Utilizing a Key Aptamer Structure-Switching Mechanism for the

Feb 12, 2016 - Aptasensing of small molecules remains a challenge as detection often .... An ATMND/SGI based label-free and fluorescence ratiometric ...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/ac

Utilizing a Key Aptamer Structure-Switching Mechanism for the Ultrahigh Frequency Detection of Cocaine Miguel A. D. Neves,† Christophe Blaszykowski,‡ and Michael Thompson*,†,‡ †

Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario, Canada M5S 3H6 Econous Systems, Inc., 80 St. George Street, Toronto, Ontario, Canada M5S 3H6

‡

Downloaded via OPEN UNIV OF HONG KONG on January 23, 2019 at 13:28:35 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Aptasensing of small molecules remains a challenge as detection often requires the use of labels or signal amplification methodologies, resulting in both difficultto-prepare sensor platforms and multistep, complex assays. Furthermore, many aptasensors rely on the binding mechanism or structural changes associated with target capture by the aptameric probe, resulting in a detection scheme customized to each aptamer. It is in this context that we report herein a sensitive cocaine aptasensor that offers both real-time and label-free measurement capabilities. Detection relies on the electromagnetic piezoelectric acoustic sensor (EMPAS) platform. The sensing interface consists of a S-(11trichlorosilyl-undecanyl)benzenethiosulfonate (BTS) adlayercoated quartz disc onto which a structure-switching cocaine aptamer (MN6) is immobilized, completing the preparation of the MN6 cocaine aptasensor (M6CA). The EMPAS system has recently been employed as the foundation of a cocaine aptasensor based on a structurally rigid cocaine aptamer variant (MN4), an aptasensor referred to by analogy as M4CA. M6CA represents a significant increase in terms of analytical performance, compared to not only M4CA but also other cocaine aptamerbased sensors that do not rely on signal amplification, producing an apparent Kd of 27 ± 6 μM and a 0.3 μM detection limit. Remarkably, the latter is in the range of that achieved by cocaine aptasensors relying on signal amplification. Furthermore, M6CA proved to be capable not only of regaining its cocaine-binding ability via simple buffer flow over the sensing interface (i.e., without the necessity to implement an additional regeneration step, such as in the case of M4CA), but also of detecting cocaine in a multicomponent matrix possessing potentially assay-interfering species. Finally, through observation of the distinct shape of its response profiles to cocaine injection, demonstration was made that the EMPAS system in practice offers the possibility to distinguish between the binding mechanisms of structure-switching (MN6) vs rigid (MN4) aptameric probes, an ability that could allow the EMPAS to provide a more universal aptasensing platform than what is ordinarily observed in the literature.

A

by a highly conserved structure (at the exception of the variable paratope region), providing antibody detection technologies with a greater degree of universality as antibodies can be substituted while maintaining the signal transduction scheme. There exists a high degree of structural diversity among aptamers, or even aptamer constructs binding the same target via different mechanisms.3,12−14 Taking the cocaine aptamer as an example, different variants can be engineered to bind via a rigid or structure-switching process. The MN4 cocaine aptamer (Figure 1A, top) is a long stem one rigid construct that binds through a mechanism in which the secondary structure is preformed in the absence of ligand, and tertiary folding occurs upon complexation.13,14 When stem one is shortened to

ptamer-based detection technologies are being developed at a frantic pace; however, the actual implementation and commercialization of aptasensors is greatly lagging behind published reports.1 Antibodies remain the “gold-standard” in analytical and medical laboratories with qualitative immunoassay-based tests being routinely performed.2 Immunoassays, although rapid, have limitations in terms of sensitivity, crossreactivity, and high costs associated with the required monoclonal antibodies. One reason for the dearth of aptasensor commercialization possibly lies in the lack of conserved structure between aptameric probes. Most aptasensor transduction schemes are designed to take advantage of the structural differences (binding mechanism) between free and ligand-bound states/conformations of aptamers,3−5 often requiring aptameric probes to be chemically tagged or modified, or the implementation of signal amplification methodologies.6−11 Antibodies, on the other hand, are characterized © 2016 American Chemical Society

Received: October 22, 2015 Accepted: February 5, 2016 Published: February 12, 2016 3098

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106

Article

Analytical Chemistry

Figure 1. (A) Secondary structures of MN4 and MN6 aptamers and their binding mechanisms with cocaine. (B) Examples of the various interfacial factors and liquid properties that influence the response of BAW piezoelectric sensors in the liquid phase. Reproduced with permission from ref 71. Copyright 2014 Sonia Sheikh.

changes occurring at the liquid/sensor interface, as depicted in Figure 1B.39,40 Several biosensing technologies have been implemented based on BAW devices including the detection of DNA oligos and proteins, as well as aptasensors against protein, DNA, and cellular targets.41−45 Literature accounts of smallmolecule BAW detection are, by comparison, much more sporadic due to the fact that such small organic entities only induce minute mass/viscoelastic changes.46−48 As a result, BAW sensors developed against small molecules rely on the implementation of complex detection strategies such as competition or displacement assays, or multistep signal amplification methodologies.18,49,50 The electromagnetic piezoelectric acoustic sensor (EMPAS), a BAW device developed in our laboratory, may constitute a more universal platform for small-molecule aptasensing.51,52 Unlike other conventional BAW sensors (i.e., the TSM), acoustic resonance is induced remotely in an electrode-free

contain three base pairs such as in MN6 construct (Figure 1A, bottom), the now unstructured aptamer in the free state binds its target via a structure-switching, ligand-induced secondary/ tertiary folding mechanism. Another structure-switching mechanism consists in splitting the aptamer into two or more strands that will come together in the presence of cocaine, inducing secondary structure formation and tertiary folding in the process. This mechanism is referred to as “split aptamer ligation” and is the foundation upon which most cocaine aptasensors are based.6,7,15−38 As evident with the cocaine aptamer example, there is a need for a more universal transduction technique that will function with different aptamer binding mechanisms. Bulk acoustic wave (BAW) sensorsthe most common being the thickness shear mode (TSM) device, better known as quartz crystal microbalance (QCM)allow for real-time monitoring of various such phenomena as mass and viscoelastic 3099

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106

Article

Analytical Chemistry

Scheme 1. Illustration of the Formation of the Sensing Surface (Steps I and II) and Subsequent Cocaine Binding by the Immobilized Structure-Switching MN6 Aptamer (Step III)

quartz piezoelectric substrate by an external magnetic field. This configuration allows for the sensor to operate at ultrahigh frequencies (>1 GHz), equating to a sensitivity up to 3 orders of magnitude greater than that of the TSM,51−53 and for surface chemistry to be carried out directly on the quartz disc rather than on plated metal electrodes.51 Functionalization of EMPAS quartz substrates is achieved using organosiloxane-based adlayers. These create an environment favorable to the subsequent controlled immobilization of biological probes, and are capable of resisting nonspecific adsorption. Quartz surfaces are located in a flow-through cell, wherein samples are injected in a running buffer, allowing the EMPAS to operate in an online and real-time detection format. Reported EMPASbased technologies include sensors for the detection of proteins54,55 and bacterial lipopolysaccharide pathogens,56 as well as the evaluation of surface fouling.57,58 Notably, the EMPAS has recently been implemented as the foundation of a small-molecule aptasensor for cocaine detection, termed M4CA.59 The sensor was based upon rigid MN4 cocaine aptamer (Figure 1A) that was immobilized onto a S-(11trichlorosilyl-undecanyl)benzenethiosulfonate (BTS) adlayercoated quartz substrate. M4CA featured a facile one-step, labeland additive-free, regenerable assay without the need for signal amplification. The determined limit of detection was, to our knowledge, one of the best reported for a cocaine aptasensor not relying on signal amplification.59 Herein, we introduce the next-generation EMPAS cocaine aptasensor (M6CA) that incorporates the structure-switching MN6 aptamer variant onto the surface configuration previously implemented (Scheme 1). The sensor produces a limit of detection comparable to cocaine aptasensors that yet rely on

multistep signal amplification, while maintaining a one-step, label- and additive-free assay format. Furthermore, we not only show that the coupling of the MN6 structure-switching aptamer with the EMPAS system enhances the sensitivity of cocaine detection, but also that the EMPAS provides a unique means to differentiate between the binding mechanism of structureswitching vs rigid aptameric probes. Together with previous work with M4CA, this study indeed demonstrates the versatility of the EMPAS system (and intercalated surface chemistry) as two aptameric probes that bind cocaine via two different mechanisms could be accommodated; perhaps offering the universal platform that aptamer-based sensing technologies need to achieve widespread adoption.

■

EXPERIMENTAL SECTION Unless otherwise specified, all reagents were purchased from Sigma-Aldrich and used as received. Furthermore, all buffers and other aqueous solutions were prepared using ultrapure distilled deionized water (ddH2O) with a measured resistivity ≥18.0 MΩ·cm. Aptamer Purification and Deprotection. 5′-Thiolmodified MN6 and MN4 (tmMN6 and tmMN4) DNA aptamer samples (aptamer-(CH2)6-S-S-(CH2)6-O-DMT) were obtained from the University of Calgary Core DNA Service (Calgary, Alberta, Canada). DNA samples were purified by denaturing (7 M urea) 20% polyacrylamide gel electrophoresis. With the use of the Elutrap Electroelution System (Whatman), the DNA samples were electroeluted from the gel; then recovered, pooled, and exchanged three times with sterilized 1 M NaCl using a 3 kDa molecular mass cutoff centrifugal filter 3100

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106

Article

Analytical Chemistry

frequency signal was allowed to restabilize. [If another injection was to be made, it would be made at this point.] The net frequency shift was then calculated to be the difference between pre- and post-injection (re)-stabilized frequency baselines. The apparent dissociation constant (Kd) and concentration limit of detection (CLoD) of the MN6 cocaine aptasensor were determined through construction of a dose−response curve with 0.5, 2.5, 5, 10, 25, 35, 50, 75, and 100 μM cocaine solutions in PBS buffer (10 mM NaHPO3, 138 mM NaCl, 2.7 mM KCl, pH 7.4) that were run on the EMPAS, in triplicate or more. [Note: only frequency shifts of the first injection (in the case of multi-injection experiments) were included in the dose−response curve.] The resultant frequency shifts were plotted against concentration and analyzed with the SigmaPlot software package (Systat Software). The Kd was determined by fitting the dose−response curve to a one-site saturation Langmuir ligand-binding model. The CLoD was determined from the standard deviation of the residuals obtained from the least-squares regression of the linear region (0.5−5 μM) of the curve.62−64

(Millipore Amicon Ultra Centrifugal Filter), and washed three times with ddH2O. The day prior to aptameric probe immobilization, the thiolprotecting group (−S−(CH2)6−O−DMT) was cleaved from the 5′-thiol modification by incubating the purified aptamer samples in dithiothreitol (DTT)-containing buffer (20 mM borate, 100 mM DTT, pH 8.5) for 1 h. Following deprotection, aptamer samples were dialyzed overnight into ddH2O using a 2 kDa cutoff Slide-A-Lyzer dialysis cassette (Thermo Scientific). Sensing Surface Preparation. AT-cut piezoelectric quartz discs were obtained from Lap-Tech Precision, Inc. (Bowmanville, Ontario, Canada) with a 13.5 mm diameter and 20.0 MHz fundamental frequency. The quartz discs were cleaned as previously described.54−56,59 Following the cleaning protocol, the quartz substrates were incubated overnight at room temperature in a humidity chamber maintained at 80% relative humidity. The cross-linker, S-(11-trichlorosilyl-undecanyl)benzenethiosulfonate (BTS), was prepared as previously described.60 All silanization glassware was pretreated with octadecyltrichlorosilane (1/20 (v/v) solution in anhydrous toluene) to prevent any undesired reaction of BTS molecules with the walls of the test tubes used during silanization. Under an anhydrous N2 atmosphere (glovebox), the trichlorosilanecontaining linking molecule was deposited onto the quartz surface through wet chemistry as previously described.54,55,59,61 BTS-silanized quartz substrates were individually immersed in 1 mL of 10 μM purified and deprotected aptamer in immobilization solvent [10 μM mercaptohexanol (MCH) dissolved in 1/1 (v/v) ddH2O/spectrograde methanol] for 16 h. [Prior to immersing quartz discs, the aptamer-containing immobilization solution was heated in a boiling water bath for 5 min, and then cooled on ice to anneal the aptamer.] The aptamer-functionalized quartz discs were then rinsed twice with ddH2O, sonicated for 5 min, and then rinsed a final time with a clean portion of ddH2O. The ddH2O rinsing protocol was then repeated with spectrograde methanol. Discs were finally dried under a stream of N2 gas, and then placed into screw-top vials in preparation for EMPAS analysis. X-ray Photoelectron Spectroscopy. Angle-resolved Xray photoelectron spectroscopy (XPS) to evaluate quartz substrate silanization (adlayer formation) and subsequent aptamer immobilization was performed with a Theta probe XPS Instrument (ThermoFisher Scientific) located at Surface Interface Ontario (University of Toronto, Toronto, Ontario, Canada). Quartz surfaces were analyzed with monochromated Al Kα X-rays at takeoff angles of 27.5, 42.5, 57.5, and 72.5° relative to the normal. The binding energy scale is calibrated to the C1s signal at 285 eV. Peak fitting and data analysis were performed with the Avantage Data System software package (ThermoFisher Scientific) provided with the instrument. Complete XPS data are tabulated in the accompanying Supporting Information (Table S1). Sensing Surface Evaluation. Samples for EMPAS analysis were prepared in PBS buffer (10 mM NaHPO3, 138 mM NaCl, 2.7 mM KCl, pH 7.4). Upon the standard EMPAS setup procedure,52 each prepared disc was inserted into the EMPAS flow-cell and PBS used as the running buffer at a flow-rate of 40 μL/min. Once the signal stabilized at the desired frequency (∼860 MHz, 43rd harmonic), the 50 μL sample loop was loaded with a freshly prepared sample and injected into the flow-cell, using a low-pressure liquid chromatography valve. Once the uninterrupted PBS buffer flow carried the cocaine sample through the flow-cell over the sensing surface, the

■

RESULTS Sensing Surface Preparation. M6CA aptasensing surfaces were built upon the interface configuration developed for M4CA, a BTS-coated quartz substrate upon which the aptameric probe in conjunction with mercaptohexanol (MCH) spacer is immobilized (Scheme 1).59 The aptameric probe implemented in this case is the structure-switching MN6 cocaine aptamer variant. Each stage of the M6CA manufacturing process was characterized using XPS surface analysis to monitor the characteristic elements of the quartz substrate (O and Si), BTS adlayer (C and S), and MN6 DNA aptamer (N). XPS data are tabulated in Table S1. As expected, the XPS acquired from the cleaned and adlayer-coated quartz discs followed the same trends previously reported for M4CA.59 Formation of the BTS adlayer caused an attenuation in the silicon and oxygen signals of the now underlying substrate, and the appearance of sulfur sulfide and sulfone peaks (at 164 and 169 eV, respectively) and of a carbon signal well above that of adventitious contamination.59 Following BTS silanization, the structure-switching tmMN6 aptamer was co-immobilized at a 1:1 molar ratio with MCH spacing molecule (Scheme 1). XPS profiles followed a similar trend to that observed when tmMN4 was immobilized under the same conditions: while the carbon signal rose slightly (and the oxygen and silicon peaks were further attenuated), a signal for nitrogen appeared and that of sulfur was only of the sulfide variety (164 eV), as seen in Table S1. These observations confirm the successful immobilization of MN6 aptameric probe (Scheme 1, reaction II) during which the sulfone-containing leaving group (Ph−SO2−) of the BTS head function is displaced by the thiol-modified aptamer (or thiol-containing MCH spacer), forming a covalent disulfide linkage with the underlying organosiloxane adlayer. Target Interaction. Discs at each stage of the manufacturing process were loaded into the EMPAS flow-cell and exposed to 100 μM cocaine in PBS buffer. During analysis, the EMPAS system was operated at the ultrahigh resonant frequency of ∼860 MHz (43rd harmonic). Bare cleaned quartz and BTScoated discs showed no response to cocaine exposure, matching previously reported results for the M4CA study.59 However, when M6CA was contacted with 100 μM cocaine in PBS buffer (Figure 2A, first injection), a large decrease in frequency was observed. The average difference between pre- and post-shift 3101

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106

Article

Analytical Chemistry

of cocaine until the 50 μM plateau. Frequency shifts were plotted against cocaine concentration to obtain a binding curve shown in Figure 3. The apparent Kd of the sensor was

Figure 3. Dose−response curve for M6CA exposed to a range of cocaine concentrations (n ≥ 3). The apparent Kd was calculated to be 27 ± 6 μM and the concentration limit of detection 0.3 μM, for a ∼0.5−5 μM dynamic range.

determined to be 27 ± 6 μM by fitting data to a one-site saturation ligand-binding model. Analysis of the dose−response curve revealed a small dynamic range spanning 0.5−5 μM. The dynamic range frequency shifts were plotted against concentration and fit to a least-squares linear regression. The regression parameters were used to determine the concentration limit of detection (CLoD), which in the case of M6CA was calculated to be 0.3 μM. Performance in a Multicomponent Sample Matrix. In an attempt to mimic real-world sample testing conditions, a multicomponent matrix (MCM) consisting of benzoic acid, glycine, glucose, neomycin, and the common cocaine metabolites benzoylecgonine and ecgonine methyl ester, each at a 100 μM concentration, was created. The sample matrix was doped with 100 μM cocaine and contacted with both M6CA and M4CA sensing surfaces producing average frequency shifts of −2126 ± 382 and −1277 ± 260 Hz, respectively (Figure 4, blue bars). Both specific adsorption signals are reduced due to matrix effects when compared to shifts obtained with cocaineonly PBS samples (−3323 ± 488 Hz). M6CA’s MCM response

Figure 2. (A) M6CA experiment outlining the sensor response and online cocaine-binding regeneration of the sensing surface upon three successive injections of cocaine in PBS. (B) Experiment with M4CA depicting three repeated injections of 100 μM cocaine in PBS, highlighting near-saturation of the sensing interface after first exposure. Reproduced with permission from ref 59. Copyright 2015 Elsevier.

baselines was determined to be −3323 ± 488 Hz, a value two times larger than the signal generated by the M4CA sensor (−1648 ± 260 Hz, Figure 2B, first injection) for the same cocaine concentration.59 When the EMPAS responses of M6CA and M4CA are compared, it is immediately clearly evident that the two sensors produce very different binding profiles (Figure 2, panel A vs B). M6CA was also tested against the common cocaine metabolites benzoyl ecgonine and ecgonine methyl ester for which this aptasensor pleasantly showed no affinity (Table S2). In our previous work, M4CA sensor was regenerated upon the injection of a chaotropic denaturant, urea, between analyte sample exposure.59 Electrochemical sensors based on short stem one cocaine aptamer constructs, such as MN6, report that electrode washing or soaking in fresh buffer solutions is sufficient for regenerating the binding capability of the sensor.6 This apparently also happened to be the case for M6CA, PBS buffer flowed over the sensing surface being sufficient to cause the aptamer to release its ligand. Figure 2A shows the progress of an M6CA experiment consisting of three repeated injection of cocaine at 100 and 25 μM (2×) concentration. This figure outlines the repeated on-line regenerability of the aptasensor’s cocaine-binding affinity. M6CA Analytical Performance. A series of cocaine concentrations ranging from 0.5 to 100 μM was contacted with the M6CA sensing surface (data tabulated in Table S2, only frequency shifts of the first cocaine injection (in the case of multi-injection experiments) were included in the dose− response curve). A clear net decrease in EMPAS resonant frequency is evident upon injection of increasing concentrations

Figure 4. M6CA and M4CA EMPAS responses to cocaine-doped and cocaine-free multicomponent matrices (n = 3). 3102

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106

Article

Analytical Chemistry

Table 1. Comparison of the Analytical Performance (in Buffer) of Cocaine Aptasensors with and without Signal Amplificationa detection method Colorimetricb

Electrochemicalb

Fluorescenceb

Surface Plasmon Resonanceb EMPAS

Colorimetric Fluorescence

Electrochemical

CLoD (μM)

strategy

Sensors without Signal Amplification Split aptamer ligation in the presence of AuNPs (SA) (TB)33 2 5 Aptamer immobilized onto microcantilever sensor surface (RA) (PB)72 Microfluidic microchip embedded with biological nanopore (SS1) 1 (PB)73 Cocaine-induced split aptamer self-assembly on electrode surface 1 (SA) (PB)32 Aptamer hybridized to AuNP-immobilized oligo. Upon cocaine 2.1 binding, AuNP no longer able to quench the aptamer-attached 74 fluorescent tag (RA) (PB) Aptamer folding-based FRET (SS1) (TB)37 10 Cocaine-induced assembly of AuNP-aptamer subunits (SA) (PB)31 1 M4CA: MN4 immobilized onto piezoelectric surface, direct binding 0.9 event causes observed signal (RA) (PB)59 M6CA (this work): MN6 immobilized onto piezoelectric surface, 0.3 added mass and tertiary structure rearrangement cause observed signal (SS1) (PB) Sensors Requiring Signal Amplification 2.8 Split aptamer ligation, streptavidin−horseradish peroxidase (HRP) catalytic oxidation signal amplification (SA) (PB)19 0.2 Minor groove binder-based energy transfer (SS1) (PB)23 Isothermal circular strand displacement amplification (SA) (TB)75

0.19

Cocaine binding releases DNA invertase conjugate that converts sucrose into glucose (SA) (PB)76 Electrochemical split aptamer ligation of enzyme-tethered strands, one strand containing HRP and the other glucose oxidase (SA) (PB)77 Electrochemical aptasensor based on Klenow fragment (KF) polymerase amplification reaction (SS1) (PB)8

1.8

Engineered biotin-labeled split aptamer-pendant DNA tetrahedral structure, streptavidin-HRP signal amplification (SA) (PB)78

0.033

0.5

0.1

limitation(s) Sample must be doped with split aptamer strands and AuNPs One of the least sensitive methods listed in this table Sample doped with split aptamer strand Sample doped with methylene blue-labeled split aptamer strand Aptamer labeled with a fluorescent tag Aptamer labeled with fluorescein and a quencher Sample doped with AuNP-labeled split aptamer strand Narrow dynamic range (∼2−50 μM) and not as sensitive as its structure-switching M6CA counterpart Narrow dynamic range (∼0.5−5 μM)

Sample doped with biotin-labeled split aptamer strand, multistep signal amplification assay Multistep assay, DNA minor groove binding fluorescent additive required Multistep signal amplification process involving DNA polymerization, fluorescent DNA binder and graphene oxide adsorption Multistep process, production of glucose detected by glucose sensor, enzyme specific additives necessary Multistep process employing dual enzyme signal amplification and requiring correct additives for enzymatic reactions to proceed Multistep process, cocaine sample must be doped with aptamer strand. KF polymerase and ferrocene-labeled DNTPs additives are necessary Multistep process, cocaine sample must be doped with aptamer strand, complicated surface preparation

a

(SS1) indicates the use of the short stem one cocaine aptamer variant. (SA) indicates the use of a split (2 stranded) cocaine aptamer variant. (RA) indicates the use of a rigid, nonstructure-switching cocaine aptamer variant. (TB) indicates assay performed in TRIS buffer. (PB) indicates assay performed in PBS buffer. bEntry reproduced with permission from ref 59. Copyright 2015 Elsevier.

aptamer folding.13,14 In contrast, the M6CA sensing interface presented herein relies on an aptameric probe that binds cocaine through a structure-switching, ligand-induced folding mechanism, wherein secondary structure formation and tertiary folding are both induced by ligand complexation.13,14 There is a 7-fold difference between aptamers’ dissociation constants (as determined in solution using isothermal titration calorimetry) with Kd = 45.3 ± 0.5 and 7 ± 1 μM for MN6 and MN4, respectively.13,14 However, once aptamers are immobilized onto EMPAS platforms, a reversal is observed with M6CA producing an apparent Kd of 27 ± 6 μM and 0.3 μM CLoD, while the values for M4CA are Kd = 46 ± 12 μM and CLoD = 0.9 μM.59 M6CA turned out to display greater sensitivity toward cocaine than M4CA (∼73 vs 18 Hz/μM), over a narrower dynamic range however (0.5−5 vs 2−50 μM). Sensor sensitivity was determined utilizing the IUPAC definition, the slope of the linear region of the analytical dose/response curve.65,66 The R2 values for the M6CA and M4CA dose−response curves are 0.9998 and 0.9950, respectively. We believe that the structureswitching binding mechanism of the MN6 aptameric probe plays a role in the increased sensitivity of M6CA over M4CA. Utilization of a structure-switching aptamer allows for M6CA’s cocaine-binding ability to be regenerated in flow (Figure 2A), whereas M4CA requires intermediate 7 M urea injections to denature the aptameric probe and cause it to

is greater than M4CA’s, whether under MCM or PBS conditions (for 100 μM cocaine). M6CA and M4CA sensors were also exposed to cocaine-free MCM resulting in frequency shifts of −248 ± 90 and −470 ± 40 Hz, respectively (Figure 4, gray bars). Interestingly, the immobilized tmMN6 aptameric probe appears to help reduce nonspecific interactions as M6CA’s nonspecific signal is on average approximately half of that produced by M4CA.

■

DISCUSSION Sensor Performance. Most aptasensors are designed to take advantage of the binding mechanism of the aptameric probe to generate a signal from target binding.3−5 The majority of cocaine aptasensors reported in the literature rely on one of the structure-switching aptamer constructs for this reason. 6,7,15−38 Our goal in the design of EMPAS-based aptasensors has always been to try and provide more universality in a sensing platform that would function with not only structure-switching aptamers but also rigid secondary structure variants. The main difference between our two EMPAS-based aptasensors, M6CA and M4CA, is the nature of the immobilized aptameric probe (Figure 1A). Previously reported M4CA utilizes the rigid MN4 cocaine aptamer construct, wherein the secondary structure is preformed prior to ligand binding, cocaine complexation inducing tertiary 3103

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106

Article

Analytical Chemistry

Figure 5. EMPAS experiments outlining the injection of cocaine followed by that of urea, on both M6CA and M4CA aptasensing platforms.

release its target. This finding is in line with literature reports on other cocaine aptasensors relying on short stem one structure-switching constructs, which document regeneration of sensor binding ability to occur after washing/rinsing with fresh buffer.6,59 The analytical performance of both M6CA and M4CA was also evaluated in a multicomponent matrix in an attempt to mimic real-world analysis conditions. Both aptasensors were capable of detecting the presence of cocaine in analyte-doped MCM (Figure 4); signal dampening compared to analysis in buffer was observed however due to matrix effects. M6CA produced a much larger specific binding signal than M4CA, however, as well as a lower nonspecific signal to cocaine-free MCM, perhaps indicating that the unstructured state of the surface-attached MN6 aptameric probe assists in increasing resistance to nonspecific interactions. The ultimate goal of the EMPAS system is to move beyond this artificial matrix toward real-life patient biosamples. In this respect, although detection in MCM constitutes a major step forward from PBS buffer, minimization of nonspecific adsorption is a priority before moving on to such more challenging samples as urine. Work in that direction is currently underway. M6CA operates in a one-step reusable assay format, wherein the direct binding of the cocaine molecule by MN6 aptamer immobilized on the piezoelectric resonator (through an intercalated organosiloxane adlayer) is sufficient to generate real-time signaling, while maintaining a facile two-step sensing surface preparation. This cocaine aptasensor does not require the use of labels, additives mixed within the analyte sample, reporter molecules, signal amplification steps, nor a combination thereof. Most literature examples of cocaine aptamer-based sensors can be divided into two categories, depending on whether they rely on signal amplification methodologies. Aptasensors that do need signal amplification require complex,

time-consuming, multistep assays with a dependence on additives and complex sensor surface configurations.11,18,49,50 Table 1 outlines many recent cocaine aptamer-based sensors that do (not) implement signal amplification methods (to note, all sensors tabulated in Table 1 performed analysis in buffer). M6CA compares well with respect to aptasensors that do not rely on signal amplification, producing to our knowledge one of the lowest CLoD values found in the literature. Some, not all, sensors that incorporate signal amplification steps are still able to produce lower limits of detection, however. Nonetheless, M6CA features the only direct binding nonsignal amplifying method capable of competing with respect to CLoD, to our knowledge. Aptamer Binding Mechanism. A common misconception is that BAW sensors’ response signal only arises from the gravimetric contribution of analyte capture upon the piezoelectric resonator. However, a growing number of literature reports outline instances, where structural rearrangement of immobilized biomolecular probes is the main factor leading to the observed change in resonant frequency.56,67−70 As seen in Figures 2 and 5, M4CA and M6CA aptasensors generate distinctly different EMPAS frequency profiles in response to cocaine. Upon cocaine binding, M4CA produces a frequency drop and immediate new baseline (i.e., no rinse-off), as shown in Figure 2B. This is to be contrasted with the trough-looking profile obtained with M6CA (Figure 2A). The M6CA overall response is similar in shape to that of M4CA during regeneration with 7 M urea (Figure 5), a treatment that denatures the aptamer−cocaine complex and causes it to release its ligand (regenerating the cocaine-binding affinity in the process). We postulate that these sudden EMPAS profile changes reflect the structural reorganization of aptameric probes: complexed MN4 denaturing upon the action of urea, releasing its ligand and refolding; MN6 forming secondary 3104

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106

Article

Analytical Chemistry

XPS analysis. We also thank Dr. Sonia Sheikh (University of Toronto, Toronto, Ontario, Canada) for useful discussion and providing Figure 1B. We also thank Dr. Alexander D. Romaschin (St. Michael’s Hospital, Toronto, Ontario, Canada) for much useful discussion and use of his facilities. The authors are grateful to the Natural Sciences and Engineering Council of Canada for support of this work.

structure and folding upon ligand binding then releasing its target due to buffer flow. Furthermore, when urea is introduced onto the M6CA sensing surface following cocaine injection and flow-induced regeneration, only a slight dip in resonant frequency is observed (Figure 5, top vs bottom second injection). This observation supports the hypothesis according to which M6CA had released its ligand due to buffer flow (but adopted a different conformation over the sensing surface), and that the frequency trough was associated with aptamer structural reorganization (so was the case for M4CA during urea-induced regeneration). The observed increase in sensitivity of M6CA over M4CA is likely due to the EMPAS system being able to detect the structural rearrangement of MN6 aptameric probe upon ligand binding. Only few examples of such phenomenon are displayed in the literature.

■

(1) Luong, J. H. T.; Male, K. B.; Glennon, J. D. Biotechnol. Adv. 2008, 26, 492−500. (2) Wu, A. H. B.; McKay, C.; Broussard, L. A.; Hoffman, R. S.; Kwong, T. C.; Moyer, T. P.; Otten, E. M.; Welch, S. L.; Wax, P. Clin. Chem. 2003, 49, 357−379. (3) Reinstein, O.; Neves, M. A. D.; Saad, M.; Boodram, S. N.; Lombardo, S.; Beckham, S. A.; Brouwer, J.; Audette, G. F.; Groves, P.; Wilce, M. C. J.; Johnson, P. E. Biochemistry 2011, 50, 9368−9376. (4) Xie, S.; Walton, S. P. Anal. Chim. Acta 2009, 638, 213−219. (5) Li, D.; Song, S.; Fan, C. Acc. Chem. Res. 2010, 43, 631−641. (6) Baker, B. R.; Lai, R. Y.; Wood, M. S.; Doctor, E. H.; Heeger, A. J.; Plaxco, K. W. J. Am. Chem. Soc. 2006, 128, 3138−3139. (7) White, R. J.; Phares, N.; Lubin, A. A.; Xiao, Y.; Plaxco, K. W. Langmuir 2008, 24, 10513−10518. (8) He, J.-L.; Yang, Y.-F.; Shen, G.-L.; Yu, R.-Q. Biosens. Bioelectron. 2011, 26, 4222−4226. (9) He, X.; Wang, G.; Xu, G.; Zhu, Y.; Chen, L.; Zhang, X. Langmuir 2013, 29, 14328−14334. (10) Freeman, R.; Sharon, E.; Teller, C.; Willner, I. Chem. - Eur. J. 2010, 16, 3690−3698. (11) Shlyahovsky, B.; Li, D.; Weizmann, Y.; Nowarski, R.; Kotler, M.; Willner, I. J. Am. Chem. Soc. 2007, 129, 3814−3815. (12) Reinstein, O.; Yoo, M.; Han, C.; Palmo, T.; Beckham, S. A.; Wilce, M. C. J.; Johnson, P. E. Biochemistry 2013, 52, 8652−8662. (13) Neves, M. A. D.; Reinstein, O.; Saad, M.; Johnson, P. E. Biophys. Chem. 2010, 153, 9−16. (14) Neves, M. A. D.; Reinstein, O.; Johnson, P. E. Biochemistry 2010, 49, 8478−8487. (15) Li, Y.; Zeng, Y.; Mao, Y.; Lei, C.; Zhang, S. Biosens. Bioelectron. 2014, 51, 304−309. (16) Smith, J. E.; Griffin, D. K.; Leny, J. K.; Hagen, J. A.; Chávez, J. L.; Kelley-Loughnane, N. Talanta 2014, 121, 247−255. (17) Yan, L.; Zhu, Z.; Zou, Y.; Huang, Y.; Liu, D.; Jia, S.; Xu, D.; Wu, M.; Zhou, Y.; Zhou, S.; Yang, C. J. J. Am. Chem. Soc. 2013, 135, 3748− 3751. (18) Shi, Y.; Dai, H.; Sun, Y.; Hu, J.; Ni, P.; Li, Z. Analyst 2013, 138, 7152−7156. (19) Nie, J.; Deng, Y.; Deng, Q.-P.; Zhang, D.-W.; Zhou, Y.-L.; Zhang, X.-X. Talanta 2013, 106, 309−314. (20) Spiropulos, N. G.; Heemstra, J. M. Artif. DNA PNA XNA 2012, 3, 123−128. (21) Das, J.; Cederquist, K. B.; Zaragoza, A. A.; Lee, P. E.; Sargent, E. H.; Kelley, S. O. Nat. Chem. 2012, 4, 642−648. (22) Deng, Q.-P.; Tie, C.; Zhou, Y.-L.; Zhang, X.-X. Electrophoresis 2012, 33, 1465−1470. (23) Zhou, J.; Ellis, A. V.; Kobus, H.; Voelcker, N. H. Anal. Chim. Acta 2012, 719, 76−81. (24) Zhang, D.-W.; Zhang, F.-T.; Cui, Y.-R.; Deng, Q.-P.; Krause, S.; Zhou, Y.-L.; Zhang, X.-X. Talanta 2012, 92, 65−71. (25) Jiang, B.; Wang, M.; Chen, Y.; Xie, J.; Xiang, Y. Biosens. Bioelectron. 2012, 32, 305−308. (26) Sharma, A. K.; Heemstra, J. M. J. Am. Chem. Soc. 2011, 133, 12426−12429. (27) Li, Y.; Ji, X.; Liu, B. Anal. Bioanal. Chem. 2011, 401, 213−219. (28) Zhou, Z.; Du, Y.; Dong, S. Biosens. Bioelectron. 2011, 28, 33−37. (29) Du, Y.; Chen, C.; Yin, J.; Li, B.; Zhou, M.; Dong, S.; Wang, E. Anal. Chem. 2010, 82, 1556−1563. (30) He, J.-L.; Wu, Z.-S.; Zhou, H.; Wang, H.-Q.; Jiang, J.-H.; Shen, G.-L.; Yu, R.-Q. Anal. Chem. 2010, 82, 1358−1364.

■

CONCLUDING REMARKS In summary, we have developed herein the next generation of EMPAS cocaine aptasensor based on structure-switching MN6 cocaine aptamer (as opposed to its MN4 rigid variant described in previous work). Outperforming its predecessor (M4CA), this M6CA aptasensor produces a CLoD that compares well with cocaine aptasensors that implement multistep signal amplification protocols, while offering a facile one-step assay and twostep sensing surface preparation. To our knowledge, this is the only cocaine aptasensing platform capable of achieving such low CLoD in a direct binding one-step detection format without sample additives or signal amplification. M6CA (as well as M4CA, to a lesser extent) was shown to be able to detect cocaine in a multicomponent matrix; however, signal dampening and nonspecific interactions were observed. Further optimization of the sensing interface should allow for the sensor to perform analysis in clinically relevant biosamples. By its nature, the acoustic wave physics of the EMPAS system also allowed to highlight inherent discrepancies in terms of binding mechanism between the rigid MN4 and structure-switching MN6 aptamers, a trait that is believed to be responsible for the increased sensitivity of M6CA over M4CA. More than this, the EMPAS was shown to constitute a versatile platform capable in practice of accommodating different aptameric probes regardless of their detection binding mechanism, an accomplishment rarely encountered in the literature. This represents the major advancement of the proof-of-concept work presented herein.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b04010. Complete M6CA XPS data and EMPAS responses (PDF)

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*Tel.: +1 416 978 3575. Fax: +1 416 978 8775. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Dr. Pasquale Benvenuto (University of Toronto, Toronto, Ontario, Canada) for useful discussion and help with 3105

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106

Article

Analytical Chemistry

(68) Feiler, A. A.; Sahlholm, A.; Sandberg, T.; Caldwell, K. D. J. Colloid Interface Sci. 2007, 315, 475−481. (69) Lee, H.-S.; Contarino, M.; Umashankara, M.; Schön, A.; Freire, E.; Smith, A. B.; Chaiken, I. M.; Penn, L. S. Anal. Bioanal. Chem. 2010, 396, 1143−1152. (70) Lee, H.-S.; Penn, L. S. Macromolecules 2008, 41, 8124−8129. (71) Sheikh, S. Antifouling and Antithrombogenic Ultrathin Surface Chemistry for Bioanalytical and Biomedical Applications, Ph.D. Thesis, University of Toronto, 2014. (72) Kang, K.; Sachan, A.; Nilsen-Hamilton, M.; Shrotriya, P. Langmuir 2011, 27, 14696−14702. (73) Kawano, R.; Osaki, T.; Sasaki, H.; Takinoue, M.; Yoshizawa, S.; Takeuchi, S. J. Am. Chem. Soc. 2011, 133, 8474−8477. (74) Ge, J.; Liu, Z.; Zhao, X. S. Chin. J. Chem. 2012, 30, 2023−2028. (75) Qiu, L.; Zhou, H.; Zhu, W.; Qiu, L.; Jiang, J.; Shen, G.; Yu, R. New J. Chem. 2013, 37, 3998−4003. (76) Zhou, Z.; Xiang, Y.; Tong, A.; Lu, Y. Anal. Chem. 2014, 86, 3869−3875. (77) Freeman, R.; Sharon, E.; Tel-Vered, R.; Willner, I. J. Am. Chem. Soc. 2009, 131, 5028−5029. (78) Wen, Y.; Pei, H.; Wan, Y.; Su, Y.; Huang, Q.; Song, S.; Fan, C. Anal. Chem. 2011, 83, 7418−7423.

(31) Golub, E.; Pelossof, G.; Freeman, R.; Zhang, H.; Willner, I. Anal. Chem. 2009, 81, 9291−9298. (32) Zuo, X.; Xiao, Y.; Plaxco, K. W. J. Am. Chem. Soc. 2009, 131, 6944−6945. (33) Zhang, J.; Wang, L.; Pan, D.; Song, S.; Boey, F. Y. C.; Zhang, H.; Fan, C. Small 2008, 4, 1196−1200. (34) Li, Y.; Qi, H.; Peng, Y.; Yang, J.; Zhang, C. Electrochem. Commun. 2007, 9, 2571−2575. (35) Li, T.; Li, B.; Dong, S. Chem. - Eur. J. 2007, 13, 6718−6723. (36) Stojanovic, M. N.; Landry, D. W. J. Am. Chem. Soc. 2002, 124, 9678−9679. (37) Stojanovic, M. N.; de Prada, P.; Landry, D. W. J. Am. Chem. Soc. 2001, 123, 4928−4931. (38) Stojanovic, M. N.; de Prada, P.; Landry, D. W. J. Am. Chem. Soc. 2000, 122, 11547−11548. (39) Stavila, V.; Volponi, J.; Katzenmeyer, A. M.; Dixon, M. C.; Allendorf, M. D. Chem. Sci. 2012, 3, 1531−1540. (40) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, 1397−1402. (41) Song, W.; Zhu, Z.; Mao, Y.; Zhang, S. Biosens. Bioelectron. 2014, 53, 288−294. (42) He, P.; Liu, L.; Qiao, W.; Zhang, S. Chem. Commun. 2014, 50, 1481−1484. (43) Shan, W.; Pan, Y.; Fang, H.; Guo, M.; Nie, Z.; Huang, Y.; Yao, S. Talanta 2014, 126, 130−135. (44) Ozalp, V. C.; Bayramoglu, G.; Erdem, Z.; Arica, M. Y. Anal. Chim. Acta 2015, 853, 533−540. (45) Wang, R.; Li, Y. Biosens. Bioelectron. 2013, 42, 148−155. (46) Chen, Q.; Wu, X.; Wang, D.; Tang, W.; Li, N.; Liu, F. Analyst 2011, 136, 2572−2577. (47) Sheng, Z.; Han, J.; Zhang, J.; Zhao, H.; Jiang, L. Colloids Surf., B 2011, 87, 289−292. (48) Dong, Z.-M.; Zhao, G.-C. Sensors 2012, 12, 7080−7094. (49) Zhang, D.-W.; Sun, C.-J.; Zhang, F.-T.; Xu, L.; Zhou, Y.-L.; Zhang, X.-X. Biosens. Bioelectron. 2012, 31, 363−368. (50) Sheng, Q.; Liu, R.; Zhang, S.; Zheng, J. Biosens. Bioelectron. 2014, 51, 191−194. (51) Thompson, M.; Ballantyne, S. M.; Cheran, L.-E.; Stevenson, A. C.; Lowe, C. R. Analyst 2003, 128, 1048−1055. (52) Ballantyne, S. M.; Thompson, M. Analyst 2004, 129, 219−224. (53) Stevenson, A. C.; Araya-Kleinsteuber, B.; Sethi, R. S.; Mehta, H. M.; Lowe, C. R. Analyst 2003, 128, 1175−1180. (54) Sheikh, S.; Sheng, J. C.-C.; Blaszykowski, C.; Thompson, M. Chem. Sci. 2010, 1, 271−275. (55) Sheikh, S.; Blaszykowski, C.; Thompson, M. Talanta 2011, 85, 816−819. (56) Thompson, M.; Blaszykowski, C.; Sheikh, S.; Romaschin, A. Biosens. Bioelectron. 2015, 67, 3−10. (57) Sheikh, S.; Yang, D. Y.; Blaszykowski, C.; Thompson, M. Chem. Commun. 2012, 48, 1305−1307. (58) Sheikh, S.; Blaszykowski, C.; Thompson, M. Surf. Interface Anal. 2013, 45, 1781−1784. (59) Neves, M. A. D.; Blaszykowski, C.; Bokhari, S.; Thompson, M. Biosens. Bioelectron. 2015, 72, 383−392. (60) Blaszykowski, C.; Sheikh, S.; Benvenuto, P.; Thompson, M. Langmuir 2012, 28, 2318−2322. (61) Benvenuto, P.; Neves, M. A. D.; Blaszykowski, C.; Romaschin, A.; Chung, T.; Kim, S. R.; Thompson, M. Langmuir 2015, 31, 5423− 5431. (62) Miller, J. N.; Miller, J. C. Statistics and Chemometrics for Analytical Chemistry, 6th ed.; Pearson Education Limited: Harlow, U.K., 2010; Vol. 46. (63) U.S. Food and Drug Administration. Guidance for Industry: Q2B Validation of Analytical Procedures: Methodology; FDA: Rockville, MD, 1996. (64) Lavagnini, I.; Magno, F. Mass Spectrom. Rev. 2007, 26, 1−18. (65) Fassel, V. A. Anal. Chem. 1976, 48, 2294−2294. (66) Ekins, R.; Edwards, P. Clin. Chem. 1997, 43, 1824−1831. (67) Wang, X.; Ellis, J. S.; Lyle, E.-L.; Sundaram, P.; Thompson, M. Mol. BioSyst. 2006, 2, 184−192. 3106

DOI: 10.1021/acs.analchem.5b04010 Anal. Chem. 2016, 88, 3098−3106