No abstract

The crystal structure of a mutant androgen receptor (AR) ligand-binding domain (LBD) in complex with the agonist 9alpha-fluorocortisol has been determined at 1.95 A resolution. This mutant AR contains two mutations (L701H and T877A) and was previously reported as a high-affinity cortisol/cortisone responsive AR (AR(ccr)) isolated from the androgen-independent human prostate cancer cell lines MDA PCa 2a and 2b (Zhao et al. Nature Med. 2000, 6, 703-6). The three-dimensional structure of the AR(ccr) LBD complexed with 9alpha-fluorocortisol shows the typical conformation of an agonist-bound nuclear receptor in which helix 12 is precisely positioned as a "lid" for the ligand-binding pocket. Binding of 9alpha-fluorocortisol to the AR(ccr) involves favorable hydrogen bond patterns on the C17 and C21 substituents of the ligand due to the mutations at 701 and 877 in the AR(ccr). Our studies provide the first structural explanation for the glucocorticoid activation of AR(ccr), which is important for the development of new therapeutic treatments for androgen-independent prostate cancer.

This study describes an efficient multiparallel automated workflow of cloning, expression, purification, and crystallization of a large set of construct variants for isolated protein domains aimed at structure determination by X-ray crystallography. This methodology is applied to MAPKAP kinase 2, a key enzyme in the inflammation pathway and thus an attractive drug target. The study reveals a distinct subset of truncation variants with improved crystallization properties. These constructs distinguish themselves by increased solubility and stability during a parallel automated multistep purification process including removal of the recombinant tag. High-throughput protein melting point analysis characterizes this subset of constructs as particularly thermostable. Both parallel purification screening and melting point determination clearly identify residue 364 as the optimal C terminus for the kinase domain. Moreover, all three constructs that ultimately crystallized feature this C terminus. At the N terminus, only three amino acids differentiate a noncrystallizing from a crystallizing construct. This study addresses the very common issues associated with difficult to crystallize proteins, those of solubility and stability, and the crucial importance of particular residues in the formation of crystal contacts. A methodology is suggested that includes biophysical measurements to efficiently identify and produce construct variants of isolated protein domains which exhibit higher crystallization propensity.