the ActD phenoxazone to stack to . In the overhanging strand the next base (T) is diverted to intercalate in the neighboring double strand , and the following base (A) is turned away to stack to the overhanging A of the other strand . The next unpaired base then stacks to the phenoxazone . There is a hydrogen bridge between the phenoxazone amino group and the deoxyribose of the diverted adenine . Otherwise the phenoxazone is held by stacking interaction to the guanines . Note the heavy kink in the backbone between the paired bases and the overhanging end . At this intercalation site of ActD one of the atibiotic's cyclic peptide chains occupies the (widened) minor groove of the DNA , the other binds just to the deoxyribose of the overhanging bases . There are specific contacts to guanosines: the sidechain threonines are hydrogen bonded to the guanines sandwiching the penoxazone . So although the DNA sequence displays no GC step as intercalation site, the "canonical" site is mimicked by diverting the adenine and thymine in the single stranded region. Thus the bonding interactions between ActD and DNA as found in dsDNA with the GC step are found in this case too. The missing C opposite the unpaired G at the overhanging end hardly matters, as there are no specific contacts between ActD and this C.

Actinomycin is known to inhibit translation more efficiently than transcription. The structure of this ActD-DNA complex gives a hint to the mechanism. In DNA polymerases the active site is filled to a great part with dsDNA. In RNA polymerases the DNA is further unwound, so the active site cleft contains only the template strand (the template strand is marked yellow, the transcript green). The RNA-DNA heteroduplex in the polymerase cleft is surrounded by single-stranded DNA. When the template contains a C followed by G requiring a G to be added to the nascent RNA chain, the topology resembles that of the above ActD-DNA complex. It is possible to model a molecule of Actinomycin into the active site of the RNA polymerase; this placement requires the heteroduplex resp. the ActD intercalation complex to be moved 4Å out of place. This prevents further activity of the polymerase.

H Robinson et al, Crystallographic analysis of a novel complex of actinomycin D bound to the DNA decamer CGATCGATCG, Biochemistry 40 (2001) 5587-5592
GMT Cheetham & TA Steitz, Structure of a transcribing T7 RNA polymerase initiation complex, Science 286 (1999) 2305-2309

7-01 - Rolf Bergmann