Electrochemical Biosensing Using Amplification-by-Polymerization

Jul 7, 2009 - EDU; phone, 919-515-2993; fax, 919-515-8920 (L.H.). ... Citation data is made available by participants in Crossref's Cited-by Linking s...
0 downloads 0 Views 1MB Size
Download PDF
Anal. Chem. 2009, 81, 7015–7021

Electrochemical Biosensing Using Amplification-by-Polymerization Yafeng Wu,†,‡ Songqin Liu,*,‡ and Lin He*,† Department of Chemistry, North Carolina State University, Raleigh, North Carolina 27695, and School of Chemistry and Chemical Engineering, Southeast University, Nanjing, 210096, P. R. China A novel signal amplification strategy for electrochemical detection of DNA and proteins based on the amplificationby-polymerization concept is described. Specifically, a controlled radical polymerization reaction is triggered after the capture of target molecules on the electrode surface. Growth of long chain polymeric materials provides numerous sites for subsequent aminoferrocene coupling, which in turn significantly enhances electrochemical signal output. Activators generated electron transfer for atom transfer radical polymerization (AGET ATRP) is used in this study for its high efficiency in polymer grafting and better tolerance toward oxygen in air. 2-Hydroxyethyl methacrylate (HEMA) and glycidyl methacrylate (GMA) are examined to provide excess hydroxyl or epoxy groups for aminoferrocene coupling. A limit of detection of 15 pM and 0.07 ng/mL is demonstrated for DNA and ovalbumin, respectively. More than 7-fold signal enhancement in ovalbumin detection has been achieved comparing to the unamplified method. In addition, a more than 5 orders of magnitude of dynamic range is achieved with a linear correlation coefficient (R2) of 0.997 for DNA, and a more than 3 orders of magnitude with R2 of 0.999 for ovalbumin. Together, the results show that the coupling of amplification-by-polymerization concept with electrochemical detection offers great promises in providing a sensitive and cost-effective solution for biosensing applications. Development of electrochemical sensors has a practical significance for their high chemical selectivity, excellent sensitivity, and more importantly, great portability.1-3 With benefits from the rapid growth in the fields of material science, biology, and electrical engineering, electrochemical biosensors have come a long way since the first Clark-type glucose sensor.4 Nevertheless, the increasing demand for screening diseases at their early stage of development calls for ultrasensitive detection of biologically * Corresponding author. E-mail, [email protected]; phone, 919-515-2993; fax, 919-515-8920 (L.H.). E-mail, [email protected]; phone, 86-25-52090613 (S.L.). † North Carolina State University. ‡ Southeast University. (1) Cash, K. J.; Ricci, F.; Plaxco, K. W. J. Am. Chem. Soc. 2009, 131, 6955– 6957. (2) Ricci, F.; Bonham, A. J.; Mason, A. C.; Reich, N. O.; Plaxco, K. W. Anal. Chem. 2009, 81, 1608–1614. (3) Freeman, R.; Li, Y.; Tel-Vered, R.; Sharon, E.; Elbaz, J.; Willner, I. Analyst 2009, 134, 653–656. (4) Clark, L. C.; Lyons, C. Ann. N.Y. Acad. Sci. 1962, 102, 29–45. 10.1021/ac9011254 CCC: $40.75  2009 American Chemical Society Published on Web 07/07/2009

relevant species at an extremely low level of expression, which inevitably leads to intense research efforts toward exploring novel means to enhance detection sensitivity. Successful strategies include the employment of new redox-active probes; the integration of enzyme-assisted signal amplification processes; and the incorporation of nanomaterials to increase the upload of electrochemical tags, etc.5-13 The last approach is particularly effective by introducing multiple redox species per binding event. Similar to nanoparticles, long chain polymeric materials with numerous chemically modifiable functional groups are capable of providing extra redox tags in the same fashion. Indeed, the use of polymer films to increase the loading of capture probes has been routinely employed.14-16 For example, polydiallyldimethylammonium chloride (PDDA) has been used to modify glassy carbon electrodes for tyrosinase immobilization in detection of 2,4-dichlorophenol (2,4-DCP).14 Similar examples include immobilization of glucose oxidase on well-defined poly(glycidyl methacrylate).15 The use of electroactive conjugated polymers as the detection probe has also been demonstrated where multifunctional polythiophene was used to electrostatically interact with DNA duplexes.17,18 The detection limit for DNA of 5 × 10-10 M was reported. We have recently described the use of dynamic polymer growth as a signal amplification method in DNA detection, in which DNA hybridization leads to immobilization of polymerization reaction centers on the surface. Subsequently triggered polymer growth results in local accumulation of monomers that (5) Wang, J. Biosensing for the 21st Century; Springer-Verlag Berlin: Berlin, Germany, 2008; Vol. 109, pp 239-254. (6) Tuncagil, S.; Odaci, D.; Yildiz, E.; Timur, S.; Toppare, L. Sens. Actuators, B 2009, 137, 42–47. (7) Chen, Z. P.; Peng, Z. F.; Luo, Y.; Qu, B.; Jiang, J. H.; Zhang, X. B.; Shen, G. L.; Yu, R. Q. Biosens. Bioelectron. 2007, 23, 485–491. (8) Wang, J.; Meng, W. Y.; Zheng, X. F.; Liu, S. L.; Li, G. X. Biosens. Bioelectron. 2009, 24, 1598–1602. (9) Tang, D. P.; Ren, J. J. Anal. Chem. 2008, 80, 8064–8070. (10) Zhang, S. S.; Zhong, H.; Ding, C. F. Anal. Chem. 2008, 80, 7206–7212. (11) Liao, W.-C.; Ho, J. A. Anal. Chem. 2009, 81, 2470–2476. (12) Chen, L. Y.; Chen, C. L.; Li, R. N.; Li, Y.; Liu, S. Q. Chem. Commun. 2009, 2670–2672. (13) Wu, Y. F.; Chen, C. L.; Liu, S. Q. Anal. Chem. 2009, 81, 1600–1607. (14) Kong, L. M.; Huang, S. S.; Yue, Z. L.; Peng, B.; Li, M. Y.; Zhang, J. Microchim. Acta 2009, 165, 203–209. (15) Xu, F. J.; Cai, Q. J.; Li, Y. L.; Kang, E. T.; Neoh, K. G. Biomacromolecules 2005, 6, 1012–1020. (16) Cui, H. F.; Ye, J. S.; Zhang, W. D.; Sheu, F. S. Biosens. Bioelectron. 2009, 24, 1723–1729. (17) Floch, F. L.; Ho, H. A.; Harding-Lepage, P.; Bedard, M.; Neagu-Plesu, R.; Leclerc, M. Adv. Mater. 2005, 17, 1251–1254. (18) Zhang, L. Y.; Sun, H.; Li, D.; Song, S. P.; Fan, C. H.; Wang, S. Macromol. Rapid Commun. 2008, 29, 1489–1494.

Analytical Chemistry, Vol. 81, No. 16, August 15, 2009

7015

alters the optical property of the substrate.19-22 This approach not only circumvents the slow kinetics and low coupling efficiency encountered using preformed polymers or nanoparticles due to their large sizes, it also provides the flexibility in tuning the degree of amplification based on the level of target molecules present and the assay throughput desired. Successful employment of this approach in DNA detection with fast turnaround, excellent sensitivity, and detector-free visualization has been demonstrated. In addition, because the reaction initiator can be attached to potentially any detection probes through well-established crosslinking reactions, the application scope of this sensing concept extends beyond DNA detection. In this report, we describe coupling of polymerization-assisted signal amplification with electrochemical detection for biosensing, in an attempt to further enhance sensing sensitivity as well as to provide an interface compatible to existing commercial sensing techniques based on electrochemical readouts. Activator generated electron transfer for atom transfer radical polymerization (AGET ATRP) was used in polymerization for its tolerance to oxygen in the air. 2-Hydroxyethyl methacrylate (HEMA) and glycidyl methacrylate (GMA) were used as the monomers to provide excess functional groups (hydroxyl or epoxy groups) for immobilization of electrochemical tags. Aminoferrocene (FcNH2) was chosen as the redox mediator for its well-characterized electrochemical behavior, small size, good stability at low potential, and ease in coupling. Detection of DNA hybridization and ovalbumin binding was demonstrated to illustrate the universal appeal of this amplification-by-polymerization sensing strategy. EXPERIMENTAL SECTION Materials. Gold electrodes were purchased from CHI Instruments, Inc. Au substrates (50 Å chrome followed by 1000 Å gold on float glass) were purchased from Evaporated Metal Films (Ithaca, NY). Dithiothreitol (DTT), triethylamine (TEA), bromoisobutyryl bromide, 11-mercaptoundecanoic acid (HS(CH2)10-COOH), polyethylene glycol (PEG)-8000, ovalbumin, bovine serum albumin (BSA), concanavalin A from Canavalia ensiformis (Con A), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC), 1,1′-carbonyldiimidazole (CDI), N-hydroxysuccinimide (NHS), ascorbic acid (Vc), 2-hydroxyethyl methacrylate (HEMA), CuCl2, tris(2-pyridylmethyl)amine (TPMA), 2,2′-bipyridyl (bpy), sodium carbonate (NaCO3), sodium hydrogen carbonate (NaHCO3), N,N-dimethylformamide (DMF), anhydrous acetone, anhydrous ethanol, sulfuric acid (H2SO4), hydrogen peroxide (H2O2), potassium dihydrogen phosphate (KH2PO4), tris(hydroxymethyl)aminomethane hydrochloride (Tris), and ethylenediaminetetraacetic acid (EDTA) were purchased from Sigma-Aldrich (St. Louis, MO). Glycidyl methacrylate (GMA) was from Alfa Aesar (Ward Hill, MA). Both GMA and HEMA were purified in house to remove (19) Lou, X. H.; Lewis, M. S.; Gorman, C. B.; He, L. Anal. Chem. 2005, 77, 4698–4705. (20) He, P.; Zheng, W. M.; Tucker, E. Z.; Gorman, C. B.; He, L. Anal. Chem. 2008, 80, 3633–3639. (21) Qian, H.; He, L. Anal. Chem. 2009, 81, 4536–4542. (22) Sikes, H. D.; Hansen, R. R.; Johnson, L. M.; Jenison, R.; Birks, J. W.; Rowlen, K. L.; Bowman, C. N. Nat. Mater. 2008, 7, 52–56.

7016

Analytical Chemistry, Vol. 81, No. 16, August 15, 2009

the inhibitor.21,23 Aminoferrocene (FcNH2) was purchased from TCI America (Portland OR). All oligonucleotides were purchased from Integrated DNA Technologies (Coralville, IA) and the sequences were listed in the corresponding schemes. T4 DNA ligase was purchased from Stratagene (La Jolla, CA). Micro Bio-Spin 30 columns were purchased from Bio-Rad Laboratories for DNA purification. Electrochemical Detection of DNA. To prepare initiatorcoupled DNA molecules,19,21 ssDNA or DNA detection probes (100 µM, 3 µL) and the conjugation buffer (1.0 M NaHCO3/ Na2CO3, pH 9.0, 5 µL) were mixed in a centrifuge tube, followed by the addition of 10 µL of the freshly prepared bromoisobutyryl NHS ester solution at 10 mg/mL and 32 µL water. After 2 h reaction at room temperature, DTT (0.01 mM, 23.5 µL) and TEA (1.5 µL) were added into the tube for another 0.5 h reaction. For the control experiments, the same ssDNA or DNA detection probe (100 µM, 3 µL) was directly mixed with TEA (1 µL), followed by a 30 min incubation with 25 µL of DTT and 21 µL of H2O for disulfide bond cleavage. Similarly, the DNA capture probe or the noncomplementary capture probe was treated with DTT and TEA in the same fashion to generate fresh thiol groups at the 3′ end. In all cases, the excess amount of reagent was removed using a Micro Bio-Spin 30 column after the reaction. The final concentrations of ssDNA with or without initiators, the DNA capture probe, and the noncomplementary capture probe were adjusted to 1 µM in KH2PO4 (1 M, pH 4.4) whereas the DNA detection probe with or without initiators was adjusted to 1 µM in a TE buffer (10 mM Tris-HCl, 1 M NaCl, 1 mM EDTA, pH 7.0). The initiator-coupled, thiolated ssDNA, the capture probe, and the noncomplementary capture probe were directly immobilized on the surface by spotting 3 µL of the solution on a clean Au electrode surface or a Au-coated glass slide for an overnight incubation. Both the electrode and the slide were kept in a humid chamber to reduce solvent evaporation. A TE buffer was used as the hybridization buffer. Target oligonucleotides at various concentrations were mixed with the initiator-coupled DNA detection probe (1 µM, 6 µL), followed by spotting on the top of the immobilized capture probes. After 2-5 h of hybridization as specified in the text, the electrode or the glass surface was rinsed with the ligation buffer containing 50 mM Tris-HCl (pH 7.5) and 7 mM MgCl2, followed by 1 h T4 ligation (5% T4 ligase, 50% PEG 8000, 10% ligation buffer, and 35% H2O, total volume 6 µL). The surface was then immersed into urea (8.3 M) for 15 min to remove nonspecifically adsorbed species prior to polymerization. During AGET ATRP detection of DNA, the modified Au electrode or the Au slide was immersed in a glass vial containing CuCl2 (0.03 mM), TPMA (0.03 mM), HEMA (800 µL, 6.58 mM), and DI H2O (800 µL). Ascorbic acid (0.06 mM, 50 µL) was added as the reducing agent to the reaction mixture, followed by immediate sealing of the container with parafilm. In most cases the reactions were conducted for 2 h unless otherwise specified. The electrode or the slide was then thoroughly rinsed and bathed in methanol for 2 h to remove nonspecifically adsorbed monomers. The glass slide was dried prior to ellipsometric measurements of the resulting film thickness. The (23) Lou, X. H.; He, L. Langmuir 2006, 22, 2640–2646.

electrode on the other hand was washed with water, acetone/ water 3:7 (v/v), acetone/water 5:5, acetone/water 7:3, and dry acetone in sequence before being immersed in a 50 µM CDI solution for 2 h. The activated electrode was then extensively washed with dry acetone prior to incubation with 20 mg/mL FcNH2 in anhydrous ethanol for 20 h at room temperature prior to electrochemical measurements, unless otherwise specified. Electrochemical Detection of Ovalbumin. To prepare initiator-conjugated proteins, the NHS-coupled initiator (10 mg/mL in DMF, 10 µL) was added to the protein target solution (ovalbumin or BSA at 10 mg/mL) where the initiator-to-protein molar ratio was controlled at 10:1. The reaction was conducted overnight to allow coupling reaching completion and excess NHS ester to hydrolyze. The concentration of the final protein solution was determined by the UV absorbance at 280 nm, then diluted to different concentrations with a sodium acetate buffer (0.1 M, pH 4.8), as specified in the text. Meanwhile, a clean Au electrode was immersed in a saturated HS-(CH2)10-COOH solution overnight. The electrode, after rinsing with methanol and water, was immersed in a MES buffer (50 mM, pH 4.5) containing EDC (50 mM) and NHS (15 mM) for 30 min, followed by immersion in 1 mg/mL protein solution (Con A or BSA) for another 30 min. After copious rinses with DI water, the same electrode was immersed in a solution containing 1 mg/mL BSA and 1 mg/mL PEG 8000 to block unoccupied adsorption sites. Initiator-coupled ovalbumin or BSA (1 mg/mL, 3 µL) was then spotted at the same location for 30 min. After washes, the modified electrode was immersed in a mixture of CuCl2 (4 mg), bpy (9.38 mg), GMA (800 µL), DMF (400 µL), and DI H2O (400 µL) in a glass container. Ascorbic acid (50 µL) was added to reduce Cu(II) ions and start the AGET ATRP reaction. After a 2.5 h reaction, the electrode was thoroughly rinsed and bathed in acetone for 2 h to remove nonspecifically adsorbed monomers, followed by immersion in 20 mg/mL FcNH2 for 20 h at room temperature prior to electrochemical detection. Instrument. Electrochemical experiments were conducted with a CV-50W voltammetric analyzer (Bioanalytical Systems, Inc.) in a conventional three-electrode cell. The working electrodes were polycrystalline gold disks (diameter 2 mm) embedded in Kel-F rods. A platinum wire and a Ag/AgCl electrode were used as the auxiliary and the reference electrodes, respectively. Prior to each experiment, the Au electrodes were polished with diamond pastes and alumina slurry of 0.05 µm particles on a polishing cloth (Buehler, Lake Bluff, IL), followed by sonicating in water and ethanol. The electrodes were then rinsed with deionized water and dried under a stream of N2, followed by subsequent surface modification. HClO4 at 0.1 M was used as the supporting electrolyte for all experiments. All electrochemical experiments were conducted under nitrogen at ambient temperature. The thicknesses of polymer films on flat Au-coated glass slides were measured with a VB-250 VASE ellipsometer. The instrument irradiated the substrates at a 70° incident angle. The reflective indices of 1.5 and 1.46 were used for the polymer films and biomolecules, respectively, in thickness calculation. A three-layer model provided by the instrument manufacturer was used to fit the experimental data.

Scheme 1. Schematic Illustration of Two Strategies Used in Coupling of Polymerization-Assisted Signal Amplification with Electrochemical Detection for Biosensinga

a (a) In DNA detection, HEMA was used as the AGET ATRP monomer and FcNH2 was coupled to the polymer through CDI. (b) In protein detection, GMA was used as the monomer and FcNH2 was directly coupled to the epoxy groups.

RESULTS AND DISCUSSION Electrochemical Detection of Polymer Growth Atop Initiator-Coupled ssDNA (ssDNA*). GMA was the chosen monomer for AGET ATRP in protein detection for its convenient one-step coupling to aminoferrocene (FcNH2). However its reactivity toward DNA molecules limited its usage in DNA detection; hence, HEMA was used instead. Both monomers provide accessible side-chain functional groups that allow straightforward coupling of FcNH2 (Scheme 1). Figure 1A illustrates the detailed experimental procedure in electrochemical detection of the positive polymer growth atop of ssDNA* (R in Figure 1A) where the same DNA sequence without the initiators, i.e., ssDNA, was used as the control to monitor nonspecific polymer adsorption (C in Figure 1A). Specifically, ssDNA or ssDNA* was directly immobilized on an Au electrode at room temperature through thiol-Au interaction. The electrode was then immersed in an AGET ATRP reaction mixture containing HEMA as the monomer, CuCl2 and TPMA as the reaction catalyst. Also added in the reaction mixture was ascorbic acid (Vc), which served as the reducing agent to bring Cu(II) complexes back to the corresponding ATRP-active, lower oxidative state of catalytic Cu(I) Analytical Chemistry, Vol. 81, No. 16, August 15, 2009

7017

Figure 1. (A) Schematic illustration of polymerization-assisted electrochemical measurement from a ssDNA*-coated electrode after polymer growth and aminoferrocene modification. (B) Cyclic voltammograms of ferrocene measured from the electrode R coated with initiator-coupled ssDNA* (- - -) or the electrode C with the same ssDNA but without initiators (s). Inset: the electrocatalytic response of ferrocene as a function of the electrode scan rate. For experimental details, see the text.

complexes (Supporting Information Scheme 1).21 A layer of poly(hydroxyethyl methacrylate) (PHEMA) film was formed on the surface of the electrode with the film thickness of approximately 18 nm after a 2 h reaction. The hydroxyl groups on the PHEMA side chains were activated by CDI, one of the most commonly used carbonylating reagent that converts free hydroxyl groups in PHEMA into imidazolyl-carbamate groups, which later reacts with aminoferrocene to form relatively stable N-alkyl carbamates (Scheme 1a).24 CDI coupling was carried out in anhydrous acetone to reduce hydrolysis. Concentrated aminoferrocene was dissolved in ethanol instead of water to slow down the competing side reaction. Figure 1B shows a typical cyclic voltammetric curve (R) from the electrode where FcNH2 was permanently anchored on the electrode grafted with PHEMA atop ssDNA*. Clear oxidation and reduction peaks at 0.33 and 0.163 V versus Ag/AgCl were observed, confirming the preserved electrochemical activity of (24) Stollner, D.; Scheller, F. W.; Warsinke, A. Anal. Biochem. 2002, 304, 157– 165.

7018

Analytical Chemistry, Vol. 81, No. 16, August 15, 2009

immobilized ferrocene.25 The ratio of the anodic and cathodic peak intensity is close to unity. A plot of the anodic peak current (ipa) against the scan rate between 0.1 and 0.5 V/s shows good linearity with an R2 value of 0.9998, typical for surfaceimmobilized electron transfer process. Surface coverage of ferrocene is estimated as 2.88 × 10-10 mol based on the equation Γ ) Q/nFA, where Q is the total charge, calculated by integrating the area of the oxidation peak; n is the moles of electrons involved during FcNH2 oxidation (n ) 1); F is the Faraday’s constant; and A is the area of the electrode calculated from the physical size of the Au disk. Given that the coverage of a close-packed monolayer coating has been reported to be 6.8 × 10-10 mol/cm2,26 a FcNH2 coupling efficiency of ∼42% is estimated. It is important to point out that surface modification, especially polymer growth, enlarges the actual electrode surface area, i.e., the A value. Therefore, the calculated surface coverage is overestimated, so is the coupling efficiency of FcNH2 in this case. The peak potential separation (∆Ep) in Figure 1B is 0.167 V. Zero peak separation between the oxidation and reduction peaks is expected for noninteracting electroactive groups attached to a surface when the electron transferring is in rapid equilibrium with the electrode.27 The peak separation observed in our system, despite that the redox species were immobilized on the surface, suggests that electron transfer is significantly hindered by the polymer chains on the surface due to the enlarged distance of the redox centers from the electrode.28,29 Electron hopping between immobilized redox moieties, hopping between the immobilized ferrocene and the electrode surface, and DNA-guided electron transfer between the immobilized ferrocene and the electrode surface are the three possible electron transfer pathways.30,31 While the exact electron transfer mechanism is still under investigation, we suspect the latter two pathways are the primary ones: The immobilized DNA molecules, hence the polymer chains, are far away from each other due to electrostatic repulsion, which leads to large areas of exposed electrode surface between DNA/polymer complexes. When ferrocene is in close proximity to the surface, direct electron hopping between immobilized ferrocene and the electrode surface occurs. When ferrocene is away from the surface due to polymer chain growth, DNA becomes the more effective route to facilitate electron transfer between the grafted polymer chains and the electrode. The observed peak separation is attributed to the movement of flexible polymer chains that introduces local motion and to the presence of the insulating polymer layer. To ensure that the signal was not originated from nonspecifically adsorbed FcNH2, a control experiment in which an electrode coated with ssDNA without initiators (C) went through the same chemical procedure. The absence of discern(25) Zhang, S. X.; Yang, W. W.; Niu, Y. M.; Sun, C. Q. Sens. Actuators, B 2004, 101, 387–393. (26) Wang, Y. F.; Yao, X.; Wang, J. X.; Zhou, F. M. Electroanalysis 2004, 16, 627–632. (27) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, 2nd ed.; Wiley: New York, 2001. (28) Choi, E. Y.; Azzaroni, O.; Cheng, N.; Zhou, F.; Kelby, T.; Huck, W. T. S. Langmuir 2007, 23, 10389–10394. (29) Martin, C. R.; Rubinstein, I.; Bard, A. J. J. Am. Chem. Soc. 1982, 104, 4817– 4824. (30) Blauch, D. N.; Saveant, J. M. J. Am. Chem. Soc. 1992, 114, 3323–3332. (31) Shao, F.; Barton, J. K. J. Am. Chem. Soc. 2007, 129, 14733–14738.

Figure 2. (A) Plots of the electrocatalytic response of ferrocene (b) and polymer film thickness (1) as a function of the AGET ATRP reaction time. (B) The electrocatalytic response of ferrocene as a function of the coupling time. All signals were collected from the electrodes coated with initiator-coupled ssDNA* after AGET ATRP and ferrocene coupling. For other experimental details, see the text.

ible peaks in the CV curve confirms that the previously observed peaks in R were associated with FcNH2 covalently bound to the hydroxylic groups of PHEMA on the electrode. The sensitivity of polymerization-assisted electrochemical detection depends on the length of the polymer chains, i.e., the number of the hydroxyl groups, grafted per biomolecular binding event and the amount of ferrocene molecules attached to each polymer chain. A living polymerization reaction, AGET ATRP offers the benefit of tuning the length of polymer chains by varying the reaction time, i.e., a longer reaction time results in a higher amount of hydroxyl groups per initiator immobilized on the electrode that are available for FcNH2 coupling. As shown in Figure 2A, the electrocatalytic response increases linearly as the polymer chain grows longer initially. The plot however starts to level off after 2 h. This slow-down in the increase of the current intensity could be due to slow-down of polymer growth from radical termination, reduced accessibility of -OH on the side chain of PHEMA for CDI or FcNH2 modification due to steric hindrance, and/or the extended distance between the attached FcNH2 to the Au electrode surface that makes electron transfer less effective when the polymer chains grow longer.31 In a parallel set of experiments, polymers were formed atop ssDNA*-coated Au glass slides in the same fashion. The polymer thicknesses at different time points were measured using ellipsometry, which reflected the absolute amount of materials formed. The results show that the change in film thickness also slowed down at the same 2 h mark and remained relatively constant at 180.5 ± 0.7 Å. The fact that the increase of redox current closely follows the physical increase of polymer mass suggests radical termination as the main cause of the signal plateau.

On the other hand, when the polymer film thickness is constant, the amount of ferrocene attached to each polymer chain can be controlled by incubating the CDI-activated Au electrode in the FcNH2 solution for different lengths of time. Figure 2B shows the catalytic current increased initially with the incubation time, suggesting the coupling of FcNH2 molecules to the polymer layer proceeded beyond skin-depth. The electrocatalytic response reached its maximum after ∼20 h incubation. A further increase in the coupling time resulted in no electrocatalytic response increase, attributed to close-to-completion coupling of FcNH2 to the accessible HEMA side chains or reduced FcNH2 coupling efficiency due to hydrolysis of imidazolylcarbamate groups over an extended reaction time. Under the optimal conditions, the electrocatalytic current was found to be proportional to the logarithm of the surface density of ssDNA* from 10 ppm to 100% surface coverage (Supporting Information Figure 1). In this experiment, the Au electrode was immersed in a solution containing ssDNA* mixed with the unmodified ssDNA of the same sequence at various ratios while the total DNA concentration was maintained at 1 µM. The small size of the initiator at the end of ssDNA* had little impact on the overall diffusion rate of the DNA molecule; thus the same mixing ratio was expected to be preserved on the surface. When the surface density of ssDNA* is below the detectable range, i.e., at 1 ppm, no discernible signal was obtained above the background. Plotting the electrocatalytic current of FcNH2 as a function of the surface density of ssDNA* in the range of 10 ppm and 100% surface coverage yields a curve where a linear fitting was obtained, with the correlation coefficient of ∼0.997. The calculated limit of detection for ssDNA* is ∼4 ppm on the surface, which is equivalent to ∼4 pM or ∼10 amol of ssDNA* in the initial solution mixture detected. Electrochemical Detection of DNA Hybridization. Scheme 2 shows the experimental procedure for DNA detection through hybridization and subsequent polymerization-assisted signal amplification. The DNA capture probe of the complementary sequence to the target DNA was first immobilized on an Au electrode (i.e., electrode R) through thiol-Au interactions. A noncomplementary capture DNA probe to the target DNA was immobilized on a separate electrode (i.e., electrode NC) in a similar fashion. Both electrodes were then immersed into the hybridization buffer containing the target DNA sequence and the initiator-labeled detection probe to form sandwiched DNA duplexes on the R electrode but not on the electrode NC. An additional control experiment was carried out to eliminate any possible side-reaction between FcNH2 and DNA by incubating a DNA capture probe-coated electrode (electrode C) with the target DNA and the detection probe without initiators, which would form similar sandwiched DNA duplexes but without locally immobilized initiators. After ligation to permanently affix initiators on the electrodes, all three electrodes were immersed into an AGET ATRP reaction mixture containing HEMA as the monomer. A layer of PHEMA film was formed on the electrode R that allowed further FcNH2 coupling via CDI after a 2 h reaction. No polymer growth was observed nor was much FcNH2 coupled on the electrodes C and NC. Analytical Chemistry, Vol. 81, No. 16, August 15, 2009

7019

Scheme 2. Polymerization-Assisted Electrochemical Detection of DNA Target through DNA Hybridization

As expected, no discernible peak is observed from either the control electrode (Figure 3A, C or NC). In contrast, pronounced electrocatalytic peaks are observed from the electrode R, albeit the overall current is ∼59% of what was previously obtained from the ssDNA*-coated electrode. Given the peak intensity is proportional to the density of the initiator-coupled DNA on the electrode surface over the range studied (Supporting Information Figure 1), this quantitative result suggests an ∼60% hybridization efficiency

Figure 3. (A) Concept-proof detection of DNA hybridization with the polymerization-assisted electrochemical method. The cyclic voltammograms were collected from the electrode R with cDNA duplexes formed and ATRP initiator immobilized (- - -), the electrode C with cDNA duplexes formed but without ATRP initiators (s), and the electrode NC with noncomplementary capture probes ( · · · ). (B) Quantitative measurements of electrocatlytic currents as a function of the concentration of target DNA. 7020

Analytical Chemistry, Vol. 81, No. 16, August 15, 2009

Scheme 3. Polymerization-Assisted Electrochemical Detection of Protein Binding

for the sandwiched DNA duplexes, an observation consistent with literature values.32 Quantitative analysis of the target DNA concentration shows the measured electrocatalytic current is also proportional to the logarithm of DNA concentration over 5 orders of magnitude from 0.1 to 1000 nM with a correlation coefficient of 0.997 (Figure 3B). The calculated LOD for target DNA detection is ∼15 pM or ∼30 amol of materials in solution, at the same level to other amplified electrochemical sensing results.18 Good dayto-day assay reproducibility was demonstrated when the same experiments were conducted on 3 different days (Supporting Information Figure 2). Electrochemical Detection of Protein Binding. The same detection concept has been extended to the assays beyond genomic detection. Scheme 3 illustrates the major steps undertaken in detection of ovalbumin, a main energy-storage protein that binds to concanavalin A (Con A) through one of its four glycosylation sites. In this assay, HS-(CH2)10-COOH (MUA) was first immobilized on a Au electrode where the carboxylic groups at the distal end of MUA were used to couple with the primary amines in Con A in a carbodiimide reaction to allow permanent attachment of the protein capture probe (as in the electrode R). The similar strategy was used to attach bovine serum albumin (BSA) on another Au electrode as the control (i.e., the electrode C). Meanwhile, ovalbumin was derivatized with the ATRP initiators at a reaction ratio of 1:10. The coupling reaction was near completion after overnight incubation where an average of eight to nine initiators was attached to each ovalbumin molecule, as quantified by electrospray mass spectrometry (data not shown). Both electrodes R and C were incubated with initiator-coupled ovalbumin for 30 min. In addition, in a separate experiment, ovalbumin was preconjugated with FcNH2 at a 1:100 reaction ratio to reach its maximal loading of ferrocene. The Fc-conjugated ovalbumin was incubated with a Con A-coated electrode (i.e., the electrode D) to provide a baseline for sensitivity comparison to the conventional electrochemical detection methods where the redox tags are directly derivatized to protein targets. Three Au-coated glass slides labeled with R, C, and D were prepared in parallel as (32) Jordan, C. E.; Frutos, A. G.; Thiel, A. J.; Corn, R. M. Anal. Chem. 1997, 69, 4939–4947.

Quantitative measurements of ovalbumin show the measured current was proportional to the logarithm of ovalbumin concentrations in the range of 0.1-500 ng/mL with a correlation coefficient of 0.999 (Figure 4B). The calculated limit of detection of ovalbumin is ∼0.07 ng/mL, comparable to the reported conventional ELISA assays where the readout signal is amplified with an enzymatic reaction.33

Figure 4. (A) Concept-proof detection of binding of ovalbumin with the polymerization-assisted electrochemical method. The cyclic voltammograms were collected from the electrode R with Con A/ovalbumin complexes formed and ATRP initiator immobilized (- - -), the electrode C with BSA immobilized (s), and the electrode D with Con A/ovalbumin-Fc complexes formed ( · · · ). (B) Quantitative measurements of electrocatlytic currents as a function of the ovalbumin concentration.

well. All three electrodes and the glass slides were immersed in an AGET ATRP reaction mixture containing glycidyl methacrylate (GMA) as the monomer. Growth of PGMA on the Au slides was confirmed by ellipsometric measurements where the Con A-attached surface R showed a film thickness change from 18.2 ± 0.5 Å after capturing ovalbumin to 155.4 ± 0.7 Å after PGMA formed (∆ thickness ) 137 Å). In contrast, the film thickness increased for less than 10 Å from nonspecific adsorption of monomers on both control slides, C and D. The pendant epoxide groups on PGMA were used for direct coupling of FcNH2. A positive redox peak was observed from the Con A-coated electrode R, whereas no catalytic wave was observed from the BSA-coated electrode C (Figure 4A). For the electrode D where Fc was preconjugated to ovalbumin, the current measured was less than 1/7 of what was obtained from the PGMA-amplified experiment, confirming unambiguously the significantly enhanced assay performance from polymer in situ growth that provided additional ferrocene binding sites for signal amplification.

CONCLUSIONS In summary, we have demonstrated here that a controlled radical polymerization reaction, AGET ATRP, can be used as an effective method to provide multiple binding sites to tag electrochemically active probes for DNA and protein biosensing. Detection limits for both DNA and ovalbumin are in the same range as the established methodologies based on enzymatic amplification; yet the use of polymerization reactions reduces the cost and improves assay robustness. The successful detection of protein or DNA binding events by electrochemical means offers a viable means to interface the sensing platform with commercial portable sensors in the future. It is important to note that extended polymer growth and long ferroncene coupling time were used in most cases in this early study. While the preliminary progress to date has shown the reaction time can be significantly reduced, continuous optimization is still ongoing in the group to improve assay throughput. ACKNOWLEDGMENT The project is partially supported by the Natural Science Foundation (Grant No. 0644865), National Natural Science Foundation of China (Grants Numbers 20675013 and 20875013), and the State Scholarship Fund of China. Technical assistance from Ms. Hong Qian and Mr. Eric Tucker in the Chemistry Department, NCSU, is acknowledged. SUPPORTING INFORMATION AVAILABLE Mechanism of AGET ATRP, quantification of the amount of ssDNA* on the surface, and reproducibility of the biosensor for detection of DNA hybridization. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review May 22, 2009. Accepted June 23, 2009. AC9011254 (33) McBride, M. T.; Gammon, S.; Pitesky, M.; O’Brien, T. W.; Smith, T.; Aldrich, J.; Langlois, R. G.; Colston, B.; Venkateswaran, K. S. Anal. Chem. 2003, 75, 1924–1930.

Analytical Chemistry, Vol. 81, No. 16, August 15, 2009

7021