Crystallography

• As aforementioned, the secondary structure of RiAFP is still unknown, thereby making the hyperactive antifreeze properties and ice-binding sites elusive. Lin et al, who have determined the secondary structure of another antifreeze protein in inchworms, have predicted that, given the alternating Alanine and Threonine repeat amino-acid residues in RiAFP (which also appear in the inchworm AFP), that RiAFP could potentially share the same structure of a parallel-stranded beta helix.
• For crystallography, large-scale expression (150 mg/mL) of the eGFP-TEV-RiAFP-His protein construct was completed in 3L E. coli cultures. The fusion protein (rather than the isolated protein) was chosen for the increased solubility and production. Tightly-packed His-trap columns were used for metal affinity chromatography to bind the poly-6-histidine tagged proteins and subsequently eluted with a gradient of imidazole-spiked buffer. SDS-PAGE Coomassie Gels were run to confirm purification and fractions that contained protein were combined and spin-concentrated. TEV protease was subsequently added at a 1:100 w/w ratio and incubated on a rocker with the fusion protein overnight at room temperature. Another SDS-PAGE Coomassie gel was run to confirm proper cleavage before loading 3 mL of the concentrated TEV, RiAFP, GFP mix onto a Hi-Load Size Exclusion Column. After size-exclusion, the most pure samples were combined and spin-concentrated to a final protein concentration of 5 mg/mL and then loaded onto crystal trays with various precipitant solutions. For the initial screen, a high-throughput Mosquito crystallization robot was used with an Emerald Biosciences Wizard grid and Hampton Research PEG grid. 8 initial crystal hits were recorded, ranging from an aggregation of needles around a nucleation center ("fuzzy balls") to a large dispersion of small, decently-ordered crystals.
• Currently, we are running pH vs. precipitating agent concentration grids, as well as buffering agent vs. precipitating agent concentration grids to optimize crystal growth. In addition, to solve the crystallographic phasing problem, we are using a quick-change, site-directed mutagenesis protocol to replace methionine residues with Se-methione for a reference point. We also plan to crystallize RiAFPs generated through MAGE to determine what structural effects are related to either enhancements or reductions in antifreeze protein activity, e.g. whether smaller/larger helix sheets might allow tighter protein-ice crystal binding to allow improved activity.

Figures