Oligonucleotide-Functionalized Gold Nanoparticles as Probes in a

Ai-Li Shiau, Pai-Chiao Cheng, Dar-Bin Shieh, Dong-Hwang Chen, Chen-His Chou, .... Ayman Saeed , Mohammad Nooredeen Abbas , Mohammad S. El-Shahawi ...
9 downloads 0 Views 150KB Size
Download PDF
Anal. Chem. 2003, 75, 4155-4160

Oligonucleotide-Functionalized Gold Nanoparticles as Probes in a Dry-Reagent Strip Biosensor for DNA Analysis by Hybridization Kyriaki Glynou,†,‡ Penelope C. Ioannou,*,‡ Theodore K. Christopoulos,*,§,| and Vassiliki Syriopoulou⊥

Medicon Hellas SA, Gerakas, Greece 15344, Laboratory of Analytical Chemistry, Department of Chemistry, University of Athens, Athens, Greece 15771, Department of Chemistry, University of Patras, Patras, Greece 26500, Institute of Chemical Engineering and High Temperature Chemical Processes, P.O. Box 1414, Patras, Greece 26500, and First Department of Pediatrics, Aghia Sophia Children’s Hospital, Athens University, Athens, Greece

The highly specific molecular recognition properties of oligonucleotides are combined with the unique optical properties of gold nanoparticles for the development of a dry-reagent strip-type biosensor that enables visual detection of double stranded DNA within minutes. The assay does not require instrumentation and avoids the multiple incubation and washing steps performed in most current assays. Gold nanoparticle reporters with oligo(dT) attached to their surface form an integral part of the strip. Biotinylated PCR products (233 bp or 495 bp) are hybridized (5 min) with a poly(dA)-tailed oligo and applied on the strip, which is then immersed in the appropriate buffer. As the buffer migrates upward, it rehydrates the nanoparticles that are linked to the target DNA through poly(dA)/(dT) hybridization. Capture of the hybrids by immobilized streptavidin in the test zone of the strip generates a characteristic red band. A second red band is formed, by hybridization, in the control zone of the strip to indicate proper test performance. The sensor offers at least 8 times higher detectability than ethidium bromide staining of agarose gels and provides confirmation of the amplified fragments. Quantitative data are obtained by densitometric analysis of the bands. As low as 2 fmol of amplified DNA were detectable by the strip sensor. Also, 500 copies of prostate-specific antigen cDNA were detected by combining PCR and the strip sensor. The sensor was used successfully for detection of hepatitis C virus in plasma samples from 20 patients. The strip detected 16 out of 16 positive samples and gave no signal for 4 samples that were negative for the virus. To our knowledge, this is the first dry-reagent system that makes use of oligonucleotide-conjugated gold nanoparticles as probes. * Corresponding authors. Phone: (+30)2107274574 (P.C.I.); (+30)2610997130 (T.K.C.). Fax: (+30)2107274750 (P.C.I.); (+30)2610997118 (T.K.C.). E-mail: [email protected] (P.C.I.); [email protected] (T.K.C.). † Medicon Hellas SA. ‡ University of Athens. § University of Patras. | Institute of Chemical Engineering and High Temperature Chemical Processes. ⊥ Aghia Sophia Children’s Hospital. 10.1021/ac034256+ CCC: $25.00 Published on Web 07/15/2003

© 2003 American Chemical Society

The analysis of specific nucleic acid sequences by hybridization has a major impact in diverse areas, such as molecular diagnosis of disease and assessment of therapy, food testing, agriculture, forensic science, and environmental testing.1 The Human Genome Project has provided a vast amount of sequence data and therefore initiated a new era of nucleic-acid-based tests. High-throughput DNA/RNA analysis techniques, however, are required in order to exploit the accumulated genetic information. For more than two decades, the analysis of specific nucleic acid sequences was carried out by electrophoresis, southern or northern transfer, hybridization with radioactive probes, and autoradiography. Although this procedure has contributed greatly to the advancement of knowledge in nucleic acid chemistry and biology, it is timeconsuming, it poses health hazards due to the use of radioactivity, and it is not amenable to automation and high-throughput analysis. Amplification techniques have transformed the way nucleic acid analysis is performed. Polymerase chain reaction (PCR), ligase chain reaction (LCR), nucleic acid sequence based amplification (NASBA), and strand displacement amplification (SDA) provide exponential amplification of minute amounts of DNA and RNA sequences from complex mixtures and, therefore, have become an essential step of the recent procedures used for nucleic acid detection and quantification.2 In recent years, considerable efforts have been put into the development of methods for analysis of PCR products. The most commonly used method, in research and routine laboratories, is based on ethidium bromide staining of DNA fragments separated by electrophoresis. High-performance liquid chromatography and capillary electrophoresis have also been used for automatable separation and detection of PCR products.3,4 Furthermore, microtiter well-based hybridization assays were developed in which the amplified DNA was hybridized and captured on the surface of the wells. After washing out the excess of reagents, the hybrids were linked to a nonradioactive detection system which usually involves an enzyme as a reporter that acts on a chromogenic, fluorogenic, or chemiluminogenic substrate.5,6 Homogeneous (1) Christopoulos, T. K. Anal. Chem. 1999, 71, 425R-438R. (2) Christopoulos, T. K. Polymerase chain reaction and other amplification systems. In Encyclopedia of Analytical Chemistry; Meyers, R. A., Ed.; Wiley: Chichester, U.K., 2000; pp 5159-5173. (3) Huber, C. G. J. Chromatogr., A 1998, 806, 3-30. (4) Righeti, P. G.; Gelfi, C. Electrophoresis 1997, 18, 1709-1714.

Analytical Chemistry, Vol. 75, No. 16, August 15, 2003 4155

hybridization assays have been developed that are based on fluorescence resonance energy transfer (FRET) and offer the advantage of real-time monitoring of the amplification process.7 The objective of the present work was to develop a strip-type biosensor, based on oligonucleotide-functionalized gold nanoparticle probes, that allows visual detection of the amplified DNA within minutes and, contrary to the described methods, does not require instrumentation. Moreover, the sensor is a dry-reagent system that does not involve multiple incubation and washing steps commonly performed in most other assays. Protein-gold nanoparticle conjugates have already been studied extensively8 and used in a large number of immunochromatographic tests.9,10 Oligonucleotide-gold nanoparticle conjugates were introduced in recent years mainly for DNA-mediated self-assembly of nanostructures in solution.11,12 Au nanoparticles conjugated to Raman dye-labeled oligonucleotides were used for detection of oligonucleotide targets by surface-enhanced Raman scattering (SERS) spectroscopy.13 Oligonucleotide-Au nanoparticle conjugates were employed for the electrochemical detection of amplified DNA using disposable electrodes.14 It was also shown that colloidal Au nanoparticles can be used for the chipbased detection of DNA hybridization employing a microscope equipped with a CCD camera.15 A molecular beacon (25-mer oligo) was constructed using an organic dye as the fluorophore and a 1.4-nm gold nanoparticle as an effective quencher. The fluorescence was restored upon binding to a complementary target.16 Oligonucleotide-Au nanoparticles were also used in a colorimetric assay, performed in solution, for detection of oligonucleotides.17 A 30-mer oligonucleotide (target DNA) was mixed with two complementary 15-mer probes that were attached to Au nanoparticles (13 nm). Upon hybridization, the target links the nanoparticles, and a polymeric network is formed. This in turn causes a red-to-purple color change of the solution. Solution freezing and thawing and subsequent transferring to a reverse-phase silica plate was required. The reported detectability was 10 fmol of oligonucleotide target. In the present work, oligonucleotide-functionalized Au nanoparticle probes constitute an integral part of a dry-reagent strip biosensor. It is demonstrated that, upon rehydration, the conjugates retain their ability to hybridize under flow conditions. The lateral-flow format provides an effective means of separation of hybridized from free probes. It should be noted that the mecha(5) Ioannou, P. C.; Christopoulos, T. K. Anal. Chem. 1998, 70, 698-702. (6) Laios, E.; Ioannou, P. C.; Christopoulos, T. K. Anal. Chem. 2001, 73, 689692. (7) Foy, C. A.; Parkes, H. C. Clin. Chem. 2001, 47, 990-1000. (8) Hayat, M. A. Colloidal gold: principles, methods, and applications; Academic Press: San Diego, CA, 1989. (9) Weller, M. G. Fresenius’ J. Anal. Chem. 2000, 366, 635-645. (10) Kasahara, Y.; Ashihara, Y. Clin. Chim. Acta 1997, 267, 87-102. (11) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature 1996, 382, 607-609. (12) Alivisatos, A. P.; Johnsson, K. P.; Peng, X.; Wilson, T. E.; Loweth, C. J.; Bruchez, M. P.; Schultz, P. G. Nature 1996, 382, 609-612. (13) Cao, Y. C.; Jin, R.; Mirkin, C. A. Science 2002, 297, 1536-1540. (14) Authier, L.; Grossiord, C.; Brossier, P.; Limoges, B. Anal. Chem. 2001, 73, 4450-4456. (15) Reichert, J.; Csaki, A.; Kohler, J. M.; Fritzsche, W. Anal. Chem. 2000, 72, 6025-6029. (16) Dubertret, B.; Calame, M.; Libchaber, A. J. Nat. Biotechnol. 2001, 19, 365370. (17) Elghanian, R.; Storhoff, J. J.; Mucic, R. C.; Letsinger, R. L.; Mirkin, C. A. Science 1997, 277, 1078-1080.

4156

Analytical Chemistry, Vol. 75, No. 16, August 15, 2003

nisms of color development in the proposed strip sensor and in the solution test of ref 17 are different. In the solution test, the optical properties of nanoparticles change due to the short interparticle distance (shorter than the particle diameter) in hybridization-induced aggregates. The color development in the strip sensor is due to the accumulation of large numbers of nanoparticles in the test zone and the control zone with no change of optical properties. MATERIALS AND METHODS Instrumentation. PCR amplification reactions were carried out in a Hybaid Omn-E thermal cycler (Middlesex, U.K.). A digital camera, Kodak DC 120, and the Gel Analyzer software for DNA and protein documentation were purchased from Kodak (New York, NY). The dispensers AirJet AJQ 3000 and BioJet BJQ 3000 as well as the Guillotine cutting module CM 4000 were from Biodot LTD (Huntigdon, U.K.). Materials. Streptavidin (SA) from Streptomyces avidinii was purchased from Sigma-Aldrich (Steinhem, Germany). Terminal transferase (TdT) was from MBI Fermentas (Vilnius, Lithuania). Ultrapure 2′-deoxyribonucleoside 5′-triphosphates (dNTPs) were purchased from HT Biotechnology (Cambridge, U.K.). Taq DNA polymerase was from Promega (Lyon, France). Sephadex G-25 Spin-pure purification columns were obtained from CPG (Lincoln Park, NJ). Gold nanoparticles (40 nm, 9 × 1010 particles/mL) were obtained from British Biocell (BB International, Cardiff, U.K.). All common reagents were from Sigma. The oligonucleotides used in this study were synthesized by MWG-Biotech (Ebersberg, Germany). A 5′-thiolated (dT)30 oligonucleotide (p1) was used for conjugation with gold nanoparticles. A (dA)30 oligonucleotide (p2) was used for construction of the control zone of the strip. The 20-mer 5′-CTCTCGTGGCAGGGCAGTCT-3′ was labeled at the 5′ end with biotin and used as an upstream primer (UPSA) for PCR of prostate specific antigen cDNA. The 20-mer 5′-GGTCGTGGCTGGAGTCATCA-3′ was used as downstream primer (DPSA) for PSA cDNA amplification. The 24-mer 5′-ATCACGCTTTTGTTCCTGATGCAG3′ was used as the PSA-specific probe (p3). The 24-mer 5′GCAGAAAGCGTCTAGCCATGGCGT-3′ (KY80) was used as the upstream primer for RT-PCR of HCV RNA (UHCV). The 24-mer 5′-CTCGCAAGCACCCTATCAGGCAGT-3′ (KY78) was 5′ biotinylated and used as the downstream primer for RT-PCR of HCV RNA (DHCV). The 26-mer 5′-CATAGTGGTCTGCGGAACCGGTGAGT-3′ (KY150) was used as a HCV-specific probe (p4). Tailing of Probes with dTTP or dATP. The 5′ thiol modified (dT)30 probe (p1) and the (dA)30 probe (p2) were tailed enzymically with dTTP and dATP, respectively. The reactions were performed in a total volume of 20 µL, containing 0.2 mol/L potassium cacodylate (pH 7.2), 0.1 mmol/L dithiothreitol (DTT), 0.1 mL/L Triton-100, 1 mmol/L CoCl2, 3.5 mmol/L dTTP (or dATP), 30 units of terminal deoxynucleotidyl transferase, and 700 pmol of probe. The reaction was carried out at 37 °C for 60 min. The tailed thiol modified poly-T probe was purified before use by using Sephadex G-25 Spin-pure columns. The specific probes for PSA (p3) and HCV (p4) were tailed with dATP as described. Preparation of Oligonucleotide Conjugated Gold Nanoparticles. Conjugation reactions were carried out by adding 0.9 nmol of purified tailed probe p1 and 8 µL of absolute pyridine to

Figure 1. Left panel: Parts of the proposed dry-reagent strip and principle of the strip-based hybridization assay for target DNA. Poly(dT) conjugated gold nanoparticles are deposited on the “conjugate pad”. Streptavidin (SA) is immobilized on the “test zone”. Poly(dA) strands are immobilized on the “control zone”. The target DNA (a biotinylated amplified fragment) is hybridized with a poly(dA)-tailed probe and loaded on the “sample application” area. The strip is immersed (“immersion pad”) in the running buffer which then migrates and rehydrates the nanoparticle conjugates that hybridize with the sample. The hybrids are captured in the test zone, giving a characteristic red line. Excess nanoparticles are captured in the control zone giving a red line. IP ) immersion pad; CP ) conjugation pad; S ) sample application area; M ) membrane; AP ) absorbent pad. The two shaded areas in the membrane represent the test zone and the control zone of the strip. Right panel: A positive and a negative striptest. In the absence of target DNA, there is no observable signal in the test zone of the strip.

10 mL of the gold nanoparticle solution (∼1.5 pmol). After standing at 4 °C for 24 h, the solution was subjected to “aging” by the addition of NaCl up to a concentration of 90 mmol/L.17 The solution was allowed to stand for another 24 h at 4 °C, and the excess of reagents was removed by centrifugation at 2800 × g for 45 min. The supernatant was discarded, and the red pellet was redispersed in 600 µL of an aqueous solution containing 30% sucrose, 0.25% Tween-20, 0.25% sodium dodecyl sulfate (SDS), and 45 mmol/L NaCl. Preparation of Dry Reagent Strips. The strip consisted of the following parts (Figure 1): a wicking pad, a glass-fiber conjugate pad, and an absorbent pad from Schleicher & Schuell (Dassel, Germany), as well as a laminated membrane from Pall & Gelman (Dreieich, Germany). The parts were assembled on a plastic adhesive backing using the Clamshell Laminator (Biodot, Huntingdon, U.K.) as follows (see Figure 1, right panel): The membrane (M, 25 mm in length) was first placed on the backing. The absorbent pad (AP, 15 mm) was then positioned above the membrane, overlapping by 2 mm. The conjugate pad (CP, 15 mm) was placed below the membrane, overlapping by 2 mm. The immersion pad (IP, 15 mm) was placed below the conjugate pad, overlapping by 2 mm. Oligonucleotide conjugated gold nanoparticles were loaded on the conjugate pad of the strip, at a density of 7.2 fmol of nanoparticles per 4 mm (strip width), by using the AirJet Dispenser AJQ 3000 and were then allowed to dry at ambient temperature. The BioJet Dispenser BJQ 3000 was employed to construct the “test zone” and the “control zone” by loading streptavidin and poly(dA) (respectively) on the membrane (Figure 1). Streptavidin was diluted in a 5% sucrose solution and loaded at a density of 1.6 µg (∼27 pmol) per 4 mm. The poly(dA)

probe p2 was diluted in water and loaded at a density of 1.2 pmol per 4 mm. The membrane was then dried at 70 °C for 20 min. Strips with a 4-mm width were cut by using the Guillotine cutting module CM 4000. The strips were stored dry at ambient temperature. Amplification of Prostate Specific Antigen cDNA. The target DNA was a 233 bp fragment synthesized by amplifying the prostate-specific antigen (PSA) cDNA using the plasmid pcDNA3.PSA18 as a template. PCR was performed in a total reaction volume of 50 µL consisting of 10 mmol/L Tris-HCl (pH 9.0), 50 mmol/L KCl, 1 g/L Triton X-100, 1.5 mmol/L MgCl2, 0.2 mmol/L dNTPs, 25 pmol each of the biotinylated upstream primer (UPSA) and the downstream primer (DPSA), 1.25 units of Taq DNA polymerase, and the plasmid. Amplification was carried out for 35 cycles: 95 °C, 15 s (1 min for the first cycle); 65 °C, 15 s; and 72 °C, 30 s. After completion of the cycles, the mixtures were incubated at 72 °C for 10 min. Subsequently, 10 µL of the PCR product was mixed with 1 µL of NaCl 0.9 mol/L and 1 pmol of the poly(dA)-tailed PSA-specific probe. The mixture was heated at 95 °C for 2 min and then incubated at 37 °C for 5 min. This step was also performed in the thermocycler. RT-PCR of Hepatitis C Virus RNA from Human Plasma. RNA from plasma specimens was isolated by using the HCV specimen preparation kit supplied with the Amplicor HCV detection kit from Roche Diagnostics (Mannheim, Germany). RT-PCR was then performed (according to the manufacturer’s instructions) by using primer combination 2 (UHCV, DHCV) of the Amplicor HCV test v. 2 amplification kit. The primer binding sites are located within the highly conserved 5′-untranslated region of the HCV genome. The size of the PCR product was 244 bp. Subsequently, 10 µL of the PCR product was mixed with 1 µL of NaCl 0.9 mol/L and 1 pmol of the poly(dA)-tailed HCV-specific probe. The mixture was heated at 95 °C for 2 min and then incubated at 37 °C for 5 min. This step was also performed in the thermocycler. Strip-Based Hybridization Assay. Portions (5 µL) of the above mixtures were applied onto the conjugate pad of the strip next to the gold nanoparticles (Figure 1). The strip’s bottom tip was then dipped into a microcentrifuge tube containing 200 µL of phosphate buffer (pH 7.4), 0.15 mmol/L NaCl, 4% glycerol, and 1% SDS. The bands were visualized within 10 min. Longer times did not affect the assay results. RESULTS AND DISCUSSION The principle of the dry-reagent DNA hybridization strip sensor is illustrated in Figure 1. Typical positive and negative strip-assays are also presented in Figure 1. Amplified DNA is end-labeled with biotin during PCR (by using a 5′ biotinylated primer) and mixed with a specific poly(dA)-tailed oligonucleotide. The solution is applied on the conjugate pad of the strip, at a position next to dried gold nanoparticles that carry poly(dT) strands on their surface. Subsequently, the strip is immersed in the running buffer. The buffer migrates upward by capillary action and rehydrates the conjugated nanoparticles which are then linked to the targetspecific probe through poly(dA)/poly(dT) hybridization. The complex migrates along the strip, and the hybrids are bound to immobilized streptavidin (test zone). The accumulation of gold (18) Emmanouilidou, E.; Ioannou, P. C.; Christopoulos, T. K.; Polizois, K. Anal. Biochem. 2003, 313, 97-105.

Analytical Chemistry, Vol. 75, No. 16, August 15, 2003

4157

nanoparticles in the test zone of the strip is visualized as a characteristic red band (plasmon absorbance at 520 nm). By migrating further, the mixture passes through the control zone of the strip where the excess (unhybridized) nanoparticle probes are captured from the immobilized poly(dA) strands, thus forming a second red band (Figure 1). In the absence of target DNA, no red band is observed in the test zone. In this case, a red band in the control zone shows that the strip-test is performed properly. The amount (in fmol) of gold nanoparticles loaded on the conjugate pad of the strip was determined by the nominal concentration of the nanoparticles (provided by the manufacturer) and the volume of the nanoparticle solution dispensed per 4 mm width of the conjugate pad. The amount of the nanoparticles was optimized in the range 2.5-10 fmol per 4 mm width. It was found that 7.2 fmol gave strong color intensities in both the test and control zones with no background. The molar ratio of thiolated poly(dT) oligonucleotide to nanoparticles in the conjugation reaction was studied in the range from 200:1 up to 1200:1. The intensity of the red band increased with the oligo/nanoparticles ratio up to 600:1. At higher ratios, no significant improvement of the intensity of the test and control zones was observed. Following conjugation, the unreacted oligonucleotide was removed by centrifugation, and the nanoparticles were resuspended, in the appropriate buffer, loaded onto the conjugate pad of the strip, and allowed to dry. The resuspension buffer contains sucrose and surfactants in order to (i) ensure stability of the nanoparticles on the strip in the dry form, (ii) facilitate their subsequent rehydration, and (iii) allow their release from the conjugate pad during the test. The amount of poly(dA)-tailed specific oligonucleotide was optimized in the range 0.36-2.8 pmol per 10 µL of PCR mixture loaded on the strip. It was observed that the intensity of the red band at the test zone increased with the amount of probe up to 1 pmol. However, higher amounts of probe resulted in lower band intensities. This is because, at high levels, the amount of poly(dA) oligonucleotide exceeds the binding capacity of the poly(dT)-conjugated nanoparticles. As a consequence, poly(dA) oligonucleotide that is hybridized to the target sequence competes with the free (unhybridized) poly(dA) oligo for binding to limited poly(dT) strands on the nanoparticles. The higher the amount of poly(dA) oligo used, the larger the fraction of nanoparticles that bind to the free probe. These nanoparticles then migrate along the entire length of the strip without being captured from immobilized streptavidin and poly(dA). As a result, the red bands of both the test zone and the control zone become fainter. The running buffer migrates along the strip, by capillary action, and carries the probes, nanoparticles, and hybrids through the test zone where the amount of immobilized streptavidin (1.6 µg, ∼27 pmol) is sufficient to bind the biotinylated target DNA as well as the unincorporated biotinylated primer from the PCR mixture. The free poly(dT)-nanoparticles are captured from the immobilized poly(dA) strands as the solution migrates through the control zone. The length of immobilized poly(dA) strands affects the intensity of the red band at the control zone. Short synthetic oligonucleotides, e.g., (dA)20, give faint bands. The band intensity increases upon tailing with dATP, giving a length of approximately 100 bases. 4158 Analytical Chemistry, Vol. 75, No. 16, August 15, 2003

Figure 2. Left panel: Detectability of the strip sensor. The target DNA is a 233 bp PCR product from prostate-specific antigen cDNA. The amounts of target DNA (femtomoles per strip) loaded on strips 1, 2, 3, 4, 5, 6, 7, and 8 are 500, 62.5, 31.2, 15.6, 7.8, 3.9, 1.9, and 0, respectively. Right panel: Agarose gel (2%) electrophoresis of the above target DNA followed by ethidium bromide staining. M ) DNA markers (φX174 digested with Bsu RI). Lanes 1, 2, 3,4, 5, 6, and 7 contain 500, 62.5, 31.2, 15.6, 7.8, 3.9, and 1.9 fmol of target DNA, respectively.

The composition of the running buffer affects critically the performance of the DNA hybridization strip-test. The running buffer should possess the following characteristics: (a) to facilitate rehydration of the dried nanoparticles and their release from the conjugate pad, (b) to migrate at a flow-rate that both provides a sufficient time for the hybridization reaction to occur and also ensures a rapid strip-test, and (c) to prevent the nonspecific binding of the nanoparticles on the membrane of the strip. The addition of carbohydrates or glycerol results in increased viscosity of the running buffer and causes a decrease of the flow-rate. Various surfactants were added in order to facilitate the release of the nanoparticles from the conjugate pad. Bovine serum albumin or casein was added for elimination of the nonspecific binding and stabilization of the nanoparticles. We found that “aged” gold nanoparticles are quite stable in the absence of protein. The presence of 4% glycerol and 1% SDS in phosphate buffered saline provided the optimum combination for clear bands, rapid tests, and background elimination. To assess the detectability of the strip-based DNA assay, we applied various amounts of PCR products (233 bp, PSA amplified DNA) ranging from 0 to 500 fmol per strip. For direct comparison, 5 µL aliquots of the same samples were subjected to agarose gel (2%) electrophoresis and ethidium bromide staining. The results are presented in Figure 2. As low as 2 fmol of target DNA was detectable with the strip whereas 16 fmol was detected by electrophoresis. Thus, the strip sensor offers 8 times higher detectability than electrophoretic analysis. The detectability of the strip sensor was further evaluated by testing larger fragments of double stranded. In particular, we tested a 495 bp fragment amplified from PSA cDNA, and the results were similar to those obtained with the 233 bp fragment. The color appears in the test zone within 3 min from the application of the solution to the strip, and the assay is completed in less than 10 min, at which time practically all the nanoparticles have migrated to the absorbent pad.

Figure 3. Densitometric analysis of the test zones of the strip. The density of the test zone is plotted against the amount of target DNA loaded on the strip.

Figure 4. Study of the specificity of the proposed strip sensor. Strip 1: Target DNA (233 bp PCR product from prostate-specific antigen cDNA) and poly(dA)-tailed probe are present in the mixture loaded on the strip. Strip 2: Same as Strip 1 but without adding the probe. Strip 3: A PCR negative (no target DNA). The poly(dA)-tailed probe is present in the mixture. Strip 4: An unrelated biotinylated PCR product (230 bp, from human β-globin gene) is present instead of the target DNA. The poly(dA)-tailed PSA probe is included in the mixture. Strip 5: A biotinylated 233 bp DNA fragment that is identical to prostate-specific antigen amplified DNA except for a 24 bp sequence located at the probe binding site.

In order to investigate whether the proposed strip sensor could provide quantitative information on target DNA, the strips were scanned (Hewlett-Packard, ScanJet 3400C desktop scanner), and the intensity of the test zone was estimated densitometrically and plotted as a function of the amount of PCR product loaded on the strip (Figure 3). It is observed that the useful analytical range extends from 2 to 30 fmol of target DNA. A plateau is reached at 62 fmol of DNA. A sample containing 500 fmol of target was also measured, and the band density was found to be the same as with the 62 fmol. Studies were performed to assess the specificity of the strip sensor (Figure 4). No signal was observed when the poly(dA) probe was not added to the PCR mixture. Although, in this case, the biotinylated target DNA is captured in the test zone, the gold nanoparticles cannot be linked to the target DNA. The PCR mixture containing only the specific poly(dA) oligonucleotide but no target DNA did not give a signal in the test zone because the hybrids formed between poly(dA) oligo and poly(dT)-nanoparticles are not captured in the test zone. Analysis performed with an unrelated PCR-amplified biotinylated DNA fragment (230 bp from human β-globin gene) gave no observable binding in the

Figure 5. Strip results (upper panel) and densitometric data (lower panel) pertaining to the detection of PCR amplified prostate-specific antigen cDNA. PCR was performed for 35 cycles, as described in the Materials and Methods section.

test zone. A biotinylated 233 bp DNA fragment that was identical to PSA target DNA except for a 24 bp sequence (located at the probe binding site)18 gave no red band in the test zone. Consequently, an intact triple hybrid between target DNA, poly(dA)tailed specific probe, and poly(dT) gold nanoparticles must be present for signal development in the test zone. To determine the reproducibility of the strip sensor, samples of 1.9, 15.6, and 31.2 fmol of PCR product were loaded each on 4 strips that had been kept dry at room temperature for one month. Densitometric analysis of the test zones of the strips gave CVs of 19%, 13%, and 7%, respectively, for the band density. Accelerated stability studies were carried out by incubating the strips at 37 and 45 °C for 111 days. It was found that the strips were functioning properly after storage for this period of time. In order to investigate the combined detectability of the polymerase chain reaction and the strip sensor, we performed PCR (50 µL reaction volume, 35 cycles) in samples containing from 500 to 5 million copies of a plasmid containing PSA cDNA. Following hybridization, 5 µL of the reaction mixture was tested by the strip sensor. The strip results along with a densitometric analysis are shown in Figure 5. As few as 500 copies of PSA cDNA can be detected. A preliminary clinical evaluation was carried out, by applying the strip sensor to the detection of hepatitis C virus (HCV) in plasma samples from 20 patients. RNA was extracted from plasma by using the Amplicor HCV specimen preparation kit. RT-PCR was then performed using the Amplicor HCV amplification kit. A 244 bp amplified DNA fragment was produced which was hybridized with a poly(dA)-tailed specific oligo and loaded on the strip. The same PCR products were assayed by using the Cobas Analytical Chemistry, Vol. 75, No. 16, August 15, 2003

4159

analyzer (Roche). In the latter system, the biotinylated PCR products are hybridized with immobilized probes, and the hybrids are detected colorimetrically in an ELISA-type assay using avidinperoxidase conjugate. The strip sensor detected 16 out of 16 samples that were found positive by the Amplicor system and gave no signal for 4 samples that were found negative by the Amplicor system. The proposed strip sensor for DNA analysis by hybridization is the first dry-reagent system which makes use of oligonucleotidefunctionalized gold nanoparticles as reporters that constitute an integral part of the strip. Oligonucleotide-tagged liposomes with encapsulated dye were used recently as reporters in lateral-flow strip assays.19,20 The probe tagged liposomes, however, were not included (as a dry reagent) in the strip but were prehybridized with the amplicons, and the mixture was then allowed to flow along the strip where the complexes were captured from immobilized probe in the test zone. Contrary to gold nanoparticles, liposomes are not commercially available, and their preparation is quite complicated. Moreover, due to the inherent instability of liposomes and their dye-encapsulated derivatives, special care should be taken at any step, from the storage of liposomes up to hybridization reaction conditions, to prevent leakage of the dye. A number of DNA hybridization assays based on oligonucleotide-functionalized Au nanoparticles were reported recently.21-23 In one configuration,21 a synthetic oligonucleotide target (e.g., a 39-mer) was first hybridized for 24 h with a shorter oligonucleotide capture probe that was immobilized on a glass slide. Then, a second oligonucleotide probe conjugated to a Au nanoparticle was allowed to hybridize for 4 h, thus forming a “sandwich”. Subsequently, a signal amplification method was used in which Ag+ was reduced by hydroquinone to silver metal at the surface of Au nanoparticles. The slide was then imaged with a scanner. The detectability of this system was found to be 50 fmol/L. In an alternative assay configuration,22 the capture oligonucleotide probe was placed between two fixed microelectrodes. Following sandwich formation, as described, silver was deposited on the nanoparticles that bridged the gap between the electrodes thus increasing the conductivity. The detectability of this assay was about 500 fmol/L. Although these assays offer superior detect(19) Esch, M. B.; Baeumner, A. J.; Durst, R. A. Anal. Chem. 2001, 73, 31623167. (20) Rule, G. S.; Montagna, R. A.; Durst, R. A. Clin. Chem. 1996, 42, 12061209. (21) Taton, T. A.; Mirkin, C. A.; Letsinger, R. L. Science 2000, 289, 1757-1760. (22) Park, S.-J.; Taton, T. A.; Mirkin, C. A. Science 2002, 295, 1503-1506. (23) Patolsky, F.; Ranjit, K. T.; Lichtenstein, A.; Willner, I. Chem. Commun. 2000, 1025-1026. (24) Huhmer, A. F.; Landers, J. P. Anal. Chem. 2000, 72, 5507-5512. (25) Obeid, P. J.; Christopoulos, T. K.; Crabtree, H. J.; Backhouse, C. J. Anal. Chem. 2003, 75, 288-295.

4160

Analytical Chemistry, Vol. 75, No. 16, August 15, 2003

abilities than the proposed strip sensor, they involve very long incubation steps and multiple washing steps. Also, these assays have been evaluated by using aqueous solutions of short synthetic oligonucleotide targets rather than DNA sequences from complex biological samples. On the contrary, we demonstrated that the combination of PCR with the proposed dry-reagent strip sensor offers a 2000 times higher detectability (500 molecules of PSA cDNA in 50 µL reaction mixture correspond to a concentration of 17 amol/L), and the overall procedure including amplification (35 cycles) and strip detection was completed in less than 2 h. It should be noted that the amplification step could become much shorter by using microfabricated devices (chips) for carrying out the PCR in minutes. Stationary or continuous-flow PCR/RT-PCR chips allow rapid heat transfer and ensure fast thermal cycling.24,25 In conclusion, the proposed strip sensor offers the following distinct advantages: (i) The detection of target DNA is significantly faster since it takes only 10 min compared to hours required for detection by agarose gel electrophoresis and other techniques. (ii) There is an 8 times higher detectability than agarose gel electrophoresis and ethidium bromide staining. (iii) The strip sensor allows for confirmation of the target DNA sequence by hybridization to a specific probe, as opposed to electrophoretic techniques that only give the size of the amplified fragment. Nonspecific PCR products are not detected because only the specific amplified sequence hybridizes with the probe. (iv) Quantitative information on the amount of target DNA in the sample can be obtained by scanning the strip (using a common desktop scanner) and analyzing the test zones densitometrically. (v) The proposed strip was designed to be universal. Indeed, the same strip can be used for detection of different DNA targets. The specificity is determined by hybridization of target DNA with the poly(dA)-tailed complementary oligo. (vi) The DNA strip-test does not require instrumentation. (vii) It is a reliable dry-reagent assay of amplified products that does not require expertise in performing the test and interpreting the results. ACKNOWLEDGMENT This work was supported by a University-Industry collaboration grant (PAVE) funded by the General Secretariat of Research and Technology (GSRT) of Greece and by Medicon Hellas SA. We thank Dr. C. Efstathiou (Laboratory of Analytical Chemistry, Department of Chemistry, University of Athens) for developing the software used in densitometric analysis of strip zones.

Received for review March 13, 2003. Accepted May 22, 2003. AC034256+