Molecularly Imprinted Solid Phase Extraction of Atrazine from Beef

AC Research

Articles Anal. Chem. 1997, 69, 803-808

Molecularly Imprinted Solid Phase Extraction of Atrazine from Beef Liver Extracts Mark T. Muldoon and Larry H. Stanker*

Food Animal Protection Research Laboratory, USDA-ARS, 2881 F & B Road, College Station, Texas 77845

Molecularly imprinted polymers were used as specific binding matrices for the solid phase extraction and cleanup of biological sample extracts. To demonstrate this, an anti-atrazine polymer was used to cleanup organic extracts of beef liver. Atrazine retention on the columns was greatest in chloroform. The binding capacity of the polymer in chloroform was 19 µmol of atrazine per gram. Purified and unpurified beef liver extracts were analyzed by both reversed-phase HPLC and ELISA. The use of molecularly imprinted solid phase extraction (MISPE) improved the accuracy and precision of the HPLC method and lowered the limit of detection (0.005 ppm). Atrazine recovery as determined by HPLC from beef liver homogenates spiked to levels from 0.005 to 0.5 ppm averaged 88.7% following MISPE and 60.9% for the unpurified extracts. Atrazine recovery as determined by ELISA averaged 92.8% following MISPE and 79.6% for the unpurified extracts. Crude tissue sample extracts interfered with both the HPLC and ELISA methods. However, the use of MISPE allowed for the rapid analysis of complex biological matrices using either method at the tolerance level of 0.02 ppm in meat products. The application of molecular imprinting technology for solid phase extraction is a new approach for the analysis of highly lipophilic low molecular weight contaminants.

Molecularly imprinted polymers (MIPs) are made by synthesizing highly cross-linked polymers in the presence of a “print” molecule. After removal of the print molecule, the polymer can be used as a selective binding medium for the print molecule or structurally related compounds. The mechanisms by which these polymers specifically bind the print molecule and related ligands are attributed to the formation of functional groups in a specific arrangment within the polymer that corresponds to the print

molecule and to the presence of shape-selective cavities.1,2 Early studies by Wulff et al. showed that MIPs could be used as solid phases in high-performance liquid chromatography (HPLC) to resolve enantiomeric mixtures of sugars.3 This technique has since been applied to the resolution of many other classes of low molecular weight compounds using several types of functionalized polymers.4 Recently, MIPs have been used in other applications of analytical chemistry. They have been used as solid phases for thin-layer chromatography (TLC)5 and in receptor assays such as the molecularly imprinted sorbent binding assay (MIA).6-9 This latter technique was similiar to radioimmunoassay; however, MIPs substituted for antibodies as the specific binding reagent. Also, MIPs have found application as specific receptors in sensors designed for in situ analysis.10,11 One of the most important attributes of MIPs is their ability to function in organic solvents. In our previous report, acetonitrile was used as the assay solvent for the development of a MIA for the s-triazine herbicide atrazine.7 Other assay solvents used in conjunction with MIPs have included methanol, chloroform, heptane, and toluene.8,12-14 In contrast to these synthetic receptors, most biological receptors, such as antibodies and enzymes, (1) Shea, K. J.; Sasaki, D. Y. J. Am. Chem. Soc. 1991, 113, 4109-4120. (2) Wulff, G.; Schauhoff, S. J. Org. Chem. 1991, 56, 395-400. (3) Wulff, G.; Vesper, W.; Grobe-Einsler, R.; Sarhan, A. Makromol. Chem. 1977, 178, 2799-2818. (4) Wulff, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1812-1832. (5) Kriz, D.; Kriz, C. B.; Andersson, L. I.; Mosbach, K. Anal. Chem. 1994, 66, 2636-2639. (6) Vlatakis, G.; Andersson, L. I.; Muller, R.; Mosbach, K. Nature 1993, 361, 645-647. (7) Muldoon, M. T.; Stanker, L. H. J. Agric. Food Chem. 1995, 43, 1424-1427. (8) Siemann, M.; Andersson, L. I.; Mosbach, K. J. Agric. Food Chem. 1996, 44, 141-145. (9) Andersson, L. I. Anal. Chem. 1996, 68, 111-117. (10) Piletsky, S. A.; Parhometz, Y. P.; Lavryk, N. V.; Panasyuk, T. L.; El’skaya, A.V. Sens. Actuators B 1994, 18-19, 629-631. (11) Kriz, D.; Mosbach, K. Anal. Chim. Acta 1995, 300, 71-75. (12) Sellergren, B.; Ekberg, B.; Mosbach, K. J. Chromatogr. 1985, 347, 1-10. (13) Kempe, M.; Mosbach, K. J. Chromatogr. A 1991, 694, 3-13.

S0003-2700(96)00464-7 This article not subject to U.S. Copyright. Publ. 1997 Am. Chem. Soc.

Analytical Chemistry, Vol. 69, No. 5, March 1, 1997 803

can function in the presence of only low amounts (below 20% v/v) of some organic solvents (e.g., methanol and acetonitrile). As a result, techniques that utilize biological receptors, such as immunoassays, must be carried out in predominantly aqueous-based solvents. Due to their compatability with organic solvents, MIPs have many potential applications for the analysis of highly lipophilic compounds, such as many pesticides and other environmental pollutants. Recently, MIPs were developed for the herbicide atrazine by several independant groups.7,8,10,15 Most applications of MIPs have concerned the analysis of compounds in either simple solvents or buffers. However, in one study, MIAs for the therapeutic drugs theophylline and diazepam were developed in which human serum was analyzed for drug levels.6 In this article, we describe the use of MIPs as the specific binding matrix for solid phase extraction of biological sample extracts. As an example, we use an atrazine MIP as the solid phase in sample extraction columns. These columns were used to purify tissue sample extracts prior to analysis and quantification by reversed-phase HPLC and enzyme-linked immunosorbent assay (ELISA). This approach represents a rapid, cost-effective method for residue analysis. EXPERIMENTAL SECTION Chemicals. Atrazine [2-chloro-4-(ethylamino)-6-(isopropylamino)-1,3,5-triazine] (>97% purity) was a gift from Ciba-Geigy (Greensboro, NC). Radiolabeled atrazine (ring-UL-14C, 18.6 mCi/ mmol) was purchased from Sigma (St. Louis, MO). Omnisolv acetonitrile, acetone, chloroform, dimethylformamide, and methanol were from EM Science (Gibbstown, NJ), and HPLC grade acetic acid and phosphoric acid (85% v/v) were from Fisher Scientific (Fair Lawn, NJ). High-purity water was obtained from a MILLI-RO4 purification system (Millipore Corp., Bedford, MA). Equipment. The multicolumn vacuum manifold was from Supelco, Inc. (Bellefonte, PA). The HPLC system was a Dionex Bio-LC (Dionex, Corp., Sunnyvale, CA), equipped with a UVvisible variable-wavelength detector (225 nm monitored). The reversed-phase columns were 4.6 mm × 25 cm LC-18 (5 µm) or LC-18-DB (deactivated) (5 µm) from Supelco, Inc. The mobile phase was 50% acetonitrile in 0.085% (v/v) phosphoric acid in water (pH 2) at a flow rate of 1.5 mL/min. Microtiter plates were Nunc Immunoplate II Maxisorp (Nunc, Roskilde, Denmark). Microtiter plate optical density (OD) measurements were made using a BioRad Model 3550 microplate reader (Bio-Rad Laboratories, Richmond, CA). Data were collected using a Macintosh II computer and Reader Driver 1.0 and Microplate Manager 1.0 software (BioRad). Other calculations utilized Excel spreadsheet software (Microsoft Corp., Redmond, WA). Polymer Synthesis. Details of polymer synthesis were previously described.7 Briefly, molar ratios of print molecule (atrazine):functional monomer (methacrylic acid, MAA):comonomer (ethylene glycol dimethyl methacrylate, EGDMA) were 1:4: 20. Polymerization was carried out at 60 °C for 24 h with 2,2′azobis[isobutyronitrile] (AIBN) as the initiator. Following grinding and print molecule extraction, polymer particles less than 25 µm were collected and sedimented in acetonitrile prior to use. Synthesis and handling of the polymers were carried out in a fume hood to avoid contact with solvents and dust particles. (14) Ramstrom, O.; Andersson, L. I.; Mosbach, K. J. Org. Chem. 1993, 58, 75627564. (15) Matsui, J.; Miyoshi, Y.; Doblhoff-Dier, O.; Takehuchi, T. Anal. Chem. 1995, 67, 4404-4408.

804

Analytical Chemistry, Vol. 69, No. 5, March 1, 1997

Enzyme-Linked Immunosorbent Assay (ELISA). The atrazine ELISA was a direct competition format using anti-atrazine monoclonal antibody (AM7B2)16 immobilized on the microtiter plate and a simazine-alkaline phosphatase tracer. Details of the procedure were previously described.17 The assay buffer was phosphate-buffered saline (pH 7) containing 0.05% (v/v) Tween 20. The monoclonal antibody was a gift from A. E. Karu, Hybridoma Facility, University of California, Berkeley, CA. Effect of Solvents on Atrazine Binding to the Molecularly Imprinted Polymer. Atrazine (10 µg) in either chloroform, acetonitrile, dimethylformamide, or water (1.0 mL) was added to a 1.5 mL polypropylene tube containing 50 mg of either imprinted or nonimprinted (synthesized without atrazine present) polymer. The tubes were vortexed and incubated 1 h at room temperature with agitation. Following centrifugation, the supernatants were analyzed by reversed-phase HPLC. The amount of atrazine added that was bound to the polymer was calculated as a residual from the amount measured in the supernatant according to the following equation:

% bound ) [(amount added - amount measured in supernatant)/amount added] × 100 (1) Molecularly Imprinted Sorbent Solid Phase Extraction (MISPE). Empty polypropylene syringe barrels (3.0 mL) were connected to a multicolumn vacuum manifold operated under negative pressure. These were packed with 250 mg of atrazineimprinted or nonimprinted (control) polymer slurried in acetonitrile. Prior to and between uses, the columns were washed successively with 10% (v/v) acetic acid/acetonitrile (10 mL), acetonitrile (20 mL), and chloroform (20 mL). The columns were preequilibrated with the application solvent prior to applying the sample. Atrazine binding and elution conditions were determined using several conditions. Elution Method 1. One microgram of atrazine in acetonitrile (0.33 mL) was applied to the column. The column was rinsed with 30 aliquots (0.33 mL) of acetonitrile. Column fractions were collected (0.33 mL) and analyzed by reversed-phase HPLC. Elution Method 2. One microgram of atrazine in chloroform (0.5 mL) was applied to the column. The column was rinsed with 20 aliquots (0.5 mL) of chloroform and 10 aliquots (1.0 mL) of 90% acetonitrile/water. Column fractions were collected (0.5 mL for chloroform, 1.0 mL for 90% acetonitrile/water), evaporated to dryness, reconstituted in acetonitrile (0.67 mL), and analyzed by reversed-phase HPLC. Column Binding Capacity. Three separate batches of imprinted and nonimprinted polymer were used. Three columns (one per batch) were packed with either imprinted or nonimprinted polymer (150 mg). After preconditioning, 10 mg of atrazine in chloroform (1 mL) containing 0.1 µCi [14C]atrazine was applied to the columns. The columns were rinsed with chloroform (30 mL) to remove all unbound atrazine, as determined by liquid scintillation counting (LSC) of the eluents. The columns were (16) Karu, A. E.; Harrison, R. O.; Schmidt, D. J.; Clarkson, C. E.; Grassman, J.; Goodrow, M. H.; Lucas, A; Hammock, B. D.; White, R. J.; Van Emon, J. M. In Immunoassays for Trace Chemical Analysis: Monitoring Toxic Chemicals in Humans, Foods, and Environment; Vanderlaan, M., Stanker, L. H., Watkins, B. E., Roberts, D. W., Eds.; ACS Symposium Series 451; American Chemical Society: Washington, DC, 1991; pp 59-77. (17) Muldoon, M. T.; Nelson, J. O J. Agric. Food Chem. 1994, 42, 1686-1692.

air-dried, and triplicate aliquots of polymer (20-30 mg) from each column were measured by LSC to determine the amount of bound atrazine. Safety: These experiments were carried out in a fume hood designated for use with radioisotopes, and proper measures were taken to prevent exposure to solvents and radioactivity. Extraction of Beef Liver. Ten grams of beef liver homogenate was extracted for 1 h with chloroform (30 mL) on a wristaction shaker. The samples were centrifuged, and a portion (20 mL) of the extract was removed and either used directly (0.33× extract) or evaporated to dryness and reconstituted in chloroform (0.67 mL, 10× extract). Evaluation of Atrazine Elution Conditions from MISPE Column for Atrazine-Spiked Beef Liver Extracts. Various atrazine elution conditions were investigated using spiked beef liver extract in order to determine the elution conditions that purified atrazine from contaminants in the tissue sample extract as determined by reversed-phase HPLC. Elution Method 2B. This method was similiar to elution method 2 with some modifications. Tissue extract (0.33 mL, 10× extract) from a 1.0 ppm atrazine-spiked liver homogenate was applied to the column. The column was rinsed with chloroform (10 mL) and eluted with 90% acetonitrile/water (10 mL). The combined chloroform and acetonitrile fractions were each evaporated to dryness, reconstituted in acetonitrile (0.33 mL), and analyzed by reversed-phase HPLC. Elution Method 3. Tissue extract (0.67 mL, 10× extract) from a 1.0 ppm atrazine-spiked beef liver homogenate was applied to the column. The column was rinsed successively with chloroform (5 mL), five aliquots (5 mL) each of 3% acetonitrile/ chloroform, 7% acetonitrile/chloroform, and 50% acetonitrile/ chloroform, and three aliquots (5 mL) of 10% acetic acid/ acetonitrile. Column fractions were collected (5 mL), evaporated to dryness, reconstituted in acetonitrile (0.67 mL), and analyzed by reversed-phase HPLC. Elution Method 4. Tissue sample extract (0.67 mL, 10× extract) from a 1.0 ppm atrazine-spiked beef liver homogenate was applied to the column. The column was rinsed successively with chloroform (5 mL), 10 aliquots (5 mL) of 3% acetonitrile/ chloroform, and five aliquots (5 mL) of 50% acetonitrile/ chloroform. Column fractions were collected (5 mL), evaporated to dryness, reconstituted in acetonitrile (0.67 mL), and analyzed by reversed-phase HPLC. Elution Method 5. Tissue sample extract (20 mL of 0.33× extract or 0.67 mL of 10× extract) from a 0.1 ppm atrazine-spiked beef liver homogenate was applied to the column. The column was rinsed with chloroform (10 mL for 0.33× extract or 5 mL for 10× extract) and eluted with 10% (v/v) acetic acid in acetonitrile (10 mL). The 10% (v/v) acetic acid in acetonitrile fractions were pooled, evaporated to dryness, reconstituted in acetonitrile (0.67 mL), and analyzed by reversed-phase HPLC using the deactivated C18 column. MISPE and Analysis of Atrazine from Spiked Beef Liver Homogenates by HPLC and ELISA. Aliquots (10 g) of beef liver homogenate were spiked with atrazine to levels of 0.5, 0.1, 0.02, 0.005, and 0.001 ppm. Five replicate sets of spiked and unspiked homogenate were extracted as described above. Five milliliters of the crude extract was evaporated to dryness and reconstituted in acetonitrile (0.167 mL, unpurified). Twenty milliliters of the extract was treated as described in elution method 5 for the 0.33× extract (MISPE-purified). For the ELISA, aliquots

Table 1. Effect of Solvents on Atrazine Binding to Imprinted and Nonimprinted Polymersa % bound solvent

EPDNb

imprinted polymer

chloroform acetonitrile dimethylformamide water

0.0 14.1 26.6 18.0

76.4 30.0 0.0 99.6

nonimprinted polymer 30.7 14.0 0.0 95.0

a Atrazine (10 µg) in chloroform, acetonitrile, dimethylformamide, or water (1 mL) was incubated with imprinted or nonimprinted polymer (50 mg) at room temperature for 1 h with agitation. Following centrifugation, supernatants were analyzed by reversed-phase HPLC. The percent atrazine bound to the polymer was calculated using eq 1. b Electron pair donor number (donicity).19

(50 µL) of the acetonitrile-reconstituted extracts were removed, evaporated to dryness, and reconstituted in assay buffer (0.5 mL, 1× extract). Two-fold dilutions of the samples were made in assay buffer on the microtiter plate and analyzed by ELISA. RESULTS AND DISCUSSION Atrazine Binding Studies. Analyte binding properties of molecularly imprinted sorbents are influenced by the type of solvent, or porogen, used in polymer synthesis and the solvent used in the particular application of the MIP.18 To determine the specific binding properties of our MIP, imprinted and nonimprinted polymers were equilibrated with atrazine in chloroform, acetonitrile, dimethylformamide, or water. Table 1 shows the results of this experiment. In organic solvents, an increase in specific binding was consistent with a decrease in the electron pair donicity (hydrogen bond acceptivity) of the solvent. It was previously shown that atrazine can hydrogen bond with MAA.15 Atrazine can also form strong hydrogen bond complexes with amides through cooperative, multipoint attachment.20 In addition, these bonds can be disrupted by polar solvents.21 Likewise, it is possible that, in acetonitrile and dimethylformamide, atrazine hydrogen bonds with the solvent and decreases its interaction with the MIP. The high nonspecific binding we observed with water was most likely a partitioning effect of the lipophilic atrazine nonspecifically into the organic polymer matrix. High nonspecific binding in aqueous solvents was recently reported with a different atrazine MIP.8 Determination of Atrazine Binding and Elution Conditions for Molecularly Imprinted Sorbent Solid Phase Extraction (MISPE) Columns. Equilibrium binding experiments showed that atrazine specifically bound to the MIP in either chloroform or acetonitrile. Therefore, these solvents were studied using the MIP in a solid phase extraction column operated under vacuum. When acetonitrile was the solvent (Figure 1), atrazine eluted from the column in a broad band, which represented 70% of the applied material. However, when chloroform was the solvent (Figure 1), all of the atrazine was retained on the column. Atrazine was eluted from the column in 90% acetonitrile/water. In addition, we found (18) O’Shannessey, D. J.; Ekberg, B.; Andersson, L. I.; Mosbach, K. J. Chromatogr. 1989, 470, 391-399. (19) Reichhardt, C. Solvent Effects in Organic Chemistry; Verlag: Weinheim, 1978; Chapter 2. (20) Welhouse, G. J.; Bleam, W. F. Environ. Sci. Technol. 1990, 27, 500-505. (21) Sellergren, B.; Lepisto, M.; Mosbach, K. J. Am. Chem. Soc. 1988, 110, 58535860.

Analytical Chemistry, Vol. 69, No. 5, March 1, 1997

805

Figure 1. Atrazine binding and elution from a MISPE column. Either acetonitrile (elution method 1, 9) or chloroform (elution method 2, O) was used as the binding and washing solvents. Following application and washing in chloroform, atrazine was eluted with 90% (v/v) acetonitrile/water. The percentage indicated in parentheses is the amount of the applied atrazine (1.0 µg) recovered in the corresponding peak.

that atrazine was bound to the column after extensive chloroform washing (100 mL). These results demonstrated that atrazine was tightly but reversibly bound to the MISPE column in chloroform. The binding capacities of imprinted and nonimprinted polymers were determined using excess radiolabeled atrazine applied to columns containing polymers derived from 3 separate syntheses. The amounts of atrazine bound to the imprinted and nonimprinted polymers were 19.0 ( 5.0 and 9.8 ( 0.4 µmol of atrazine per gram of polymer, respectively. Specific atrazine binding was nearly 2 times greater than nonspecific binding. Similiar results were found when the polymers were used in an MIA format.7 MISPE and Analysis of Atrazine from Spiked Beef Liver Homogenates by HPLC. Crude chloroform tissue extracts were less complex than either acetonitrile or ethyl acetate extracts as determined by HPLC (monitored at 225 nm; data not shown). Chloroform was also compatible with the MIP and, therefore, was used as the extraction solvent for subsequent experiments. Initially, elution method 2B was evaluated for purifying 10× extracts from beef liver homogenates but was unsuccessful in removing the major interfering components in the sample. Therefore, elution method 3 was used to determine whether atrazine could be enriched in the sample by changing the polarity of the solvent and flushing contaminants from the column. Figure 2 shows the results of this experiment. Using a gradient of increasing polarity from chloroform to 10% acetic acid/acetonitrile, increasing total amounts of atrazine were eluted with each solvent. However, atrazine was significantly enriched in the 50% acetonitrile/chloroform and 10% acetic acid/acetonitrile fractions (data not shown). Figure 2 also shows that the MIP appeared to bind atrazine with at least three different binding avidities. These were characterized by the solvent polarity required to desorb atrazine from each class of binding site. Multiple classes of binding sites were previously described for another atrazine MIP using Scatchard analysis of receptor binding and were attributed to the formation of noncovalent interactions between atrazine and single or multiple MAA molecules.15 Here, we present data that support this mechanism with our MIP using a different analysis technique. 806 Analytical Chemistry, Vol. 69, No. 5, March 1, 1997

Figure 2. Elution of atrazine from a MISPE column using an extract from atrazine-spiked beef liver homogenate. Extract from a 1.0 ppm atrazine-spiked beef liver homogenate (10×) was applied to the column, and atrazine was eluted by increasing the amount of acetonitrile (AcN) in chloroform (CHCl3). Tightly bound atrazine was eluted with acetic acid (HAc) in acetonitrile. The percentage indicated in parentheses is the amount of the applied atrazine (6.67 µg) recovered in the corresponding peak.

To determine the extent to which atrazine could be purified from the 10× extract, chloroform and 3% acetonitrile/chloroform washing steps were followed by a 50% acetonitrile/chloroform elution (elution method 4). Using this condition, 49% of the atrazine eluted with the washing solvent, and the remaining 51% eluted with the elution solvent. The HPLC chromatograms of the crude extract and the 50% acetonitrile/chloroform fraction from this extract are shown in Figure 3. The chromatograms clearly illustrate the effectiveness of MISPE for sample extract purification. Atrazine was significantly enriched, and most of the impurities were removed from the sample. In addition, after removal of the solvent in the wash fraction, a brown, oily material remained, which suggested that this initial wash removed the majority of the fats and lipids present in the organic liver extract. However, due to the relatively low analyte recovery obtained with this elution condition, it was not further pursued since it was unsuitable for quantitative analysis. In an attempt to streamline the MIPSE procedure, we tested the effect of sample volume on analyte recovery. Therefore, either 0.67 mL of a 10× extract or 20 mL of a 0.33× extract was applied to the column. This was followed by a simple chloroform wash to remove fats and lipids and elution with 10% acetic acid/ acetonitrile. Atrazine recoveries were 103% and 93% for the 0.67 and 20 mL extracts, respectively, from a 0.1 ppm atrazine-spiked beef liver homogenate. This result suggested that acceptable analyte recovery could be maintained despite the use of a sample extract that was nearly 30 times more dilute. Figure 4 shows the HPLC chromatograms from the crude extract and the extract following MISPE (20 mL application volume). The packing material used for HPLC analysis for these and subsequent experiments was deactivated C18, which resulted in improved resolution of the tissue extract contaminants and atrazine. Although the level of sample extract cleanup was not as extensive using this protocol compared to using a 3% acetonitrile/chloroform washing step (elution method 5), it was significant, and analyte recovery was nearly quantitative at this spike level. Upon

Table 2. HPLC Results from the Analysis of Unpurified and MISPE-Purified Extracts of Atrazine-Spiked Beef Liver Homogenates amount found, ppm spike level, ppm

mean

unpurified % recovery

%CV

MISPE-purified mean % recovery %CV

0.000 0.001 0.005 0.020 0.100 0.500

0.006 0.005 0.003 0.012 0.064 0.281

533.4 62.4 61.2 63.6 56.2

147.5 146.8 137.2 75.2 28.3 30.4

0.000 0.000 0.005 0.019 0.085 0.372

0.0 99.0 96.6 84.6 74.6

18.6 31.5 6.7 13.1

Table 3. ELISA Results from the Analysis of Unpurified and MISPE-Purified Extracts of Atrazine-Spiked Beef Liver Homogenates amount found, ppm spike level, ppm Figure 3. HPLC chromatograms of extracts from a 1.0 ppm atrazine-spiked beef liver homogenate before (A) and after MISPE (B) with acetonitrile/chloroform wash step. Extract from a 1.0 ppm atrazine-spiked beef liver homogenate (10×) was applied to the column and rinsed with 3% (v/v) acetonitrile/chloroform. Atrazine was eluted with 50% (v/v) acetonitrile/chloroform (elution method 4). The arrow indicates the atrazine peak. HPLC analysis was performed using a standard C18 column (non-end-capped).

Figure 4. HPLC chromatograms of extracts from a 0.1 ppm atrazine-spiked beef liver homogenate before (A) and after MISPE (B). Extract from a 0.1 ppm atrazine-spiked liver homogenate (0.33×) was applied to the column and rinsed with chloroform. Atrazine was eluted with 10% (v/v) acetic acid/acetonitrile. The arrow indicates the atrazine peak. HPLC analysis was performed using a deactivated C18 column (end-capped).

evaporation, the mean weight of the chloroform-extractable solids present in the crude extract was 3.2 mg/g liver tissue. After MISPE, 91% of the extractable solids were eliminated. This was further evidence that MISPE significantly enriched the analyte from nonspecific materials in the extract. The streamlined MISPE protocol was used for subsequent experiments. MISPE and Analysis of Atrazine from Spiked Beef Liver Homogenates by HPLC and ELISA. Beef liver homogenates

0.000 0.001 0.005 0.020 0.100 0.500

unpurified mean % recovery 0.000 0.000 0.004 0.012 0.082 0.472

0.0 82.2 59.5 82.0 94.5

%CV

mean

15.8 10.4 29.2 15.9

0.001 0.001 0.005 0.019 0.082 0.438

MISPE-purified % recovery %CV 100.0 105.5 95.0 83.3 87.5

145.0 223.6 30.9 23.4 16.6 9.1

were fortified to various levels of atrazine and extracted, and the dilute crude extracts were subjected to the MISPE procedure using elution method 5. These and unpurified extracts were analyzed by both HPLC and ELISA. For the HPLC method (Table 2), the MISPE procedure resulted in improvements in accuracy and precision and lowered the limit of detection.22 Following MISPE, atrazine recoveries at spike levels of 0.005 ppm and above were greater than 80%. Atrazine was not detected in either the unspiked or in the 0.001 ppm spiked samples. In all the unpurified extracts, atrazine determination was inaccurate. The average %CV values for the determination of atrazine in unpurified and MISPEpurified extracts were 94.2% and 17.5%, respectively. An explanation for the observed improvements in the accuracy and precision of the HPLC determination can be seen in Figure 4. The elimination of important interfering components in the tissue extract by MISPE resulted in baseline resolution of the atrazine peak. This allowed for low amounts of analyte to be properly quantified by HPLC. The MISPE procedure also improved the accuracy of the ELISA method at or above atrazine spike levels of 0.005 ppm (Table 3). Following MISPE, atrazine recoveries for these extracts were above 80%. However, the precision of the ELISA was not improved by the MISPE procedure. At or above 0.005 ppm, the average %CV values for the determination of atrazine in unpurified and MISPE-purified extracts were 17.8% and 20.0%, respectively. Following MISPE, the unspiked and 0.001 ppm spiked samples gave background levels of approximately 0.001 ppm atrazine. Therefore, this method could not be used at tissue levels below 0.005 ppm. The improvement in recovery observed with MISPE probably resulted from the elimination of fats and lipids from the extract that, upon solvent exchange (which is required for ELISA (22) Keith, L. H.; Crummett, W.; Deegan, J.; Libby, R. A.; Taylor, J. K.; Wentler, G. Anal. Chem. 1983, 55, 2210-2218.

Analytical Chemistry, Vol. 69, No. 5, March 1, 1997

807

analysis), would sequester the lipophilic analyte in the insoluble matrix (unpurified extracts were opaque) and reduce the amount of analyte available for antibody binding. The MISPE procedure resulted in the reliable quantification of atrazine with either the HPLC or ELISA method at the tolerance level of 0.02 ppm atrazine in meat products.23 Without MISPE, the effectiveness of either method at this level would have been marginal. The use of a dilute sample extract also showed that MISPE could remove and concentrate the analyte from a large sample volume. For the 0.005 ppm tissue extract, atrazine was present in the dilute extract at a concentration of 1.7 ppb. This feature may have important implications for environmental analytical situations, where the analyte of interest may be present at low levels, requiring the processing of large sample volumes. The successful use of MISPE to clean up complex biological matrices as determined by two different analytical methods clearly demonstrated that this technique is robust and may be generally applicable.

CONCLUSIONS We have applied molecular imprinting technology for residue analysis in complex biological samples. This is the first report of the use of MIPs as the solid phase matrix for the solid phase extraction and cleanup of organic tissue extracts. The application of MIPs in this mode resulted in improvements in analyte quantification by two different determinative methods, HPLC and ELISA. Although solid phase extraction methods are well established for many low molecular weight compounds, including the s-triazine herbicides,24 this approach offers a new alternative for the analysis of highly lipophilic analytes that may not be amenable to traditional analytical techniques.

Received for review May 9, 1996. Accepted November 12, 1996.X AC9604649

(23) Food Safety and Inspection Service, Science and Technology Evaluation Branch: Washington, DC, p 1.5. (24) Thurman, E. M.; Meyer, M.; Pomes, M.; Perry, C. A.; Shwab, A. P. Anal. Chem. 1990, 62, 2043-2048.

808

Analytical Chemistry, Vol. 69, No. 5, March 1, 1997

X

Abstract published in Advance ACS Abstracts, February 1, 1997.