An ultrasensitive ELISA to assay femtomolar level SARS-CoV-2 antigen based on specific peptide and tyramine signal amplification

Abstract

The SARS-CoV-2 spreading rapidly has aroused catastrophic public healthcare issues and economy crisis worldwide. It plays predominant role to rapidly and accurately diagnose the virus for effective prevention and treatment. As an abundant transmembrane protein, spike protein (SP) is one of the most valuable antigenic biomarkers for diagnosis of COVID-19. Herein a phage expression of WNLDLSQWLPPM peptide specific to SARS-CoV-2 SP was screened. Molecular docking revealed that the isolated peptide binds to major antigenic epitope locating at S2 subunit with hydrogen bonding. Taking the specific peptide as antigen sensing probe and tyramine signal amplification (TSA), an ultrasensitive "peptide-antigen-antibody" ELISA (p-ELISA) was explored, by which the limit of detection (LOD) was 14 fM and 2.8 fM SARS-CoV-2 SP antigen for first TSA and secondary TSA, respectively. Compared with the LOD by the p-ELISA by direct mode, the sensitivity with 2nd TSA enhanced 100 times. Further, the proposed p-ELISA method can detect SARS-CoV-2 pseudoviruses down to 10 and 3 TCID₅₀/mL spiked in healthy nasal swab sample with 1st TSA and 2nd TSA, separately. Thus, the proposed p-ELISA method with TSA is expected to be a promising ultrasensitive tool for rapidly detecting SARS-CoV-2 antigen to help control the infectious disease.

Graphical Abstract

1. Introduction

Since the end of 2019, the SARS-CoV-2 induced coronavirus disease 2019 (COVID-19) pandemic has killed over 6 million people worldwide and plummeted the global economy [1]. Frequent testing still plays a paramount role in curbing the virus spread [2]. Although real-time polymerase chain reaction (RT-PCR) based nucleic acid testing is the gold standard for COVID-19, it requires expensive instruments and professional operators [3], [4]. SARS-CoV-2 belongs to the genus beta coronavirus and is mainly composed of four structural proteins: spike (S) protein (SP), nucleocapsid protein (NP), envelope protein and membrane protein [5]. SP is a trimeric membrane protein consisting of S1 subunit and S2 subunit, which mediates host recognition and membrane fusion [6], [7]. S1 protein contains receptor binding domain (RBD) [8], which is a functional area of the interaction between the SP and the host cell membrane to recognize the host cell surface receptor for transmission [9]. As an abundant viral transmembrane protein, SARS-CoV-2 SP has a different amino acid sequence from other coronaviruses, which has high immunogenicity and specificity and is considered as the most suitable antigen for direct detection of virus particle [10], [11]. As the current prevalent SARS-CoV-2 strain, the Omicron variant is significantly less pathogenic and virulent than the original strain and variants such as Delta and has caused a lower proportion of severe illness and death, but it spread more swiftly [12], [13]. Antigen testing is an important diagnostic tool for detecting infection and has been shown to help curb the spread of SARS-CoV-2 [14]. The World Health Organism recommended viral antigen detection, which has the merits to identify those people who are most likely at risk of spreading the disease [15]. So far, many methods for SP or S1 antigen testing have been developed, such as enzyme-linked immunosorbent assay (ELISA) [16], [17], lateral flow immunoassay (LFIA) [18], molecular-imprinted biosensor [19], electrochemical immunosensors [20], [21] and chemiluminescence immunoassay (CLIA)[22]. However, the sensitivity for most methods is unsatisfactory, despite that they are based on antibodies or aptamers to recognize the antigens. Therefore, it is vital to develop ultrasensitive antigen assay to diagnose SARS-CoV-2 virus infection [23].

For low-abundance targets, the method with low sensitivity often leads to false-negative, while the traditional methods can no longer meet the high sensitivity requirements. For instance, ELISA is a common antigen diagnostic method, nevertheless, it experienced unsatisfactory sensitivity in comparison with other methods. Specific molecular probes alone are not enough to achieve fast and accurate diagnosis. Therefore, a simple but robust signal amplification method is paramount. Significantly, it is the key issue to integrate powerful signal amplification and biorecognition elements with good affinity and specificity. Phage display technology has been used as one of the most common methods for bioscreening of specific ligands [24], [25] for biosensors [26], [27], [28], [29]. For example, we recently isolated a specific phage as bifunctional probe with antigen recognition and signal amplification to develop CLIA to detect low to 78 pg/mL SARS-CoV-2 S1[22]. Apparently, this strategy is still felt dissatisfied due to its limited sensitivity. Tyramine signal amplification (TSA) technology is a kind of enzymatic assay that use horseradish peroxidase (HRP) for in situ labeling of target proteins densely. The main principle of TSA is to use the peroxidase reaction of tyramine, whereby tyramine is covalently bound under HRP-catalyzed H₂O₂, producing a large amount of enzymatic product that binds to tryptophan, tyrosine and histidine residues of the protein [30]. However, most immunoassays based on TSA techniques used nanomaterials to increase tyramine loading. For example, strategies such as coating HRP on gold nanoparticles (Au NPs) with tyramine to form tyramine-HRP repeats[30], enriching tyramine-HPR conjugates on functionalized Au NPs [31], had been reported. However, these experimental steps are tedious and not conducive to flexible application. Meanwhile, peptides have significant advantages, such as good stability, ease of synthesis, and low-cost [32]. Consequently, they have been widely applied in biosensing, tumor therapy and templated synthesis [33], [34], [35], [36], [37].

In this work, we screened a new phage monoclonal expression of WNLDLSQWLPPM peptide by phage display, which was identified by specificity test and affinity testing as well as antigenic epitope analysis by molecular docking. Then by combining the specific peptide as an antigen recognition probe and tyramine-based signal amplification, an ultrasensitive peptide-ELISA (p-ELISA) assay was explored, capable of detecting low to 0.4 pg/mL (2.8 fM) for SARS-CoV-2 SP antigen via secondary TSA mode, by which the sensitivity is enhanced 100-fold compared with the direct mode (40 pg/mL). Further, the as-proposed method can detect SARS-CoV-2 pseudoviruses down to 3 TCID₅₀ (median tissue culture infective dose)/mL spiked in nasal swab samples. This method is universal and could be modularly applied to other sensitive ELISA platforms. Further, the as-proposed p-ELISA with TSA is a potential ultrasensitive method to assay SARS-CoV-2 S antigen to help control the epidemics.

2. Experimental section

2.1. Molecular docking

The interaction between peptide Pn and SARS-CoV-2 S protein was inferred by molecular docking. The three-dimensional (3D) structure of S protein is available in the protein data bank (PDB:7FG7) [38]. The peptide was constructed and further optimized with OPLS3 force field, which was also employed to optimize SARS-CoV-2 S protein (removing unnecessary ligands and water molecules). At the same time, the known site-binding experiments were used to limit the docking range (Ser686-Pro1140). Finally, the induced fit docking method was adopted and the optimal results of different docking regions were exported.

2.2. Microscale thermophoresis (MST) analyzing the affinity

SARS-CoV-2 SP with serial dilutions and peptide Pn labeling 5-carboxy fluorescein (FAM) (Pn-FAM) were diluted with PBS (pH 7.4). Equivolume of Pn-FAM solution with constant concentration was mixed with SARS-CoV-2 SP with varying concentrations and incubated for 30 min protected from light. The above mixtures were separately added to premium capillaries and subsequently subjected to MST analysis, which was performed by a reported method [39], [40], [41] using a Monolith NT.115 (NanoTemper Technologies, Germany). The thermophoresis measurement was performed at 20% excitation power and 40% MST power. The dissociation constant (K _d) was obtained on basis of the MO affinity analysis software.

2.3. Establishment of a peptide-based ELISA (p-ELISA) assaying SARS-CoV-2 SP antigen

In brief, peptide Pn was fixed overnight at 4 °C on microtiter plate, which was then washed 3 times with 0.05% PBST and blocked with 2% bovine serum albumin (BSA) for 2 h. After washing, SARS-CoV-2 SP antigen was incubated with shaking at room temperature (RT), to which anti-SARS-CoV-2 RBD monoclonal antibody (mAb) was added. After that, IgG-conjugated horseradish peroxidase (HRP) (IgG-HRP) mAb was added and incubated with shaking at RT for 1.5 h. Color development solution was added to the wells. Finally, 50 μL of 2 M H₂SO₄ was added to interrupt the color development reaction, and the optical density at 452 nm (OD_452 nm) was measured with microplate reader. The procedure for optimizing of the conditions is described in Supplementary data.

2.4. p-ELISA experiment with TSA mode

To improve the sensitivity of the assay, we used biotin-tyramine/HRP-labeled streptavidin (SA-HRP) for signal amplification. To research the optimal conditions, 200 pg/mL of SARS-CoV-2 S antigen was applied for the assay under the ELISA experimental conditions. The concentration of biotin-tyramine, the reaction time of tyramine-biotin and SA-HRP dilution were optimized. The optimal conditions obtained were used to detect different concentrations of SARS-CoV-2 S antigen. The reaction time of biotin-tyramine for the secondary TSA was 10 min, and the others were the same as above.

2.5. Detection of SARS-CoV-2 pseudovirus by simulation of real samples

SARS-CoV-2 pseudovirus was first inactivated (65 °C, 30 min) in biosafety cabinet[22]. Then the inactivated pseudoviruses were added to healthy nasal swab samples. ELISA experiments were performed under the optimal conditions described above.

3. Results and discussion

3.1. Screening and identification of phage bound to SARS-CoV-2 SP antigen

A 12-peptide pIII phage display library was used to biopan phage clones that bind to SARS-CoV-2 SP (Scheme S1 and Supplementary experimental section). From the 2nd round on, the phage reversion rate increased with biopanning rounds (Table S1), indicating the effective enrichment of the phage bound to SARS-CoV-2 SP. Then 30 phage monoclonals were randomly selected and amplified to extract DNA for sequencing, from which six peptide sequences were obtained and peptide P1 with sequence of WNLDLSQWLPPM appeared most frequently (16/30) (Table S2). The binding ability of the six phages to SARS-CoV-2 SP was examined by phage ELISA assay. Compared with other phages, P1-phage had a higher absorbance value and was significantly higher than that of the control (Fig. S1), indicating that P1-phage binds to SARS-CoV-2 SP with good affinity.

The binding of P1-phage to various antigens (each 2 µg/mL) was tested ( Fig. 1). The OD_452 nm value for SARS-CoV-2 SP is remarkably higher than those for MERS-CoV SP and SARS-CoV-2 SP. That is, there doesn’t exist cross-reactivity between P1-phage and other coronavirus S proteins as well as some randomly chosen tumor markers and antibodies. More interestingly, the OD_452 nm value for SARS-CoV-2 S2 is a little less than that value for SARS-CoV-2 SP, but much higher than that for SARS-CoV-2 S1, suggesting phage P1 majorly binding to S2 subunit. On the other hand, the signal for SARS-CoV-2 S1 is still obviously higher than those values for other antigens, indicating that P1 may also bind to SARS-CoV-2 S1 loosely. This may be reasonable, considering that SARS-CoV-2 SP consisting of S1 subunit and S2 subunit. Meanwhile P1 did not cross-react with SP from other coronaviruses, and neither bound to randomly selected tumor markers and antibodies, suggesting that phage-displaying peptide P1 had high specificity. Peptide P1 was synthesized by linking a short spacer GGGSKKKC at its C-terminus to obtain peptide Pn for the following experiments.

Fig. 1.

Selectivity assay of P1-phage with various antigens. Error bars represent the standard deviation of three replicates.

3.2. Molecular docking

We adopted flexible docking to construct the 3D structure of complex of the peptide Pn and S protein. The optimal output docking result is shown in Fig. 2A. Apparently, the amino acid residues (Gln7, Trp8, Leu9, Pro10, Pro11, Met12) of the peptide bind to subunit S2, while a fraction of amino acids (Trp1, Asn2) bind to subunit S1 by hydrogen bond (Fig. 2B, Fig. S2). Multiple hydrogen bonds form between amino acid residues of peptides and S proteins. Specifically, Gln7 forms two hydrogen bonds with Arg1039 and Asp1041, Trp8 forms one hydrogen bond with Asp1041, while Met12 forms two hydrogen bonds with Lys1045 to generate complex (Fig. S2), indicating that Gln7, Trp8, Met12 are possible key amino acids. This molecular docking results agree well with our specificity testing.

Fig. 2.

(A) 3D structure of complex of the peptide Pn and S protein. (B) Predicted interaction of peptide Pn with S protein. blueviolet representing peptide, grey representing S protein, while the amino acids residues on peptide binding with S1 subunit are labeled by green.

3.3. Affinity analysis by MST

FAM fluorescent labeling Pn peptide (Pn-FAM) was synthesized for MST measurements. MST verified a direct binding between Pn peptide and SARS-CoV-2 SP. The subsequent analysis of the MST curve ( Fig. 3) revealed a K _d value of 61 ± 7 nM.

Fig. 3.

The binding of peptide Pn and SARS-CoV-2 SP was analyzed by MST. Data points indicate the difference in normalized fluorescence generated by Pn-FAM and SARS-CoV-2 SP. The curve indicates the calculated fit. Error bars represent the standard deviation of three replicates.

3.4. Establishment and condition optimization of p-ELISA assay in direct mode

In this study, the peptide Pn was used as a capture probe, an antibody (anti-SARS-CoV-2 RBD mAb) bound to the RBD domain of S1 subunit was selected as the detection antibody, by which a “peptide-antigen-antibody” sandwich p-ELISA was constructed to detect SARS-CoV-2 SP antigen ( Scheme 1).

Scheme 1.

Schematic illustrating the p-ELISA. (A) direct mode, (B) TSA mode.

To improve the sensitivity of the assay, the main experimental parameters were optimized. In ELISA experiments, the coating conditions are crucial. If proteins or peptides are not immobilized onto the well, or if the coating efficiency is not high, the performance will be directly affected [42]. First, the effect of coating buffers was investigated. It was found that for immobilization of peptide, 50 mM CB buffer (pH 8.6) was significantly better than other buffers ( Fig. 4A). The OD_452 nm reached a peak when peptide dosage increased to 4 µg/mL, which did not increase as the peptide concentration continued to increase (Fig. 4B), indicating that saturation fixation was achieved at 4 µg/mL. The optimal binding time of antigen reached at 1.5 h (Fig. 4C). The detection antibody is the key parameter in ELISA. The OD_452 nm values of the positive group with SARS-CoV-2 S antigen and the negative group without SARS-CoV-2 S antigen were recorded separately to calculate the positive-to-negative ratio (P/N). The highest P/N ratio was observed at 0.2 μg/mL anti-SARS-CoV-2 RBD mAb (Fig. 4D). Interestingly, although the signal of the positive group increased with anti-SARS-CoV-2 RBD mAb dosage, the P/N value decreased, which may be due to the increase in the background value caused by the antibodies. In addition, the signal reached the optimum for anti-IgG-HRP mAb dilution ratio of 1:20000 (Fig. 4E).

Fig. 4.

Optimization of p-ELISA conditions. (A) effect of coating solutions, including phosphate buffered saline (PBS, pH 7.4), Tris–buffered saline (TBS, pH 7.5), TBS (pH 8.0), Na₂CO₃/NaHCO₃ buffer (CB, pH 8.6), and CB (pH 9.6). (B) effect of Pn peptide dosage. (C) effect of incubation time of the antigen. (D) effect of anti-SARS-CoV-2 RBD mAb dose. (E) effect of anti-IgG-HRP dilution ratio. The red lines corresponding to P/N ratio. Error bars represent the standard deviation of three replicates.

3.5. Optimization of a p-ELISA assay with TSA

In order to enhance the sensitivity of peptide ELISA assay, we further introduced TSA technique to p-ELISA (Scheme 1B). A tyramine-biotin complex is involved to the final step of the direct mode of p-ELISA protocol. When biotinylated tyramine is catalyzed by HRP and deposited on the site signal to be amplified, SA-HRP is added to take advantage of the high affinity between streptavidin and biotin, thus introducing more HRP into the wells of the positive microplate where the sample is tested, and producing a stronger color reaction with the chromogenic 3,3′,5,5′‑tetramethylbenzidine. In this way, the signal difference between varying antigen concentrations becomes significantly and the assay sensitivity can be improved.

In order to better utilize biotin-tyramine for signal amplification, we first optimized the conditions based on the direct mode p-ELISA assay. The P/N value reached the highest when 10 μg/mL SA-HRP was applied ( Fig. 5A). The reaction time of biotin-tyramine in the wells is also an important factor. The incubation of 20 min reached the optimal P/N ratio (Fig. 5B). Additionally, the background values did not change much with varying SA-HRP dilution (Fig. 5C). The P/N value reached a plateau at 1:20000 dilution of SA-HRP.

Fig. 5.

Optimization of conditions of p-ELISA in first TSA mode. (A) effect of biotin-tyramine dosage, (B) effect of reaction time, (C) effect of SA-HRP dilution. Error bars represent the standard deviation of three replicates.

3.6. p-ELISA based SARS-CoV-2 SP antigen assay by direct mode or by TSA mode

Based on the optimal conditions, the responses of three p-ELISA detection modes as a function of different SP antigen standard concentrations were tested (Fig.S3). On the basis of Fig.S3, the data points with linear relationships were selected to draw calibration curves, respectively ( Fig. 6). In the direct mode, the ΔOD_452 nm value (the OD_452 nm difference between positive group and the negative blank group) increased with increasing antigen concentration (Fig. 6A), from which ΔOD_452 nm value is linear within 40–10000 pg/mL SARS-CoV-2 SP. In TSA mode, for the first TSA, a calibration curve was established (Fig. 6B) with a linear range of 2–400 pg/mL. In the secondary TSA mode, the linear range of the assay was 0.4–10 pg/mL (Fig. 6C). From Fig.S3, the actual lower limits of detection (LOD) were determined to be 40 pg/mL (280 fM), 2 pg/mL (14 fM) and 0.4 pg/mL (2.8 fM) for direct mode, first TSA mode and secondary TSA mode, separately. Apparently, the detection with TSA mode is more sensitive than direct detection mode. Streptavidin has a very high affinity to bind biotin specifically. Each streptavidin can bind four biotins, and the specific reaction between biotin and SA-HRP can reduce the spatial site hindrance of the reaction and makes the reaction amplify significantly. After the first signal amplification, a large number of HRP and streptavidin are newly bound, and the aromatic amino acid residues on these proteins can provide binding sites for biotinylated tyramine to bind again. In Table S3 we collected the detection performance of different methods for SARS-CoV-2 S antigen. The LOD was reported for aptamer based electrochemical biosensor (66 pg/mL) [43], molecular-imprinted biosensor (100 pg/mL) [19], surface-enhanced resonance Raman scattering (10 pg/mL) [44], nanobody-based ELISA (0.147 ng/mL) [16], and LFIA (0.1 ng/mL) [45]. Obviously the p-ELISA significantly excelled over most methods reported so far. The p-ELISA with secondary TSA model also confirms that these newly bound HRP and streptavidin can provide binding sites for the next round binding of biotinylated tyramine. Actually, even 3rd, 4th, 5th TSA could be performed as well. Thus, after several rounds of such cyclic amplification, a large number of HRP molecules can be bound, resulting in a geometrically amplified or even wirelessly amplified signal for its detection, and for detection sensitivity.

Fig. 6.

SARS-CoV-2 SP detection by p-ELISA with different amplification. (A), direct mode, (B) first TSA mode, (C) secondary TSA mode. Error bars represent the standard deviation of three replicates.

3.7. Sensitivity test of p-ELISA for SARS-CoV-2 pseudovirus

Typically, SARS-CoV-2 pseudovirus is assembled by human immunodeficiency virus type I based on vector infiltration of 293 T cells expressing SARS-CoV-2 SP on its surface [45]. The pseudovirus surface resembles that of SARS-CoV-2, which can enter human cells but cannot replicate itself, allowing for used as a model virus [46]. We inactivated the pseudovirus and then collected nasal sub leachate from healthy individuals for different dilution. The pseudovirus was then tested in the nasal substrate using a p-ELISA with direct mode or TSA mode. Pseudoviruses down to 60 TCID₅₀/mL can be detected by direct mode ( Fig. 7A), despite the fact that other components in the nasal sample interfered with the identification of the analyte. Excitingly, the LOD with 1st TSA is as low as 10 TCID₅₀/mL (Fig. 7B), which is a 6-fold increase in sensitivity compared to the direct mode. After 2nd TSA, as low as 3 TCID₅₀/mL pseudoviruses can be detected (Fig. 7C), suggesting a 20-fold increase in sensitivity compared to the direct mode. Although in terms of sensitivity enhancement effect, pseudovirus detection was weaker than that for SARS-CoV-2 S antigen, which may be due to the fact that pseudovirus was adulterated with nasal subsamples of healthy human. Compared to reverse transcription loop-mediated isothermal amplification (54 TCID₅₀/mL) [47], nanobody-ELISA (16 TCID₅₀/mL) [16] and nanozyme chemiluminescence paper (360 TCID₅₀/mL) [45], our as-proposed p-ELISA method exhibited the best sensitivity for pseudoviruses.

Fig. 7.

SARS-CoV-2 pseudovirus detection by p-ELISA with different amplification. (A), direct mode, (B) first TSA mode, (C) secondary TSA mode. The red dashed line is the LOD determined by the mean value of negative controls (absence of SARS-CoV-2 pseudovirus) plus 3 times the standard deviation. Error bars represent the standard deviation of three replicates.

4. Conclusion

To summarize, a phage monoclonal expressing the peptide with the sequence WNLDLSQWLPPM specific to SARS-CoV-2 SP was obtained through four rounds of screening of phage display. Subsequently, a p-ELISA assay was established based on the selected peptide and TSA, which has excellent sensitivity and great flexibility. The LODs for SARS-CoV-2 S antigen were 40, 2, 0.4 pg/mL for p-ELISA with direct mode, 1st TSA, 2nd TSA mode, respectively. The LODs for pseudoviruses in nasal formula were 60, 10, and 3 TCID₅₀/mL for p-ELISA with direct mode, 1st TSA and 2nd TSA mode, separately. Therefore, this work would provide a novel inexpensive and ultrasensitive method for rapid diagnosis of SARS-CoV-2. Further, the biopanned peptides have shown great potential as "diagnostic antibodies". Finally, the integration of the selected specific peptide with TSA will be prospected in a wider range of applications.

CRediT authorship contribution statement

Junchong Liu: Methodology, Investigation, Writing – original draft. Shuang Pang: Methodology, Investigation, Writing – original draft. Mingyang Wang: Investigation. Haipeng Yu: Investigation. Pengxin Ma: Investigation. Tao Dong : Investigation. Zongmei Zheng: Investigation. Yiming Jiao: Investigation. Yaru Zhang: Investigation. Aihua Liu: Conceptualization, Funding acquisition, Supervision, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Biographies

Mr. Junchong Liu is a graduate student at Qingdao University who is currently pursuing Master’s degree in Bioengineering. His research interests are peptide based biosensing.

Miss Shuang Pang is a graduate student at Qingdao University who is currently pursuing Master’s degree in Microbiology. She is interested in phage display and biosensing application.

Dr Aihua Liu received his PhD degree from Tohoku University, Japan (2004) with major in biosensing. Then he worked in the National Institute of Advanced Industrial Science & Technology, Japan under the Japanese Society for the Promotion of Sciences (JSPS) fellowship (2004–2006). After that he moved to the US and conducted his postdoc research in Michigan State University, University of Oklahoma and University of Texas at Arlington. He was appointed to the Professor of Qingdao Institute of Bioenergy & Bioprocess, Chinese Academy of Sciences, where he led the Laboratory of Biosensing (2010–2016). In 2016, he was appointed as Professor and Director of Institute for Chemical Biology & Biosensing, Qingdao University. His research interests cover microbial surface display, nanozyme and their applications in biosensors, IVD and nanomedicine.

Appendix A. Supplementary material

Supplementary material

.

Data Availability

The authors do not have permission to share data.