Bioconjugate Chem. 1995, 6,691-694
691
Synthesis of a Fluorescent Analog of Polychlorinated Biphenyls for Use in a Continuous Flow Immunosensor Assay Paul T. Charles, David W. Conrad,+ Megan S. Jacobs, J o h n C. Bart,' and Anne W. Kusterbeck" Center for Bio/Molecular Science and Engineering, Code 6900, Naval Research Laboratory, 4555 Overlook Avenue, S.W., Washington, DC 203754348, Received March 29, 1995@
A synthetic scheme has been developed for the preparation of a dye-labeled analog of polychlorinated biphenyls. The reaction of 2,3,5-trichlorophenol with 3-bromopropylamine hydrobromide under basic conditions was used to introduce a free primary amine group into the parent compound by formation of a stable ether linkage. Reaction of this amine with the succinimidyl ester of a sulfoindocyanine dye resulted in amide bond formation to produce a fluorescently-labeled product. The dye conjugate was used to charge a column containing immobilized antibodies against polychlorinated biphenyls. Upon application of samples containing various concentrations of polychlorinated biphenyls, the fluorescent analog was displaced from the column in amounts proportional to the concentration of analyte. Concentrations of polychlorinated biphenyl as low as 1 ppm were measurable using this system.
INTRODUCTION
Polychlorinated biphenyls (PCBs) are a class of some 209 compounds (called congeners) which are distinguished by their degree of chlorination (1). PCBs were used in the United States in a wide variety of industrial applications, such as transformer and capacitor dielectric fluids, printing inks, and pesticides (21, until the federal government began to regulate their manufacture and use in the mid-1970s (3). One of the more common forms of PCBs in industrial use is Monsanto's Aroclors, which are mixtures of many PCB congeners that contain a specific overall degree of chlorination (e.g., Aroclor 1260 is a mixture of PCBs with a n average percentage of chlorine by weight equal to 60%). Given their known toxicity (4-7) and suspected carcinogenicity (8)in humans, PCBs have become one of the most important environmental pollutants targeted for analysis and remediation. The most frequently employed method for PCB analysis currently is gas chromatography (GC) with either electron capture detection (ECD) (9)or mass spectroscopic (MS) detection (10). Unfortunately, both GC-ECD and GC-MS, while being extremely sensitive and accurate, have several important drawbacks for widespread environmental analyses. These techniques require samples to be sent to off-site labs where highly-trained personnel usually take 1-2 weeks to complete the analysis. Additionally, these tests generally cast hundreds of dollars per sample, regardless of whether an environmental contaminant is present or not. Our device does not require the use of high-vacuum pumps necessary for GC-MS and is quite easily taken out into the field. Sample analysis a t the site, along with the fact that each sample takes less than 5 min to analyze, cuts the turnaround time from weeks to minutes. Furthermore, negative samples can be run virtually cost-free, as only samples ,that contain the analyte in question deplete the column. Thus, one person with
minimal training can provide very cost-effective, on-site analysis compared with analytical labs. Other techniques have also been developed with the goal of providing less expensive, on-site analysis, including simple photometric tests (11)and immunoassays (12-14). Nevertheless, the present method is more cost effective in the end than handheld test kits and does not require reagent addition or timed incubations. In this paper, we describe the synthesis of a dye-labeled analog of a PCB used in a fluorescence-based continuous flow immunosensor (15).This marks the first time that extremely hydrophobic analytes like PCBs have been analyzed in this fashion, as previous studies in this lab concentrated on more hydrophilic species (16,171. Briefly, the continuous flow immunosensor is a semiautomated system in which the analyte-containing medium is allowed to flow through a column containing matriximmobilized antibodies against the analyte. These antibodies are incubated with a fluorescent dye-labeled analog of the analyte before the analysis begins. As the analyte passes through the antibody matrix, some dyelabeled antigen is displaced from the antibody in favor of binding the analyte; the analog is then detected in a fluorometer located downstream from the column. We chose to use a derivative of 2,3,5-trichlorophenol as our PCB analog because the antibodies we employed in this work were produced using a trichlorophenyl hapten and not a polychlorinated biphenyl hapten. Cy5.29 was selected as the fluorescent dye because it has a large ('0.28) quantum yield, is water soluble due to the two sulfonate groups, and its emission maximum is far into the red (667 nm) where little interference is expected from naturally-fluorescent species in the water samples. In addition to a description of the synthesis of this dyelabeled antigen, preliminary results for the detection of various Aroclors will be presented. EXPERIMENTAL PROCEDURES
National Research Council (NRC) Postdoctoral Fellow.
* American Society for Engineering Education (ASEE) Post+
doctoral Fellow. * Author to whom correspondence should be addressed. Phone: (202)404-6042.Fax: (202)767-9594. Abstract published in Advance ACS Abstracts, October 1, 1995. @
The 2,3,5-trichlorophenol (Aldrich), 3-bromopropylamine hydrobromide (Aldrich), Cy5.29-OSu sulfoindocyanine dye (Biological Detection Systems, Inc., Pittsburgh, PA), polyclonal chicken anti-PCB IgY (18) antibodies (O.E.M. Concepts, Toms River, NJ), Aroclors 1248, 1254, and 1260 (ChemService, Westchester, PA), Aroclor 1242
Not subject to U S . Copyright. Published 1995 by American Chemical Society
Charles et al.
692 Bioconjugate Chem., Vol. 6, No. 6, 1995
(Ultra Scientific, North Kingstown, RI), activated column matrix (Emphaze chromatography beads, 3M), Triton X-405, reduced, (Aldrich), and phosphate-buffered saline (PBS) (NaC1 120 mM, KC1 2.7 mM, monobasic and dibasic phosphate buffer salts 10 mM, pH 7.4) (Sigma) were all used as received. TLC was performed using precoated silica gel 60 (EM Science) or CIS(Whatman) glass plates each of 0.25 mm thickness. Melting points were determined using a capillary melting point apparatus (MEL-TEMP 11, Laboratory Devices) and are uncorrected. Chromatography was performed using an HPLC system consisting of two pumps (Model 510, Waters), an injector (Model U6K, Waters), and a photodiode array detector (Model 996, Waters). Reversedphase columns (Nova-Pak Cg, or pBondapak CIS,Waters, Radial-Pak 8 x 100 mm) with linear gradients between water (0.2% acetic acid, pH 3.0) and methanol (0.2% acetic acid) were used in all experiments. Detection of the column effluent was achieved a t 280 or 550 nm. Fluorescence detection of effluent from the immunosensor column was provided by a fluorometer with a 12 pL flow cell (JASCO, Model 821-FP). For a complete description of the flow immunosensor components see ref 17. IHNMR spectra were recorded using a 250 MHz spectrometer (IBM AM 250, Bruker). MS (high and low resolution) were performed by Shrader Analytical and Consulting Laboratories, Detroit, MI. (2,3,5-Trichlorophenoxy)propylamineHydrochloride (1). To 0.5 g (2.5 mmol) of 2,3,5-trichlorophenol in 9.0 mL of ethanol was added 0.55 g (2.5 mmol) of bromopropylamine hydrobromide. After the pH of the solution was adjusted to 12 by the dropwise addition of 4 M NaOH, the reaction mixture was allowed to reflux for 5 h. The solvent was removed under reduced pressure to give a crude yellow product. The residue was extracted with ether (3 x 20 mL), and the combined extracts were dried with anhydrous magnesium sulfate and filtered. Anhydrous HC1 was bubbled through the solution for 1 min to yield the hydrochloride as a white precipitate. The product was purified by HPLC using a CSreversed-phase column and a linear solvent gradient between water/ methanol (50/50)(containing 0.2% acetic acid) and methanol (containing 0.2% acetic acid) in 6.0 min a t a flow rate of 2.0 mumin. Detection a t 280 nm showed the product to elute 1.92 min after the start of the gradient. Removal of the HPLC solvents under vacuum yielded 0.37 g (51%) of compound 1 (TLC, silica gel, methanol, Rf = 0.2, positive ninhydrin test): mp 118 "C; 'H NMR (CD30D) 6 7.21 (d, lH, arom, J = 2.2 Hz), 7.06 (d, l H , arom, J = 2.2 Hz), 4.20 (t, 2H, OCH2, J = 5.6 Hz), 3.25 (t, 2H, CH2NH2, J = 6.7 Hz), 2.18 (quint, 2H, CHZCHZCHZ, J = 5.7 Hz); low-resolution mass spectrum (direct probe) calcd for C9Hl1Cl2NO(freebase - C1 H) 220, found 220. (2,3,5-Trichlorophenoxy)propyl-Cy5.29(2). To a solution of 7.7 mg (26.5 mmol) of (2,3,5-trichlorophenoxy)propylamine hydrochloride in 400 p L of sodium borate buffer (12.5 mM, pH 9.3) was added 5.1 mg (6.44 pmol) of Cy5.29-OSu. After the mixture was stirred for 2 h, 800 mL of water was added and the pH of the solution was adjusted to 6.0 with glacial acetic acid. Purification was by HPLC using a CS reversed-phase column and a linear solvent gradient between watedmethanol (50/50) (containing 0.2% acetic acid) and methanol (containing 0.2% acetic acid) in 6.0 min a t a flow rate of 2.0 mL/min. Detection a t 550 nm showed the product to elute 7.96 min after the start of the gradient. Removal of the HPLC solvents under vacuum yielded 5.2 mg (87%)of compound 2 (TCPA-Cy51 (TLC, CIS, methanoywater (70:30), Rf = 0.7): lH NMR (CD30D) 6 8.3 (t, 2H, J = 13 Hz, p, p' protons of bridge), 7.8-7.9 (s + m, 4H, 4-H, 4'-H, 6-H,
+
6'-H), 7.0-7.4 (m, 4H, 7-H, 7'-H and 4-H, 6-H of trichlorophenoxy moiety), 6.6 (t, l H , J = 12 Hz, y proton of bridge), 6.3 (dd, 2H, J = 14 Hz, a,a' protons of bridge), 4.1-4.3 (dt m, 8H, a-,a'- CHZ and NHCH2, CH201, 3.2 (t, 2H, J = 6.8 Hz, CH2C(0)),2.1 (quint, 2H, J = 7.0 Hz, NHCHZCH~CHZO), 1.1-2.0 (s m, 21H, 5 CH2 groups, with a s a t 1.7 for 2 (CH312);low-resolution mass spectrum (positive ion, magic bullet) calcd for C ~ Z H ~ & KN308Sz 929, found 930 (MH), 914 (MH, Na substituted for K), 892 (M+ H, H substituted for K). Coupling of Anti-PCB Antibody to Column Matrix. The support used for the immobilization of the antiPCB antibody was Emphaze beads. These beads are about 60 pm in diameter and contain reactive azalactone groups that react with amines on the antibody to open the lactone ring and form a very stable amide bond between bead and antibody. Emphaze beads (0.24 g) were suspended in 4.0 mL of buffer (0.1 M sodium carbonate, 0.6 M sodium citrate, pH 8.5), mixed with a vortexer and sonicated for 5 min. The beads were centrifuged at 4000 rpm for 5 min, and the supernatant was removed. A 0.78 mL solution of polyclonal chicken anti-PCB antibodies (1.0 mg/mL) in the above buffer was added to the beads. The suspension was mixed for 1 h and centrifuged a t 4000 rpm for 5.0 min with subsequent removal of supernatant. A 2.0 mL solution of Tris buffer (1.0 M, pH 8.0) was added to the antibody-coated support to deactivate all sites on the beads that did not react with the antibodies. The antibody-derivatized support was sequentially washed for 15 min with each of the following solutions: (1)phosphate-buffered saline (PBS), (pH 7.3, 2.0 mL), (2) 0.1 M NaCl(2.0 mL), and (3) PBS, (pH 7.3, 5 x 2.0 mL). The supernatant was assayed by measuring the absorbance a t 280 nm in order to determine the amount of unbound antibody. From this measurement and the initial concentration of protein, the amount of antibody covalently linked to the beads was calculated. Incubation of Antibody Binding Sites with (2,3,5Trich1orophenoxy)propyl-Cy5.29(2). Microcolumns (7.5 x 52 mm) were packed with 100 pL of the antibodymodified beads. After the column was rinsed with PBS (5 x 2.0 mL), the column support was incubated overnight a t 4 "C in a solution of 2 (100 pL, 1.0-5.0 pM in PBS). Before measurements were taken using the column, it was rinsed with a degassed solution of PBS, pH = 7.4, containing 0.1% Triton X-405, reduced, and 15% ethanol (buffer A). Detection of Aroclors. The Aroclor samples (in methanol) were dried by evaporation under a stream of nitrogen and then redissolved in buffer A. Serial dilution using buffer A was used to prepare concentrations of Aroclors ranging from 0.5 ppm to 20 ppm for testing in the flow immunosensor (15). Triplicate samples of Aroclors (100 pL) were injected over the column starting with the lowest concentration and continuing to more concentrated samples.
+
+
+
+
+
RESULTS AND DISCUSSION
Synthetic Aspects. Scheme 1outlines the synthesis of compound 1. It is a Williamson ether synthesis in which the sodium salt of 2,3,5-trichlorophenol was reacted with 3-bromopropylamine hydrobromide to give 3-(2,3,5-trichlorophenoxy)propylamine.This species was then converted to the hydrochloride salt to afford compound 1 in 51% yield. Compound 2 was prepared following the procedure outlined in Scheme 2. The sulfoindocyanine dye Cy5.29 is commercially available as the primary amine-reactive succinimidyl ester (Cy5.29-OSu). Formation of the amide
Synthesis of a Fluorescent Analog of PCBs
Bioconjugate Chem., Vol. 6,No. 6,1995 693
Scheme 1
a,
cn
s
2 1
Concentration of Aroclor 1254 (ppm) Figure 1. Dose-response curve for Aroclor 1254. The error bars represent 1standard deviation, and each point on the curve is the average of three replicates.
Scheme 2
Table 1. Limit of Detection for Aroclors Aroclor 1242 1248
o=c 0
limit of detection (PPm) 10 5
Aroclor 1254 1260
limit of detection (PPm) 4 1
O,&\O
Table 2. Fluorescence Detector Response for Negative Controls negative control 2,4-dichlorophenol 2,4,6-trichlorophenol 2,4,5-trichloroaniline 2,4,5-trichlorophenoxyacetic acid 2,3,5,6-tetrachlorophenol 2,3,4,5,6-pentachlorophenol a
I
iHz o=c NH-CH2-CH~-CHz-O
CI
2
bond was easily accomplished, and the product 2 was easily separated from the hydrolyzed Cy5.29 acid via HPLC using a reversed-phase CScolumn. The yield for this reaction was 87%. Column Preparation. Immobilization of the antiPCB antibody on a solid support by the outlined procedure yielded an 89% coupling efficiency, as determined by UV absorbance measurements a t 280 nm. The antibody was then exposed to a solution of TCPA-Cy5 (1.0-5.0 pM in PBS) overnight in order to allow the dyelabeled antigen to bind to the antibody. Before the columns were used to detect PCBs, unbound and nonspecifically bound dye conjugates were washed off the column by the initiation of buffer flow through the column. By monitoring the fluorescence (Aex = 635 nm and A,, = 661 nm) of the eluant, a stable base line was determined to be achieved when the fluorescence change was less than 0.0002 AFU/min. Aroclor Detection. Samples of buffer A spiked with
concn 20 ppm 20 ppm 20ppm 20 ppm
response small negative peaka small negative peak" none none
16 ppm 20 ppm
small negative peaka small negative peaka
Fluorescence response signal below base line level.
various concentrations of Aroclor 1242, 1248, 1254, and 1260 have been analyzed using the continuous flow immunosensor. Table 1shows the detection limit for the various Aroclors using the TCPA-Cy5 conjugate described above. The detection limit is defined as the lowest concentration of analyte that could be reproducibly detected using our sensor. Parameters such as flow rate, identity and concentration of the organic cosolvent, identity and concentration of the surfactant, antibody affinity, and dye conjugate have not been fully optimized, and thus, additional experiments could quite possibly yield lower detection limits. A typical dose response curve for Aroclor 1254 is shown in Figure 1. A linear relationship between the integrated peak area of the fluorescence response and the concentration of the Aroclor is observed in the 2-20 pg/mL (220 ppm) region. This leads to a detection limit of 4 ppm, which is already equal to the sensitivity demonstrated (14)by commercially-available immunoassay test kits. Table 2 contains the data for the negative control experiments, showing that the anti-PCB antibodies we used in these experiments have a n excellent ability to discriminate between polychlorinated phenyl compounds and PCBs. In the case of each negative control, the concentration listed is merely the highest tested (chosen to be 10 times more concentrated than the Aroclors). Most likely, higher concentrations of these species will eventually lead to positive signals from the immunosensor, but we do not expect to encounter such large concentrations of chlorinated aromatics in aqueous samples.
Charles et al.
694 Bioconjugate Chem., Vol. 6,No. 6, 1995
Since polyclonal antibodies were used in this study, it is possible that a sample analyzed on a column calibrated with Aroclor 1260 would not analyze for the same quantity of PCB as the same sample assayed on a column standardized with Aroclor 1242. This is due to the possibility that only a few of the many congeners in a n Aroclor may actually be recognized by the antibody and, thus, give rise to the measured signal. However, this problem is common to other antibody-based assays as well. Studies with the disposable immunoassay kits have shown that analyses of field samples can be reasonably accurate if care is taken to use the proper Aroclor as the standardizing agent (14).Immunoassays such as the continuous flow immunosensor or test kits are designed to process large quantities of samples on-site so that PCB-contaminated sites can be identified. Additional testing of the small subset of samples that test above the regulatory limit can then be done by laboratories employing traditional techniques in order to confirm the sites that need remediation. CONCLUSIONS
A major obstacle in the development of fluorescencebased immunoassays for small molecules is the conjugation of the dye to the hapten in such a way that the binding affinity for the antibody is maintained. The synthesis of (2,3,5-trichlorophenoxy)propyl-Cy5.29as a fluorescent analog of PCBs has allowed us to extend the continuous flow immunosensor technology into the realm of environmental detection of chlorinated organics. Although the sensor already has the necessary sensitivity to compete with the other portable immunoassays which are currently available, we plan to continue our investigation in this area in hopes of further optimizing the antibody-antigen binding interaction. New antibodies and other dye-labeled PCB analogs are currently being studied to this end, and future plans include the testing of sensor response towards specific PCB congeners and its effect on the overall analysis of Aroclors. ACKNOWLEDGMENT
The authors thank Frances Ligler and Linda Judd for their helpful comments and suggestions. David Conrad graciously acknowledges the financial support of the National Research Council in the form of a postdoctoral fellowship. John Bart thanks the American Society for Engineering Education for his postdoctoral fellowship. This work was supported by the Office of Naval Research. LITERATURE CITED (1) Erickson, M. D. (1986) Analytical Chemistry of PCBs, Butterworth Publishers, Boston. (2) Durfee, R. L., Contos, G., Whitmore, F. C., Barden, J. D., Hackman, E. E., and Westin, R. A. (1976) PCBs in the United States-Industrial Use and Environmental Distributions. Prepared for the U.S. Environmental Protection Agency, Office of Toxic Substances, Report No. EPA 560/6-76-005 (NTIS NO. PB-252012).
(3) United States Congress, Toxic Substances Control Act, Public Law 94-469, Oct 11, 1976. (4) McKinney, J. D. (1976) Toxicology of Selected Symmetric Hexachlorobiphenyl Isomers: Correlating Biological Effects with Chemical Structure. Proceedings of the National Conference on Polychlorinated Biphenyls, Nov 14-21,1975, Chicago, IL, Environmental Protection Agency, Washington, DC. (5) Biocca, M., Moore, J. A., Gupta, B. N., and McKinney, J. D. (1976) Toxicology of Selected Symmetrical Hexachlorobiphenyl Isomers. I. Biological Responses in Chicks and Mice. Proceedings of the National Conference on Polychlorinated Biphenyls, Nov 14-21, 1975, Chicago IL, Environmental Protection Agency, Washington, DC. (6) Calandra, J. C. (1976) Summary of Toxicological Studies on Commercial PCB’s. Proceedings of the National Conference on Polychlorinated Biphenyls, Nov 14-21, 1975, Chicago, IL, Environmental Protection Agency, Washington, DC. (7) Safe, S. (1984) Polychlorinated Biphenyls (PCBs) and Polybrominated Biphenyls (PBBs): Biochemistry, Toxicology, and Mechanism of Action. CRC Crit. Rev. Toxicol. 13, 319395. (8) Kimbrough, R. D. (1980) Occupational Exposure. Halogenated Biphenyls, Terphenyls, Napthalenes, Dibenzodioxins and Related Compounds (R. D. Kimbrough, Ed.) pp 333-398, ElsevierNorth-Holland Biomedical Press, New York. (9) Environmental Protection Agency (1984) Organochlorine Pesticides and PCBs-Method 608. Fed. Reg. 49(209), 89104. (10) Longbottom, J. E., and Lichtenberg, J. J., Eds. (1982) Methods for Organic Chemical Analysis of Municipal and Industrial Wastewater. U S . Environmental Protection Agency, Report No. EPA-600/4-82-057. (11) Vo-Dinh, T., Pal, A., and Pal, T. (1994) Photoactivated Luminescence Method for Rapid Screening of Polychlorinated Biphenyls. Anal. Chem. 66, 1264-1268. (12) Newsome, W. H., and Shields, J. B. (1981) Radioimmunoassay of PCBs in Milk and Blood. Int. J. Environ. Anal. Chem. 10, 295-304. (13) Mattingly, P. G., and Brashear, R. J. (1992) Reagents and Method for Detecting Polychlorinated Biphenyls. U. S. Patent 5,145,790. (14) Waters, L. C., Smith, R. R., Stewart, J. H., Jenkins, R. A., and Counts, R. W. (1994) Evaluation of Two Field Screening Test Kits for the Detection of PCBs in Soil by Immunoassay. J . AOAC Int. 77, 1664-1671. (15) Kusterbeck,A. W., Wemhoff, G. A., Charles, P. T.,Yeager, D., Bredehorst, R., Vogel, C.-W., and Ligler, F. S. (1990) A Continuous Flow Immunoassay for Rapid and Sensitive Detection of Small Molecules. J. Immunol. Methods 135,191197. (16) Ogert, R. A., Kusterbeck, A. W., Wemhoff, G. A., Burke, R., and Ligler, F. S. (1992) Detection of Cocaine using the Flow Immunosensor. Anal. Lett. 25, 1999-2019. (17) Whelan, J. P., Kusterbeck, A. W., Wemhoff, G. A., Bredehorst, R., and Ligler, F. S. (1993) Continuous-Flow Immunosensor for Detection of Explosives. Anal. Chem. 65, 35613565. (18) The terminology of “IgY”for the antibody used in this study is the designation for a “IgG-like” protein that is found in both the chicken serum and egg yolk. This is the species which O.E.M. Concepts sells as its anti-PCB antibody.
BC950063K
