Effect of crystal freezing and small-molecule binding on internal cavity size in a large protein: X-ray and docking studies of lipoxygenase at ambient and low temperature at 2.0 A resolution.

Flash-freezing is a technique that is commonly used nowadays to collect diffraction data for X-ray structural analysis. It can affect both the crystal and molecular structure and the molecule's surface, as well as the internal cavities. X-ray structural data often serve as a template for the protein receptor in docking calculations. Thus, the size and shape of the binding site determines which small molecules could be found as potential ligands in silico, especially during high-throughput rigid docking. Data were analyzed for wild soybean lipoxygenase-3 (MW 97 kDa) at 293 and 93 K and compared with the results from studies of its molecular complexes with known inhibitors, structures published by others for a derivative of the same enzyme (98 K) or a topologically close isozyme lipoxygenase-1 (at ambient temperature and 100 K). Analysis of these data allows the following conclusions. (i) Very small changes in the relative orientation of the molecules in the crystal can cause major changes in the crystal reciprocal lattice. (ii) The volume of the internal cavities can ;shrink' by several percent upon freezing even when the unit-cell and the protein molecular volume show changes of only 1-2%. (iii) Using a receptor structure determined based on cryogenic data as a target for computational screening requires flexible docking to enable the expansion of the binding-site cavity and sampling of the alternative conformations of the crucial residues.

One site fits both: a model for the ternary complex of folate + NADPH in R67 dihydrofolate reductase, a D2 symmetric enzyme.

R67 dihydrofolate reductase (DHFR) is a novel enzyme that confers resistance to the antibiotic trimethoprim. The crystal structure of R67 DHFR displays a toroidal structure with a central active-site pore. This homotetrameric protein exhibits 222 symmetry, with only a few residues from each chain contributing to the active site, so related sites must be used to bind both substrate (dihydrofolate) and cofactor (NADPH) in the productive R67 DHFR.NADPH.dihydrofolate complex. Whereas the site of folate binding has been partially resolved crystallographically, an interesting question remains: how can the highly symmetrical active site also bind and orient NADPH for catalysis? To model this ternary complex, we employed DOCK and SLIDE, two methods for docking flexible ligands into proteins using quite different algorithms. The bound pteridine ring of folate (Fol I) from the crystal structure of R67 DHFR was used as the basis for docking the nicotinamide-ribose-Pi (NMN) moiety of NADPH. NMN was positioned by both DOCK and SLIDE on the opposite side of the pore from Fol I, where it interacts with Fol I at the pore's center. Numerous residues serve dual roles in binding. For example, Gln 67 from both the B and D subunits has several contacts with the pteridine ring, while the same residue from the A and C subunits has several contacts with the nicotinamide ring. The residues involved in dual roles are generally amphipathic, allowing them to make both hydrophobic and hydrophilic contacts with the ligands. The result is a 'hot spot' binding surface allowing the same residues to co-optimize the binding of two ligands, and orient them for catalysis.