A molecularly imprinted polymer for 17α-ethynylestradiol evaluated by immunoassay
3
A molecularly imprinted polymer for 17a-ethynylestradiol evaluated by immunoassay Irene Idziak* and Amina Benrebouh MDS Pharma Services, Advanced Analytical R&D Dept., 2350 Cohen St., Montréal, Quebec, Canada H4R 2N6. E-mail: [email protected] Received 21st February 2000, Accepted 2nd June 2000 Published on the Web 29th June 2000 Molecularly imprinted polymers have been synthesized using 17a-ethynylestradiol as template. The affinity and the specificity of the polymers towards the template were evaluated by radioligand immunoassay. The polymer prepared in toluene with 4-vinylpyridine as functional monomer and ethyleneglycol dimethacrylate as crosslinker was found to have a strong and highly selective affinity for 17a-ethynylestradiol. A comparison with the natural antibody indicated that the recognition properties of the polymer towards several structurally related steroids were similar or superior to those of the antibody. The results show that it is possible to make a successful polymeric receptor site for a lipophilic molecule that can provide only two potential points of hydrogen bonding interaction. Introduction Molecular imprinting is an emerging technique used to synthesize polymers with specific receptor sites. These materi- als are constructed around a chosen template which is subsequently removed, leaving cavities that are complementary in shape and electronic environment to the template. They have found a wide range of applications, such as chromatographic separations, 1 solid phase extraction, 2 catalysis, 3 chemosensors 4 and artificial antibodies. 5,6 In order to achieve optimal rebinding of the imprint molecule careful consideration has to be given to the design of the polymer. The polymer needs to be rigid enough to retain the memory of the imprint and it must have functional groups within the cavity that can attractively interact with the template molecule. Molecular recognition can be provided by a combina- tion of these two factors plus more subtle forces, such as hydrophobic interactions, the effect of which is harder to predict. The polymers are typically prepared with ethyleneglycol dimethacrylate or trimethylolpropane trimethacrylate 7 as cross- linking monomers. This yields rigid macroporous materials. 8 Functional monomers, such as methacrylic acid, acrylamides or vinylpyridine are most often used to participate in hydrogen bonding with the template and are copolymerized with the crosslinkers. If the template contains carboxy or amino groups ionic interactions are possible and should lead to stronger binding. It remains a challenge, however, to make a molecularly imprinted polymer (MIP) that is capable of specific recognition, using a lipophilic template molecule with few functional groups. 9 Many steroids fall into that category and in spite of their clinical and analytical importance they have only been used in a few instances to make MIPs. 10,11 In our quest for a convenient but sensitive method for the detection of 17a- ethynylestradiol in biological samples we decided to prepare molecular imprints and evaluate their potential by immu- noassay. Experimental Materials Ethyleneglycol dimethacrylate (EDMA), 4-vinylpyridine (4-VP), methacrylic acid (MAA), 1,1A-azobis(cyclohexane- carbonitrile) (VAZO®), 17a-ethynylestradiol, estradiol, es- trone, mestranol, mifepristone, testosterone and anhydrous grade solvents were purchased from Sigma-Aldrich, (Oakville, Ontario, Canada) and used as received. Mandel Scientific Co (Guelph, ON, Canada) supplied 17a-[6,7- 3 H(N)]-ethynylestra- diol (specific activity 44 Ci mmol 21 ); liquid scintillation cocktail (Ultima Gold) was from Packard (Meriden, CT, USA). The rabbit anti-ethynylestradiol antibody and the specificity data for the antiserum were supplied by Wein Laboratories Inc. (Succasunna, NJ, USA). Preparation of the molecular imprinted polymers Ethynylestradiol imprinted polymers were synthesized with either 4-vinylpyridine, methacrylic acid, or a mixture of these compounds as functional monomers and ethyleneglycol dime- thacrylate as crosslinker. The corresponding blank polymers were made simultaneously (Table 1). Polymer P1 was prepared in the following manner. Ethynyl- estradiol (80 mg, 0.27 mmol) was suspended in toluene (1.8 ml) in a borosilicate test tube and put into solution by the addition of 4-VP (260 ml, 2.4 mmol). EDMA (1.8 ml, 9.6 mmol) and the initiator, VAZO® (30 mg, 0.12 mmol), were subsequently added. The solution was sonicated for 1 min and sparged with nitrogen. The test tube was then covered with parafilm and placed in an ice bath. The reaction mixture was allowed to reach room temperature while being irradiated at 365 nm for 24 h. The resulting hard polymer was ground in a Retsch RM100 Mortar Grinder (Brinkmann Instruments Canada Ltd., Mississauga, ON, Canada) and water sieved to particles < 25 mm, which were filtered off through a fluted paper filter. The template was subsequently removed by continuous extraction with methanol Table 1 Polymerization mixtures Polymer Ethynyl- estradiol/mmol 4-VP/ mmol MAA/ mmol EDMA/ mmol P1 0.27 2.4 — 9.6 B1 — 2.4 — 9.6 P2 0.27 1.8 1.2 9.6 B2 — 1.8 1.2 9.6 P3 0.27 — 3.0 9.6 B3 — — 3.0 9.6 This journal is © The Royal Society of Chemistry 2000 DOI: 10.1039/b002350g Analyst, 2000, 125, 1415–1417 1415 Published on 29 June 2000. Downloaded by University of Glasgow Library on 17/08/2013 09:32:20. View Article Online / Journal Homepage / Table of Contents for this issue
Transcript of A molecularly imprinted polymer for 17α-ethynylestradiol evaluated by immunoassay
A molecularly imprinted polymer for 17a-ethynylestradiolevaluated by immunoassay
Irene Idziak* and Amina Benrebouh
MDS Pharma Services, Advanced Analytical R&D Dept., 2350 Cohen St., Montréal, Quebec,Canada H4R 2N6. E-mail: [email protected]
Received 21st February 2000, Accepted 2nd June 2000Published on the Web 29th June 2000
Molecularly imprinted polymers have been synthesized using 17a-ethynylestradiol as template. The affinity andthe specificity of the polymers towards the template were evaluated by radioligand immunoassay. The polymerprepared in toluene with 4-vinylpyridine as functional monomer and ethyleneglycol dimethacrylate as crosslinkerwas found to have a strong and highly selective affinity for 17a-ethynylestradiol. A comparison with the naturalantibody indicated that the recognition properties of the polymer towards several structurally related steroids weresimilar or superior to those of the antibody. The results show that it is possible to make a successful polymericreceptor site for a lipophilic molecule that can provide only two potential points of hydrogen bonding interaction.
Introduction
Molecular imprinting is an emerging technique used tosynthesize polymers with specific receptor sites. These materi-als are constructed around a chosen template which issubsequently removed, leaving cavities that are complementaryin shape and electronic environment to the template. They havefound a wide range of applications, such as chromatographicseparations,1 solid phase extraction,2 catalysis,3 chemosensors4
and artificial antibodies.5,6
In order to achieve optimal rebinding of the imprint moleculecareful consideration has to be given to the design of thepolymer. The polymer needs to be rigid enough to retain thememory of the imprint and it must have functional groupswithin the cavity that can attractively interact with the templatemolecule. Molecular recognition can be provided by a combina-tion of these two factors plus more subtle forces, such ashydrophobic interactions, the effect of which is harder topredict.
The polymers are typically prepared with ethyleneglycoldimethacrylate or trimethylolpropane trimethacrylate7 as cross-linking monomers. This yields rigid macroporous materials.8Functional monomers, such as methacrylic acid, acrylamides orvinylpyridine are most often used to participate in hydrogenbonding with the template and are copolymerized with thecrosslinkers. If the template contains carboxy or amino groupsionic interactions are possible and should lead to strongerbinding. It remains a challenge, however, to make a molecularlyimprinted polymer (MIP) that is capable of specific recognition,using a lipophilic template molecule with few functionalgroups.9 Many steroids fall into that category and in spite oftheir clinical and analytical importance they have only beenused in a few instances to make MIPs.10,11 In our quest for aconvenient but sensitive method for the detection of 17a-ethynylestradiol in biological samples we decided to preparemolecular imprints and evaluate their potential by immu-noassay.
Experimental
Materials
Ethyleneglycol dimethacrylate (EDMA), 4-vinylpyridine(4-VP), methacrylic acid (MAA), 1,1A-azobis(cyclohexane-
carbonitrile) (VAZO®), 17a-ethynylestradiol, estradiol, es-trone, mestranol, mifepristone, testosterone and anhydrousgrade solvents were purchased from Sigma-Aldrich, (Oakville,Ontario, Canada) and used as received. Mandel Scientific Co(Guelph, ON, Canada) supplied 17a-[6,7-3H(N)]-ethynylestra-diol (specific activity 44 Ci mmol21); liquid scintillationcocktail (Ultima Gold) was from Packard (Meriden, CT, USA).The rabbit anti-ethynylestradiol antibody and the specificitydata for the antiserum were supplied by Wein Laboratories Inc.(Succasunna, NJ, USA).
Preparation of the molecular imprinted polymers
Ethynylestradiol imprinted polymers were synthesized witheither 4-vinylpyridine, methacrylic acid, or a mixture of thesecompounds as functional monomers and ethyleneglycol dime-thacrylate as crosslinker. The corresponding blank polymerswere made simultaneously (Table 1).
Polymer P1 was prepared in the following manner. Ethynyl-estradiol (80 mg, 0.27 mmol) was suspended in toluene (1.8 ml)in a borosilicate test tube and put into solution by the addition of4-VP (260 ml, 2.4 mmol). EDMA (1.8 ml, 9.6 mmol) and theinitiator, VAZO® (30 mg, 0.12 mmol), were subsequentlyadded. The solution was sonicated for 1 min and sparged withnitrogen. The test tube was then covered with parafilm andplaced in an ice bath. The reaction mixture was allowed to reachroom temperature while being irradiated at 365 nm for 24 h. Theresulting hard polymer was ground in a Retsch RM100 MortarGrinder (Brinkmann Instruments Canada Ltd., Mississauga,ON, Canada) and water sieved to particles < 25 mm, which werefiltered off through a fluted paper filter. The template wassubsequently removed by continuous extraction with methanol
Table 1 Polymerization mixtures
PolymerEthynyl-estradiol/mmol
4-VP/mmol
MAA/mmol
EDMA/mmol
P1 0.27 2.4 — 9.6B1 — 2.4 — 9.6P2 0.27 1.8 1.2 9.6B2 — 1.8 1.2 9.6P3 0.27 — 3.0 9.6B3 — — 3.0 9.6
This journal is © The Royal Society of Chemistry 2000
DOI: 10.1039/b002350g Analyst, 2000, 125, 1415–1417 1415
Publ
ishe
d on
29
June
200
0. D
ownl
oade
d by
Uni
vers
ity o
f G
lasg
ow L
ibra
ry o
n 17
/08/
2013
09:
32:2
0.
View Article Online / Journal Homepage / Table of Contents for this issue
in a Soxhlet apparatus for 6 h. The extraction was stoppedovernight and continued the next day for a further 6 h. Thepolymer was dried at 60 °C for 2 d. Polymers P2 and P3 wereprepared in the same way. The corresponding blank polymers,B1, B2 and B3 were prepared simultaneously, omittingethynylestradiol.
Immunoassay
The polymer was suspended in the incubation solvent andsonicated for homogeneity. Of this suspension, 570 ml wastransferred to a glass tube followed by 30 ml of 17a-[6,7-3H(N)]-ethynylestradiol solution (specific activity 44 Cimmol21). The volume was brought up to 700 ml with solvent orappropriate solutions of competing ligand. The tubes wereshaken overnight in an orbital shaker at 150 rpm. Aftercentrifugation (Beckman J6-MI, Beckman Coulter Canada Inc.,Mississauga, ON, Canada) for 20 min, supernatant (350 ml) wasadded to liquid scintillation cocktail (10 ml). The radioactivitywas measured using a Tri-Carb Model 1900 TR b-radiationcounter (Packard Instruments, Downers Grove, IL, USA).
Results and discussion
MIPs designed for use in immunoassays must have high qualitybinding sites in order to meet the necessary standard ofspecificity. The affinity of the polymer for the template isultimately determined by the change in free energy of thesystem. A certain loss of entropy is inevitable when a templatemolecule binds to the polymer. In the case of ethynylestradiol(Fig. 1) the entropy change is not large because it is a rigidmolecule due to its fused 4-ring structure. The loss of entropymust be offset by an increase in enthalpy which is dependent onthe template’s ability to attractively interact with the polymermatrix. Hydrogen bonding between the template and suitablefunctional monomers in non-polar solvents is a means of
assuring the strength and fidelity of the binding. However, forhydrophobic template molecules, such as ethynylestradiol,where the number of functionalities capable of hydrogenbonding is limited, it is to be expected that all potentialattractive forces must be carefully optimized to make a highaffinity imprint. Although the results of other researchers in thisfield can be used to advantage,12,13 optimization still involvesexperimental trial and error.
The binding characteristics of the MIPs P1, P2 and P3 (Table1) were evaluated by determining the amount of polymer thatwas needed to bind 50% of added radio-labeled ethynylestradiolin a chosen solvent. Similar tests were performed at the sametime on the blanks B1, B2 and B3 to evaluate the specificity ofthe MIPs. Preliminary binding assays indicated that the 4-VPpolymer P1 or the 4-VP plus MAA polymer P2 bound thetemplate more strongly and with more specificity than did theMAA polymer P3. Since the phenolic hydrogen on the A-ringhas an acidic character capable of interaction with basicmoieties this was reasonable.14 Although the use of a mixture ofMAA and VP has been shown in certain cases to result inimproved recognition15 we found that the binding strength andthe specificity of polymer P1 were slightly better than that ofP2.
The choice of solvent used for polymerization and forbinding proved to be the factor that was most influential in fine-tuning the performance of the anti-ethynylestradiol MIP.Solvents, such as methylene chloride, THF or acetonitrile wereinitially used as porogens because of their non-polarity and theirability to bring all the polymerization ingredients into solution.In these solvents rebinding was poor or non-existent. Changingthe binding solvent to toluene greatly increased the affinity ofthe template for the imprinted polymers, as well as for theblanks, indicating that the interactions were non-specific. Theaddition of methylene chloride to the toluene increased thespecificity, but decreased the binding strength. These results arean indication that toluene promotes the affinity of the templatetowards the polymer matrix, not unlike the hydrophobicinteractions of proteins in water that are a part of nature’srecognition mechanism.
While ethynylestradiol is sparingly soluble in toluene(approximately 4 mg ml21), further tests revealed that theaddition of 4-VP resulted in total solubility of the requiredamount of template (40 mg ml21). This suggested the formationof a complex between template and functional monomer, aprerequisite for an optimal imprint.16 Indeed, the 4-VP MIPspolymerized in toluene exhibited strong binding characteristicsin the same solvent, as well as excellent specificity. Only 160mg ml21 was sufficient to bind 50% of radio-labeled ethynyles-tradiol, while the corresponding blank polymer B1 did not showany binding at polymer concentrations of up to 2 mg ml21. Thisshows an improvement on the previously reported sensitivityand selectivity of an anti-ethynylestradiol MIP, where 30 mg ofpolymer was required to bind 44% of radioligand, while thesame amount of blank polymer still bound 11%.10
The IC50 value of MIP P1, determined by measuring theamount of non-radiolabeled ethynylestradiol that was requiredto displace 50% of the radiolabeled species, was found to be 68ng ml21 (0.23 mM). As little as 5 ng ml21 of non-radiolabeledethynylestradiol was still able to displace 10% of the radi-oligand, while 2 mg ml21 resulted in 90% displacement.
The same competitive binding assay was used to evaluate thecrossreactivity with a variety of structurally related compounds.The results of the P1 assays and the crossreactivities exhibitedby natural ethynylestradiol antibodies are shown in Fig. 1. Acomparison reveals that the crossreactivities for the naturalantibody and the MIP are similar, except for mestranol. Here theMIP shows no crossreactivity in contrast to the natural antibodywhich is unable to distinguish between ethynylestradiol andmestranol. Thus the specificity of the MIP not only rivals that ofthe antibody, but exceeds it. The MIP’s lack of recognition of
Fig. 1 Structurally related compounds used in competitive binding assays.The crossreactivity values for the MIP (P1) were determined by dividing theIC50 (the amount of non-radiolabeled ethynylestradiol that was required todisplace 50% of the radiolabeled species) of ethynylestradiol by that of thecompeting ligand. No data were available for the recognition of mifepris-tone by the natural antibody.
1416 Analyst, 2000, 125, 1415–1417
Publ
ishe
d on
29
June
200
0. D
ownl
oade
d by
Uni
vers
ity o
f G
lasg
ow L
ibra
ry o
n 17
/08/
2013
09:
32:2
0.
View Article Online
mifepristone, estrone, mestradiol and testosterone was reason-able in view of the structural differences. Mifepristone is toolarge to fit, due to the presence of the 11b-dimethylpyridinemoiety. Estrone has a 17-keto functionality, which is notcapable of hydrogen bonding to the functional monomer andalso flattens the 5-membered D-ring. Mestradiol has a methoxyinstead of a hydroxy group on the A-ring, affecting the size aswell as the electronic environment of this potential point ofrecognition. Similarly, the testosterone molecule is sufficientlydifferent from ethynylestradiol to account for the lack ofbinding. The low crossreactivity of estradiol (2.0%), however,was unexpected. The only difference between the two mole-cules is the presence of the ethynyl group which apart fromrequiring space was not expected to significantly interact withthe polymer matrix since the proton residing on the triple bondis only capable of very weak hydrogen bonding. The shape ofthe 5-membered ring should not be altered to any extent by thesubstitution of the ethynyl group by a hydrogen atom and this isunlikely to account for the diminished binding. However, theabsence or presence of the ethynyl group should affect bindingstrength if a p–p electron stacking interaction occurs betweenthe triple bond and the ring of the 4-VP molecule that isinvolved in the hydrogen bonding to the 17b-hydroxy group.Alternatively, the binding at the D-ring may require two 4-VPmolecules; one to hydrogen bond with the 17b-hydroxy andanother to interact separately with the triple bond. Similarly,these types of stacking forces might contribute to the recogni-tion between 4-VP and the template A-ring. We are currentlyconducting further tests to investigate the validity of thesehypotheses.
In summary, our results show that it is feasible to make ahighly specific molecular imprint of a lipophilic molecule withfew functional groups. It is possible to predict the performanceof the MIP by observing the effect of the functional monomer onthe solubility of the template in the chosen polymerization
solvent. A non-polar rebinding solvent can promote theoptimization of attractive forces, other than hydrogen bonding,between the template and the polymer.
Acknowledgement
The authors thank the Departments of Organic Synthesis andImmunochemistry for their support and the use of theirfacilities.
References
1 M. Kempe and K. Mosbach, J. Chromatogr., A, 1995, 691, 317.2 B. Sellergren, Trends Anal. Chem., 1999, 18, 164.3 J. V. Beach and K. Shea, J. Am. Chem. Soc., 1994, 110, 379.4 A. L. Jenkins, O. M. Uy and G. M. Murray, Anal. Chem., 1999, 71,
373.5 G. Vlatakis, L. I. Andersson, R. Müller and K. Mosbach, Nature,
1993, 361, 645.6 L. I. Andersson, Anal. Chem., 1996, 68, 111.7 M. Kempe and K. Mosbach, Tetrahedron Lett., 1995, 36, 3563.8 A. Guyot and M. Bartholin, Prog. Polym. Sci., 1982, 8, 277.9 M. J. Whitcombe, M. E. Rodriguez, P. Villar and E. N. Vulfson, J.
Am. Chem. Soc., 1995, 117, 7105.10 L. Ye, P. A. G. Cormack and K. Mosbach, Anal. Commun., 1999, 36,
35.11 O. Ramström, L. Ye and K. Mosbach, Chem. Biol., 1996, 3, 471.12 G. Wulff, Angew. Chem. Int. Ed. Engl., 1995, 34, 1811.13 K. Ensing and Th. De Boer, Trends Anal. Chem., 1999, 18, 138.14 S. H. Cheong, S. McNiven, A. Rachkov, R. Levi, K. Yano and I.
Karube, Macromolecules, 1997, 30, 1317.15 O. Ramström, L. I. Andersson and K. Mosbach, J. Org. Chem., 1993,
58, 7562.16 D. Spivak, M. A. Gilmore and K. J. Shea, J. Am. Chem. Soc., 1997,
119, 4388.
Analyst, 2000, 125, 1415–1417 1417
Publ
ishe
d on
29
June
200
0. D
ownl
oade
d by
Uni
vers
ity o
f G
lasg
ow L
ibra
ry o
n 17
/08/
2013
09:
32:2
0.
View Article Online
