Environ. Sci. Technol. 2003, 37, 4733-4736
Detection of 2,4,6-Trinitrotoluene in Environmental Samples Using a Homogeneous Fluoroimmunoassay ELLEN R. GOLDMAN,* TIFFANY J. COHILL, CHARLES H. PATTERSON, JR., GEORGE P. ANDERSON, ANNE W. KUSTERBECK, AND J. MATTHEW MAURO Center for Bio/Molecular Science and Engineering, U.S. Naval Research Laboratory, Washington, D.C. 20375
We have tested both soil and water environmental samples for 2,4,6-trinitrotoluene (TNT) using a simple homogeneous assay. This assay is based on changes in fluorescence emission intensity when a fluorescently labeled TNT analogue, bound to an anti-TNT antibody, is competitively displaced by TNT. Fluoroimmunoassay results for TNT concentrations in diluted acetone extracts prepared from archived soils were in good agreement with the results from high-performance liquid chromatography analysis of the same sample extracts. In addition, assays of TNT-spiked environmental well water gave results essentially identical with assays conducted in a TNTspiked laboratory buffer. The homogeneous fluoroimmunoassay, which is rapid, simple, sensitive, and amenable to high throughput screening, shows promise for near realtime evaluation of TNT contamination in environmental samples.
Introduction Manufacture, storage, and demilitarization of weapons have led to the contamination of soil and groundwater with the explosives contained in the ordnance (1, 2). The explosives tend to be extremely persistent in soils and often migrate into surrounding aquifers, further increasing the level of contamination and requiring years of environmental remediation. Complete remediation involves analysis of samples at an off-site laboratory, a costly, time-consuming process that may result in inadequate site characterization when estimating explosive levels due to sample heterogeneity and the relatively small sample number analyzed at any one site (1, 3). Before, during, and after the remediation process, onsite analytical methods can be an important tool in determining the nature and distribution of contamination. Also, by screening a larger number of samples at potentially contaminated sites for only the most frequent explosive, 2,4,6trinitrotoluene (TNT), the extent of contamination can often be estimated (1). Numerous methods have been developed and tested in the past 10 years to meet the need for rapid on-site determinations of TNT levels in soil and groundwater. Colorimetric methods developed by Jenkins and Walsh (5) and based on a color change following derivatization of TNT have been commercialized (Ensys, Millipore Corp.) and * Corresponding author phone: 202-404-6052; fax: 202-767-9594; e-mail:
[email protected]. 10.1021/es034328e Not subject to U.S. Copyright. Publ. 2003 Am. Chem. Soc. Published on Web 09/18/2003
validated by the U.S. EPA for field use (6). Further adaptation of colorimetric analysis to sequential injection spectroscopy for higher throughput and testing of environmental samples has recently been reported (7). A gas chromatographic method, based on laboratory gas chromatography protocols and designed to be field-portable, was shown to be robust and highly accurate in on-site explosives determinations in soil (8). Like other laboratory methods, immunological methods have evolved over the years from their traditional use in medical diagnostics to significant applications in explosives detection, human exposure, and environmental monitoring (9, 10). Enzyme-linked immunosorbant assays (ELISA) and rapid immunofiltration assays for TNT in soil and water using simplified ELISAs have been described (11, 12), and commercial immunoassay kits for TNT in soil (Dtech, Strategic Diagnostics) have received EPA method approval (13). These methods often involve multiple steps and analysis times ranging from 20 min to 2 h (4, 14-19). More recent methods for on-site TNT analysis in environmental matrixes, such as continuous-flow immunoassays, can provide results in only a few minutes (20, 21) but may suffer from significant sample throughput limitations inherent to the flow format. Despite these advances, a critical evaluation of commercial and emerging immunoassay methods for field analysis concluded that these methods were not yet precise and accurate enough to eliminate the requirement for further laboratory confirmation (14). The primary problems involved matrix effects, nonspecific interactions, and issues with sample heterogeneity. These methods, however, were found to be generally useful for yes/no, qualitative results. Because of their low cost, selectivity, and sensitivity, further development of immunoassay field methods was encouraged. We have developed a rapid, simple, and sensitive method for analysis of dissolved TNT that should be amenable to high throughput formats (22). The method is based on the change in the fluorescence emission intensity of a fluorescently labeled TNT analogue that occurs upon its dissociation from an anti-TNT antibody. TNT concentrations can be determined by monitoring the fluorescence decrease that occurs when unlabeled TNT in test samples competes with the fluorescent analogue for binding to the antibody (Figure 1). Unlike standard enzyme-based immunoassays, the format allows detection within minutes without the need for further reagents or color development. In the present study, we have applied this homogeneous assay to the detection of TNT extracted from field soil samples. Extracts containing TNT were analyzed by the homogeneous assay method and by high-performance liquid chromatography (HPLC). In addition, groundwater from the same area was spiked with TNT and tested in the homogeneous assay to examine the utility of the method for evaluating TNT levels directly in environmental water samples.
Materials and Methods Materials. The fluorescent-labeled TNT analogue cyanine diaminopentane-labeled trinitrophenyl (Cy5-DAP-TNP) was prepared as previously described (23). Monoclonal anti-TNT antibody A1.1.1 was purchased from Strategic Biosolutions, Newark, DE. A TNT standard solution was purchased from Radian International. The five soil samples tested were archived soils taken from Umatilla Army Depot Activity (UMDA), Hermiston, OR, and were provided by H. Craig (U.S. EPA, Region 10). All soil samples used were dry and well-homogenized. Groundwater was from a background monitoring well also located at UMDA. VOL. 37, NO. 20, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4733
FIGURE 1. Schematic of the assay. Fluorescent emission decreases as the antibody binds TNT in place of the fluorescent TNT analogue. Spectra were acquired at 5 nm resolution on a Tecan Safire fluorescent plate reader. Soil Extractions. The modified acetone extraction from Shriver-Lake et al. (2) was used to extract TNT from soil samples. Briefly, extracts were prepared by mixing 2.0 g of each soil sample with 10 mL of acetone in glass vials. Each sample was shaken for 3 min and then filtered through a 0.22 µm syringe filter (Nylon Acrodisc, Gelman) into a clean glass vial. Acetone extracts were used immediately for analysis or sealed and kept in the dark at -20 °C for up to 4 weeks prior to analysis. Immediately prior to TNT analysis, small volumes of acetone extracts were transferred to glass test tubes and the acetone was removed by brief treatment with a stream of nitrogen gas. TNT in the test tubes was then dissolved in phosphate-buffered saline (PBS: 200 mM NaCl, 2.7 mM KCl, 8.2 mM Na2HPO4, 4.2 mM NaH2PO4, 1.15 mM K2HPO4; pH 7.4) prior to further dilution for fluoroimmuno analysis. The volume of PBS used to dissolve the TNT from the acetone extract in these experiments was 10 times the volume of the original acetone sample, creating a 1/10 dilution of the TNT in the extract. Identical splits of acetone extracts analyzed by the homogeneous fluoroimmunoassay method were subsequently used for TNT quantitation by HPLC. TNT Analysis in Spiked Well Water Samples versus Spiked PBS. Known volumes of a standard TNT solution were spiked directly into water from a background well at UMDA or into PBS. Final TNT concentrations in the prepared samples ranged from 1.6 to 5000 µg/L. TNT-spiked PBS samples were run simultaneously with spiked well water samples (i.e., in other wells in the same microtiter plate). Homogeneous Fluoroimmunoassay for TNT. Assays were performed following the published protocol for the platebased homogeneous assay for the detection of TNT (22). Briefly, at least 1 day prior to conducting analyses, white 96-well plates (Microfluor 2, Dynex) were blocked with 300 µL of a blocking solution consisting of 1% bovine serum albumin (BSA) in Tris-buffered saline (TBS) [50 mM TrisHCl, 105 mM NaCl, pH 7.5] at 4 °C. The blocking buffer was discarded, and 25 µL of 100 nM Cy5-DAP-TNP TNT analogue plus 25 µL of 5 µg/mL A1.1.1 anti-TNT antibody was added to each test, standard or control well. A 50 µL aliquot of each PBS-diluted unknown sample, standard TNT dilution, or buffer control was then added to the appropriate wells. To generate the assay standard curve, standard TNT dilutions of 1.6, 3.1, 6.3, 12.5, 25, 50, 500, and 5000 µg/L were run on each plate along with the TNT unknowns. For each unknown, dilutions of 1/10, 1/100, 1/1000, and 1/10000 from the initial TNT sample were assayed to help ensure that one or more concentration readings occurred in the linear range of the standard curve. Sample fluorescence was read immediately 4734
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 20, 2003
FIGURE 2. Results from TNT standards spiked into PBS. A standard curve was generated for the linear portion of these data. after final plate preparation (within ca. 5 min). Samples, standards, and PBS blank controls were performed in at least triplicate. Fluorescence emission was read using a Tecan SpectraFluor microplate reader with a 620 nm excitation filter and a 670 nm emission filter. The data were plotted as the difference of the average signal in the control wells that contained only PBS (or background well water) and the average fluorescent signal in the sample or standard wells. HPLC Methods for TNT Analysis. Each soil sample was tested by a standard HPLC method for analysis of TNT in extracts of soil samples (24), using spectrophotometric detection of TNT standards to produce a calibration curve. Dilutions of soil sample extracts for HPLC analysis were chosen based on the results of prior homogeneous fluoroimmunoassays. Calculation of mg/kg (ppm) of TNT in solid samples was performed by multiplying the determined TNT concentration of each acetone extract in mg/L by 5, the dilution factor for diluting the TNT present in 2 g of soil into 10 mL of acetone.
Results and Discussion Homogeneous Fluoroimmunoassays on Soil Field Samples. Construction of a standard curve (Figure 2) allowed direct calculation of TNT levels in acetone extracts derived from contaminated soil samples. A summary of the calculated mean TNT concentration for each soil extract from the immunoassay and the corresponding result from HPLC analysis of the same samples is shown in Table 1. From the results of analysis of each acetone extract, the ppm TNT in each contaminated soil sample can be calculated. The results
TABLE 1. TNT Concentrations Determined for Acetone Extracts of Soil Using Immunoassay and HPLC Analysis soil sample G16-L2A G18-L1-A G18-L3-A G51-L1-A G55-X-A
homogeneous assay (mg/L)a,b 7.5 ( 0.4 14 ( 4 2.4 ( 1 2.2 ( 1 116 ( 20
HPLC (mg/L)b 9.9 ( 0.6 15.6+ 0.6 2.0 ( 0.09 1.7 ( 0.02 153 ( 0.7
HPLC (ppm)c
RPD
49.5 78 10 8.5 765
-28 -11 18 22 -28
a Values shown are the calculated mean TNT concentration for each soil extract and were determined from 6-9 replicates. b Concentration of TNT in acetone extracts. c Values calculated for the concentration of TNT in soil samples.
FIGURE 4. Comparison of TNT spiked into PBS and TNT spiked into a background monitoring well located at UMDA.
FIGURE 3. Comparison of the homogeneous TNT assay and HPLC methods on soil extracts using linear regression. shown in Table 1 assume 100% extraction of the explosive into acetone; the 3 min acetone extraction used in these experiments has been shown to give at least 70% extraction efficiency versus a more extensive extraction procedure (25). Obviously, incomplete extraction introduces an additional complicating factor into quantitative analysis of these types of samples, and it is also likely that soil from different contaminated sites varies in extraction efficiency. TNT extraction that is incomplete and of variable degrees, however, is fundamentally a sampling problem and is unrelated to the utility of the presently described assay methodology. As shown in Table 1, the results of the homogeneous fluoroimmunoassay correlated well with the TNT concentrations determined by the standard HPLC method. The accuracy of the homogeneous assay was examined by two methods: relative percentage difference (RPD) (1, 2) and linear regression analysis. In RPD analysis, the smaller the RPD value, the more accurate the new method (in this case, homogeneous fluoroimmunoassay) relative to the established method (HPLC in this case). For the present samples, RPDs ranged from -28 to 22 (Table 1). These results indicate a good agreement between the two methods because they are within the RPD value of (50 generally considered acceptable (2). Using linear regression analysis (Figure 3), a true agreement of the homogeneous fluoroimmunoassay results versus the HPLC results would result in a slope and correlation coefficient of 1.0. Although the correlation of our data was excellent (0.9997), the slope differed from 1 (0.75) and the intercept was slightly different from zero (1.02). Uncertainty in the TNT concentrations determined by HPLC was less than 7% of the average value for all samples (Table 1). Any unknown systematic or random errors involved in applying the homogeneous method could explain the minor deviation of the slope and intercept from ideal values.
TNT Screening versus Quantitation in Soil Samples. Although the present homogeneous fluoroimmunoassay method for analysis of TNT demonstrated quantitative results comparable with HPLC, it should also be useful for rapid preliminary analysis of TNT contamination at sites in any stage of remediation. It is possible to complete the assay of each sample in 5 min or less after extracting TNT from soil samples. By testing only a single concentration of acetone extract from each soil sample in a preliminary screening mode, more samples could be processed on each microtiter plate. In addition to standard TNT samples and no-TNT control wells, 24 samples can be run in triplicate on each 96-well plate. Using triplicate samples on four plates, 96 samples could easily be prepared and analyzed in 30 min compared to an estimated 33 h to conduct a single analysis of 24 samples by HPLC. Soil acetone extracts that are strongly positive in a preliminary screening mode could be tested in a second quantitation stage by analyzing a dilution series to obtain readings on the linear part of the standard curve. The simplicity of the fluoroimmunoassay should also make it amenable to automation using robotic pipetting and delivery to a fluorescent plate reader. Furthermore, it has been demonstrated that the anti-TNT A1.1.1 antibody is functional in 5% (v/v) acetone (2). We found that the homogeneous fluoroimmunoassay performs identically in PBS and in PBS with a final acetone concentration of 5% (data not shown), so in automated on-site assays acetone extracts of soil could be diluted 1/20 directly into an assay buffer for screening, thus avoiding the acetone removal step we used in the present laboratory studies. An examination of cross-reactivities with other nitroaromatics and other explosives in this assay showed that the results are consistent with the cross-reactivity profile of the antibody used in the assay (22). This homogeneous fluoroimmunoassay has worked with five monoclonal anti-TNT antibodies examined as well as with several anti-TNT singlechain antibody Fv fragments (ERG, TC unpublished observation). Ultimate limits of detection and cross-reactivity profiles of the assay could be tuned by antibody choice. Homogeneous Fluoroimmunoassay on Spiked Well Water Samples. When uncontaminated well water from UMDA was spiked with TNT to explore the utility of the homogeneous assay for direct evaluation of environmental water samples, results were essentially identical with those of spiked PBS samples over the range of 1.56-5000 µg/L TNT (Figure 4). Well water at UMDA has high levels (8000-40 000 µg/L) of nitrates, low turbidities, and low levels of humics (4). Although nitrates can interfere with some colorimetric TNT assay methods, they do not appear to have any effect on the present antibody-based TNT detection method (4). It has been found, however, that high amounts of organic materials in water may compete with antibodies for binding VOL. 37, NO. 20, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4735
sites (14). High concentrations of humics may thus interfere with the present homogeneous assay method. Other field validation studies using antibody-based biosensors for environmental analysis have found that immunoassay methods may exhibit more significant matrix effects. When a large population of complex environmental samples with multiple contaminants was tested, the results for TNT determinations were found to be imprecise and biased high (26). Clearly, it will be necessary to evaluate more samples using water having varying compositions. In the evaluation of the method for TNT measurements in solid soil samples, however, solvent extractions are required. Humics or other interferents may not be a problem if they do not partition into acetone under the extraction conditions. Future Work. This study has demonstrated use of the homogeneous fluoroimmunoassay for detection and quantitation of TNT in environmental samples. In all soil samples tested, the method compared well quantitatively to standard laboratory HPLC analysis. The method also performed well in TNT-spiked well water samples. Unlike field immunoassay kits, multiple samples and serial dilutions can be run at one time, allowing for improved assay calibration and statistical analysis. In the future, both soil and water samples with a range of properties will need to be examined to explore sample matrix effects and validate the assay format for onsite testing.
Acknowledgments Thanks go to Lisa Shriver-Lake and Paul Charles for helpful discussion. T.J.C. was supported by the National Science Foundation through the NRL summer research program for HBCU/MI/TCU undergraduates. This work was funded by the Office of Naval Research. The views, opinions, and/or findings described in this report are those of the authors and should not be construed as official Department of the Navy positions, policies, or decisions.
(5) (6) (7) (8)
(9) (10) (11) (12) (13) (14)
(15) (16)
(17) (18) (19) (20) (21) (22)
(23)
Literature Cited (1) Crockett, A. B.; Jenkins, T. F.; Craig, H. D.; Sisk, W. E. Overview of on-site analytical methods for explosives in soil; Special Report No. 98-4; U.S. Cold Regions Research and Engineering Laboratory (CRREL): Hannover, NH, 1998. (2) Shriver-Lake, L. C.; Patterson, C. H.; van Bergen, S. K. Field Anal. Chem. Technol. 2000, 4, 239. (3) Crockett, A. B.; Craig, H. D.; Jenkins, T. F. Field Sampling and Selecting On-Site Analytical Methods for Explosives in Water; EPA/600/S-99/002; U.S. Environmental Protection Agency: Washington, DC, 1999. (4) Craig, H.; Fergison, G.; Markos, A.; Kusterbeck, A.; Shriver-Lake, L.; Jenkins, T.; Thorne, P. Field Proceedings of the Great PlainsRocky mountain Hazardous Substance Research Center (HSRC)/ Waste Education and Research Consortium (WERC) Joint Conference on the Environment; Great Plains/Rocky Mountain
4736
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 20, 2003
(24) (25)
(26)
Hazardous Substances Research Center: Manhattan, KS, 1996; p 204. Jenkins, T. F.; Walsh, M. E. Talanta 1992, 39, 419. U.S. Environmental Protection Agency Office of Solid Waste Method 8515. Echols, R. T.; Christensen, M. M.; Krisko, R. M.; Aldstadt, J. H., III. Anal. Chem. 1999, 71, 2739. Hewitt, A. D.; Jenkins, T. F.; Ranney, T. A. Field Gas Chromatograph/Thermionic Detector System of On-Site Determination of Explosives in Soils; CRREL TR-01-9; CRREL: Hannover, NH, 2001. Rogers, K. R.; Gerlach, C. L. Environ. Sci. Technol. 1999, 33, 500A. Van Emon, J. M.; Gerlach, C. L. Anal. Chim. Acta 1998, 376, 55. Keuchel, C.; Weil, L.; Niessner, R. Anal. Sci. 1992, 8, 9. Keuchel, C.; Niessner, R. Fresenius’ J. Anal. Chem. 1994, 350, 538. U.S. Environmental Protection Agency Office of Solid Waste Method 4050. Thorne, P. G.; Myers, K. F. Evaluation of Commercial Enzyme Immunoassays for the Field Screening of TNT and RDX in Water; Special Report 97-32; CRREL: Hannover, NH, 1997. Julicher, P.; Mussenbrock, E.; Renneberg, R.; Cammann, K. Anal. Chim. Acta 1995, 315, 279. D Tech TNT Explosives Field Test Kit Literature; Strategic Diagnostics Inc.: Newark, DE, www.sdix.com (accessed Aug 2003). TNT RaPID Assay Literature; Strategic Diagnostics Inc.: Newark, DE, www.sdix.com (accessed Aug 2003). Lan, E. H.; Dunn, B.; Zink, J. I. Chem. Mater. 2000, 12, 1874. Heiss, C.; Weller, M. G.; Niessner, R. Anal. Chim. Acta 1999, 396, 309. Charles, P. T.; Gauger, P. R.; Patterson, C. H.; Kusterbeck, A. W. Environ. Sci. Technol. 2000, 34, 4641. Gauger, P. R.; Holt, D. B.; Patterson, C. H.; Charles, P. T.; ShriverLake, L. C.; Kusterbeck, A. W. J. Hazard. Mater. 2001, 83, 51. Goldman, E. R.; Anderson, G. P.; Lebedev, N.; Lingerfelt, B. M.; Winter, P. T.; Patterson, C. H.; Mauro, J. M. Anal. Bioanal. Chem. 2003, 375, 471. Kusterbeck, A. W.; Patterson, C. H.; Gauger, P. R. In Current Protocol for Field Analysis Chemistry; Lopez-Avila, V., Barcelo, D., Beckert, W., Goheen, S. C., Jinno, K., Keith, L. H., Rittenburg, J. H., Eds.; John Wiley & Sons: New York, 2002; pp 2C11.12C11.14. U.S. Environmental Protection Agency SW-846 Method 8330. Jenkins, T. F.; Schumacher, P. W.; Mason, J. G.; Thorne, P. G. On-site analysis for high concentrations of explosives in soil: Extraction kinetics and dilution procedures; Special Report No, 96-10; CRREL: Hannover, NH, 1996. Foley, G. J.; Reichle, D. E.; Marqusee, J. Environmental Technology Verification Report: Explosives Detection Technology; EPA/ 600-R-00/045; U.S. Environmental Protection Agency: Washington, DC, 2002.
Received for review April 10, 2003. Revised manuscript received August 15, 2003. Accepted August 19, 2003. ES034328E
