Covalent Attachment of Biomacromolecules to Plasma-Patterned and

Sep 18, 2012 - Department of Control and Instrumentation Engineering, Korea University Sejong Campus, Jochiwon-eup, Yeongi-gun, Chungnam, 339-700, Rep...
0 downloads 7 Views 3MB Size
Download PDF
Article pubs.acs.org/bc

Covalent Attachment of Biomacromolecules to Plasma-Patterned and Functionalized Carbon Nanotube-Based Devices for Electrochemical Biosensing Joon Hyub Kim,† Joon-Hyung Jin,‡ Jun-Yong Lee,† Eun Jin Park,§ and Nam Ki Min*,†,∥ †

Department of Biomicrosystem Technology and ‡Department of Electrical Engineering, Korea University, Anam-dong, Seongbuk-gu, Seoul, 136-701, Republic of Korea § Department of Physics and Astronomy, Seoul National University, Shillim-dong, Kwanak-ku, Seoul, 151-747, Republic of Korea ∥ Department of Control and Instrumentation Engineering, Korea University Sejong Campus, Jochiwon-eup, Yeongi-gun, Chungnam, 339-700, Republic of Korea ABSTRACT: The interface between biomacromolecules and carbon nanotubes (CNTs) is of critical importance in developing effective techniques that provide CNTs with both biomolecular recognition and signal transduction through immobilization. However, the chemical inertness of CNT surfaces poses an obstacle to wider implementation of CNTs in bioanalytical applications. In this paper, we present a review of our recent research activities related to the covalent attachment of biomacromolecules to plasma-patterned and functionalized carbon nanotube films and their application to the fabrication of electrochemical biosensing devices. The SWCNT films were spray-deposited onto a miniaturized three-electrode system on a glass substrate and activated using highly purified atomic oxygen generated in radiofrequency plasma; this introduced oxygen-containing functional groups into the SWCNT surface without fatal loss of the original physicochemical properties of the CNTs. The carboxylated SWCNT electrodes were then selectively modified via amidation or esterification for covalent immobilization of the biomacromolecules. The plasmatreated SWCNT-based sensing electrode had an approximately six times larger effective area than the untreated SWCNT-based electrode, which significantly amplified the amperometric electrochemical signal. Finally, the efficacy of plasma-functionalized SWCNT (pf-SWCNT) as a biointerface was examined by immobilizing glucose oxidase, Legionella pneumophila (L. pneumophila)-specific antibodies, L. pneumophila-originated DNAs, and thrombin-specific aptamers on the pf-SWCNT-based three-electrode devices. The pf-SWCNT films were found to support direct covalent immobilization of the above-listed biomacromolecules on the films and to thereby overcome the many drawbacks typically associated with simple physisorption. Thus, pf-SWCNT sensing electrodes on which biomacromolecules were covalently immobilized were found to be chemically stable and have a long lifetime.

■

INTRODUCTION Carbon nanotubes (CNTs) are being strongly considered as electrode materials with enormous potential for high-sensitivity electrochemical biosensors owing to their desirable and distinctive characteristics, including high electrocatalytic activity,1,2 promotion of electron-transfer reactions of biomolecules,3,4 and reduced surface fouling effects.5,6 Although the use of CNTs for biosensing applications is promising, their development is still in the nascent stage, and there are many challenges to overcome for the successful commercialization of CNT-based sensor technologies. CNTs themselves are not directly used to detect biomolecules from biological samples because they cannot specifically recognize target biomolecules and do not provide any measurable signals for detecting biomolecules. Therefore, it is of critical importance to develop techniques that can endow CNTs with both a molecular recognition capability and a signal transduction function, which requires well-defined functionalization or modification of CNT © 2012 American Chemical Society

surfaces. Immobilization of biomolecules and the preservation of their biological activity may produce functional surfaces suitable for use in the development of biosensors. Biomolecule immobilization characteristics can be controlled by changing electrode surface parameters such as chemistry and size. Essentially, two approaches are available for the immobilization of biomolecules: noncovalent attachment and covalent binding. A representative noncovalent interaction is physical adsorption of biomolecules around the nanotubes; this can avoid the destruction of the π-conjugated skeleton and electronic properties of CNTs.7 Although this process is relatively simple, it is unfortunately quite unstable in terms of its chemistry, because the adsorption strength between the CNT and biomolecules is considerably weak, making it impossible to Received: May 24, 2012 Revised: September 5, 2012 Published: September 18, 2012 2078

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086

Bioconjugate Chemistry

■

Article

EXPERIMENTAL SECTION Plasma-Functionalized SWCNT Electrode Preparation. Plasma-patterned pf-SWCNT-based three-electrode biosensing devices were prepared on inexpensive glass substrates with low melting points. The SWCNT working electrode on the Pt support, Ag/AgCl reference electrode, and Pt thin-film counter electrode were integrated into the electrode system by conventional microfabrication and the plasma process presented in this work. Scheme 1 shows a drawing of the

achieve the desired long-term stability, and the adsorbed biomolecules are subject to being denatured on the hydrophobic CNT surface. Furthermore, the nonspecific binding of proteins to the CNTs causes the specificity of a sensing electrode surface to markedly deteriorate. To overcome these problems, a variety of techniques have been developed for a stronger conjugation of CNTs with biorecognition molecules. The most reliable immobilization method for biomolecules on CNT is the covalent cross-linking of protein molecules after they are modified with some functional groups, such as a carboxylic acid group. Covalent coupling of the proteins to carboxylic acid functionalized CNTs is probably a more robust approach that can ensure a more stable attachment without any functional degradation of the biomolecules of interest. There are usually two approaches for introducing these functional groups into CNT surfaces.8 The first is oxidative acid treatment, which is often used for the purification and chemical functionalization of CNTs. The second approach involves the exposure of the electrode surface to a plasma atmosphere. Both treatments have previously been shown to introduce oxygen-containing functional groups into CNTs.8 However, they have different effects on CNT surfaces owing to their differing functionalization mechanisms. Treatment of CNTs with strong acids under oxidative conditions introduces functional groups such as carboxylic acids predominantly at the more reactive, open end of the CNTs rather than at their side walls.9 This procedure is an easy method to remove metallic impurities from the CNTs, although it requires a long processing time. Acid treatment with sonication usually shortens the tube length and thereby greatly reduces the aspect ratio of the CNTs. In contrast, O2 plasma can be used to treat the CNT surface more effectively without causing any severe damage and to leave the CNT surface intact by optimizing the plasma treatment condition. Plasma treatment influences the entire surface of the CNT film. This implies that evenly distributed surface functionalization is available with O2 plasma treatment, unlike in conventional acid treatment. Thus, many carboxylic groups can be introduced into the CNT sidewalls as well as at the open ends. Although previous studies discuss the O2 plasma treatment of CNTs, a systematic study remains to be conducted on the effectiveness of O2 plasma-functionalized (pf)-SWCNT films in terms of the immobilization of biorecognition molecules such as antibodies, enzymes, and DNA strands. In this study, we closely examined the effects of O2 plasma treatment on the overall biointerfacial properties of biomolecule-immobilized electrodes based on pf-SWCNT for biosensing applications. To demonstrate the effectiveness and bioelectroactivity of pf-SWCNT electrodes for the versatile covalent immobilization of biomolecules, miniaturized SWCNT-based three-electrode systems consisting of the CNT working electrode, an Ag/AgCl reference electrode, and a Pt counter electrode were designed and fabricated on glass substrates. Glucose oxidase, a Legionella-specific antibody, Legionella pneumophila (L. pneumophila)-originated probe DNA, and a peptide aptamer were covalently immobilized on the plasma-treated SWCNT working electrodes. The efficacy of the immobilization of biorecognition molecules was also extensively characterized. The results suggest that all the biomolecules listed above can be theoretically targeted to achieve some degree of immobilization efficiency.

Scheme 1. Drawing of Fully Integrated Electrochemical Three-Electrode System Composed of Ag/AgCl Thin-Film Reference Electrode, Pt Thin-Film Counter Electrode, and the pf-SWCNT-Patterned Working Electrode

miniaturized SWCNT-based electrode design. The dimensions of each electrode region were carefully determined in accordance with the required current ranges of each electrode. Typically, Pt conducting tracks are 0.25 mm wide and the working electrode is 1.1 mm in diameter with a CNT island patterned on a prepatterned Pt thin film of 1 mm diameter. The dimensions of the reference electrode were 250 μm × 250 μm. The area of the counter electrode was approximately 3.358 mm2. The counter electrode, owing to its relatively large surface area, allowed a wide range of current flows between the counter and working electrodes and a wide input range for the operational potential. The output pads were used to connect the electrodes on the substrate to a pin on a packaged chip or were interfaced to the external circuitry via a three-channel connector that was specifically designed to facilitate signal processing to the external circuitry. All the output pads were positioned near the edge of the substrate. Purified SWCNTs were purchased from Hanwha Nanotech Co. (Korea) and used without further purification. According to the information provided by the manufacturer, the carbon nanotubes were synthesized by an arc-discharge method using iron as catalysts. After acid and heating purification, the SWNTs contained more than 70 wt % carbon nanotubes and fewer than 30% impurities. To prepare the solutions for the spray-coated films, 3 mg SWCNT suspended in 150 mL dichlorobenzene was tip-sonicated for 20 min to obtain an even suspension, and the isolated and dispersed SWCNTs were then separated from the aggregated SWCNTs and the insoluble material by ultracentrifugation (20 000 g, 20 min, 4 °C). The 80-nm-thick SWCNT film was spray-deposited over the substrate using a previously prepared SWCNT suspension. The film thickness of the SWCNT was finely controlled by adjusting the flow rate of the SWCNT suspension while maintaining as constant all other spraying conditions, including 2079

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086

Bioconjugate Chemistry

Article

Figure 1. SEM image of the fully integrated electrochemical three-electrode system (a) and the boundary between the pf-SWCNT working electrode and nitride passivation area (b).

the substrate temperature. Following SWCNT film deposition, the unwanted SWCNT layer deposited around the effective working electrode region was removed by 50 W O2 plasma etching at an oxygen flow rate of 1.689 × 10−2 Pa·m3·s−1 for 5 min. Figure 1a,b shows SEM images of the fabricated SWCNT three-electrode device and magnified images of the edges of the patterned SWCNT working electrode and Ag/AgCl reference electrode, respectively. The clear and well-defined areas of each electrode confirm that the two different O2 plasma conditions used in this study were effective and excellent for producing the SWCNT-patterned three-electrode system. The final step in the electrode fabrication process involved functionalization of the plasma-patterned SWCNT working electrode on the prepatterned Pt support. The chemical properties of the SWCNT surface greatly affect the sensing performance of biosensors because a reliable biointerface should be prepared to provide the SWCNT-modified electrode surface with a better immobilization environment. Strong interaction between biorecognition molecules and the target sensing electrode enhances the sensitivity and selectivity of biosensors with a long lifetime. We employed O2 plasma treatment to meet these chemical as well as physical requirements. When exposed to the radiofrequency (RF) field, molecular oxygen (O2) breaks down into monatomic oxygen (O): O+ and O−. The O atom is chemically the most active element in the plasma phase and readily reacts with any organic hydrocarbon. Atomic oxygen free radicals interact with the surface of the SWCNTs; they break C−C bonds and create active sites for the bonding of functional groups in the plasma. The experimental conditions of conventional plasma treatment can be optimized depending on the generally required plasma conditions for the chemical and physical activation of the intact SWCNT layer. The gas flow rate of molecular oxygen and substrate temperature were fixed in this study at 1.689 × 10−2 Pa·m3·s−1 and 25 °C, respectively. The RF plasma power and plasma treatment time were optimized to ranges of 10−60 W and 10−60 s, respectively. Changes in the chemical composition of the SWCNT surface after plasma treatment were investigated by X-ray photoelectron spectroscopy (XPS). Figure 2a−d shows examples of the XPS survey scans identifying elements present in the SWCNT films before and after O2 plasma treatment at 20 W for 20 s. These were taken for all the samples prior to individual elemental scans (C1s, O1s, etc.). The C1s core-level peak position of the carbon atoms was observed at approximately 285 ± 0.5 eV. The peak position for oxygen was at approximately 533.4 ± 0.5 eV (Figure 2a,b). Functionalized

Figure 2. XPS spectra of the SWCNT film before (a and c) and after (b and d) O2 plasma treatment for 20 s.

SWCNTs had a higher oxygen content of approximately 34.09−42.32 at % relative to untreated SWCNTs (16.37 at %), as can be noted from Table 1. This implies that the oxygen plasma treatment effectively grafted oxygen species onto the SWCNTs. Figure 2c,d displays magnified C1s core level spectra before and after plasma treatment; the pf-SWCNT film had at least five different components even though there were only three before the treatment. The main peak, observed at 285.01 ± 0.03 eV, was assigned to sp2-hybridized graphite-like carbon atoms and hydrogen-bound carbon atoms. Another peak, observed at around 285.77 ± 0.08 eV, was assigned to sp3hybridized carbon atoms. These two peaks clearly reveal the presence of carbon atoms in the absence of oxygen atoms. The three remaining peaks at 287.15 ± 0.21, 288.70 ± 0.13, and 2080

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086

Bioconjugate Chemistry

Article

Table 1. XPS Results of the pf-SWCNT at Various Plasma Treatment Times C1s (% ratio)

O1s (% ratio)

plasma treatment time (s)

CC (sp2)

CC (sp3)

sp3/sp2

C−O−C

−CO

−COO

COx (total)

−OH

−CO or C−O−C

O (total)

0 10 30 60

73.15 52.27 46.66 40.79

5.00 5.88 5.51 6.43

6.83 11.24 11.81 15.77

3.54 5.56 5.78 5.62

1.44 3.22 3.11 2.83

2.46 2.81 2.64

4.98 11.31 12.03 11.18

1.01 4.27 8.99 7.80

15.86 26.28 26.80 33.80

16.87 30.55 35.78 41.60

289.74 ± 0.10 eV were considered to be due to carbon atoms bound to one or more oxygen atoms.10,11 Because electronegative oxygen atoms induce carbon atoms to carry a partially positive charge, extra bonding energy is required owing to the Coulomb force. Hence, the peaks were at a higher bonding energy level. These oxygen-containing carbon atoms can be assigned as those of a hydroxyl group (CO), carbonyl group (CO), and carboxyl group (−COO), respectively. The untreated SWCNTs showed a dominant C1s peak at 285 eV. They also exhibited other peaks that could be attributed to the partial oxidation of CNTs by the acid purification procedure. However, these species were relatively few in number. The ratio of sp2 to sp3 carbon changed according to O2 plasma treatment time; the relevant data are summarized in Table 1. Interaction of O2 plasma with a graphite surface breaks the CC bonds and changes the sp2 hybridization of the C atoms to create defects on the surfaces of carbon nanotubes. Therefore, sp2 carbon decreases as plasma treatment time increases. The main purpose of the O2 plasma treatment of CNTs in terms of chemistry is to introduce oxygen-containing functional groups into the CNT surface. These functional groups are known to promote direct electron transfer of the electrocatalytic activity of redox-active biomolecules.1,12 The density of such groups strongly depends on the applied plasma power and treatment time. In order to estimate the quantitative effect of the key plasma process parameters on the electrochemical properties of the SWCNT electrodes and optimize the plasma process conditions, cyclic voltammetry (CV) for SWCNTbased electrodes prepared under different plasma conditions was performed in a 3 M KCl solution containing 10 mM K3[Fe(CN)6]. Figure 3a,b shows the CV diagrams of the Fe(CN)63‑/Fe(CN)64‑ redox couple of SWCNT films before and after O2 plasma treatment, respectively. The untreated SWCNT sample showed a relatively small peak current (ip,CV; subscript CV means cyclic voltammetry) and a wide range of ΔEp from 220 to 446 mV depending on the applied scan rate. However, after the O2 plasma treatment, the faradaic ip,CV increased considerably with ΔEp close to 70 mV. The maximum ip,CV was observed when the SWCNT electrodes were plasma-treated at 20 W for 20 s. Under this condition, the smallest and most stable ΔEp was obtained. The decrease in ΔEp indicates that the O2 plasma-functionalized CNTs had a faster electron transfer capability than the untreated SWCNTs. The ip,CV increased as the plasma power and treatment time increased to 20 W and 20 s, respectively, and then gradually decreased as the plasma conditions became increasingly unfavorable. The SEM and AFM analyses demonstrated that the CNT surface was chemically etched or destroyed under strong plasma conditions, resulting in the loss of its superior inherent properties, such as a high conductivity and large surface area. This result indicates that, even though samples treated with stronger plasma have a higher concentration of carbonyl/carboxyl groups than samples treated at 20 W for 20

Figure 3. CV characteristics of the SWCNT before the O2 plasma treatment at 10 (1), 20 (2), 50 (3), 100 (4), 200 (5), and 300 mV·s−1 (6) (a), and after the O2 plasma treatment at various plasma power (scan rate = 100 mV·s−1) (b). The electrochemical characteristics of the SWCNT depending on the applied plasma conditions are summarized in (c). Closed rectangles represent ΔEp as a function of the applied RF power. Anodic and cathodic ip,CV are represented by open and closed circles, respectively.

s, they are actually not suitable for sensing electrode materials because the SWCNT surface is seriously damaged by the severe plasma conditions. Figure 3c is the electrochemical optimization process for the SWCNT functionalization in terms of ΔEp and ip,CV showing that O2 plasma treatment at 20 W for 20 s gives optimal performance of the SWCNT film electrode. The effective area of the pf-SWCNT-based working electrode was estimated by analyzing cyclic voltammograms obtained in a 3 M KCl solution containing 10 mM K3[Fe(CN)6]. Figure 4 shows cyclic voltammograms of the Fe(CN)63‑/Fe(CN)64‑ redox reactions occurring on a fully integrated SWCNT-based working electrode, and the effective area of the SWCNTpatterned working electrode was compared with those of an untreated SWCNT-based electrode and a bare Pt thin-film 2081

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086

Bioconjugate Chemistry

Article

plasma treatment significantly affects the physicochemical properties of the electrode surface. The optimized pfSWCNT electrode was found to have an approximately six times larger effective area than an untreated SWCNT-based electrode, which resulted in a significant increase in ip. Immobilization of Biomacromolecules. Three different biorecognition moleculesi.e., glucose oxidase (GOD), Legionella-specific antibody, and L. pneumophila-originated probe DNAwere employed to demonstrate the usefulness of the plasma treatment for effectively immobilizing these biomolecules on an SWCNT-based sensing electrode. Enzymes are proteins that catalyze chemical reactions and have a quaternary structure with a specific amino acid sequence. Immobilized enzymes have been widely used for analysis in clinical diagnosis, environmental monitoring, and the food industry. Generally, the choice of a suitable immobilization strategy is determined by the physicochemical properties of both supporting surface and the enzyme of interest. Formation of amide bonding between the N-terminal of an enzyme and the carboxylic acid-modified substrate appears to be the most frequently used covalent bonding method. The covalent bonding method, which has a relatively strong bond strength, shows better sensitivity and long-term stability. However, the immobilization method needs to be carefully selected because chemical bonding, unlike physical adsorption, may occasionally denature enzymatic three-dimensional conformation. Oxygen plasma treatment encourages the CNT surface to be functionalized with oxygen-containing chemical functional groups such as carboxylic acid, carbonyl, and hydroxyl groups. Therefore, enzyme immobilization by a pf-CNT is categorized as a covalent bonding method. Enzymes, excluding membrane proteins that function in a lipid bilayer, generally function in an aqueous phase, and amino acids with hydrophilic side-chains construct the outer surface of the enzyme. Simultaneously, the inner part of the enzyme is composed of hydrophobic amino acids. This implies that the intrinsically hydrophobic CNT surface should be chemically modified for a more facile immobilization of enzymes. Plasma-based surface treatment has another advantage in terms of adjustment of the plasma power and treatment time to effectively prepare an appropriate surface condition of CNTs. Scheme 2a schematically shows the glucose sensing mechanism by glucose oxidase (GOD) immobilized on the surface of the pf-SWCNT electrode. The pf-SWCNT film was immersed for 2 h in an aqueous solution containing 75 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and 15 mM N-hydroxysulfosuccinimide (sulfo-NHS). Chemical activation of carboxylic acid groups on the pf-SWCNT surface facilitated amide bond formation during the covalent immobilization of GOD on the pf-SWCNT film. The GOD concentration used for immobilization was 500 units/mL, and the prepared glucose sensing device was sufficiently dried under aerobic conditions. The immobilized GOD decomposed glucose molecules into gluconolactone and hydrogen peroxide. The latter was further decomposed into water and oxygen molecules; two electrons were liberated from each glucose molecule by providing the working electrode with 0.6 V vs Ag/ AgCl. Antibodies belonging to a class of protein molecules called immunoglobulins (Ig), also known as γ-globulins, are large Yshaped proteins composed of two specific long amino-acid chains and another two short amino-acid chains. Active sites of an antibody that selectively bind to corresponding antigens are generally located at both ends of the Y-shaped amino acid

Figure 4. Cyclic voltammograms of pf-SWCNT-modified Pt thin-film electrode in a 3 M KCl solution containing 10 mM K3Fe(CN)6 (a). The scan rates are 10 (1), 50 (2), 200 (3), and 300 mV s−1(4). Cyclic voltammograms of bare Pt electrode were also shown for comparison (b). The plots based on the Randles-Sevcik equation are drawn for the pf-SWCNT-based electrode in red and the bare Pt electrode in black (c). Closed circles denote anodic and cathodic peak currents of pfSWCNT-modified Pt thin film, and those of the bare Pt thin film are represented by open circles.

electrode with a thickness of 200 nm. Smaller redox peaks with higher ΔEp were observed for the untreated SWCNT working electrode compared to the other two electrodes. However, the voltammetric currents greatly increased with better resolution after the SWCNT working electrode was treated with O2 plasma. As shown in Figure 4a, cyclic voltammograms of the pf-SWCNT sensing electrode showed a diffusion-controlled reversible reaction with 70 mV of peak potential separation (ΔE) at all scan rates. They are actually comparable to those of bare Pt electrode shown in Figure 4b. The faradaic ip for a reversible electron transfer is given by the Randles−Sevcik eq 1. i p = (2.69 × 105)n3/2AD01/2C0v1/2

(1)

where n is the charge transfer number, A is the surface area of the working electrode (cm2), D0 is the diffusion coefficient (7.6 × 10−6 cm2 s−1 for a Fe(CN)63‑/Fe(CN)64‑ redox couple), C0 is the bulk concentration of the redox species (mol cm−3), and v is the scan rate (V s−1). The plot of ip as a function of v1/2 is linear as long as the diffusion predominantly governs the mass transfer process in the electrical double layer region, as shown in Figure 4c. This linear dependency is a useful diagnostic indicator for characterizing a redox system. The effective working electrode area can actually be determined from either the anodic or cathodic slopes of the curves shown in Figure 4c. The presence of oxygen-containing functional groups at the surface of the pf-SWCNT working electrode because of the O2 2082

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086

Bioconjugate Chemistry Scheme 2.

Article

a

a

(a) Schematic illustration of glucose sensing mechanism occurring on the SWCNT-patterned working electrode on which glucose oxidase molecules are covalently immobilized via EDC/sulfo-NHS chemistry. (b) Schematic drawing of a primary anti-PAL-immobilized pf-SWCNT working electrode. HRP-labeled secondary anti-PAL shall be conjugated with the primary anti-PAL by specific interaction of PAL antigen to both the primary and the secondary anti-PAL. (c) Covalent immobilization scheme and working mechanism of a pf-SWCNT-based DNA sensor. The digtagged oligonucleotide links HRP to the hybridized DNA strands. (d) Schematic illustration of working mechanism of a pf-SWCNT-based aptasensor for monitoring molecular thrombins.

electrode substrate and incubated for 10 min at 37 °C after the immobilized electrodes were rinsed with the washing solution. In the presence of HRP and H2O2, 3,3′,5,5′-tetramethyl benzidine (TMB) was oxidized to form TMB+ and finally TMB2+. Therefore, the more anti-PAL antigen molecules that are present, the more HRP-conjugated monoclonal antibodies that become conjugated to the anti-PAL antigen, which would result in a more enhanced oxidative current flow being observed. Deoxyribonucleic acid (DNA), another major group of biomacromolecules, is composed of pentose, phosphate, purine, and pyrimidine bases and include adenine, guanine, cytosine, and thymine. Nucleosides composed of a base and pentose are elongated by the formation of a phosphate backbone to produce single-stranded DNA (ssDNA). ssDNA may hybridize with cDNA (cDNA) via hydrogen bonding between bases to form double-stranded DNA (dsDNA). Because the DNA backbone does not have any chemical functional group for covalent bonding, the single-stranded probe DNA (pDNA) molecule needs a linker containing the amine terminal for amide bonding to the pf-CNT surface. Voltammetry-based electrochemical quantification is possible as long as labeling with electroactive species is available. Inherently heat-resistive DNA molecules, unlike generally observed enzymes and antibodies, show excellent long-term stability. The same scheme used to prepare an immunosensor was still available for immobilization of probe DNA on the pf-SWCNT surface, as shown in Scheme 2c. The target DNA (Polymerase Chain Reaction (PCR) product) was a 101-mer composed of base pairs (bp) between 86 and 186 of the mip gene (708 bp) from L. pneumophila. According to the NCBI (National Center for Biotechnology Information) nucleotide database, the mip sequences are conserved among 114 L. pneumophila strains, with only four of them having one mismatch in either of the forward or reverse primers used to amplify the 101-mer target,

chains. Immobilized enzymes decompose the substrates into a couple of products, and some of those products subsequently undergo redox reactions on the working electrode surface. Cyclic voltammetry is the most effective method of quantifying the redox products. On the other hand, in the working mechanism of an antigen−antibody binding eventunlike most enzyme−substrate interactionsmany secondary antibodies are surrounding an antigen to suppress its pathogenic activity. The resulting antigen−antibody complex is also bound to the immobilized primary antibodies. No redox process is involved in the antigen−antibody interaction. Hence, conductometry or electrochemical impedance spectroscopy (EIS), which detects the electrical conductivity change, are more commonly used in immunobiosensing. Nevertheless, labeling secondary antibodies with a redox species, including horseradish peroxidase (HRP), allows cyclic voltammetry to be available with an amplified electrochemical signal. Scheme 2b shows an example of antibody immobilization on the SWCNT surface. EDC/sulfo-NHS chemistry was also available, and the activated pf-SWCNTs reacted with amine groups of Legionellaspecific, polyclonal primary antipeptidoglycan-associated lipoprotein (PAL) antibodies to form amide bonds. A working solution of the polyclonal anti-PAL was prepared by dissolving 2 g of polyclonal anti-PAL into 50 mM carbonate−bicarbonate buffer (pH 9.6) solution for 1 h at 37 °C to a concentration of 2 g mL−1. A 200 μL portion of the solution was applied to the activated SWCNTs. The resulting anti-PAL modified SWCNT electrodes were rinsed with a previously prepared washing solution, and active sites of the anti-PAL were blocked in a freshly prepared phosphate-buffered saline (PBS) solution containing 1% BSA for 1 h at 37 °C. Various dilutions of PAL with concentrations of 0.08−10 ng mL−1 were carefully prepared and added to the surface of each electrode, followed by incubation for 1 h at 37 °C. Then, HRP-conjugated monoclonal antibodies (200 L of 2 g mL−1) were added to the 2083

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086

Bioconjugate Chemistry

Article

chemical reaction occurring on a working electrode is equal to the current density (J) divided by nF, where n and F are the number of electrons involved in the electrochemical reaction and the Faraday constant (96 485 C·mol−1), respectively. A higher concentration of the target glucose or GOD molecules present on the sensing electrode results in a larger oxidative current flow owing to the enhanced rate constant. Physically adsorbed GOD molecules are easily rinsed away from the electrode surface during the electrochemical measurements, which results in a dramatic reduction of the sensitivity of the electrode to the target glucose. In addition, the untreated SWCNT surface does not provide the working electrode with a perfect surface for GOD immobilization. As shown in Figure 5,

but they are highly divergent among 240 non-pneumophila species. The mip gene has been suggested to be a viable target for the specific detection of L. pneumophila via the DNA amplification method. The immobilized probe DNA, 101-mer target DNA, and reporter DNA sequences were [5′-amine modification-TAG CTA CAG ACA AGG ATA AGT TGT CTT-3′], [5′-TTT AGC CAT TGC TTC CGG ATT AAC ATC TAT GCC TTG ATT TTT AAA ATT CTT CCC CAA ATC GGC ACC AAT GCT ATA AGA CAA CTT ATC CTT GTC TGT AGC TA-3′], and [5′-TCC GGA AGC AAT GGC TAA A-digoxigenin (dig) modification-3′], respectively. A 200 μL portion of a 100 nM probe DNA solution was applied to the pf-SWCNT surface and left 2 h to covalently immobilize the probe DNA strand to the surface. Then, a 5 μL portion of a hybridization solution containing synthetic target DNA or PCR product was dropped for hybridization and incubated at room temperature for 30 min. Immediately afterward, the hybridized substrates were thoroughly rinsed with deionized water, and the dig-tagged oligonucleotides were applied. Finally, a 5 μL portion of a 1 μg/mL antidig-HRP solution was applied to be conjugated with the dig-tags. All reactions were performed at room temperature (25 ± 1 °C). Aptamers are small nucleic acid sequences that can be relatively easily selected in vitro, and they have three major advantages in biomedical applications.13 Aptamers target a broad range of biological macromolecules, including proteins, nucleic acids, and even whole cells, and they allow easy fabrication of a new type of biosensors targeting various biological molecules. Once an aptamer strand is selected, the aptamer sequence is ready to be synthesized by use of commercially available sources with high reproducibility and purity. Three-dimensional changes in the aptamer’s molecular structure caused by an aptamer−target binding event provides a novel biosensor design with high detection sensitivity and selectivity. To create functionalized aptamers, ferrocene (Fc)and Au-containing nanoparticle groups are usually modified on the electrode.14−16 However, in routine laboratory work, the aptamer-based sensing method does not provide high sensitivity and rapid detection.17−19 Scheme 2d schematically describes the working mechanism of a pf-SWCNT-based aptasensor for monitoring thrombin molecules. Thrombin, a serine protease, has a very important role in the blood coagulation cascade to ultimately stem blood loss. Amineterminated single-stranded DNA (ssDNA) modified with Fc is immobilized on the pf-SWCNT surface at the first step. Then, aptamer strands are hybridized with the immobilized cDNA strands to form much stiffer double-stranded DNAs. Once target thrombin molecules approach, the double-stranded DNAs are dehybridized, and the liberated aptamer strands become bound to the thrombin molecules because of their much stronger binding interaction. The dehybridized singlestranded DNA complementary to the aptamer strands are now flexible and become self-hybridized to form a hairpin structure under a weak basic condition (approximately pH 7.5). The hairpin structure renders electrochemically active Fc labels to be closer to the working electrode surface and generates strong electrochemical redox signals.

Figure 5. Chronoamperograms of GOD immobilized pf-SWCNT electrode at 0.6 V vs Ag/AgCl in a glucose concentration of 0 (1), 0.1 (2), 0.5 (3), 1 (4), 5 (5), and 10 (6) mM. The inset shows calibration curves of various SWCNT-based sensing systems. GOD adsorbed SWCNT layers before and after O2 plasma functionalization are represented by closed (1) and open (2) rectangles, respectively. SWCNT layers covalently modified with GOD molecules before and after O2 plasma functionalization are represented by closed (3) and open (4) circles, respectively. All data points shown in the inset were acquired at 30 s from the chronoamperograms.

the pf-SWCNT covalently modified with GOD molecules showed the best sensitivity for the detection of glucose molecules with a sensitivity of 5.3 μA·mM−1 (R2 = 0.9626) compared to the other three electrodes. pf-SWCNT-Based Immunosensor for Detecting Legionella Species. The enzyme-linked immunosorbent assay (ELISA) is prevalent as a basic immunosensing tool; HRP molecules are the most commonly used labeling agent in a substrate solution containing 3,3′,5,5′-tetramethylbenzidine dihydrochloride (TMB) and hydrogen peroxide. Although Kim et al. reported that L. pneumophila-specific peptidoglycanassociated lipoprotein (PAL) is an excellent target antigen for monitoring Legionnaires’ disease by the ELISA-based diagnostic test,20 the ELISA assay still requires additional pretreatment steps for the separation and purification of PAL fusion from MBP-PAL. Figure 7 shows square-wave voltammetry (SWV) diagrams of antibody-immobilized electrodes in a TMB substrate solution at 100 mV·s−1 for monitoring MBPPAL (Figure 6a) and PAL (Figure 6b) taken from L. pneumophila. Their concentrations were varied from 0.01 to 100 ng·mL−1 to increase the number by ten times. HRPcatalyzed TMB oxidation in the presence of hydrogen peroxide generated two electrons with two corresponding oxidation peaks. The first peak, shown around 0.29 V vs Ag/AgCl,

■

RESULTS AND DISCUSSION Glucose Sensing Based on the pf-SWCNT Working Electrode. The amount of immobilized biocatalyst GOD on the SWCNT surface strongly affects the rate of glucose decomposition reaction. Generally, the rate of the electro2084

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086

Bioconjugate Chemistry

Article

Figure 7. DPV diagrams and corresponding calibration curve of the pfSWCNT-based DNA sensor for monitoring L. pneumophila in 10 mM PBS containing 20 μM MB. The hybridization reaction was performed by applying a target solution containing 100 μM target DNA.

Figure 6. SWV diagrams of pf-SWCNT-modified electrode for direct detection of MBP-PAL (a) and PAL (b). The concentrations of MBPPAL and PAL were varied from 0.01 to 100 ng·mL−1 with increasing the number by ten times.

revealed better linear dependence of the sensitivity on PAL concentrations than the other peak, shown at 0.44 V with respect to Ag/AgCl. The slopes of the PAL-conjugated and MBP-PAL-conjugated electrodes in a target concentration range of 0.01 to 100 ng·mL−1 were 2.03 × 10−2 μA·mm−2 and 3.16 × 10−2 μA·mm−2 per concentration decade, respectively, with a lower detection limit of 5 pg·mL−1. This implies that the pf-SWCNT electrode-based immunosensor shows a sensitivity 1.56 times better for direct monitoring of MBP-PAL without further purification or separation steps and a detection limit almost 1000 times lower than the standard ELISA assay. pf-SWCNT-Based DNA Sensor for Monitoring Legionella Species. The pf-SWCNT-based DNA sensor clearly separated and detected the L. pneumophila-originated target DNA by differential pulse voltammetry (DPV) in a PBS solution (pH 7.4) containing 20 μM methylene blue (MB). However, the untreated SWCNT could not quantitatively differentiate the dsDNA from the ssDNA. The electrochemically active MB was strongly bound to the free guanine base of the DNA backbone. A difference in ip,DPV of approximately 1.8 times was observed for dsDNA and the ssDNA in pf-SWCNTbased DNA sensing with a 50 mV peak potential shift. The DPV-based calibration curve shown in Figure 7 demonstrated that ip,DPV of the pf-SWCNT-based DNA sensor linearly decreased as the target DNA concentration was increased from 10 pM to 100 nM. The resulting sensitivity was approximately 96.6 μA·mM−1 with a correlation coefficient of 0.9786. Negative slopes were quite common in this work because ssDNA is more electrically conductive than dsDNA. pf-SWCNT-Based Thrombin Aptasensor. Faradaic peaks from the SWV diagrams for the redox reactions of the labeled Fc were observed at around −0.6 V vs Ag/AgCl. The ssDNA is more conductive than dsDNA; therefore, the faradaic peak of Fc-labeled ssDNA-modified pf-SWCNT was decreased after hybridization with the aptamer strands, as shown in Figure 8, from curve (a) to curve (b). Dehybridization of aptamer strands by the presence of thrombin molecules induced a stronger binding between the aptamer strands and thrombin. However, ip,SWV was still small (curve (c)), possibly because nonspecifically adsorbed aptamer−thrombin complexes still existed on the electrode surface. After thorough rinsing with hybridization buffer, the Fc-labeled ssDNA modified pf-

Figure 8. SWV characteristics of an aptasensor in a 3 M KCl solution containing 10 mM K3Fe(CN)6. (a) Fc-labeled ssDNA-modified pfSWCNT electrode; (b) aptamer-hybridized electrode; (c) aptamerdehybridized electrode; (d) Fc-labeled ssDNA self-hybridized electrode. Inset shows the plots of ip,SWV as a function of thrombin concentration in fg·mL−1.

SWCNT freely formed a unique hairpin structure to allow the dangling Fc molecules to be closer to the working electrode surface and enhanced ip,SWV (curve (d)). Consequently, ip,SWV increased with increasing thrombin concentration in an incubation solution, and the correlation between ip,SWV and thrombin was linear from 10 fg·mL−1 to ∼1 μg·mL−1.

■

CONCLUSIONS In this study, we introduced various biomacromolecules into the pf-SWCNT working electrode. The SWCNT film was spray-coated, and patterned onto the working electrode substrate. Then, the surface was treated with O2 plasma to prepare the chemically and physically active surface for the immobilization of GOD molecules, L. pneumophila-specific antibodies, L. pneumophila-originated DNAs, and thrombinspecific aptamers. The O2 plasma-activated pf-SWCNT layer was evenly covered with carboxylic acid groups over the entire surface of the SWCNT film with minimal alterations of the surface morphology that allows uniform immobilization of the biomacromolecules. Covalent immobilization of biomacromolecules via EDC/sulfo-NHS chemistry provided reliable attachment of biomacromolecules to the sensing electrode surface, even though the structural distortion of threedimensional conformation of the biomacromolecules should 2085

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086

Bioconjugate Chemistry

Article

Pireaux, J. J. (2006) Oxygen functionalisation of MWNT and their use as gas sensitive thick-film layers. Sens. Actuators, B-Chem. 113, 36−46. (11) Kim, J. A., Seong, D. G., Kang, T. J., and Youn, J. R. (2006) Effects of surface modification on rheological and mechanical properties of CNT/epoxy composites. Carbon 44, 1898−1905. (12) Atlas, R. M. (1999) Legionella: from environmental habitats to disease pathology, detection and control. Environ. Microbiol. 1, 283− 293. (13) Song, S. P., Wang, L. H., Li, J., Zhao, J. L., and Fan, C. H. (2008) Aptamer-based biosensors. TrAC, Trends Anal. Chem. 27, 108− 117. (14) Lu, Y., Li, X. C., Zhang, L. M., Yu, P., Su, L., and Mao, L. Q. (2008) Aptamer-based electrochemical sensors with aptamer-complementary DNA oligonucleotides as probe. Anal. Chem. 80, 1883−1890. (15) Mir, M., Vreeke, M., and Katakis, L. (2006) Different strategies to develop an electrochemical thrombin aptasensor. Electrochem. Commun. 8, 505−511. (16) He, P. L., Shen, L., Cao, Y. H., and Lia, D. F. (2007) Ultrasensitive electrochemical detection of proteins by amplification of aptamer-nanoparticle bio bar codes. Anal. Chem. 79, 8024−8029. (17) Zhang, Y. L., Huang, Y., Jiang, J. H., Shen, G. L., and Yu, R. Q. (2007) Electrochemical aptasensor based on proximity-dependent surface hybridization assay for single-step, reusable, sensitive protein detection. J. Am. Chem. Soc. 129, 15448−+. (18) Li, B. L., Wang, Y. L., Wei, H., and Dong, S. J. (2008) Amplified electrochemical aptasensor taking AuNPs based sandwich sensing platform as a model. Biosens. Bioelectron. 23, 965−970. (19) Zheng, J., Feng, W., Lin, L., Zhang, F., Cheng, G., He, P., and Fang, Y. (2007) A new amplification strategy for ultrasensitive electrochemical aptasensor with network-like thiocyanuric acid/gold nanoparticles. Biosens. Bioelectron. 23, 341−347. (20) Kim, M. J., Sohn, J. W., Park, D. W., Park, S. C., and Chun, B. C. (2003) Characterization of a lipoprotein common to Legionella species as a urinary broad-spectrum antigen for diagnosis of Legionnaires’ disease. J. Clin. Microbiol. 41, 2974−2979.

be taken into consideration. Unlike physisorption, covalent attachment avoids the desorption of biomacromolecules and maintains long-term chemical stability. Additionally, the amide bond formed between the N-terminal of the biomacromolecule and the carboxylated C-terminal of the pf-SWCNT is a much stronger chemical bond than a single bond, because the carbonyl carbon and nitrogen atoms in the amide bond have a partial double-bond character. The double-bond characteristic prohibits the amide bond from free rotation around the principal bond axis, resulting in a mechanically robust biosensing element. Indeed, even distribution of chemical functional groups on the SWCNTs surface and covalent attachment of biomacromolecules to the surface are the major advantages in use of O2 plasma treatment for SWCNTs. The O2 plasma-based patterning and functionalization of SWCNT films presented in this work demonstrated relatively easy immobilization of biomacromolecules, and the potential possibility of mass production makes it promising to use a fabrication method based on O2 plasma-treated SWCNTs in the design of a variety of noble sensing devices.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +82-02-3290-3991. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by grant no. K2090300181211E0100-01710 from the Korea Foundation for International Cooperation of Science and Technology.

■

REFERENCES

(1) Azamian, B. R., Davis, J. J., Coleman, K. S., Bagshaw, C. B., and Green, M. L. H. (2002) Bioelectrochemical single-walled carbon nanotubes. J. Am. Chem. Soc. 124, 12664−12665. (2) Lin, Y. H., Yantasee, W., and Wang, J. (2005) Carbon nanotubes (CNTs) for the development of electrochemical biosensors. Front. Biosci. 10, 492−505. (3) Guiseppi-Elie, A., Lei, C. H., and Baughman, R. H. (2002) Direct electron transfer of glucose oxidase on carbon nanotubes. Nanotechnology 13, 559−564. (4) Nugent, J. M., Santhanam, K. S. V., Rubio, A., and Ajayan, P. M. (2001) Fast electron transfer kinetics on multiwalled carbon nanotube microbundle electrodes. Nano Lett. 1, 87−91. (5) Vashist, S. K., Zheng, D., Al-Rubeaan, K., Luong, J. H. T., and Sheu, F. S. (2011) Advances in carbon nanotube based electrochemical sensors for bioanalytical applications. Biotechnol. Adv. 29, 169−188. (6) Gabay, T., Ben-David, M., Kalifa, I., Sorkin, R., Abrams, Z. R., Ben-Jacob, E., and Hanein, Y. (2007) Electro-chemical and biological properties of carbon nanotube based multi-electrode arrays. Nanotechnology 18. (7) Chen, R. J., Zhang, Y. G., Wang, D. W., and Dai, H. J. (2001) Noncovalent sidewall functionalization of single-walled carbon nanotubes for protein immobilization. J. Am. Chem. Soc. 123, 3838−3839. (8) Dumitrescu, L., Wilson, N. R., and Macpherson, J. V. (2007) Functionalizing single-walled carbon nanotube networks: Effect on electrical and electrochemical properties. J. Phys. Chem. C 111, 12944− 12953. (9) Chou, A., Bocking, T., Singh, N. K., and Gooding, J. J. (2005) Demonstration of the importance of oxygenated species at the ends of carbon nanotubes for their favourable electrochemical properties. Chem. Commun., 842−844. (10) Ionescu, R., Espinosa, E. H., Sotter, E., Llobet, E., Vilanova, X., Correig, X., Felten, A., Bittencourt, C., Van Lier, G., Charlier, J. C., and 2086

dx.doi.org/10.1021/bc300275z | Bioconjugate Chem. 2012, 23, 2078−2086