2.1. Structural Models

The atomic models of the DNA binding domains (DBD) of AR- and GR complexed to their respective response element were prepared using the crystallographic structures 1R4I and 1R4R, respectively. In order to achieve consistency with the AR(DBD)-DNA complex, the guanine in the spacer region of the GR(DBD)-DNA complex, was mutated in silico to cytosine. The response elements in the two complexes are thus 5′-CC AGAACAtcaTGTTCT GA-3′ (DR, for AR) and 5′-CC AGAACAtcaAGAACA GA-3′ (IR, for GR), respectively. The residues listed in bold are the core response elements including the two half sites, HS1 and HS2, respectively, the spacer is given in small letters. We have constructed two atomic models of the SPARKI receptor, one based on the structure of the AR-DNA complex (1R4I) and one on the structure of the GR-DNA complex (1R4R). In the AR-based model, termed SpAR, residues in the second zinc-binding motif of AR that differ from GR (highlighted in green in Figure 1A), were replaced with the corresponding residues of the GR protein, as in the experimental mutation (Schauwaers et al., 2007). These residues are located at the dimerization interface (see Figures 1A,C). The second model, termed SpGR, is based on the GR protein in which the residues of the first zinc-binding motif of GR that differ from AR, which are part of the DNA-binding interface, were mutated to those of AR. The resulting sequence of the proteins in both Sparki models, SpAR and SpGR is thus identical, however, their initial structures differ, since these are based on two different crystal structures.

Both SPARKI models were furthermore modeled in complex with both DNA sequences, DR and IR, respectively. Therefore, a total of six models, i.e., AR-DR, GR-IR, SpAR-DR, SpAR-IR, SpGR-DR, and SpGR-IR have been simulated.