Multi-objective optimization and Design of
Experiments as tools to tailor molecularly imprinted
polymers specific for glucuronic acid
Stephanie Kunath1), Nataliya Marchyk2), Karsten Haupt2), Karl-Heinz Feller1)
1)University
of Applied Sciences Jena, Department of Medical Engineering and Biotechnology, Jena, Germany
2)Enzyme and Cell Engineering Laboratory, UMR CNRS 6022, Compiègne University of Technology, Compiègne, France
INTRODUCTION
DoE
Molecularly imprinted polymers (MIPs) are tailor-made biomimetic receptors showing
high affinity and selectivity for target molecules.
 Block diagram: influencing factors (x1 - x3)
objective parameters (y1 - y4)
O
They are obtained by polymerization of functional and cross-linking monomers in the
presence of molecular templates, and contain binding sites with "memory" for the size,
shape and chemical functionality of the template molecule.
NH
HN
NH
H
O
H
C
OH
H
O
O
O
OH
OH
O
NH2
•Bmax – maximal absorption of GA at its saturating concentration (measure of capacity)
•K50 – polymer concentration at half maximum absorption of GA (measure of affinity)
•B2 – relative absorption of GA at 2 mg/mL of polymer
•IF – imprinting factor
Figure 1. Schematic Representation of the Molecular Imprinting Procedure
MIP
=
COMPLEX SYSTEM
 3 influencong factors with 2 factor levels
=
Compositional / operational factors dependency
Influencing factors
Factor levels
MAM, eq to analyte
3
7
CL, %
60
83
AIBN, mol% of double bonds 0,56
1
 Multi-objective optimization based on Derringer and Suich (Derringer, G. et al. 1980)
- Optimization of MIP improving parameters via desirability functions;
- Surface response plots;
MIP performance is a function of :
- Monomers [nature, distribution, amount]
- Initiator [nature, amount]
- Porogenic solvent
- Physical form
- Type of initiation [UV/VIS/ ∆]
RESULTS
= Important to understand factors interaction and dependancy
GOOD
BAD
Objectives
We present a multi-objective optimization of binding properties of MIPs based on a
Design of Experiments (DoE) plan. The power of the applied optimization methods is
illustrated using a polymer imprinted withthe template glucuronic acid as a model.
Model template glucuronic acid
As a model target a biologically important molecule was used, glucuronic aid, an
important building block of glycoconjugates such as e.g. hualuronan and chondroitin
sulfate which represent parts of proteoglycons in the intercellular matrix (glycocalix)
(Kobata, A. et al. 2005) and of cartilage. Glucuronic acid is also part of many glycoconjugate
drug metabolites (Baillie, T.A. ey al. 2002)
EXPERIMENTAL
To optimize the MIP :
MIP synthesis
Target template
HO
Functional monomers
NH
O
O
NH2
OH
HN
OH
OH
Β-D-glucuronic acid (GA)
Cross-linking monomer
O
O
O
Methacrylamide (MAM)
Porogen
O
S
O
Acrylamido benzamidine (AAB)
Radical initiator
CN
CN
N
N
O
Ethyleneglycol dimethacrylate
Anhydrous DMSO
Predicted best polymer: 83% CL, 3eq MAM, 0.65 mol% AIBN
NH2
OH
O
 Lower GA MAM equivalent, Increase cross-linking degree (A)
 Use 0.65 mol% of initiator with low MAM concentration (B)
and high cross-linking degree (C)
Comparison of optimized polymer with
the starting point polymer:
‘Affinity’ :
6 times ↗
‘Capacity’ : 40% ↗
IF :
1.3 times ↗
Equilibrium binding isotherms of the predicted
polymer (M-MIP, R-reference polymer) in
methanol:water (9:1 vol)
AIBN
- The ratio GA : AAB = 1:1 was kept constant;
- Polymerization by precipitation, 50°C, overnight
- Polymers were synthesized according to a full factorial experimental design plan
based on 3 influencing factors (crosslinking degree (CL), monomer (MAM) and
initiator (AIBN) concentrations);
MIP evaluation
- Equilibrium binding experiments with radioligand GA[14C] in methanol:water (9:1);
- Analysis of binding parameters with ANOVA, regression analysis and multi-objective
optimization method based on a desirability approach.
CONCLUSIONS
• Comprehensive multi-objective optimization of the binding properties of MIPs has been
performed;
•Cause-effect-correlations of factors were determined.
• In addition, an optimal glucuronic acid binding polymer composition was also found for use
in pure water as the solvent (not shown) ;
Acknowledgements
The authors thank EFRE (TNA VI-1/2009) and the region of Picardy, France (Glycosense project) for financial
support.