Development of a chemiluminescent ELISA and a ... - Springer Link

4 downloads 0 Views 240KB Size Report
Nov 18, 2009 - gold-based LFIA for TNT detection. S. Girotti & S. Eremin & A. Montoya & M. J. Moreno &. P. Caputo & M. D'Elia & L. Ripani & F. S. Romolo &.
Anal Bioanal Chem (2010) 396:687–695 DOI 10.1007/s00216-009-3264-0

ORIGINAL PAPER

Development of a chemiluminescent ELISA and a colloidal gold-based LFIA for TNT detection S. Girotti & S. Eremin & A. Montoya & M. J. Moreno & P. Caputo & M. D’Elia & L. Ripani & F. S. Romolo & E. Maiolini

Received: 21 September 2009 / Revised: 21 October 2009 / Accepted: 22 October 2009 / Published online: 18 November 2009 # Springer-Verlag 2009

Abstract To identify the explosive used in a terrorist attack, or to obtain an early sign of environmental pollution it is important to use simple and rapid assays able to detect analytes at low levels, possibly on-site. This is particularly true for TNT (2,4,6-trinitrotoluene), one of the most employed explosives in the 20th century and at the same time, because of its toxicity, a well known pollutant. In this work we describe the development of an indirect competitive ELISA with chemiluminescent detection (CL-ELISA) and of a lateral-flow immunoassay (LFIA) based on colloidal gold nanoparticle labels. A commercially available monoclonal antibody was used and 13 specially synthesized conjugates were tested. We optimized the assay by determining the optimal concentration of monoclonal antibody and conjugates and the influence of various non-specific factors such as: tolerance to organic solvents at different concentrations, the washing and competitive step time, and the crossreactivity with related compounds. The sensitivity and S. Girotti (*) : P. Caputo : E. Maiolini Dipartimento di Scienza dei Metalli, Elettrochimica e Tecniche Chimiche, Università di Bologna, Via San Donato 15, 40127 Bologna, Italy e-mail: [email protected]

S. Eremin Department of Chemical Enzymology, Faculty of Chemistry, M.V. Lomonosov Moscow State University, Moscow 119992, Russia A. Montoya : M. J. Moreno Instituto Interuniversitario de Investigación en Bioingeniería y Tecnología Orientada al Ser Humano, Universidad Politécnica de Valencia, Camino de Vera, s/n, 46022 Valencia, Spain

reproducibility of the CL-ELISA were good (LOD and IC50 values in the ng mL−1 range, and CV value about 7%). It has been applied to real samples of various materials involved in a controlled explosion of an “improvised explosive device”. Three extraction procedures were tested on these samples, all employing methanol as the solvent. The lateral flow immunoassay (LFIA), developed by using the same immunoreagents, reached a detection limit of 1 μg mL−1 when tested on the same samples analysed by CL-ELISA. Keywords 2,4,6-Trinitrotoluene (TNT) . Chemiluminescence . ELISA . LFIA . Immunoassay

Introduction 2,4,6-Trinitrotoluene (TNT) is one of the most often used explosives, because it is simple and relatively safe to M. D’Elia Gabinetto Regionale di Polizia Scientifica per l’Emilia Romagna, Via Volto Santo 3, 40123 Bologna, Italy L. Ripani Reparto Investigazioni Scientifiche (RIS) Carabinieri Roma, Viale Tor di Quinto, 151, 00191 Roma, Italy F. S. Romolo Dipartimento di Medicina Legale - Università di Roma “La Sapienza”, Viale Regina Elena, 336, 00161 Roma, Italy F. S. Romolo Institut de Police Scientifique, Université de Lausanne, Batiment Batochimie, 1015 Lausanne, Switzerland

688

manufacture, and has high explosive power, and above all because its high chemical stability and low sensitivity to impact and friction make it safe to handle [1]. TNT poses a threat not only because of its possible use by terrorists, but also because it is a well known environmental toxic pollutant [2, 3]. For these reasons the possibility of detecting this compound in traces, especially on-site, far from laboratories, is of great interest in the field of prevention of crime, in the immediate identification of the explosives used in terrorist attacks [4], or in revealing an early sign of environmental pollution. Many methods have been reported for TNT detection, including the standard EPA methods based on highperformance liquid chromatography (HPLC) [5], or gas chromatography [6] with various detection systems, for example mass spectrometry [7, 8] and spectroscopy [8–13], luminescence [14, 15] and immunoassay [16–21]. A novel hybrid electrochemical–colorimetric sensing platform has recently been developed [22]. Various extraction methods have been applied, for example solid-phase microextraction coupled with gas-chromatographic analysis [23, 24]. In a recent review Smith [25] provides an exhaustive list of the literature describing the use of biosensors and biologically-inspired system for explosives detection underling the need for fast, highly specific, reliable, and lowcost tests in this field. This objective also includes the development of physical or chemical techniques for the detection of latent materials at the crime scene according to the concept of diagnostic field tests. A significant advantage of the use of diagnostic field tests is the ability to deal with “dissipating evidence”, for example explosive traces on the hands of suspects. If time is lost, there is a risk of losing such evidence, which is liable to deteriorate rapidly [26] and so various portable instruments have been developed [9], such as the portable ion mobility spectrometer [27]. Several diagnostic field tests for TNT are based on the formation of coloured Meisenheimer and Janowsky anions in alkaline acetone or methanol [28–31]. The use of the antibodies in detection methods can help to achieve a goal mentioned above, which is why they are used in different systems and formats. Among laboratory methods based on antibody reactions, various ELISA have been described and commercialized [19, 20, 32] but none employs chemiluminescent detection and among the already developed on-site methods only Smith [33] employed a lateral-flow device. In recent years, use of lateral-flow assays for qualitative and/or semi quantitative on-site analysis has increased substantially, because they are rapid, one-step, and low cost tests [34]. They are now used to detect different analytes in several fields: clinical [35], food [36], agricultural and environmental [37, 38]. The molecule which will

S. Girotti et al.

work as the recognition element, often a primary or secondary antibody, can be labelled with different nanoparticles: colloidal gold [37, 39], latex [39], silica [40], and carbon [41], etc. In this work a commercially available monoclonal antibody and 13 different specially synthesized hapten conjugates have been tested to develop, first, an indirect competitive enzyme-linked immunosorbent assay with chemiluminescent detection (CL–ELISA) and, second, a lateral-flow immunoassay (LFIA) based on a colloidal gold nanoparticles label. Preliminary results were presented at the International Conference “Chimia 2009— New Trends in Applied Chemistry”, May 13–16, 2009, Constanta, Romania [42]. The reliability of the assays was determined by applying them to the analysis of real samples represented by different materials from a controlled TNT test explosion of an “improvised explosive device” (IED).

Material and methods Samples and reagents Standard TNT solutions were purchased from AccuStandard (New Haven, USA) and were used to prepare spiked soil samples: 0.4 g TNT-free soil samples were spiked with 100 μL of 0.5, 1, and 5 μg mL−1 standard. We also analysed shelf surfaces, plastic, and tissue debris produced by the test explosion of 20 g TNT charge inserted into a video cassette. The cassette was wrapped with tissue and inserted into an A4 envelope to simulate a mail bomb, placed on a metal shelf and activated by an N.8 electric blasting cap. Mouse anti-TNT monoclonal antibodies A1.1.1 were from Strategic Diagnostic (Newark DE, USA); horseradish peroxidase-labelled goat anti-mouse immunoglobulins were from Dako (Glostrup, Denmark); rabbit anti-mouse, goat anti-rabbit immunoglobulin, and tetrachloroauric acid were from Sigma. Carbonate–bicarbonate buffer (0.05 mol L−1, pH 9.6; 15 mmol L−1 Na2CO3, 30 mmol L−1 NaHCO3), phosphate buffer (PBS; 10 mmol L−1, pH 7.4; 137 mmol L−1 NaCl, 2.7 mmol L−1 KCl, 10 mmol L−1 Na2HPO4, 2 mmol L−1 KH2PO4); PBS–Tween20 (PBST; 10 mmol L−1 PBS; 0.05% Tween20), PBS–gelatine 1X (PBSG 1X; 1 mmol L−1 PBS, 0.5% fish gelatine) PBS–gelatine 2X (PBSG 2X; 2 mmol L−1 PBS, 1% fish gelatine), and borate buffer (BB; 0.2 mol L−1, pH 8.5; 50 mmol L−1 Na2B4O7.10H2O; 200 mmol L−1 H3BO3) were used. Black polystyrene high-binding plates were from Costar (Cambridge, USA). Nitrocellulose membranes HIFlow Plus HFB1200225 were from Millipore (Billerica, MA, USA).

Development of a chemiluminescent ELISA and a colloidal gold-based LFIA for TNT detection

689

2-Amino-4,6-dinitrotoluene and 4-amino-2,6-dinitrotoluene were kindly supplied by S. Eremin. All other chemicals and organic solvents were of reagent grade.

The tissue (E) and plastic (F) debris (three specimens of each material) were directly immersed in methanol: 0.02 g tissue and 0.17 g plastic were dipped in 500 μL methanol. Extraction of the residues of the explosive from all the samples was performed by use of three different procedures:

Conjugate synthesis

& &

The conjugates were prepared according to standard procedures, as already described [43, 44]. Two conjugates were synthesized starting from trinitrobenzenesulfonate (TNBS), three from trinitrobenzene (TNB), two from aminodinitrotoluene (ADNT), and six from nitrophenylalkyl acid (NPA). The haptens were conjugated to different proteins or macromolecules [43, 44]: ovoalbumin (OVA) (TNBS–OVA, TNB–OVA, NPA-3–OVA, NPA-4–OVA, and NPA-5–OVA), bovine serum albumin (BSA) (TNBS– BSA, TNB–BSA, NPA-3–BSA, NPA-4–BSA, and NPA-5– BSA), soybean trypsin inhibitor (STI) (TNB–STI and ADNT–STI) and aminodextran (DA) (TNBS–DA). The characteristics of these and the concentrations tested are reported in Table 1. Sample extraction Spiked samples were left for 3 h at room temperature then immersed in 500 μL of methanol. Four different areas on the shelf surface, labelled A, B, C, and D (Fig. 1) were divided into subsamples of area 1 cm2 and three samples were collected from different portions of each area using cotton swabs wetted with methanol. Each swab was then dipped in 500 μL methanol.

&

simple shaking for 3 min (method “S”); shaking for 3 min followed by immersion in hot water (almost 70 °C) for 3 min (method “SW”); shaking for 3 min followed by sonication (Starsonic 60 Liarre; Casalfiumanese, BO, Italy) for 1 h in an ice bath (“SS” method).

All samples were analysed both by CL-ELISA and lateral-flow immunoassay after suitable dilution, and by gas chromatography as reference method to confirm the results of the two developed immunoassays. Chemiluminescent ELISA Assays were performed in 96-well microplates as an indirect competitive format. The assay was optimized by determining first the optimum concentration of the immunoreagents, conjugates, and antibody, by a checkerboard titration in the concentration ranges 0.25–2 μg mL−1 for conjugates and 0.25–1 μg mL−1 for anti-TNT monoclonal antibody. The tolerance to the organic solvents methanol and acetone at 1, 3, 5, and 10% in the final dilution was tested. The calibration curves were constructed in the range 0.001–1000 ng mL−1. The microplate was coated with 1 μg mL−1 TNB–OVA conjugate in 0.05 mol L−1 carbonate–bicarbonate buffer,

Table 1 Characteristics and concentrations of the conjugates and TNT monoclonal antibodies tested for optimisation of CL–ELISA Conjugate

Initial concentration of conjugate (mg mL−1)

Tested concentration of conjugate (μg mL−1)

Tested concentration of TNT monoclonal antibody (μg mL−1)

TNB–OVAa TNB–BSAa TNB–STIa ADNT–STIa NPA–5–OVAa NPA–5–BSAa NPA–4–OVAa NPA–4–BSAa NPA–3–OVAa

28.5 22 27 17 7.8 6.2 9 3.6 1.8

0.25; 0.25; 0.25; 0.25; 0.25; 0.25; 0.25; 0.25; 0.25;

0.25; 0.25; 0.25; 0.25; 0.25; 0.25; 0.25; 0.25; 0.25;

0.33; 0.33; 0.33; 0.33; 0.33; 0.33; 0.33; 0.33; 0.33;

0.5; 0.5; 0.5; 0.5; 0.5; 0.5; 0.5; 0.5; 0.5;

1 1 1 1 1 1 1 1 1

NPA–3–BSAa ADNT–OVAa TNBS–OVAb TNBS–DAb

3.4 68 1 1

0.25; 0.5; 1; 2 0.5; 1 0.1; 0.2; 0.5; 1; 2 0.025; 0.05; 0.1; 0.2; 1; 2

0.25; 0.25; 0.25; 0.25;

0.33; 0.33; 0.33; 0.33;

0.5; 0.5; 0.5; 0.5;

1 1 1 1

a

Prepared according to Ref. [44]

b

Prepared according to Ref. [43]

0.5; 0.5; 0.5; 0.5; 0.5; 0.5; 0.5; 0.5; 0.5;

1; 1; 1; 1; 1; 1; 1; 1; 1;

2; 2; 2; 2; 2 2 2 2 2

4; 4; 4; 4;

8; 8; 8; 8;

10 10 10 10; 20

690

S. Girotti et al.

stirring. When the solution changed colour from yellow to cherry red it was left to boil for 5 min more, cooled, and filtered. The pH was adjusted to 7.5 with 50 mmol L−1 carbonate buffer, pH 9.6 and 0.05% (m/v) sodium azide was added. The solution was stored at 4 °C for several months. Colloidal gold–antibody conjugation

Fig. 1 Points of sample collection from exploded shelf using a cotton swab

pH 9.6 (100 μL/well) and incubated overnight at room temperature (RT). Each well was washed three times with 200 μL PBST and the uncoated sites were blocked by adding 200 μL/well PBS-G 1X and incubating at 37 °C for 90 min. After washing, 50 μL/well of 1:40,000 dilution of antiserum in PBSG 2X and 50 μL/well standard solutions or sample extracts were added and incubated for 1 h at RT. The microplate was washed and 100 μL/well goat antimouse IgG-HRP diluted 1:2000 in PBSG 1X (according to the manufacturer’s guidelines) were added, before incubating for 90 min at RT. Finally, after washing, 100 μL luminescent mixture (45 μL luminol 1 mmol L−1, 10 μL p-iodophenol 0.5 mmol L−1 (both from Sigma, Germany), 9845 μL BB 0.2 mol L−1, pH 8.5 and 100 μL of 1 mmol L−1 hydrogen peroxide (Merck, Germany)) were added to each well and the absorbance at 425 nm was immediately recorded by means of a Victor 1420 luminometer (Wallac–Perkin– Elmer, Waltham, MA, USA). The chemiluminescent data, expressed as relative luminescent units (RLU) were normalized according to the expression: %B=B0 ¼ 100ðA  Aexcess Þ=ðA0  Aexcess Þ and then mathematically fitted to a four-parameter logistic equation using Sigmaplot software (SPSS) version 8.0, as already described [45]. The cross-reactivity with other related nitroaromatic compounds was evaluated by following the same procedure and calculating the corresponding IC50 percentage ratios. Lateral-flow immunoassay Colloidal gold preparation Gold nanoparticles of 40 nm were obtained according to the method of Frens (1973) [46]. Briefly, 50 mL 0.01% tetrachloroauric acid (HAuCl4) was heated to boiling then a solution of 1% trisodium citrate was added with constant

The best dilution of antibody to be used for conjugation was determined by titration. To the selected dilution of antibody 1 mL colloidal gold and 100 mL BB 20 mmol L−1 pH 8.5 were added, stored 1 h at room temperature under continuous stirring, and then blocked by addition of 0.2 mol L−1 BB pH 8.5 containing 1% BSA. After centrifugation, the pellet was resuspended in 0.2 mol L−1 BB–1% BSA pH 8.5, again centrifuged, and resuspended in 0.2 mol L −1 BB–0.1% BSA pH 8.5. After further centrifugation the pellet was resuspended in 0.2 mol L−1 BB pH 8.5 containing 1% BSA, 2% sucrose, 0.25% Tween20, and 0.05% sodium azide, and stored at 4 °C for several months. All centrifugations were carried out at 25,000 g for 30 min at 10 °C by use of a Sorvall RC2-B Superspeed centrifuge (Thermo Fisher Scientific, USA). Preparation of the strips A nitrocellulose membrane strip 10 cm long and 2.5 cm wide was cut and fixed on the Easy Printer LPM02 printer device (Advanced Sensor Systems, India) for coating of the test and control lines. Test and control lines, respectively, were coated with 100 μL TNBS–OVA 250 μg mL−1 and 100 μL goat anti-rabbit immunoglobulin 200 μg.mL−1, both dissolved in PBS 1X, distributing 1 μL solution mm−1. The membrane strips were stored overnight at 37 °C and then cut into strips 0.5 cm wide. Assay procedure The lateral-flow assay was performed by dipping the strip in a well containing 50 μL running buffer (PBS 1X–1% BSA), 10 μL rabbit anti mouse-colloidal gold, and 3 μL mouse anti-TNT antibody. The upper part of the strip was in contact with a piece of filter paper which forced the complete migration of the liquid throughout the membrane. After 5 min two red lines were visible, confirming the assay had been conducted correctly. The sample extracts were added into the well containing the reagents described above. The presence of the analyte sought was revealed by a decrease of the intensity of the colour of the test line, until its complete disappearance, to a degree directly proportional to the amount of the analyte in the sample.

Development of a chemiluminescent ELISA and a colloidal gold-based LFIA for TNT detection

691

Table 3 Mean recovery of TNT spiked soils extracted with different methods TNT spiked (ngmL−1)

Extraction methoda

TNT recovered (ngmL−1)

CV (%)

Mean recovery (%)

500

S SW SS S SW

255 181 222 889 878

9.5 8.1 11.8 10.3 9.3

51.0 36.2 44.4 88.9 87.8

SS S SW SS

921 4650 5435 3645

7.5 9.9 7.3 7.8

92.1 93.0 108.7 72.9

1000

5000

Fig. 2 Typical competitive curve for TNT using four different hapten conjugates in 10% methanol: dashed lines with dots, TNBS–OVA; dashed lines with triangles, TNB–OVA; dashed lines with dots and squares, TNB–BSA, dots and diamonds, NPA-4–BSA. Error bars: ±standard deviation (n=3)

Gas chromatography assay A Fractovap 4160 gas chromatograph (Carlo Erba Instruments, Italy) equipped with a split-splitless injector and an RTX-1 MS (10 m, 0.25 mm, 0.25 μm) column, and coupled with a model 850 electron-capture detector (ECD) (Carlo Erba Instruments) was used for analysis of the sample extracts. The flow was 1 mL min−1 and the split 1:5. The injector temperature was set at 175 °C, and the initial oven temperature was set at 45 °C per 1 min. The temperature ramp was set at 25° min−1 to 200 °C and maintained for 5 min at 300 °C. Borwin software ver. 1.61 (Jasco Europe, LC, Italy) was used to collect and process the data. Each sample previously analysed by CL-ELISA was injected (injection volume 3 μL). Standard curves were generated Table 2 Cross reactivity of several compounds tested by CL–ELISA Compound

Cross reactivity (%)

2,4,6-Trinitrotoluene (TNT) 2-Amino-4,6-dinitrotoluene 4-Amino-2,6-dinitrotoluene 2,6-Dinitrotoluene (2,6-DNT) 1,3,5-Trinitrobenzene (1,3,5-TNB) 1,3-Dinitrobenzene (1,3-DNB) 2,4,6-Trinitrobenzoic acid Pentaerythritoltetranitrate (PETN) Cyclomethylenetrinitramine (RDX) 3,5-Dinitroaniline (3,5-DNA) Picric acid Chloramphenicol

100 16 0.1 2.3 9.4 0.9 NDa NDa NDa NDa NDa NDa

a

Not detectable

a

S, simple shaking for 3 min; SW, shaking for 3 min then immersion in hot water (almost 70 °C) for 3 min; SS, shaking for 3 min then sonication for 1 h in an ice bath

by injection of TNT solutions. The TNT peak was confirmed by the standard addition method. The method, according to EPA 8095 [6], is capable of detecting the target compound in the range 0.003–0.5 ng mL−1 and of quantitative analysis in the range 0.03–5 ng mL−1.

Results and discussion Chemiluminescent ELISA The results of checkerboard titration showed that only seven (TNBS–OVA, TNBS–DA TNB–OVA, TNB–BSA, TNB–STI, NPA-4–OVA, NPA-4–BSA; see Table 1) of the 13 tested conjugates were able to provide a useful calibration curve, and that only four had an IC50 value less than 30 μg mL−1 (Fig. 2). The maximum sensitivity was achieved by use of the following conditions: 1 μg mL −1 TNB–OVA and 0.33 μg mL−1 mouse anti-TNT antibody; 90 min as the optimum period for blockage of un-coated sites and for incubation with the labelled antibody; and methanol as extracting solvent at a final concentration of 1 or 10%, because the competitive calibration curves obtained with

Fig. 3 LFIA of TNT standard at different concentrations: 1, no TNT; 2, 100 μg mL−1 TNT; 3, 1 μg mL−1 TNT; 4, 0.1 μg mL−1 TNT

692

Fig. 4 LFIA for 2-amino-4,6-DNT (2) and 4-amino-2,6-DNT (3) compared to the blank sample (1)

these two concentrations were not significantly different. The IC50 and limit of detection (LOD) values obtained with 1% methanol were 2.50 ng mL−1 and 0.22 ng mL−1, respectively, and with 10% methanol of 2.97 ng mL−1 and 0.41 ng mL−1. The reproducibility, expressed as the coefficient of variation (CV), was 7% in both cases. In order to reduce dilution of the extracts by water to a minimum we preferred to use 10% as final dilution of methanol (Fig. 2). The cross reactivity of some TNT-related compounds was tested; the results are reported in Table 2. Low cross reactivity was observed for 2-amino-4,6-dinitrotoluene (16%) and 1,3,5-trinitrobenzene (1,3,5-TNB) (9.4%), and very low cross reactivity for 2,6-dinitrotoluene (2,6-DNT), 1,3dinitrobenzene (1,3-DNB), and 4-amino-2,6-dinitrotoluene. For all other tested compounds cross reactivity was negligible or absent. These results agree fairly well with those obtained by Zeck et al. [47] using the same antibody. No interferences with the assay were detected in presence of the methanol extracts from blank soil and material samples. The results from analysis of spiked soils are shown in Table 3. The different extraction methods

Fig. 5 LFIA of 10 μL of extracts from soil sample spiked with standard TNT: 1, blank soil; 2, TNT 0.5 μg mL−1; 3, TNT 1 μg mL−1; 4, TNT 5 μg mL−1. a, “SS” extraction; b, “SW” extraction; c, “S” extraction

S. Girotti et al.

gave almost the same results, the differences were not significant, and these data confirmed those observed by Jenkins [32] who stated: “in general, a concentration of at least 70% of that attained after 18-hour ultrasonic extraction is achieved after three minutes of manual extraction”. This result gives very important support to the suitability of a lateral-flow assay, because an on-site method needs a simplified, but effective, extraction. This is true for three minutes shaking in methanol, which can be performed without problem on-site. Lateral-flow immunoassay After optimisation, the best conditions for these assays were use of 250 μL mL−1 TNBS–OVA (hapten conjugate) in

Fig. 6 LFIA of 10 μL of real sample extracts: a, surface A; b, surface B; c, surface C; d, surface D; e, tissue debris; f, plastic debris. 1, Blank; 2, SS extraction; 3, SW extraction; 4, S extraction

Development of a chemiluminescent ELISA and a colloidal gold-based LFIA for TNT detection

PBS 1X as test line and 200 μL mL−1 goat anti rabbit antibody (GAR) in PBS 1X as control line. In order to discover the detection limit of the system we added 1 μL of standard TNT at concentrations in the range 100–0.01 μg mL−1 to the strip and the minimum amount leading to a visible decrease of the test line was 1 μg mL−1 (no. 3 in Fig. 3). The LFIA analyses confirmed the cross reactivity data obtained for CL-ELISA. As an example, in Fig. 4 the cross reactivity effects of 2-amino-4,6-dinitrotoluene and 4amino-2,6-dinitrotoluene, two of the more characteristic residues of TNT, are shown: only the first compound, partially recognised as in CL-ELISA, produced a small decrease of the test line colour. Results obtained for spiked soils, all related to the three different extraction methods, are depicted in Fig. 5a, b, and c, and those corresponding to tissue and plastic debris in Fig. 6a–f. Only for sample F there was no agreement with the ELISA determinations, because the decrease of the test line colour was less than expected from its TNT concentration, confirmed by gas chromatography. The lateral-flow immunoassay developed in this research is particularly suitable for on-site testing, and is more practical if we make a comparison with colour field tests for TNT. These suffer from several problems, for example

693

interference from of common anions, colour instability, and water intolerance. The absorbance at the characteristic wavelength of the anion is reduced drastically when the concentration of water in the analyte is 20% [30]. To avoid this problem Üzer et al. introduced extraction of TNT into an organic solvent mixture [29] and the LOD of the method was 0.38 μmol L−1 TNT in the organic phase. The procedure published by Üzer et al. was later modified by solid-phase extraction (SPE) of the Meisenheimer anion [30], lowering the LOD. It is interesting to compare the performance of our lateral-flow immunoassay with a commercial test system based on a non-immunoassay method developed by Jenkins [31], called EnSys TNT Soil Kit. Typically about 10 samples of soil can be run in about 40 min using this kit (extraction time is 2–10 min per sample plus test run time of approximately 2 min). The analysis is based on spectrophotometric analysis, giving an assay range from 1 to 30 μg mL−1 total TNT in soil, and has a minimum detectable amount (MDL) slightly lower than 1 μg mL−1 for 2,4-dinitrotoluene, 1,3,5-trinitrobenzene, 1,3-dinitrobenzene, and tetryl. The LFIA developed in our study is easily recognised to be a more rapid and more selective analytical detection system, having similar sensitivity for TNT. Positive samples can later be quantified with our CL-ELISA or other analytical methods.

Table 4 Comparison of CL–ELISA and GC–ECD assessment of the TNT content of different extracts of real samples Samplea

Extraction methodb

TNT determined by CL–ELISA (ng mL−1)

ELISA repeatability (CV, %)

TNT determined by GC–ECD (ng mL−1)

GC–ECD repeatability (CV, %)

A

SS SW S SS SW S SS SW S SS SW S SS SW S

320,680 57,400 103,200 4,157 4,086 4,445 159,120 147,540 170,570 6,680 10,180 10,740 573 9,268 6,861

17 15 19 14 18 16 20 17 17 13 15 16 11 11 10

294,368 44,906 97,430 3,947 3,961 4,173 150,047 134,230 149,491 7,340 8,736 8,964 354 8,854 6,251

13 18 11 12 15 14 21 11 13 16 17 12 10 10 9.1

SS SW S

244 36 59

9.8 11 9.2

207 34 45

7.8 15 14

B

C

D

E

F

a b

A, surface A; B, surface B; C, surface C; D, surface D; E, tissue debris; F, plastic debris

S, simple shaking for 3 min; SW, shaking for 3 min then immersion in hot water (almost 70 °C) for 3 min; SS, shaking for 3 min then sonication for 1 h in an ice bath

694

S. Girotti et al.

Gas chromatography

References

The results obtained by CL-ELISA and by GC–ECD from plastic and tissue debris are reported in Table 4. The values determined by the CL-ELISA are in agreement with those from gas chromatographic analysis of the same extracts. General overestimation of TNT content by CL-ELISA compared with the GC results is apparent. This characteristic is not a drawback of CL-ELISA, because identification of the presence of TNT by this assay is unaffected and helps to avoid false negative results.

1. Urbanski T (1964) Chemistry and technology of explosives, vol 1. Pergamon Press, Oxford, p 265 2. Yinon J (1990) Toxicity and metabolism of explosives. CRC Press, Boca Raton, FL 3. Purohit V, Basu AK (2000) Mutagenicity of nitroaromatic compounds. Chem Res Toxicol 13:673–692 4. Moore DS, Goodpaster JV (2009) Explosives analysis. Anal Bioanal Chem 395:245–246 5. EPA. Method 8330 (2000) Nitroaromatics and nitramines by High Performance Liquid Chromatography (HPLC), http://www.epa. gov/waste/hazard/testmethods/pdfs/8330b.pdf. Accessed 14 Sept 2009 6. EPA. Method 8095 (2000) Explosives by gas chromatography, US Environmental Protection Agency. http://www.epa.gov/epaoswer/ hazwaste/test/pdfs/8095.pdf. Accessed 14 Sept 2009 7. Song L, Bartmess JE (2009) Liquid chromatography/negative ion atmospheric pressure photoionization mass spectrometry: a highly sensitive method for the analysis of organic explosives. Rapid Commun Mass Spectrom 23:77–84 8. Zhang Y, Ma X, Zhang S, Yang C, Ouyang Z, Zhang X (2009) Direct detection of explosives on solid surfaces by low temperature plasma desorption mass spectrometry. Analyst 134:176–181 9. Moore DS, Scharff RJ (2008) Portable Raman explosives detection. Anal Bioanal Chem 393:1571–1578 10. De Lucia FC Jr, Gottfried JL, Miziolek (2009) AW Evaluation of femtosecond laser-induced breakdown spectroscopy for explosive residue detection. Opt Express 17:419–425 11. Gottfried JL, De Lucia FC, Jr MCA, Miziolek AW (2009) Laserinduced breakdown spectroscopy for detection of explosives residues: a review of recent advances, challenges, and future prospects. Anal Bioanal Chem 395:283–300 12. Leahy-Hoppa MR, Fitch MJ, Osiander R (2009) Terahertz spectroscopy techniques for explosives detection. Anal Bioanal Chem 395:247–257 13. Pacheco-Londoño LC, Ortiz-Rivera W, Primera-Pedrozo OM, Hernández-Rivera SP (2009) Vibrational spectroscopy standoff detection of explosives. Anal Bioanal Chem 395:323–335 14. Meaney MS, McGuffin VL (2008) Luminescence–based methods for sensing and detection of explosives. Anal Bioanal Chem 3391:2557–2576 15. Pittman TL, Thomson B, Miao W (2009) Ultrasensitive detection of TNT in soil, water, using enhanced electrogenerated chemiluminescence. Anal Chim Acta 632:197–202 16. Anderson GP, Lamar JD, Charles PT (2007) Development of a luminex based competitive immunoassay for 2, 4, 6-trinitrotoluene (TNT). Environ Sci Technol 8:2888–2893 17. Anderson GP, Goldman ER (2008) TNT detection using llama antibodies and a two-step competitive fluid array immunoassay. J Immunol Methods 339:47–54 18. Wang J, Liu G, Wu H, Lin Y (2008) Sensitive electrochemical immunoassay for 2, 4, 6-trinitrotoluene based on functionalized silica nanoparticle labels. Anal Chim Acta 610:112–118 19. Krämer PM, Kremmer E, Weber CM, Ciumasu IM, Forster S, Kettrup AA (2005) Development of new rat monoclonal antibodies with different selectivities and sensitivities for 2, 4, 6trinitrotoluene (TNT) and other nitroaromatic compounds. Anal Bioanal Chem 382:1919–1933 20. Goldman ER, Hayhurst A, Lingerfelt BM, Iverson BL, Geirgiou G, Anderson GP (2003) 2, 4, 6-Trinitrotoluene detection using recombinant antibodies. J Environ Monit 5:380–383 21. Ciumasu IM, Krämer PM, Weber CM, Kolb G, Tiemann D, Windisch S, Frese I, Kettrup AA (2005) A new, versatile field immunosensor for environmental pollutants: development and

Conclusions The objective of developing and optimizing a CL-ELISA and a lateral-flow assay for detection of TNT from different materials was achieved. The first method is highly sensitive whereas the second enables analysis of samples directly onsite. For this purpose our systems seem more suitable than commercial tests based on colour reactions and spectrophotometric analysis. Comparing the LOD value reached by our CL-ELISA, 0.2–0.6 ng mL−1, with those obtained by other ELISA formats [19, 20, 25], we can conclude that the sensitivity of our assay was similar to those previously reported but the ability of the chemiluminescent reagents to detect lower concentrations of horseradish peroxidase (HRP) enabled reduction of the optimum antibody and conjugate concentrations. The developed LFIA had an LOD of 1 μg mL−1, which is good sensitivity for semi-quantitative on-site methods, in agreement with those obtained with some commercial kits, for example the Envirogard TNT (Millipore Corporation), the TNT RaPID Assay (Ohmicron Environmental Diagnostic, USA), and the D TECH TNT explosives field test kit (Strategic Diagnostic, USA) with LOD in the ranges 0.2– 15.0 μg mL−1, 0.25–5 μg mL−1, and 5–45 ng mL−1, respectively. The results from evaluation of real samples carried out by the CL-ELISA and the LFIA were confirmed by GC– ECD, showing that the first can be useful as a quantitative laboratory-based method and the second is a useful on-site screening tool, enabling discrimination between samples containing or not containing the analyte.

Acknowledgments This work was supported by grants from the Ministry of University and University of Bologna (PRIN 2006033429: “New Analytical Tools for Security and Criminal Investigations: Trace Detection and Identification of Explosives and Related Compounds”) and “Fundamental Oriented Research” and Grant 08-03-90301 of the Russian Foundation for Basic Research “New strategy for robust and sensitive chemiluminescent immuno-bio-technology for detection of chlorinated herbicides in environment”.

Development of a chemiluminescent ELISA and a colloidal gold-based LFIA for TNT detection

22.

23.

24.

25.

26. 27.

28. 29.

30.

31. 32.

33.

34.

35.

proof of principle with TNT, diuron, and atrazine. Biosens Bioelectron 21:354–364 Forzani ES, Lu D, Leright MJ, Aguilar AD, Tsow F, Iglesias RA, Zhang Q, Lu J, Li J, Tao N (2009) A hybrid electrochemical– colorimetric sensing platform for detection of explosives. J Am Chem Soc 131:1390–1391 Halasz A, Groom C, Zhou E, Paquet L, Beaulieu C, Deschamps S, Corriveau A, Thiboutot S, Ampleman G, Dubois C, Hawari J (2002) Detection of explosives and their degradation products in soil environments. J Chromatogr A 963:411–418 Mayfield HT, Burr E, Cantrell M (2006) Analysis of explosives in soil using solid phase microextraction and gas chromatography. Anal Lett 39:1463–1474 Smith RG, D’Souza N, Nicklin S (2008) A review of biosensors and biologically-inspired systems for explosives detection. Analyst 133:571–584 Almog J (2006) Forensic science does not start in the lab: the concept of diagnostic field tests. J Forensic Sci 51:1228–1234 Ewing RG, Atkinson DA, Eiceman GA, Ewing GJ (2001) A critical review of ion mobility spectrometry for the detection of explosives and explosive related compounds. Talanta 54:515– 529 Jenkins TF, Walsh ME (1992) Development of field screening methods for TNT, 2, 4-DNT and RDX in soil. Talanta 39:419–428 Üzer A, Erçağ E, Apak R (2005) Selective spectrophotometric determination of TNT in soil and water with dicyclohexylamine extraction. Anal Chim Acta 534:307–317 Üzer A, Erçağ E, Apak R (2008) Selective colorimetric determination of TNT partitioned between an alkaline solution and a strongly basic Dowex 1–X8 anion exchanger. Forensic Sci Int 174:239–243 Strategic Diagnostic Inc (2009) Explosives. http://www.sdix.com/ ProductSpecs.asp?nProductID=21. Accessed 14 Sept 2009 Jenkins TF, Schumacher PW, Mason JG, Thorne PG (1996) Onsite analysis for high concentrations of explosives in soil. Extraction kinetics and dilution procedures. Special Report No 96-10; CRREL: Hanover, NH, USA Smith RG (2007) A review of biosensors for explosives detection. Proceedings of the 1st UK-US Conference on Chemical and biological sensors and detectors. London, UK, April 2007 Posthuma-Trumpie GA, Korf J, van Amerongen A (2009) Lateral flow (immuno)assay: its strengths, weaknesses, opportunities and threats. A literature survey. Anal Bioanal Chem 393:569–582 Nadanaciva S, Willis JH, Barker ML, Gharaibeh D, Capaldi RA, Marusich MF, Will Y (2009) Lateral-flow immunoassay for

36. 37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

695

detecting drug-induced inhibition of mitochondrial DNA replication and mtDNA-encoded protein synthesis. J Immunol Methods 343:1–12 Krska R, Molinelli A (2009) Rapid test strips for analysis of mycotoxins in food and feed. Anal Bioanal Chem 393:67–71 Wang S, Zhang C, Zhang Y (2009) Lateral flow colloidal goldbased immunoassay for pesticide. Methods Mol Biol 504:237– 252 Zhao Y, Zhang G, Liu Q, Teng M, Yang J, Wang J (2008) Development of a lateral flow colloidal gold immunoassay strip for the rapid detection of enrofloxacin residues. J Agric Food Chem 56:12138–12142 Nielsen K, Yu WL, Lin M, Davis SA, Elmgren C, Mackenzie R, Tanha J, Li S, Dubuc G, Brown EG, Keleta L, Pasick J (2007) Prototype single step lateral flow technology for detection of avian influenza virus and chicken antibody to avian influenza virus. J Immunoassay Immunochem 28:307–318 Xia X, Xu Y, Zhao X, Li Q (2009) Lateral flow immunoassay using europium chelate-loaded silica nanoparticles as labels. Clin Chem 55:179–182 Posthuma-Trumpie GA, Korf J, van Amerongen A (2008) Development of a competitive lateral flow immunoassay for progesterone: influence of coating conjugates and buffer components. Anal Bioanal Chem 392:1215–1223 Maiolini E, Girotti S, Ferri E, Caputo P, Guarnieri G, Eremin SA, Montoya A, Moreno MJ, D’Elia M (2009) Development of chemiluminescent methods for explosives detection. Ovidius Univ Ann Chem 20:57–60 Moreno MJ, Plana E, Montoya A, Caputo P, Manclus JJ (2007) Application of a monoclonal-based immunoassay for the determination of imazalil in fruit juices. Food Addit Contam 24:704–712 Mart’ianov AA, Dzantiev BB, Zherdev AV, Eremin SA, Cespedes R, Petrovic M, Barcelo D (2005) Immunoenzyme assay of nonylphenol: study of selectivity and detection of alkylphenolic non-ionic surfactants in water samples. Talanta 65:367–374 Girotti S, Maiolini E, Ghini S, Ferri E, Fini F, Nodet P, Eremin S (2008) quantification of thiram in honeybees: development of a chemiluminescent ELISA. Anal Lett 41:46–55 Frens G (1973) Controlled nucleation for the regulation of the particle size in monodisperse gold solutions. Nat Phys Sci 241:20–22 Zeck A, Weller MG, Niessner R (1999) Characterization of a monoclonal TNT-antibody by measurement of the crossreactivities of nitroaromatic compounds. Fresenius’ J Anal Chem 364:113–120