Accessible Published by De Gruyter November 12, 2014

Nucleic acid electrochemical and electromechanical biosensors: a review of techniques and developments

Ruben Rosario and Raj Mutharasan

Abstract

This review comprises the last decade’s development on experimental techniques for electrochemical and electromechanical sensing of nucleic acids, which originate from pathogenic bacteria, parasites, and viruses commonly found in food, water, and medical context. The electrochemical devices that are of primary interest are those that use voltammetry for detecting DNA and RNA-associated electrochemically active molecules at the working electrode. Attograms of nucleic acids have been reported to be detectable with electrochemical sensors in a batch-mode measurement arrangement. The mass-sensing electromechanical devices sense nucleic acids at the femtogram levels in a flow format without a molecular technique for amplifying target strand using polymerase chain reaction. Both underlying physics and methods of various studies are summarized, with discussion on limitations and potentials. We call attention to the need for sensors that not only detect but also confirm detection, as false negatives are not acceptable when one measures pathogenic species.

Introduction

Biosensors present a portable, economical alternative for detection of pathogenic bacteria, parasites, and viruses. Rapid, economical detection is vital when population is threatened by contaminated food, parasite-laden water, and viruses. Waterborne parasites alone account for one fifth of child mortality annually (Baldursson and Karanis 2011). Cryptosporidium and Giardia lamblia are the main culprits in 95% of parasite outbreaks. Cyanobacteria produces hepatotoxins, which are potential carcinogens that contaminate surface waters (Falconer and Humpage 2005). Foodborne illnesses annually cause 128,000 hospitalizations and 3,000 deaths in the USA (Centers for Disease Control and Prevention 2014). Salmonella, Campylobacter, Shigella, and Listeria are the most common bacterial causes of foodborne illnesses monitored by the CDC. People infected with oncogenic variations of human papillomavirus (HPV), hepatitis C (HCV), or hepatitis B (HBV) cannot seek treatment until they are diagnosed, usually with a test processed at a central laboratory facility. In these and other applications, biosensors offer an alternate option for rapid diagnosis and, therefore, a timely corrective response.

Current methods for detection of waterborne parasites are labor intensive and require trained professionals. The example of waterborne parasite detection will illustrate the effort required. The method prescribed by the Environmental Protection Agency (EPA) for detection of waterborne parasites is to filter 10 liters of water, collect the target species by immunomagnetic precipitation, and then enumerate via plating (EPA Method 1623). Methods for detecting foodborne bacteria are even involved: the Food and Drug Administration (FDA) and US Department of Agriculture (USDA), for detecting Listeria, currently require a 52-h enrichment of bacteria from food sample before plating and enumeration (Hitchins and Jinneman 2011). The outbreaks are as costly as they are deadly.

Rapid and sensitive detection of the infectious agents before they are widely dispersed is needed to prevent human and monetary losses (Todd and Notermans 2011). Currently, the FDA and EPA require plating, followed by polymerase chain reaction (PCR) for detection of pathogen. Plating is labor intensive and has a long lag time before obtaining results. PCR, while theoretically sensitive enough for detecting a single strand of DNA in a sample, is not as sensitive when the target is present in a complex matrix such as food samples. Biosensors offer the advantages of reduced sample preparation and extraordinary detection sensitivity. Recently, biosensor researchers have shifted their focus to DNA or RNA-based detection, as they are more definitive and less likely to yield false negatives. Furthermore, nucleotides are more resilient at high temperatures over a range of pH values. DNA sensors can also genotype a virus for tailoring treatment, as in the case of HPV infection (Lereau et al. 2013). Hence, this review is limited to nucleic acid-based detection of pathogens.

Role of biosensors

Many types of biosensors have been reported in the literature, but all of them use a recognition molecule with a specific affinity for the target. Antibodies, aptamers, phages, and single-stranded DNA are common recognition molecules. Nucleic acid-based biosensors that target DNA or RNA offer the advantages of detection at high sensitivity while operating in a complex background and short time to results. Piezoelectric sensors, for example, were able to detect low concentrations of cells without enrichment or sample treatment, and the assay required only 2 h (Rijal and Mutharasan 2013). The sensitivity of these devices can be enhanced further by targeting ribosomal RNA (rRNA), which is present in high copy number in target organisms. Biosensors use a wide variety of transducing mechanisms including optical, electrochemical, thermometric, electromechanical, and magnetic elements.

The most studied devices in the past decade use optical, electrochemical, and electromechanical techniques. Optical sensors using surface plasmon resonance (SPR) and surface-enhanced Raman spectroscopy (SERS) have been an actively investigated platform. Comprehensive analysis of sensing methods of SPR and SERS can be found elsewhere, and a brief description is given here (Driscoll et al. 2013, Sipova and Homola 2013). An SPR sensor contains a metal-coated, highly refractive prism. An incident laser excites electromagnetic modes called surface plasmons at the metal-dielectric (metal-fluid) interface, which decay exponentially into both media. When an analyte binds or is very near (100–600 nm) the metal surface, the resulting refractive index change produces changes in the coupling angle, coupling wavelength, or phase of the incident light. SERS devices typically use gold (or silver) surfaces and nanoparticles for increasing the intensity of Raman spectroscopy peaks. The intensity of the methylene violet peak, for example, can be increased by a factor as much as 1014 with colloidal silver (Kneipp et al. 1997). Studies using SPR or SERS aimed at detecting nucleic acids of foodborne pathogens, waterborne parasites, and viruses, however, are not numerous. Studies are reported to date that examined SPR or SERS-based detection of foodborne bacteria, waterborne parasite, and common viruses and are summarized in Table 1. One notes that only three of the investigations used nonsynthetic nucleotide targets (Zezza et al. 2006, Joung et al. 2008, Zhang et al. 2012). Furthermore, SPR and SERS-based studies reported rarely investigate detection in complex media (Driscoll et al. 2013, Sipova and Homola 2013). Hence, we will limit the review to electrochemical and electromechanical biosensors in this paper, as they have made significant progress for practical applications over the last decade. The operating principles, detection limits, and time to results (TTR) are discussed for these methods. Prior to examining biosensors, we will briefly review traditional culture and molecular techniques, as they are the current method of choice.

Table 1

Optical detection of pathogenic DNA/RNA.

Optical sensorTargetSample matrixAnalysis time (h)Limit of detectionReferences
SPRSalmonella ampliconsPCR mix4.5102 CFU/mlZhang et al. 2012
SERSWest Nile virus oligonucleotidesBuffer3Not reportedHarpster et al. 2009
SPRSynthetic E. coli rRNABuffer358 pg/mlJoung et al. 2008
SPRS. aureus rRNA from cell lysateCell lysate37×105 CFU/mlJoung et al. 2008
SERSSynthetic HBV DNABuffer1–250 aMLi et al. 2013
SPRSynthetic S. aureus, E. coli, B.abortus DNABuffer0.25100 pMPiliarik et al. 2009
SPRF. culmorum ampliconsPCR mix1050 fgZezza et al. 2006

Traditional techniques

Food samples are generally monitored for pathogens using biological analytic techniques such as enzyme-linked immunosorbent assay (ELISA) and plating following enrichment in specialized and selective media. Chromogens present in the specialized media indicate types of metabolic activity specific to the target organism. The enrichment phases for the plating methods last between 48 and 52 h depending on the bacteria, according to the FDA’s Biological Analytical Manual (Hitchins and Jinneman 2011). The FDA, for example, recommends that Escherichia coli be identified in milk samples by incubation in violet red bile agar for 18–24 h followed by a check for purple-red colonies of a minimum diameter. ELISA was one of the earliest available commercial methods for rapid identification of biological entities. Antibodies targeting the pathogen are typically immobilized on a solid substrate and then exposed to the target organism. After a rinsing step, horseradish peroxidase-conjugated antibody is exposed to the well, and it binds to the antigen, creating an antibody-antigen-antibody sandwich. A substrate for the enzyme is added in a subsequent step to induce a color change that is readily visible or detected via fluorescence. ELISA’s detection limit is on the order of 103 cfu/ml sample. Thus, enrichment is required for food samples, which have lower pathogen concentration (Gehring et al. 2004). Waterborne parasites are not amenable to plating. The EPA’s “Method 1623” is a time-intensive process requiring positive control verification with spiked water samples to reduce or avoid false negatives. The above-described traditional methods require multiple laboratory and manual steps where error and contamination are a significant concern.

PCR is a very effective method and is, in fact, necessary for diagnosis and monitoring of antiviral treatment efficacy for hepatitis infections (Ghany et al. 2009). Quantitative PCR (qPCR) has been found to be a highly sensitive technique, but polymerase enzymes are inhibited by serum proteins and proteins found in food samples (Doyle et al. 2007, Kermekchiev et al. 2009). The inhibitors reduce amplification efficiency, which not only reduces sensitivity, but also causes false negatives (Doyle et al. 2007). Extensive multistep sample preparation is required for food samples prior to qPCR. For example, E. coli and Salmonella were detected in milk at concentrations as low as 5–10 cfu/ml of milk sample (Soejima et al. 2012). Since the PCR sample volume was 50 μl, the actual concentration is, in fact, much higher in the sampled PCR reaction mixture. The researchers achieved the high sensitivity by incubating the initial sample to enrich the culture while simultaneously degrading the PCR inhibitors with Proteinase K. After centrifuging the sample into a pellet, only a small portion could be used for the assay lest too much PCR inhibitors were introduced to the PCR master mix. Furthermore, the sample was concentrated to improve performance. PCR experimental conditions such as cycle temperatures and primer design are engineered using the kinetics and thermodynamics of nucleic acids in solution. Duplex formation within biosensor apparatus, however, occurs on a surface.

Considerations for probe immobilization and hybridization strategies for biosensors

DNA hybridization on a surface is very distinct from hybridization in solution in terms of kinetics and thermodynamics. (Peterson et al. 2001, Gong and Levicky 2008, Springer et al. 2010). The general strategy for detection of target nucleic acids with biosensors is nearly always the same. A single-stranded DNA (ssDNA) probe complementary to the target nucleic acid chains of interest (ssDNA or ssRNA) is immobilized on a surface. Then, the sample containing the target nucleic acid is introduced, and the signal for hybridization is transduced by optical, electrochemical, electromechanical, or other mechanisms.

Hybridization to a surface probe is found to be slower than in solution. The following example is illustrative. At 20°C in Tris buffer with 0.5 m NaCl, two complementary 25-mer ssDNA chains (1 μm each) hybridize almost completely (99% efficiency) within 41 s in solution. Under identical conditions, when ssDNA was immobilized onto the gold-coated surface of a prism (10 mm×17.2 mm), only 22.6% of the target DNA hybridized in that time (Gao et al. 2006). Hybridization, in fact, equilibrated at ∼25% under those conditions. A pseudo-Langmuir model was assumed for the surface reaction. The legitimacy of this assumption will be discussed later. There are several factors that cause the observed differences in surface hybridization kinetics and thermodynamics.

The most obvious cause of reduced kinetics and reaction extent is the larger magnitude of electrostatic repulsion between the phosphate backbones of the probe and target: the nucleic acid charge concentrations are much more focused at the functionalized surface than in solution. A careful researcher can screen these electrostatic interactions by using a high cation concentration in the hybridization buffer ([Na+]=1 m) (Springer et al. 2010). The reader should note that it has been reported that Mg2+ cations are exceptional in shielding electrostatic repulsion compared to Na+ (Springer et al. 2010). Most probes, however, do not hybridize to targets even at high ionic strength if probe density, SP (measured in chains cm-2), is not carefully controlled. Several investigators have reported that hybridization efficiency (max percentage of probes hybridized) is capped at 25–40% if probe density exceeds 2×1012 chains cm-2 (Peterson et al. 2001, Hagan and Chakraborty 2004). Figure 1 illustrates the effect of probe density on hybridization. The result reported is for hybridization of an immobilized thiolated 24-mer ssDNA probe with its complementary strand as measured by SPR. The measured signal for the hybridization is given as a function of time for several probe densities at 1 m NaCl in TE buffer. Biosensor experimentalists can limit deposition of probes by using low salt concentrations in their immobilization buffer and limiting probe incubation times. The surface concentration of charge-balancing cations is another factor that limits hybridization speed and efficiency.

Figure 1 Biotinylated 24-mer ssDNA capture probes were immobilized onto a surface plasma resonance (SPR) sensor surface at varying surface densities. This immobilized layer was then exposed to the complementary target strands with a concentration of 1 μm in 1 m NaCl with TE buffer. The hybridization rate and efficiency is clearly higher at lower probe densities. This image was reproduced from Peterson et al. (2001) with permission from Oxford University Press.

Figure 1

Biotinylated 24-mer ssDNA capture probes were immobilized onto a surface plasma resonance (SPR) sensor surface at varying surface densities. This immobilized layer was then exposed to the complementary target strands with a concentration of 1 μm in 1 m NaCl with TE buffer. The hybridization rate and efficiency is clearly higher at lower probe densities. This image was reproduced from Peterson et al. (2001) with permission from Oxford University Press.

Local cation concentration gradients can present a thermodynamic barrier to target DNA hybridization. Where SP is in the range of 1012–1013 chains/cm2, and typical 20-mer probes are roughly 10 nm long, the local concentration of phosphates will be ∼200 mm. Cations from solution balance the phosphate charges, and if this local cation concentration (Cc,s) is much more than the bulk concentration (Cc,b), a strong thermodynamic barrier inhibits binding of target ssDNA or ssRNA. More cations would move to surface to balance the negative charges, acting against the concentration gradient (Gong and Levicky 2008). Hence, we define a parameter, Π, given in equation (1) for classifying hybridizing conditions into three regimes: pseudo-Langmuir, suppressed hybridization, and nonhybridizing.

(1)Π=Cc,sCc,b (1)
(2)Cc,s=SPNhNA (2)

The parameter, Π, is the ratio of Cc,s and Cc,b, and Cc,s is defined by equation (2). The surface cation concentration is a function of probe surface density, SP; the number of phosphates per chain, N; the height of the probe layer, h; and Avogadro’s number, NA. Figure 2 gives the boundaries of the hybridization regimes of 18-mer DNA as a function of Π. When Cc,s exceeds Cc,b (Π>1), the concentration gradient allows for very little hybridization. Figure 3A is illustrative of pseudo-Langmuir, suppressed hybridization, and nonhybridization regimes observed at various ionic strengths and probe density. Figure 3B gives the extent of hybridization for the same set of experimental conditions. Pseudo-Langmuir behavior is observed when the extent of hybridization reaches a constant value, usually at lower SP values; for example, see response at CB=0.33 m and 1 m in Figure 3B. The suppressed hybridization regime is the intermediate regime where nucleic acids hybridize to a limited extent, but there are strong interactions between the probe sites. The Π formula provides a general framework for thermodynamics of hybridization at a surface that is useful in biosensor experiments. Let us now examine the kinetics of hybridization.

Figure 2 The extent of hybridization, x, is given as a function of Π. The target/probe strands consisted of 20-mer ssDNA. The experiments were performed in sodium phosphate buffer. The target strands were labeled with ferrocene, and cyclic voltammetry was used to measure the quantity of hybridized target strands. The most striking transition is noted where Π=2.2, which is consistent with the hypothesized transition zone of Π≈1. When Π>2.2, the osmotic pressure from the high surface concentration of cations relative to the bulk inhibits hybridization. This figure was reproduced with permission from Gong and Levicky (2008).

Figure 2

The extent of hybridization, x, is given as a function of Π. The target/probe strands consisted of 20-mer ssDNA. The experiments were performed in sodium phosphate buffer. The target strands were labeled with ferrocene, and cyclic voltammetry was used to measure the quantity of hybridized target strands. The most striking transition is noted where Π=2.2, which is consistent with the hypothesized transition zone of Π≈1. When Π>2.2, the osmotic pressure from the high surface concentration of cations relative to the bulk inhibits hybridization. This figure was reproduced with permission from Gong and Levicky (2008).

Figure 3 (A) This figure divides the data points from the hybridization experiments into three regimes: nonhybridizing, suppressed hybridization, and pseudo-Langmuir hybridization. These experiments were performed with varying probe densities and solution ionic strength. Probe density, SP, was measured in numbers of chains per cm2, and ionic strength was measured in terms of concentration of phosphate, CP. The concentration of Na+ in solution is approximately 1.67 CB. (B) This plot gives the extent of hybridization, x, measured using the same method as Figure 2. The pseudo-Langmuir hybridization behavior is apparent for the CB=0.33 m and 1 m phosphate trends where the x reaches a plateau for varying SP. The trends for C=0.012 m and 0.012 m are clearly in the nonhybridizing regimes except for the lowest values of SP. This figure was reproduced with permission from Gong and Levicky (2008).

Figure 3

(A) This figure divides the data points from the hybridization experiments into three regimes: nonhybridizing, suppressed hybridization, and pseudo-Langmuir hybridization. These experiments were performed with varying probe densities and solution ionic strength. Probe density, SP, was measured in numbers of chains per cm2, and ionic strength was measured in terms of concentration of phosphate, CP. The concentration of Na+ in solution is approximately 1.67 CB. (B) This plot gives the extent of hybridization, x, measured using the same method as Figure 2. The pseudo-Langmuir hybridization behavior is apparent for the CB=0.33 m and 1 m phosphate trends where the x reaches a plateau for varying SP. The trends for C=0.012 m and 0.012 m are clearly in the nonhybridizing regimes except for the lowest values of SP. This figure was reproduced with permission from Gong and Levicky (2008).

The hybridization of DNA strands in solution follow second-order kinetics. When there is no secondary structure to the single stranded DNA (e.g., hairpin structure), the rate constant, kon, will be ∼105m-1 s-1 at 20–25°C with [Na+]=0.5–1 m (Hagan and Chakraborty 2004, Sekar et al. 2005, Gao et al. 2006, Rauzan et al. 2013). Complementary strands hybridize virtually completely within 3 s at micromolar concentrations. The second-order rate kinetics will mean that the time for the reaction will be longer at lower concentrations. If an oligonucleotide contains a hairpin, kon can be reduced by more than an order of magnitude. The rate constant for DNA hybridization on a surface exhibits pseudo-Langmuir behavior, and is given by:

(3)dCPTdt=kaCT(σ-CPT)-kdCPT (3)
(4)ks=kaCT+kd (4)
(5)CPT(t)=CPTeq(1-exp(-kst)) (5)

where CPT, CT, σ, ka, kd, and ks are the surface hybridized target concentration, bulk target concentration, concentration of probe in the immobilized layer, association rate constant, disassociation rate constant, and effective rate constant. Equation (3) can be integrated to produce equation (5) when CT is in great excess. The equilibrium surface concentration of hybridized probes is designated as CPTeq. The effective rate constant is typically in the 0.5–5×104m-1 s-1 range for experiments with ∼20-mer oligonucleotides, so they are usually at least an order of magnitude less than the second-order rate constants of solution (Hagan and Chakraborty 2004, Gao et al. 2006). The local concentration of surface probe is much higher than it would otherwise be in the homogenous reaction, and thus, the hybridization usually reaches equilibrium in 10–15 min (Dell’Atti et al. 2007, Uludag et al. 2008, Liao and Ho 2009, Sun et al. 2009, Li et al. 2013). This time is true when CT in the nanomolar to micromolar concentration range.

Electrochemical biosensors

Sensing principle

Electrochemical biosensors use two general methods for transducing nucleic acid hybridizing events. This class of biosensors can transduce adsorption of insulating molecules as an increase in electrical impedance. The second method is where the electrochemical sensor transduces oxidation (or reduction) of electroactive species as electrical current. Unlabeled nucleic acids, however, are not sufficiently electrochemically active for sensitive detection. Two methods have been investigated to overcome this difficulty, namely, two-step sandwich binding and the use of intercalating redox indicators. The first approach uses a labeled reporter probe that binds to the available sites of the target nucleotide, which has hybridized to its complementary immobilized capture probe. This reporter probe forms the top layer sandwich. The bottom layer is made of the capture probe immobilized onto the electrode surface. The reporter strand is covalently labeled with an electrochemically active species such as a metal nanoparticle or an enzyme that produces a redox couple. The second approach is to use an electrochemically active molecule that intercalates with the hybridized capture and target nucleotides. The agents of both methods, whether they are labeled reporter probes or intercalating molecules, will be called redox indicators in this paper for the sake of simplicity.

The biosensor detects redox indicator concentration and changes in mass transport of the redox indicator to the electrode. For example, mass transport rate decreases as target species bind to the recognition molecule on the electrode. Depending on washing and reacting steps, the redox indicator may be in solution, immobilized onto the electrode, or intercalated within hybridized DNA strands. Electrochemical biosensors monitor faradaic current from the redox indicators as the potential of the working electrode is manipulated (Jan 2012). The faradaic current increases with concentration of the redox indicator. This general technique is called voltammetry.

Experimental design

Electrochemical biosensors typically use a three-electrode cell (Figure 4). A standard cell is an anaerobic glass container that is maintained oxygen-free by nitrogen so oxidation occurs exclusively at the electrode. The cell is thermostated as the potential of the reference electrode is a function of temperature. Inert electrolyte in solution within the cell ensures the electric field induces minimal migration of electroactive ions. Sensing of nucleotide hybridization is designed to occur at the working electrode, which is typically a planar piece of metal or a screen-printed electrode. Two general sensing methods are used, namely, potentiostatic and galvanostatic control. In the former, the potential of the working electrode is maintained at a constant value while the current is monitored. In the latter, we maintain the current at a constant value while the potential is monitored. An auxiliary electrode such as a platinum wire or mesh completes the circuit. The potential of the working electrode is measured with respect to a reference electrode. The Ag∣AgCl reference is a common choice for sensing in aqueous systems.

Figure 4 Electrochemical cell with three-electrode arrangement. A planar electrode is used as the working electrode, and a metal mesh serves as the auxiliary electrode. Inert gas is sparged to prevent oxidation by oxygen.

Figure 4

Electrochemical cell with three-electrode arrangement. A planar electrode is used as the working electrode, and a metal mesh serves as the auxiliary electrode. Inert gas is sparged to prevent oxidation by oxygen.

Pulse voltammetry

Pulse voltammetry is the practice of measuring current while making pulse changes in the applied working electrode potential. The most prevalent methods are square wave voltammetry (SWV) and differential pulse voltammetry (DPV). Both incrementally increase or decrease the applied working electrode potential in a step manner. The potential waveforms of the applied potential for differential pulse and square wave voltammetry are given in Figure S-1 and S-3 in the supplementary materials. Both techniques have been shown to be capable of sensing target nucleotides at the femtomole level (Pinijsuwan et al. 2008, Liao and Ho 2009). Here, we describe the techniques briefly, but more details can be found elsewhere (Osteryoung and Osteryoung 1985, Mithran and Werasak 2012). In square wave and differential pulse voltammetry, the current is sampled after the step changes in potential are imposed with sufficient delay, so that capacitive current decays to zero. Then, the difference between the currents sampled before and after each step change, ΔI, is plotted as a function of applied potential. The current signal increases as the potential approaches the E1/2, a close approximation of redox potential of the redox indicator, and peaks with a magnitude proportional to the potential of the indicator concentration. A typical current-potential profile for differential pulse voltammetry is shown Figure S-2 (supplementary materials). The experimental details for DPV is given in greater detail in Lane and Hubbard (1976).

The reporter probe labels vary; the nucleotides may be conjugated to horseradish peroxidase, liposomes encapsulating metal ions, or other species that produce an electrochemically active species (Kerman et al. 2004). The principle of differential pulse voltammetry, however, is always the same. The current is proportional to the rate of oxidation, which is diffusion limited; therefore, the concentration of label, for example, at the electrode surface is proportional to the current. Other forms of voltammetry such as cyclic voltammetry may not be as sensitive in detecting redox species, but are useful for sensor characterization (Goyal et al. 2007).

Cyclic voltammetry

In cyclic voltammetry, the applied potential is manipulated according to the saw tooth wave form (Figure S3 and S4), and the resulting current is measured. The measurements are done in solution with a redox couple. A typical voltammogram for cyclic voltammetry is shown in the supplementary materials (Figure S-5). All sampled current is a combination of faradaic and substantial capacitive current as the potential is swept continuously. There are two current peaks corresponding to oxidation and reduction of the redox couple. The distance between the peak currents and their magnitudes are useful for electrode characterization (Kissinger and Heineman 1983). The separation of the peak current potentials, ΔEp, is approximately 57.0/n mV for reversible, mass transport-limited reactions; n is the number of electrons involved in electron transfer (Brett 1993). As the electrode undergoes modification with insulating molecules, access to the electrode surface diminishes, and the peak currents, as a consequence, decrease. The value of ΔEp also increases if electron transfer is reaction limited. The voltammogram given in Figure S-5 illustrates these effects as insulating alkylthiols adsorb onto a gold electrode. This form of voltammetry confirms that various modifications at the electrode take place with or without labeled species. The technique, however, is not as sensitive as pulse voltammetry as there is always a capacitive current contribution to the peak current signals.

Impedimetric genosensing

Electrochemical impedance spectroscopy is another approach for detecting nucleic acid hybridization events; an excellent description of the technique was provided by Bard and Faulkner (Bard and Faulkner 2001). During the measurement, the working electrode is maintained at a constant DC potential to maintain a constant ratio of the redox couple [e.g., (FeCN6)3-/4-] and a small AC potential is applied at a various frequencies. The current over-potential relations are linear for the small perturbations in AC current so that the current only propagates at the same frequency as the applied potential. A potentiostat measures the equivalent resistance and capacitance of the cell at each frequency for determining the resistance to charge transfer, labeled as Rct, at the electrode interface. The Rct parameter increases during sensing as the reaction rate slows. Resistance to charge transfer is a critical parameter and functions as a sensing signal. Modification of the electrode alters the mass transport characteristics of the cell. For example, adsorption of a species that inhibits transport of the redox couple will increase charge transfer resistance. Therefore, Rct is used as a transduction signal for sensing.

Chronoamperometry

Chronoamperometry is the practice of measuring current as a function of time with step changes in the potential applied to the working electrode. In double step chronoamperometry, the potential assumes the wave form featured in the lower panel of Figure S-6 (see supplementary section). The current is measured immediately after the potential step as the redox indicator is oxidized or reduced. The bulk concentration of the electroactive species can be calculated if the reaction is diffusion controlled and reversible (i.e., Nernstian). Further details may be found in Bard and Faulkner (2001).

Applications

While there have been a number of investigations using electrochemical biosensors for detecting foodborne and waterborne pathogens in the last decade, we highlight the important ones in Table 2. The cases that are more relevant to “real-world” detection are the ones that test with complex samples such as tissue, cell culture (without the aid of PCR), or food. It should be noted that many studies utilize synthetic nucleotides (Aguilar and Fritsch 2003, Heidenreich et al. 2010, Sun et al. 2012). Some of the more striking innovations in DNA detection have occurred in the field of virus detection, especially with respect to probe design. Researchers who examined sensing of viruses investigate various capture probes that are tagged with redox species (Lereau et al. 2013, Grabowska et al. 2014a,b). In one approach, polythiolated DNA probe tagged with ferrocene was evaluated (Lereau et al. 2013). When the capture probe was not hybridized, the ferrocene molecule on the flexible ssDNA is more mobile and migrates to the electrode surface. When the target DNA hybridizes with the probe, a stiff double helix is formed, which significantly decreases the DPV signal due to relatively immobile ferrocene. Use of peptide nucleic acids (PNA) as capture probes has also been investigated, as it provides for neutral DNA analogs, which lowers electrostatic repulsion, and form “triplex” with dsDNA (Ahour et al. 2012, 2013). We examine three examples to illustrate the techniques that are extremely sensitive, conducted without amplification, or use novel probes.

Table 2

Electrochemical detection of pathogenic DNA/RNA.

Detection techniqueTargetSample matrixAnalysis time (h)Limit of detectionReferences
DPVE. coli ss-DNA50 μl sample of magnetically separated ssDNA75 cfu/mlAnderson et al. 2013
SWVE. coli ssDNA1 μl buffer∼10.75 amol nucleotide in 1 μl sampleLiao and Ho 2009
Potentiometric sensorE. coli rRNA4 μl buffer110 cfu (0.2 amol 16S rRNA) in 4 μl sampleWu et al. 2009
Cyclic voltammetryC. parvum mRNA500 μl lysate162 μg mRNA (37 oocysts)/1 mlAguilar and Fritsch 2003
DPVL. monocytogenes ssDNA5 μl PCR mix∼10.145 fmol ssDNA in 5 μl sampleSun et al. 2012
ChronoamperometryE. coli“Meat juice” lysate∼71 cfu/mlHeidenreich et al. 2010
Impedance spectroscopySalmonella choleraesuis ampliconPCR mix∼31 nMaBerdat et al. 2008
DPVStaphylococcus saprophyticus DNACell lysate0.51 cfu/μlLam et al. 2012
Cyclic voltammetryB. anthracis ampliconPCR mix∼ 410 pg/μlPal and Alocilja 2010
DPVPlasmid with HCV viral core DNA6 μl buffer3–49.5 pg/μlAhour et al. 2013
DPVHCV ds-DNA6 μl buffer39.63 pMAhour et al. 2012
SWVH5N1BufferNot reported21 fMGrabowska et al. 2014a
DPVSynthetic HCV DNABuffer with complex background110 fMLereau et al. 2013
SWVAmplified HPV DNA from tissuePCR mix2 (+PCR)∼100 pMCivit et al. 2012
DPVHBV ssDNA10 μl buffer30.300 pMZheng et al. 2014

aLOD established with synthetic oligonucleotide.

Square wave voltammetry was used in Liao and Ho (2009) for detecting ssDNA corresponding to the gene rfbE in E. coli O157:H7. Thiolated ssDNA capture probe was coupled to the surface of a gold-coated screen-printed electrode (SPE). A competitive binding assay scheme outlined in Figure 5A was used. The target ssDNA and a competing reporter probe consisting of complementary ssDNA conjugated to a Ru3+-containing liposome were incubated with the electrode simultaneously. The electrode was rinsed, dried, and transferred to Tris-HCl buffer for electrochemical measurement.

Figure 5 (A) The above scheme presents a competitive binding assay where DNA-tagged liposomes compete with the DNA target for available capture probe sites. (B) The peak currents sampled from square wave voltammetry is given for various concentrations of the target ssDNA. The current diminishes with increasing concentrations of ssDNA. There is a roughly log-linear relation between the two variables for concentrations between 1 and 106 fmol. This figure was reproduced with permission from Liao and Ho (2009). Copyright © 2009 American Chemical Society.

Figure 5

(A) The above scheme presents a competitive binding assay where DNA-tagged liposomes compete with the DNA target for available capture probe sites. (B) The peak currents sampled from square wave voltammetry is given for various concentrations of the target ssDNA. The current diminishes with increasing concentrations of ssDNA. There is a roughly log-linear relation between the two variables for concentrations between 1 and 106 fmol. This figure was reproduced with permission from Liao and Ho (2009). Copyright © 2009 American Chemical Society.

The current measured in square wave voltammetry assay increases as the potential of the SPE nears the redox potential of Ru3+ at which point it reaches its maximum. The magnitude of the peak current is proportional to the number of reporter probes. Hence, the current will decrease in the presence of increased unlabeled ssDNA. There is a log-linear relationship between peak current and concentration of the unlabeled ssDNA, as shown in Figure 5B. The limit of detection can be obtained by subtracting three times the standard deviation from the average current for the control response in the absence of rfbE ssDNA. In the study reported by Liao et al., a limit of detection of 0.75 amol per assay was achieved, which translates to 5 μl of 0.15 pM sample. Such small volume expedites the hybridization assay, but may prove problematic for analysis of real food or environmental samples that are often in the milliliter and liter ranges, respectively. The largest sample volume of nucleotides of the cited investigations, in fact, was used in Aguilar and Fritsch (2003) with a sample of volume of 500 μl.

In the study reported by Aguilar and Fritsch, an electrochemical approach targeted 121-mer mRNA of the hsp70 heat shock protein in Cryptosporidium parvum (Aguilar and Fritsch 2003). A sample of 2.6×106 oocysts/ml was heat shocked for 10 min to induce transcription of hsp70. A capture DNA probe was immobilized onto an aminated Au/SiO2 wafer. Then, 500 μl of a 50-μg/ml solution of heat-shocked oocysts was incubated with the functionalized wafer for 1 h for hybridization to be completed. Subsequently, incubation with a reporter probe consisting of 42-base ssDNA conjugated to alkaline phosphatase was used. The modified wafer was rinsed and transferred to a solution containing the substrate for the enzyme, p-aminophenyl phosphate (PAPP). Over a 12-h period, alkaline phosphatase generated the electroactive species, p-aminophenyl, which was measured by cyclic voltammetry. The authors established that there was very limited cross-reactivity with several common pathogens such as Cr. lamblia, Listeria monocytogenes, Campylobacter lari, E. coli O157:H7, and Salmonella typhi. A calibration curve was established using synthetic hsp RNA as a target. The peak currents varied linearly in the 5- to 50-μg/ml concentration range, and the limit of detection was calculated as 2 μg/ml (146 nm), which is modest compared to the work of Liao et al. who obtained a detection limit of 1 amol (Liao and Ho 2009).

The limit of detection for Cryptosporidium is poor in comparison to the femtomol and attomol values in the E. coli experiments. The 2-μg/ml limit of detection was established using a synthetic target as in all the work summarized in Table 2, with the exception of (Wu et al. 2009) where they used bacterial lysate with no PCR. The limits of detection given in Table 2 would be more relevant to food safety and environmental monitoring if they were extracted from experiments that used actual samples of the pathogens as in Wu et al. (2009). When authors reported in Table 2 the use of food samples or cell cultures, they use them only to establish selectivity. The hsp70 detection experiment is notable as it targeted a nucleic acid sequence as opposed to an immuno-based sensor. Aguilar and Fritsch used oocysts concentration on the order of 106 oocysts/ml. Therefore, PCR was necessary for the detection of the parasite in a real water sample, especially because the infective dose is ∼30 oocysts.

The study by Lereau et al. must be noted for its novel probe design to achieve high sensitivity (Lereau et al. 2013). The researchers utilized a novel polythiolated 15-mer ssDNA probe with three ferrocene residue at the end of the strand to achieve a 10-fM limit detection for a 105-mer oligonucleotide corresponding to the 3a hepatitis C virus genotype (Lereau et al. 2013). The polythiol modification permitted greater signal stability as the probe was less likely to degrade and detach from the gold sensor surface. When the 105-mer ssDNA synthetic target hybridized to the probe, the ferrocene molecules moved away from the electrode surface, decreasing the current measured by DPV. A scheme depicting the technique is given in Figure 6A. The 10-fM limit of detection was achieved with 50 ng/ml sperm DNA and 1 pM non-complementary 105-mer ssDNA in the background. The DPV voltammogram for this measurement is given in Figure 6B. This electrochemical method would have proven to be more robust for clinical genotyping of HCV if they used HCV amplicons instead of synthetic nucleotides.

Figure 6 (A) Scheme for probe design where 15-mer polythiolated ssDNA probe with three ferrocene molecules grafted near the end of the chain. When the 105-mer complementary nucleotide hybridizes, the ferrocene molecules move from the surface, reducing the current measured by differential pulse voltammetry (DPV). (B) The peak current measured by DPV decreases over the course of an hour after t=0 when the complementary 105-mer oligonucleotide is added to the hybridization buffer (sodium phosphate 20 mm, NaClO4 250 mm). These images were reproduced with permission from Lereau et al. (2013). Copyright © 2013 American Chemical Society.

Figure 6

(A) Scheme for probe design where 15-mer polythiolated ssDNA probe with three ferrocene molecules grafted near the end of the chain. When the 105-mer complementary nucleotide hybridizes, the ferrocene molecules move from the surface, reducing the current measured by differential pulse voltammetry (DPV). (B) The peak current measured by DPV decreases over the course of an hour after t=0 when the complementary 105-mer oligonucleotide is added to the hybridization buffer (sodium phosphate 20 mm, NaClO4 250 mm). These images were reproduced with permission from Lereau et al. (2013). Copyright © 2013 American Chemical Society.

Electrochemical sensors, in combination with PCR techniques, are a powerful approach to detecting nucleic acid sequences in batch assays. The cyclic voltammetry and impedimetric assays also provide characterization of adsorbed monolayers. An alternative sensing approach is to use electromechanical sensors to detect nucleic acids with flow-through sample analysis, where the sample volume is larger than the 1- to 50-μl volumes reported for the electrochemical sensors.

Electromechanical biosensors

Sensing principle

Quartz crystal microbalances (QCM) and resonating cantilevers are the most prevalent electromechanical devices in nucleic acid-based biosensing. Both sensor types are composed of piezoelectric materials; an applied AC potential generates a cyclic change in the shape of the material. When the excited potential frequency coincides with inherent mechanical resonant frequency, the piezoelectric material undergoes larger than normal change in its electrical properties, specifically the real part of its electrical impedance. Such a change is easily measured by monitoring the resulting current in the circuit. The QCM and resonating cantilevers operate at resonant frequencies particular to their geometry, and they transduce added mass at their surface as a decrease in resonant frequency. The measured resonant frequency decreases as a consequence of target nucleotides hybridizing to the complementary capture probes immobilized on the surface. Non-targets do not hybridize with the capture probe. Preamplification of the target can be done by PCR, and postsensing amplification is done by further secondary hybridization. Despite the similarities in their signal transduction and piezoelectric nature, the differences in the construction of the two devices are quite considerable and are further discussed.

A quartz crystal microbalance is composed of a thin disk cut from a single quartz crystal. Gold electrodes on both sides impose an oscillating electric field normal to the surface of the crystal wafer. The direction of oscillation depends on the orientation of the crystal lattice in the electric field (Deakin and Buttry 1989). The AT-cut, typical of QCM, produces a wafer with displacements orthogonal to the electric field and is often described as the shear mode or parallel to the surface. The wafer’s resonant frequency is a function of its thickness with values between 2 and 20 MHz. The resonant frequency peak of the crystal is so sharp that frequency counters can measure resonant frequency differences as small as 0.1 Hz.

A piezoelectric cantilever, on the other hand, is made of ceramic piezoelectric material coated with electrodes on either side. Common piezoelectric materials include lead zirconate titanate, zinc oxide, barium titanate, and aluminum nitride (Dubois and Muralt 1999, 2001, Wada et al. 1999, Shrout and Zhang 2007). A cantilever, as a three-dimension, distributed mass system has four distinct resonant mode shapes, which include the lateral, extension, flexural (bending), and torsional modes. These mode shapes are summarized in Figure 7.

Figure 7 The variety of resonant mode shapes is pictured here. Some modes are a combination of the above mode shapes. This image was reprinted from Song and Bhushan (2006) with permission from Elsevier.

Figure 7

The variety of resonant mode shapes is pictured here. Some modes are a combination of the above mode shapes. This image was reprinted from Song and Bhushan (2006) with permission from Elsevier.

Most cantilever-based sensing in literature uses the flexural resonant modes, even though other modes do respond to mass changes (Ilic et al. 2000, 2004, Waggoner and Craighead 2007). An impedance analyzer sweeps across a range of excitation frequencies in the region of a resonant mode to record frequency shifts. The principle of impedance-based resonance detection relies on the fact that the accumulation of charge is due to the deformation of the piezoelectric material’s crystal lattice. Furthermore, the magnitude of cantilever deflection increases as the exciting frequency of the AC potential nears the device’s resonant frequency. The magnitude of charge accumulation and its rate of change with respect to time subsequently increase as the cantilever approaches resonance. The admittance (Y), or inverse of the impedance (Z), can be utilized for detecting a resonant mode from this impedance response. The relationship between admittance and charge accumulation is given in Lee et al. (1999):

(6)Y=IV=IQ+ICV=1Vdqpdt+(d(CV)Vdt+1R) (6)

The applied voltage and total current output are represented by V and I. The total measured current has a piezoelectric (Ip) and capacitive (IC) component. The Ip component is the time derivative of the accumulated charge (qp), which is defined in as:

(7)qp=-d31EZpw0ld2v(x)dx2dx (7)

The parameters d31 and E are the piezoelectric constant and Young’s modulus, respectively. The distance between the plane of zero strain and the neutral plane of the piezoelectric layer is given as Zp. The cantilever width, displacement at position x and cantilever length are represented by w, v(x), and l. While an impedance analyzer is used to sweep a range of excitation frequencies, the described relations between admittance and resonant frequency are used to detect resonance, as is illustrated in Figure 8. The piezoelectric device current is at its maximum at resonance. The phase angle of the output current also peaks at resonance as the piezoelectric material deviates from purely capacitive behavior.

Figure 8 The capacitive and piezoelectric components of current output are plotted as a function of excitation frequency. The corresponding admittance values are also plotted. This figure was reprinted from Lee et al. (1999) with permission from Elsevier.

Figure 8

The capacitive and piezoelectric components of current output are plotted as a function of excitation frequency. The corresponding admittance values are also plotted. This figure was reprinted from Lee et al. (1999) with permission from Elsevier.

Experimental design

The QCM apparatus is composed of an oscillator circuit connected to an AT-cut quartz crystal. A frequency counter monitors the frequency, while the sample fluid is exposed to one side of the Au-coated crystal. QCM and dynamic cantilevers are capable of detecting nucleotides in a flow loop as hybridization occurs, unlike the batch electrochemical method discussed in the previous sections. Such a flow loop arrangement for QCM is shown in Figure 9A. A similar arrangement is used for piezoelectric cantilevers with the exception that an impedance analyzer is used for resonant frequency measurement as the quality factor (Q) of the resonant frequency peaks is modest (Q∼30–60) and is not as high as that of QCM (Q∼103–104).

Figure 9 (A) A QCM apparatus for DNA sensing is shown. The closed flow loop contains a heater to denature PCR products. This figure was reprinted from Wu et al. (2007) with permission from Elsevier. (B) This sensogram is typical for QCM detection of DNA. (A) After the probe is immobilized, the baseline is recorded. (B) The target ssDNA is introduced in hybridization buffer and circulated for 10–20 min. (C) The sensor is rinsed with hybridization buffer. The noise is a result of pressure fluctuation. (D) The stable signal for the hybridized DNA is recorded. (E) The DNA is denatured with 100 mm HCl. (F) The frequency returns the prehybridization value. Reprinted from Dell’Atti et al. (2007) with permission from Elsevier.

Figure 9

(A) A QCM apparatus for DNA sensing is shown. The closed flow loop contains a heater to denature PCR products. This figure was reprinted from Wu et al. (2007) with permission from Elsevier. (B) This sensogram is typical for QCM detection of DNA. (A) After the probe is immobilized, the baseline is recorded. (B) The target ssDNA is introduced in hybridization buffer and circulated for 10–20 min. (C) The sensor is rinsed with hybridization buffer. The noise is a result of pressure fluctuation. (D) The stable signal for the hybridized DNA is recorded. (E) The DNA is denatured with 100 mm HCl. (F) The frequency returns the prehybridization value. Reprinted from Dell’Atti et al. (2007) with permission from Elsevier.

Quartz crystal microbalance

As discussed in the previous section, the quartz crystal of the QCM is a thin AT-cut disk with gold electrodes on both sides designed to promote shear-mode deformation (Martin et al. 1991). The resonant frequency of the disk, due to the nature of its vibration, depends on the thickness (typically 300 μm for a 5-MHz crystal). A smaller thickness leads to higher resonant frequencies and greater sensitivity, but its mechanical fragility may become limiting. The displacement profiles along the thickness of the crystal are given in Figure S-7. The resonant frequency change of a QCM in the gas phase is given by the Sauerbrey equation (Lucklum and Hauptmann 2000):

(8)Δff=-2f02ZcqΔmA (8)

The change in resonant frequency, Δff is a function of the initial resonant frequency, f0, added mass per area, Δm/A, and the characteristic impedance of the crystal, Zcq. The characteristic impedance is defined by the mechanical and electrical properties of the crystal. The QCM resonant frequency response to a Newtonian fluid is described by:

(9)Δfl=-2f02Zcqρlηl4πf0 (9)

That is, the shift down in resonant frequency is directly proportional to the square root of the density and viscosity of the liquid. The inertial effects of an adsorbed film and viscous fluid loading can be treated as additive and are often modeled as:

(10)Δf=-2f02Zcq(ΔmA+ρlηl4πf0) (10)

If the adsorbed film layer exhibits viscoelastic characteristics, the resonant frequency shift calculations are more mathematically complex, and the reader may consult reference Lucklum and Hauptmann (2000). In most biological assays, the target is present in an aqueous fluid, and thus, the Newtonian behavior is a good approximation.

A 5-MHz QCM is sensitive at 1 ng/cm2 of analyte (Rodahl et al. 1995). Amplification is required to achieve higher sensitivity; alternatively, a higher resonant frequency crystal may be used. When PCR is used to amplify the target nucleotide, a combined limit of detection was reported as 100 fg genomic DNA (Sun et al. 2006). Sandwich binding of nucleotide-tagged gold nanoparticles can quadruple the frequency response and sensitivity (Chen et al. 2008). The sensor platform is commercially available and is well- characterized. Manufacturers of QCM products include Standard Research Systems, Q-Sense, and International Crystal Manufacturing. A typical QCM frequency response for DNA detection is provided in Figure 9B with temporal responses for the different phases of hybridization, washing, and sensor regeneration.

Dynamic cantilever sensors

Dynamic cantilevers exhibit multiple modes of resonance including bending, torsional, and lateral modes. The bending modes, in general, are the most extensively used for sensing. A resonant-mode cantilever transduces added mass as decrease in resonant frequency. As a distributed mass system, the displacement and resonant frequencies are described by a set of governing partial differential equations (Sader 1998). Researchers, however, frequently approximate a dynamic cantilever as a spring-mass system (single harmonic oscillator) that exhibits only one mode. For the simple spring-mass system, the relation between added mass, Δm and resonant frequency is given by (Fritz 2008):

(11)Δm=k4π2(1fm2-1f02) (11)

The parameters k, fm, and f0 represent the spring constant, resonant frequency after mass loading and before mass loading. The spring constant is a combination of the Young’s modulus, cantilever length, and the distribution of mass. A sketch of resonating cantilever operating in a bending mode is given in Figure 10. The figure also includes the frequency response to an added mass.

Figure 10 A rough sketch of a resonating cantilever is presented where a capture probe is immobilized on its surface. The resonant frequency of the cantilever shifts downward in response to the added mass of the target strand.

Figure 10

A rough sketch of a resonating cantilever is presented where a capture probe is immobilized on its surface. The resonant frequency of the cantilever shifts downward in response to the added mass of the target strand.

Dynamic cantilevers in a liquid medium respond to motion of the surrounding fluid like the QCM device. Unlike the QCM device, the cantilever operates out-of-plane so there is an inertial component in addition to the viscous forces; the fluid exerts an opposing force as the cantilever “pushes” the fluid. The relative contribution of the viscous and inertial forces is characterized by the Reynolds’ number of the cantilever:

(12)Re=ρωb24η (12)

The cantilever resonant frequency, cantilever width, fluid density, and viscosity are given as ω, b, ρ, and η, respectively. The inertial contribution dominates when Re>>1. It is clear from equation (12) that inertial forces dominate as resonant frequency and/or cantilever width increase. In such a case, resonant frequency of a cantilever in an inviscid medium of density, ρ, is given as (Sader 1998):

(13)ffluidfvac=(1+πρb4ρch)-1/2 (13)

where the ratio of the resonant frequencies in fluid (ffl) and vacuum (fvac) are a function of fluid density (ρ), density of the cantilever (ρc), cantilever width (b), and cantilever thickness (h). If the Reynolds number decreases as a result of increasing viscosity or smaller b, the viscous contribution grows more significant, and two things occur: the resonant frequency shifts further downward, and the peak widens as cantilever motion is damped (Sader 1998, Dorignac et al. 2006, Fritz 2008). The quality factor, Q, is a measure of the energy lost in one period of cantilever vibration:

(14)Q=EdriveElost=f0Δf1/2 (14)

The energy to drive the cantilever in one cycle and the energy lost in the cycle are given as Edrive and Elost. The resonant frequency and peak width at half amplitude are f0 and Δf1/2. The quality factor, thus, decreases as the cantilever is further damped. The mathematical relations that describe the viscous effects are presented in greater detail elsewhere (Sader 1998, Chon et al. 2000).

Applications

In Table 3, a summary of the investigations are given for the detection of nucleic acids of foodborne bacteria, waterborne parasites, and viruses. It must be noted that the cases that are most applicable for commercial testing of food and blood are the ones that demonstrate detection capability in complex media such as serum, food matrix, cell extracts, or tissue samples (Dell’Atti et al. 2007, Johnson and Mutharasan 2013, Sharma and Mutharasan 2013). Some of the studies in Table 3 use PCR as a first step to amplify the original DNA in sample, while others sample the genomic extract without any further treatment. We examine the details of two investigations to illustrate the methods used for each sensor type.

Table 3

Electromechanical detection of foodborne and waterborne pathogenic DNA/RNA.

Detection techniqueTargetSample matrixAnalysis time (h)Limit of detectionReferences
QCM with PCRE. coli O157:H7Buffer/PCR mix2N/AWu et al. 2007
QCM with PCRE. coli O157:H7Buffer/PCR mix24 ng/l PCR product/8 cfuSun et al. 2009
QCM with PCRE. coli O157:H7Juice/milk/beef2.55120 cfu/mlChen et al. 2008
QCML. monocytogenesMilk10 cfu/mlZhou et al. 2010
QCM with PCRB. anthracisPCR mix3.33.5×102 cfu/mlHao et al. 2011
QCM with PCRS. epidermisPCR mix4–51.3×103 cfu/mlXia et al. 2008
CantileverE. coli O157:H7Beef wash1–2720 cfuRijal and Mutharasan 2013
CantileverL. monocytogenesProteinous background1.5700 cfuSharma and Mutharasan 2013
CantileverCyanobacteriaCell extract250 cfu/mlJohnson and Mutharasan 2013
QCM with PCRAmplified HPV from tissuePCR mix with background DNA0.5 (+PCR)30 nMDell’Atti et al. 2007
QCM with PCRVaccina virus DNA50 μl PCR mix0.5 (+PCR)25 ng/50 μlKleo et al. 2011
QCM with PCRDengue virus DNA from mosquito cellsPCR mix1.5 (+PCR)2 Plaque-forming units/mlChen et al. 2009
QCMSynthetic HSV DNABuffer∼452 pMUludag et al. 2008

Chen et al. used PCR in conjunction with QCM for detecting E. coli O157:H7 with a combined limit of detection of 1.2×104 cfu/ml (Chen et al. 2008). All surface modifications and measurements were carried out with the sensor installed in a flow cell. An ssDNA probe complementary to the eaeA gene target was immobilized onto the gold-coated QCM surface for detection. Far lower limit of detection of 1.2×102 cfu/ml was realized by amplifying the QCM signal with a second gold nanoparticle-labeled ssDNA complementary to the collected target eaeA gene captured on the sensor. The limit of detection corresponds to the concentration in the original cell suspensions prior to DNA extraction and the PCR amplification step. The authors demonstrated that their method was specific to the target by conducting sensing experiments with DNA strands derived from other bacterial strains. The presence of extraneous strands caused negligible change in sensor response. The Au nanoparticle amplification reduced the possibility of false negatives. The investigators reported that the frequency response was halved when beef samples inoculated with E. coli O157:H7 were used, but did not quantify the detection limit for the complex matrix.

Investigators using a piezoelectric cantilever sensor achieved a detection limit of 7×102L. monocytogenes cells without an amplifying PCR step (Sharma and Mutharasan 2013). All modification and measurements were conducted with the sensor installed in a flow cell. The genomic DNA extract of the cells was introduced after immobilizing complementary ssDNA capture probes on the sensor surface. Gold nanoparticles were used to amplify the frequency response, as illustrated in Figure 11A. The transient sensor response in the presence of the genomic DNA extract is given in Figure 11B. The performance of the cantilever in the presence of extraneous genetic and protein using a Listeria sample mixed with the genomic extract of E. coli JM101, which was composed of roughly 25% cell protein. The extraneous E. coli material lessened the detection limit to 7×103 cells/ml of L. monocytogenes. The detection limit of 7×102 cells is impressive given that no PCR amplification was required, in spite of the 25% protein background of the genomic extract. All of the QCM entries of Table 2 used PCR to realize the reported low detection limits.

Figure 11 (A) The signal for a QCM or resonating cantilever can be amplified with gold nanoparticles conjugated with DNA complementary to the target strand. The overall frequency change is multiplied as the mass of nanoparticle is substantially greater than the target nucleotide. This image was taken with permission from Johnson and Mutharasan (2013). Copyright © 2013 American Chemical Society. (B) The cantilever sensor response to increasing concentrations of Listeria genomic DNA extract, followed by a dramatic decrease in resonant frequency with the addition of functionalized gold nanoparticles. This image was reproduced with permission from Sharma and Mutharasan (2013). Copyright © 2013 American Chemical Society.

Figure 11

(A) The signal for a QCM or resonating cantilever can be amplified with gold nanoparticles conjugated with DNA complementary to the target strand. The overall frequency change is multiplied as the mass of nanoparticle is substantially greater than the target nucleotide. This image was taken with permission from Johnson and Mutharasan (2013). Copyright © 2013 American Chemical Society. (B) The cantilever sensor response to increasing concentrations of Listeria genomic DNA extract, followed by a dramatic decrease in resonant frequency with the addition of functionalized gold nanoparticles. This image was reproduced with permission from Sharma and Mutharasan (2013). Copyright © 2013 American Chemical Society.

Conclusions and outlook

The electrochemical sensors feature very impressive detection limits, but they have three significant drawbacks. None of the electrochemical sensor results given in Table 2 were obtained as the hybridization process was taking place. Hybridization was only confirmed after a rinsing step and transfer of the DNA-exposed electrode to an electrochemical cell. Such a method is prone to variability. Second, they rely on PCR or enrichment steps to accomplish detection limits as low as 1 cfu/ml. PCR can present limitations with complex food samples, as was discussed earlier. Last, the electrochemical sensors only sample very small volumes on the order of 5–10 μl of the nucleotide samples. In spite of the small sample sizes, the reported limits of detection are impressive, and amplification via redox indicators is a sound means of preventing false negatives. This is an actively investigated field, as evidenced by the substantial number of entries for the few pathogens of interest in this review.

The electromechanical DNA sensors can measure larger sample volumes because they are amenable for measurement in a flow format. Furthermore, all surface modifications and hybridization schemes can be done with the devices installed in the flow loop. The QCM sensors require PCR and amplification to accomplish acceptable detection limits. The cantilever sensors have detected concentrations on the order of 100 cfu/ml without an enrichment procedure or an amplification step.


Corresponding author: Raj Mutharasan, Department of Chemical and Biological Engineering, Drexel University, Philadelphia, PA 19104, USA, e-mail:

Acknowledgments

The authors are grateful for the generous support of NSF Grant CBET- 1159841 which provided the funds for the reported work.

References

Aguilar, Z. P.; Fritsch, I. Immobilized enzyme-linked DNA-hybridization assay with electrochemical detection for Cryptosporidium parvum hsp70 mRNA. Anal. Chem.2003, 75, 3890–3897. Search in Google Scholar

Ahour, F.; Pournaghi-Azar, M. H.; Alipour, E.; Hejazi, M. S. Detection and discrimination of recombinant plasmid encoding hepatitis C virus core/E1 gene based on PNA and double-stranded DNA hybridization. Biosens. Bioelectron.2013, 45, 287–291. Search in Google Scholar

Ahour, F.; Pournaghi-Azar, M. H.; Hejazi, M. S. An electrochemical approach for direct detection and discrimination of fully match and single base mismatch double-stranded oligonucleotides corresponding to universal region of hepatitis C virus. Anal. Methods2012, 4, 967–972. Search in Google Scholar

Anderson, M. J.; Miller, H. R.; Alocilja, E. C. PCR-less DNA co-polymerization detection of Shiga like toxin 1 (stx1) in Escherichia coli O157:H7. Biosens. Bioelectron.2013, 42, 581–585. Search in Google Scholar

Baldursson, S.; Karanis, P. Waterborne transmission of protozoan parasites: review of worldwide outbreaks – An update 2004–2010. Water Res.2011, 45, 6603–6614. Search in Google Scholar

Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Principles and Applications; New York: John Wiley, 2001. Search in Google Scholar

Berdat, D.; Rodriguez, A. C. M.; Herrera, F.; Gijs, M. A. M. Label-free detection of DNA with interdigitated micro-electrodes in a fluidic cell. Lab Chip2008, 8, 302–308. Search in Google Scholar

Brett, C. B. O. A.M. Electrochemistry: Principles, Methods, and Applications; New York; Oxford: Oxford University Press; 1993; p 427. Search in Google Scholar

Chen, S.-H.; Wu, V. C. H.; Chuang, Y.-C.; Lin, C.-S. Using oligonucleotide-functionalized Au nanoparticles to rapidly detect foodborne pathogens on a piezoelectric biosensor. J. Microbiol. Methods2008, 73, 7–17. Search in Google Scholar

Chen, S.-H.; Chuang, Y.-C.; Lu, Y.-C.; Lin, H.-C.; Yang, Y.-L.; Lin, C.-S. A method of layer-by-layer gold nanoparticle hybridization in a quartz crystal microbalance DNA sensing system used to detect dengue virus. Nanotechnology2009, 20, 215501–215510. Search in Google Scholar

Chon, J. W. M.; Mulvaney, P.; Sader, J. E. Experimental validation of theoretical models for the frequency response of atomic force microscope cantilever beams immersed in fluids. J. Appl. Phys.2000, 87, 3978–3988. Search in Google Scholar

Civit, L.; Fragoso, A.; Hoelters, S.; Duerst, M.; O’Sullivan, C. K. Electrochemical genosensor array for the simultaneous detection of multiple high-risk human papillomavirus sequences in clinical samples. Anal. Chim. Acta2012, 715, 93–98. Search in Google Scholar

Deakin, M. R.; Buttry, D. A. Electrochemical applications of the quartz crystal microbalance. Anal. Chem.1989, 61, 1147A–1154A. Search in Google Scholar

Dorignac, J.; Kalinowski, A.; Erramilli, S.; Mohanty, P. Dynamical response of nanomechanical oscillators in immiscible viscous fluid for in vitro biomolecular recognition. Phys. Rev. Lett.2006, 96, 186105. Search in Google Scholar

Doyle, M. P.; Beuchat, L. R. Food Microbiology: Fundamentals and Frontiers; 3rd Edition; Washington, DC: ASM Press, 2007; p. 1038. Search in Google Scholar

Dell’Atti, D.; Zavaglia, M.; Tombelli, S.; Bertacca, G.; Cavazzana, A. O.; Bevilacqua, G.; Minunni, M.; Mascini, M. Development of combined DNA-based piezoelectric biosensors for the simultaneous detection and genotyping of high risk human papilloma virus strains. Clin. Chim. Acta2007, 383, 140–146. Search in Google Scholar

Driscoll, A. J.; Harpster, M. H.; Johnson, P. A. The development of surface-enhanced Raman scattering as a detection modality for portable in vitro diagnostics: progress and challenges. Phys. Chem. Chem. Phys.2013, 15, 20415–20433. Search in Google Scholar

Dubois, M. A.; Muralt, P. Properties of aluminum nitride thin films for piezoelectric transducers and microwave filter applications. Appl. Phys. Lett.1999, 74, 3032–3034. Search in Google Scholar

Dubois, M. A.; Muralt, P. Stress and piezoelectric properties of aluminum nitride thin films deposited onto metal electrodes by pulsed direct current reactive sputtering. J. Appl. Phys.2001, 89, 6389–6395. Search in Google Scholar

Falconer, I. R.; Humpage, A. R. Health risk assessment of cyanobacterial (blue-green algal) toxins in drinking water. Int.J. Environ. Res. Public Health2005, 2, 43–50. Search in Google Scholar

Fritz, J. Cantilever biosensors. Analyst2008, 133, 855–863. Search in Google Scholar

Gehring, A. G.; Irwin, P. L.; Reed, S. A.; Tu, S. I.; Andreotti, P. E.; Akhavan-Tafti, H.; Handley, R. S. Enzyme-linked immunomagnetic chemiluminescent detection of Escherichia coli O157:H7. J. Immunol. Methods2004, 293, 97–106. Search in Google Scholar

Gao, Y.; Wolf, L. K.; Georgiadis, R. M. Secondary structure effects on DNA hybridization kinetics: a solution versus surface comparison. Nucleic Acids Res.2006, 34, 3370–3377. Search in Google Scholar

Ghany, M. G.; Strader, D. B.; Thomas, D. L.; Seeff, L. B. Diagnosis, management, and treatment of hepatitis C: an update. Hepatology2009, 49, 1335–1374. Search in Google Scholar

Gong, P.; Levicky, R. DNA surface hybridization regimes. Proc. Natl. Acad. Sci.2008, 105, 5301–5306. Search in Google Scholar

Goyal, R. N.; Gupta, V. K.; Bachheti, N. Fullerene-C60-modified electrode as a sensitive voltammetric sensor for detection of nandrolone – An anabolic steroid used in doping. Anal. Chim. Acta2007, 597, 82–89. Search in Google Scholar

Grabowska, I.; Singleton, D. G.; Stachyra, A.; Gora-Sochacka, A.; Sirko, A.; Zagorski-Ostoja, W.; Radecka, H.; Stulz, E.; Radecki. A highly sensitive electrochemical genosensor based on Co-porphyrin-labelled DNA. J. Chem. Commun.2014a, 50, 4196–4199. Search in Google Scholar

Grabowska, I.; Stachyra, A.; Gora-Sochacka, A.; Sirko, A.; Olejniczak, A. B.; Lesnikowski, Z. J.; Radecki, J.; Radecka, H. DNA probe modified with 3-iron bis(dicarbollide) for electrochemical determination of DNA sequence of Avian Influenza Virus H5N1. Biosens. Bioelectron.2014b, 51, 170–176. Search in Google Scholar

Hagan, M. F.; Chakraborty, A. K. Hybridization dynamic of surface immobilized DNA. J. Chem. Phys.2004, 120, 4958–4968. Search in Google Scholar

Hao, R.-Z.; Song, H.-B.; Zuo, G.-M.; Yang, R.-F.; Wei, H.-P.; Wang, D.-B.; Cui, Z.-Q.; Zhang, Z.; Cheng, Z.-X.; Zhang, X.-E. DNA probe functionalized QCM biosensor based on gold nanoparticle amplification for Bacillus anthracis detection. Biosens. Bioelectron.2011, 26, 3398–3404. Search in Google Scholar

Harpster, M. H.; Zhang, H.; Sankara-Warrier, A. K.; Ray, B. H.; Ward, T. R.; Kollmar, J. P.; Carron, K. T.; Mecham, J. O.; Corcoran, R. C.; Wilson, W. C.; Johnson, P. A. SERS detection of indirect viral DNA capture using colloidal gold and methylene blue as a Raman label. Biosens. Bioelectron.2009, 25, 674–681. Search in Google Scholar

Heidenreich, B.; Poehlmann, C.; Sprinzl, M.; Gareis, M. Detection of Escherichia coli in Meat with an Electrochemical Biochip. J. Food Prot.2010, 73, 2025–2033. Search in Google Scholar

Hitchins, A. D., Jinneman, K. In Bacteriological Analytical Manual; FDA, 2011. Search in Google Scholar

Ilic, B.; Czaplewski, D.; Craighead, H. G.; Neuzil, P.; Campagnolo, C.; Batt, C. Mechanical resonant immunospecific biological detector. Appl. Phys. Lett.2000, 77, 450–452. Search in Google Scholar

Ilic, B.; Craighead, H.; Krylov, S.; Senaratne, W.; Ober, C.; Neuzil, P. Attogram detection using nanoelectromechanical oscillators. J. Appl. Phys.2004, 95, 3694–3703. Search in Google Scholar

In CDC Estimates of Foodborne Illness in the United States; Centers for Disease Control and Prevention, 2014. Search in Google Scholar

Jan, L. In Electrochemical DNA Biosensors; Singapore: Pan Stanford Publishing; 2012; pp 1–27. Search in Google Scholar

Johnson, B. N.; Mutharasan, R. A cantilever biosensor-based assay for toxin-producing cyanobacteria Microcystis aeruginosa using 16S rRNA. Environ. Sci. Technol.2013, 47, 12333–12341. Search in Google Scholar

Joung, H.-A.; Lee, N.-R.; Lee, S. K.; Ahn, J.; Shin, Y. B.; Choi, H.-S.; Lee, C.-S.; Kim, S.; Kim, M.-G. High sensitivity detection of 16s rRNA using peptide nucleic acid probes and a surface plasmon resonance biosensor. Anal. Chim. Acta2008, 630, 168–173. Search in Google Scholar

Kerman, K.; Kobayashi, M.; Tamiya, E. Recent trends in electrochemical DNA biosensor technology. Meas. Sci. Technol.2004, 15, R1. Search in Google Scholar

Kermekchiev, M. B.; Kirilova, L. I.; Vail, E. E.; Barnes, W. M. Mutants of Taq DNA polymerase resistant to PCR inhibitors allow DNA amplification from whole blood and crude soil samples. Nucleic Acids Res.2009, 37, e40. Search in Google Scholar

Kissinger, P. T.; Heineman, W. R. Cyclic voltammetry. J. Chem. Educ.1983, 60, 702. Search in Google Scholar

Kleo, K.; Kapp, A.; Ascher, L.; Lisdat, F. Detection of vaccinia virus DNA by quartz crystal microbalance. Anal. Biochem.2011, 418, 260–266. Search in Google Scholar

Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R.; Feld, M. S. Single molecule detection using Surface-Enhanced Raman Scattering (SERS). Phys. Rev. Lett.1997, 78, 1667–1670. Search in Google Scholar

Lam, B.; Fang, Z.; Sargent, E. H.; Kelley, S. O. Polymerase chain reaction-free, sample-to-answer bacterial detection in 30 minutes with integrated cell lysis. Anal. Chem.2012, 84, 21–25. Search in Google Scholar

Lane, R. F.; Hubbard, A. T. Differential double pulse voltammetry at chemically modified platinum electrodes for in vivo determination of catechol amines. Anal. Chem.1976, 48, 1287–1293. Search in Google Scholar

Lee, C.; Itoh, T.; Suga, T. Self-excited piezoelectric PZT microcantilevers for dynamic SFM – with inherent sensing and actuating capabilities. Sens. Actuators A Phys.1999, 72, 179–188. Search in Google Scholar

Lereau, M.; Fournier-Wirth, C.; Mayen, J.; Farre, C.; Meyer, A.; Dugas, V.; Cantaloube, J.-F.; Chaix, C.; Vasseur, J.-J.; Morvan, F. Development of innovative and versatile polythiol probes for use on ELOSA or electrochemical biosensors: application in hepatitis C virus genotyping. Anal. Chem.2013, 85, 9204–9212. Search in Google Scholar

Li, M.; Cushing, S. K.; Liang, H.; Suri, S.; Ma, D.; Wu, N. Plasmonic nanorice antenna on triangle nanoarray for surface-enhanced raman scattering detection of hepatitis B virus DNA. Anal. Chem.2013, 85, 2072–2078. Search in Google Scholar

Liao, W.-C.; Ho, J.-A. A. Attomole DNA Electrochemical Sensor for the Detection of Escherichia coli O157. Anal. Chem.2009, 81, 2470–2476. Search in Google Scholar

Lucklum, R.; Hauptmann, P. The quartz crystal microbalance: mass sensitivity, viscoelasticity and acoustic amplification. Sens. Actuators B Chem.2000, 70, 30–36. Search in Google Scholar

Martin, S. J.; Granstaff, V. E.; Frye, G. C. Characterization of a quartz crystal microbalance with simultaneous mass and liquid loading. Anal. Chem.1991, 63, 2272–2281. Search in Google Scholar

Mithran, S.; Werasak, S. In Electrochemical DNA Biosensors; Singapore: Pan Stanford Publishing, 2012; pp. 163–204. Search in Google Scholar

Osteryoung, J. G.; Osteryoung, R. A. Square wave voltammetry. Anal. Chem.1985, 57, 101A–110A. Search in Google Scholar

Pal, S.; Alocilja, E. C. Electrically active magnetic nanoparticles as novel concentrator and electrochemical redox transducer in Bacillus anthracis DNA detection. Biosens. Bioelectron.2010, 26, 1624–1630. Search in Google Scholar

Peterson, A. W.; Heaton, R. J.; Georgiadis, R. M. The effect of surface probe density on DNA hybridization. Nucleic Acids Res.2001, 29, 5163–5168. Search in Google Scholar

Piliarik, M.; Parova, L.; Homola, J. High-throughput SPR sensor for food safety. Biosens. Bioelectron.2009, 24, 1399–1404. Search in Google Scholar

Pinijsuwan, S.; Rijiravanich, P.; Somasundrum, M.; Surareungchai, W. Sub-femtomolar electrochemical detection of DNA hybridization based on latex/gold nanoparticle-assisted signal amplification. Anal. Chem.2008, 80, 6779–6784. Search in Google Scholar

Rauzan, B.; McMichael, E.; Cave, R.; Sevcik, L. R.; Ostrosky, K.; Whitman, E.; Stegemann, R.; Sinclair, A. L.; Serra, M. J.; Deckert, A. A. Kinetics and thermodynamics of DNA, RNA, and hybrid duplex formation. Biochemistry2013, 52, 765–772. Search in Google Scholar

Rijal, K.; Mutharasan, R. A method for DNA-based detection of E. coli O157:H7 in a proteinous background using piezoelectric-excited cantilever sensors. Analyst2013, 138, 2943–2950. Search in Google Scholar

Rodahl, M.; Höök, F.; Krozer, A.; Brzezinski, P.; Kasemo, B. Quartz crystal microbalance setup for frequency and Q-factor measurements in gaseous and liquid environments. Rev. Sci. Instrum.1995, 66, 3924–3930. Search in Google Scholar

Sader, J. E. Frequency response of cantilever beams immersed in viscous fluids with applications to the atomic force microscope. J. Appl. Phys.1998, 84, 64–76. Search in Google Scholar

Sekar, M. M. A.; Bloch, W.; St John, P. M. Comparative study of sequence-dependent hybridization kinetics in solution and on microspheres. Nucleic Acids Res.2005, 33, 366–375. Search in Google Scholar

Sharma, H.; Mutharasan, R. hlyA gene-based sensitive detection of Listeria monocytogenes using a novel cantilever sensor. Anal. Chem.2013, 85, 3222–3228. Search in Google Scholar

Shrout, T. R.; Zhang, S. J. Lead-free piezoelectric ceramics: alternatives for PZT? J. Electroceram.2007, 19, 113–126. Search in Google Scholar

Sipova, H.; Homola, J. Surface plasmon resonance sensing of nucleic acids: a review Anal. Chim. Acta2013, 773, 9–23. Search in Google Scholar

Soejima, T.; Minami, J.; Iwatsuki, K. Rapid propidium monoazide PCR assay for the exclusive detection of viable Enterobacteriaceae cells in pasteurized milk. J. Dairy Sci.2012, 95, 3634–3642. Search in Google Scholar

Song, Y.; Bhushan, B. Simulation of dynamic modes of atomic force microscopy using a 3D finite element model. Ultramicroscopy2006, 106, 847–873. Search in Google Scholar

Springer, T.; Sipova, H.; Vaisocherova, H.; Stepanek, J.; Homola, J. Shielding effect of monovalent and divalent cations on solid-phase DNA hybridization: surface plasmon resonance biosensor study. Nucleic Acids Res.2010, 38, 7343–7351. Search in Google Scholar

Sun, H.; Zhang, Y.; Fung, Y. Flow analysis coupled with PQC/DNA biosensor for assay of E. coli based on detecting DNA products from PCR amplification. Biosens. Bioelectron.2006, 22, 506–512. Search in Google Scholar

Sun, H.; Choy, T. S.; Zhu, D. R.; Yam, W. C.; Fung, Y. S. Nano-silver-modified PQC/DNA biosensor for detecting E. coli in environmental water. Biosens. Bioelectron.2009, 24, 1405–1410. Search in Google Scholar

Sun, W.; Qi, X.; Zhang, Y.; Yang, H.; Gao, H.; Chen, Y.; Sun, Z. Electrochemical DNA biosensor for the detection of Listeria monocytogenes with dendritic nanogold and electrochemical reduced graphene modified carbon ionic liquid electrode. Electrochim. Acta2012, 85, 145–151. Search in Google Scholar

Todd, E. C. D.; Notermans, S. Surveillance of listeriosis and its causative pathogen, Listeria monocytogenes. Food Control2011, 22, 1484–1490. Search in Google Scholar

Uludag, Y.; Li, X.; Coleman, H.; Efstathiou, S.; Cooper, M. A. Direct acoustic profiling of DNA hybridisation using HSV type 1 viral sequences. Analyst2008, 133, 52–57. Search in Google Scholar

Wada, S.; Suzuki, S.; Noma, T.; Suzuki, T.; Osada, M.; Kakihana, M.; Park, S. E.; Cross, L. E.; Shrout, T. R. Enhanced piezoelectric property of barium titanate single crystals with engineered domain configurations. Jpn. J. Appl. Phys. Part 1-Regul. Pap. Short Notes Rev. Pap.1999, 38, 5505–5511. Search in Google Scholar

Waggoner, P. S.; Craighead, H. G. Micro- and nanomechanical sensors for environmental, chemical, and biological detection. Lab Chip2007, 7, 1238–1255. Search in Google Scholar

Wu, V. C. H.; Chen, S.-H.; Lin, C.-S. Real-time detection of Escherichia coli O157:H7 sequences using a circulating-flow system of quartz crystal microbalance. Biosens. Bioelectron.2007, 22, 2967–2975. Search in Google Scholar

Wu, J.; Chumbimuni-Torres, K. Y.; Galik, M.; Thammakhet, C.; Haake, D. A.; Wang, J. Potentiometric detection of DNA hybridization using enzyme-induced metallization and a silver ion selective electrode. Anal. Chem.2009, 81, 10007–10012. Search in Google Scholar

Xia, H.; Wang, F.; Huang, Q.; Huang, J.; Chen, M.; Wang, J.; Yao, C.; Chen, Q.; Cai, G.; Fu, W. Detection of Staphylococcus epidermidis by a quartz crystal microbalance nucleic acid biosensor array using Au nanoparticle signal amplification. Sensors2008, 8, 6453–6470. Search in Google Scholar

Zezza, F.; Pascale, M.; Mule, G.; Visconti, A. Detection of Fusarium culmorum in wheat by a surface plasmon resonance-based DNA sensor. J. Microbiol. Methods2006, 66, 529–537. Search in Google Scholar

Zhang, D.; Yan, Y.; Li, Q.; Yu, T.; Cheng, W.; Wang, L.; Ju, H.; Ding, S. Label-free and high-sensitive detection of Salmonella using a surface plasmon resonance DNA-based biosensor. J. Biotechnol.2012, 160, 123–128. Search in Google Scholar

Zheng, J.; Chen, C.; Wang, X.; Zhang, F.; He, P. A sequence-specific DNA sensor for Hepatitis B virus diagnostics based on the host–guest recognition. Sens. Actuator B Chem.2014, 199, 168–174. Search in Google Scholar

Zhou, H.; Gao, Z.; Luo, G.; Han, L.; Sun, S.; Wang, H. Determination of listeria monocytogenes in milk samples by signal amplification quartz crystal microbalance sensor. Anal. Lett.2010, 43, 312–322. Search in Google Scholar

Supplemental Material

The online version of this article (DOI: 10.1515/revac-2014-0017) offers supplementary material, available to authorized users.

Received: 2014-6-5
Accepted: 2014-8-30
Published Online: 2014-11-12
Published in Print: 2014-12-1

©2014 by De Gruyter