Electrochemical DNA sensors based on spatially distributed redox mediators: challenges and promises

Bioreceptor layer assembling

The biosensor assembling is schematically outlined in Fig. 1. The surface layer bearing mediators and DNA probes or aptamers was stepwise deposited by self-organization of the charged components followed by their covalent modification. The quantities of the reactants added to the growing layer are controlled by the charge of the primary layer attached to the electrode surface. The deposition of the other polyelectrolytes depends on their own charge and size but to a much lower extent by their concentrations in the solution and incubation period. The final layer facing the solution is further modified by aminated redox label and biospecific receptor via carbodiimide binding.

Fig. 1:

Electrostatic assembling of DNA sensors with 3D net of redox centers. (a) Deposition of the polymeric NR layer; (b) electrostatic accumulation of negatively charged carboxylated macrocycles; (c) carbodiimide binding of monomeric NR and aminated aptamer molecules; (d) reaction with thrombin as target analyte; (e) electronic exchange in 3D net of the surface layer.

Spatial separation of the redox centers in the film obtained is a most important factor affecting the biosensor performance. On the one hand, the redox centers introduced in the surface layer should retain their ability to electron exchange. On the other hand, their surface density should leave enough space for analyte molecules penetrating the layer in the biochemical recognition event. Introduction of bulky non-conductive molecules in the redox net suppresses the electron exchange between reduced and oxidized redox centers of the label and hence decrease the current recorded after incubation in the sample tested.

The 3D redox net described can be obtained by co-polymerization or co-deposition of different electrochemically active species [39], [42], [43] or by casting mesoporous materials made of non-conductive material followed by their filling with redox species [44], [45]. However, such alternatives require precise tuning of the layer content and high regularity of the final product structure. Taking into account multistep protocol of the surface layer deposition and necessity to introduce biochemical receptors in the pores of such materials, the advantages of the protocols described seem ambiguous.

Macrocyclic receptors belonging to calixarenes and their thioanalogs are mostly applied in analytical chemistry as specific sorbents and ionophores toward metal cations [46]. Guest ions are commonly entrapped in the pseudo-cavity formed by substituents at the lower rim of the macrocyclic platform that pre-determines relative positioning of the functional groups in accordance with the size and charge of a guest. Meanwhile, the same macrocycles can be employed in the structure of the biospecific layers as bioreceptor carriers. Various functional groups including those of redox labels can be covalently attached to the terminal amino or carboxylic groups of the macrocycle substituents. In both cases, macrocyclic moiety separates the biochemical binding sites and decreases steric limitations of their interactions with the analyte molecules. Similar behavior can be expected from other representatives of metacyclophanes, e.g. pillar[n]arenes [47] or cucurbit[m]uriles [48]. The pillar[5]arene has the size comparable to that of thiacalix[4]arene but involves 10 hydroxyl groups, which can be modified with redox centers. Although the rotation of hydroquinone units around methylene bridges does not provide the formation of stable configurations, any position of the macrocycle units offers potential binding sites for an analyte from both sides of the central core. This makes the accessibility of their terminal functions similar to that of 1,3-alternate of (thia)calix[4]arene. The convenient synthesis of pillar[5]arene proposed by T. Ogoshi et al. in 2008 [49] has offered new opportunities of chemical functionalization of this promising class of receptor precursors. Cucurbit[7]uril was reported in the assembly of DNA and aptasensors as a part of universal immobilization platform due to its ability to form host-guest complexes with ferrocenemethylammonium cation where the guest retains its redox activity recorded by voltammetry [50]. Chemical structures of the macrocyclic carriers applied in the assembly of aptasensors are presented in Fig. 2.

Fig. 2:

Chemical structures of carboxylated derivative of thiacalix[4]arene (cone conformation) and pillar[5]arene used in the work.

The use of macrocyclic carriers of the redox labels provides their spatial separation from each other but also increases the resistance of the electron transfer that influences the voltammetric response and its changes in the presence of receptors and guest molecules. This was for the first time observed for electrochemical behavior of Neutral red (NR) covalently attached to the macrocycle by carbodiimide binding (3). The NR molecule is involved in reversible transfer of two electrons and one hydrogen ion (4) corresponded to the classical pair of the peaks on voltammogram.

Contrary to redox species directly attached to the electrode, the NR introduced in the macrocycle retains the shape of the cyclic voltammogram typical for diffusionally free mediators. Moreover, the appropriate peak currents changed proportionally to the square root from the scan rate though no free mediators were added to the solution. Such a behavior confirmed the electron exchange between the oxidized and reduced forms of the NR dye present in the surface layer (see Fig. 1e).

The coulometric measurements as well as direct comparison of the peak currents found for the NR molecules, being free and immobilized on the macrocyclic support, made it possible to conclude that about 30% of the carboxylic groups of the thiacalix[4]arene derivatives were substituted with the NR functions, probably, due to steric limitations of carbodiimide binding reaction. This means, the number of remained carboxylic groups is quite sufficient for both electrostatic assembling in the surface layer and binding with appropriate aptamer molecules. Certainly, implementation of various aptamers toward different species affects the voltammograms of the NR in some different manner but all of the resulting voltammograms show similarity in the relative height of the cathodic and anodic NR peaks and peak potentials. Cathodic peaks are sufficiently higher than anodic peaks and are better resolved. For this reason, their shifts better represent the affine interactions in the layer.

Implementation of the macrocyclic carriers results also in a slow restore of the biosensor after the measurement. In many cases, the peaks on the NR voltammogram progressively decreased in the series of potential runs performed with the same biosensor with intermediate solution stirring in open circuit mode. The addition of the K₃[Fe(CN)₆] solution accelerated the reversed oxidation of the NR groups in the middle of the layer and decreased the measurement time to 2–3 min. In optimal conditions, the voltammetric characteristics of the NR groups attached to the macrocyclic supports were well reproducible with 1.5%–2.5% deviation of the cathodic peak current and allowed up to 50 measurements with no losses of the signal.

3.2. Thrombin aptasensor

Human α-thrombin called also as is a blood clothing factor is a multifunctional serine protease, which plays an important role in the procoagulant and anticoagulant functions [51], [52]. The monitoring of the α-thrombin level is important in cardiovascular disease therapy. Thrombin is often used as a model of the DNA-protein interactions and particularly in the development of the novel signal transduction principles. In biosensors, aptamers are used for specific thrombin recognition. Aptamers are short consequences of the DNA oligonucleotides that are synthesized de novo using combinatorial chemistry approach and then selected against an analyte using affine chromatography. The appropriate protocol called SELEX [53] results in specifying unique molecule with most efficient binding to the analyte molecule. In case of α-thrombin, two DNA aptamers have been described, one of which is 15-mer sequence 5′-GGT TGG TGT GGT TGG-3′ specific to fibrinogen binding exosite (FIBRI) [52] whereas another one specifically binds the heparin binding exosite (HEPA) [54]. The binding motif of both DNA aptamers is formed by G-quartet structure. The FIBRI based aptamers are most intensively studied in the biosensor assemblies. To avoid significant limitations of the analyte access, the poly(dT) tail was introduced in its structure together with terminal amino group for carbodiimide binding to the carboxylate group of the macrocyclic carrier. The aptasensor toward α-thrombin consisted of the poly(NR) layer obtained by electropolymerization performed by multiple potential cycling. Then, polycarboxylated thiacalix[4]arene was casted on the polymer surface. The following treatment with EDC-NHS, common reagents used in carbodiimide binding, and the NR/aptamer solution [see scheme (3)] make it possible to assemble the biorecognition layer. The incubation of the aptasensor in α-thrombin solution suppressed the cathodic NR peak on voltammogram. The analyte concentration range covers sub-micromolar concentrations (see also Fig. 1 for more details) [55]. The deposition of the layers was controlled by electrochemical impedance spectroscopy (EIS). The charge transfer resistance measured in the presence of ferricyanide ions as redox probe increased with each layer and thrombin concentration. The AFM study indicates the role of hydrogen bonds in the formation of specific morphology of the active surface: first, deposition of agglomerates onto the polymeric NR layer takes place. Carbodiimide binding destroys the system of hydrogen bonds of carboxylic groups of the thiacalix[4]arene so that the relief of the surface becomes more flattened. The following interaction with α-thrombin changes the surface layer to a much lower extent [56].

Analytical characteristics of the aptasensors are compared with those obtained by traditional approaches in Table 1. One could see the use of 3D net of redox mediators made it possible to decrease the LOD tenfold against similar aptasensors based on electropolymerized phenothiazine dyes even though latter coatings were additionally modified to reach maximal response. Thus, polymeric Methylene blue (MB) layer was obtained onto multiwalled carbon nanotubes for better electric wiring and poly(Methylene green) layer was synthesized in the presence of DNA as template. Its acidic removal and substitution with aptamer improved their electrostatic interactions. In all the cases target interaction with thrombin resulted in suppression of the polymer redox activity recorded by potential cycling. Nevertheless, the appropriate changes on voltammograms were much smaller than those observed in case of the redox labels attached to the macrocycle core. As a result, the α-thrombin quantification was achieved only with EIS parameters (increase of the charge transfer resistance).

The importance of steric factors in the aptasensor signal generation was confirmed by the influence of the thiacalix[4]arene configuration on the aptasensor performance [56]. In the voltammetric mode, partial cone configuration showed minimal LOD (0.3 nM against 0.8 and 1.0 for cone and 1,3-alternate). In the EIS measurement, best results were obtained with 1,3-alternate (LOD 0.05 vs. 0.5 and 0.3 for cone and partial cone, respectively). Indeed, the EIS measurements detect density of the packing in the surface layer caused by the target interaction. In these conditions, symmetric 1,3-alternate has taken advantage over other configurations. In voltammteric mode, steric factors determine mostly distribution of the aptamer and the NR molecules around the macrocycle core. In partial cone configuration, they become closer to each other in the final product and hence the NR electrochemistry appeared more sensitive to the α-thrombin binding.

All the aptasensors described were tested on spiked samples of blood serum and standard albumin solutions. No significant influence was found in both cases.

3.3. Mycotoxins determination

Mycotoxins are a group of secondary metabolites of fungi exerting serious toxic effect on vertebrates and other living beings. Most dangerous mycotoxins are related to Aspergillus, Fusarium, and Penicillium spp. [59]. Aflatoxins occur in nuts, cereals and rice. The maximum permitted amount of aflatoxins established by European Community legislation is 4 μg/kg (2 μg/kg of aflatoxin B₁) in groundnuts, nuts, dried fruits and processed products for direct human consumption [60]. Ochratoxin A is produced by Aspergillus ochraceus and Penicillium verrucosum sp. mainly after harvesting on insufficiently dried cereal. WHO established maximal permissible OTA content of 5 μg/kg in cereals and 3 μg/kg in cereal products [61].

The determination of aflatoxin B₁ was performed using previously described scheme of the surface layer assembling [62]. The aminated aptamer against aflatoxin B₁ with terminal amino group and C₁₂ linker H₂N-C₁₂-5′-GTT GGG CAC GTG TTG TCT TGT GTC TCG TGC CCT TCG CTA GGC CCA CA-3′ was covalently attached to the polycarboxylated thiacalix[4]arene in partial cone conformation by carbodiimide binding together with the NR molecules. The biospecific components were deposited onto the poly(NR) layer obtained by electropolymerization. Resulting redox activity of the NR depended on the aflatoxin binding that suppressed the cathodic peak current similarly to that in case of the α-thrombin aptasensor. The dynamic concentration was 0.1–100 nM and LOD 0.1 nM. The EIS measurements decreased the LOD to 0.05 nM. These values are comparable or below those obtained for fluorescent immuno- and aptamer assay. It should be mentioned that the aflatoxin molecule is sufficiently smaller and appropriate aptamer bigger than those used in above mentioned α-thrombin aptasensors. To some extent, this results in higher distortion of the shape of the NR voltammogram obtained with the covalently attached NR against that recorded from its aqueous solution. Implementation of rather small aflatoxin molecule in the surface layer would less affect the electron exchange parameters that bulky thrombin molecules. Nevertheless, the current shift was observed in a similar concentration range in both cases (α-thrombin and aflatoxin B₁). The additional effect of the analyte on hydrophobic-hydrophilic balance of the surface layer can be taken into account. Hydrophobic molecules being entrapped in the surface layer would prevent access of small and highly charged ferricyanide ions used as redox probe in EIS measurements.

Similar behavior was established for the aptasensor toward ochratoxin A where thiolated aptamer molecules were attached to silver nanoparticles as redox labels [63]. The particles were obtained by chemical reduction of AgNO₃ solution with tetrasubstituted thiacalix[4]arene bearing catechol units. After reduction, appropriate quinone products are placed on the metal surface preventing aggregation of the nanoparticles and stabilizing them in solution and on the electrode surface. The scheme of the surface layer assembling and signal generation is presented in Fig. 3. Ochratoxin provokes switching the aptamer conformation from linear one to G-quadruplex. This makes the surface layer denser and decreases the rate of ferricyanide transfer in the EIS measurements. The LOD achieved (0.05 nM) was lower than that obtained with the same aptamer chemisorbed on bare golden electrode [64].

Fig. 3:

The formation of 3D net from silver nanoparticles obtained by chemical reduction of Ag⁺ ions with thiacalix[4]arene bearing catechol fragments followed by covalent binding of thiolated aptamer and ochratoxin A (OTA) binding due to conformational switch of aptamer.

3.4. Pillar[5]arene based aptasensors

As was mentioned above, the structure and size of the pillar[5]arene molecule is very similar to that of calix[4]arene except higher number of terminal functional groups and bigger flexibility of the macrocyclic core. This offer goop opportunities for the development of similar 3D layers with distributed redox probes on the pillar[5]arene platform. This idea was proved in the design of aptasensor against cytochrome c based on its specific binding with the aptamer 5′-NH₂-CCG TGT CTG GGG CCG ACC GGC GCA TTG GGT ACG TTG TTG C-3′. Cytochrome c is an electron-carrying mitochondrial protein that is often considered as a model for the investigation of redox paths in biological membranes [65]. The release of mitochondrial cytochrome c indicates the cell apoptosis and can be used for screening efficiency of potential anti-cancer drugs [66]. Decrease in the cytochrome c redox activity detects anti-respiratory poisons [67]. The formation of the biospecific layer was performed in a manner previously described in detail for the α-thrombin aptasensors. The decacarboxylated pillar[5]arene (see the structure in Fig. 1) was electrostatically accumulated on the poly(NR) film. After that, monomeric NR and aminated aptamer were covalently attached to the terminal groups by carbodiimide binding (Fig. 4).

Fig. 4:

Biospecific layer assembling and target analyte binding with the aptasensor based on polycarboxylated pillar[5]arene bearing NR and aptamer toward cytochrome c.

The specific interaction with cytochrome c was monitored by voltammetric signal of NR and by EIS parameters [68]. The mechanism of the signal and the relation between the electrochemical response and analyte binding were indirectly confirmed by similarity in the aptamer-cytochrome c association constant (2.2 nM) and the appropriate I₅₀ (1–10 nM) values. It was also interesting to see if the performance of the aptasensor depended on the sequence of the surface layer formation. For this purpose, covalent attachment of the NR molecules to pillar[5]arene was performed either prior to its deposition on the poly(NR) layer or by the treatment of the macrocyclic carriers already accumulated in the sensor interface. Neither charge transfer resistance and nor LOD altered significantly in these two protocols. Contrary to that, the nominal ratio of reactants bonded on pillar[5]arene core and absolute concentration of the aptamer affect the analytical characteristics of the cytochrome c determination to some extent. In optimal conditions, the aptasensor described makes it possible to detect down to 0.02 nM cytochrome c. The signal does not interfere with bovine serum albumin and some other components of biological fluids. The aptasensor was tested in spiked samples of artificial blood plasma.

3.5. Factors affecting the performance of the aptasensors based on 3D mediator net

The approach described above for the DNA aptasensors has undisputable advantages, i.e. simple assembling, universal platform for various analytes detection, no necessity in deep transformation of the targets/aptamers prior to tier use, etc. Nevertheless, some limitations related to the choice of redox labels and underlying film exist.

Redox label. All the experiments described are based on the application of the NR, a dye often applied in the electrochemical biosensors due to excellent mediation properties [69]. From the point of view of biosensor assembling, it is important that the NR contains primary amino group consumed in carbodiimide binding. However, substitution of the NR with thionine exerting similar redox properties and bearing two primary amino groups resulted in almost full loss of the response. Probably, this can be explained by lower size of the redox label molecule that retains ability to electron exchange after binding a target analyte. The importance of steric factors was shown also for the NR based biosensors: the use of the thiacalixarene derivative with carboxylic groups attached via shorter linker than that shown in Fig. 1 resulted in about threefold lower sensitivity of the α-thrombin detection [56].

Underlying film. The electron flux near the electrode interface is determined by the gradient of the potential. All the experiments were performed in the presence of dissolved oxygen as a natural electron acceptor. Nevertheless, in some cases, the recovery of the biosensor between the potential runs limited the total duration of the experiments. The peak currents were restored for 3–5 min in open circuit mode. The addition of ferricyanide ions to the solution accelerated the process and improved the measurement-to-measurement repeatability of the response, which reached about 1.3%–1.5% (10 runs). The substitution of the poly(NR) film with some other polymerized layers decreased the gradient of the potential because the formal redox potential of both monomeric and polymeric NR forms is rather low (about −450 mV for thrombin aptasensors). Polyphenothiazine derivatives show the formal redox potential of about 250–400 mV. As a result, the use of poly(Methylene green) layer decreased the sensitivity of the thrombin determination against that of the aptasensor based on poly(NR) (see Table 1). Polythionine in combination with thiacalix[4]arene bearing thionine groups showed no response to aptamer-analyte interactions.

Carbon black is another support that might be useful in such systems. Indeed, the combination of carbon black with pillar[5]arene showed reversible electrochemical behavior and was successfully applied in the assembly of various biosensors [70], [71]. The deposition of decacarboxylated pillar[5]arene bearing aptamer toward cytochrome c and the NR made it possible to obtain the response toward the analyte, but the appropriate LOD value was significantly higher than that obtained with poly(NR) underlying layer (0.02 nM and 1 nM) [72]. This might be due to higher roughness of the carbon black against polymeric film and lower regularity of the bioreceptor layer. As a result, insulating layer of the analyte does not fully prevent electrochemical reactions between the electrode and redox labels.

4.1. E-Sensors

Starting from first publications, the conditions for reliable sensitive detection of “signal-on” scheme were specified on the example of the MB label. First, it was found that the maximal difference in the current displaying the electron exchange in closed and opened forms of a pinhole DNA probe corresponded to the maximal density of the DNA probes immobilized on the electrode surface. High density (about 10¹² probe molecules per cm²) corresponds to the regular monolayer structure of the surface layer and prevents electron exchange between neighboring redox labels that compensates for decrease in the number of labels directly contacted with the electrode surface [76]. The mechanism when the current depends on the frequency of such collisions is considered as most probable against that assuming tunneling the electrons along the ds-DNA part of the hybridization product [77]. As a result, the relative decrease of the current was more pronounced for the high density even though the efficiency of hybridization step was higher for medium density of the surface coverage [76]. Similar effect was found for the size of the loop: the bigger the number of nucleotides in the non-hybridized part of the pinhole sequence the higher sensitivity of the hybridization detection. Direct comparison of the results obtained with MB and ferrocene indicated small advantages of ferrocene in sensitivity of the signal while MB showed higher stability and repeatability of the response in a series of measurements with the same probe [78]. The recovery of the E-sensors is easily attained by gentle heating the solution and washing the sequences released from pinhole probes.

The comparison of various redox “reports” has been prolonged later on [79]. It was confirmed that MB exhibited high stability of repeated measurements performed even in blood. Other samples used for the E-sensor testing included urine, saliva, soil and beer [80]. In all the samples mentioned the E-sensor demonstrated excellent repeatability of the measurement results performed with the same DNA probe and minimal matrix influence on the response. Other labels, e.g. Nile blue, anthraquinone and ferrocene derivative suffered from inherent instability of Au–S bonds observed at the potentials below −0.6 V and above 0.4 V [79].

The following improvement of the E-sensor performance with the MB label is achieved by reduction of the number of nucleotides in the linker between the loop and electrode from 30–50 to about 20 [81] and by introduction of anionic carboxylate group in the dye moiety [82]. The first one is explained by diffusional control of the electron transfer reaction in unfolded stem-loop probe and the second one by prevention of the electrostatic interactions of the positively charged MB molecules with phosphate residues at the minor grooves of the ds-DNA helix. Optimized E-sensors were validated on the detection of cancer biomarker TP53 gene and investigation of single nucleotide polymorphism.

Concerning other labels, the sensitivity of the E-sensor in the papilloma virus detection was significantly improved by the attachment of four ferrocene units to terminal groups of the probe [83]. The relative decay of the signal recorded by DPV also decreased with reduction of the complementary sequence of the loop from 50 to 30 nucleotides. A similar effect was reached by application of PAMAM G4 dendrimer bearing ferrocene units [84]. The modifiers were placed on the electrode covered with multiwalled carbon nanotubes and polypyrrole layer. The biosensor was validated on detection of the DNA sequence related to rpoB gene of Mycobacterium tuberculosis in the real PCR samples.

The complication of DNA probe to be able to form two interconnected loops increased sensitivity of target sequence detection [85]. In this biosensor, MB attached to terminus of the sequence is liberated in the target interaction while in closed form it is fixed far from the electrode by so called “pseudoknot” structure (Fig. 5b). Optimal length of DNA duplex fragments between two loops and density of the DNA probe on the electrode were specified. The E-sensor was utilized for direct measurements in such a complex media as blood serum. Later on, similar pseudoknot aptamer labeled with the MB was successfully utilized for IgE determination with LOD of 60 pM. The aptamer was immobilized on planar screen-printed electrode and the signal was recorded using square-wave voltammetry [86].

Both ferrocene and MB labels were utilized in the dual mode E-sensor. The target sequence reacts with the loop labeled with ferrocene near the electrode together with the primer bearing MB on the opposite end of the sequence. In such interactions, the ferrocene oxidations current increased and those of MB decreased proportionally to the target DNA concentration. The use of strand displacement by polymerase reaction releases the target sequence and hence increases sensitivity of the analysis (LOD 28 fM) [87]. Besides amplification effect, the use of dual label mode can improve the robustness of the E-sensor measurement [88]. For this reason, one label (MB) is attached to the terminal group of the probe while second one (ferrocene) is introduced in the sequence near the electrode. In target interaction and followed unfolding of the loop the MB signal decreases while that of ferrocene remains constant. The difference in the label signals is much more reliable in quantification of the target sequence effect.

Similar amplification is described in [89]. Biotinylated stem-loop probe is treated with a target sequence and strand displacement polymerase to increase the number of probes unfolded. Then magnetic nanoparticles modified with streptavidin are added and finally the alkaline phosphatase is immobilized on them. The DPV signal recorded is related to the oxidation of p-aminophenol, the product of enzymatic hydrolysis of p-aminophenylphosphate. The LOD of 0.1 fM was reached for 26-mer target DNA sequence. Besides alkaline phosphatase, horseradish peroxidase was applied in similar sensor attached to the probe-sequence duplex via avidin-biotin binding [90]. The current of benzoquinone reduction yielded in enzymatic oxidation of hydroquinone was measured within the analyte concentration from 0.1 to 1000 nM. Reusability of the biosensor was achieved by its treatment with 8 M urea solution. HRP modification via digoxigenin – anti-digoxigenin coupling was applied in the E-sensor based on biotinylated stem-loop probe immobilized on streptavidin coated electrode [91]. The reaction was monitored by enzymatic oxidation of tetramethylbenzidine and its cathodic recovery. The E-sensor was used for determination of the PCR amplicons from the uidA gene of E. coli (250 base pairs) with lower quantification limit of 30 pg.

Reversed mechanism of the signal generation was realized in [92]. In this work, capture probe was immobilized together with unfolded stem-loop probe in a manner providing their hybridization and preventing folding. In the presence of target sequence, the displacement takes place and the stem-loop probe is liberated from the complex with capture DNA and converts into the loop form. As a result, terminal MB groups are fixed on the distance of direct electron exchange with the electrode and its current increases with the target sequence content (LOD 1 pM). Besides other parameters, the length of the duplex formed between the stem-loop probe and capture probe is important. Short length (4–8 base pairs) does not provide rigidity of the construct while a long distance (12 or 14 base pairs) complicates the displacement reaction. Similar effect can be obtained if labeled short DNA sequence is attached to that released from the complex with capture probe (LOD 10 fM) [93].

Enantiomeric assay has been performed with E-sensor based on short ss-DNA probe consisting a human telomeric fragment rich with guanine bases [94]. The probe was immobilized by 3′ terminal SH group to Au electrode and contained ferrocene label at 5′ end. In Li+ solution, the sequence was unfolded so that ferrocene unit could be involved in electron exchange with the electrode. The only one helical enantiomer [Ni₂L₃]⁴⁺consisting of three bis(pyridylimine) ligands L (5) stimulated formation of quadruplex formation that forced ferrocene unit to leave electrode surface. The signal recorded by alternating AC voltammetry decreased fivefold.

Contrary to that, similar experiment with cysteine rich fragment resulted in de crease of the current by 30% only.

4.2. E-AB and TRE sensors

The mechanism of signal formation for aptamer based biosensors does not differ dramatically from that already considered for E-sensors. Mostly, two alternative mechanisms affecting label signal are considered, i.e. changes in collision intensity and aptamer folding. In the first case, implementation of target analytes suppresses the electron transfer with the label by limitation of the flexibility of the complex formed. In general, such mechanism is very similar to that considered in detail in Section “Co-immobilization of redox labels and bioreceptors on multifunctional supramolecular carriers” for the aptasensors based on multifunctional supramolecular carriers. Second mechanism which probably dominates among the majority of the E-AB sensors takes into account folding the aptamer initiated by interaction with the label and removal of the label from the proximity of the electrode [95]. The analytical characteristics, biosensor assembling and signal generation details are summarized in Table 2.

One could see, almost all of the E-AB sensors are based on thiolated aptamers that are immobilized on Au electrode or Au nanoparticles obtained prior to immobilization step by electrodeposition. The Au-S binding provides site-specific immobilization that leaves free for conformation changes about the whole aptamer/probe molecule. The monitoring of intrinsic redox-activity of juglone is the only example of alternative approach to this step of biosensor assembling [107].

First E-sensors utilized “signal-off” protocol of signal measurement. Nevertheless, the opposite changes of the current (“signal-on” mode) are more attractive from the point of view of measurement accuracy and reproducibility. False results of “signal-off” protocol can be also caused by partial distortion of the surface layer and removal of nanoparticles with attached aptamers. The “signal-on” mode is reached by liberation of labeled DNA sequences attached to the electrode, which could either restore their flexibility or transfer in a closed pinhole form. In such schemes (see Fig. 5c, d) ds-DNA formed with an aptamer and labeled auxiliary sequence are used and the addition of an analyte results in dissociation of hybridized parts of the sequences. In many E-AB sensors, both “signal-on“ and “signal-off” modes are used, commonly with different labels, e.g. ferrocene and MB. Such an approach called as ratiometric improves the robustness of the results because the changes in the signal of differently labeled reactants are shifted in the sample tested in opposite directions. Besides, the use of two labels makes it possible to detect simultaneously two different analytes on the same support (electrode) [105]. Most of the aptasensors are devoted to multiple use and the recovery of initial state as well as its stability in such manipulations should be always checked by appropriate measurements. In rather simple E-AB sensors with a single pinhole sequence, gentle heating of the sensor can attain the recovery after the measurement. In case of displacement schemes, guanidine chloride, urea and some other denaturing agents. Contrary to conventional DNA- and immunosensors, enzymes are rarely used as labels due to complication of the protocol and inconformity to idea of reagent-free detection with E- (E-AB) sensors. Nevertheless, alkaline phosphatase and horseradish peroxidase have been described. Their attachment to the complexes is mostly achieved by avidin (streptavidin) – biotin binding.

Among other amplification approaches, specific biochemical paths affecting length and nature of DNA sequences showed advantages in reaching sub-nanomolar LOD values [102]. Strand displacement polymerase and exonuclease liberate target sequences in increase the number of folded aptamers on the surface or increase the distance of electron transfer in “signal-on” and “signal-off” modes of measurement.

Most of the examples presented in Table 2, describe validation of the E-AB sensors in reals samples. In some cases, the devices are adapted for mostly simple and inexpensive application in the framework of point-of-care diagnostics. Thus, paper based electrodes were described for the detection of thrombin, serine protease known as blood clotting factor [101]. Continuous microfluidic system with four planar electrodes was designed for cocaine detection [97]. Other sensors utilize screen-printed electrodes modified with Au nanoparticles as cheaper alternative to Au disk electrodes [99], [103].

Besides pinhole DNA- and aptasensors, conformational changes and limitation of collisions are reached in hybrid materials including ss-DNA coupled with mesoporous silica nanoparticles. They can be considered as nanocontainers filled with appropriate indicators and capped with DNA sequences. Such an approach was called as target responsive encapsulation (TRE). Conformational changes caused by hybridization of DNA probe or folding aptamers result in appropriate shifts of the signal referred to the availability of the pore content or their permeability for the oligonucleotides involved in reactions described. The universal character of the strategy proposed was demonstrated by determination of thrombin, ATP and hybridization of 26-mer DNA sequence related to the Alzheimer’s disease [108]. The signal of the MB released was measured with DPV. Simultaneous use of two labels (MB and ferrocene) with appropriate sequences sensitive to Hg²⁺ and Ag⁺ made it possible to detect both metals able to form complexes with two thymine (T) residues (T-Hg²⁺-T as example) as example. On the same basis, logic gates mimicking Boolean logic functions were described.

Mutated apolipoprotein E gene associated with Alzheimer’s disease was detected with “signal-off” DNA sensor based on novel hybrid material, graphene – mesoporous silica particles conjugated with ferrocene-carboxylic acid [109]. Prior to contact with target sequence, the support of the sensor was saturated with the MB molecules that are caged in the pores as nanocontainers and prevented from release by ds-DNA immobilized on the surface. The reaction with target DNA sequence released redox indicator from the pores and increases appropriate signal on DPV voltammogram. In parallel, the ferrocene signal was recorded to avoid rude mistakes related to the deterioration of the nanoparticles and their removal from the transducer. The ratio of the label peak currents linearly depended on the logarithm of the target sequence concentration from 0.01 pM to 10 nM (LOD 10 fM).