Challenges of Electrochemical Impedance Spectroscopy in Protein

Apr 13, 2009 - Di Kang , Sheng Sun , Martin Kurnik , Demosthenes Morales , Frederick W. Dahlquist , and Kevin W. Plaxco. Journal of the American Chemi...
11 downloads 0 Views 2MB Size
Download PDF
Anal. Chem. 2009, 81, 3944–3949

Challenges of Electrochemical Impedance Spectroscopy in Protein Biosensing A. Bogomolova,* E. Komarova, K. Reber, T. Gerasimov, O. Yavuz, S. Bhatt, and M. Aldissi Fractal Systems Inc., 108 Fourth Street, Belleair Beach, Florida 33786 Electrochemical impedance spectroscopy (EIS) measurement, performed in the presence of a redox agent, is a convenient method to measure molecular interactions of electrochemically inactive compounds taking place on the electrode surface. High sensitivity of the method, being highly advantageous, can be also associated with nonspecific impedance changes that could be easily mistaken for specific interactions. Therefore, it is necessary to be aware of all possible causes and perform parallel control experiments to rule them out. We present the results obtained during the early stages of aptamer-based sensor development, utilizing a model system of human alpha thrombin interacting with a thiolated DNA aptamer, immobilized on gold electrodes. EIS measurements took place in the presence of iron ferrocyanides. In addition to known method limitations, that is, inability to discriminate between specific and nonspecific binding (both causing impedance increase), we have found other factors leading to nonspecific impedance changes, such as: (i) initial electrode contamination; (ii) repetitive measurements; (iii) additional cyclic voltammetry (CV) or differential pulse voltammetry (DPV) measurements; and (iv) additional incubations in the buffer between measurements, which have never been discussed before. We suggest ways to overcome the method limitations. Electrochemical Impedance spectroscopy (EIS) becomes increasingly popular for electrochemical measurements of molecular interactions and is widely used in a variety of biosensing applications. The theory, principles, and applications of EIS to biosensing have been summarized in a recent review.1 Traditional applications of EIS include immunosensing protocols, utilizing electrodeimmobilized antibodies to register protein-protein interactions,2-4 or to detect smaller molecules, such as antibiotics5 in complex biological media. There is a growing number of publications on the use of EIS for genosensing utilizing electrodes with immobilized ss-DNA probes which detect complementary ss-DNA * To whom correspondence should be addressed. E-mail: abogomol@ hotmail.com. (1) Lisdat, F.; Scha¨fer, D. Anal. Bioanal. Chem. 2008, 391, 1555–1567. (2) Barton A. C.; Davis F.; Higson S. P. Anal. Chem. 2008, 80, 9411-9416. (3) Tsekenis, G.; Garifallou, G. Z.; Davis, F.; Millner, P. A.; Gibson, T. D.; Higson, S. P. Anal. Chem. 2008, 80, 2058–2062. (4) Kassanos, P.; Iles, R. K.; Bayford, R. H.; Demosthenous, A. Physiol. Meas. 2008, 29, S241-254. (5) Tsekenis, G.; Garifallou, G. Z.; Davis, F.; Millner, P. A.; Pinacho, D. G.; Sanchez-Baeza, F.; Marco, M. P.; Gibson, T. D.; Higson, S. P. Anal. Chem. 2008, 80, 9233-9239.

3944

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

target molecules through hybridization.6-11 The uses of EIS are not limited to protein-protein or DNA-DNA interactions. Electrodes with immobilized dsDNA were used to detect specific binding of anticancer drug cisplatin.12 The method is very sensitive and can be used for a “label-free” detection of a wide range of molecular recognition events happening on the electrode surface. Development of the “artificial tongue” utilizing nanostuctured conducting polymer films and EIS protocol has been reported.13 Using EIS, it was possible to detect such small molecules as L-histidine and fenvalerate using molecularly imprinted polymers as sensing layers on the electrodes,14,15 or to detect copper and lead ions on a cysteamine-EDTA-modified electrode.16 Although typically the biomolecular recognition protocol is carried out in solution, it was possible to detect the enzyme collagenase by EIS in aerosol form, through digestion of electrodebound gelatin in air.17 The EIS method also has a good potential for cell biology/microbiology applications requiring surface characterization with attached cells or bacteria in vitro or in vivo. Thus, cell proliferation of Saccharomyces cerevisiae on gold electrodes was studied beyond the monolayer coverage,18 and formation of bacterial biofilms of Staphylococcus epidermidis was monitored by EIS on medical implants.19,20 (6) Weng, J.; Zhang, J.; Li, H.; Sun, L.; Lin, C.; Zhang, Q. Anal. Chem. 2008, 80, 7075–7083. (7) Davis, F.; Nabok, A. V.; Higson, S. P. Biosens. Bioelectron. 2005, 20, 1531– 1538. (8) Davis, F.; Hughes, M. A.; Cossins, A. R.; Higson, S. P. Anal. Chem. 2007, 79, 1153–1157. (9) Peng, H.; Soeller, C.; Vigar, N.; Kilmartin, P. A.; Cannell, M. B.; Bowmaker, G. A.; Cooney, R. P.; Travas-Sejdic, J. Biosens. Bioelectron. 2005, 20, 1821– 1828. (10) Peng, H.; Soeller, C.; Vigar, N. A.; Caprio, V.; Travas-Sejdic, J. Biosens. Bioelectron. 2007, 22, 1868–1873. (11) Lucarelli, F.; Tombelli, S.; Minunni, M.; Marrazza, G.; Mascini, M. Anal. Chim. Acta 2008, 609, 139–159. (12) Yan, F.; Sadik, O. A. J. Am. Chem. Soc. 2001, 123, 11335–11340. (13) Riul, A. R.; Malmegrim, R. R.; Fonseca, F. J.; Mattoso, L. H. Artif. Organs. 2003, 27, 469–472. (14) Liao, H.; Zhang, Z.; Nie, L.; Yao, S. J. Biochem. Biophys. Methods 2004, 59, 75–87. (15) Gong, J. L.; Gong, F. C.; Kuang, Y.; Zeng, G. M.; Shen, G. L.; Yu, R. Q. Anal. Bioanal. Chem. 2004, 379, 302–307. (16) Shervedani, R. K.; Farahbakhsh, A.; Bagherzadeh, M. Anal. Chim. Acta 2007, 587, 254–262. (17) Saum, A. G.; Cumming, R. H.; Rowell, F. J. Biosens. Bioelectron. 2000, 15, 305–313. (18) Heiskanen, A. R.; Spe´gel, C. F.; Kostesha, N.; Ruzgas, T.; Emne´us, J. Langmuir. 2008, 24, 9066–9073. (19) Dunlop, P.; Oliver, L.; Byrne, T.; McAdams, E. Stud. Health Technol. Inf. 2005, 117, 213–217. (20) Oliver, L. M.; Dunlop, P. S.; Byrne, J. A.; Blair, I. S.; Boyle, M.; McGuigan, K. G.; McAdams, E. T. Conf. Proc. IEEE Eng. Med. Biol. Soc. 2006, 1, 535–538. 10.1021/ac9002358 CCC: $40.75  2009 American Chemical Society Published on Web 04/13/2009

The method provides unique advantages compared to other electrochemical methods, such as high sensitivity, signal quantification ease, and ability to separate the surface binding events from the solution impedance. Impedance data are recorded in the range of frequencies, using alternating current (AC) of small amplitude (thus the EIS is often referred to as AC Impedance). Compared to other electrochemical methods, such as cyclic voltammetry (CV) or differential pulse voltammetry (DPV), which are also used to characterize molecular interactions on the surface of electrodes, AC Impedance is less destructive to the measured biological interactions because it is performed in a very narrow range of small potentials. Electrochemically inert species can be conveniently measured by AC Impedance when the measurement is performed in the presence of a redox agent, such as iron ferrocyanide(s), which undergoes oxidation and reduction at the surface of the electrode at a certain potential applied during the measurement. The surface availability for redox reaction decreases (while impedance increases) with the increased binding of the analyte. However, high sensitivity of the method results in potential drawbacks, which have to be taken into account to obtain reliable and reproducible data. Overlooking the system peculiarities can result in false-positive or false-negative results. Below is a brief report of potential drawbacks of the system which we discovered during the early steps of aptamer-based sensor development, using a well-known model system of human alpha-Thrombin, interacting with a specific aptamer, selected in ref 21. There have been several Thrombin EIS-based sensors reported recently,22-26 which use the same materials and experimental design as we did in this work. In our opinion, the drawbacks of the system we report are quite universal, and we feel the need to caution other researchers, who enter the field of electrochemical sensing, inspired by the seeming ease of AC Impedance applications. EXPERIMENTAL SECTION Materials. BSA, Streptavidin, HEPES, KCl, NaCl, NaOH, ethanol, K3Fe(CN)6, K4Fe(CN)6, mercaptoethanol were purchased from Sigma-Aldrich. PBS buffer pH 7.4 and DNase-free, RNase-free deionized sterile water were purchased from Invitrogen. Human alpha Thrombin was purchased from Haematologic Technologies (Essex Junction, VT) and used within 2 weeks. Electrodes and Electrochemical Cells. (a) Gold disk electrodes (1.6 mm diameter Au disk, embedded in CTFE plastic body, 7.5 cm length x 6 mm OD) and Ag/AgCl reference electrodes were purchased from Bioanalytical Systems (BAS, West Lafayette, IN); BAS voltammetry cell was used with the gold disk working electrode, Pt wire counter electrode, and Ag/AgCl reference electrode in a standard three-electrode setup. Pt wire (1 mm thick) (21) Baldrich, E.; Restrepo, A.; O’Sullivan, C. K. Anal. Chem. 2004, 76, 7053– 7063. (22) Radi, A. E.; Acero Sanchez, J. L.; Baldrich, E.; O’Sullivan, C. K. Anal. Chem. 2005, 77, 6320–6323. (23) Du, Y.; Li, B.; Wei, H.; Wang, Y.; Wang, E. Anal. Chem. 2008, 80, 5110– 5117. (24) Cho, M.; Kim, Y.; Han, S. Y.; Min, K.; Rahman, M. A.; Shim, Y. B.; Ban, C. BMB Reports 2008, 41, 126–131. (25) Li, X.; Shen, L.; Zhang, D.; Qi, H.; Gao, Q.; Ma, F.; Zhang, C. Biosens. Bioelectron. 2008, 23, 1624–1630. (26) Deng, C.; Chen, J.; Nie, Z.; Wang, M.; Chu, X.; Chen, X.; Xiao, X.; Lei, C.; Yao, S. Anal. Chem. 2009, 81, 739–745.

was purchased from Alfa Aesar (Ward Hill, MA). Prior to use, the gold disk working electrode was cleaned by polishing with diamond paste using a BAS polishing kit, rinsed in ethanol and deionized water, which was followed by electrochemical cleaning in 0.5 M NaOH or 0.5 M H2SO4, and rinsed in deionized water. (b) To prepare gold foil electrodes, Au foil (0.5 mm thickness, Alfa Aesar) was cut to 1′′ × 1′′ pieces. Gold-sputtered slides (SigmaAldrich) were cut to the same size. Gold foil or gold-sputtered slide electrodes were mounted at the bottom of a custom-made electrochemical cell with a 3 mm diameter opening for the working electrode,27 with Pt wire as a counter electrode and Ag/AgCl reference electrode. Prior to use, the gold foil electrodes were cleaned in piranha solution (70% concentrated H2SO4, 30% H2O2, see safety considerations) for 5-10 min and rinsed thoroughly with deionized water, while the gold-sputtered slides were rinsed in ethanol and deionized water, followed by electrochemical cycling in 50 mM H2SO4 for 20-40 cycles in the range of -0.9 to 0.9V, 100 mV/s., and rinsed in deionized water. Oligonucleotide Probes and Immobilization. Two probes containing the well-studied human Thrombin-binding aptamer (5′-GGTTGGTGTGGTTGG-3′) 21-26,28 were used. Thiolated probes were custom synthesized at Operon Technologies (Huntsville, AL). probe T: 5′ GGTTGGTGTGGTTGG--(CH2)3-S-S-(CH2)3-OH 3′ probe Ts: 5′ HO-(CH2)6-S-S-(CH2)6--TCTCTCTCTCTCTCTCTCTCTCTCTCTCTCGGTTGGTGTGGTTGG -3′ The 15-mer probe T was modified with thiol on the 3′ end, while the 45-mer probe Ts contained a spacer and thiol modification on the 5′ end. The probes were immobilized on gold electrodes by self-assembly from 1 µM solution in 1 M NaCl for 2 h at room temperature or for 16 h at 4 °C, either in dithiol (as supplied) or in the deprotected form (following treatment with DTT and column purifications as instructed by Operon). For probe immobilization, precleaned dry gold electrodes were placed in a humidified chamber, and the probe solution was added dropwise on top of the electrode. Following immobilization, the electrodes were rinsed several times with 1 M NaCl and incubated with 100 mM β-mercaptoethanol in 1 M NaCl for 15 min. After rinsing several times with 1 M NaCl and with the binding buffer (10 mM Hepes-NaOH, pH 8, 50 mM KCl) the electrodes with immobilized probes were immediately used for further binding experiments or for AC Impedance measurements. Binding Experiments. After probe immobilization and orientation with ME treatment, the electrodes were incubated with one of the following: 0.1 mM Thrombin (3.56 µg/mL) as a specific binder, or 35.6 µg/mL streptavidin (as a nonspecific binder, 10× by weight to the specific one), or 1% BSA (blocking), all in 10 mM HEPES-NaOH, pH 8, 50 mM KCL, for 0.5-2 h at room temperature. The electrodes were washed thoroughly with the binding buffer and used immediately for the AC Impedance measurements. More than one binding step could be carried out on the same electrode. Electrochemical Equipment and Measurements. All electrochemical measurements were performed using CHI-760B Potentiostat (CH Instruments). The measurements were per(27) Komarova, E.; Aldissi, M.; Bogomolova, A. Biosens. Bioelectron. 2005, 21, 182–189. (28) Ikebukuro, K.; Okumura, Y.; Sumikura, K.; Karube, I. Nucleic Acids Res. 2005, 33, e108.

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

3945

formed in the measurement buffer, containing 10 mM HEPES pH 8, 0.5-20 mM K4Fe(CN)6, 0.5-20 mM K3Fe(CN)6, a total of 50 mM K+ adjusted with KCl. Cyclic voltammetry was recorded in the range [-0.2 V to +0.6 V] prior to probe immobilization to determine the midpoint between oxidation and reduction which was chosen as an applied potential for further AC Impedance measurements (the electrodes were recleaned after CV prior to probe immobilization). AC impedance spectra for each electrode were recorded after probe immobilization and/or after binding/blocking steps in the same measurement buffer, with DC bias potential of ∼0.192 V in a frequency range of 0.1 Hz to 100 kHz and AC amplitude of 0.005 V. Prior to AC Impedance measurements, the electrodes were rinsed in the measurement buffer, and after the measurement, if used for further binding experiments, they were rinsed several times with the binding buffer (without ferrocyanides). AC Impedance spectra were presented as a Nyquist plot (Z′, real impedance, versus Z′′, imaginary impedance) using CH Instruments software. Although the obtained spectra were fitted using Randles equivalent circuit CH Instruments software, the figures below contain the experimental spectra, not the fitted ones. Safety Considerations. Piranha solution (70% concentrated H2SO4, 30% H2O2) is extremely dangerous and should be handled with extreme caution. Gloves and eye protection required for handling. The solution is prepared in glass containers under the hood by adding peroxide to acid; it should be kept in the open container and never stored. Mixing piranha with organic compounds may cause an explosion. RESULTS AND DISCUSSION The desired scenario of typical label-free biosensing experiments is to measure the binding of specific analyte (DNA or protein) on the working electrode, previously modified with a specific probe, such as an oligonucleotide (for DNA detection), an oligonucleotide aptamer, or a peptide (for protein detection). AC Impedance spectra are recorded in the presence of the redox agent, undergoing electrochemical reaction at the surface of the working electrode, using a three-electrode system. The spectra are recorded before and after the binding of the analyte. The spectra are typically presented in the form of a Nyquist plot (where Z′ is the real and Z′′ is the imaginary part of impedance, see ref 1 for detailed method explanation) and overlaid to pinpoint the differences. An ideal Nyquist plot reveals the semicircle (with the diameter, corresponding to the charge transfer resistance, Rct, produced by redox reactions at the interface with the electrode) followed by a diagonal straight line (corresponding to the impedance of the current due to diffusion from the solution to the interface). The presence of the electrode-bound probe and/ or analyte will result in the appearance of the larger semicircle (visually separate from the small semicircle, corresponding to the electrode). The increase in the semicircle diameter (hence Rct value) characterizes the increased surface coverage upon either specific or nonspecific analyte binding, resulting in less availability for a redox reaction. We have found additional causes of Rct increase besides actual binding on a protein or DNA. They include: (i) initial electrode contamination; (ii) repetitive measurements; (iii) additional CV or DPV measurements; and (iv) additional incubations in the buffer between 3946

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

measurements. Our experimental data and troubleshooting suggestions are presented below. Initial Electrode Characterization. It is very important to start with a perfectly clean electrode, which can be confirmed by AC Impedance spectra. A freshly cleaned gold electrode will reveal a classical Nyquist plot, with the semicircle corresponding to Rct of pure Au electrode and the diagonal line corresponding to diffusion. Different electrodes with the same surface area will produce identical spectra. Appearance of additional semicircles corresponds to the remaining impurities on the surface of the electrode and will compromise further binding studies. CV of the ferrocyanides on the clean electrode reveals clear oxidation and reduction peaks, with positions differing slightly depending on the reference electrode used. For highest sensitivity, it is important to select an applied potential for EIS from the corresponding CV of the clean electrode as the average between oxidation and reduction potentials. After any electrochemical measurement it is necessary to reclean the electrode prior to any further experiments. Freshly cleaned electrodes reveal the absence or low protein binding (Supporting Information, Figure S-1A); however, if an oligonucleotide probe, immobilized protein, or any impurities (including those accumulated from prolonged incubation of a clean electrode in the buffer) are initially present on the electrode, nonspecific protein binding will result in Rct semicircle increase and might obscure the specific binding (Supporting Information, Figure S-1B). For best reliability, one can verify the cleanliness of the electrode by AC Impedance, and then reclean it immediately before further use, for example, for probe immobilization. Performing Repetitive Measurements on the Same Electrode. One expects to register protein binding through the increase in Rct semicircle. For example, Figure 1A shows Rct increase which can be attributed to binding of Thrombin to a specific aptamer probe, previously immobilized on the electrode. However, it is important to know that Rct of the modified electrode will keep increasing if the measurements are repeated on the same electrode (Figure 1B). Such “repeated measurement-related” Rct increase is more pronounced with higher ferrocyanide concentrations in the measuring buffer. We have found that using 1-2 mM of each or of either K3Fe(CN)6 and K4Fe(CN)6 (instead of typically used 5-50 mM) is sufficient to obtain quantifiable Impedance spectra and allows us to minimize the “repeated measurementrelated” Rct increase. Another reason for measurement-related Rct increase is the rearrangement of the charged probe or probe-analyte complex due to the applied positive potential during AC Impedance measurement (∼+0.19V). Since the typical AC Impedance measurement takes from 5 to 15 min, such effect can be considerable. Thus, even shorter application times of the positive potential result in significant Impedance increase of electrodes modified with oligonucleotide probes (Supporting Information, Figure S-2). Shortening the measurement time by using the narrow range of AC Impedance frequency (just enough to obtain the semicircle to be quantified), along with limiting the

Figure 1. Impedance increase can be caused by either protein binding or repeated measurement. (A, B), Impedance measurements of the same electrode (Au foil). (s) electrode with immobilized deprotected probe Ts; (b) same electrode, after incubation with 1 µM Thrombin in 10 mM HEPES, 50 mM KCl, pH 8, for 30 min, RT; (O, 4, g) three consecutive repeated Impedance measurements performed every 20 min, respectively, after Thrombin incubation; AC Impedance recorded at 0.192 V, [0.1 Hz-100 kHz], AC amplitude 0.005 V, in 10 mM Hepes-NaOH, pH 8, 5 mM K4Fe(CN)6, 5 mM K3Fe(CN)6, 40 mM KCl.

Figure 2. CV measurement is more damaging to the biological complex than AC Impedance. (A) Overlay of 3 consecutive CV measurements at different steps of an electrode modification (Au foil). (s) Electrode 1 (Au foil), modified with deprotected probe T (in 1 M NaCl, 2 h., RT), followed by 100 mM Mercaptoethanol, in 1 M NaCl, 15 min; (0) Electrode 1, after 1% BSA in 10 mM HEPES, 50 mM KCl, pH 8 for 1 h, RT; (b) Electrode 1, after incubation with 100 nM Thrombin in 10 mM HEPES, 50 mM KCl, pH 8, 1 h, RT. CV performed for 2 cycles, from -0.2 V to + 0.6 V, 0.1 V/s. (B) Overlay of 3 sets of consecutive AC Impedance measurements of the same electrode. Set 1: (s) Electrode 1 (Au foil), modified with deprotected probe T (in 1 M NaCl, 2 h., RT), followed by 100 mM mercaptoethanol, in 1 M NaCl, 15 min., (AC Impedance measurement 1 before CV); (4) 2nd measurement, before CV. Set 2: (0) electrode 1, after 1% BSA in 10 mM HEPES, 50 mM KCl, pH 8 for 1 h, RT; (9) 2nd measurement; (×) 3rd measurement (after CV). Set 3: (O) Electrode 1, after incubation with 100 nM Thrombin in 10 mM HEPES, 50 mM KCl, pH 8, 1 h, RT; (b) 2nd measurement. (*) 3rd measurement (after CV); AC Impedance recorded at 0.192 V, from 1 Hz to 100 kHz, AC amplitude 0.005 V, in 10 mM Hepes-NaOH, pH 8, 2 mM K4Fe(CN)6, 2 mM K3Fe(CN)6, 46 mM KCl.

repeated number of measurements, performed on the same electrode, will be helpful to minimize the measurement-related Rct changes. Performing Additional CV or DPV Measurements. CV is routinely used along with AC Impedance to characterize protein binding. The decreasing area of the cyclic voltammogram, such as the one shown in Figure 2A, is used as evidence of protein binding on the electrode. A typical overlay presents three steps (1) probe immobilization; (2) blocking with an irrelevant protein; (3) specific protein binding. However, the CV measurement by itself causes the increase of the electrode Rct, probably causing rearrangement of the surface-bound compounds due to the range of the applied potentials. The CV-related Rct increase is higher compared to the increase caused by additional AC Impedance measurements (Figure 2B).

The three sets of spectra overlaid in Figure 2B are as follows: set 1 was recorded after probe immobilization, set 2 after blocking with BSA, and set 3 after incubation with Thrombin. The AC Impedance spectra were recorded twice after probe immobilization (producing similar spectra in set 1, Figure 2B), followed by CV (recorded in Figure 2A). This was followed by incubation with BSA and two AC Impedance measurements (producing similar spectra), followed by CV and the third AC Impedance measurement, which revealed a CV-related Rct increase in set 2, Figure 2B. The next step was the incubation with Thrombin, and two AC Impedance measurements, followed by CV and the third AC Impedance measurement, which also revealed noticeable Rct increase after CV in set 3, Figure 2B. In this experiment, the nonspecific Rct increase, related to additional AC Impedance measurement, is minimal, while the one caused by CV is Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

3947

Figure 3. Impedance increase on the blank electrode can be caused just by incubations in the buffer. (A) Consecutive measurements of electrode A (Au foil). Set 1: (s) blank electrode, after incubation in 10 mM HEPES, 50 mM KCl, pH 8 for 20 h, 4 °C; (]) same as (s), 2nd measurement. Set 2: (+) after incubation in 10 mM Hepes-NaOH, pH 8, 2 mM K4Fe(CN)6, 2 mM K3Fe(CN)6, 46 mM KCl for 1 h, RT; (2) same as (+), 2nd measurement; (*) same as (+), 3rd measurement; Set 3: (4) after incubation in 10 mM Hepes-NaOH, pH 8, 2 mM K4Fe(CN)6, 2 mM K3Fe(CN)6, 46 mM KCl for second hour, RT; (f) same as (4), 2nd measurement; (g) same as (4), 3rd measurement. (B) Electrode B (Au foil), parallel incubations performed in ferrocyanide-free buffer (10 mM HEPES, 50 mM KCl, pH8). AC Impedance recorded at 0.192 V, from 0.1 Hz to 100 kHz, AC amplitude 0.005 V, in 10 mM Hepes-NaOH, pH 8, 2 mM K4Fe(CN)6, 2 mM K3Fe(CN)6, 46 mM KCl.

noticeably larger. DPV causes similar changes as CV and should be used with caution or avoided. Performing Additional Incubations in the Buffer between Measurements. Even more worrisome is that nonspecific increase of Rct values can be obtained in the absence of protein or DNA binding by just incubating a gold electrode in the measurement buffer (Figure 3). Incubation-related Rct increase is more pronounced than the measurement-related increase, discussed above. Here, the two clean gold foil electrodes (A) and (B) were incubated in the buffer (10 mM HEPES, 50 mM KCl, pH 8) for 20 h, and Impedance spectra were recorded twice for each electrode after such incubation. It is important to notice that overnight incubation resulted in the appearance of a second small semicircle in addition to the one corresponding to freshly cleaned gold (set 1). Then the electrode (A) was incubated in the buffer containing ferrocyanides (10 mM Hepes-NaOH, pH 8, 2 mM K4Fe(CN)6, 2 mM K3Fe(CN)6, 46 mM KCl) for 1 h., while electrode (B) was incubated in a ferrocyanide-free buffer (10 mM Hepes-NaOH, pH 8, 50 mM KCl) for 1 h., and Impedance spectra were recorded 3 times for each electrode (set 2). The 1 h incubations were repeated for each electrode in the same buffers, and three Impedance measurements for each electrode followed (set 3). From this experiment one can see clearly that incubations in ferrocyanide-containing buffers should be avoided, since they result in a significant nonspecific Rct increase, like in Figure 3A, because of interaction of ferrocyanides with gold. One should minimize multiple incubations and multiple measurements on the same electrode, keeping in mind that even insignificant nonspecific Rct increase (like in Figure 3B) can compromise the real data, resulting from specific protein binding. For example, in a typical biosensing experiment one can think of “blocking” the surface of the electrode with a nonspecific protein, and then using such an electrode to bind the specific analyte. However, this would produce false-positive “specific” binding (Figure 4), unless the incubation-related Rct increase is ruled out with the help of the parallel control experiments. Here, three gold foil electrodes underwent similar modification 3948

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

Figure 4. Protein binding after blocking: nonspecific Rct increase can be caused by incubation in a blocking agent or in plain buffer (overlay of three parallel experiments). (s) Electrode 1 (Au foil), modified with probe T (in 1 M NaCl, 2 h., RT), followed by 100 mM Mercaptoethanol, in 1 M NaCl, 15 min., after blocking in 1% BSA in 10 mM HEPES, 50 mM KCl, pH8 for 1 h, RT; (4) electrode 2, same modification; (+) electrode 3, same modification; (O) electrode 1, after incubation with 100 nM Thrombin in 10 mM HEPES, 50 mM KCl, pH 8, 1 h, RT; (2) electrode 2, after incubation with 1% BSA in 10 mM HEPES, 50 mM KCl, pH 8, 1 h, RT; (*) electrode 3, after incubation in 10 mM HEPES, 50 mM KCl, pH8, 1 h, RT; AC Impedance recorded at 0.192 V, from 0.1 Hz to 100 kHz, AC amplitude 0.005 V, in 10 mM Hepes-NaOH, pH 8, 2 mM K4Fe(CN)6, 2 mM K3Fe(CN)6, 46 mM KCl.

with a Thrombin-specific aptamer probe, followed by “blocking” in 1% BSA, after which the first Impedance measurement was recorded once for each electrode. Initial Rct values were very close for the three electrodes. During the second binding step, one electrode was incubated with Thrombin, the second was again incubated in BSA, the third one in the ferrocyanide-free buffer, and after that Impedance was remeasured. All three electrodes revealed very similar Rct increases regardless whether they were incubated in a specific protein, nonspecific protein, or plain buffer (Figure 4). In our opinion, the only way to avoid such false-positive results, produced by incubation-related Rct increase, is to run parallel control experiments using plain buffer and nonspecific analytes, as well

as use electrodes modified with nonspecific probes. Even with these precautions, the method will remain prone to falsenegative results, that is, inability to register a specific signal due to high experimental noise. The need to perform parallel control experiments with nonspecific probes/analytes, along with the need to use multiple experiments to obtain reliable data for specific binding lead us to the idea of designing an array electrode, suitable for AC Impedance measurements. We have reported this work briefly in ref 29 and are in the process of preparation of a detailed publication. By performing multiple parallel measurements, along with careful selection of probe immobilization conditions, it is possible to use AC Impedance technique to obtain qualitative data on protein binding (Supporting Information, Figures S-5, S-6). A detailed manuscript on array electrodes is in preparation. CONCLUSIONS High sensitivity of EIS allows registration of subtle changes produced by interaction of the biomolecules in close proximity to the electrode, such as DNA-DNA, DNA-protein, and protein-protein. On the other hand, possessing high sensitivity, this method can register irrelevant changes as well, such as (29) Gulledge J. S., Cannons A., Komarova E., Reber K., Aldissi M., Bogomolova A. Proceedings of the 108th Gen. ASM Meeting, Boston, MA, June 1-5, 2008.

changes in position of DNA and/or protein molecules relative to the electrode. Such nonspecific changes can be caused by the measurement itself due to the attraction/repulsion of the charged biological molecules to/from the surface of the electrode. The method is also very sensitive to any kind of electrode contamination. However, AC Impedance is less damaging to the biological interactions if compared to routinely used CV or DPV methods. Knowing the possible causes, which can lead to nonspecific Impedance changes, it is possible to overcome those causes or minimize their effects. The need for multiple control experiments, run in parallel, to account for these effects, cannot be underestimated. ACKNOWLEDGMENT This work was supported by US Army Contract No W911NF05-C-0019. SUPPORTING INFORMATION AVAILABLE Further details are given in Figures S-1 to S-6. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review February 2, 2009. Accepted March 31, 2009. AC9002358

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

3949