Biosensors and Bioelectronics 55 (2014) 149–156
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
An ultrasensitive homogeneous aptasensor for kanamycin based on upconversion fluorescence resonance energy transfer Hui Li, De-en Sun, Yajie Liu, Zhihong Liu n Key Laboratory of Analytical Chemistry for Biology and Medicine (Ministry of Education), College of Chemistry and Molecular Sciences, Wuhan University, Wuhan 430072, China
art ic l e i nf o
a b s t r a c t
Article history: Received 11 November 2013 Received in revised form 29 November 2013 Accepted 29 November 2013 Available online 10 December 2013
We developed an ultrasensitive fluorescence resonance energy transfer (FRET) aptasensor for kanamycin detection, using upconversion nanoparticles (UCNPs) as the energy donor and graphene as the energy acceptor. Oleic acid modified upconversion nanoparticles were synthesized through a hydrothermal process followed by a ligand exchange with hexanedioic acid. The kanamycin aptamer (50 -NH2AGATGGGGGTTGAGGCTAAGCCGA-30 ) was tagged to UCNPs through an EDC–NHS protocol. The π–π stacking interaction between the aptamer and graphene brought UCNPs and graphene in close proximity and hence initiated the FRET process resulting in quenching of UCNPs fluorescence. The addition of kanamycin to the UCNPs–aptamer–graphene complex caused the fluorescence recovery because of the blocking of the energy transfer, which was induced by the conformation change of aptamer into a hairpin structure. A linear calibration was obtained between the fluorescence intensity and the logarithm of kanamycin concentration in the range from 0.01 nM to 3 nM in aqueous buffer solution, with a detection limit of 9 pM. The aptasensor was also applicable in diluted human serum sample with a linear range from 0.03 nM to 3 nM and a detection limit of 18 pM. The aptasensor showed good specificity towards kanamycin without being disturbed by other antibiotics. The ultrahigh sensitivity and pronounced robustness in complicated sample matrix suggested promising prospect of the aptasensor in practical applications. & 2013 Elsevier B.V. All rights reserved.
Keywords: Kanamycin Aptasensor Upconversion fluorescence Graphene Fluorescence resonance energy transfer
1. Introduction Kanamycin, an important antibiotic that perturbs protein synthesis as a result of misreading the genetic code and inhibiting translation by binding to the 30S subunit of ribosomal, has been extensively used in human and veterinary medicine to treat infections caused by Gram-positive and Gram-negative bacteria (Fourmy et al., 1998). However, an overdosage of kanamycin is likely to cause kinds of rather serious side effects such as ototoxicity, nephrotoxicity and antibiotic resistance (Oertel et al., 2004). Therefore, carefully monitoring kanamycin concentration in body fluid is necessary and urgent to avoid drug abuse and to assure human health. Up to now, a variety of methods including high performance liquid chromatography (Morovján et al., 1998; Megoulas and Koupparis, 2005; Wang and Peng, 2009), enzyme-linked immunosorbent assay (Watanabe et al., 1999; Loomans et al., 2003; H.W. Chen et al., 2008), capillary electrophoresis (Kaale et al., 2001, 2003; Long et al., 2003), surface plasmon resonance (Raz et al., 2009; Frasconi et al., 2010) and electrochemical immunosensor (Zhao et al., 2011; Wei et al., 2012)
n
Corresponding author. Tel.: +86 27 87217886; fax: +86 27 68754067. E-mail address: [email protected] (Z. Liu).
0956-5663/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bios.2013.11.079
have been developed to detect kanamycin in animal-derived food and biological samples. Nevertheless, in consideration of the inherent imperfectness involved in these technologies like time-consuming, tedious operation or high consumption of reagents, new strategies for kanamycin assay are still desired. In the past several years, aptasensor has emerged as a simple yet efficient technique attracting increasing attention. It is referred to as a new category of biosensor constructed with aptamers as recognition elements. Various targets including metal ions (Miyake et al., 2006; Liu et al., 2007), small organic molecules (Ellington and Szostak, 1992; Shoji et al., 2007), proteins (Wang et al., 2011; Zhan et al., 2012) and even whole cells or microorganisms (Y.Q. Chen et al., 2008; Raddatz et al., 2008) have been detected using their specific aptasensors. Aptamers offer many advantages over antibodies such as low batch–batch variability, small size, low immunogenicity, easy chemical modification and good chemical stability (Keefe et al., 2010). In view of the noticeable advantages of aptamers, Song et al. (2011) screened a singlestranded DNA (ssDNA) aptamer showing high specificity and affinity for kanamycin and then constructed a simple colorimetric aptasensor. Afterwards electrochemical (Zhu et al., 2012; Daprà et al., 2013) and fluorometric (Leung et al., 2013) kanamycin aptasensors were also developed on the basis of this aptamer.
150
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
Despite the achievements made in the aptamer-based kanamycin assay, the sensitivity of these aptasensors, however, still has considerable space to improve considering that the blood drug concentration could be at very low levels. Besides, the practical use of such aptasensors, especially in complicated biological sample matrix (such as body fluid), has not yet been well established. Fluorescence resonance energy transfer (FRET) is recognized as a sensitive and reliable technique for homogeneous bioassays (Jares-Erijman and Jovin, 2003; Schuler and Eaton, 2008). FRET is a non-radiative energy transfer process and occurs between the energy donors and acceptors in close proximity (normally 1–10 nm) through long-range dipole–dipole interactions (Sapsford et al., 2006). Since the energy transfer efficiency (E) is inversely proportional to the sixth power of the distance between the energy and acceptor, a donor-to-acceptor distance change will sharply alter the efficiency resulting in ultrahigh sensitivity in bioassays. Another important merit is that the through-space interaction is mostly independent of intervening solvents and biomolecules, which makes it possible to specifically recognize target molecules in a complicated sample (Lakowicz, 2006). Nonetheless, conventional FRET technique using down-conversion fluorophores as energy donors usually suffer from background interferences which mainly include (1) the co-excitation of energy donor and acceptor due to the small Stokes shift of most fluorescent dyes; (2) the autofluorescence and scattering light from coexisting biomolecules (Frangioni, 2003). As such, the assay sensitivity may be reduced and the application of downconversion FRET technique in complicated matrixes is restricted. To circumvent these restrictions, upconversion nanoparticles (UCNPs) have been introduced to FRET as energy donors (Kuningas et al., 2005a). The near infrared (NIR) excitation and anti-Stokes emission natures of UCNPs endow this kind of fluorophore with strong ability to overcome the above-mentioned shortcomings, so that the UC-FRET technique has shown a promising prospect in homogeneous bioassay in biological samples (Kuningas et al., 2005b, 2006, 2007; Rantanen et al., 2007, 2008, 2009). Herein, we developed an ultrasensitive UC-FRET-based aptasensor for homogeneous determination of kanamycin in human serum, using upconversion nanoparticles as the energy donor and graphene as the energy acceptor. Graphene is an efficient dark quencher and has been proved as a good energy acceptor for UCNPs (Zhang et al., 2011; Wu et al., 2012). Another reason for choosing graphene as energy acceptor is that the aptamer can be facily assembled on its surface through π–π stacking interaction between the nucleobases and the two-dimensional (2D) graphitic carbon nanomaterial, such that the covalent label can be omitted (Xing et al., 2012). The aptasensor featuring flexible configuration and easy operation provided a detection limit of 9 pM in buffered solutions, which was lowered by orders of magnitude as compared to other reported kanamycin aptasensors. What is more, the sensor was applied in diluted serum sample with a detection limit of 18 pM, suggesting its high robustness and practical usage.
2. Materials and method 2.1. Materials Kanamycin was obtained from Amresco (USA). Amine modified kanamycin aptamer (50 -NH2-AGATGGGGGTTGAGGCTAAGCCGA-30 ) was supplied by Sangon Biotechnology Co., Ltd. (Shanghai, China). 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride and N-hydroxysuccinimide were purchased from Sigma-Aldrich. Graphene oxide was from Sinocarbon Graphene Marketing Center (Beijing, China). The other reagents were all from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). The solvents and
reagents were used as received without further purification. All aqueous solutions were prepared in ultrapure water (purified by Milli-Q biocel from Millipore China Ltd.). 2.2. Instrumentation The crystal phase of UCNPs was identified by a Bruker D8 Discover X-ray diffractometer (XRD) with 2θ range from 101 to 701 at a scanning rate of 41/min, with Cu Ka irradiation (k¼ 1.5460 Å). The size and morphology of hexanedioic acid (HDA)-modified NaYF4: Yb, Er upconversion nanoparticles was characterized by HITACHI H-7000FA transmission electron microscopy (100 kV accelerating voltage). FT-IR spectra of oleic acid-UCNPs, HDAUCNPs and graphene oxide after modification were measured on a Nicolet iS10 FT-IR Spectrometer (Thermo Scientific, USA) with the KBr pellet technique. UV–vis absorption measurements were conducted on a UV2550 UV–vis spectrophotometer (Shimadzu Scientific Instruments Inc.). The upconversion fluorescence spectra were recorded on a RF5301 fluorescence spectrophotometer (Shimadzu Scientific Instruments Inc.) using a 980 nm diode continuous-wave (CW) laser (Beijing Hi-Tech Optoelectronic Co., Ltd.) as the external excitation source, with the power being set at 850 mW. 2.3. Synthesis of oleic acid-modified NaYF4: Yb, Er upconversion nanoparticles Oleic acid-coated NaYF4: Yb, Er was synthesized according to a reported procedure (Cui et al., 2011). Briefly, 0.6 mmol of lanthanide oxides Ln2O3 (Y:Yb:Er ¼0.78:0.2:0.02 in mol) were dissolved in hot nitric acid (65 1C) to acquire Ln(NO3)3, and the solvent was evaporated after a 6 h reaction. NaOH (1.2 g, 30 mmol), H2O (5 mL), ethanol (10 mL), and oleic acid (20 mL) were mixed under agitation to form a homogeneous solution. Subsequently, 0.6 mmol (total amounts) of rare-earth nitrate (1.2 mL, 0.5 mol/L Ln(NO3)3) aqueous solution was added under magnetic stirring. Then 1.0 M aqueous NaF (4 mL) solution was added dropwise to the above solution. After agitating for another 10 min, the mixture was transferred to a 50 mL autoclave, sealed, and hydrothermally treated at 160 1C for 12 h. Afterwards the autoclave was cooled to room-temperature naturally, and the products were collected at the bottom of the container by cyclohexane, washed with cyclohexane/ethanol (1:6 v/v) for three times and centrifuged to obtain powder products. 2.4. Ligand-exchange of oleic acid-modified NaYF4: Yb, Er nanoparticles with HDA Ligand-exchange of UCNPs was performed according to the literature report with some modifications (Zhang et al., 2009). The diethylene glycol (DEG, 30 mL) containing HDA (1.5 g) was heated up to 110 1C for 30 min with vigorous stirring under nitrogen. A chloroform solution of oleic acid-modified UCNPs (40 mg, 2 mL) was injected into the above solution, then heated to 240 1C and kept at this temperature for about 5 h until the solution became clear. After the solution was cooled to room temperature, excess ultrapure water was added. The HDA-UCNPs were isolated by centrifugation and decantation. Finally the product was washed three times with pure water and redispersed in 4 mL of deionized water for further use. 2.5. Preparation of hydrothermal converted graphene from graphene oxide The conversion was accomplished through a hydrothermal route (Zhou et al., 2009). The pH of 40 mL of 0.5 mg/mL graphene
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
151
oxide aqueous solution was adjusted to 8.0 with 1 M NaOH. Then it was transferred to a Teflon-lined autoclave and heated at 180 1C for 6 h. Afterwards the autoclave was cooled to room temperature naturally. The hydrothermally treated graphene was found to precipitate at the bottom of the autoclave as a black powder, possibly due to the low solubility in super-fluidic water under hydrothermal conditions. The as-obtained graphene could be readily dispersed by ultrasonification in water.
same experimental procedures. For the determination of kanamycin in human serum, freshly collected serum sample from a healthy volunteer (provided by School of Medicine, Wuhan University) was 100-fold diluted with Tris–HCl buffer without further processing, and the same assay procedure as in the aqueous solution was followed. Standard addition method was adopted to determine the concentration of kanamycin in real serum samples.
2.6. Preparation of sulfonated graphene
3. Results and discussion
The sulfonic acid groups were introduced by a simple sulfonation method (Si and Samulski, 2008). For short, 40 mg of sulfanilic acid and 15.8 mg of sodium nitrite were first dissolved in 10 mL of NaOH solution (0.25%), and then added to 10 mL of 0.1 M HCl on ice bath under stirring. After reaction for 10 min, the aryl diazonium salt solution was added to 40 mL of 0.5 mg/mL graphene dispersion and was kept stirring for 2 h on ice bath. After filtration and rinsing, the purified sulfonated graphene was diluted to 0.5 mg/mL and stored at 4 1C.
3.1. Principle of the UC-FRET based aptasensor for kanamycin
2.7. Attachment of kanamycin aptamer to upconversion nanoparticles The amine modified kanamycin aptamer was covalently conjugated to HDA-UCPs following the standard EDC–NHS conjugation protocol (Hermanson, 2008). Firstly, 2 mg of HDA-UCPs was dissolved in 4 mL of MES buffer (10 mM, pH 5.5), and then 0.32 mg (0.4 mM) of EDC HCl and 0.46 mg (1 mM) of NHS were added to the solution to activate the carboxyl groups on HDA. The mixture was incubated at room temperature with gentle shaking for 45 min. After centrifugation, the activated HDA-UCNPs were washed with pure water for three times and then dispersed in 4 mL of HEPES buffer (10 mM, pH 7.2) containing 2 nmol amino modified kanamycin aptamer. The reaction lasted overnight at room temperature with slow shaking. In order to block the excess NHS, 10 mg of Tris was added into the reaction mixture. Subsequently, the kanamycin aptamer labeled HDA-UCNPs were collected by centrifugation and washed with ultrapure water for three times and finally diluted in 4 mL of Tris–HCl buffer (10 mM, pH 8). The 0.5 mg/mL UCNPs–aptamer complex was stored at 4 1C for further use.
The UC-FRET aptasensor was constructed based on the conformation change of the kanamycin aptamer before and after interacting with kanamycin. Generally, aptamers are singlestranded DNA or RNA nucleic acids which were screened in vitro through a process named as Systematic Evolution of Ligands by Exponential Enrichment (SELEX) (Ellington and Szostak, 1990). Up to now, a variety of aptamers have been selected towards a wide range of targets, including metal ions (Miyake et al., 2006; Liu et al., 2007), proteins (Wang et al., 2011; Zhan et al., 2012), peptides (Mendonsa and Bowser, 2005), toxins (Tang et al., 2007) and so on. The occurrence of binding between aptamers and their corresponding targets is usually accompanied by the conformation change of aptamers. The π–π stacking interaction usually happens between random-coiled aptamers and the graphitic 2D nanomaterial with sp2 carbon atoms and a large planar structure, which offers the advantage of simplifying experimental approaches. In the absence of kanamycin, the random-coiled aptamers are spontaneously adsorbed onto graphene surface through π–π stacking and thus bring the UCNPs (energy donor) and graphene (energy acceptor) in close proximity, leading to the quenching of UNCPs emission (Scheme 1). After the introduction of kanamycin into the UCNPs–graphene FRET system and the preferential binding between the aptamer and kanamycin, the aptamer is transformed to a hairpin structure which possesses much lower affinity toward graphene (Song et al., 2011; Zhu et al., 2012). Under this circumstance, the distance between UCNPs and graphene is enlarged blocking the energy transfer. As a consequence, the fluorescence emission of UCNPs is restored in a kanamycin concentration dependent manner.
2.8. Kanamycin detection in aqueous solution and diluted human serum In order to select an appropriate concentration of graphene to do the following fluorescence recovery experiments, a fixed amount of UCNPs–aptamer (0.012 mg/mL) was incubated with increasing amounts of sulfonated graphene in 10 mM Tris–HCl buffer (pH 8, 10 mM NaCl) at room temperature for 1.5 h, and then upconversion fluorescence measurements were carried out. In a typical FRET assay process, various concentrations of kanamycin were first mixed with the UCNPs–aptamer (0.012 mg/mL) conjugates in 10 mM Tris–HCl buffer (pH 8, 10 mM NaCl) and the mixture was incubated at 4 1C for 2 h. Afterwards sulfonated graphene was added individually into the above mixtures with an ultimate concentration of 0.016 mg/mL followed by incubation for another 1.5 h. Finally, the fluorescence intensity of the reaction mixture was recorded under the excitation of 980 nm. To examine the specificity of the FRET aptasensor, a list of other antibiotics including ampicillin, streptomycin sulfate, gentamicin, aureomycin, oxytetracycline and tetracycline were added into the UCNPs– aptamer–graphene system in place of kanamycin following the
Scheme 1. Schematic illustration of the kanamycin aptasensor on the basis of FRET from aptamer-linked UCNPs to graphene.
152
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
3.2. Characterization of the energy donor and energy acceptor To realize the above design, we first synthesized hexanedioic acid modified NaYF4: Yb, Er and sulfonated graphene. We used this kind of UCNPs because the rare earth ions co-doped NaYF4 nanocrystals are so far known to be the most efficient upconversion phosphors due to the low phonon energy of NaYF4 as host matrix. It is worth noting that in the lanthanide-doped UCNPs, it is the doping ion (Er3þ ) that acts as the emitter. Taking into consideration the distance dependence of FRET process, the size of particles can have remarkable effect on the energy transfer efficiency. It is understandable that a smaller particle size will be more favorable for the energy transfer since more emitting ions are located near the surface of the particles within the effective FRET distance. Compared to most UCNPs-based FRET assays previously reported by our lab and others (Wang et al., 2011; Kuningas et al., 2005a, 2005b, 2006, 2007; Rantanen et al., 2007, 2008, 2009; Zhang et al., 2011; Peng et al., 2011; Yuan and Liu, 2012; Wang et al., 2012, 2013), where the UNCPs were normally larger than 50 nm in diameter, we hereby prepared relatively smaller oleic acid modified-NaYF4: Yb, Er nanoparticles through a developed route: a hydrothermal synthesis of oleic acid-coated UCNPs followed by a ligand exchange process with hexanedioic acid. The TEM image (Fig. 1a) reveals spherical particles with the diameter ranging from 10 to 20 nm and good dispersibility in aqueous solution. The X-ray diffraction (XRD) pattern indicates that the synthesized NaYF4: Yb, Er nanoparticles were mainly with cubic phase (Fig. 1b). The existence of oleic acid and hexanedioic acid
molecules on the surface of the nanoparticles were confirmed by FT-IR analysis (Fig. 1c). The peak at ν¼3008 cm 1, attributable to the presence of ¼C–H stretching vibration, indicated the existence of oleic acid (the upper curve in Fig. 1c), while its disappearance in the lower curve verified the occurrence of ligand exchange process. What is more, the appearance of the stretching mode of the –COOH group at 1735 cm 1 indicated the presence of free –COOH groups on hexanedioic acid modified UCNPs. It is known that the conjugated network of graphene oxide can be largely destroyed in the preparation process from exfoliation of graphite through oxidation with strong acids. Therefore, in order to sufficiently make use of the π–π stacking interaction between nucleic acids and the carbon nanomaterial, it is necessary to recover the conjugated sp2 network by reduction of graphene oxide (Zhou et al., 2009). On the other hand, however, the reduced graphene oxide (RGO) always exhibits lowered water dispersibility, which is unfavorable for homogeneous sensing. To balance the two aspects, we further modified the RGO through sulfonation. The surface functionalization of RGO with sulfonic groups was confirmed by FT-IR spectra (Fig. 1d) with the peaks at 1170 cm 1, 1122 cm 1 and 1037 cm 1(two νS–O and one νS–phenyl ) in the upper curve. 3.3. Construction of the UC-FRET aptasensor In this system the aptamer chains are tagged to UCNPs surface, as shown in Scheme 1. To preclude possible steric hindrance
Fig. 1. (a) TEM image of HDA-modified NaYF4: Yb, Er upconversion nanoparticles. (b) XRD pattern of NaYF4: Yb, Er nanocrystals: Δ, cubic phase (JCPDS file no. 77-2042). (c) FT-IR spectra of oleic acid coated-UCNPs (the upper curve) and HDA modified UCNPs (the lower curve). (d) FT-IR spectra of RGO before (the lower curve) and after (the upper curve) modification with sulfonic groups.
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
153
Fig. 2. (a) UV–vis absorption spectra of UCNPs and UCNPs–kanamycin aptamer conjugate (0.5 mg/mL). (b) Fluorescence quenching of UCNPs–kanamycin aptamer (0.012 mg/mL) with various concentrations of sulfonated graphene (0, 0.003, 0.008, 0.013, 0.016 mg/mL). (c) Time dependence of the fluorescence quenching degree with 0.012 mg/mL UCNPs–kanamycin aptamer and 0.016 mg/mL sulfonated graphene. All experiments were performed in Tris–HCl buffer (10 mM, 10 mM NaCl, pH 8.0) under excitation at 980 nm.
caused by the relatively large nanoparticles during the recognition of aptamer by the target, hexanedioic acid was selected as the surface ligand of UCNPs because it offers a four-methylene flexible spacer. The exterior free carboxyl of hexanedioic acid was employed to covalently link with the aptamer, and the successful conjugation was confirmed by the 260 nm absorption on the UV–vis spectrum (Fig. 2a). To investigate the energy transfer between the donor–acceptor pair, different amounts of graphene were added into a fixed concentration of UCNPs–aptamer (0.012 mg/mL). After simply incubating the mixture for a short while, a graphene amount-related quenching of UCNPs fluorescence was observed (Fig. 2b). The driving force causing the donorto-acceptor assembly was assigned mainly to π–π stacking of the nucleobases on graphene. Since both the phosphate backbone of the DNA chain and the sulfonated graphene surface were negatively charged, electrostatic attraction was excluded. Such a design of electric charge was able to ensure the separation of the aptamer from graphene surface after interacting with the target. The time dependence of fluorescence quenching (Fig. 2c) revealed that the interaction reached to equilibrium in about 10 min, which was a quite quick event that could help to reduce the time consumption of the assay. A graphene concentration (0.016 mg/mL) corresponding to the quenching efficiency of 47% was selected for subsequent experiments. Although higher quenching efficiencies could be obtained by increasing the concentration of graphene, the use of
excessive quencher would contribute to nonspecific quenching which is unfavorable for subsequent fluorescence restoration. 3.4. Kanamycin determination in buffer solution Kanamycin detection was first conducted in Tris–HCl buffer with the above proposed UC-FRET aptasensor. The introduction of kanamycin into the UCNPs–aptamer–graphene FRET system led to fluorescence restoration of UCNPs (Fig. 3a). It relied on the conformation change of kanamycin aptamer into a hairpin structure after the binding between kanamycin and aptamer, which greatly weakened the π–π stacking interaction. In this condition the energy donor and acceptor were separated apart from each other, and thus the FRET process was inhibited. Because of the homogeneous nature of FRET, the assay was a very simple and straightforward procedure without any separation or repeated reagent addition. A linear relationship was obtained between the increasing fluorescence intensity of UCNPs and the logarithm of kanamycin concentration in the range from 0.01 to 3 nM (Fig. 3b). The detection limit was 9 pM, which was calculated as the concentration corresponding to 3 times of the SD (standard deviation) of the background signal from seven independent measurements. The sensitivity of this UC-FRET aptasensor was greatly enhanced compared to the reported kanamycin aptasensors (Song et al., 2011; Zhu et al., 2012; Daprà et al.,
154
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
Fig. 3. (a) Fluorescence recovery of the biosensor with the introduction of different amounts of kanamycin (0, 0.01, 0.03, 0.1, 0.3, 1, 3 nM) in Tris–HCl buffer (10 mM, 10 mM NaCl, pH 8.0). (b) The linear relationship between the fluorescence recovery (at 546 nm) and the logarithm of kanamycin concentration within the range of 0.01–3 nM in Tris–HCl buffer, data were presented as average7 SD from three independent measurements. (c) Fluorescence recovery of the biosensor with the introduction of different amounts of kanamycin (0, 0.03, 0.05, 0.08, 0.5, 1, 3 nM) in 100-fold diluted human serum. (d) The linear relationship between the fluorescence recovery (at 546 nm) and the logarithm of kanamycin concentration within the range of 0.03–3 nM in the diluted serum, data were presented as average7 SD from three independent measurements. All experiments were performed in the presence of 0.012 mg/mL UCNPs–kanamycin aptamer and 0.016 mg/mL sulfonated grapheme. Excitation wavelength: 980 nm.
2013; Leung et al., 2013). Such a performance improvement of the proposed sensor, in our opinion, also benefited from the adoption of the four-methylene spacer of hexanedioic acid. Compared to the situation where the probes (e.g., aptamers) are directly linked to the nanoparticles, the spacer acts as an isolation region eliminating (at least partly) the steric effect the particles may have on the probing process. Note that the length of the spacer can have contradictory influences on the sensing, i.e., a longer spacer provides more complete elimination of the steric hindrance yet the energy transfer efficiency may be reduced as well. In this sense, we believe that the selection of surface ligand can be another necessary job in the future for such UCNPs-based FRET sensors. 3.5. Specificity of the UC-FRET aptasensor towards kanamycin In order to examine the specificity of this UC-FRET aptasensor towards kanamycin, some analogous antibiotics including ampicillin, streptomycin sulfate, gentamicin, aureomycin, oxytetracycline and tetracycline were introduced individually into the aptasensor instead of kanamycin in the aqueous solution. Even when these substances were with a concentration (1 nM) 10 times higher than that of kanamycin (0.1 nM), they did not cause obvious restoration of the
UNCPs fluorescence intensity (Fig. 4). The pronounced selectivity of assay was attributed to the highly specific recognition between the aptamer and the target. At the same time, this experimental result also confirmed that the conformation alteration of the aptamer played a key role in the sensing. 3.6. Detecting kanamycin in serum sample Determination of kanamycin concentration in body fluids is essential for clinical diagnosis and survey. We chose human serum as a representative sample to examine the practical applicability of the UC-FRET aptasensor. Serum is a known complicated biological matrix containing various biomolecules which tend to raise autofluorescence and scattering light in optical sensing. Although, as described above, the FRET efficiency is basically independent of the coexisting biomolecules, those background optical signals can cause serious interference to down-conversion fluorescence-based assays. The freshly collected clean (which means no kanamycin contained) serum sample was 100-fold diluted with 10 mM Tris–HCl buffer to maintain the pH, which was used as the assay medium to check if a linear calibration could still be set up. With an identical assay procedure as that in aqueous solution, a similar fluorescence restoration was observed (Fig. 3c) and a linear relationship between the fluorescence intensity
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
155
Fig. 4. (a) Normalized fluorescence spectra of the sensor in the presence of different antibiotics. The highest curve is the spectrum in the presence of 0.1 nM kanamycin, and the other seven curves (which are overlapped) are the spectra of the sensor and that in the presence of 1 nM other antibiotics (ampicillin, streptomycin sulfate, gentamicin, aureomycin, oxytetracycline and tetracycline). (b) Relative fluorescence intensity ((Fother antibiotics F0)/(Fkanamycin F0)) of the UC-FRET aptasensor in the presence of different substances, where F0 is the fluorescence intensity in the absence of kanamycin or other antibiotics. Data were presented as average7SD from three independent measurements. The concentration of the other antibiotics was 1 nM, and the concentration of kanamycin was 0.1 nM. Experiments were conducted in Tris–HCl buffer (10 mM, 10 mM NaCl, pH 8.0) under excitation at 980 nm.
Table 1 Determination of kanamycin in spiked serum samples. Sample no.
Added (nM)a
Found (nM)
Recovery (%)
RSD (%) n¼ 3
1 2 3
0.8 1 3
0.75 0.93 3.08
94 93 103
1.5 0.6 1.7
a
Mean value of three determinations by the biosensor.
and the logarithm of kanamycin concentration was also obtained in the range from 0.03 nM to 3 nM, with the detection limit being calculated as 18 pM (Fig. 3d). It is also seen that the detection in serum sample exhibited good reproducibility. Though the linear range was slightly narrowed down and the detection limit was higher than that in the aqueous solution, the aptasensor still showed enough robustness in the serum sample which could be attributed to three merits of the sensor: (1) the homogeneous feature of FRET technique, (2) the circumvention of background interference by UCNPs and (3) the firm and stable aptamer-target recognition. Then the standard addition method was applied to check the validity of the method. The recovery of kanamycin in three spiked serum samples and the RSD levels (n¼ 3) are presented in Table 1. The above results indicated that this UC-FRET aptasensor for kanamycin could find practical applications hereafter.
4. Conclusions In summary, we have constructed an ultrasensitve aptasensor for kanamycin detection based on the fluorescence resonance energy transfer from upconversion nanoparticles to graphene. Combining the advantages of aptamers and the UC-FRET technique, the aptasensor exhibited excellent sensing performances in both aqueous solutions and body fluid sample. The homogeneous sensor also had the merits of simple configuration, straightforward operation and time and cost efficiency, suggesting promising prospect in practical use.
Acknowledgment This work was supported by the National Natural Science Foundation of China (Nos. 21075094, 21375098), the National
Basic Research of China (973 program, No. 2011CB933600) and the Program for New Century Excellent Talents in University (NCET-11-0402). References Chen, H.W., Medley, C.D., Sefah, K., Shang, G.D., Tang, Z.W., Meng, L., Smith, J.E., Tan, W.H., 2008. ChemMedChem 3, 991–1001. Chen, Y.Q., Wang, Z.Q., Wang, Z.H., Tang, S.S., Zhu, Y., Xiao, X.L., 2008. J. Agric. Food Chem. 56, 2944–2952. Cui, S.S., Chen, H.Y., Gu, Y.Q., 2011. J. Phys. Conf. Ser. 277, 012006. Daprà, J., Lauridsen, L.H., Nielsen, A.T., Rozlosnik, N., 2013. Biosens. Bioelectron. 43, 315–320. Ellington, A.D., Szostak, J.W., 1990. Nature 346, 818–822. Ellington, A.D., Szostak, J.W., 1992. Nature 355, 850–852. Fourmy, D., Yoshizawa, S., Puglisi, J.D., 1998. J. Mol. Biol. 277, 333–345. Frangioni, J.V., 2003. Curr. Opin. Chem. Biol. 7, 626–634. Frasconi, M., Vered, R.T., Riskin, M., Willner, I., 2010. Anal. Chem. 82, 2512–2519. Hermanson, G.T., 2008. Bioconjugate Techniques, second ed. Academic, San Diego. Jares-Erijman, E.A., Jovin, T.M., 2003. Nat. Biotechnol. 21, 1387–1395. Kaale, E., Schepdael, A.V., Roets, E., Hoogmartens, J., 2001. J. Chromatogr. A 924, 451–458. Kaale, E., Schepdael, A.V., Roets, E., Hoogmartens, J., 2003. Electrophoresis 24, 1119–1125. Keefe, A.D., Pai, S., Ellington, A., 2010. Nat. Rev. Drug Discov. 9, 537–550. Kuningas, K., Rantanen, T., Karhunen, U., Lövgren, T., Soukka, T., 2005a. Anal. Chem. 77, 2826–2834. Kuningas, K., Rantanen, T., Ukonaho, T., Lövgren, T., Soukka, T., 2005b. Anal. Chem. 77, 7348–7355. Kuningas, K., Ukonaho, T., Päkkilä, H., Rantanen, T., Rosenberg, J., Lövgren, T., Soukka, T., 2006. Anal. Chem. 78, 4690–4696. Kuningas, K., Päkkilä, H., Ukonaho, T., Rantanen, T., Lövgren, T., Soukka, T., 2007. Clin. Chem. 53, 145–146. Lakowicz, J.R., 2006. Principles of Fluorescence Spectroscopy, third ed. SpringerVerlag, Berlin, Heidelberg. Leung, K.H., He, H.Z., Chan, D.S.H., Fu, W.C., Leung, C.H., Ma, D.L., 2013. Sens. Actuators B: Chem. 177, 487–492. Liu, J., Brown, A.K., Meng, X., Cropek, D.M., Istok, J.D., Waston, D.B., Lu, Y., 2007. Proc. Natl. Acad. Sci. USA 104, 2056–2061. Long, Y.H., Hernandez, M., Kaale, E., Schepdael, A.V., Roets, E., Borrull, F., Calull, M., Hoogmartens, J., 2003. J. Chromatogr. B 784, 255–264. Loomans, E.E.M.G., Wiltenburg, J.V., Koets, M., Amerongen, A.V., 2003. J. Agric. Food Chem. 51, 587–593. Megoulas, N.C., Koupparis, M.A., 2005. Anal. Chim. Acta 547, 64–72. Mendonsa, S.D., Bowser, M.T, 2005. J. Am. Chem. Soc. 127, 9382–9383. Miyake, Y., Togashi, H., Tashiro, M., Yamaguchi, H., Oda, S., Kudo, M., Tanaka, Y., Kondo, Y., Sawa, R., Fujimoto, T., Machinami, T., Ono, A., 2006. J. Am. Chem. Soc. 128, 2172–2173. Morovján, G., Csokán, P.P., Konda, L.N., 1998. Chromatographia 48, 32–36. Oertel, R., Neumeister, V., Kirch, W., 2004. J. Chromatogr. A 1058, 197–201. Peng, J.H., Wang, Y.H., Wang, J.L., Zhou, X., Liu, Z.H., 2011. Biosens. Bioelectron. 28, 414–420.
156
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
Raddatz, M.S.L., Dolf, A., Endl, E., Knolle, P., Famulok, M., Mayer, G., 2008. Angew. Chem. Int. Ed. 47, 5190–5193. Rantanen, T., Päkkilä, H., Jämsen, L., Kuningas, K., Ukonaho, T., Lövgren, T., Soukka, T., 2007. Anal. Chem. 79, 6312–6318. Rantanen, T., Järvenpää, M.L., Vuojola, J., Kuningas, K., Soukka, T., 2008. Angew. Chem. Int. Ed. 47, 3811–3813. Rantanen, T., Järvenpää, M.L., Vuojola, J., Arppe, R., Kuningas, K., Soukka, T., 2009. Analyst 134, 1713–1716. Raz, S.R., Bremer, M.G.E.G., Haasnoot, W., Norde, W., 2009. Anal. Chem. 81, 7743–7749. Sapsford, K.E., Berti, L., Medintz, I.L., 2006. Angew. Chem. Int. Ed. 45, 4562–4588. Schuler, B., Eaton, W.A., 2008. Curr. Opin. Struct. Biol. 18, 16–26. Shoji, A., Kuwahara, M., Ozaki, H., Sawai, H., 2007. J. Am. Chem. Soc. 129, 1456–1464. Si, Y.C., Samulski, E.T., 2008. Nano Lett. 8, 1679–1682. Song, K.M., Cho, M., Jo, H., Min, K., Jeon, S.H., Kim, T., Han, M.S., Ku, J.K., Ban, C., 2011. Anal. Biochem. 415, 175–181. Tang, J.J., Yu, T., Guo, L., Xie, J.W., Shao, N.S., He, Z.K., 2007. Biosens. Bioelectron. 22, 2456–2463. Wang, L.F., Peng, J.D., 2009. Chromatographia 69, 519–522. Wang, Y.H., Bao, L., Liu, Z.H., Pang, D.W., 2011. Anal. Chem. 83, 8130–8137. Wang, Y.H., Shen, P., Li, C.Y., Wang, Y.Y., Liu, Z.H., 2012. Anal. Chem. 84, 1466–1473.
Wang, Y.H., Wu, Z.J., Liu, Z.H., 2013. Anal. Chem. 85, 258–264. Watanabe, H., Satake, A., Kido, Y., Tsuji, A., 1999. Analyst 124, 1611–1615. Wei, Q., Zhao, Y.F., Du, B., Wu, D., Li, H., Yang, M.H., 2012. Food Chem. 134, 1601–1606. Wu, S.J., Duan, N., Ma, X.Y., Xia, Y., Wang, H.X., Wang, Z.P., Zhang, Q., 2012. Anal. Chem. 84, 6263–6270. Xing, X.J., Liu, X.G., He, Y., Luo, Q.Y., Tang, H.W., Pang, D.W., 2012. Biosens. Bioelectron. 37, 61–67. Yuan, Y.X., Liu, Z.H., 2012. Chem. Commun. 48, 7510–7512. Zhan, L., Peng, L., Yu, Y., Zhen, S.J., Huang, C.Z., 2012. Analyst 137, 4968–4973. Zhang, C.L., Yuan, Y.X., Zhang, S.M., Wang, Y.H., Liu, Z.H., 2011. Angew. Chem. Int. Ed. 50, 6851–6854. Zhang, Q.B., Song, K., Zhao, J.W., Kong, X.G., Sun, Y.J., Liu, X.M., Zhang, Y.L., Zeng, Q.H., Zhang, H., 2009. J. Colloid Interface Sci. 336, 171–175. Zhao, Y.F., Wei, Q., Xu, C.X., Li, H., Wu, D., Cai, Y.Y., Mao, K.X., Cui, Z.T., Du, B., 2011. Sens. Actuators B 155, 618–625. Zhou, Y., Bao, Q.L., Tang, L.A.L., Zhong, Y.L., Loh, K.P., 2009. Chem. Mater. 21, 2950–2956. Zhu, Y., Chandra, P., Song, K.M., Ban, C., Shim, Y.B., 2012. Biosens. Bioelectron. 36, 29–34. Zhu, Y., Chandra, P., Ban, C., Shim, Y.B., 2012. Electroanalysis 24, 1057–1064.
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
An ultrasensitive homogeneous aptasensor for kanamycin based on upconversion fluorescence resonance energy transfer Hui Li, De-en Sun, Yajie Liu, Zhihong Liu n Key Laboratory of Analytical Chemistry for Biology and Medicine (Ministry of Education), College of Chemistry and Molecular Sciences, Wuhan University, Wuhan 430072, China
art ic l e i nf o
a b s t r a c t
Article history: Received 11 November 2013 Received in revised form 29 November 2013 Accepted 29 November 2013 Available online 10 December 2013
We developed an ultrasensitive fluorescence resonance energy transfer (FRET) aptasensor for kanamycin detection, using upconversion nanoparticles (UCNPs) as the energy donor and graphene as the energy acceptor. Oleic acid modified upconversion nanoparticles were synthesized through a hydrothermal process followed by a ligand exchange with hexanedioic acid. The kanamycin aptamer (50 -NH2AGATGGGGGTTGAGGCTAAGCCGA-30 ) was tagged to UCNPs through an EDC–NHS protocol. The π–π stacking interaction between the aptamer and graphene brought UCNPs and graphene in close proximity and hence initiated the FRET process resulting in quenching of UCNPs fluorescence. The addition of kanamycin to the UCNPs–aptamer–graphene complex caused the fluorescence recovery because of the blocking of the energy transfer, which was induced by the conformation change of aptamer into a hairpin structure. A linear calibration was obtained between the fluorescence intensity and the logarithm of kanamycin concentration in the range from 0.01 nM to 3 nM in aqueous buffer solution, with a detection limit of 9 pM. The aptasensor was also applicable in diluted human serum sample with a linear range from 0.03 nM to 3 nM and a detection limit of 18 pM. The aptasensor showed good specificity towards kanamycin without being disturbed by other antibiotics. The ultrahigh sensitivity and pronounced robustness in complicated sample matrix suggested promising prospect of the aptasensor in practical applications. & 2013 Elsevier B.V. All rights reserved.
Keywords: Kanamycin Aptasensor Upconversion fluorescence Graphene Fluorescence resonance energy transfer
1. Introduction Kanamycin, an important antibiotic that perturbs protein synthesis as a result of misreading the genetic code and inhibiting translation by binding to the 30S subunit of ribosomal, has been extensively used in human and veterinary medicine to treat infections caused by Gram-positive and Gram-negative bacteria (Fourmy et al., 1998). However, an overdosage of kanamycin is likely to cause kinds of rather serious side effects such as ototoxicity, nephrotoxicity and antibiotic resistance (Oertel et al., 2004). Therefore, carefully monitoring kanamycin concentration in body fluid is necessary and urgent to avoid drug abuse and to assure human health. Up to now, a variety of methods including high performance liquid chromatography (Morovján et al., 1998; Megoulas and Koupparis, 2005; Wang and Peng, 2009), enzyme-linked immunosorbent assay (Watanabe et al., 1999; Loomans et al., 2003; H.W. Chen et al., 2008), capillary electrophoresis (Kaale et al., 2001, 2003; Long et al., 2003), surface plasmon resonance (Raz et al., 2009; Frasconi et al., 2010) and electrochemical immunosensor (Zhao et al., 2011; Wei et al., 2012)
n
Corresponding author. Tel.: +86 27 87217886; fax: +86 27 68754067. E-mail address: [email protected] (Z. Liu).
0956-5663/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bios.2013.11.079
have been developed to detect kanamycin in animal-derived food and biological samples. Nevertheless, in consideration of the inherent imperfectness involved in these technologies like time-consuming, tedious operation or high consumption of reagents, new strategies for kanamycin assay are still desired. In the past several years, aptasensor has emerged as a simple yet efficient technique attracting increasing attention. It is referred to as a new category of biosensor constructed with aptamers as recognition elements. Various targets including metal ions (Miyake et al., 2006; Liu et al., 2007), small organic molecules (Ellington and Szostak, 1992; Shoji et al., 2007), proteins (Wang et al., 2011; Zhan et al., 2012) and even whole cells or microorganisms (Y.Q. Chen et al., 2008; Raddatz et al., 2008) have been detected using their specific aptasensors. Aptamers offer many advantages over antibodies such as low batch–batch variability, small size, low immunogenicity, easy chemical modification and good chemical stability (Keefe et al., 2010). In view of the noticeable advantages of aptamers, Song et al. (2011) screened a singlestranded DNA (ssDNA) aptamer showing high specificity and affinity for kanamycin and then constructed a simple colorimetric aptasensor. Afterwards electrochemical (Zhu et al., 2012; Daprà et al., 2013) and fluorometric (Leung et al., 2013) kanamycin aptasensors were also developed on the basis of this aptamer.
150
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
Despite the achievements made in the aptamer-based kanamycin assay, the sensitivity of these aptasensors, however, still has considerable space to improve considering that the blood drug concentration could be at very low levels. Besides, the practical use of such aptasensors, especially in complicated biological sample matrix (such as body fluid), has not yet been well established. Fluorescence resonance energy transfer (FRET) is recognized as a sensitive and reliable technique for homogeneous bioassays (Jares-Erijman and Jovin, 2003; Schuler and Eaton, 2008). FRET is a non-radiative energy transfer process and occurs between the energy donors and acceptors in close proximity (normally 1–10 nm) through long-range dipole–dipole interactions (Sapsford et al., 2006). Since the energy transfer efficiency (E) is inversely proportional to the sixth power of the distance between the energy and acceptor, a donor-to-acceptor distance change will sharply alter the efficiency resulting in ultrahigh sensitivity in bioassays. Another important merit is that the through-space interaction is mostly independent of intervening solvents and biomolecules, which makes it possible to specifically recognize target molecules in a complicated sample (Lakowicz, 2006). Nonetheless, conventional FRET technique using down-conversion fluorophores as energy donors usually suffer from background interferences which mainly include (1) the co-excitation of energy donor and acceptor due to the small Stokes shift of most fluorescent dyes; (2) the autofluorescence and scattering light from coexisting biomolecules (Frangioni, 2003). As such, the assay sensitivity may be reduced and the application of downconversion FRET technique in complicated matrixes is restricted. To circumvent these restrictions, upconversion nanoparticles (UCNPs) have been introduced to FRET as energy donors (Kuningas et al., 2005a). The near infrared (NIR) excitation and anti-Stokes emission natures of UCNPs endow this kind of fluorophore with strong ability to overcome the above-mentioned shortcomings, so that the UC-FRET technique has shown a promising prospect in homogeneous bioassay in biological samples (Kuningas et al., 2005b, 2006, 2007; Rantanen et al., 2007, 2008, 2009). Herein, we developed an ultrasensitive UC-FRET-based aptasensor for homogeneous determination of kanamycin in human serum, using upconversion nanoparticles as the energy donor and graphene as the energy acceptor. Graphene is an efficient dark quencher and has been proved as a good energy acceptor for UCNPs (Zhang et al., 2011; Wu et al., 2012). Another reason for choosing graphene as energy acceptor is that the aptamer can be facily assembled on its surface through π–π stacking interaction between the nucleobases and the two-dimensional (2D) graphitic carbon nanomaterial, such that the covalent label can be omitted (Xing et al., 2012). The aptasensor featuring flexible configuration and easy operation provided a detection limit of 9 pM in buffered solutions, which was lowered by orders of magnitude as compared to other reported kanamycin aptasensors. What is more, the sensor was applied in diluted serum sample with a detection limit of 18 pM, suggesting its high robustness and practical usage.
2. Materials and method 2.1. Materials Kanamycin was obtained from Amresco (USA). Amine modified kanamycin aptamer (50 -NH2-AGATGGGGGTTGAGGCTAAGCCGA-30 ) was supplied by Sangon Biotechnology Co., Ltd. (Shanghai, China). 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride and N-hydroxysuccinimide were purchased from Sigma-Aldrich. Graphene oxide was from Sinocarbon Graphene Marketing Center (Beijing, China). The other reagents were all from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). The solvents and
reagents were used as received without further purification. All aqueous solutions were prepared in ultrapure water (purified by Milli-Q biocel from Millipore China Ltd.). 2.2. Instrumentation The crystal phase of UCNPs was identified by a Bruker D8 Discover X-ray diffractometer (XRD) with 2θ range from 101 to 701 at a scanning rate of 41/min, with Cu Ka irradiation (k¼ 1.5460 Å). The size and morphology of hexanedioic acid (HDA)-modified NaYF4: Yb, Er upconversion nanoparticles was characterized by HITACHI H-7000FA transmission electron microscopy (100 kV accelerating voltage). FT-IR spectra of oleic acid-UCNPs, HDAUCNPs and graphene oxide after modification were measured on a Nicolet iS10 FT-IR Spectrometer (Thermo Scientific, USA) with the KBr pellet technique. UV–vis absorption measurements were conducted on a UV2550 UV–vis spectrophotometer (Shimadzu Scientific Instruments Inc.). The upconversion fluorescence spectra were recorded on a RF5301 fluorescence spectrophotometer (Shimadzu Scientific Instruments Inc.) using a 980 nm diode continuous-wave (CW) laser (Beijing Hi-Tech Optoelectronic Co., Ltd.) as the external excitation source, with the power being set at 850 mW. 2.3. Synthesis of oleic acid-modified NaYF4: Yb, Er upconversion nanoparticles Oleic acid-coated NaYF4: Yb, Er was synthesized according to a reported procedure (Cui et al., 2011). Briefly, 0.6 mmol of lanthanide oxides Ln2O3 (Y:Yb:Er ¼0.78:0.2:0.02 in mol) were dissolved in hot nitric acid (65 1C) to acquire Ln(NO3)3, and the solvent was evaporated after a 6 h reaction. NaOH (1.2 g, 30 mmol), H2O (5 mL), ethanol (10 mL), and oleic acid (20 mL) were mixed under agitation to form a homogeneous solution. Subsequently, 0.6 mmol (total amounts) of rare-earth nitrate (1.2 mL, 0.5 mol/L Ln(NO3)3) aqueous solution was added under magnetic stirring. Then 1.0 M aqueous NaF (4 mL) solution was added dropwise to the above solution. After agitating for another 10 min, the mixture was transferred to a 50 mL autoclave, sealed, and hydrothermally treated at 160 1C for 12 h. Afterwards the autoclave was cooled to room-temperature naturally, and the products were collected at the bottom of the container by cyclohexane, washed with cyclohexane/ethanol (1:6 v/v) for three times and centrifuged to obtain powder products. 2.4. Ligand-exchange of oleic acid-modified NaYF4: Yb, Er nanoparticles with HDA Ligand-exchange of UCNPs was performed according to the literature report with some modifications (Zhang et al., 2009). The diethylene glycol (DEG, 30 mL) containing HDA (1.5 g) was heated up to 110 1C for 30 min with vigorous stirring under nitrogen. A chloroform solution of oleic acid-modified UCNPs (40 mg, 2 mL) was injected into the above solution, then heated to 240 1C and kept at this temperature for about 5 h until the solution became clear. After the solution was cooled to room temperature, excess ultrapure water was added. The HDA-UCNPs were isolated by centrifugation and decantation. Finally the product was washed three times with pure water and redispersed in 4 mL of deionized water for further use. 2.5. Preparation of hydrothermal converted graphene from graphene oxide The conversion was accomplished through a hydrothermal route (Zhou et al., 2009). The pH of 40 mL of 0.5 mg/mL graphene
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
151
oxide aqueous solution was adjusted to 8.0 with 1 M NaOH. Then it was transferred to a Teflon-lined autoclave and heated at 180 1C for 6 h. Afterwards the autoclave was cooled to room temperature naturally. The hydrothermally treated graphene was found to precipitate at the bottom of the autoclave as a black powder, possibly due to the low solubility in super-fluidic water under hydrothermal conditions. The as-obtained graphene could be readily dispersed by ultrasonification in water.
same experimental procedures. For the determination of kanamycin in human serum, freshly collected serum sample from a healthy volunteer (provided by School of Medicine, Wuhan University) was 100-fold diluted with Tris–HCl buffer without further processing, and the same assay procedure as in the aqueous solution was followed. Standard addition method was adopted to determine the concentration of kanamycin in real serum samples.
2.6. Preparation of sulfonated graphene
3. Results and discussion
The sulfonic acid groups were introduced by a simple sulfonation method (Si and Samulski, 2008). For short, 40 mg of sulfanilic acid and 15.8 mg of sodium nitrite were first dissolved in 10 mL of NaOH solution (0.25%), and then added to 10 mL of 0.1 M HCl on ice bath under stirring. After reaction for 10 min, the aryl diazonium salt solution was added to 40 mL of 0.5 mg/mL graphene dispersion and was kept stirring for 2 h on ice bath. After filtration and rinsing, the purified sulfonated graphene was diluted to 0.5 mg/mL and stored at 4 1C.
3.1. Principle of the UC-FRET based aptasensor for kanamycin
2.7. Attachment of kanamycin aptamer to upconversion nanoparticles The amine modified kanamycin aptamer was covalently conjugated to HDA-UCPs following the standard EDC–NHS conjugation protocol (Hermanson, 2008). Firstly, 2 mg of HDA-UCPs was dissolved in 4 mL of MES buffer (10 mM, pH 5.5), and then 0.32 mg (0.4 mM) of EDC HCl and 0.46 mg (1 mM) of NHS were added to the solution to activate the carboxyl groups on HDA. The mixture was incubated at room temperature with gentle shaking for 45 min. After centrifugation, the activated HDA-UCNPs were washed with pure water for three times and then dispersed in 4 mL of HEPES buffer (10 mM, pH 7.2) containing 2 nmol amino modified kanamycin aptamer. The reaction lasted overnight at room temperature with slow shaking. In order to block the excess NHS, 10 mg of Tris was added into the reaction mixture. Subsequently, the kanamycin aptamer labeled HDA-UCNPs were collected by centrifugation and washed with ultrapure water for three times and finally diluted in 4 mL of Tris–HCl buffer (10 mM, pH 8). The 0.5 mg/mL UCNPs–aptamer complex was stored at 4 1C for further use.
The UC-FRET aptasensor was constructed based on the conformation change of the kanamycin aptamer before and after interacting with kanamycin. Generally, aptamers are singlestranded DNA or RNA nucleic acids which were screened in vitro through a process named as Systematic Evolution of Ligands by Exponential Enrichment (SELEX) (Ellington and Szostak, 1990). Up to now, a variety of aptamers have been selected towards a wide range of targets, including metal ions (Miyake et al., 2006; Liu et al., 2007), proteins (Wang et al., 2011; Zhan et al., 2012), peptides (Mendonsa and Bowser, 2005), toxins (Tang et al., 2007) and so on. The occurrence of binding between aptamers and their corresponding targets is usually accompanied by the conformation change of aptamers. The π–π stacking interaction usually happens between random-coiled aptamers and the graphitic 2D nanomaterial with sp2 carbon atoms and a large planar structure, which offers the advantage of simplifying experimental approaches. In the absence of kanamycin, the random-coiled aptamers are spontaneously adsorbed onto graphene surface through π–π stacking and thus bring the UCNPs (energy donor) and graphene (energy acceptor) in close proximity, leading to the quenching of UNCPs emission (Scheme 1). After the introduction of kanamycin into the UCNPs–graphene FRET system and the preferential binding between the aptamer and kanamycin, the aptamer is transformed to a hairpin structure which possesses much lower affinity toward graphene (Song et al., 2011; Zhu et al., 2012). Under this circumstance, the distance between UCNPs and graphene is enlarged blocking the energy transfer. As a consequence, the fluorescence emission of UCNPs is restored in a kanamycin concentration dependent manner.
2.8. Kanamycin detection in aqueous solution and diluted human serum In order to select an appropriate concentration of graphene to do the following fluorescence recovery experiments, a fixed amount of UCNPs–aptamer (0.012 mg/mL) was incubated with increasing amounts of sulfonated graphene in 10 mM Tris–HCl buffer (pH 8, 10 mM NaCl) at room temperature for 1.5 h, and then upconversion fluorescence measurements were carried out. In a typical FRET assay process, various concentrations of kanamycin were first mixed with the UCNPs–aptamer (0.012 mg/mL) conjugates in 10 mM Tris–HCl buffer (pH 8, 10 mM NaCl) and the mixture was incubated at 4 1C for 2 h. Afterwards sulfonated graphene was added individually into the above mixtures with an ultimate concentration of 0.016 mg/mL followed by incubation for another 1.5 h. Finally, the fluorescence intensity of the reaction mixture was recorded under the excitation of 980 nm. To examine the specificity of the FRET aptasensor, a list of other antibiotics including ampicillin, streptomycin sulfate, gentamicin, aureomycin, oxytetracycline and tetracycline were added into the UCNPs– aptamer–graphene system in place of kanamycin following the
Scheme 1. Schematic illustration of the kanamycin aptasensor on the basis of FRET from aptamer-linked UCNPs to graphene.
152
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
3.2. Characterization of the energy donor and energy acceptor To realize the above design, we first synthesized hexanedioic acid modified NaYF4: Yb, Er and sulfonated graphene. We used this kind of UCNPs because the rare earth ions co-doped NaYF4 nanocrystals are so far known to be the most efficient upconversion phosphors due to the low phonon energy of NaYF4 as host matrix. It is worth noting that in the lanthanide-doped UCNPs, it is the doping ion (Er3þ ) that acts as the emitter. Taking into consideration the distance dependence of FRET process, the size of particles can have remarkable effect on the energy transfer efficiency. It is understandable that a smaller particle size will be more favorable for the energy transfer since more emitting ions are located near the surface of the particles within the effective FRET distance. Compared to most UCNPs-based FRET assays previously reported by our lab and others (Wang et al., 2011; Kuningas et al., 2005a, 2005b, 2006, 2007; Rantanen et al., 2007, 2008, 2009; Zhang et al., 2011; Peng et al., 2011; Yuan and Liu, 2012; Wang et al., 2012, 2013), where the UNCPs were normally larger than 50 nm in diameter, we hereby prepared relatively smaller oleic acid modified-NaYF4: Yb, Er nanoparticles through a developed route: a hydrothermal synthesis of oleic acid-coated UCNPs followed by a ligand exchange process with hexanedioic acid. The TEM image (Fig. 1a) reveals spherical particles with the diameter ranging from 10 to 20 nm and good dispersibility in aqueous solution. The X-ray diffraction (XRD) pattern indicates that the synthesized NaYF4: Yb, Er nanoparticles were mainly with cubic phase (Fig. 1b). The existence of oleic acid and hexanedioic acid
molecules on the surface of the nanoparticles were confirmed by FT-IR analysis (Fig. 1c). The peak at ν¼3008 cm 1, attributable to the presence of ¼C–H stretching vibration, indicated the existence of oleic acid (the upper curve in Fig. 1c), while its disappearance in the lower curve verified the occurrence of ligand exchange process. What is more, the appearance of the stretching mode of the –COOH group at 1735 cm 1 indicated the presence of free –COOH groups on hexanedioic acid modified UCNPs. It is known that the conjugated network of graphene oxide can be largely destroyed in the preparation process from exfoliation of graphite through oxidation with strong acids. Therefore, in order to sufficiently make use of the π–π stacking interaction between nucleic acids and the carbon nanomaterial, it is necessary to recover the conjugated sp2 network by reduction of graphene oxide (Zhou et al., 2009). On the other hand, however, the reduced graphene oxide (RGO) always exhibits lowered water dispersibility, which is unfavorable for homogeneous sensing. To balance the two aspects, we further modified the RGO through sulfonation. The surface functionalization of RGO with sulfonic groups was confirmed by FT-IR spectra (Fig. 1d) with the peaks at 1170 cm 1, 1122 cm 1 and 1037 cm 1(two νS–O and one νS–phenyl ) in the upper curve. 3.3. Construction of the UC-FRET aptasensor In this system the aptamer chains are tagged to UCNPs surface, as shown in Scheme 1. To preclude possible steric hindrance
Fig. 1. (a) TEM image of HDA-modified NaYF4: Yb, Er upconversion nanoparticles. (b) XRD pattern of NaYF4: Yb, Er nanocrystals: Δ, cubic phase (JCPDS file no. 77-2042). (c) FT-IR spectra of oleic acid coated-UCNPs (the upper curve) and HDA modified UCNPs (the lower curve). (d) FT-IR spectra of RGO before (the lower curve) and after (the upper curve) modification with sulfonic groups.
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
153
Fig. 2. (a) UV–vis absorption spectra of UCNPs and UCNPs–kanamycin aptamer conjugate (0.5 mg/mL). (b) Fluorescence quenching of UCNPs–kanamycin aptamer (0.012 mg/mL) with various concentrations of sulfonated graphene (0, 0.003, 0.008, 0.013, 0.016 mg/mL). (c) Time dependence of the fluorescence quenching degree with 0.012 mg/mL UCNPs–kanamycin aptamer and 0.016 mg/mL sulfonated graphene. All experiments were performed in Tris–HCl buffer (10 mM, 10 mM NaCl, pH 8.0) under excitation at 980 nm.
caused by the relatively large nanoparticles during the recognition of aptamer by the target, hexanedioic acid was selected as the surface ligand of UCNPs because it offers a four-methylene flexible spacer. The exterior free carboxyl of hexanedioic acid was employed to covalently link with the aptamer, and the successful conjugation was confirmed by the 260 nm absorption on the UV–vis spectrum (Fig. 2a). To investigate the energy transfer between the donor–acceptor pair, different amounts of graphene were added into a fixed concentration of UCNPs–aptamer (0.012 mg/mL). After simply incubating the mixture for a short while, a graphene amount-related quenching of UCNPs fluorescence was observed (Fig. 2b). The driving force causing the donorto-acceptor assembly was assigned mainly to π–π stacking of the nucleobases on graphene. Since both the phosphate backbone of the DNA chain and the sulfonated graphene surface were negatively charged, electrostatic attraction was excluded. Such a design of electric charge was able to ensure the separation of the aptamer from graphene surface after interacting with the target. The time dependence of fluorescence quenching (Fig. 2c) revealed that the interaction reached to equilibrium in about 10 min, which was a quite quick event that could help to reduce the time consumption of the assay. A graphene concentration (0.016 mg/mL) corresponding to the quenching efficiency of 47% was selected for subsequent experiments. Although higher quenching efficiencies could be obtained by increasing the concentration of graphene, the use of
excessive quencher would contribute to nonspecific quenching which is unfavorable for subsequent fluorescence restoration. 3.4. Kanamycin determination in buffer solution Kanamycin detection was first conducted in Tris–HCl buffer with the above proposed UC-FRET aptasensor. The introduction of kanamycin into the UCNPs–aptamer–graphene FRET system led to fluorescence restoration of UCNPs (Fig. 3a). It relied on the conformation change of kanamycin aptamer into a hairpin structure after the binding between kanamycin and aptamer, which greatly weakened the π–π stacking interaction. In this condition the energy donor and acceptor were separated apart from each other, and thus the FRET process was inhibited. Because of the homogeneous nature of FRET, the assay was a very simple and straightforward procedure without any separation or repeated reagent addition. A linear relationship was obtained between the increasing fluorescence intensity of UCNPs and the logarithm of kanamycin concentration in the range from 0.01 to 3 nM (Fig. 3b). The detection limit was 9 pM, which was calculated as the concentration corresponding to 3 times of the SD (standard deviation) of the background signal from seven independent measurements. The sensitivity of this UC-FRET aptasensor was greatly enhanced compared to the reported kanamycin aptasensors (Song et al., 2011; Zhu et al., 2012; Daprà et al.,
154
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
Fig. 3. (a) Fluorescence recovery of the biosensor with the introduction of different amounts of kanamycin (0, 0.01, 0.03, 0.1, 0.3, 1, 3 nM) in Tris–HCl buffer (10 mM, 10 mM NaCl, pH 8.0). (b) The linear relationship between the fluorescence recovery (at 546 nm) and the logarithm of kanamycin concentration within the range of 0.01–3 nM in Tris–HCl buffer, data were presented as average7 SD from three independent measurements. (c) Fluorescence recovery of the biosensor with the introduction of different amounts of kanamycin (0, 0.03, 0.05, 0.08, 0.5, 1, 3 nM) in 100-fold diluted human serum. (d) The linear relationship between the fluorescence recovery (at 546 nm) and the logarithm of kanamycin concentration within the range of 0.03–3 nM in the diluted serum, data were presented as average7 SD from three independent measurements. All experiments were performed in the presence of 0.012 mg/mL UCNPs–kanamycin aptamer and 0.016 mg/mL sulfonated grapheme. Excitation wavelength: 980 nm.
2013; Leung et al., 2013). Such a performance improvement of the proposed sensor, in our opinion, also benefited from the adoption of the four-methylene spacer of hexanedioic acid. Compared to the situation where the probes (e.g., aptamers) are directly linked to the nanoparticles, the spacer acts as an isolation region eliminating (at least partly) the steric effect the particles may have on the probing process. Note that the length of the spacer can have contradictory influences on the sensing, i.e., a longer spacer provides more complete elimination of the steric hindrance yet the energy transfer efficiency may be reduced as well. In this sense, we believe that the selection of surface ligand can be another necessary job in the future for such UCNPs-based FRET sensors. 3.5. Specificity of the UC-FRET aptasensor towards kanamycin In order to examine the specificity of this UC-FRET aptasensor towards kanamycin, some analogous antibiotics including ampicillin, streptomycin sulfate, gentamicin, aureomycin, oxytetracycline and tetracycline were introduced individually into the aptasensor instead of kanamycin in the aqueous solution. Even when these substances were with a concentration (1 nM) 10 times higher than that of kanamycin (0.1 nM), they did not cause obvious restoration of the
UNCPs fluorescence intensity (Fig. 4). The pronounced selectivity of assay was attributed to the highly specific recognition between the aptamer and the target. At the same time, this experimental result also confirmed that the conformation alteration of the aptamer played a key role in the sensing. 3.6. Detecting kanamycin in serum sample Determination of kanamycin concentration in body fluids is essential for clinical diagnosis and survey. We chose human serum as a representative sample to examine the practical applicability of the UC-FRET aptasensor. Serum is a known complicated biological matrix containing various biomolecules which tend to raise autofluorescence and scattering light in optical sensing. Although, as described above, the FRET efficiency is basically independent of the coexisting biomolecules, those background optical signals can cause serious interference to down-conversion fluorescence-based assays. The freshly collected clean (which means no kanamycin contained) serum sample was 100-fold diluted with 10 mM Tris–HCl buffer to maintain the pH, which was used as the assay medium to check if a linear calibration could still be set up. With an identical assay procedure as that in aqueous solution, a similar fluorescence restoration was observed (Fig. 3c) and a linear relationship between the fluorescence intensity
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
155
Fig. 4. (a) Normalized fluorescence spectra of the sensor in the presence of different antibiotics. The highest curve is the spectrum in the presence of 0.1 nM kanamycin, and the other seven curves (which are overlapped) are the spectra of the sensor and that in the presence of 1 nM other antibiotics (ampicillin, streptomycin sulfate, gentamicin, aureomycin, oxytetracycline and tetracycline). (b) Relative fluorescence intensity ((Fother antibiotics F0)/(Fkanamycin F0)) of the UC-FRET aptasensor in the presence of different substances, where F0 is the fluorescence intensity in the absence of kanamycin or other antibiotics. Data were presented as average7SD from three independent measurements. The concentration of the other antibiotics was 1 nM, and the concentration of kanamycin was 0.1 nM. Experiments were conducted in Tris–HCl buffer (10 mM, 10 mM NaCl, pH 8.0) under excitation at 980 nm.
Table 1 Determination of kanamycin in spiked serum samples. Sample no.
Added (nM)a
Found (nM)
Recovery (%)
RSD (%) n¼ 3
1 2 3
0.8 1 3
0.75 0.93 3.08
94 93 103
1.5 0.6 1.7
a
Mean value of three determinations by the biosensor.
and the logarithm of kanamycin concentration was also obtained in the range from 0.03 nM to 3 nM, with the detection limit being calculated as 18 pM (Fig. 3d). It is also seen that the detection in serum sample exhibited good reproducibility. Though the linear range was slightly narrowed down and the detection limit was higher than that in the aqueous solution, the aptasensor still showed enough robustness in the serum sample which could be attributed to three merits of the sensor: (1) the homogeneous feature of FRET technique, (2) the circumvention of background interference by UCNPs and (3) the firm and stable aptamer-target recognition. Then the standard addition method was applied to check the validity of the method. The recovery of kanamycin in three spiked serum samples and the RSD levels (n¼ 3) are presented in Table 1. The above results indicated that this UC-FRET aptasensor for kanamycin could find practical applications hereafter.
4. Conclusions In summary, we have constructed an ultrasensitve aptasensor for kanamycin detection based on the fluorescence resonance energy transfer from upconversion nanoparticles to graphene. Combining the advantages of aptamers and the UC-FRET technique, the aptasensor exhibited excellent sensing performances in both aqueous solutions and body fluid sample. The homogeneous sensor also had the merits of simple configuration, straightforward operation and time and cost efficiency, suggesting promising prospect in practical use.
Acknowledgment This work was supported by the National Natural Science Foundation of China (Nos. 21075094, 21375098), the National
Basic Research of China (973 program, No. 2011CB933600) and the Program for New Century Excellent Talents in University (NCET-11-0402). References Chen, H.W., Medley, C.D., Sefah, K., Shang, G.D., Tang, Z.W., Meng, L., Smith, J.E., Tan, W.H., 2008. ChemMedChem 3, 991–1001. Chen, Y.Q., Wang, Z.Q., Wang, Z.H., Tang, S.S., Zhu, Y., Xiao, X.L., 2008. J. Agric. Food Chem. 56, 2944–2952. Cui, S.S., Chen, H.Y., Gu, Y.Q., 2011. J. Phys. Conf. Ser. 277, 012006. Daprà, J., Lauridsen, L.H., Nielsen, A.T., Rozlosnik, N., 2013. Biosens. Bioelectron. 43, 315–320. Ellington, A.D., Szostak, J.W., 1990. Nature 346, 818–822. Ellington, A.D., Szostak, J.W., 1992. Nature 355, 850–852. Fourmy, D., Yoshizawa, S., Puglisi, J.D., 1998. J. Mol. Biol. 277, 333–345. Frangioni, J.V., 2003. Curr. Opin. Chem. Biol. 7, 626–634. Frasconi, M., Vered, R.T., Riskin, M., Willner, I., 2010. Anal. Chem. 82, 2512–2519. Hermanson, G.T., 2008. Bioconjugate Techniques, second ed. Academic, San Diego. Jares-Erijman, E.A., Jovin, T.M., 2003. Nat. Biotechnol. 21, 1387–1395. Kaale, E., Schepdael, A.V., Roets, E., Hoogmartens, J., 2001. J. Chromatogr. A 924, 451–458. Kaale, E., Schepdael, A.V., Roets, E., Hoogmartens, J., 2003. Electrophoresis 24, 1119–1125. Keefe, A.D., Pai, S., Ellington, A., 2010. Nat. Rev. Drug Discov. 9, 537–550. Kuningas, K., Rantanen, T., Karhunen, U., Lövgren, T., Soukka, T., 2005a. Anal. Chem. 77, 2826–2834. Kuningas, K., Rantanen, T., Ukonaho, T., Lövgren, T., Soukka, T., 2005b. Anal. Chem. 77, 7348–7355. Kuningas, K., Ukonaho, T., Päkkilä, H., Rantanen, T., Rosenberg, J., Lövgren, T., Soukka, T., 2006. Anal. Chem. 78, 4690–4696. Kuningas, K., Päkkilä, H., Ukonaho, T., Rantanen, T., Lövgren, T., Soukka, T., 2007. Clin. Chem. 53, 145–146. Lakowicz, J.R., 2006. Principles of Fluorescence Spectroscopy, third ed. SpringerVerlag, Berlin, Heidelberg. Leung, K.H., He, H.Z., Chan, D.S.H., Fu, W.C., Leung, C.H., Ma, D.L., 2013. Sens. Actuators B: Chem. 177, 487–492. Liu, J., Brown, A.K., Meng, X., Cropek, D.M., Istok, J.D., Waston, D.B., Lu, Y., 2007. Proc. Natl. Acad. Sci. USA 104, 2056–2061. Long, Y.H., Hernandez, M., Kaale, E., Schepdael, A.V., Roets, E., Borrull, F., Calull, M., Hoogmartens, J., 2003. J. Chromatogr. B 784, 255–264. Loomans, E.E.M.G., Wiltenburg, J.V., Koets, M., Amerongen, A.V., 2003. J. Agric. Food Chem. 51, 587–593. Megoulas, N.C., Koupparis, M.A., 2005. Anal. Chim. Acta 547, 64–72. Mendonsa, S.D., Bowser, M.T, 2005. J. Am. Chem. Soc. 127, 9382–9383. Miyake, Y., Togashi, H., Tashiro, M., Yamaguchi, H., Oda, S., Kudo, M., Tanaka, Y., Kondo, Y., Sawa, R., Fujimoto, T., Machinami, T., Ono, A., 2006. J. Am. Chem. Soc. 128, 2172–2173. Morovján, G., Csokán, P.P., Konda, L.N., 1998. Chromatographia 48, 32–36. Oertel, R., Neumeister, V., Kirch, W., 2004. J. Chromatogr. A 1058, 197–201. Peng, J.H., Wang, Y.H., Wang, J.L., Zhou, X., Liu, Z.H., 2011. Biosens. Bioelectron. 28, 414–420.
156
H. Li et al. / Biosensors and Bioelectronics 55 (2014) 149–156
Raddatz, M.S.L., Dolf, A., Endl, E., Knolle, P., Famulok, M., Mayer, G., 2008. Angew. Chem. Int. Ed. 47, 5190–5193. Rantanen, T., Päkkilä, H., Jämsen, L., Kuningas, K., Ukonaho, T., Lövgren, T., Soukka, T., 2007. Anal. Chem. 79, 6312–6318. Rantanen, T., Järvenpää, M.L., Vuojola, J., Kuningas, K., Soukka, T., 2008. Angew. Chem. Int. Ed. 47, 3811–3813. Rantanen, T., Järvenpää, M.L., Vuojola, J., Arppe, R., Kuningas, K., Soukka, T., 2009. Analyst 134, 1713–1716. Raz, S.R., Bremer, M.G.E.G., Haasnoot, W., Norde, W., 2009. Anal. Chem. 81, 7743–7749. Sapsford, K.E., Berti, L., Medintz, I.L., 2006. Angew. Chem. Int. Ed. 45, 4562–4588. Schuler, B., Eaton, W.A., 2008. Curr. Opin. Struct. Biol. 18, 16–26. Shoji, A., Kuwahara, M., Ozaki, H., Sawai, H., 2007. J. Am. Chem. Soc. 129, 1456–1464. Si, Y.C., Samulski, E.T., 2008. Nano Lett. 8, 1679–1682. Song, K.M., Cho, M., Jo, H., Min, K., Jeon, S.H., Kim, T., Han, M.S., Ku, J.K., Ban, C., 2011. Anal. Biochem. 415, 175–181. Tang, J.J., Yu, T., Guo, L., Xie, J.W., Shao, N.S., He, Z.K., 2007. Biosens. Bioelectron. 22, 2456–2463. Wang, L.F., Peng, J.D., 2009. Chromatographia 69, 519–522. Wang, Y.H., Bao, L., Liu, Z.H., Pang, D.W., 2011. Anal. Chem. 83, 8130–8137. Wang, Y.H., Shen, P., Li, C.Y., Wang, Y.Y., Liu, Z.H., 2012. Anal. Chem. 84, 1466–1473.
Wang, Y.H., Wu, Z.J., Liu, Z.H., 2013. Anal. Chem. 85, 258–264. Watanabe, H., Satake, A., Kido, Y., Tsuji, A., 1999. Analyst 124, 1611–1615. Wei, Q., Zhao, Y.F., Du, B., Wu, D., Li, H., Yang, M.H., 2012. Food Chem. 134, 1601–1606. Wu, S.J., Duan, N., Ma, X.Y., Xia, Y., Wang, H.X., Wang, Z.P., Zhang, Q., 2012. Anal. Chem. 84, 6263–6270. Xing, X.J., Liu, X.G., He, Y., Luo, Q.Y., Tang, H.W., Pang, D.W., 2012. Biosens. Bioelectron. 37, 61–67. Yuan, Y.X., Liu, Z.H., 2012. Chem. Commun. 48, 7510–7512. Zhan, L., Peng, L., Yu, Y., Zhen, S.J., Huang, C.Z., 2012. Analyst 137, 4968–4973. Zhang, C.L., Yuan, Y.X., Zhang, S.M., Wang, Y.H., Liu, Z.H., 2011. Angew. Chem. Int. Ed. 50, 6851–6854. Zhang, Q.B., Song, K., Zhao, J.W., Kong, X.G., Sun, Y.J., Liu, X.M., Zhang, Y.L., Zeng, Q.H., Zhang, H., 2009. J. Colloid Interface Sci. 336, 171–175. Zhao, Y.F., Wei, Q., Xu, C.X., Li, H., Wu, D., Cai, Y.Y., Mao, K.X., Cui, Z.T., Du, B., 2011. Sens. Actuators B 155, 618–625. Zhou, Y., Bao, Q.L., Tang, L.A.L., Zhong, Y.L., Loh, K.P., 2009. Chem. Mater. 21, 2950–2956. Zhu, Y., Chandra, P., Song, K.M., Ban, C., Shim, Y.B., 2012. Biosens. Bioelectron. 36, 29–34. Zhu, Y., Chandra, P., Ban, C., Shim, Y.B., 2012. Electroanalysis 24, 1057–1064.
