Introduction

Cardiovascular disease (CVD) is the leading cause of death in most countries around the world, which can eventually lead to acute heart failure (HF) and possible death [1, 2]. Brain natriuretic peptide (BNP), as a biologically active natural hormone, is a quantitative marker of HF owing to allow risk stratification for emergency department patients for medical judgment [3,4,5]. Therefore, it is highly accurate for diagnosing HF. Some methods for detecting BNP, such as conventional spectroscopy [6], surface plasmon resonance (SPR) [7, 8], and electrochemical method [9], have been developed to provide various possibilities for clinical diagnosis of CVD.

The enzyme-linked immunosorbent assay (ELISA) based on antigen-antibody interaction has been widely used in various analytical fields since its development to replace radioimmunoassay, which is not only highly specific but also capable of detecting ultralow concentrations of antigen [10, 11]. However, relying solely on an antibody to bind antigen to detect a target has certain downsides. In addition to variations in antibodies produced in different batches, it may be challenging to result in specific monoclonal antibodies, particularly against non-immunogenic molecules, which are prone to false positives in the assay [12]. Therefore, in a variety of different options, it is an ideal approach to combine the modular aptamer with a specific antibody instead of the target capture agent. On the one hand, DNA aptamers have high intrinsic chemical stability and can bind to recognition molecules with high specificity and high affinity [13, 14]. On the other hand, aptamers are easily modified and labeled by various reporter molecules and functional groups, so that the interface or nanoparticles can be easily functionalized [15, 16], improving the availability of the detection means. The dual recognition of aptamers and antibodies to targets provides a more direct and effective mode [7, 17].

Taking advantage of the distinctive physical and chemical properties of nanoparticles, many studies on functionalized nanoparticles have been reported in biomolecular detection [18,19,20]. Among various types of nanoparticles, magnetic nanoparticles (MagNPs) have attracted more and more attention by using external static magnetic fields to effectively separate analytes in complex media [21, 22]. This exceptional advantage of MagNPs simplifies the cumbersome and complex traditional experimental approach compared with classical purification ways which are particularly complicated and time-consuming. However, most of the MagNPs exhibit weak electrical conductivity. Thus, the sensitivity is limited in electrochemical amperometric applications. The combination of gold nanoparticles (AuNPs) to MagNPs not only can ensure the rapid pre-enrichment, but also can improve the electrical conduction of poor conductors, thanks to magnetic properties of MagNPs and the superior conductivity of AuNPs. Therefore, magnetic gold nanocomposite can substantially improve the sensitivity of electrochemical sensing.

Previous studies have reported that the dissociation constant (Kd) value of the antibody against the antigen is 0.33 nM [23], and the Kd value of the aptamer to the target is 12.5 nM toward BNP [24], showing excellent individual binding ability. In this study, a triple recognition strategy based on target-induced magnetic gold nanocomposite was proposed. Two kinds of nanoparticles are used to construct electrochemical method. First, MagNPs modified with anti-BNP are combined with the target BNP to form MagNP-Ab@BNP. Then, MagNP-Ab@BNP is separated from the complex environment by a permanent magnet to avoid the degradation of BNP by neutral endopeptidase. Apt-MB, which is modified on AuNPs, specifically binds the C-terminal of BNP in the MagNP-Ab@BNP complex, and a large nanocomposite is formed by cascade conjugation in this process. After magnetic separation, the nanocomposite in solution can be captured and fixed on the surface of gold electrode (GE) by hybridization between Apt and cDNA. The contents of BNP are quantitatively determined by collecting the electrochemical signals of the MB molecules on AuNP-Apt-MB (Fig. 1). This method reveals high sensitivity and wide dynamic range of detection, demonstrating practical possibilities in clinical testing applications.

Fig. 1
figure 1

Schematic of triple recognition electrochemical immunoassay based on functional nanoparticles for BNP

Full size image

Experimental section

Reagents and apparatuses

N-Hydroxysulfosuccinimide (NHS), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), dithiothreitol (DTT), mercaptoundecanoic acid (MUA), bovine hemoglobin (BHb), myoglobin (Mb), ascorbic acid (AA), 6-mercapto-1-hexanol (MCH), sodium citrate, and hydrogen tetrachloroaurate trihydrate (HAuCl4·3H2O) were purchased from Sigma-Aldrich (Shanghai, China, http://www.aladdin-e.bioon.com.cn/). Fe3O4 magnetic nanoparticles obtained from AMI Life Science Inc. (AMO-Mag, Gimpo, Korea, http://amilifesciences.com/). BSA was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China, http://www.reagent.com.cn). BNP and anti-BNP were obtained from Sangon Biotech Co., Ltd. (Shanghai, China, https://www.sangon.com/). All of the above chemical reagents are of analytical grade. DNA sequences were synthesized by Sangon Biotech. Co., Ltd. (Shanghai, China, https://www.sangon.com/) and listed below. Apt-MB: 5′-HS-TTT TTT TAA ACG CTC AAA GGA CAG AGG GTG CGT AGG AAG GGT ATT CGA CAG GAG GCT CAC A-MB-3′; complementary DNA (cDNA): 5′-SH-TGT GAG CCT CCT GTC GAA TAC CCT TCC TAC GCA CCC TCT GTC CTT TGA GCG TTT AAA AAA A-3′. All the solutions are from deionized water (DW) of Millipore purified water system (> 18 mΩ cm, Millipore, USA).

Electrochemical measurements were performed using an Autolab PGSTAT128N system (Eco Chemise B.V., The Netherlands). The UV-vis spectra were recorded on a Shimadzu UV-2450 spectrophotometer. Z-potential and dynamic light scattering (DLS) were performed from Malvern Nano Zetasizer (MAL1202556). Transmission electron microscopic (TEM) image was recorded from a JEM-2010F microscope (Japan). Fourier transform infrared (FT-IR) spectra were generated by an FT-IR spectroscopy (Bruker 70, Germany) instrumentation.

Preparation of AuNP-apt-MB and MagNP-Mb

AuNPs were synthesized in a one-pot process. The colloidal solution of AuNP was synthesized by classical citrate reduction of HAuCl4·3H2O [25]. Specific experimental details and the modification of Apt-MB were presented in the Supporting Information. Details on the preparation of MagNPs and modification of antibodies were provided in the Supporting Information.

Preparation of functionalized nanoconjugate

Initially, 100 μL of MagNP-Ab was mixed with BNP solution in centrifuge tube and stirred slightly for 20 min at 37 °C. Then MagNPs were suspended in 100 μL phosphate-buffered saline (PBS) after washing. Next, 200 μL excess of the AuNP-Apt-MB was added to the tube and reacted for 30 min at 37 °C with gentle shaking. After washing with PBS for three times, the MagNP-Ab@BNP@AuNP-Apt-MB nanoconjugate was collected by a permanent magnet under the centrifuge tube and then resuspended in 100 μL PBS.

Electrode surface modification

First, the electrode was polished using fine-grained sandpaper (5000) to regenerate GE surface, and polished with 0.3 and 0.05 μm polishing powders. Then, the piranha solution (30% H2O2:H2SO4 = 1:3) was covered on the surface of GE for 5 min to eliminate organic and inorganic substances that might be adsorbed on the electrode surface. After the electrode was washed with DW, cyclic voltammetry (CV) activation (20 cycles, 0–1.6 V) was carried out in a 0.5-M H2SO4 solution to remove the oxide film on the surface of the electrode. Finally, the electrode was washed gently with DW and dried with nitrogen.

A total of 90 μL of DTT (1 mM) was mixed with 10 μL cDNA (1 μM) and reacted in the dark for 1 h, which converted disulfide bonds into sulfhydryl groups and facilitated the fixation of nucleic acids on GE surface. A total of 50 μL DNA were modified electrode surface for 2 h, and the unbound DNA was removed by washing with DW for three times. Next, the cDNA-functionalized GE was placed in 200 μL MCH (0.5 mM) solution for 1 h to obtain a regular array of DNA monolayer. MCH used here was a good blocking agent, which not only backfilled the Au substrate but also resisted non-specific binding. Then, 50 μL nanocomposite was used to modify the surface of GE for 30 min. At this point, all the reactants had been modified to the surface of the electrode.

Electrochemical measurements

Electrochemical measurements were performed with a three-electrode system (CE, SCE, and WE). Electrochemical measurement methods included CV, differential pulse voltammetry (DPV), and electrochemical impedance spectroscopy (EIS). CV and EIS of the biosensor were treated in 8 mL [Fe (CN)6]3−/4− solution. The scanning potential of CV was − 0.1 to 0.5 V, and the scanning rate was set to 100 mV/s. The EIS had a frequency range of 0.1 Hz to 100 kHz. DPV was carried out in PBS with a potential of − 0.6 to 0.1 V, a pulse amplitude of 50 mV, a pulse width of 50 ms, and a pulse period of 0.1 s. Each concentration measured was repeated three times, and the average was taken as the result.

Application to real samples

Human serum samples were collected from the First Affiliated Hospital of Zhengzhou University (Zhengzhou, China) and they used without any pre-treatment. Recovery experiments were performed at BNP serum samples spiked with 1 pg mL−1, 10 pg mL−1, 100 pg mL−1, 1 ng mL−1, and 10 ng mL−1.

Results and discussion

Characterization of materials

The formation of MagNPs and AuNPs were confirmed by TEM imaging and DLS (Fig. 2 a and b). The TEM image from Fig. 2a clearly shows that the size of the core-shell MagNPs is 320 nm, and the black near-spherical Fe3O4 core is surrounded by a uniform SiO2 shell of 23-nm thickness to accommodate subsequent antibody modification. The particle size distribution of MagNPs was analyzed by DLS (inset). The mean equivalent hydrodynamic radius of MagNPs is 310 nm. The synthesized AuNPs are uniform spherical and monodisperse with a diameter of 26 ± 0.5 nm. DLS (inset of Fig. 2b) reveals that AuNP particle size distribution is in a narrow range, with 78.7% of the particle sizes within 20 to 32 nm and a single peak at 26 nm (Fig. 2b). To evaluate the coupling of MagNPs to antibodies and the modification of DNA onto AuNPs, FT-IR spectroscopy was employed to analyze the IR spectra that correspond to MagNPs, MagNP-Ab, AuNPs, and AuNP-Apt-MB (Fig. 2 c and e), which were used to identify the specific functional groups of the effective components based on the peak positions in the region of infrared radiation. The strong stretching vibration of N-H located at 3385 cm−1 and the C=O of the amide bond at 3421 cm−1 demonstrate successful modification of the antibody on MagNP-Ab; in-plane bending of the C-H in functional group of the antibody also shows the coupling of the antibody to the MagNPs (Fig. 2c). As shown in Fig. 2e, the C=O bond disappears in the DNA-modified AuNPs, which is attributed to the combination of Apt-MB and AuNPs to remove the trisodium citrate as the stabilizer [26]. The presence of the -SH bond indicates the successful modification of DNA on the surface of AuNPs. In addition, we studied the assembly of AuNPs and Apt-MB by UV-vis. The maximum plasma absorption of the bare AuNPs occurred at 519 nm, as shown in Fig. 2d. The absorption peak at this wavelength corresponds to a typical AuNPs of 25 nm, which is also consistent with the visual image in the TEM. The maximum plasma absorption of the AuNPs-Apt-MB appeared at 522 nm, the covalent bond of -SH further broadened the absorption band, showing a slight red shift. Finally, the charge changed after surface modification of nanoparticles which were studied by Z-potential (Fig. 2f). Coupling MagNPs with positively charged antibodies resulted in a change in potential from − 44.7 to − 14.2 mV. DNA modification on the surface of the gold nanoparticles caused a 5.2-mV negative potential shift of the charge. These results confirm that AuNP-Apt-MB and MagNP-Ab are synthesized successfully, and both dispersion solutions have good physical stability.

Fig. 2
figure 2

TEM and DLS data (inset) of synthetic MagNPs (a) and AuNPs (b). (c) FT-IR spectra of MagNPs before (black) and after (red) antibody coupling. (d) UV-vis spectrum of AuNPs before (black) and after (red) Apt-MB modification. (e) FT-IR spectra before (black) and after (red) aptamer modification of AuNPs. (f) Z-potential before and after functionalization of nanoparticles

Full size image

Electrochemical properties

Depending on the functionalization steps, Fig. 3a shows the different changes in the imaginary part of impedance (-Z″) as a function of the real part of impedance (Z′) [27, 28]. The bare GE presented a small Rct value (the diameter of semi-circular Nyquist plots) and a large redox peak of [Fe (CN)6]3−/4−, respectively, in the process, indicating that GE has a small electron transfer resistance and a strong interfacial electron transfer ability. After the surface of the electrode was modified with cDNA and MCH, respectively, the Rct value increased due to enhancive blockage of the redox probe, [Fe (CN)6]3−/4− through electrode surface. Similarly, the redox peak gradually decreased, as indicated by CV shown in Fig. 3b. Finally, the nanocomposite bonded to cDNA on the electrode surface, the Rct value was decreased while the redox peak increased. Due to the presence of AuNPs in the nanocomposite, the surface of the electrode had enhanced electron transport capabilities. The above results show that this method has been successfully constructed. The electrochemical properties of the constructed system were studied using CV, and the electrochemical signal of MB was collected in 0.1 M PBS, as shown in Fig. S1. The height of the reduction peak at − 0.3 V was used as a measure for the quantification of the target analyte.

Fig. 3
figure 3

a EIS of bare GE, cDNA/GE, MCH/cDNA/GE, nanocomposite/MCH/cDNA/GE in 5 mM [Fe (CN)6]3−/4− containing 0.1 M KCl (the experimental points (the symbols) and the fitting curve (the solid lines)). Inset: Randles equivalent circuit using a constant phase element. b CV characterization of different modified GEs in 5 mM [Fe (CN)6]3−/4− containing 0.1 M KCl

Full size image

Optimization of the experimental parameters

The following parameters were optimized: (a) the concentration of BNP antibody, (b) pH of the reaction system, (c) temperature, (d) reaction time. Respective text and figures on optimizations are given in the Electronic Supporting Material (Fig. S2). In short, the following experimental conditions were found to give best results: (a) optical concentration, 20 μg mL−1; (b) pH, 7; (c) best temperature, 37 °C; (d) reaction time, 20 min.

Sensor response for BNP determination

Under the optimized conditions, the performance of the triple recognition immunoassay was evaluated with different concentrations of the target BNP solution (contains key concentrations 100 pg mL−1 and 400 pg mL−1 for determining HF disease status). There is no charge transfer during the specific binding of MagNP-Ab and BNP, as well as the coupling with subsequent AuNP-Apt-MB, but only the formation of nanocomplex, which mean that the detection of BNP depended on the oxidation signal of MB on the nanocomplex already fixed to the electrode surface. During this process, the more BNP are captured, the larger nanocomposite is formed (containing more MB), resulting in an increase in peak current. Different concentrations of BNP were detected using this detection strategy as presented in Fig. 4a, the DPV oxidation currents of MB increased proportionally with the BNP concentrations increasing, corresponding to the measurement principle and process described above. As illustrated in Fig. 4b, the results show a good linear relationship between the oxidation peak currents and logarithm of BNP concentration in the range of 1 to 104 pg mL−1. Meanwhile, the linear regression equation is Y (μA) = 0.240 lg CBNP + 0.272 (R2 = 0.996) with a low detection limit of 0.56 pg mL−1 (S/N = 3), according to the means of 3’s blank criterion, which is significantly lower than the clinical detection index (100 pg mL−1). The sensitivity of the method was found to be 176.23 μA μM−1 cm−2. Good linear range and low detection limit indicate that the triple recognition strategy has an ideal detection effect for the analyte. This method is compared with other literature methods for detecting BNP and shows good analytical performance as summarized in Table 1. The combination of MagNPs and AuNPs improves the conductivity of the complex while simplifying the pre-enrichment step. The dual targeting of antibodies and aptamers to BNP enhanced the specificity of recognition.

Fig. 4
figure 4

Analytical performances of the triple recognition strategy. a DPV responses of this method. After incubation with different concentrations of BNP. b Linear relationship for BNP determination in the range of 1 pg mL−1 to 10 ng/mL in 0.1 M PBS. Voltage range, − 0.6 to 0.1 V; pulse amplitude, 50 mV; pulse width, 50 ms; pulse period, 0.1 s. Reference electrode, saturated calomel electrode (SCE)

Full size image
Table 1 The analytical performances of the biosensor compared with literature methods for BNP determination
Full size table

Selectivity and stability studies

According to the mechanism of the immunoassay strategy, this method should be ineffective against other interferers. In order to study the selectivity of our approach, four kinds of interferences (AA, BHb, Mb, BSA) were tested using this detection system. It is notable that the interferences’ concentration used here is much higher than target BNP. Obvious changes in current response are observed for target BNP compared with the interfering substances (Fig. 5a). These results demonstrate that the triple recognition strategy enables excellent selectivity for determining BNP. To research the storage stability of the assay, this immunosensors were stored in a dry environmental (4 °C) for different times, and different concentrations of BNP (100 pg mL−1 and 10 ng mL−1) were tested. It is observed that the electrochemical signal retained about 93.4% of its original value (Fig. 5b and S3), showing that this strategy has excellent stability.

Fig. 5
figure 5

a Selectivity of the electrochemical assay toward five kinds of interferences. Concentrations of BNP were 100 pg mL−1, while others were 10 pg mL−1. b The stability of the proposed method

Full size image

Determination of BNP in serum samples

To evaluate the analytical performance of the immunoassay in real samples, we investigated the potential utility of our method for detecting BNP in clinical serum samples. Serum samples with BNP concentrations of 1 pg mL−1, 10 pg mL−1, 100 pg mL−1, 1 ng mL−1, and 10 ng mL−1 were prepared according to the procedure described in the experimental section. Three sets of parallel samples were tested for each concentration, and the results were taken as the average of the three sets of experiments. The measured sample concentration was calculated by substituting the signal value into the obtained linear regression equation (Fig. S4, R2 = 0.991). Statistical comparison and analysis show that the detection strategy has good consistency in different matrices (Fig. S5), which indicates that the detection strategy has predictable application value. As shown in Table 2, the relative standard deviation (RSD) of the detection results are less than 6.31%, and the recovery are in the range of 92.22–104.2%, indicating that the measured experimental values are well matched with the true concentration of BNP in serum samples. The great magnetic concentration property of MagNP-Ab helps to capture analyte directly in the complex biological samples without the purification process. To sum up, this strategy has potential clinical application and can be used for the detection of BNP in clinical serum samples.

Table 2 The result of BNP determination in clinical serum samples (n = 3)
Full size table

Conclusions

In this work, an enhanced voltammetry immunoassay using a triple recognition system is developed for determination of BNP. MagNPs and Apt-MB-modified AuNPs are used as separable identifier and signal amplifier, respectively. The double recognition of BNP by aptamers and antibodies lead to the formation of MagNP-Ab@BNP@AuNP-Apt-MB nanocomposite, which simplifies the pre-separation procedure and enhances conductivity. This strategy exhibits excellent bioactivity and can be used to detect BNP in clinical serum samples. However, this method requires voltammetry, and electrochemical reduction is irreversible, which may limit its applicability to single assays. In general, this detection platform holds good promise for HF-related BNP detection and can be further exploited for sensing applications in disease diagnosis.