(704j) Molecular Dynamic Simulation of Protein Devices: Pegylated Proteins and Protein Microarrays

Protein based technological devices have great potential to change how we harness the power of biology. Two examples of such protein devices are PEGylated proteins and protein microarrays. PEGylation is a process primarily used in pharmaceutical therapeutics that covalently bonds polyethylene glycol (PEG) polymer chains onto the protein's primary structure with the propose of reducing the immunogenicity of the drug by slowing renal filtration thus increasing protein's efficiency in the body. Protein microarrays, or proteins that have been covalently tethered to a solid substrate, have potentially transformative applications for detection assays in fields as diverse as proteomics, national defense and biocatalysis.

Creating the aforementioned proteins-based devices requires placing the protein in a situation different than its naturally-occurring environment, and functionalization often renders the molecule inactive. However, at certain sites, functionalization has little effect on stability and can even enhance protein activity. Because the current experimental methodology requires a laborious and expensive guess-and-check approach, being able to predict good attachment site locations that result in stabilization would greatly reduce experimental costs and facilitate rapid development of next-generation devices.

Recently, we developed a coarse-grain molecular simulation method that utilizes MBAR machine learning algorithms to accurately predict the full range of stabilizing to destabilizing attachment sites. In this presentation, we explore an in-depth analysis of β-lactimase (penicillinase), and how PEGylation and microarray tethering affects this protein's stability.

The results of an in silco screen of all experimentally-relevant surface attachment sites show that for both protein devices, device attachment site strongly affects thermostability, and experiments confirm these results. Moreover, by screening the same protein for both PEGylated proteins and protein microarrys, we show that optimal site locations for one protein functionalization approach is not predictive of the other.

Additional research in PEGylation has yielded new Go-model coarse-grain parameters for polyethylene glycol. Both a purely-repulsive and an attractive force field were used to asses PEG effects on protein stability. The results from these simulations indicate that PEG/protein structures formed in simulation contradict previous theory that PEG stability primarily derives from wrapping around the protein and instead the PEG polymer is attracted to itself and remains removed from the protein body.

Taken as a whole, these results reveal a highly-detailed view of protein-PEG and protein-surface biophysics and reveal the synergistic nature of our computational/experimental approach for the in silico rational design of protein-based devices.