|M.Sc Student||Raim Vladimir|
|Subject||Defining Descriptors of Proteins for Predicting Molecular|
|Department||Department of Chemical Engineering||Supervisor||Professor Simcha Srebnik|
|Full Thesis text|
Molecular imprinting is a technique that is used to create artificial receptors by the formation of a polymer network around a template molecule, creating a molecularly imprinted polymer (MIP). These receptors may be used in any application that requires molecular recognition and they are cost-effective. However, while small molecules (0.5 kDa) are good imprintogens - their MIPs have high imprinting factors (IF) and selectivity (α), MIPs imprinted by proteins (40 kDa) showed significantly smaller IF and α. Multifunctionality, conformational flexibility, large size of the proteins and aqueous polymerization environment are some obstacles for successful protein imprinting.
Most of the existing works are devoted to elucidating the influence of polymerization conditions such as temperature, amount and kind of solvent, functional monomers and cross-linkers on efficiency of PIPs. Less attention has been paid to the influence of the properties of a template molecule itself on PIP efficiency. The main purpose of this work is to answer the question: “Given all experimental conditions equal, can protein X be imprinted better than protein Y?” To answer this question, data on IF and α was gathered from literature for differently sized imprintogens. To describe the molecular shape of a protein we employed a method of shape histograms that allows comparing molecular shape of a protein to other proteins as well as to any geometrical 3-dimensional shape. Shape histograms together with other known protein properties such as molecular weight, isoelectric point, solvent excluded surface area and volume were used as protein descriptors. Using data in the compiled tables we looked for correlation between the descriptors of a template and IF and α of the respective PIP. To automate this task we developed a program, named ProtShape, that operates with available protein structures in RCSB protein data bank.
We have found a clear trend between the linear size of the template and efficiency of imprinted polymer: the smaller the template the higher IF is achieved. Comparison of protein shapes to corrugated spheres led us to a conclusion that more rough surface of a protein correlates with lower IF of the PIP. Nonetheless, statistical analysis of template-competitor proteins in competitive adsorption experiments showed that MW/EIP descriptor long used as a tool to explain results of these experiments indeed is the best of all other molecular descriptors examined and their combinations.