It’s definitely possible to have a superposition of states: https://www.ncbi.nlm.nih.gov/m/pubmed/28291761/
You need to use alternate conformers to generate the different states of the crystal. Thanks, Nick On 13 Dec 2017, at 20:12, Matthew Bratkowski <mab...@cornell.edu<mailto:mab...@cornell.edu>> wrote: Hello all, I am working on a ligand binds near the active site of the protein, such that part of the ligand would clash with part of the natural substrate. I recently co-crystallized the enzyme with both molecules and solved the crystal structure to high resolution (around 1.4 angstrom). Surprisingly, the structure appears to contain both molecules. A few atoms from both molecules are located only ~1.4 A apart and are clashing (although not overlapping). The electron density between them looks connected, but based on the two groups that are clashing (a methyl group and a carbonyl oxygen), I do not think that a covalent adduct occurs. I had a few questions. 1) My guess is that the crystal is "sampling" two different conformational states and that both are visible due to the high diffraction resolution. The substrate contains a ring that shows a characteristic "hole" in the electron density and binds in the exact substrate binding site, suggesting that it is not a different molecule (no molecules with ring structures were included in the sample, crystallization buffer, or cry-protectant). One of the two proteins in the ASU contains electron density for whole substrate, while the other site has only density around the ring. However, a sizable amount of red FoFc density is present around the substrate, suggesting that it is only partially occupied. Does this explanation seem plausible? 2) How would I go about modeling these two molecules in the structure? Should I include both molecules (in their entirety) in the structure? I suspect that neither the ligand nor substrate are completely occupied, so should I modify the occupancies to reflect this? Thanks, Matt
