Microfabricated Tin–Film Electrodes for Protein and DNA Sensing

Oct 16, 2013 - Xue-cheng Sun , Chong Lei , Lei Guo , Yong Zhou. Microchimica Acta ... Xuecheng Sun , Shaotao Zhi , Chong Lei , Yong Zhou. Biomedical ...
0 downloads 0 Views 3MB Size
Download PDF
Technical Note pubs.acs.org/ac

Microfabricated Tin−Film Electrodes for Protein and DNA Sensing Based on Stripping Voltammetric Detection of Cd(II) Released from Quantum Dots Labels Christos Kokkinos,*,† Anastasios Economou,*,‡ Panagiota S. Petrou,§ and Sotirios E. Kakabakos§ †

Laboratory of Analytical Chemistry, Department of Chemistry, University of Ioannina, 45110 Ioannina, Greece Laboratory of Analytical Chemistry, Department of Chemistry, University of Athens, 157 71 Athens, Greece § Immunoassay/Immunosensors Lab, Institute of Nuclear & Radiological Sciences & Technology, Energy & Safety, NCSR “Demokritos”, Aghia Paraskevi, Athens 153 10, Greece ‡

S Supporting Information *

ABSTRACT: A novel disposable microfabricated tin−film electrochemical sensor was developed for the detection of proteins and DNA. The sensor was fabricated on a silicon wafer through photolithography to define the sensor geometry followed by tin sputtering. A sandwich-type immunoassay with biotinylated reporter antibody was employed for the determination of prostate-specific antigen (PSA) in human serum samples. For the detection of C533G mutation of the RET gene, biotinylated oligonucleotide probes were used. The biotinylated biomolecular probes were labeled with streptavidin (STV)-conjugated CdSe/ZnS quantum dots (QDs); quantification of the analytes was performed through acidic dissolution of the QDs and stripping voltammetric detection of the Cd(II) released. The proposed QD-based electrochemical sensor overcomes the limitations of existing voltammetric sensors and provides a mercury-free sensing platform with scope for mass-production and further potential for application in clinical diagnostics.

M

Nevertheless, the existing applications referring to QDs as labels for electrochemical detection of proteins and DNA make use of electroplated mercury− or bismuth−film electrodes (MFEs and BiFEs, respectively).11−21 MFEs, despite their advantages for electroanalysis, are toxic and considered undesirable.22,23 On the other hand, electroplated BiFEs have several drawbacks such as the need of a Bi(III) plating solution, the requirement of a conductive substrate, the careful control of the solution pH, the requirement for cleaning/regeneration of the electrode surface, and the strong dependence of the metal− film morphology on the electroplating conditions.24,25 In this work, we report a novel microfabricated tin sensor to perform QD-based electrochemical assays for the detection of proteins and DNA. This new type of sensor exhibits exceptional sensitivity and selectivity for the voltammetric stripping detection of Cd(II). The application of microengineering technology for the fabrication of the proposed sensor overcomes all the limitations of electroplating and enables the mass-production of inexpensive and disposable devices with small size.24,25 The bioanalytical potential of the tin sensor is

ethods that enable sensitive, selective, and rapid detection of proteins and DNA are important tools in bioanalytical chemistry.1,2 In the quest for rapid, simple, and inexpensive methods for onsite assays with high sensitivity, electrochemical techniques have been developed.3 In particular, the combination of electrochemistry and nanoparticles (NPs) serving as labels provides an elegant way to detect DNA and proteins in connection with highly sensitive electrochemical stripping analysis.4,5 Gold NPs (AuNPs) are the most common NP labels used in electrochemical detection of proteins and DNA;6−8 after the target biomolecules are labeled with AuNPs, the latter are chemically oxidized, and the Au(III) released is detected by stripping voltammetry. However, the voltammetric detection of Au(III) (normally performed at carbon-based electrodes) lacks in sensitivity while labeling with AuNPs does not allow multiple detection of more than one biomolecules in a single assay. Quantum dots (QDs) offer an attractive alternative to metal NPs as electrochemical labels. The most important advantages of QDs over AuNPs are the higher sensitivity of detection (achieved by selecting a proper combination of the electrode material and composition of nanocrystals) and the possibility to perform multi-analyte assays in a single run using different QDs.3,9,10 © 2013 American Chemical Society

Received: September 2, 2013 Accepted: October 16, 2013 Published: October 16, 2013 10686

dx.doi.org/10.1021/ac402783t | Anal. Chem. 2013, 85, 10686−10691

Analytical Chemistry

Technical Note

Scheme 1. Schematic Procedure of the (A) Electrochemical Hybridization Assay for the Detection of C533G Mutation of RET Gene and (B) Electrochemical Immunoassay for the Detection of PSA in Human Serum

Figure 1. (A) Schematic of the process for the fabrication of tin sensors, (B) photograph of two tin sensors, (C) SEM image of the tin sensor surface, and (D) comparative stripping voltammograms of a solution containing 40 μg L−1 of Cd(II) on an in situ electroplated MFE on glassy carbon, an in situ electroplated BiFE on glassy carbon, and a microfabricated tin sensor. Supporting electrolyte: 0.1 mol L−1 acetate buffer (pH 4.5); deposition potential: −1.20 V; deposition time: 120 s.

demonstrated through the development of microtitration well assays for the determination of the cancer marker PSA in

human serum and for the detection of C533G mutation of RET gene [which is one of the most frequently encountered 10687

dx.doi.org/10.1021/ac402783t | Anal. Chem. 2013, 85, 10686−10691

Analytical Chemistry

Technical Note

Fabrication of the Sensors. The fabrication of the sensors is schematically illustrated in Figure 1A. Silicon wafers (5″ in diameter, 500 μm in thickness) (Si-Mat silicon materials, Kaufering, Germany) were covered with a 1080 nm-thick layer of SiO2 by means of wet thermal oxidation. Then, a 1 μm thick layer of AZ5214 photoresist (EZ−EM Materials) was spincoated onto the wafer at 7000 rpm for 1 min followed by heating at 95 °C for 10 min. The wafer was then illuminated though an appropriately designed mask with an Hg lamp (AZ210 Mega, U.K.) for 60 s, immersed in AZ726 MIF developer (EZ−EM Materials) for 90 s, washed with doubly distilled water, and dried under nitrogen. Then, a layer of Sn was sputtered on the wafer at a nominal thickness of 300 nm from a Sn target (power of 40 W). The unexposed photoresist film was removed through sonication in acetone for 5 min (liftoff step). A second photolithographic step was performed to define the sensing area and to isolate it from the grip area (Supporting Information); the final diameter of the sensing area of the working electrodes was 5 mm (Figure 1B). Connection of the sensor to the potentiostat was accomplished by a crocodile clip gently attached to the grip area (Scheme 1). Assays. The procedures for the immunoassay for total PSA; the preparation of the BSA−oligonucleotide conjugate, and the hybridization assay are described in detail in the Supporting Information. Stripping Voltammetric Analysis. After the immunoassay or the hybridization assay in the microwells, 50 μL of a 0.05 mol L−1 HNO3 solution were added in each well, and the wells were sonicated for 5 min to release Cd(II) from the QDs. The HNO3 solution containing the dissolved QDs was transferred to a voltammetric cell containing 3 mL of acetate buffer (0.1 mol L −1 , pH 4.5) (Scheme 1) and the electrolytic preconcentration of Cd(II) was carried out at −1.20 V for a predefined time period in stirred solution. After Cd(II) accumulation on the sensor, a square wave voltammetric scan (initial potential, −0.95 V; final potential, −0.68 V; frequency, 50 Hz; pulse height, 40 mV; step increment, 4 mV) was applied to the working electrode and the voltammogram was recorded. Then, the electrode was cleaned from remaining traces of Cd for 10 s at −0.68 V before the next analytical cycle (Figure S2 of the Supporting Information).

mutations in the Greek population related to predisposition/ diagnosis of Multiple Endocrine Neoplasia Type 2 (MEN2)].26 For the DNA assay, a capture bovine serum albumin (BSA) conjugate of an oligonucleotide corresponding to the wild-type sequence of the C533G mutation of the RET gene was hybridized with the biotinylated target oligonucleotide (corresponding to either the wild or the mutant type sequence), followed by labeling with STV-conjugated CdSe/ ZnS QDs (Scheme 1A). For the detection of PSA in human serum, a two-site immunoassay was developed using a mouse monoclonal anti-PSA capture antibody and a biotinylated antiPSA reporter antibody (Scheme 1B); after the immunoreaction, the surface-coupled biotinylated reporter molecule was labeled with STV-conjugated CdSe/ZnS QDs. In both protein and DNA assays, the QDs were dissolved in nitric acid, and the released Cd(II) was quantified at the tin sensor by anodic stripping voltammetry (Scheme 1).

■

EXPERIMENTAL SECTION Reagents. Mouse monoclonal antibodies against human total PSA (Medix Biochemica OY, Kauniainen, Finland and Scripps Laboratories, San Diego, MA, respectively) were used as capture (clone # 8301) and reporter antibodies (clones # 8312 and BP005S), respectively. Normal horse serum was also obtained from Scripps Laboratories. STV-conjugated CdSe/ ZnS QDs (Qdot 585 streptavidin conjugate, 1 μmol L−1) were purchased from Life Technologies (Carlsbad, CA). PSA from human semen (≥99% SDS−PAGE), BSA, and Tween 20 were from Sigma−Aldrich (St. Louis, MO). Sulfosuccinimidyl-4-(Nmaleimidomethyl)cyclohexane-1-carboxylate (sulfo-SMCC), 2iminothiolane·HCl (Traut’s reagent), EZ-Link sulfo−NHS− biotin, and Dextran desalting columns were obtained from Pierce (Rochford, IL). Oligonucleotides were purchased by VCB Biotech (Wien, Austria) and were HPLC-purified. The sequences of the 5′-end amine-modified probes corresponding to the wild-type sequence of the G533C mutation and to the irrelevant sequence were 5′-GTGTGGCGGCCTGGG-3′ and 5′-GTCCAAAGCGAGCAAG-3′, respectively; the sequences of the 5′-end biotinylated target probes corresponding to wildand mutant-type sequence of G533C mutation were 5′CCCAGGCCGCCACAC-3′ and 5′-CCCAGGACGCCACAC-3′, respectively. All the other reagents were purchased from Merck (Darmstadt, Germany). Microtitration polystyrene wells were purchased from Nalge Nunc International (Rochester, NY). Instrumentation. All electrochemical experiments were carried out with an electrochemical analyzer PGSTAT101 (Metrohm Autolab, Utrecht, The Netherlands) connected to a personal computer. Control of the potentiostat, data acquisition, and data manipulation were performed with NOVA 1.8 (Metrohm Autolab). Measurements were performed in a standard glass voltammetric cell equipped with a saturated calomel reference and a Pt wire auxiliary electrode; the setup of the electrodes in the cell is illustrated in Scheme 1. The morphology of the tin deposit was observed with a scanning electron microscope (Leo 440, Carl Zeiss, Germany). The thin film deposition system was the CV401 (Cooke Vacuum Products, South Norwalk, CT) and the tin target was of 99.9% purity (Williams Advanced Materials, Buffalo, NY). An Ascent Fluoroscan fluorescence microtitration plate reader (Labsystems, Finland) equipped with an excitation filter at 390 nm and an emission filter at 584 nm was used for the measurements of the fluorescence signals.

■

RESULTS AND DISCUSSION The polarization window of the tin sensor in the 0.1 mol L−1 acetate buffer (pH 4.5) was assessed through DC voltammetry by scanning the potential in the range from −1.40 to −0.50 V (Figure S1 of the Supporting Information). At potentials more negative than −1.15 V, a sharp increase in the current was observed due to the reduction of hydrogen cations to hydrogen while at potentials more positive than −0.64 V a large oxidation current was observed as the tin on the surface of the sensor started to oxidize. A flat polarization curve was obtained in the potential range from −1.15 to −0.64 V. Since Cd(II) produced an anodic stripping peak with maximum at −0.77 V, the proposed sensor was suitable as a Cd(II) sensor. Inspection by SEM (Figure 1C) revealed that the tin−film had a rather rough surface, which is characteristic for metals with low melting points deposited by sputtering, leading to an increase in the active surface area of the sensor.27 The microfabricated tin sensor was compared to in situ electroplated BiFE and MFE in terms of their respective stripping response for Cd(II) using square-wave anodic stripping voltammetry. Comparative voltammograms obtained 10688

dx.doi.org/10.1021/ac402783t | Anal. Chem. 2013, 85, 10686−10691

Analytical Chemistry

Technical Note

Figure 2. Effect of (A) the deposition time and (B) the deposition potential on the stripping peak current values obtained at the tin sensor for a 10 ng mL−1 PSA calibration solution (red line) or a 10 nmol L−1 oligonucleotide solution (black line). Each point represents the mean value ± sd (n = 3). (C) Plot of the stripping peak current values versus the PSA concentration (voltammograms for PSA concentrations: 0−50 ng mL−1 are provided as the inset). (D) Plot of the stripping peak current values versus the DNA concentration (voltammograms for the DNA concentrations: 0−300 nmol L−1 are provided as the inset). Each point in the calibration plots represents the mean value ± sd (n = 3) after subtraction of the mean signal corresponding to nonspecific absorption.

at the different electrodes are illustrated in Figure 1D, indicating that the sensitivity of Cd(II) detection at the tin sensor was 2.5 and 3 times higher than on the MFE and BiFE electrodes, respectively. To increase the sensitivity and selectivity of the assay, and at the same time to reduce the signal due to nonspecific binding, the duration of the blocking step, the washing and the reaction times, and compositions of buffers were optimized. It was found that 2 h blocking with a solution containing 10 mg mL−1 BSA and subsequent washing with the respective washing buffer could effectively minimize the signal due to nonspecific binding. To obtain the maximum response using the lower possible amount of QDs, the concentration of STV-conjugated QD solution was optimized. It was found that the maximum ratio of specific to nonspecific signal was obtained using 10 nmol L−1 of STV-conjugated QDs. The effect of the electrolytic deposition time and deposition potential was studied using a 10 ng mL−1 PSA calibration solution and a 10 nmol L−1 oligonucleotide solution. For deposition times up to 600 s, the peak current increased almost linearly with the deposition time, while at longer deposition times the current values leveled off (Figure 2A). As the deposition potential became more negative, the signal started to increase reaching plateau values for potentials equal to, or more negative than, −1.20 V (Figure 2B). On the basis of these results, a deposition time of 600 s and a deposition potential of −1.20 V were finally selected for further experiments.

Figure 2 (panels C and D) illustrate typical stripping voltammograms and calibration plots obtained for different PSA and DNA concentrations under optimal experimental conditions. The linear response of the assays extended up to 20 ng mL−1 for PSA and 30 nmol L−1 for DNA, with a correlation of determination of R2 = 0.997 in both cases. The limits of detection (calculated as the concentration of biomolecule corresponding to three times the standard deviation of signal due to nonspecific absorption) were 0.12 ng mL−1 for PSA (equivalent to 6 pg in the 50 μL sample) and 0.08 nmol L−1 for DNA (equivalent to 4 fmol in the 50 μL sample). The limits of detection of this method for PSA and DNA are comparable with those of QD-based electrochemical assays employing MFEs electroplated in situ.12,21 The within-sensor reproducibility [expressed as the % relative standard deviation at the same sensor (n = 8)] was lower than 3.6% over the whole calibration range for both DNA and PSA analysis. The between-sensor reproducibility (expressed as the % relative standard deviation at five different sensors) was lower than 12.3% over the whole calibration range for both DNA and PSA analysis. The average lifetime of each sensor was 10 measurements, thus enabling operation in the semi-disposable mode (Figures S3 and S4 of the Supporting Information). The electrochemical assays for PSA and DNA were compared to the respective ones based on fluorescence detection (Figure S5A and S5B of the Supporting Information). The fluorescence protein and DNA assays presented lower linearity and higher detection limits (0.24 ng mL−1 for PSA and 10689

dx.doi.org/10.1021/ac402783t | Anal. Chem. 2013, 85, 10686−10691

Analytical Chemistry

Technical Note

0.18 nmol L−1 for DNA), demonstrating the enhanced capabilities of the proposed electrochemical sensor for bioanalytical assays. The accuracy of PSA determination in human serum samples using the tin sensors was determined through recovery experiments. For this purpose, two human serum samples from male donors were spiked with known concentrations of PSA. The PSA concentration in the untreated and the spiked samples was determined using the calibration curves of Figure 2C. The % recovery values ranged from 95 to 108% (n = 3), demonstrating satisfactory accuracy of the PSA determination in human serum samples (Table 1).

■

Table 1. Recovery of PSA in Spiked Human Serum Samples

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

sample no

amount added (ng mL−1)

amount determined (ng mL−1)

recovery %

1

0 2.0 5.0 10 0 2.0 5.0 10

0.29 (±0.02) 2.20 (±0.13) 5.46 (±0.33) 11.20 (±0.58) 1.04 (±0.07) 3.15 (±0.15) 6.16 (±0.29) 10.85 (±0.64)

− 95 103 108 − 105 102 98

2

microengineering fabrication approaches. The new tin sensors are environmentally friendly, they overcome the disadvantages of existing voltammetric sensors and biosensors, and offer high sensitivity for the detection of proteins and DNA in complex matrices such as human serum. Extension of this work is in progress for multianalyte detection in a single assay using different types of QDs and for multiplexed detection directly in microwells using miniaturized 3-electrode devices (Supporting Information).

ASSOCIATED CONTENT

S Supporting Information *

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Tel:+306972937675. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The research project is implemented within the framework of the Action ≪Supporting Postdoctoral Researchers≫ of the Operational Program “Education and Lifelong Learning” (Action’s Beneficiary: General Secretariat for Research and Technology), and is cofinanced by the European Social Fund (ESF) and the Greek State.

In DNA sensors aiming at the detection of mutations, the “discrimination ratio” [defined as the ratio of the signal obtained from the mismatching sequences (wild/mutant) to the signal received for the fully complementary sequence (wild/ wild)] should be as high as possible in order to be able to distinguish between wild and mutant sequences in the sample.28 The specific mutation studied here is a single nucleotide polymorphism resulting from replacement of a cytosine (C) by an adenine (A) molecule in the DNA sequence. This difference between the wild and the mutant sequence results in only a slight destabilization of the hybridized mismatching sequences (wild/mutant) as compared to the fully matching ones (wild/wild). Thus, in order, to increase the “discrimination ratio” obtained upon hybridization of immobilized wild-type sequence with the wild and mutant type sequence, the sensors were washed with solutions of sequentially reduced ionic strength (from 1 × HEN to 0.063 × HEN) after hybridization in 1 × HEN buffer (0.1 mol L−1 HEPES buffer, 0.005 mol L−1 NaCl, 1 mmol L−1 EDTA, pH 8.0). Due to the electrostatic nature of interaction that keeps together the DNA strands, it is expected that washing with solutions of reduced ionic strength would affect the hybridization of the fully complementary sequences to a lesser extent than the mismatching sequences, thus increasing the “discrimination ratio”. The “discrimination ratio” achieved for the specific mutations was 5.9 (±0.4) (n = 3), sufficiently high to ensure discrimination between mismatching and fully-matching sequences (Figure S6 of the Supporting Information).

■

REFERENCES

(1) Wang, J. Electroanalysis 2007, 19, 769−776. (2) Liu, A.; Wang, K.; Weng, S.; Lei, Y.; Lin, L.; Chen, W.; Lin, X.; Chen, Y. TRAC−Trends Anal. Chem. 2012, 37, 101−111. (3) Siangproh, W.; Dungchai, W.; Rattanarat, P.; Chailapakul, O. Anal. Chim. Acta 2011, 690, 10−25. (4) Palchetti, I.; Mascini, M. Anal. Bioanal. Chem. 2012, 402, 3103− 3114. (5) Wang, J. Microchim. Acta 2012, 177, 245−270. (6) Gonzalez-Garcia, M. B.; Fernandez-Sanchez, C.; Costa-Garcia, A. Biosens. Bioelectron. 2000, 15, 315−321. (7) Dequaire, M.; Degrand, C.; Limoges, B. Anal. Chem. 2000, 72, 5521−5528. (8) Liao, K.; Cheng, J.; Li, C.; Liu, R.; Huang, H. Biosens. Bioelectron. 2009, 24, 1899−1904. (9) Asefa, T.; Duncanc, C. T.; Sharmac, K. K. Analyst 2009, 134, 1980−1990. (10) Liu, G.; Lin, Y. Talanta 2007, 74, 308−317. (11) Yanga, M.; Javadib, A.; Gongb, S. Sens. Actuators, B 2011, 155, 357−360. (12) Wang, J.; Liu, G.; Wu, H.; Lin, Y. Small 2008, 4, 82−86. (13) Han, M.; Gao, X.; Su, J.; Nie, S. Nat. Biotechnol. 2001, 19, 631− 635. (14) Hansen, J. A.; Wang, J.; Kawde, A.; Xiang, Y.; Gothelf, K. V.; Collins, G. J. Am. Chem. Soc. 2006, 128, 2228−2229. (15) Dong, H.; Yan, F.; Ji, H.; Wong, D. K.; Ju, H. Adv. Funct. Mater. 2010, 20, 1173−1179. (16) Xiang, Y.; Zhang, H.; Jiang, B.; Chai, Y.; Yuan, R. Anal. Chem. 2011, 83, 4302−4306. (17) Yang, C.; Xu, C.; Wang, X.; Xiao, Z.; Gu, B. Electrochem. Acta 2012, 60, 50−54. (18) Lu, D.; Wang, J.; Wang, L.; Du, D.; Timchalk, C.; Barry, R.; Lin, Y. Adv. Funct. Mater. 2011, 21, 4371−4378. (19) Qian, J.; Dai, H.; Pan, X.; Liu, S. Biosens. Bioelectron. 2011, 28, 314−319.

■

CONCLUSIONS AND OUTLOOK In this work, a novel disposable tin microsensor is described for the sensitive electrochemical assays of proteins and DNA using QDs as labels. The proposed detection approach is based on stripping voltammetric detection of Cd(II) released after the immuno− and hybridization assays. This electrochemical sensing platform combines the advantages of QD labels (in terms of electrochemical signal enhancement) and of thin−film 10690

dx.doi.org/10.1021/ac402783t | Anal. Chem. 2013, 85, 10686−10691

Analytical Chemistry

Technical Note

(20) Du, D.; Ding, J.; Tao, Y.; Li, H.; Chen, X. Biosens. Bioelectron. 2008, 24, 863−868. (21) Marin, S.; Merkoci, A. Nanotechnology 2009, 20, 055101. (22) Mercury Time to Act, United Nations Environment Programme, 2013. http://www.unep.org/PDF/PressReleases/ Mercury_TimeToAct.pdf. (23) United Nations Environment Programme. http://www.unep. org/hazardoussubstances/Mercury/Negotiations/tabid/3320/Default. aspx. (24) Kokkinos, C.; Economou, A.; Raptis, I.; Speliotis, T. Electrochem. Commun. 2007, 9, 2795−2800. (25) Kokkinos, C.; Economou, A.; Raptis, I.; Speliotis, T. Electrochem. Commun. 2009, 11, 250−253. (26) Kaldrymides, P.; Mytakidis, N.; Anagnostopoulos, T.; Vassiliou, M.; Tertipi, A.; Zahariou, M.; Rampias, T.; Koutsodontis, G.; Konstantopoulou, I.; Ladopoulou, A.; Bei, T.; Yannoukakos, D. Clin. Endocrinol. 2006, 64, 561−566. (27) Musil, J.; Matous, J.; Vlcek, J.; Koudl, L.; Muller, K. Czech. J. Phys. 1994, 44, 565−574. (28) Relogio, A.; Schwager, C.; Richter, A.; Ansorge, W.; Valcarcel, J. Nucleic Acid Res. 2002, 30, e51.

10691

dx.doi.org/10.1021/ac402783t | Anal. Chem. 2013, 85, 10686−10691