A GFP-based strategy for overexpression, purification and characterization of membrane proteins and the crystallographic structure of murine aminoacylase 3 with substrates

The primary objective of my graduate studies was to establish a GFP-based methodology for membrane proteins with a C-terminus located in the periplasm of Escherichia coli (Cout). In 2005, when I started my PhD, the first few reports of using GFP to monitor membrane protein expression and purification had emerged. However, these methods were limited to membrane proteins whose C-terminus is localized to the cytoplasm of E. coli (Cin). This restriction is due to the fact that GFP does not fluoresce in the periplasm. With 30% of all membrane proteins having Coout topology, there was a void to fill, so we set out to establish that, for Cout proteins, a C-terminally fused GFP can be repositioned to the cytoplasm by inserting the single transmembrane segment protein, Glycophorin A (GpA), between the membrane protein of interest and the GFP. I tested this GpA-GFP fusion on eleven test proteins, resulting in high expression as indicated by strong GFP fluorescence. I also demonstrated that the GpA-GFP fusion could be used with fluorescence-detection size exclusion chromatography for identifying suitable detergents for protein solubilization and stability. In the process of working with these constructs we also learned more about which types of protein expression conditions should be screened for increased yields, and that expression without the GpA-GFP fusion often drastically reduces expression levels. Despite the great success I had with expressing and purifying five of the eleven test proteins in large yields, I was unable to proceed to full-scale crystallization trials with these proteins. Instead, I had the opportunity to work on the structure of the soluble protein, aminoacylase 3 (AA3).;AA3 is a hydrolytic enzyme that catalyzes the deacetylation of N-acetylated aromatic amino acids and mercapturic acids like N-acetyl-S-1,2-dichlorovinyl-L-cysteine (NA-DCVC). NA-DCVC is the metabolite of the haloalkene tricholorethlene and AA3 deacetylation has been indirectly linked to acute renal failure in mammals. In collaboration with Dr. Alexander Pushkin we were able to determine the structure of murine AA3 (mAA3) and the non-hydrolyzing mutant, E177A-mAA3, with two substrates: N-acetyl-L-tyrosine and NA-DCVC. Using the mAA3 structures with substrates we were able to determine that the substrates bind through a hydrogen bonding network to the N-acetyl-alpha-amino carboxylic acid (NAACA) substrate component and the side chain constituent is held in place by a number of van der waals interactions. These finding have led us to propose that mAA3 substrate recognition is a dynamic process involving many residues, giving mAA3 the ability to accommodate a broad range of substrates.