One of the hazards of recombinant protein expression is that you don't always get what you want. In the case of the Hermes transposase, the active hexamer was proteolytically cleaved during expression to a smaller species, and this is what we crystallized. In this assembly, we have two Hermes 79-612 monomers (one in red and the other in orange). Intertwined with the N-terminal domains of each Hermes monomer is another Hermes fragment consisting of residues 79-162 (shown in green).
The central catalytic domain of Hermes has a DDE RNase-like fold as seen in retroviral integrases and the Tn5, MuA, and Mos1 transposases. However, a large all-helical domain is inserted into this fold. Its structure does not resemble that of any other protein, and it seems likely that it plays an important role in the formation or stabilization of DNA hairpins on flanking DNA during Hermes transposition.
We used the structure shown here to model a hexameric assembly of Hermes and to model where DNA might bind. You can see our current effort here.