Multi-objective optimization and design of experiments as tools to tailor molecularly imprinted polymers specific for glucuronic acid

We present a multi-objective optimization of the binding properties of a molecularly imprinted polymer (MIP) which specifically binds glucuronic acid (GA). A design of experiments approach is used to improve four different parameters that describe the binding properties of the polymer. Eleven different methacrylamide-co-ethyleneglycol dimethacrylate polymers imprinted with GA were synthesized according to a full factorial experimental design plan with 3 influencing factors (degree of cross-linking, molar equivalent of monomer to template and initiator concentration). These polymers were characterized by adsorption of the radiolabeled target analyte in methanol:water 9:1. The binding parameters were computed to optimize the polymer composition, taking into account four objective variables: the maximum binding capacity at high (Bmax) and low (B2) analyte concentrations, the equilibrium constant K50, and the imprinting factor (IF, binding to MIP/binding to NIP). With the multi-objective optimization method based on a desirability approach the composition of a twelfth "ideal" polymer could be predicted. This predicted polymer with highest "desirability" was synthesized with a composition of 0.65 mol% of initiator and a 1:4:20 ratio of template:functional monomers:cross-linker (T:M:X) (80% of cross-linking), and found to be the overall best MIP. Improvements over the original starting polymer were a 6 times lower K50, which corresponds to higher affinity, 20% higher capacity at low analyte concentration (B2), 40% higher capacity (Bmax) and 1.3 times increased imprinting factor (IF). Binding assays were also performed in aqueous solvents. Good binding properties were obtained in pure water with an imprinting factor of 3.2. Thus, this polymer is potentially applicable to biological samples like urine where glucuronides occur.