Article pubs.acs.org/ac
Effects of DNA Probe and Target Flexibility on the Performance of a “Signal-on” Electrochemical DNA Sensor Yao Wu and Rebecca Y. Lai* 651 Hamilton Hall, University of NebraskaLincoln, Lincoln, Nebraska 68588-0304, United States S Supporting Information *
ABSTRACT: We report the effect of the length and identity of a nontarget binding spacer in both the probe and target sequences on the overall performance of a folding-based electrochemical DNA sensor. Six near-identical DNA probes were used in this study; the main differences between these probes are the length (6, 10, or 14 bases) and identity (thymine (T) or adenine (A)) of the spacer connecting the two target binding domains. Despite the differences, the signaling mechanism of these sensors remains essentially the same. The methylene blue (MB)-modified probe assumes a linear unstructured conformation in the absence of the target; upon hybridization to the target, the probe adopts a “close” conformation, resulting in an increase in the MB current. Among the six sensors, the T14 and A14 sensors showed the largest signal increase upon target hybridization, highlighting the significance of probe flexibility on sensor performance. In addition to the target without a midsequence spacer, 12 other targets, each with a different oligo-T or oligo-A spacer, were used to elucidate the effect of target flexibility on the sensors’ signaling capacity. For all six sensors, hybridization to targets with a 2- or 3-base spacer resulted in the largest signal increase. Higher signal enhancement was also observed with targets with an oligo-A spacer. For this sensor design, addition of a long nontarget binding spacer to the probe sequence is advantageous, as it provides flexibility for optimal target capture. The length of the spacer in the target sequence, however, should be adequately long to enable efficient hybridization yet does not introduce undesirable electrostatic and crowding effects.
N
conditions. While many researchers, including our group, have successfully converted “signal-off” sensors into “signal-on” sensors via the use of frequency- or time-gating, there are merits in designing “signal-on” E-DNA sensors.13−15 To date, a myriad of “signal-on” E-DNA sensors have been developed.16−21 One of the simplest yet highly effective designs is the sensor reported by Immoos et al.; the signaling mechanism of this sensor is essentially the reverse of the original “‘signal-off”’ E-DNA sensor.22 Target binding induces the linear DNA probe to adopt a “close” conformation, which alters the electron transfer kinetics, resulting in an increase in the redox current. However, for this sensor to function as designed, flexibility of the probe is essential. In the previous study, the two target capture domains of the DNA probe were linked via a poly(ethylene glycol) (PEG) spacer.22 Recently, we have developed a more controlled method to link the two target capture domains; our sensor features a flexible spacer
ucleic acid based diagnostics has evolved significantly over the past decades. A wide range of novel technologies are currently available for ultrasensitive detection of DNA and RNA.1−3 Optical and electrochemical sensing platforms are the most sought after, given that many are capable of real time detection of the target sequences.3−6 Among the currently available electrochemical sensors, a unique class of DNA sensors, the electrochemical DNA (E-DNA) sensor, have gained substantial popularity since it was reported by Fan et al. in 2003.7 The signaling of this sensor relies on the target binding-induced conformational and dynamics change of the ferrocene-modified stem-loop DNA probe, resulting in a detectable reduction in the redox current. The merits of this sensor design include high sensitivity and selectivity, in addition to being reagentless and reusable.8,9 These sensors are also compatible with different probe immobilization methods and sensor substrates.10−12 Despite the various advantages, this sensor is inherently a “signal-off” sensor, which is not ideal for most biosensing applications. Disadvantages include limitations in the signaling capacity, in which only a maximum of 100% signal suppression can be attained under any experimental © XXXX American Chemical Society
Received: July 22, 2014 Accepted: August 10, 2014
A
dx.doi.org/10.1021/ac5027226 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Table 1. DNA Probes T6: T10: T14: A6: A10: A14:
5′5′5′5′5′5′-
HS-C6-TTAGCTCCAA HS-C6-TTAGCTCCAA HS-C6-TTAGCTCCAA HS-C6-TTAGCTCCAA HS-C6-TTAGCTCCAA HS-C6-TTAGCTCCAA
TTTTTT TACGCCACC T-MB-3′ TTTTTTTTTT TACGCCACC T-MB-3′ TTTTTTTTTTTTTT TACGCCACC T-MB-3′ AAAAAA TACGCCACC T-MB-3′ AAAAAAAAAA TACGCCACC T-MB-3′ AAAAAAAAAAAAAA TACGCCACC T-MB-3′
Table 2. DNA Targets 17-base target without a spacer: M0: 5′-TTGGAGCTGGTGGCGTA-3′ 18-base target with 1A insertion: MA1: 5′-TTGGAGCTAGGTGGCGTA-3′ 19-base target with 2A insertion: MA2: 5′-TTGGAGCTAAGGTGGCGTA-3′ 20-base target with 3A insertion: MA3: 5′-TTGGAGCTAAAGGTGGCGTA-3′ 21-base target with 4A insertion: MA4: 5′-TTGGAGCTAAAAGGTGGCGTA-3′ 22-base target with 5A insertion: MA5: 5′-TTGGAGCTAAAAAGGTGGCGTA-3′ 23-base target with 6A insertion: MA6: 5′-TTGGAGCTAAAAAAGGTGGCGTA-3′ 1-base mismatch target: MA3-1M: 5′-TTGGACCTAAAGGTGGCGTA-3′ 2-base mismatch target: MA3-2M: 5′-TTGGAGCTAAAGCTGCCGTA-3′
18-base MT1: 19-base MT2: 20-base MT3: 21-base MT4: 22-base MT5: 23-base MT6:
target with 1T insertion: 5′-TTGGAGCTTGGTGGCGTA-3′ target with 2T insertion: 5′-TTGGAGCTTTGGTGGCGTA-3′ target with 3T insertion: 5′-TTGGAGCTTTTGGTGGCGTA-3′ target with 4T insertion: 5′-TTGGAGCTTTTTGGTGGCGTA-3′ target with 5T insertion: 5′-TTGGAGCTTTTTTGGTGGCGTA-3′ target with 6T insertion: 5′-TTGGAGCTTTTTTTGGTGGCGTA-3′
■
consisting of 10 consecutive thymine (T) bases.23 Similar to most previously developed “signal-off” and “signal-on” E-DNA sensors, this sensor is reagentless and reusable. It is also very sensitive, achieving a detection limit of 200 fM. It is capable of single−base mismatch discrimination and is selective enough for real time detection of the target sequence in realistically complex biological samples such as saliva. The design is versatile and can be tailored for detection of essentially any nucleic acid sequence. It is worth noting that one of the main advantages of this design, when compared to the sensor fabricated with a PEG-based spacer, is the incorporation of a DNA-based spacer, which is both tunable and less synthetically challenging. While our previously published sensor, as designed, has good sensitivity and specifcity,23 we have yet to evaluate the effect of the length and identity of the spacer in the probe on sensor performance. Thus, in this study we have systematically characterized the effect of the spacer length on the sensors’ signaling capacity, specificity, and selectivity. A total of six DNA probes, each with a different oligo-T or oligo-A spacer, were investigated. In addition to the full-complement DNA target that does not contain a spacer (M0), 12 other targets, each with a different midsequence T or A insertion, were used to elucidate the effect of target flexibility on sensor performance. Both alternating current (AC) voltammetry and cyclic voltammetry (CV) were used as sensor interrogation techniques, in addition to providing evidence to support the proposed signaling mechanism. A fundamental understanding of the effect of probe and target flexibility on sensor response could offer new insights into the design of future generations of E-DNA sensors.
EXPERIMENTAL PROCEDURES AND METHODS Materials. 6-Mercapto-1-hexanol (C6-OH), tris(2carboxyethyl)phosphine hydrochloride (TCEP), sodium hydroxide (NaOH), potassium chloride (KCl), magnesium chloride (MgCl2), calcium chloride (CaCl2), 8 M guanidine hydrochloride (GHCl), and bovine calf serum were used as received (Sigma-Aldrich, St. Louis, MO). Synthetic stimulated human parotid saliva was purchased from US Biocontract Inc. (San Diego, CA). All other chemicals were of analytical grade. All of the solutions were made with deionized water (DI) purified through a Synergy Ultrapure Water System (18.2 MΩ· cm, Millipore, Billerica, MA). The interrogation buffer was either a physiological buffer solution (Phys2) consisting of 20 mM Tris, 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, and 1 mM CaCl2 (pH 7.4) or 25% saliva (with 75% Phys2). Six linear DNA probes purchased from Biosearch Technologies, Inc. (Novato, CA) were used as received (see Table 1). The DNA probes were modified at the 5′-end with a C6disulfide (HO-(CH2)6-S-S-(CH2)6-5′DNA) linker and at the 3′ end with a MB redox label. The probes have two target binding domains (underlined), including an 8-base region at the 5′-end and a 9-base region at the 3′-end (Figure S1 of the Supporting Information (SI)). Fifteen target DNA sequences purchased from Integrated DNA Technologies (Coralville, IA) were used without further purification (see Table 2). The midsequence spacer is underlined, and the mismatch bases are bold underlined. E-DNA Sensor Fabrication. Prior to sensor fabrication, gold disk electrodes with a geometric area of 0.0314 cm2 (CH Instruments, Austin, TX) were polished with 0.1 μm diamond slurry (Buehler, Lake Bluff, IL), rinsed with DI water, and sonicated in a low-power sonicator for ∼5 min to remove bound particulates. They were then electrochemically cleaned B
dx.doi.org/10.1021/ac5027226 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
For the full complement targets, the ratio between the MB peak current in the presence and absence of the target DNA was used to calculate the % signal enhancement (SE) (eq 3).
by a series of oxidation and reduction cycles in 0.5 M H2SO4. The real surface area of each electrode was estimated on the basis of the amount of charge consumed during the reduction of the gold surface oxide in 0.05 M H2SO4 using a reported value of 400 μC cm−2.24 The roughness factor (real area/ geometric area) of the electrodes used in this study ranged from 1.0 to 1.5. Fabrication of the sensors involved several steps. First, 1 μL of the 200 μM DNA probe solution was mixed with 1 μL of 10 mM TCEP; this solution was left at room temperature (∼23 °C) for 1 h to reduce the disulfide bonds. The solution was then diluted with a 10 mM phosphate buffer saline supplemented with 0.1 M NaCl (pH 7.4). The diluted solution of the six DNA probes (1 μM T6, 0.2 μM T10, 1 μM T14, 1 μM A6, 2 μM A10, and 1 μM A14) was drop casted onto freshly cleaned gold electrodes for 1 h. The electrodes were then rinsed with water and subsequently passivated with 2 mM C6-OH overnight to displace nonspecifically bound DNA probes. The density of electroactive DNA probes on the electrode surface, Γ, was determined by integrating the charge under the MB reduction peak in CV scans collected at slow scan rates (eq 1). Γ = Q /nFA
Signal Enhancement (%) = [(I − I0)/I0] × 100
where I is the baseline-subtracted peak current obtained in the presence of the target, and I0 is the baseline-subtracted peak current in the target-free solution. For the two mismatch targets, the ratio between the MB peak current in the absence and presence of the target DNA was used to calculate the % signal suppression (SS) (eq 4). Signal Suppression (%) = [(I0 − I )/I0] × 100
(4)
where I0 is the baseline-subtracted peak current in the targetfree solution, and I is the baseline-subtracted peak current obtained in the presence of the target. Sensor regeneration was achieved by rinsing the electrode surface with DI water for 40 s. Sensor interrogation− regeneration experiments were performed in 25% synthetic human saliva using MA3 (1.0 μM) as the target. All experiments were performed at room temperature and without mechanical stirring. Unless mentioned otherwise, all experimental results presented here are averaged from three different sensors.
■
(1)
RESULTS AND DISCUSSION Sensor Design. The design and signaling mechanism of the six sensors are shown in Scheme 1. Fabrication of these sensors
where Q is the integrated charge of the reduction peak in the CV scans, n is the number of electrons transferred per redox event (n = 2 for MB), F is the Faraday′s constant, and A is the real electrode area. Γ for each of the six systems is presented as an average value from CVs recorded at three different scan rates (20, 50, and 100 mV s−1). MB peak currents obtained in CV at different scan rates were used to determine the apparent diffusion coefficient (Dapp) for the two best performing sensors (i.e., T14 and A14 sensors) upon hybridization to M0 and the six A targets using the following equation.25 Iopen /Iclose = 0.286Dapp1/2 /(n1/2ν1/2l)
(3)
Scheme 1. Design and Signaling Mechanism of the Six Sensors Used in This Study
(2)
where Iopen is the MB peak current from the probe in the singlestranded state before target addition, Iclose is the MB peak current from the probe in the “close” conformation after the addition of the target, Dapp is the apparent diffusion coefficient of MB, n is the number of electrons (n = 2 for MB), ν is the voltammetric scan rate, and l is the length of the DNA probe (T14 and A14 probes). Electrochemical Measurements. Electrochemical measurements were performed at room temperature using a CHI 1040A Electrochemical Workstation (CH Instruments, Austin, TX). The E-DNA sensors were characterized by AC voltammetry over a wide range of frequencies (1−2000 Hz) using an amplitude of 25 mV. CV was used to determine the surface probe coverage and Dapp. DNA probe-modified gold disk electrodes were used as working electrodes. A platinum wire was used as the counter electrode and a Ag/AgCl (3 M KCl) electrode served as the reference electrode (CH Instruments, Austin, TX). Prior to target interrogation, the probe-modified electrodes were allowed to equilibrate in Phys2 or 25% saliva for at least 20 min. AC voltammograms were collected every 5 min after the addition of the target until signal saturation had been achieved (i.e., stable MB peak current). The target concentration was 1.0 μM for all hybridization experiments.
requires direct immobilization of a thiolated and MB-modified DNA probe onto a gold electrode, and subsequent electrode passivation with C6-OH. The structure of the dual-labeled DNA probes used in this study is shown in Figure S1. The DNA probes are linear unstructured probes, all containing a 9base target recognition region at the 3′-end and an 8-base target recognition region at the 5′-end. The two regions are connected through an oligo-T or oligo-A spacer that is 6, 10, or 14 bases in length. Two additional T(s) are added to the 5′end of the six DNA probes to improve probe flexibility. The two target recognition regions are designed to be fully complementary to the 17-base target without a spacer (M0). As designed, in the absence of the target, the probe is quite flexible; electron transfer between the tethered MB and the electrode is thus less efficient. Target hybridization induces formation of a “‘close’” probe structure, confining MB close to the electrode surface for efficient electron transfer. This change C
dx.doi.org/10.1021/ac5027226 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 1. AC voltammograms of the T14 (A) and A14 (B) sensors in the absence, presence of 1.0 μM MA3, and after sensor regeneration. All scans were recorded in Phys2 at 600 Hz.
Figure 2. %SE for all six sensors upon hybridization to the 13 targets used in this study. The number indicates the number of bases in the spacer.
D
dx.doi.org/10.1021/ac5027226 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
electron transfer kinetics after target hybridization is much faster when compared to that in the unhybridized state. Since the %SE is calculated using both pre- and post-hybridization MB currents, it is also dependent on the applied frequency. For these sensors, 600 Hz was the optimal interrogation frequency and was thus used in most parts of this study. In addition to MA3, we systematically evaluated each sensor’s response to 12 other targets (Figure 2). Apart from M0, all other targets have a midsequence oligo-T or oligo-A spacer that serves to minimize steric hindrance to surface binding. For the T6, T10, and T14 sensors, hybridization to MT3 resulted in the highest %SE, followed by MT2. The lowest %SE was observed with M0, followed by MT6 (Figure 2A−C). These results are not unexpected; hybridization to the target without a spacer is unfavorable because of both steric and electrostatic interferences. However, targets with a very long spacer (e.g., MT5, MT6) are not ideal for this application either. Although the extra bases in the spacer should not affect the hybridization between the 3′-end of the target sequence and the distal end of the probe, they could affect the binding between the 5′-end of the target sequence and the proximal end of the probe. This interaction is necessary for the formation of the “close” structure, which gives rise to the increase in the MB current (Scheme 1). It is worth noting that the T14 sensor showed higher %SE for all seven targets when compared to both T6 and T10 sensors, highlighting the advantages of including a longer spacer between the two target recognition regions of the probe. In addition to the T targets, we also evaluated these three sensors’ response to the A targets (Figure 2A−C). As can be seen, all three sensors responded better to the A targets. This effect was more pronounced for the T6 sensor; for example, only ∼326%SE was obtained for MT3, whereas ∼832%SE was seen with MA3 (Figure 2A). However, this was not the case for targets with a 6-base spacer; MA6 showed substantially lower %SE than MT6. Overall, these results indicate that all three sensors are better suited for detection of A targets; they are also more adept at differentiating between A targets with a different spacer length (i.e., better at differentiating between MA3 and MA6 than MT3 and MT6). To further understand the reasons behind the better signal enhancement seen with the A targets, we interrogated the A6, A10, and A14 sensors with the same T and A targets (Figure 2D−F). These three sensors responded rather poorly to the T targets, in particular, the A6 sensor. Independent of the target spacer length, the A6 sensor showed the lowest %SE for all T targets when compared to both A10 and A14 sensors. Similar to the T14 sensor, the A14 sensor showed the highest %SE for all targets (Figure 2C and F). For these sensors, hybridization to MT6 actually led to a reduction in the MB current (i.e., “signal-off” behavior), suggesting the target’s inability to efficiently bind to the proximal end of the probe to generate the “close” conformation. It is worth noting that the A6, A10, and A14 sensors responded significantly better to the A targets when compared to the T targets. This effect was the most prominent for the A6 sensor, in which only ∼187%SE was observed with MT3, whereas hybridization to MA3 resulted in ∼863%SE. These results prove that the interactions between the oligo-A (or oligo-T) spacer and the midsequence inserted T(s) (or A(s)), if present, do not contribute to the enhanced signal change seen with the A targets. The sensors’ preference for the A targets alludes to the fundamental differences between oligo-T and oligo-A. The exact reason behind this behavior is not clear, given that T is a smaller base and oligo-T is known to
in probe conformation also leads to a measurable increase in Dapp of the tethered MB, resulting in an increase in the redox current.22,25 In addition to M0, targets with an additional oligoT (MT1−6) or oligo-A (MA1−6) spacer were used in this study. The goal here is 2-fold; first, we aim at understanding the effect of both the length and identity of the spacer in the probe sequence on the overall sensor performance. We also intend to elucidate the impact of the type of midsequence insertion in the target on the signaling capacity of the sensor. Both aspects are equally important for enabling optimal sensor performance, which is critical for this sensor, a sensor that is amenable to real world applications. Sensor Characterization Using Alternating Current Voltammetry. AC voltammetry is one of the most commonly used sensor interrogation techniques for this class of foldingand dynamics-based electrochemical biosensors; thus, it was first used to characterize these six sensors. In the absence of the target DNA, we observed a well-defined MB peak at ∼ −0.27 V (vs Ag/AgCl) for all six sensors; this potential is consistent with the reduction potential of MB in this medium (Figure 1 and Figure S2 of the SI). A large increase in the MB current was observed in the presence of MA3, a target sequence with a 3A spacer, suggesting formation of the presumed “close” probe structure. The %SE ranged from 650 to 1250%, with T14 and A14 sensors showing the largest signal change (Figure 1). Like most E-DNA sensors, these sensors can be easily regenerated via a simple 40-s deionized water rinse.9,15 Sensor regeneration efficiency was close to 100%, as evidenced by the overlapping AC voltammograms recorded before hybridization and after sensor regeneration. Optimal probe coverages for target hybridization for the six sensors are shown in Figure S3 of the SI. Sensors with a higher probe coverage were not used because of the low %SE; this effect is likely the result of steric hindrance that is known to affect target hybridization. Among the six sensors, the T10 sensor had the lowest probe coverage; this coverage, however, was found to be optimal for target capture for this sensor. It is worth noting that all six sensors showed fast binding kinetics; signal saturation can be achieved in ∼60 s. Despite the difference in %SE, the response time is very similar among the sensors (Figure S4 of the SI). This response time is faster than most sensors of this class, which require 15−45 min to reach signal saturation. This improvement could be attributed to the slightly lower probe coverage, as well as the enhanced probe flexibility brought on by the incorporation of a short spacer (e.g., a 3A spacer).9,15 To understand the sensing mechanism, we examined the AC frequency-dependent current response of all six sensors before and after hybridization with MA3 (Figure S5 of the SI). For this class of sensors, the AC peak current increases with increasing frequency, and the current decreases at higher frequencies when the electron transfer events cannot keep up with the oscillating potential.26,27 The “threshold” frequency where the MB current decreases drastically is inherently dependent on a rate-limited reaction, which is the elastic diffusion of the tethered MB on the electrode.28−31 This rate-limiting reaction is also known to be affected by the surface probe coverage, as well as the design of the sensor. For all six sensors, in the absence of the target, the MB current increased between 1 and 25 Hz, followed by a sharp decrease at frequencies beyond 25 Hz. This profile is consistent with observations from E-DNA sensors fabricated with linear unstructured probes.32 In the presence of MA3, the MB peak current increased rapidly with frequency and plateaued at ∼800 Hz, suggesting that the E
dx.doi.org/10.1021/ac5027226 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
behave similarly to an ideal self-avoiding walks (SAW) polymer at intermediate-to-high monovalent ion concentrations.33 The rigidity of oligo-T is mostly due to steric restrictions on backbone conformation.34 Oligo-A, however, is capable of forming an extensively stacked structure at low temperature. Previous studies have shown that even at room temperature two-thirds of the A bases are stacked.34,35 Thus, it is possible that the stacking of the inserted A bases stabilizes the interactions between the target and the surface-immobilized probe. Further studies are required to confirm this hypothesis. Nevertheless, these sensors have demonstrated capabilities in differentiating between targets with a different spacer sequence. In addition to the difference in %SE, hybridization kinetics also differs depending on the target sequence. All six sensors responded more slowly (∼5 min) to targets without or with a very long spacer (e.g., M0, MA6), especially when compared to targets with an optimal spacer length (e.g., MA3). This is expected given that targets with an optimal spacer length have minimal steric hindrance for hybridization. Our results allude to the application of these sensors in discriminating between a wild type target and a mutated target with a midsequence base insertion. Mixed sequence spacers are also of interest, and these experiments are currently underway in the laboratory. Sensor Characterization Using Cyclic Voltammetry. Although AC voltammetry has been the electrochemical interrogation technique of choice, CV is an equally effective technique for both sensor interrogation and characterization. We further characterized these sensors using CV, and their response to MA3 is shown in Figure S6 of the SI. In the absence of the target, we observed a large voltammetric hysteresis (i.e., peak-to-peak separation), indicating slow electron transfer kinetics of MB. A reduction in the voltammetric hysteresis was clearly evident in the presence of MA3. An improvement in the electron transfer kinetics is expected given that the Dapp of the tethered MB should increase upon target hybridization. To verify this, we employed a previously developed method to determine the Dapp of MB for the two best performing sensors, the T14 and A14 sensors, in the presence of M0 and the six A targets. Shown in Figure 3 are the corresponding %SE and Dapp for the two sensors in the presence of the targets. The Dapp increased substantially with increasing spacer length in the target sequence; however, hybridization to targets with a very long spacer such as MA6 resulted in a lower Dapp. More importantly, the change in the Dapp clearly reflects the change in the %SE recorded in the presence of the targets. These results allude to the effect of probe conformational change, in this case, the formation of a “close” structure, on the elastic diffusion of the MB label.28 Sensor Specificity. Sensor specificity is crucial for all real world sensing applications; here we evaluated all six sensors’ responses to MA3-1M and MA3-2M, targets with a 1-base and 2-base mismatch, respectively. These two targets have the same core sequence as MA3, the target with a 3A spacer. As shown in Figures 4A, C and S7, all six sensors behaved as “signal-off” sensors in the presence of MA3-1M, a target with an internal C−C mismatch at the 5′-end of the sequence. The %SS ranged from 67 to 85%, with the T10 sensor showing the largest %SS. This “‘signal-off”’ behavior is expected, since MA3-1M is capable of hybridizing to the 9-base region at the 3′-end of the probe, but the 1-base mismatch inhibits binding of the target to the 8-base region at the 5′-end of the probe, the main interaction that enables formation of the “‘close’” structure. However, formation of the 9-base duplex at the 3′-end reduces
Figure 3. %SE and Dapp for the T14 (A) and A14 (B) sensors in the presence of M0 and the six A targets. The number indicates the number of bases in the spacer.
probe flexibility, altering the apparent diffusion of the tethered MB, resulting in a large reduction in the MB current. While hybridization to MA3-1M gave rise to a large reduction in the MB current, hybridization to MA3-2M, a target with two internal C−C mismatches at the 3′-end of the sequence, did not result in a substantial change in the MB current in AC voltammetry (Figures 4B, D and S8). The %SS ranged from 6 to 9%, with the T14 sensor showing the largest %SS. The presence of the two mismatch bases could inhibit hybridization between the 3′-end of the target sequence and the distal end of the probe. While hybridization between the 5′-end of the sequence and the proximal end of the probe is not affected by the mismatch bases, this interaction should not have a major effect on the signaling of this sensor, given that the MB label is located at the distal end of the probe. A minor change in the probe flexibility, however, is anticipated, which could result in a slight decrease in the MB current. It is worth mentioning that the AC frequency-dependent profiles obtained in the presence of MA3-1M and MA3-2M are very different from that recorded in the presence of MA3, the perfect match target (Figures S5F and S9 of the SI). The profiles of the other five sensors in the presence of these two mismatch targets are very similar to the ones shown in Figure S9 (data not shown). In addition to AC voltammetry, CV was also used to evaluate sensor specificity. Shown in Figure S10 are the sensors’ responses to the mismatch targets, MA3-1M and MA3-2M, when interrogated using CV at a scan rate of 100 V s−1. Hybridization to MA3-1M led to a large reduction in the size of F
dx.doi.org/10.1021/ac5027226 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 4. AC voltammograms of the T14 and A14 sensors in the absence and presence of 1.0 μM MA3-1M (A, C) and 1.0 μM MA3-2M (B, D). Shown on the left is the presumed signaling mechanism responsible for the observed signal change.
Figure 5. Interrogation−regeneration plots for the sensors with an oligo-T (A) and oligo-A (B) spacer. The sensors were interrogated with 1.0 μM MA3 in 25% saliva at an AC frequency of 600 Hz.
the A14 sensor, we observed ∼1151 and ∼953%SE in pure buffer and 25% saliva, respectively. Apart from the slight difference in %SE, other important sensor properties such as binding kinetics, regenerability, and reusability remained unchanged. Independent of the interrogation solvent, the T14 and A14 sensors showed higher %SE than the rest of the sensors. Even after exposure to this complex solvent, all sensors regenerated close to 100% after a simple, 40-s rinse with DI water. The sensors were reused twice after the first regeneration; only a slight change in the %SE was evident after each regeneration cycle.
the MB peaks, which corresponds well with the signal reduction seen in AC voltammetry. Similar to the results shown in AC voltammetry, only a minor reduction in the size of the redox peaks was evident in the presence of MA3-2M for all six sensors. Overall, independent of the electrochemical interrogation technique, all six sensors have demonstrated excellent specificity for the perfect match target. Sensor Selectivity and Reusability. Good sensor selectivity is required for all sensing applications, in particular, for this class of reagentless and reusable sensors that can potentially be used in point-of-care diagnostics. Here we interrogated all six sensors with MA3 in a realistically complex medium such as 25% synthetic human saliva (Figure 5). When compared to the %SE obtained in pure buffer, the %SE obtained in 25% saliva was slightly lower for all six sensors. This could be due to the combined effects of reduction in the ionic strength and an increase in solvent viscosity. For example, for
■
CONCLUSION
In summary, we have systematically evaluated the effect of the length and identity of a nontarget binding spacer in both the probe and target sequences on the overall performance of a folding-based electrochemical DNA sensor. Among the six G
dx.doi.org/10.1021/ac5027226 | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
(19) Xiao, Y.; Lubin, A. A.; Baker, B. R.; Plaxco, K. W.; Heeger, A. J. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 16677−16680. (20) Yu, Z.; Lai, R. Y. Anal. Chem. 2013, 85, 3340−3346. (21) Yu, Z.; Lai, R. Y. Chem. Commun. 2012, 48, 10523−10525. (22) Immoos, C. E.; Lee, S. J.; Grinstaff, M. W. J. Am. Chem. Soc. 2004, 126, 10814−10815. (23) Wu, Y.; Lai, R. Y. Chem. Commun. 2013, 49, 3422. (24) Angerstein-Kozlowska, H.; Conway, B. E.; Hamelin, A.; Stoicoviciu, L. J. Electroanal. Chem. 1987, 228, 429−453. (25) Farjami, E.; Campos, R.; Ferapontova, E. E. Langmuir 2012, 28, 16218−16226. (26) Sumner, J. J.; Weber, K. S.; Hockett, L. A.; Creager, S. E. J. Phys. Chem. B 2000, 104, 7449−7454. (27) Sumner, J. J.; Creager, S. E. J. Phys. Chem. B 2001, 105, 8739− 8745. (28) Abi, A.; Ferapontova, E. E. J. Am. Chem. Soc. 2012, 134, 14499− 144507. (29) Huang, K.-C.; White, R. J. J. Am. Chem. Soc. 2013, 135, 12808− 12817. (30) Anne, A.; Demaille, C. J. J. Am. Chem. Soc. 2006, 128, 542−557. (31) Anne, A.; Demaille, C. J. J. Am. Chem. Soc. 2008, 130, 9812− 9823. (32) Ricci, F.; Lai, R. Y.; Plaxco, K. W. Chem. Commun. 2007, 36, 3768−3770. (33) Sim, A. Y. L.; Lipfert, J.; Herschlag, D.; Doniach, S. Phys. Rev. E 2012, 86, 021901. (34) Mills, J. B.; Vacano, E.; Hagerman, P. J. J. Mol. Biol. 1999, 285, 245−257. (35) Inners, L. D.; Felsenfeld, G. J. Mol. Biol. 1970, 50, 373−389.
probes used in this study, the probes with a 14-base spacer are best suited for this sensor design; this long spacer provides adequate flexibility for improved target capture. Among the 13 targets investigated, targets with a 2- or 3-base spacer are ideal as they provide adequate flexibility yet do not introduce unfavorable steric interference and surface crowding. Furthermore, hybridization between the probes and the A targets is more efficient, presumably because of the stacking effect, which is a characteristic of oligo-A. Despite the difference in the probe spacer sequence, all six sensors are capable of differentiating between the perfect match and mismatch targets. Like most EDNA sensors, these sensors are regenerable and reusable even after exposure to a complex medium such as diluted saliva. Owing to these unique attributes, these sensors can potentially be used as diagnostic tools for point-of-care testing.
■
ASSOCIATED CONTENT
* Supporting Information S
Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Fax: +1 402 472 9402. Tel: +1 402 472 5340. E-mail: rlai2@ unl.edu. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This research was supported by the National Science Foundation (CHE-0955439). The authors would like to thank Anita J. Zaitouna for the helpful discussions.
■
REFERENCES
(1) Sassolas, A.; Leca-Bouvier, B. D.; Blum, L. J. Chem. Rev. 2008, 108, 109−139. (2) Vercoutere, W.; Akeson, M. Curr. Opin. Chem. Biol. 2002, 6, 816−822. (3) Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2003, 21, 1192−1199. (4) Odenthal, K. J.; Gooding, J. J. Analyst 2007, 132, 603−610. (5) Dai, N.; Kool, E. T. Chem. Soc. Rev. 2011, 40, 5756−5770. (6) Fan, X.; White, I. M.; Shopova, S. I.; Zhu, H.; Suter, J. D.; Sun, Y. Anal. Chim. Acta 2008, 620, 8−26. (7) Fan, C.; Plaxco, K. W.; Heeger, A. J. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 9134−9137. (8) Mao, Y.; Luo, C.; Ouyang, Q. Nucleic Acids Res. 2003, 31, e108. (9) Yang, W.; Lai, R. Y. Langmuir 2011, 27, 14669−14677. (10) Caňete, S. J. P.; Lai, R. Y. Chem. Commun. 2010, 46, 3941− 3943. (11) Lai, R. Y.; Lee, S.-H.; Soh, H. T.; Plaxco, K. W.; Heeger, A. J. Langmuir 2006, 22, 1932−1936. (12) Yang, W.; Gerasimov, J. Y.; Lai, R. Y. Chem. Commun. 2009, 20, 2902−2904. (13) White, R. J.; Plaxco, K. W. Anal. Chem. 2010, 82, 73−76. (14) Farjami, E.; Clima, L.; Gothelf, K.; Ferapontova, E. E. Anal. Chem. 2001, 83, 1594−1602. (15) Lai, R. Y.; Walker, B.; Stormberg, K.; Zaitouna, A. J.; Yang, W. Methods 2013, 64, 267−275. (16) Xiao, Y.; Qu, X.; Plaxco, K. W.; Heeger, A. J. J. Am. Chem. Soc. 2007, 129, 11896−11897. (17) March, G.; Noël, V.; Piro, B.; Reisberg, S.; Pham, M.-C. J. Am. Chem. Soc. 2008, 130, 15752−15753. (18) Rowe, A. A.; Chuh, K. N.; Lubin, A. A.; Miller, E. A.; Cook, B.; Hollis, D.; Plaxco, K. W. Anal. Chem. 2011, 83, 9462−9466. H
dx.doi.org/10.1021/ac5027226 | Anal. Chem. XXXX, XXX, XXX−XXX