Anal. Chem. 2003, 75, 6381-6387
Articles
Application of a Molecularly Imprinted Polymer for the Effective Recognition of Different Anti-Epidermal Growth Factor Receptor Inhibitors LiLi Zhu, Lirong Chen, and Xiaojie Xu*
College of Chemistry and Molecular Engineering, Modern Research Center for Traditional Chinese Medicine, Peking University, Beijing 100871, People’s Republic of China
A molecularly imprinted polymer (MIP) was prepared using (E)-piceatannol, a natural potential anti-epidermal growth factor receptor (EGFR) inhibitor, as the template and 4-vinylpyridine as the functional monomer. The template was isolated from a Chinese traditional Tibetan medicinal herb, Caragana jubata, by a solid-phase extraction procedure. The crude extract of this herb was loaded on the MIP column for the binding test, and two different compounds besides the template itself were specifically recognized by the polymer, which were identified to be butein and quercetin possessing potent antiEGFR tyrosine kinase activities with IC50 values of 10 and 15 µM, respectively. Affinity and selectivity for these inhibitors and another three compounds coexisting with the template in this herb were evaluated in the chromatographic mode. For the first time, the affinity of a molecularly imprinted polymer was investigated to be correlative to the bioactivities of the analytes. The chromatographic behavior of the analytes was consistent with their activity values: the more active inhibitor was retained longer on the MIP. This research work afforded us a new approach for the effective recognition of novel anti-EGFR inhibitors from herbs by using the MIP as the receptor mimic to assay the bioactivities of reserved components, which will be very helpful in the direct separation of lead candidates for anticancer drugs. Epidermal growth factor receptor (EGFR) plays a fundamental role in the regulation of cell growth.1 However, under certain conditions, as a result of overexpression, mutation, or coexpres* Corresponding author. Tel: 86-10-62757456. Fax: 86-10-62751708. E-mail:
[email protected]. (1) Fahad, A. A.; Jinzi, J. W.; Kit, S. L. Biopolymers 1998, 47, 197-223. 10.1021/ac026371a CCC: $25.00 Published on Web 10/28/2003
© 2003 American Chemical Society
sion of the ligand and the receptor, EGFR can become hyperactivated; the result of this is uncontrolled cell proliferation.2-4 EGFR is overexpressed in numerous tumors,5 including those in the brain, lungs, ovaries, bladder, head, and neck. Selective inhibitors of this receptor are, therefore, of increasing interest as mediators of cell growth (e.g., in psoriasis6) and as potential anticancer drugs.7 Traditional isolation or separation of lead compounds from natural resources followed by the bioassay guidance has always played an important role in drug development. However, it is often time-consuming and inefficient for this procedure as a result of the poor affinity and selectivity of conventional materials (e.g. silica gel, polyamide, ion-exchange types, and reversed-phase column). New types of sorbent materials, such as highly cross-linked polymers and chemically modified polymers, are therefore currently being developed to allow more effective extraction and separation. Molecular imprinting is an emerging technique for preparing the polymeric materials possessing highly selective and affinitive properties.8-16 The technique involves complexation in a solution (2) Plowman, G. D.; Ullrich, A.; Shawver, L. K. Drug News Perspect. 1994, 7, 334-337. (3) Salomon, D. S.; Brandt, R.; Ciadiello, F.; Normanno, N. Crit. Rev. Oncol. Haematol. 1995, 19, 183-232. (4) Gullick, W. J. Br. Med. Bull. 1991, 47, 87-98. (5) Woodburn, J. R. Pharmacol. Ther. 1999, 82, 241-250. (6) Levitzki, A. FASEB J. 1992, 6, 3275-3282. (7) Dobrusin, E. M.; Fry, D. W. Annu. Rep. Med. Chem. 1992, 27, 169-178. (8) Wulff, G.; Sarhan, A. Angew. Chem., Int. Ed. Engl. 1972, 11, 341-344. (9) Mosbach, K. Trends Biochem. Sci. 1994, 19, 9-14. (10) Nicholls, I. A.; Andersson, L. I.; Mosbach, K.; Ekberg, B. Trends Biotechnol. 1995, 13, 47-51. (11) Olsen, J.; Martin, P.; Wilson, I. D.; Jones, G. R. Analyst 1999, 124, 467471. (12) Martin, P.; Wilson, I. D.; Jones, G. R. J. Chromatogr., A 2000, 889, 143147. (13) Popp, P.; Paschke, A. Chromatography 1999, 49, 686-690.
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003 6381
Figure 1. The chemical structures of the template and other five analytes.
of a target molecule (template) with functional monomers through either covalent or noncovalent bonds followed by a polymerization reaction with an excess of cross-linkers. Removal of the templates leaves behind specific recognition sites that are complementary to the template in terms of its shape, size, and functionality in the polymer network. Using molecularly imprinted polymers (MIPs) to trap active inhibitors from herbs is proved by our previous work,17,18 and the current study to be straightforward and efficient as a method that is complementary to the traditional isolation. First, MIPs possess an inherent predetermined selectivity to the template and those analytes chemically similar to the target molecule. Despite their low or trace contents in herbs, the target molecule and other related compounds could still be specifically recognized as a result of the significant affinity of MIPs. Second, MIPs have already been identified as stable artificial mimics suitable for the substitution of receptors and antibodies in assays and sensors.19-22 A MIP prepared with a known anti-EGFR inhibitor was applied herein as the receptor alternative for high-throughput screening. Active compounds could be directly recognized from the crude extract of the herb without bioassay guidance, such as ELISA, and thus, plenty of time and outlay in screening out a large number of interferences without bioactivities would be saved. Moreover, the most remarkable advantages of this material are the high stability, low cost, and ease of preparation. In this current work, a molecularly imprinted polymer was prepared with (E)-piceatannol, a potent natural anti-EGFR tyrosine kinase inhibitor, as the template. And using the MIP to simulate the receptor, two other active anti-EGFR inhibitors as well as the template itself were directly separated from this herb, while other compounds with (14) Sellergren, B.; Andersson, L. I. Methods Companion Methods Enzymol. 2000, 22, 92-106. (15) Andersson, L. I. J. Chromatogr., B 2000, 745, 3-13. (16) Owens, P. K.; Karlsson, L.; Lutz, E. S. M.; Andersson, L. I. Trends Anal. Chem. 1999, 18, 146-154. (17) Jianchun, X.; Lili, Z.; Xiaojie, X. Anal. Chem. 2002, 74, 2352-2360. (18) Jianchun, X.; Lili, Z.; Hongpeng, L.; Li, Z.; Chongxi, L.; Xiaojie X. J. Chromatogr., A 2001, 934, 1-11. (19) Vlatakis, G.; Andersson, L. I.; Mu ¨ ller, R.; Mosbach, K. Nature 1993, 361, 645-647. (20) Piletsky, S. A.; Piletskaya, E. V.; Panasyuk, T. L.; El’syaka, A. V.; Levi, R.; Karube, I.; Wulff, G. Macromolecules 1998, 31, 2137-2140. (21) Mirsky, V. M.; Hirsch, T.; Piletsky, S. A.; Wolfbeis, O. S.; Angew. Chem., Int. Ed. Engl. 1999, 38, 1108-1110. (22) Surugiu, I.; Ye, L.; Yilmaz, E.; Dzgoev. A.; Danielsson. B.; Mosbach, K.; Haupt, K. Analyst 1999, 125, 13-16.
6382
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003
low inhibitory activities were removed first with little reservation. This study is the first attempt of the application of a MIP in direct recognition of different anti-EGFR inhibitors from natural resources according to their bioactivities. EXPERIMENTAL SECTION Materials. The template molecule, (E)-piceatannol, and two other anti-EGFR inhibitors, butein and quercetin, as well as three other compounds were isolated from Caragana jubata and identified by their corresponding spectroscopic data by our group. The chemical structures of these analytes are shown in Figure 1. Ethylene glycol dimethacrylate (EDMA), 4-vinylpyridine, and azobis (isobutyronitrile) (AIBN) were purchased from Acros Organics (Geel, Belgium). Before use, the EDMA was distilled under vacuum after being extracted with 10% sodium hydroxide brine and dried over anhydrous magnesium sulfate. AIBN was recrystallized in methanol. Methanol, acetonitrile and tetrahydrofuran were of HPLC grade. Glacial acetic acid was of analytical grade. The water was demineralized and purified by a Millipore system. HPLC Analysis. Evaluation of the selectivity and affinity of the MIP was performed with a HPLC system consisting of a quaternary pump, a variable-wavelength detector, an on-line vacuum degasser, and a 20-µL manual injector (Hewlett-Packard, Palo Alto, CA). Methanol was used as the eluent at a flow rate of 0.5 mL min-1. The analytes were detected by UV adsorption at 254 nm. The sample volume was 20 µL. The column temperature was ambient. Identification of Analytes by Mass Spectrometry. A Mariner electrospray ionization/time-of-flight (ESI-TOF) mass spectrometry (PE Biosystems) was used to identify the analytes. For ESI, the mass spectrometer was operated in the negative mode using the following conditions: spray tip potential, 4800 v; nozzle potential, 100 v; detector voltage, 1950 v; nozzle temperature, 140 °C; quad temperature, 140 °C. All the other parameters were set as default values. Extraction of the Template from the Herb. As described in our previous work,23 using quercetin-imprinted polymer as the sorbent material in a solid-phase extraction (SPE) procedure, (E)piceatannol, a potent anti-EGFR tyrosine kinase inhibitor, was directly separated form the crude extract of C. jubata. In that (23) Lili, Z.; Xiaojie, X. J. Chromatogr., A 2003, 991, 151-158.
Table 1. Chromatographic Parameters for Piceatannol and Other Analytes
K′ (MIP) K′ (NIP) R (MIP) R (NIP) IF
(E)-piceatannol
butein
quercetin
butin
calycosin
benzylphthalide
9.71 2.05 1 1 4.74
5.57 2.07 1.74 0.99 2.69
4.14 2.01 2.35 1.02 2.05
0.49 0.45 19.8 4.55 1.09
0.27 0.33 36.0 6.21 0.82
0.14 0.28 69.4 7.32 0.50
study, quercetin was chosen as the template because it is a known natural anti-EGFR inhibitor24 and is easily obtained. This compound has been reported to exist in a large number of herbs, including a Chinese traditional Tibetan medicinal herb, C. jubata. The primary objective of using the quercetin-imprinted polymer in the SPE procedure was to extract novel active compounds from C. jubata, and piceatannol was isolated as a result of its longer retention time on the polymer than other unrelated components. Then the bioassay evaluation by ELISA showed that piceatannol displayed potent anti-EGFR tyrosine kinase activity, with an IC50 as low as 4.9 µM. Therefore, a large amount of piceatannol was concentrated by the MISPE procedure for further investigation. Preparation of Polymers. The template molecule piceatannol (1 mmol) and functional monomer 4-vinylpyridine (5 mmol) were dissolved into acetonitrile-tetrahydrofuran (3:1, v/v, 10 mL) in a test tube, and the cross-linker EDMA (40 mmol) and the initiator AIBN (42 mg) were added. The tube was sealed under N2 atmosphere, after the mixture had been degassed, it was placed in a 60 °C water bath and incubated for 24 h. The rigid polymers were ground in a mortar and sieved to pass through a 30-µm sieve. Fine particles were removed by decantation in acetone, and the remains were dried under vacuum. The polymer particles were dry-packed in a stainless steel column (150 × 2.1 mm), and the MIP column was washed on-line with methanol; methanol-acetic acid (9:1, v/v); and finally, with methanol until a stable baseline was obtained in order to eliminate the template molecule. As a control, nonimprinted polymer (NIP) in the absence of the template during the polymerization was also prepared and treated in the identical manner. Another control polymer was imprinted with the template molecule calycosin, which displayed low anti-EGFR tyrosine kinase activity. The process for preparation of this polymer was similar to that for the piceatannol-imprinted one, and the functional monomer was also 4-vinylpyridine. This polymer was washed using the same procedures as described above in order to remove the template. Characterization of the Polymers by Chromatographic Mode. HPLC evaluation of the MIP column for the analytes was performed at room temperature. The sample size injected was 20 µL, and the concentration of sample was 0.1 mmol L-1. The capacity factor k′ was calculated using the equation k′ ) (tR - t0)/t0, where tR is the retention time of a sample, and t0, the time to elute the void marker acetone. The imprinting effect (IF) was defined by the equation IF ) k′(MIP)/ k′(NIP), where k′ (MIP) is the capacity factor of the molecularly imprinted polymer, and k′ (NIP) is that of the nonimprinted polymer. The relative retention value (R) was calculated using the equation R ) k′(template)/k′(analyte).
Sample Preparation for the Separation Test. Air-dried roots of C. jubata were extracted by macerating with 85% ethanol at room temperature. The solvent was evaporated in a vacuum and then partitioned between H2O, petroleum ether, CHCl3, EtOAc, and n-butanol, successively. Bioassay-guided results suggested that the CHCl3 and EtOAc extracts were efficient in inhibiting the tyrosine kinase activity of EGFR. Therefore, in the following study, the mixture of these two extracts of C. jubata was investigated. The extract sample (5 mg) was evaporated to dryness in a vacuum centrifuge (RE 52A, Shanghai, China) and then redissolved in 5 mL of methanol to 1 mg mL-1; thereafter, it was diluted to 0.1 mg mL-1 with methanol before loading on the MIP column. EGFR Inhibition Assay by ELISA. To investigate the feasibility of mimicking the receptor with the MIP, the retention behavior of the compounds on the polymer was contrasted to their bioactivities against EGFR. All six analytes were therefore evaluated for their anti-EGFR tyrosine kinase inhibitory activities by ELISA. The procedure of enzyme reactions was carried out as described in the literature.25 EGFR and EGF as well as other reagents (e.g., ATP, antibodies, and the substrate peptide) used in the ELISA were bought from Sigma (U.S.A.). Assays were conducted in a 96-well microtiter plate, and a Bio-Rad plate reader was used to measure the absorbance at 490 nm. For each compound, three independent dose-response curves were constructed, and the IC50 value was computed. The reported value is the average, and the variation was generally (10%.
(24) Graziani, Y.; Erickson, E.; Erickson, R. L. Eur. J. Biochem. 1983, 135, 583588.
(25) David, W. F.; Alan, J. K.; Amy, M.; Linda, A. A.; James, M. N.; Wilbur, R. L.; Richard, W. C.; Alexander, J. B. Science 1994, 265, 1093-1095.
RESULTS AND DISCUSSION Binding Characteristics of the Polymers. During the development of imprinted polymers for binding assays or sensors, it is convenient to initially use a well-established method, such as chromatography or equilibrium binding assays, with the radiolabeled analyte to determine the affinity and performance of the polymers. In this way, any possible doubts as to whether the polymer is actually imprinted, which may arise when the polymer is used under nonoptimal conditions dictated by the final application, may be eliminated. In our model system, a polymer imprinted with (E)-piceatannol was prepared. The selectivity test of the polymer was carried out for the main compounds coexistent with the template in the extract of this herb. The relative retention values (R) and the imprinting efficiency values (IF) are listed in Table 1. The retention of piceatannol on the MIP (K′ ) 9.71) compared to that on the NIP (K′ ) 2.05) clearly indicated the presence of templated binding sites in the imprinted polymer. The parameters showed in Table 1 also demonstrated that the polymer exhibited stronger affinity to quercetin and butein than the other three compounds.
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003
6383
Figure 2. The chromatographic trace of the herb extract on the column of the piceatannol-imprinted polymer. The sample volume and concentration injected were 20 µL and 0.1 mmol L-1, the mobile phase is methanol, and the flow-rate is 0.5 mL min-1.
Figure 3. The mass spectra of the eluted fraction corresponding to peak a in Figure 2.
The results of the selectivity test may give aspects of the molecular recognition mechanism, which was partly reflected in the relative retention values (R). The higher relative retention values for butin, calycosin, and benzylphthalide showed pronounced nonspecific interactions occurring between the samples and the polymer matrix. The chemical memory (spatial arrangement of the complementary functionality) of the polymer network may partially account for the weaker retention of these three compounds: compared to butein and quercetin, butin and calycosin possess fewer functional groups capable of hydrogen bonding interaction with the polymer, whereas the fewest functional groups of benzylphthalide as well as the smallest spatial structure of this molecule lead to its poorest matching to the microcavities of polymers based on the shape and functional groups formed by imprinting. Between quercetin and butein, the chemical structure of the latter seems to be more similar to the template, piceatannol, which may explain the slightly stronger retention of this compound on the polymer. As the solvent acetonitrile was adopted to investigate the selectivity characters of the polymer, the retention difference between the analytes was more obvious. But the template was observed to specifically retain on the MIP column too strongly to be washed out until after 90 min. When using methanol as the solvent to perform the binding test, the results were satisfactory despite that the chromatographic peaks appeared broad and asymmetric. One important factor contributing to peak broadening of the MIP column may be the heterogeneous character of the polymer. It was supposed that like polyclonal antibodies, the imprinted polymer contains a heterogeneous population of sites. Low affinity sites predominate in the MIP, but only a relatively small fraction have a very high affinity.26,27 Another cause of poor column efficiency is the inhomogeneous particle size. The (26) Wulff, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1812-1832. (27) Kempe, M.; Mosbach, K. Anal. Lett. 1991, 24, 1137-1145.
6384
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003
preparation of uniformly sized particles for the stationary phase will improve column efficiency remarkably. As the more polar and protic solvent, methanol-acetic acid (9:1, v/v) was used. A symmetrical chromatographic peak was obtained, but the imprinting effect displayed was much weaker. It is obvious that the selection of an appropriate solvent for the affinity test of MIP is important. Application of the MIP To Separate Other Anti-EGFR Inhibitors from the Herb. Most of the extraction by MIPs was investigated in the off-line mode; in other words, the high selectivity and affinity of the MIPs were used as a cartridge for matrix discrimination followed by analysis and identification. In this work, the sample of C. jubata crude extract was directly loaded onto the MIP column for the following separation by off-line mode. The chromatographic trace of the crude extract on the MIP column is shown in Figure 2. From the chromatographic trace, it can be observed that three main components, a, b, and c were retained longer than others, and a large number of compounds were eluted early in the elution with much weaker reservation. The washing-out fractions corresponding to the peak tops of a, b ,and c were collected and evaporated to dryness under vacuum, and then the three fractions were detected with an ESI-TOF mass spectrometry after being redissolved in 200 µL of methanol. The resulting traces are shown in Figures 3-5. Herein, the mass spectrometry was chosen to identify the analytes because the procedure of MS detection is more convenient and faster than a chromatographic experiment since the time for elution of the template from the MIP column was as long as ∼75 min. In addition, this analytical method possesses higher response and sensitivity than other systems, such as a UV/vis detector. The main three components, a, b, and c, that were isolated were shown by mass measurement to be quercetin ([M - H]-
Figure 4. The mass spectra of the eluted fraction corresponding to peak b in Figure 2.
Figure 5. The mass spectra of the eluted fraction corresponding to peak c in Figure 2.
ion at 301.04), butein ([M - H]- ion at 271.07) and piceatannol ([M - H]- ion at 243.07), respectively. From the mass spectra, it could be found that there was no other molecular peak except the three analytes themselves (the corresponding dimmer of the main molecular ion was also displayed in Figures 3 and 5), which meant that each inhibitor could be purified by the separation of the MIP column as these three peaks were collected correspondingly. The template molecule was retained longest on the polymer and could be discriminated from the matrix interferences of the crude extract of the herb. The other two anti-EGFR inhibitors, butein and quercetin, were also separated from other abundant interferences resulting from their stronger retention on the column. But the chromatographic peaks of these two compounds could not be absolutely detached from each other (the baselines of the chromatographic peaks for these two compounds were not separated from each other completely), since there was no significant difference between their retention characteristics, which was consistent with the results described in the above section. Although acetonitrile was adopted as the solvent, these two inhibitors could be separated completely from each other with the better selectivity of the MIP in an eluent of lower polarity, and the retention time for the analytes was also prolonged. The chromatographic behavior of the MIP column demonstrates the potential application of this polymer in the recognition and concentration of novel active compounds. Evaluation of the MIP as the Receptor Binding Mimics. As described above, the main purpose of the current study was to investigate the application of MIP as the receptor alternative for direct separation of active compounds from the herb, and the efficiency of the polymer cavities for substituting the binding pocket of the receptor should be taken into account. It is wellknown that the binding pocket of a receptor could recognize different inhibitors according to their bioactivities: the more active
Table 2. Comparison of the IF Values for Six Analytes to Their Bioactivities compd
imprinting effect (IF)
inhibitory activity (IC50)
(E)-piceatannol butein quercetin butin calycosin benzylphthalide
4.74 2.69 2.05 1.09 0.82 0.50
4.9 µM 10 µM 15 µM 1.1 mM 1.2 mM 59.0 mM
molecule could bind to the pocket more strongly. The retention effect of the polymer expressed by the IF value for each compound was thus contrasted to their inhibitory activity values (IC50), and the results are listed in Table 2. The larger IF value reveals the stronger affinity of the polymer to the analyte; the smaller IC50 value corresponds to the higher bioactivity of the compound against the enzyme. From the inhibitory data it could be concluded that the first three compounds possessed potent bioactivities against EGFR tyrosine kinase, whereas the last three showed rather low activities; especially, the compound benzylphthalide was regarded as bearing no activity, with IC50 as high as 59.0 mM. Accordingly, only the first three analytes were retained on the polymer, and the others were washed out early in the elution with little preservation. In addition, the first three molecules belonging to different types of chemical structures were effectively recognized by the MIP according to their bioactivities: the template with the highest activity (IC50 ) 4.9 µM) was bound to the polymer most strongly; butein (IF ) 2.69) was retained a little longer than quercetin (IF ) 2.05), which was consistent with its slightly higher inhibitory activity. In other words, the polymer imprinted with a known anti-EGFR inhibitor, piceatannol, could recognize different Analytical Chemistry, Vol. 75, No. 23, December 1, 2003
6385
Table 3. Chromatographic Parameters for Quercetin and Another Two Inhibitors on the Quercetin-Imprinted Polymer
K′ (MIP) K′ (NIP) R (MIP) R (NIP) IF inhibitory activity (IC50)
quercetin
butein
(E)-piceatannol
6.98 2.00 1 1 3.49 15 µM
2.07 1.28 3.37 1.56 1.62 10 µM
1.96 1.43 3.56 1.40 1.37 4.9 µM
Figure 6. Comparison of two MIPs in their recognition abilities for different anti-EGFR inhibitors.
anti-EGFR inhibitors according to their bioactivities: the more active compound possessed lager IF value, which meant that it was retained on the polymer longer and, therefore, could be separated from the other components. This work demonstrated that it is feasible to use a MIP for separating active inhibitors directly from the crude extract of the herb, which would be very helpful in discovering lead compounds or drug candidates. Comparison of the Efficiency of Three MIPs for Recognizing Different Inhibitors. We have proposed that the unique application of MIPs studied in this work was to simulate the receptor for effective recognition of different anti-EGFR inhibitors. Three MIPs prepared with two potent inhibitors (piceatannol and quercetin) as well as another template molecule with low inhibitory activity (calycosin) were thus compared for their efficiency. In our previous study,23 the selectivity of the quercetin-imprinted polymer for these three inhibitors, piceatannol, butein, and quercetin, was evaluated in the chromatographic mode. The results (as listed in Table 3) showed that the template, quercetin, was retained longest on this polymer, and butein was retained longer than piceatannol, despite its lower bioactivity. Since the two templates (piceatannol and quercetin) were both inhibitors against EGFR, the effects of these two MIPs in recognizing the three main anti-EGFR inhibitors existed in C. jubata are contrasted in Figure 6, in which the factor -ln(IC50) was adopted to indicate the bioactivity of the analytes. The bioactivity was higher because this factor was larger. From this comparison, it could be found that the piceatannol-imprinted polymer was able to recognize three inhibitors according to their bioactivities, whereras the polymer imprinted by quercetin could not, although it extracted these active compounds from the herb through a SPE procedure. In other words, the corresponding polymer imprinted by the more active inhibitor, (E)-piceatannol, could mimic the receptor more effectively and, thus, recognize the inhibitors more exactly. This conclusion is reasonable con6386 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003
Figure 7. The chromatographic trace for the herb extract on the HPLC column of calycosin-imprinted polymer. The sample volume and concentration injected were 20 µL and 0.1 mmol L-1. The mobile phase consisted of solvent A (methanol) and solvent B (methanolacetic acid, 9:1, v/v) with the following gradient: 0 min, A, 100%; 60 min, B, 100%; 120 min, B, 100%. The flow rate is 0.5 mL min-1.)
sidering that the resulting cavities in the polymer network will be more comparable to the binding pocket of the receptor in terms of the shape, size, and arrangement of functional groups, since the template with higher inhibitory activity could bind to the pocket more efficiently. Moreover, another control MIP prepared with a nonretaining compound such as calycosin was also evaluated for its efficiency for recognizing different inhibitors present in C. jubata. And the results showed that the retention was lost for piceatannol and improved for calycosin. Because this polymer was utilized in a column for HPLC analysis, those three analytes displaying potent inhibitory activities against EGFR tyrosine kinase, including piceatannol, butein, and quercetin, were washed out early in the elution process with weak retention on the polymer, while two other compounds (benzylphthalide and calycosin) possessing little binding affinities to the piceatannol-imprinted polymer were retained on this polymer. Especially for the template molecule calycosin, the retention effect was improved greatly. The chromatographic trace for the HPLC analysis is shown in Figure 7. Also these analytes were identified by ESI-TOFMS, as described above. The analytes eluted in the asymmetric peak a were piceatannol, butein, and quercetin; peak b was identified as benzylphthalide, and c was the template, calycosin. Obviously, the polymer imprinted with piceatannol that displayed the highest potent inhibitory activity against EGFR could mimic the receptor for recognizing different inhibitors most effectively, but the recognition of the one prepared with the template molecule calycosin that showed low anti-EGFR activity was unsuccessful. It could be expected that using one representative inhibitor possessing the highest activity as the template to synthesize the corresponding polymer, the efficiency may be improved further and, thus, could be applied to more complicated systems, which is another program in process in our group. CONCLUSIONS Taking the two important characteristics, one steric memory (size and shape), and the other chemical memory (spatial arrangement of the complementary functionality), of the MIP into account, the application of this material in trapping different types of active compounds from the herb and recognizing the inhibitors according to their bioactivities, as it has turned out in the present work, has proven to be straightforward and feasible. Certainly, the shape of the cavity (van der Waals complementary) and correct
arrangement of the functional groups within the cavity play an important role in molecular recognition. From a structure-activity relationship (SAR) point of view, most inhibitors against the same receptor often adopt a similar or even common binding model and, therefore, are inclined to possess similar structures in terms of size, shape, and functional groups. For the binding pocket of one enzyme, only those molecules belonging to some size (too large could not enter into the pocket and too small could not occupy the binding sites efficiently) may enter easily, maintain stability, and thus, be potential inhibitors. Moreover, the possible interactions between the enzyme and the inhibitors, such as a hydrogen bond, ionic interactions, and hydrophobic effects, are also utilized in molecular imprinting. To date, molecularly imprinted materials have been used in a range of applications. They are increasingly used as selective supports in liquid chromatography,28 thin-layer chromatography,29 capillary electrophoresis30 and solid-phase extraction.31 Most of the range of analytes mainly focuses on the template molecule itself, the chiral isomer of the template, or the congener compounds different in several groups from the template, whereas using MIP to extract different types of compounds with a biological function (e.g., bioactivities) similar to the template from herb has rarely been reported. This paper is the first attempt to use a polymer imprinted by an active inhibitor against EGFR to mimic the receptor in recognizing different inhibitors according to their bioactivities. The effective separation of different inhibitors from the herb by the piceatannol-imprinted polymer in the presence of a number (28) Siemann, M.; Andersson, L. I.; Mosbach, K. J. Antibiotics 1997, 50, 8991. (29) Kriz, D.; Kriz, C. B.; Andersson, L. I.; Mosbach, K. Anal. Chem. 1994, 66, 2636-2639. (30) Nillson, K.; Lindell, J.; Norrloew, O.; Sellergren, B. J. Chromatogr., A 1994, 680, 57-61. (31) Sellergren, B. Trends Anal. Chem. 1999, 18, 164-174.
of interferences afforded us a useful approach for the discovery of anticancer lead compounds, which will be very beneficial in drug development, especially when there are no feasible methods for bioactivity evaluation. The MIP could commendably simulate the receptor for further bioassay. Moreover, because of the inherent advantages of MIPs, for example, physical robustness; resistance to elevated temperatures and pressures; and inertness toward acids, bases, metals ions; and organic solvents, this material can be well exploited in a large number of samples, including the crude extract of herbs. Therefore, the effective components of herbs could be trapped and analyzed directly using MIPs as the sorbent material using only a simple pretreatment. On one hand, the known active compounds with trace contents in the crude extract of herbs could be concentrated using the MISPE procedure; on the other hand, the polymer imprinted with known inhibitors could trap other novel lead compounds with relative chemical structures to the template directly from the herb. Therefore, using MIPs to trap efficient components from natural resources becomes fast and effective as a cogent supplementary means to the traditional isolation. Although there were only six compounds that were recognized from a traditional Chinese medicine in this study, it is believed that this method could be applied to more complicated and more extensive systems, such as synthetic combinatorial libraries or other natural resources. It is anticipated that using a MIP prepared with a known lead compound that is expensive or difficult to obtain could trap different candidates that may be very simple or cost-efficient and that have similar bioactivities.
Received for review November 30, 2002. Accepted April 30, 2003. AC026371A
Analytical Chemistry, Vol. 75, No. 23, December 1, 2003
6387
