Introduction

The rapid spread of the novel human coronavirus 2019 (COVID-19) and its morbidity have created an urgent need for rapid and sensitive diagnostics. Conventional methods for virus detection include lateral flow strips, enzyme-linked immunosorbent assay, and polymerase chain reaction (PCR) [1]. Among these techniques, PCR is the gold standard for detecting coronaviruses from various types of biological specimens. The main drawbacks of thermal cyclers include long analysis time, expensive instrument setup, and the need for skilled personnel [2]. Various virus detection methods, including electrochemical biosensors, fluorescence biosensors, surface-enhanced Raman scattering-based biosensors, colorimetric biosensors, chemiluminescence biosensors, surface plasmon resonance-based biosensors, and magnetic biosensors, have been developed for SARS-CoV-2 [3]. Among these detection principles, electrochemical biosensors can consist of potentiometric, amperometric, conductometric, voltammetric, polarographic, impedimetric, capacitive, or piezoelectric elements [4]. Moreover, biosensors incorporated with different recognition elements such as imprinted polymers, enzymes, antibodies, DNA, and aptamers can be applied for virus detection [5]. These excellent reviews have summarized the background of the principles and mechanisms for virus detection using biosensors.

Electrochemical biosensors can be used as point-of-care devices because of their low cost, selectivity, and ability to be miniaturized [6]. Electrochemical biosensors to detect SARS-CoV-2, such as screen-printed electrodes (SPEs) and field-effect transistors (FET), also have been widely used [7]. The electrochemical sensing of coronavirus amplicons from PCR has been developed using printed-circuit-board (PCB) electrodes [8]. In addition, SPEs have advantages such as easy production, low cost, disposability, sensitivity, and miniaturization [9, 10]. Ehsan et al. chose SPEs as impedance immunosensors to detect coronavirus in nasopharyngeal fluid samples. Antibody immobilization on a carbon electrode provides a low limit of quantification (0.25 fg mL−1) of patients’ samples [11]. SPEs coated with gold nanoparticles are applied to immobilize antibodies against the coronavirus nucleocapsid with a linear range from 1.0 pg mL−1 to 100 ng mL−1 and a limit of detection of 0.4 pg mL−1 [12]. The Cu2O-modified SPE is used to detect the spike protein of SARS-CoV-2 using charge transfer resistance. The assay range is around 0.25 fg mL−1 to 1 μg mL−1 within 20 min, and it exhibits a recovery rate of ~ 97–103% in different artificial saliva, artificial nasal, and universal transport media [13]. These studies indicate the potential and power of disposable electrochemical biosensors. However, these studies must also address the issues of complex manufacturing and the time needed for analysis if they are to be applicable for point-of-care testing.

Magnetic beads (MB) integrated with SPE have been applied to detect different biomarkers [14, 15]. In this study, the facile detection of spike protein in serum, urine, and saliva was developed using MB-based disposable SPE in combination with a portable potentiostat. The antibody-peroxidase-loaded MBs are the detection platform for electrochemical assays. The MBs enable simple washing and easy separation on the working electrode using a magnet. The assembly of the MBs and enzymatic reaction on the electrode surface can increase the measurement sensitivity. The electrochemical signal is generated by peroxidase using hydrogen peroxide and hydroquinone as the substrates. We systematically characterized the platform of MBs and SPEs by scanning electron microscopy (SEM), thermogravimetric analysis (TGA), and a superconducting quantum interference device (SQUID). The performance of these immunosensors was evaluated using square wave voltammetry (SWV) to detect spike protein added to saliva, urine, and serum.

Experimental

Reagents and materials

MB-NH2 and MB-COOH (U-108, U-118) were purchased from TAN Bead (Taoyuan, Taiwan). Hydroquinone (HQ), glutaraldehyde, sodium phosphate dibasic, sodium phosphate monobasic, 3-aminophenyl boronic acid hemi sulfate salt (APBA, 95%), polyethylene glycol (PEG, mw:3350), dextran sulfate (DS), glucose, bovine serum albumin (BSA), vitamin C, and human serum were from Sigma-Aldrich (MO, USA). DMEM was purchased from Elabscience (TX, USA). Urea was from J.T. Baker (Leicestershire, UK) and γ-cyclodextrin (γ-CD) was from Wako (Tokyo, Japan). DNase was from Worthington Co. (Lakewood, USA), and RNase was from GenMark (Taipei, Taiwan). Biotinylated detection antibody for SARS-COV-2 spike protein (S2) (Lot#0,918,201,221), streptavidin-HRP (Lot#G01152188), standard spike protein (Lot#C07152088), and ELISA kit were obtained from RayBio (GA, USA). Hydrogen peroxide (H2O2, 30%) was from Fluka (MI, USA). All other chemicals were of extra pure analytical grade and used without further purification. Phosphate buffer solution (0.1 M and pH 7.0) was the diluent to study the reactions. Water was freshly prepared using the ultrapure water system (Sartorius Lab instruments GmbH & Co, Gottingen, Germany). SPEs containing a counter, working, and Ag reference electrodes were from Zensor (TE-100, Taichung, Taiwan).

Apparatus

The morphology and size of modified MBs were characterized using a scanning electron microscope (SEM, HITACHI S-4700, Tokyo, Japan), transmission electron microscope (TEM, JM-1011, JEOL, Tokyo, Japan), and ZetaSizer® Nano ZS 90 (Malvern, Worcestershire, UK). The carbon-coated copper grid was used to load the nanomaterials for TEM analysis. Thermal decomposition of modified MBs was analyzed using Q50 Analyzer (TA Instruments, DE, USA) in a nitrogen environment at various heating rates. The electrochemical properties of the SPE were analyzed using a portable electrochemical simulator (Zensor, ECAS-100). All the UV spectra were recorded by a microplate reader (SpectroStar, BMG, Ortenberg, Germany).

Preparation of layer-by-layer modified MBs

A 100 µL aliquot of the MB-NH2 was transferred into a 1.5 mL Eppendorf tube. The MBs were washed twice with 1 mL of 0.1 M phosphate buffer solution for 10 min, and the supernatant was discarded. Amine groups of MBs were activated by 60 min continuous vortex in 100 µL of glutaraldehyde (4 ×) solution. Then APBA was captured onto the activated beads for 60 min at 25 °C under continuous vortex. Separately, 100 × biotin antibody and 100 × of streptavidin-HRP were mixed in a tube and incubated for 45 min at 4 °C. The MB/APBA were washed twice with 0.1 M phosphate buffer solution, then 100 µL Ab-HRP solution was added to the MB/APBA and incubated for 30 min at 25 °C to form MB/APBA/Ab-HRP. Then, the MB/APBA/Ab-HRP was further blocked in 5% glucose (GLU) for 30 min. Subsequently, the MB/APBA/Ab-HRP/GLU was washed twice with 1 mL of 0.1 M phosphate buffer solution (pH 7.0). After the wash and separation, 100 µL spike protein in serum was added to MB/APBA/Ab-HRP/GLU and incubated for 30 min at room temperature.

Fabrication and assay of the MB-based immunosensor

The fabrication of the MB-based immunosensor is outlined in Fig. 1. First, an SPE was cleaned by sonication with DDW (10 s), and a neodymium magnet on the backside of the SPE was used to attract MB/APBA/Ab-HRP/GLU on the working electrode. Five microliters of modified MB (MB/APBA/Ab-HRP/GLU) was added to the SPE. The solution of 5.0/5.0 mM H2O2/HQ solution (95 μL) was added to the SPE surface and waited for 5 s for SWV measurement. SWV measurement of HRP signal on SPE was obtained by applying a potential range of -0.4 ~ 0.2 V in a H2O2/HQ solution as previously described [16].

Fig. 1
figure 1

Schematic illustration for spike protein detection using the MB-based electrochemical immunosensor. MB/APBA/Ab-HRP/GLU/SP was synthesized using layer-by-layer modification. Ab-HRP was prepared by incubating a biotinylated antibody with streptavidin-HRP. 3-Aminophenyl boronic acid: APBA, glucose: GLU, spike protein: SP. Ab-HRP binds preferentially to APBA on MB. Unbound APBA was blocked by GLU. Spike protein binds to Ab-HRP to form the immune complex, which can impede the substrate H2O2 and electron mediator HQ from reaching HRP’s active site and electrode, and thus decrease in electrochemical signal

Full size image

Optimization of measuring parameters and signal definition

To optimize measurement, operation factors such as concentrations of antibody, HRP, H2O2/HQ, incubation time of antibody, reaction time of MBs were evaluated. Only one parameter was varied, and the other parameters were fixed. The SWV signal was obtained using 10 ng mL−1 serum spiked protein in this experiment. Each signal was repeated three times (n = 3). Based on the signal/blank (S/B) ratio, the optimal parameters were selected. The S/B ratio was calculated using the previous definition [17, 18]:

$$\begin{array}{l}S/B\;\mathrm{ratio}=\Delta{\mathrm I}_{10}/\Delta{\mathrm I}_0\\\Delta{\mathrm I}_0={\mathrm I}_{\mathrm{PBS}}-{\mathrm I}_{\mathrm{HS}}\\\Delta{\mathrm I}_{10}={\mathrm I}_{\mathrm{PBS}}-{\mathrm I}_{\mathrm{HS}+\mathrm{spike}\;\mathrm{protein}}\end{array}$$

where IPBS is the signal of MBs treated with phosphate buffer solution, IHS is the signal of MBs treated in human serum without spike protein, IHS + spike protein is the signal of MBs treated with 10 ng mL−1 spike protein in human serum. The SWV signals of MBs were obtained using the 5.0 mM H2O2/5.0 mM HQ solution. The principle of the S/B calculation was demonstrated in Fig. S1.

SWV and EIS assay procedure

Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were used to study the different steps of immobilization, whereas the quantitative measurements were done using SWV. CV was cycled from − 0.5 to 0.6 V with a scan rate of 50 mV s−1. SWV was recorded from 200 to − 400 mV with the increment time of 5 ms, amplitude of 75 mV, sample width of 6 ms, and pulse period of 100 ms. EIS was run with amplitude 100 mV, frequency range 1 ~ 1000 Hz, and initial potential 20 mV. The real and imaginary components from impedance data of EIS were plotted as Nyquist plots. A semicircle in the Nyquist plot corresponds to the electron-transfer process, and the diameter of the semicircle is equal to the resistance of charge transfer (Rct). The EIS data of the modified MBs were also regressed by software (Zsimpwin) using the resistor–capacitor-resistor equivalent circuit to obtain the resistance and capacitance of the MB-loaded electrode. Histogram of charge transfer resistance and peak current of different MBs were studied. Assays were performed in triplicate, and the error bar represented the standard deviation of the measurement. The Ks by the CV method was evaluated using the Nicholson method as described previously [17].

Association constant analysis

In order to understand spike protein capture by the immobilized antibody on MBs, the association constant (Ka) for the binding between spike protein and the antibody was determined using EIS and the Langmuir isotherm assumptions such as monolayer adsorption and equal binding energy for all binding sites. θ represents the occupied ratio of antibody binding sites, and Ka is related to the equilibrium adsorption between spike protein and antibody on MBs. The Ka and θ were calculated using the Rct data from EIS as described previously [17].

Statistical analysis

In this study, all data are reported as means ± standard deviations (SD) in three repeats. The mean and SD of measurements were calculated using Microsoft Excel to evaluate the reproducibility. The data are expressed as mean ± standard deviation from three repeats. To predict the limit of detection (LOD) and recovery rate (%), the following formulas were used [19]:

$$\mathrm{LOD}=3\times \frac{SD}{m}$$

where SD is the standard deviation of the lowest concentration of spike protein and m is the slope of the calibration curve.

$$\mathrm{Recovery }(\mathrm{\%})=\frac{\mathrm{found}}{\mathrm{added}}\times 100$$

Results and discussion

Principle of SARS-CoV-2 biosensor

The detection mechanism and fabrication steps of the developed MB-based immunosensor for the detection of COVID-19 spike protein are demonstrated in Fig. 1. Layer-by-layer modified MBs are used to fasten the assay through simultaneous capture and concentration of the targets with the aid of a magnet. The APBA modified MBs (MB/APBA) can immobilize Ab-HRP via diol-boronic ester bond to prepare the MB/APBA/Ab-HRP and then glucose treatment to obtain MB/APBA/Ab-HRP/GLU. HRP is one of the most extensively studied and commonly used enzymes for the construction of H2O2 biosensors. Generally, direct transfer of an electron between HRP and an electrode is difficult because the active sites of HRP are deeply buried in a thick protein shell and because the large distance between the active sites and the electrode surface will slow down the electron transfer. Electron transfer via a mediator like HQ, however, is more effective for establishing an electrical connection between the redox centers and the electrode [20]. In this work, H2O2 was a substrate reduced by Ab-HRP, and HQ was used as a mediator for electron transduction. Previously, chorionic gonadotrophin has been proved to decrease the catalytic activity of the Ab-HRP on the detection probe and reduce the current signal on the immunosensor [21]. In another paper, when the antigen of parathyroid hormone increased, the catalytic activity (Kcat) of Ab-HRP on SPCE decreased, and the signal current decreased [17]. These papers suggest the mechanism as follows. When the concentration of antigen increased, the thickness of immuno-complexes on MBs increased. The non-electroactive complex was a blocking layer against the mass transfer of H2O2 to HRP sites and also impeded the transfer of the electron mediator HQ to the electrode. EIS is an effective method for probing the features of an electrode using an impedance signal. We also found an increase in the Rct values after the formation of immuno-complexes on MBs when an increase in spike protein concentrations. This could be ascribed to the formation of the antigen–antibody complex, which could increase the impedance and hinder the transfer of electrons to the electrode. This explained the spike protein concentration was inversely proportional to the current signal.

Characterization of modified MBs

The morphology of modified MBs was investigated by scanning electron microscope (SEM) and transmission electron microscope (TEM) imaging. Figure 2(A–D) shows micro-scale images of the original MBs and modified MBs with a 1.0 µm scale bar. All images were obtained with 15 kV voltage. MBs are typically sphere shape, as seen in Fig. 2(A); MB/APBA has an aggregated cloud-like structure, as in Fig. 2(B), and MB/APBA/Ab-HRP have expanded clusters, possibly from the interaction of Ab-HRP with increased size, as shown in Fig. 2(C). An aggregated globular structure was observed in Fig. 2(D). In addition, the various sizes of the MBs were characterized using a Zeta sizer. An increase in size from 978 to 1041 nm was indicated the conjugation of APBA on the MB surface. Similarly, after immobilizing Ab-HRP, the size increased to 1535 nm and finally reached 2674 nm after adsorption of spike protein (Table S1). To select MBs, we performed a test with two kinds of MB, one containing the NH2 group and the other containing the COOH group.

Fig. 2
figure 2

SEM images of layer-by-layer modified MBs AMB-NH2, B MB/APBA, C MB/APBA/Ab-HRP, D MB/APBA/Ab-HRP/GLU/spike protein with scale bar: 1 µm, and E size distribution of modified MBs

Full size image

We evaluated HRP adsorption using two kinds of MBs modified by APBA. The crosslinkers tested for MB-NH2 and MB-COOH were glutaraldehyde and EDC/NHC, respectively. We measured the APBA remaining in the solution using UV absorbance at 233 nm. By using the APBA standard as a reference, the concentration of APBA was evaluated and plotted, as in Fig. S2(A). To confirm the effect of correct orientation, glutaraldehyde is used as a cross-linker to conjugate Ab-HRP to magnetic beads without APBA. The MB-NH2 was directly activated using a glutaraldehyde solution to form MB-NH2/GA. The Ab-HRP was added to the MB-GA and incubated for 30 min, and a TMB assay was used to evaluate the HRP activity of the MBs. Figure S2(B) shows the comparison of the calibration curve of MB with and without the APBA layer. MB-NH2/GA/APBA/Ab-HRP had higher activity (slope) than MB-NH2/GA/Ab-HRP. The APBA can covalently bond to 1,2- and 1,3-diols of the carbohydrate moiety of glycoproteins that can immobilize the Ab-HRP on the MBs. Previously, antibodies were immobilized using their lysine, arginine, and glutamine residues by random conjugation of glutaraldehyde on electrode surfaces, resulting in unsatisfactory activity in antigen targeting [22]. The immobilization of the oriented antibody with the proper targeting activity has been achieved using APBA [23]. Owing to the oriented immobilization of Ab-HRP via boronic acid groups on MB, more activity of the Ab-HRP was preserved in this study. Glutaraldehyde randomly conjugates the Ab-HRP using the amino-acid residues with the amines on MB-NH2, resulting in the decrease of HRP activity. The presence of APBA on the MB surface help to preserve more Ab-HRP activity, as shown in Fig. S2(B). Figure S3(A–D) shows TEM images of the modified MBs. The two MBs (MB-NH2, MB/APBA) had a similar pattern. However, the comparison shows that MB-NH2 looks less aggregated with a smaller size. Upon immobilization, a round aggregated structure with increased size was observed. Antibody immobilization and signal amplification are two key factors in target detection at ultra-low concentrations. The boronic acid in APBA forms cyclic esters with 1,2- and 1,3-diols on the carbohydrate moiety of glycoproteins that can immobilize the antibody on the MBs [24, 25]. APBA can maintain the correct antibody orientation and improve antibody targeting efficiency [26]. Therefore, we adopted APBA as a ligand to the immobilized antibody on MB to maintain the correct orientation of antibodies.

The thermal decomposition of MBs was analyzed under a nitrogen atmosphere and a heating rate of 10 °C min −1. The TGA result is shown in Fig. 3(A). There are two major endothermic procedures, found at 334 and 380 °C, respectively. The first weight loss region ranging from 61 to 271 °C is responsible for 2% weight loss. The first weight-loss region shall be named as the thermal release of crystal water in MB-APBA. Second weight loss indicates high decomposition temperature, this weight loss region is regarded as the destruction and decomposition of the APBA layer under thermal effect. On the other hand, the TGA analysis of MB-NH2 showed only one decomposition process, starting at 280 °C. For a further understanding, the magnetic properties of our MB were measured using SQUID. The magnetization curve of MB and MB/APBA were measured and compared in Fig. 3(B).

Fig. 3
figure 3

A Thermogravimetric analysis of modified MBs; B magnetic hysteresis of MB-NH2 and MB/APBA

Full size image

Electrochemical characterization

EIS has been performed to characterize the stepwise fabrication procedures of this immunosensor, and the effect of electrode modification can be reflected on Rct as reported by Ren et al. [27]. Impedance results during the MB modification are shown in Fig. 4(A). In Fig. 4(B), the MB-free electrode shows the lowest value of Rct (1238 Ω). After conjugating MB-NH2 with APBA, the Rct value increased to 3316 Ω. Then, there was a significant increase in the Rct value after the antibody was immobilized on the MB surface. After blocking with glucose, the MB/APBA/Ab-HRP/GLU had the highest Rct. The solution resistance and the double-layer capacitance were similar after layer-by-layer modification on MBs, suggesting these parameters were not affected by the APBA conjugation and antibody immobilization [10, 28]. However, Rct values were significantly influenced by the binding of spike protein on MBs. The hindrance to the electron transfer process, blocking of a redox reaction, and the hamper of electron tunneling to the electrode attributed to the increasing thickness of the immuno-complex formed on MB surfaces [29]. EIS is a non-destructive and label-free measurement for the immuno-complex by applying the small amplitude perturbation to collect the impedance information of the electrode.

Fig. 4
figure 4

A Impedance results of layer-by-layer modified MBs using 3 mM ferricyanide in phosphate buffer solution. B Histogram of charge transfer resistance of modified MBs on bare SPE. Error bar represents the standard deviation from three repeats n = 3

Full size image

Thus, Rct results confirmed the successful fabrication of the immunosensor. The kinetic parameters of modified MBs, such as electron transfer rate constant (Ks, cm s−1), were compared after modification (Fig. S4 and Table S2). The oxidation and reduction peak currents in CV from different electrode surfaces can be theoretically attributed to electron transfer rates (Ks) and Rct. The MB-free SPE had higher Ks; upon modification, the values of the Ks decreased. MB-NH2 has a Ks value of 0.19 cm s−1, and upon coating with APBA, the Ks value decreases slightly to 0.18 cm s−1. After Ab-HRP was immobilized, there was a significant decrease in Ks (0.17 cm s−1). Finally, after glucose blocking the value decreased to 0.13 cm s−1 as in Fig. S4(B). The reduction of electron transfer rate after the layer-by-layer modification was observed.

Optimization of the experimental parameter

Optimization of HRP and antibody dilution

The antibody, as a sensing element, must have a strong immobilization on the MBs with a good homogenous orientation to detect spike protein efficiently. Moreover, the modified MBs must be stable enough to bear the modification condition and the electrochemical assay. For HRP optimization, various dilution from 50 to 200 × was tested. The delta current (ΔI) rose as HRP folds increased to 100 × and then decreased. One 100-–fold dilution had the highest S/B ratio, as shown in Fig. S5(A). Based on the S/B ratio, 100 × HRP was selected as the optimized HRP fold. Similarly, for antibody dilution, different antibody folds (50 to 200 × were chosen) and an optimized HRP fold (100 ×) were selected for this test. The delta current gradually decreased as the antibody fold increased, as noted in Fig. S5(B). The optimized condition for the antibody fold was 100 × as it had the highest S/B ratio.

Optimization of incubation time, H2O2/HQ concentration, and blocking process

The effects of incubation time and H2O2/HQ concentration were investigated for sensitive electrochemical detection of COVID-19 spike protein. Given the incubation time of the antibody, immobilization plays a key role in binding with the spike protein. To improve the total assay time, the incubation times of Ab-HRP were tested. To optimize incubation time, 10, 20, 30, and 60 min were tested. The optimized Ab-HRP fold was selected for this test. The SWV signal indicated that 30 min had the highest S/B ratio, as seen in Fig. S5(C). The optimization process led to an incubation period of 30 min for the 100 × dilution HRP and 100 × Ab concentration. Similarly, to optimize the H2O2/HQ concentration, 2.5/2.5, 5/5, 10/10, and 20/20 mM were tested. The delta current (ΔI) was increased with H2O2/HQ concentration, as seen in Fig. S5(D). The highest S/B ratio occurred at 5/5 mM of H2O2/HQ. Therefore, it was selected for further testing.

Previously, boronic acid-functionalized MB (BA-MB) has been reported to isolate immunoglobulin G synthesized by recombinant cells [30]. Glucose has the BA binding ability via the cis-diol linkage and affects the antibody purification by BA-MB. Borlido et al. also demonstrate that glucose can compete the BA on MB with the glycoproteins [31]. Using this concept, we first immobilized the Ab-HRP and sequentially blocked the free APBA binding sites using the optimized glucose concentration to avoid the nonspecific binding from the samples. The blocking process is very crucial for immunosensor preparation because it protects the immobilized antibodies and avoids nonspecific binding of impurities on electrodes. Therefore, six blocking agents were added to the MB surface and evaluated for their ability to prevent nonspecific binding. The chemicals tested included 5% dextran sulfate, bovine serum albumin, phosphate buffer solution, gamma-cyclodextrin, glucose, polyethylene glycol. Based on experimental data, the S/B ratio of glucose was higher than BSA shown in Fig. S6. Therefore, glucose was chosen to block the free APBA on the MB surface. The optimized parameters are summarized in Table 1. In addition, other related parameters, such as the APBA concentration and the incubation time of the spike protein, were tested using a HRP activity assay. The optimized APBA concentration on MBs was 0.1%, and the incubation time of spike protein was found to be 30 min (Fig. S7).

Table 1 Summary of optimization of experimental parameters
Full size table

Detection of spike protein

The analytical performance of the MB-based immunosensor was investigated by SWV. The dose–effect of spike protein ranging from 3.12 to 200 ng mL−1 was evaluated. We investigated the sensor’s ability to detect the spike protein of SARS-CoV-2 added in human saliva, urine, and serum. The SWV spectra of saliva, urine, and serum samples are shown in Fig. 5(A, B, C), and their standard curves are shown in Fig. 5(D, E, F). The current signal decreased with increasing doses of spike protein because higher concentrations of antigen decreased the current signal of the HRP-labeled antibody. This was because the antigen occupied the limited binding sites of antibody-HRP to form an immunocomplex. The current was inversely proportional to the concentration of spike protein, and there was a linear dependence on spike protein concentrations in the range of 3.12 to 200 ng mL−1. This yielded sensitivities of 4.44, 2.49, and 0.84 µA ng−1 mL and LODs of 0.20, 0.31, and 0.54 ng mL−1 in human saliva, urine, and serum, respectively. The dose-dependent SWV signals could be explained by Rct results. The thickness of immuno-complexes on MBs increased when the spike protein increased. The inert complex hindered the H2O2/HQ reaction with HRP and also reduced the electron transfer of the redox mediator to the SPE. Previous reports also demonstrate the targets of immunosensors decrease the activity of the HRP-labelled antibody on the immunosensor and reduce the current signal [21]. A detailed comparison of LOD is shown in Table S3. The data and recovery of SWV, ELISA, and EIS are summarized in Table 2. The developed MBs can also be applied to the colorimetric assay. In the absence of spike protein, the TMB was converted by MB/APBA/Ab-HRP/GLU and became blue. The dose–effect of spike protein on EIS is shown in Fig. 6(A). The Rct value increases with an increase in the dose of the spike protein. The linear relationship between the concentration of spike protein and the Rct value is shown in Fig. 6(B). Based on the equation, we evaluated the association constant (Ka) and found it to be 4.11 ng−1 mL.

Fig. 5
figure 5

SWV signals generated from spike protein added in A saliva, B urine, C serum. Calibration curves based on peak currents from spike protein added in D saliva, E urine, F serum using MB-based immunosensors. Detection was carried out on the working surface of SPE by placing an external magnet and loading 5 µL of sample and 95 µL of HQ/H2O2. Each data point represents the mean ± SD of three separate measurements were obtained by using the same SPE. The working potential was around 92 mV

Full size image
Table 2 Comparison of recovery yield of spike protein by EIS, SWV, and ELISA method*
Full size table
Fig. 6
figure 6

A Impedance results for the measurement of the association constant using spike protein, B standard curve of △RCT(Ci)/RCT(C0) versus spike protein concentration to obtain the associate constant

Full size image

Interference test

To study the selectivity of the fabricated immunosensors, human serum spiked with 10 ng mL−1 of spike protein was evaluated in the presence of potentially interfering compounds in serum such as vitamin C, serum albumin, urea, RNase, DNase, trypsin, immunoglobulin, and cell culture medium (DMEM). Spike protein (10 ng mL−1) was used as the positive control for the SWV method. Spike protein had the highest S/B ratio (control). In the presence of an interferent, the MB-based immunosensor still showed a similar current response. Urea and DNase would reduce the signal currents and S/B ratios. The other interferents did not significantly affect the spike protein detection of MB-based immunosensor (Fig. 7).

Fig. 7
figure 7

SWV histogram of the MB-based immunosensor (MB/APBA/Ab-HRP/GLU) under the interfering compounds. The tested spike protein concentration in human serum was 10 ng mL−1. n = 3

Full size image

Stability and storage test

The developed MBs were stored at 4 °C for 49 days to evaluate their storage stability. MB/APBA was treated by Ab-HRP and 5% glucose to freshly prepare MB/APBA/Ab-HRP/GLU. The storage stability of modified MB was evaluated based on the S/B ratio and SWV signal in Fig. 8(A). After 49 days of storage of MB/APBA, the S/B ratio was 3.66, which indicated it has great sensing ability. For MB/APBA/Ab-HRP/GLU, the S/B ratio on Day 1 was 3.44, and after 49 days of storage, it declined to 2.85. The activity and targeting of antibody-peroxidase on MBs could be retained after 49 days of storage, as shown in Fig. 8(B). More than 80% of the S/B ratio had been maintained as compared with freshly prepared MBs. This suggests that the antibody-immobilized MBs had good stability.

Fig. 8
figure 8

Stability tests of modified MBs A MB/APBA and B MB/APBA/Ab-HRP/GLU after 49-day storage at 4 °C. SWV was run with amplitude 75 mV, pulse period 100 mV, and initial potential (200 to − 400 mV) using H2O2/HQ. The tested spike protein concentration was 10 ng mL−1

Full size image

A comparison of the published immunosensors to this study is summarized in Table 3. These reported immunosensors adopted the EIS, DPV, SWV, and CV methods to quantify the targets on SARS-CoV-2. For example, Eissa and Zourob integrated sample collection and detection into a single platform using diazonium-modified SPEs for nucleocapsid detection. They achieved a limit of detection of 0.8 pg mL−1 for SARS-CoV-2, and there was no significant cross-reactivity against the influenza antigen [32]. SPEs coated with graphene oxide and gold nanostars can detect the viral spike protein, with a superior limit of detection of 1.68 × 10−22 μg mL−1 [33]. In these papers, nanomaterials such as MBs and gold nanoparticles are used to amplify the electrochemical signal using different mechanisms, including a large active surface, extraordinary electron-transfer ability, and multivalent affinity with the enzyme and antibody. Very recently, an electrochemical assay was developed for Nucleocapsid protein using primary antibody-labeled MB to capture antigen and secondary antibody-labeled phosphatase to detect antigen and generate enzymatic product naphthol [34]. The carbon black modified SPE then converted the naphthol to the current signal. This novel device of Fabiani et al. provides low LOD and rapid analysis. However, two antibodies and several wash steps were needed as compared to the single antibody and simultaneous detection in the present study. Our proposed MB-based immunosensor could detect spike protein using the SWV and EIS methods with a fast procedure and acceptable sensitivity. However, these following challenges still need to be carefully investigated in future studies. This developed method is limited to the viremia stage, where viruses enter the bloodstream and have access to the rest of the body. Because the clinical samples contain proteins, cells, and biomolecules that can interfere with the selectivity and sensitivity of the developed biosensors, the evaluation using in patient’s samples from nasopharyngeal swabs can prove the clinical application of this system. Finally, the antibody used in this study has the limits of being heat sensitive and unstable to long-term storage, which will need further improvement.

Table 3 Comparison of various nanomaterial-based electrochemical methods for detection of SARS-CoV-2
Full size table

Conclusion

We developed a magneto immunosensor for detecting the SARS-CoV-2 spike protein in three body fluids by combining MB-based SPEs and a portable potentiostat. The sensitive and reliable detection was achieved in a label-free format by monitoring the change in the SWV current upon binding with the spike protein. This MB can detect spike protein by electrochemical methods with good stability and reproducibility. This portable device can also be used to confirm the conventional analysis of spike protein by lateral flow and ELISA-like methods. This platform’s advantages include low-cost MBs, disposable electrodes, label-free detection, and facile assay in contrast to the expensive PCR instruments. We demonstrated rapid electrochemical analysis of spike protein with an acceptable sensitivity, which has the potential to be used for point-of-care diagnostics.