Triplex DNA-based Bioanalytical Platform for Highly Sensitive Homogeneous Electrochemical Detection of Melamine

Melamine detection has attracted much attention since the discovery of the damage of melamine to human health. Herein, we have developed a sensitive homogeneous electroanalytical platform for melamine detection, which is relied on the formation of triplex molecular beacon integrated with exonuclease III (Exo III)-mediated signal amplification. The formation of triplex molecular beacon was triggered by the recognition and incorporation of melamine to the abasic (AP) site contained in the triplex stem. The stem of the triplex molecular beacon was designed to have a protruding double-strand DNA, which can be recognized and hydrolyzed by Exo III for releasing methylene blue (MB)-labeled mononucleotide. These released MB molecules exhibit high diffusivity toward indium tin oxide electrode with negative charge, thus producing a significantly increased electrochemical response. Taking advantages of the high binding affinity of the DNA triplex structure containing AP sites towards melamine and the unique features of Exo III, this sensing platform is capable for sensitive and selective melamine assay with a detection limit as low as 8.7 nM. Furthermore, this strategy shows good applicability for melamine assay in real samples. Therefore, this strategy broadens the application of triplex DNA and presents a new method for sensitive detection of melamine.

Recently, triplex DNA has attracted much attention because of its indispensable in many biological processes, such as gene translation, gene expression, DNA transcription, and replication 31 . The triplex structure is constructed through binding the third DNA strand to the double-helical DNA via Hoogstern hydrogen bonds [32][33][34] . The formation of triplex structure can be regulated by external triggers such as target molecules, pH, and metal ions [35][36][37] , making it a potentially powerful tool for numerous applications in biological chemistry and medicine 38,39 . For example, triplex DNA has been utilized as recognition motifs in the design of various biosensing platforms and DNA nanomachines through the target-induced triplex DNA formation and dissociation 40,41 . Apart from the sequence-specific recognition of double-stranded DNA, the incorporation of the third DNA strand into the stem of molecular beacons to form hairpin-like DNA structure is an alternative strategy for the construction of biosensors [42][43][44] . For instance, triplex molecular beacons have been utilized to design melamine aptamer by introducing an abasic (AP) site into the triplex stem 43,44 . In this aptamer, the integrated AP site could provide a hydrophobic vacancy for binding melamine through hydrogen bonding interaction between thymine and melamine. Based on this aptamer, this group has developed selective fluorescence biosensors for melamine based on Mg 2+ -dependent DNAzyme and silver nanoclusters, respectively 43,44 . Most recently, Wang and Lin et al. have constructed a sensitive and selective electrochemical biosensor for melamine determination by designing AP site into triplex DNA, in which four DNA strands can form a T-shaped DNA nanostructure in the presence of melamine. This biosensor exhibits excellent performance for melamine detection and can be used for testing the migration of melamine from melamine bowl 45 .
In this work, we present a homogeneous electroanalytical strategy for sensitive assay of melamine based on the formation of triplex molecular beacon and exonuclease III (Exo III)-assisted signal molecules release. This approach employs a hairpin DNA (HP) and a 3′-methylene blue (MB) labeled DNA (MB-DNA) containing AP sites as the melamine recognition part, Exo III as a tool enzyme, and an indium tin oxide (ITO) electrode with negative charge as the working electrode. In the presence of melamine, the formation of triplex molecular beacon occurs and then MB molecules are released from MB-DNA by Exo III, resulting in a significantly increased electrochemical signal of MB. Taking advantages of the high selectivity of the DNA triplex structure containing AP sites towards melamine and the unique features of Exo III, this proposed strategy is capable of detecting melamine with high sensitivity and selectively.

Results and Discussion
Design of Melamine Detection Strategy. Figure 1 schematically illustrates the strategy for sensitive homogeneous electroanalysis of melamine, which is based on a simple system of HP, MB-DNA, and Exo III. HP is an ingeniously designed hairpin DNA, which consists of four sequential domains, namely I, II, I′, and II′. The base sequence of domain I is complementary to that of domain I′, while domains II and II′ are short poly(dT) sequences. So, HP alone has a hairpin structure with a protruded fragment (domain II′). MB-DNA is a 3′-methylene blue (MB) labeled DNA, which contains two sequential domains: a poly(dA) sequence with two AP sites (domain II″) and a same base sequence to domain I′. Exo III, which displays sequence dependence, can catalyze the stepwise removal of 3′ mononucleotides from blunt or recessed 3′-termini of double-stranded DNA [46][47][48][49] . In the absence of melamine, MB-DNA cannot reconfigure HP because of the limited stability of triplex stem with AP sites. Thus, MB-DNA maintains single strand structure and gives a very weak electrochemical signal of MB, because both negatively charged MB-DNA and ITO electrode have strong electrostatic repulsion. However, in the presence of melamine, melamine can locate in the AP sites of the triplex DNA structure and form a stable T-melamine-T structure through hydrogen bonding interaction 43,44 . Thus, MB-DNA can hybridize with HP to form a molecular beacon with a triplex stem (domains II, II′, and II″) and a protruding double strand (domains I and I′). This duplex with blunt 3′ termini is then recognized and cleaved by Exo III, leading to the digestion of MB-DNA in the protruding duplex. This digestion process releases MB-labeled mononucleotide, melamine, and HP. The released melamine and HP can hybridize with another MB-DNA and trigger a new digestion process. The released MB-labeled mononucleotide, which has less negative charge than that of MB-DNA, possesses higher diffusivity toward ITO electrode owing to the reduced electrostatic repulsion. As a result, a remarkably increased electrochemical signal would be obtained. Therefore, this strategy realized the sensitive detection of melamine in a "signal-on" mode.

Feasibility Study.
To verify the feasibility of melamine assay, a series of control experiments were preformed and their electrochemical responses were recorded. As shown in Fig. 2, no electrochemical signal was detected from the solution of melamine (curve a) due to the absence of electrochemical signal molecule. In the presence of HP, MB-DNA, and melamine, a negligible DPV current of MB molecule was observed (curve b), and the intensity of the current was almost the same as that obtained from MB-DNA solution (curve c), indicating that the negatively charged MB-DNA and the triplex DNA structure were hard to reach the surface of ITO electrode owing to the strong electrostatic repulsion between them. Additionally, when only Exo III was mixed with MB-DNA, the DPV current was almost unchanged (curve d), because the single strand MB-DNA can resist the digestion of Exo III. Nevertheless, because of the existence of equilibrium between different molecular conformations of HP, small portion of HP may be in a single-stranded conformation, which would hybridize with MB-DNA and produce blunt 3′ termini, causing the cleavage of some MB-DNA by Exo III. As a result, addition of HP and Exo III to MB-DNA solution led to a slightly increased DPV current (curve e). However, a remarkably increased DPV current of MB was obtained (curve f) when melamine was added into the reaction system, indicating the successful hybridization between HP and MB-DNA, as well as the subsequent digestion of MB-DNA by Exo III.
Optimization of Assay Conditions. In this melamine sensing strategy, the electrochemical signal was generated by the released MB-labeled mononucleotide, which originated from MB-DNA. To ensure the high sensitivity, the amounts of MB-DNA should be sufficient, thus 1 μM MB-DNA was added into the reaction solution during the following experiments. Furthermore, as the formation of Hoogsteen TA·T triplex structure is sensitive to pH value, the influence of pH was optimized by analyzing the electrochemical signals in the presence of 50 μM melamine. As shown in Fig. 3A, the intensity of DPV peak current obviously increased when the pH value increased from 6.8 to 7.2, but further increasing the pH value resulted in dramatically decreased peak intensities. The highest DPV peak current was acquired at pH = 7.2, which was chosen as the optimal pH value. To achieve the best analytical performance, the concentration of Exo III, the reaction time of Exo III cleavage, and melamine incorporation were also optimized, respectively. Figure 3B reveals that a higher DPV signal was obtained when a higher concentration of Exo III was adopted. But the DPV peak current increased slowly when the concentration was higher than 0.75 U/μL. Thus, 0.75 U/μL was selected as the optimal Exo III concentration. Figure 3C shows the effect of reaction time of Exo III cleavage on this melamine sensing strategy. Obviously, the DPV peak currents increased remarkably as the time of Exo III cleavage extended to 60 min. Afterward, the DPV peak currents did not change much with increased cleavage time, suggesting that the release of MB-labeled mononucleotide from the MB-DNA was almost finished at enzymatic reaction time of 60 min. Similarly, obviously increased DPV peak currents were observed with the extension of the melamine incorporation time and almost reach the saturation value at 5 h (Fig. 3D). Therefore, 60 min and 5 h were adopted as the optimal Exo III cleavage time and melamine incorporation time, respectively. Sensitivity of Melamine Assay. The performance of this approach for melamine assay was investigated by analyzing the electrochemical response of the sensing system towards different concentrations of melamine. As shown in Fig. 4A, the DPV current of MB molecule was sensitive to melamine concentration and gradually increased with the increase of melamine concentration. The relationship between melamine concentration and the intensity of DPV peak current is presented in Fig. 4B. Obviously, melamine concentration ranging from 0 to 1000 μM could be directly measured by this sensing strategy. Furthermore, the inset of Fig. 4B shows a good linear relationship between the intensity of peak current and the logarithm of melamine concentration ranging from 50 nM to 500 μM. The correlation equation was determined to be i p = 118.7 log C melamine + 72.0 (R 2 = 0.988),  where, i p is the intensity of peak current (nA), C melamine is the melamine concentration (10 −8 M). The detection limit towards melamine was estimated to be 8.7 nM (1.1 ppb), which is lower than that of the above mentioned biosensors (Table 1) and the maximum level permitted by China (1 ppm for infant formulas), USA and European Union (2.5 ppm), and India (1 ppm for infant formulas), respectively 50, 51 . Selectivity of Melamine Analysis. The selectivity of the proposed strategy for melamine assay was evaluated by testing the electrochemical response of this sensing system toward several potential interfering substances, such as cytidine, guanosine, methionine, glycine, lactose, vitamin C, vitamin B 2 , Zn 2+ , and K + (Fig. 5). Compared with the significant increase of the DPV peak current upon melamine addition, there were only weak electrochemical signals observed for most other species. Therefore, this homogeneous electroanalytical strategy exhibited excellent selectivity for melamine detection and has great potential for real sample analysis application.

Real Sample Analysis.
To investigate the applicability of this electroanalysis method for real sample assay, melamine in animal feed was tested by using this strategy. As melamine concentration in the animal feed was lower than the detection limit of this method, standard addition experiments were preformed. The electrochemical responses of the spiked samples were analyzed and the results are listed in Table 2. Analytical results showed that the melamine concentrations in the spiked real samples were statistically close to the concentration of melamine added and their recoveries are ranging from 90% to 97%. Although these recoveries are lower than 100%, they are still in the recovery range permitted by Chinese National Standards (60~120%, GB/T27404-2008). Thus, random deviation of measurement is speculated to be the main reason for the inconsistency between melamine concentration detected and added. Moreover, the complex matrix such as animal feed has no evident influence on this homogeneous electroanalytical strategy, suggesting the good anti-interference capability and the potential application of this method for the analysis of melamine in real samples.

Conclusions
In summary, a homogeneous electroanalysis platform for sensitive melamine assay was developed based on melamine-triggered formation of triplex molecular beacon and Exo III-assisted signal amplification. Taking advantage of the high binding affinity of AP site-containing triplex structure towards melamine and the unique properties of Exo III, numerous MB-labeled mononucleotides were released, which results in significantly increased electrochemical signals. All results consistently confirmed that the new electrochemical platform was applicable for melamine detection with high sensitivity and selectively. Additionally, this strategy exhibits good potential for the analysis of melamine in real samples. Thus, this strategy broadens the application of triplex DNA and presents a new strategy for the sensitive and selective detection of melamine.  -GTC TGT TTT TTT TCT CAG ACT CTT TTT TTT-3′ MB-DNA: 5′-AGA AA X AA X ACA GAC-MB-3′ X denotes the AP site of spacer C3.

Methods
The MB-DNA and HP were directly used and diluted in 50 mM Tris-HCl (pH = 7.2, containing 100 mM NaCl, 5 mM MgCl 2 ) to give the stock solution of 10 µM.
Melamine Assay. The melamine assay was performed in 50 µL Tris-HCl reaction buffer containing 1 μM HP, 1 µM MB-DNA, and target melamine with different concentrations. This reaction solution was incubated at 25 °C for 5 h, followed by the addition of 37.5 U Exo III and incubation at 37 °C for 60 min before the electrochemical measurements.
Sensing Melamine in Pig Feed. Pig feed, a mixed powder mainly consists of corn, soybean, bran, fish meal, minerals (such as ZnSO 4 , KI, FeSO 4 , CuSO 4 ), and vitamins (such as vitamin A/B 2 /C/D/E/K), was chosen as the real sample matrix to evaluate the practical application of the as-proposed approach for melamine detection. Firstly, 1.0 g pig feed was mixed with 20 mL Tris-HCl reaction buffer in a centrifuge tube, and the mixture was ultrasonicated and shaken for 10 min, respectively. After repeating this process for two times, the mixture was centrifugated at 15,000 rpm for 10 min, then the supernatant was collected and further filtered with 0.45 μm filter before melamine assay. Finally, different concentrations of melamine were added into the feed samples to prepare the melamine spiked samples. Each concentration was repeated for 5 times, and the result was an average of 5 measurements.
Electrochemical Measurement. Differential pulse voltammetric (DPV) measurements were carried out on an Autolab Electrochemical Workstation (Metrohm, Switzerland) with a conventional three-electrode system: an ITO electrode as the working electrode, a platinum wire as the counter electrode, and an Ag/AgCl electrode as the reference electrode. It should be noted that ITO electrode with negative charge was obtained by using a previously reported pretreatment method 29 .  Table 2. Recovery data for feed samples added with different concentrations of melamine (n = 5).