ARTICLE pubs.acs.org/ac
Highly Sensitive Detection of Exocytotic Dopamine Release Using a Gold-Nanoparticle-Network Microelectrode Kelly L. Adams, Bikash Kumar Jena, Stephen J. Percival, and Bo Zhang* Department of Chemistry, University of Washington, Seattle, Washington 98195-1700, United States
bS Supporting Information ABSTRACT: Here we report a new type of microelectrode sensor for single-cell exocytotic dopamine release. The new microsensor is built by forming a gold-nanoparticle (AuNP) network on a carbon fiber microelectrode. First a gold surface is obtained on a carbon fiber microdisk electrode by partially etching away the carbon followed by electrochemical deposition of gold into the pore. The gold surface is chemically functionalized with a sol-gel silicate network derived from (3-mercaptopropyl)trimethoxysilane (MPTS). A AuNP network is formed by immobilizing Au nanoparticles onto the thiol groups in the sol-gel silicate network. The AuNP-network microelectrode has been characterized by scanning electron microscopy (SEM) and steady-state voltammetry. The AuNP-network microelectrode has been used for amperometric detection of exocytotic dopamine secretion from individual pheochromocytoma (PC12) cells. The results show significant differences in the kinetic peak parameters including shorter rise time, decay time, and half-width as compared to a bare carbon fiber electrode equivalent. These results indicate AuNP-network microelectrodes possess an excellent sensing activity for single-cell exocytotic catecholamine release, specifically dopamine. Moreover, key advantageous properties inherent to bare carbon fiber microelectrodes (i.e., rigidity, flexibility, and small size) are maintained in addition to an observed prolonged shelf life stability and resistance to cellular debris fouling and dopamine polymerization.
and photoresist19 on the electrode in hopes of enhancing electrode sensitivity and/or selectivity.20-22 Electrode response time must be carefully considered when selecting an appropriate electrode modification, for some modifications may compromise temporal resolution while improving sensitivity and/or selectivity. Metal nanoparticles have received considerable interest because of their unusual optical, electronic, and catalytic properties.23-28 In particular, gold nanoparticles have attracted great attention due to their inert behavior and weak adsorption properties.29 Furthermore, Au nanoparticles have been successfully used for bioanalytical applications because of low cytotoxicity, high affinity with molecules of thiol/amine containing groups, and offering suitable platforms for surface immobilization of a wide range of enzymes and biomolecules.30,31 Tailoring the Au nanoparticles onto electrode surfaces has been extensively utilized in electroanalytical and electrocatalytic applications.32-34 The unique activity of Au nanoparticles is superior to its bulk counterpart because of its high surface area and interface-dominated properties.35 Here we report a new microelectrode sensor for single-cell exocytotic dopamine release, which is based on a AuNP network formed on a single carbon fiber microelectrode. This new type of microelectrode has been used for the amperometric detection of dopamine secretion from single PC12 cells. The sensitivity of this electrode, assessed by the kinetic peak parameters including rise
E
lectrochemical detection using microelectrodes has emerged as an increasingly important technique for studies of single biological cells.1-3 Specifically in neuroscience, electrochemical methods have been used to gain further knowledge of dynamic processes at single nerve cells which would lead to better and more descriptive models of the neurotransmission process. A key dynamic event in neuronal communication is exocytosis, a process that has been widely investigated for several decades4,5 and can be summarized as intracellular vesicles fusing with the cell membrane to release their contents.6 Methods to observe and to quantify individual exocytotic events have traditionally revolved around electron microscopy and patch-clamp methods.7 However, electrochemistry using carbon fiber microelectrodes has been extremely useful in the detection of numerous easily oxidizable neurochemicals (e.g., dopamine, epinephrine, 5-hydroxytryptamine, and histamine) released from single cells8,9 on the millisecond time scale when used in the constant potential amperometric mode.9 Moreover, carbon fiber microelectrodes offer numerous desirable features, including biocompatibility with cells, small probe size, and highly resistant to mechanical strain.10 Despite these valuable advantages, a great deal of work has been carried out to develop new microelectrodes for measurements in vitro. To maximize electrode sensitivity, the electrode surface should be free of adsorbed molecules, such as (but not limited to) proteins and oxidized products.11,12 Many attempts have been reported to modify the carbon microelectrode by forming a modification layer, such as Nafion,12-15 4-sulfobenzenediazonium tetrafluoroborate,16 single-walled carbon nanotubes,17,18 r 2010 American Chemical Society
Received: October 1, 2010 Accepted: December 3, 2010 Published: December 22, 2010 920
dx.doi.org/10.1021/ac102599s | Anal. Chem. 2011, 83, 920–927
Analytical Chemistry
ARTICLE
time, decay time, and half-width, has been carefully compared to a bare carbon fiber electrode. The improved performance in terms of prolonged shelf life stability and resistance to cellular debris and dopamine oxide adsorption fouling are also discussed.
background current, the detection algorithm assigned a baseline for the peak to the level of the background current. Double peaks and those with jagged tops (i.e., more than one peak) were manually excluded. Scanning Electron Microscopy. SEM images were obtained using a field-emission microscope (FEI Sirion) at the Nanotech User Facility located at the University of Washington. The FEI Sirion SEM was equipped in-house with an energy dispersive spectrometer (EDAX) and was used for elemental analysis. All samples were sputter-coated with a thin layer of Au/Pd or carbon for SEM imaging. Samples were not coated prior to EDAX elemental analysis. Gold Nanoparticle Synthesis. Citrate-stabilized Au nanoparticles were prepared by adding 0.64 mL of 1.15% trisodium citrate and freshly prepared 0.08% NaBH4 (0.32 mL) in 1% trisodium citrate to 30 mL of water containing 1% HAuCl4 (0.32 mL) and stirring the solution for 10 min at room temperature.37 Preparation of MPTS Sol. The (3-mercaptopropyl)trimethoxysilane (MPTS) sol was prepared by dissolving MPTS, methanol, and water (as 0.1 M HCl) in a molar ratio of 1:3:3 and stirring the mixture vigorously for 30 min.33 Cell Culture. PC12 cells were purchased from the American Type Culture Collection (Manassas, VA) and maintained as previously described.38 Briefly, PC12 cells were grown on mouse collagen IV-coated culture dishes (BD Biosciences; Bedford, MA) in phenol red-free RPMI 1640 (Mediatech, Manassas, VA) supplemented with 10% equine serum (HyClone, Logan, UT), 5% fetal bovine serum (HyClone, Logan, UT), and 100 units/ mL penicillin/streptomycin (HyClone, Logan, UT). Cells were kept in a 7% CO2 atmosphere at 37 °C and subcultured every 7 to 9 days or when cell confluency was reached. Cells were used for all experiments between 4 and 7 days after subculturing. Fabrication of Carbon Fiber Microelectrodes. Carbon fiber microelectrodes were fabricated following previously reported procedures with slight modifications.38 Briefly, an isolated 5-μmdiameter fiber was aspirated into a borosilicate glass capillary (1.2 mm O.D., 0.69 mm I.D., Sutter Instrument Co., Novato, CA). Capillaries were subsequently pulled with a commercial micropipet puller (Model P-97, Sutter Instrument Co., Novato, CA) and sealed with epoxy (Epoxy Technology, Billerica, MA) and beveled (Model BV-10, Sutter Instrument Co., Novato, CA) to 45°. Fabrication of AuNP-Network Microelectrodes. Figure 1 illustrates the process for the fabrication of AuNP-network microelectrodes. In the first step, a carbon fiber electrode with a 5-μm diameter was fabricated as described above and beveled to 45°. Next, the carbon fiber was electrochemically etched in isotonic saline to produce a recessed carbon fiber microelectrode for subsequent gold deposition. The recess depths obtained for this study ranged from 1 to 30 μm. The electrochemical etching of the carbon fiber was carried out by applying a 1-Hz, 4-V triangular waveform versus a silver quasireference electrode using a function generator (33220A, Agilent Technologies). This etching process was repeated up to three times until a visible etch was observed with the aid of an optical microscope (Figure SI1). Electrodes that were not visibly etched after three 1-min cycles were discarded. Gold was then electrochemically deposited inside this recess in a one-compartment, two-electrode cell at -1.5 V versus Pt for 15 min. The recessed electrode was sonicated in the potassium aurocyanide solution for ∼1 min prior to electrodeposition. Electrodes were visually inspected to determine if the recess was filled with Au (Figure SI1). Additional 15-min electrodeposition cycles were carried out if an electrode was only partially filled
’ EXPERIMENTAL SECTION Chemicals and Reagents. Potassium aurocyanide solution used for the electrodeposition of gold was purchased from Technic Inc. (Cranston, RI). Ferrocene (Fe(C5H5)2, Fluka), tetran-butylammonium hexafluorophosphate (TBAPF6, Aldrich), and (3-cyanopropyl)dimethylchlorosilane (Fluka) were of reagent grade quality or better and used without further purification. All other chemicals were of analytical grade and used as received from Sigma without further purification. Single cell experiments were performed in an isotonic saline solution prepared with 150 mM NaCl, 5 mM KCl, 1.2 mM MgCl2, 5 mM glucose, 10 mM HEPES, and 2 mM CaCl2, and the pH was adjusted to 7.4 with concentrated NaOH. All aqueous and organic solutions were made using 18 MΩ 3 cm water from a Barnstead NanoPure purification system (Thermo Scientific) and acetonitrile, respectively. Electrochemical Instruments. Cyclic voltammograms (CVs) were recorded at 50 mV/s versus a Ag/AgCl reference electrode (Bioanalytical Science, Inc.) using a computer-controlled Dagan Chem-Clamp voltammeter/amperometer (Dagan Corporation, Minneapolis, MN). All CVs were recorded using virtual instrumentation written in-house with LabView (National Instruments). Single-Cell Experimental Setup. Electrochemical recordings at single cells were made on an inverted optical microscope (IX71, Olympus) positioned on a vibration isolation table (Model 63-563, Technical Manufacturing Corporation, Peabody, MA) inside a home-built Faraday cage. The working electrode was gently positioned onto a single cell using a micromanipulator (Model MHW-3, Narishige, Inc., East Meadow, NY). A slight deformation in the outline of the cell confirmed the close proximity of the electrode to the cell surface. Exocytosis was stimulated at approximately 45-s intervals with a 5-s, 30-psi pulse (Eppendorf/ Brinkmann Instruments, Hauppauge, NY) from a glass micropipet containing isotonic saline with elevated Kþ (100 mM). Culture dishes were warmed using an automatic temperature controller (Model TC-344B, Warner Instrument Corp., Hamden, CT) fitted to the microscope stage, and all experiments were performed at 37 ( 1 °C. A constant potential (þ700 mV) was applied to the working electrode with respect to a single Ag/AgCl quasireference electrode placed in the cell bathing solution. Single-Cell Data Acquisition and Data Analysis. Amperometric data were recorded with an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA) and monitored in real-time via the AxoScope software package (Version 10.2, Molecular Devices). The output was digitized at 5 kHz and filtered at 2 kHz with an internal four-pole Bessel filter and analyzed without subsequent filtering. Amperometric peaks were identified and characterized with the MiniAnalysis software detection algorithm (Synaptosoft, Decatur, GA). Peaks were detected if both the amplitude of local maxima and the area under the curve exceeded a threshold of five times the root-mean-squared noise for a flat, 2-s recording acquired at the beginning of each experiment.36 Peaks were visually inspected to confirm that electrical noise was not included and to include peaks manually that were not detected due to their proximity in the current trace. Overlapping events were discarded if the baseline for each could not be determined. In cases where peaks appeared to rise above a broader 921
dx.doi.org/10.1021/ac102599s |Anal. Chem. 2011, 83, 920–927
Analytical Chemistry
ARTICLE
Figure 2. SEM images of a (A) sol-gel-modified and (B) AuNP-network electrode.
thoroughly with acetonitrile and allowed to dry in air. Then a thiol-terminated sol-gel network of MPTS was formed at the electrode surface via tip submersion for 20 min into MPTS sol.41 Electrodes were then rinsed with copious amounts of ethanol followed by DI water to remove any physically absorbed MPTS. Then electrode tips were immersed overnight into a solution of Au nanoparticles. Because of the strong affinity of Au on thiol,42 Au nanoparticles were immobilized into the sol-gel silicate network forming a AuNP-network microelectrode. Electron hopping between the Au nanoparticles and between the nanoparticles and the Au substrate is believed to be responsible for the electrochemical responses observed on the nanoparticles. Prior to use as amperometric sensors for dopamine detection at PC12 cells, all electrodes were rinsed with DI water and characterized in a 100-μM dopamine solution. Electrodes lacking a stable voltammetric response were discarded.
Figure 1. Schematic representation of the major fabrication steps for a AuNP-network microelectrode. A bare carbon fiber microdisk electrode (A) is first etched to form a recessed electrode (B). Gold is electrochemically deposited into the recess and polished to a smooth disk (C). A three-dimensional gel network is applied (D) for subsequent Au nanoparticle immobilization (E). (Schematic not drawn to scale and is drawn with a 90° beveled electrode for simplicity.).
’ RESULTS AND DISCUSSION
with gold. Electrodes that were not successfully filled with Au after three 15-min electrodeposition cycles were discarded. The electrodeposited gold typically overflowed the cavity and formed a gold ball beyond the confines of the recess. This excessive gold ball was beveled away, leaving a polished Au microdisk electrode suitable for subsequent modification with a sol-gel network. A three-dimensional gel network for gold nanoparticle immobilization (Figure 1D) was applied to the gold surface with slight modifications as previously described.33 The glass insulating layer surrounding the gold microdisk electrodes was first passivated in order to prevent unwanted MPTS and gold nanoparticle adsorption during later stages of electrode fabrication. Passivation was achieved by submerging the electrode tips into a 2% (by volume) solution of (3-cyanopropyl)dimethylchlorosilane in acetonitrile for a minimum of 18 h.8,39,40 Electrodes were rinsed
SEM Characterization of AuNP-Network Microelectrodes. SEM has been employed to characterize the AuNP-network microelectrode. Note that as a surface technique, SEM does not reveal the 3-D structure of the Au-NP network formed on the microelectrode surfaces. However, it does show some clear differences in the surface features of a sol-gel-modified Au microelectrode before and after modification with Au nanoparticles. A representative SEM image of a control, “sol-gel-only” electrode free of Au nanoparticles is shown in Figure 2A. Overall, the surface of the MPTS gel was smooth and contained only minor topographical features. Contrastingly, the representative SEM image of an AuNP-network microelectrode (Figure 2B) revealed an electrode surface with a more rough landscape. These visibly spherical features were 5 rms), which could yield a slightly larger number of dopamine molecules. As we have shown in Figure 6, the amperometric signals on the network electrodes are, in general, higher in magnitude and faster than the ones observed on bare carbon fiber electrodes. We believe that the higher/faster exocytotic events observed on the AuNP-network microelectrode are most likely due to the unique structure/surface properties of the nanoparticle-network electrodes. The fact that there is only a small difference (∼10%) in the total number of DA molecules indicates that the major differences in the amperometric signals are due to electrochemical kinetics. The 3-D structure of the nanoparticle-network electrodes and the unique properties of Au nanoparticles could be responsible for the enhanced kinetics. When a carbon fiber disk electrode is used, dopamine molecules diffuse in parallel to the cell surface in the small space between the cell and the electrode until they are fully oxidized. The total time of an exocytotic event shown on the amperometric trace is therefore dependent on several factors, including the secretion kinetics, the oxidation kinetics of dopamine on the electrode (electron-transfer kinetics), and the diffusion of dopamine after they are released from the intracellular vesicles. In the case of the AuNP-network electrode, the diffusion of dopamine can happen in parallel to the
Figure 7. Cyclic voltammograms for a representative (A) AuNP-network electrode and (B) bare carbon fiber electrode in 100 μM dopamine before (solid line) and after (dashed line) amperometric detection at single PC12 cells.
Table 1. Summary of the Shelf Storage Stability for Two Different Representative AuNP-Network Electrodes Undergoing Either Short Term (2 to 10 days) or Long Term (38 to 46 days) Storagea
a
% of i (day 0) ( SEM
short-term
long-term
105 ( 2
106 ( 3
The current for dopamine oxidation recorded at þ600 mV was normalized to the initial current recorded on the day immediately following Au nanoparticle immobilization and then subsequently expressed as a percentage.
cell surface in the space between the cell and the electrode as well as perpendicular to the cell into the network of the Au nanoparticles. Therefore, the molecular transport of dopamine could be enhanced after the secretion of dopamine from the cell, which could lead to an increased consumption of the dopamine and a shorter exocytotic event on the amperometric trace. However, more experiments are obviously needed in the future to further investigate this observation. Electrode Stability and Long-Term Storage. Other attractive qualities of AuNP-network electrodes are durable electrode sensitivity after electrochemically sampling from several PC12 cells (Figure 7A) as well as their long-term shelf stability (Table 1). AuNP-network electrode sensitivity was mostly conserved following exposure to multiple biological cell surfaces, as shown in Figure 7A, unlike the traditional carbon fiber electrode counterpart (Figure 7B). It has recently been shown that the carbon electrode surface can be renewed by oxidative etching via repeated potential sweeps.48 The AuNP-network electrode, 925
dx.doi.org/10.1021/ac102599s |Anal. Chem. 2011, 83, 920–927
Analytical Chemistry however, does not require such treatment, as it maintains its sensitivity. Additionally, we believe the added gel layer on our AuNP-network microelectrode serves as a protective barrier from cellular debris adhesion and direct molecular adsorption to the electrode surface, factors believed to contribute to the lowgrade fouling generally observed when using bare carbon fiber electrode probes. Moreover, electrode shelf life in open air was assessed by comparing the measured current in a 100 μM dopamine solution immediately following AuNP immobilization to the measured current obtained on subsequent days. Table 1 summarizes the average percentage of original current measured following shortterm (2 to 10 days) and long-term (38 to 46 days) storage after AuNP immobilization. As the percentage values indicate for both storage durations, the measured current remained fairly constant without any necessary electrode pretreatment. We believe the small (
