A Metasurface Plasmonic Analysis Platform Combined with Gold Nanoparticles for Ultrasensitive Quantitative Detection of Small Molecules

Food safety related to drug residues in food has become a widespread public concern. Small-molecule drug residue analysis often relies on mass spectrometry, thin-layer chromatography, or enzyme-linked immunosorbent assays (ELISA). Some of these techniques have limited sensitivity and accuracy, while others are time-consuming, costly, and rely on specialized equipment that requires skilled operation. Therefore, the development of a sensitive, fast, and easy-to-operate biosensor could provide an accessible alternative to conventional small-molecule analysis. Here, we developed a nanocup array-enhanced metasurface plasmon resonance (MetaSPR) chip coupled with gold nanoparticles (AuNPs) (MSPRAN) to detect small molecules. As sulfamethazine drug residues in poultry eggs may cause health issues, we selected this as a model to evaluate the feasibility of using MSPRAN for small-molecule detection. The MSPRAN biosensor employed competitive immunoassay technology for sulfamethazine detection. The limit of detection was calculated as 73 pg/mL, with sensitivity approximately twice that of previously reported detection methods. Additionally, the recovery rate of the biosensor, tested in egg samples, was similar to that measured using ELISA. Overall, this newly developed MSPRAN biosensor platform for small-molecule detection provides fast and reliable results, facile operation, and is relatively cost-effective for application in food safety testing, environmental monitoring, or clinical diagnostics.


Introduction
With economic improvements, food safety awareness has increased worldwide. However, food safety incidents continue to occur frequently and thus stricter testing and technical requirements are called for by regulatory authorities [1,2]. Many food safety issues are caused by small-molecule drug residues and mycotoxins that can have a negative impact on human health and the environment [3,4]. Because of insufficient monitoring and research, antibiotic residues frequently contaminate eggs and pose health challenges [4]. Sulfamethazine is one of the commonly used antibiotic drugs in poultry breeding. The excessive use of sulfamethazine can leave drug residues in food [5], and their residues in poultry eggs may cause health issues, such as allergic reactions, destruction of the hematopoietic system, and emergence of drug-resistant strains [5,6]. phosphate-buffered saline (PBS) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Sulfamethazine, chloramphenicol, enrofloxacin, ofloxacin, florfenicol, and fipronil were obtained from R-Biopharm Rhone (Darmstadt, Germany). Sulfamethazine antibodies and antigens were obtained from Guangzhou youkangduo Biotechnology Co., Ltd. (Guagnzhou, China). A sulfamethazine ELISA kit was purchased from MLBIO Biotechnology Co., Ltd. (Shanghai, China). All chemicals were of analytical grade and required no further purification.

Fabrication of the MetaSPR Chip and Biosensor
The replica-molding technique was used for biosensor fabrication according to the methods described in our previous reports [24,25]. Firstly, the mold was photo-and ionetched to produce a tapered nanocup array on a silicon wafer containing nanopillars with a diameter, depth, and period of 200, 450, and 300 nm, respectively. Following this, the mold was made hydrophobic using hexylsilane, dried in a vacuum dryer for 12 h, and then a layer of UV-curable polymer solution was applied. A polyethylene terephthalate sheet was placed on top, followed by UV-irradiation for 5 min. The polyethylene terephthalate sheet was then peeled off to form a polymeric nanocone array structure. The MetaSPR chip was mass-produced by depositing multiple layers of metal (15 nm titanium, 70 nm silver, and 20 nm gold layers) using electric spray technology. Finally, the biosensor was fabricated by cutting the MetaSPR chip into 12.4 cm × 8.3 cm sheets, which were affixed to a standard open-bottom 96-well plate, made using a 3D printer (Object 30 primer™; Stratasys Ltd., Rehovot, Israel). The physical and optical properties of the MetaSPR chip were recorded using a smartphone camera, scanning electron microscope, and a generic microplate reader.

Synthesis of AuNP and Labeled Antibodies
Redox reactions were performed to synthesize colloidal AuNPs using a previously reported method [26]. HAuCl4 (300 mL, 1 mM) was heated to boiling point, whereafter 15 mL of 75 mM citrate were added. The mixture was boiled for another 15 min, resulting in a dark red colloidal AuNP solution, and was then cooled to room temperature. The physical properties of AuNP solution were characterized by transmission electron microscopy. The sulfamethazine antibody was immobilized onto the surface of AuNPs using a previously described method [12,27]. Briefly, 1.5 mL colloidal AuNP solution was adjusted to pH 7.4 with 1 M tris (pH 9.27), followed by the addition of different concentrations of sulfamethazine antibodies (final concentrations 0.38 µg/mL, 0.76 µg/mL, 1.52 µg/mL, and 2.28 µg/mL) in 10 mM PBS (pH 7.4) to optimize the labeled AuNP density, and the mixture was incubated for 15 min. Then, 150 µL 10% bovine serum albumin (BSA) was used to block the AuNPs for 15 min, followed by centrifugation at 5400× g for 20 min. The supernatant was removed, and the precipitate was resuspended in R2 buffer (20 mM Tris (pH 7.3), 0.3% sucrose, and 0.05% PEG 20000), which was then stored at 4 • C until further use. In addition, solutions with different antibody labeling densities were evaluated to distinguish between a blank (basic buffer, 20 mM tris, 20 mM EDTA, 0.5% Tween-20) and a sample containing 1 ng/mL sulfamethazine, to determine the optimal labeling density.

Functionalizing the Surface of the MetaSPR Chip
A self-assembly process was adopted for forming carboxyl groups on the surface of the MetaSPR chip to immobilize sulfamethazine antigens. Each chip well was filled with 3-mercaptopropionic acid (50 µL, 50 mM) and left at 37 • C for 30 min, whereafter the wells were washed twice with isopropanol and deionized water. For immobilization, an 1-ethyl-(3-dimethylaminopropyl)-carbodiimide hydrochloride/N-hydroxysuccinimide mixture (50 µL, 400 mM/100 mM) was used to activate the carboxyl groups on the surface of the chip and then placed in 2-morpholinoethanesulphonic acid (pH 5.0) for 5 min at room temperature. Then, 50 µL of 20 µg/mL sulfamethazine antigens were added to the chip wells and left at 37 • C for 15 min. Subsequently, the chip wells were blocked with 130 µL of 1% BSA in PBS and incubated for 30 min at 37 • C. Thereafter, the liquid in the chip wells was discarded. The 96-well biosensor was subsequently dried in an oven at 37 • C for 8 min, sealed, and stored at 4 • C until further use.

Measurement of Small Molecules Using the MSPRAN Biosensor
A competitive immunoassay method was used for small-molecule detection. Sulfamethazine (50 µL, concentrations ranging from 0 to 32 ng/mL) diluted in R1 buffer (20 mM tris, 20 mM EDTA, 0.5% TW20, 1% BSA) was added to the immobilized sulfamethazine antigens in each chip well of the biosensor. The starting point of the reaction was measured using a generic microplate reader at full wavelength (500-700 nm). Then, 5 µL aliquots of AuNP-labeled sulfamethazine antibodies were added to different concentrations of sulfamethazine solution in each chip well, followed by incubation for 10 min at 37 • C in a thermostatic metal bath with shaking (7000 rpm). The terminal point of the reaction was then measured, and the output signal was obtained by subtracting the starting point value from the terminal point.

Small-Molecule Detection in Whole Egg
To verify the capability of the 96-well MSPRAN biosensor to detect small molecules, sulfamethazine-spiked whole eggs were chosen as a real-world sample. Unspiked eggs were used as a control. For spiked sample preparation, 10 mL acetonitrile, 2 g sodium chloride, 3 g anhydrous sodium sulfate, and different final concentrations of sulfamethazine (2 ng/mL, 4 ng/mL, and 16 ng/mL) were mixed with 5 g homogenized egg. The mixture was then centrifuged at 2500× g for 4 min, whereafter 2 mL of the supernatant were collected and incubated with 300 mg anhydrous magnesium sulfate, 100 mg Npropylethylenediamine, 100 mg C18, and 15 mg graphitized carbon black for 30 s with shaking. After incubation, the mixture was centrifuged at 2500× g for 4 min, and 1 mL supernatant was collected for nitrogen drying. The precipitate was redissolved in 500 µL R1 buffer and used for sulfamethazine detection. To compare accuracy, both the MSPRAN biosensor and an ELISA kit were used.

Principle of Small-Molecule Detection Using the MSPRAN Biosensor
When evanescent waves of incident light are coupled with a surface plasmon wave, they begin to oscillate, generating an SPR effect [28]. Due to the periodic nanocup structures of the MetaSPR chip, the SPR effect is produced without the need for another coupling element [16]. The MetaSPR chip consisted of multiple metal layers (15 nm titanium, 70 nm silver, and 20 nm gold layers), among which titanium ensured adhesion between the noble metal and chip substrate, while silver and gold allowed for good plasmon resonance [24,29,30]. Since molecule analysis with the MetaSPR was based on mass transfer, the biosensor required the assistance of other materials. AuNPs served as an amplification tool because they can significantly shorten analysis times due to their coupling effect and high molecular mass [12,31]. The recognition function of AuNPs conjugated to antibodies is mediated by noncovalent adsorption mechanisms, such as ionic, electrostatic, and hydrophobic interactions [32]. Considering this, a competitive immunoassay for smallmolecule detection using the MSPRAN biosensor coupled with the AuNP amplifier was developed ( Figure 1). In this method, small molecules in solution compete with antigens immobilized on the surface of the MetaSPR chip for binding sites on antibody-labeled AuNP. With an increase in small-molecule concentration, the amount of antibody-labeled AuNPs binding to antigens immobilized on the chip decreased, resulting in weak signaling. AuNPs binding to antigens immobilized on the chip decreased, resulting in weak signaling.

Characterization of MetaSPR Chip and AuNPs
According to the image obtained with the camera of a smartphone, the chip showed a high optical response with a rainbow-like color (Figure 2a), indicating high sensitivity. Scanning electron microscopy of the top view of the chip showed a highly uniform nanocone array (Figure 2b), indicating that batches differed little and that the chips had high repeatability for mass production. In addition, the transmission electron micrograph showed a uniform size distribution of AuNPs (Figure 2c), indicating that particle size may have little influence on small-molecule detection. To further evaluate the sensitivity of the MetaSPR chip, different concentrations of sucrose solutions (0% to 5%, corresponding to refractive index (RI) values of 1.3325-1.3404) were used to measure absorption spectra with a generic microplate reader. The MetaSPR was shown to be highly sensitive, because it could measure an RI change of 0.0003 (Figure 3a

Characterization of MetaSPR Chip and AuNPs
According to the image obtained with the camera of a smartphone, the chip showed a high optical response with a rainbow-like color (Figure 2a), indicating high sensitivity. Scanning electron microscopy of the top view of the chip showed a highly uniform nanocone array (Figure 2b), indicating that batches differed little and that the chips had high repeatability for mass production. In addition, the transmission electron micrograph showed a uniform size distribution of AuNPs (Figure 2c), indicating that particle size may have little influence on small-molecule detection. AuNPs binding to antigens immobilized on the chip decreased, resulting in weak signaling.

Characterization of MetaSPR Chip and AuNPs
According to the image obtained with the camera of a smartphone, the chip showed a high optical response with a rainbow-like color (Figure 2a), indicating high sensitivity. Scanning electron microscopy of the top view of the chip showed a highly uniform nanocone array (Figure 2b), indicating that batches differed little and that the chips had high repeatability for mass production. In addition, the transmission electron micrograph showed a uniform size distribution of AuNPs (Figure 2c), indicating that particle size may have little influence on small-molecule detection. To further evaluate the sensitivity of the MetaSPR chip, different concentrations of sucrose solutions (0% to 5%, corresponding to refractive index (RI) values of 1.3325-1.3404) were used to measure absorption spectra with a generic microplate reader. The MetaSPR was shown to be highly sensitive, because it could measure an RI change of 0.0003 (Figure 3a-c). Moreover, after subtracting the optical density (OD) of water from the OD of the different sucrose solutions (Figure 3d-f), light intensity decreases and increases around 570 nm and 590 nm, respectively. Consequently, the OD change served as an output signal through subtraction of two wavelengths (OD590-OD575). The two-wavelength OD values correlated positively with the concentration of the sucrose solutions (R 2 = 0.999). To further evaluate the sensitivity of the MetaSPR chip, different concentrations of sucrose solutions (0% to 5%, corresponding to refractive index (RI) values of 1.3325-1.3404) were used to measure absorption spectra with a generic microplate reader. The MetaSPR was shown to be highly sensitive, because it could measure an RI change of 0.0003 (Figure 3a-c). Moreover, after subtracting the optical density (OD) of water from the OD of the different sucrose solutions (Figure 3d

Assessing Antibody Immobilization and Optimization of Testing Conditions
Before conducting the competitive immunoassays for small-molecule detection the MSPRAN biosensor, the AuNP antibody-labeling density and testing buffer wer timized. Different antibody labeling densities were evaluated for small-molecule d tion, and a final concentration of 2.28 μg/mL sulfamethazine antibody labeled on th face of AuNPs was found to clearly distinguish the blank and a test sample (Figur Thereafter, the basic buffer (20 mM tris, 20 mM EDTA, 0.5%TW20) was added to va reagents to improve the reaction response. However, the addition of basic buffer tested reagents did not show any improvement or interference ( Figure 4b).
Next, different concentrations of BSA were incubated with basic buffer, and th dition of BSA appeared to improve the response unit significantly (approximately 6 ( Figure 4c). The reaction value with addition of 1% BSA was the highest and allowe a clear distinction between the blank and the test sample. Therefore, an R1 buffer con ing BSA (20 mM tris, 20 mM EDTA, 0.5%TW20, and 1%BSA) was selected for the com itive immunoassay.

Assessing Antibody Immobilization and Optimization of Testing Conditions
Before conducting the competitive immunoassays for small-molecule detection using the MSPRAN biosensor, the AuNP antibody-labeling density and testing buffer were optimized. Different antibody labeling densities were evaluated for small-molecule detection, and a final concentration of 2.28 µg/mL sulfamethazine antibody labeled on the surface of AuNPs was found to clearly distinguish the blank and a test sample (Figure 4a). Thereafter, the basic buffer (20 mM tris, 20 mM EDTA, 0.5% TW20) was added to various reagents to improve the reaction response. However, the addition of basic buffer to all tested reagents did not show any improvement or interference (Figure 4b).

Assessing Antibody Immobilization and Optimization of Testing Conditions
Before conducting the competitive immunoassays for small-molecule detection using the MSPRAN biosensor, the AuNP antibody-labeling density and testing buffer were optimized. Different antibody labeling densities were evaluated for small-molecule detection, and a final concentration of 2.28 μg/mL sulfamethazine antibody labeled on the surface of AuNPs was found to clearly distinguish the blank and a test sample (Figure 4a). Thereafter, the basic buffer (20 mM tris, 20 mM EDTA, 0.5%TW20) was added to various reagents to improve the reaction response. However, the addition of basic buffer to all tested reagents did not show any improvement or interference (Figure 4b).
Next, different concentrations of BSA were incubated with basic buffer, and the addition of BSA appeared to improve the response unit significantly (approximately 6-fold) (Figure 4c). The reaction value with addition of 1% BSA was the highest and allowed for a clear distinction between the blank and the test sample. Therefore, an R1 buffer containing BSA (20 mM tris, 20 mM EDTA, 0.5%TW20, and 1%BSA) was selected for the competitive immunoassay.  Next, different concentrations of BSA were incubated with basic buffer, and the addition of BSA appeared to improve the response unit significantly (approximately 6-fold) (Figure 4c). The reaction value with addition of 1% BSA was the highest and allowed for a clear distinction between the blank and the test sample. Therefore, an R1 buffer containing BSA (20 mM tris, 20 mM EDTA, 0.5% TW20, and 1%BSA) was selected for the competitive immunoassay.

Establishment of a Standard Curve for Small-Molecule Quantitative Analysis
Sulfamethazine is a synthetic antibiotic with broad-spectrum antibacterial activity, good curative effect, and low cost and is widely used in the poultry industry, with residues in food posing health issues. Therefore, sulfamethazine was selected as a model drug for assessing the ability of the 96-well MSPRAN biosensor to detect small molecules. The MSPRAN biosensor can be detected by a general microplate reader or a WeSPR100 Multifunctional Molecular Analyzer (Shanghai, China), making it cost-effective and applicable to resource-poor areas and private laboratories. With an increase in sulfamethazine concentration, OD values increased at around 581 nm and decreased at around 608 nm wavelengths (Figure 5a). In addition, the two-wavelength OD values (OD 608 -OD 581 ) decreased in a dose-dependent manner from 0 ng/mL to 32 ng/mL (Figure 5b). There was a good correlation between the concentration and these two-wavelength OD values according to a four-parameter logistic (4-PL) regression equation (R 2 = 0.999; Figure 5c), and that equation can serve as a standard curve to calculate the limit of detection (LOD) value. The equation was as follows: Sulfamethazine is a synthetic antibiotic with broad-spectrum antibacterial activity, good curative effect, and low cost and is widely used in the poultry industry, with residues in food posing health issues. Therefore, sulfamethazine was selected as a model drug for assessing the ability of the 96-well MSPRAN biosensor to detect small molecules. The MSPRAN biosensor can be detected by a general microplate reader or a WeSPR100 Multifunctional Molecular Analyzer (Shanghai, China), making it cost-effective and applicable to resource-poor areas and private laboratories. With an increase in sulfamethazine concentration, OD values increased at around 581 nm and decreased at around 608 nm wavelengths (Figure 5a). In addition, the two-wavelength OD values (OD608-OD581) decreased in a dose-dependent manner from 0 ng/mL to 32 ng/mL (Figure 5b). There was a good correlation between the concentration and these two-wavelength OD values according to a four-parameter logistic (4-PL) regression equation (R 2 = 0.999; Figure 5c), and that equation can serve as a standard curve to calculate the limit of detection (LOD) value. The equation was as follows: In addition, an ELISA kit was used to verify the MSPRAN biosensor by evaluating the correlation between a sulfamethazine concentration (0-162 ng/mL) and reaction signal (Figure 5d), which yielded R 2 = 0.999, a result that was similar to the result obtained with the MSPRAN biosensor. These results showed that the MSPRAN biosensor could serve as a tool for the quantitative detection of sulfamethazine. Moreover, this result indicates that further studies are warranted to investigate the capabilities of this technique for real sample analysis.

MSPRAN Performance Evaluation for Sulfamethazine Detection
To verify the sensitivity of the MSPRAN biosensor, the LOD values were calculated from the standard curve according to methods described in a previous study [33]. The OD value of a blank sample minus three times the standard deviation was interpolated into the standard curve to obtain the corresponding concentration values, which served as the In addition, an ELISA kit was used to verify the MSPRAN biosensor by evaluating the correlation between a sulfamethazine concentration (0-162 ng/mL) and reaction signal (Figure 5d), which yielded R 2 = 0.999, a result that was similar to the result obtained with the MSPRAN biosensor. These results showed that the MSPRAN biosensor could serve as a tool for the quantitative detection of sulfamethazine. Moreover, this result indicates that further studies are warranted to investigate the capabilities of this technique for real sample analysis.

MSPRAN Performance Evaluation for Sulfamethazine Detection
To verify the sensitivity of the MSPRAN biosensor, the LOD values were calculated from the standard curve according to methods described in a previous study [33]. The OD value of a blank sample minus three times the standard deviation was interpolated into the standard curve to obtain the corresponding concentration values, which served as the LOD value. The LOD for sulfamethazine detected by the MSPRAN biosensor was  Table 1. LOD value of this study was lower than those of other detection methods, indicating that the MSPRAN biosensor was highly sensitive. In addition, substrate selectivity was also assessed by testing different non-specific small molecules, such as chloramphenicol, enrofloxacin, ofloxacin, florfenicol, and fipronil, at a concentration of 50 ng/mL. None of the non-specific molecules tested exhibited a measurable reaction on the sulfamethazine detection platform (data not shown). In addition, repeatability was tested by performing measurements at different concentrations of sulfamethazine, and data were characterized as the coefficient of variation (CV) values. The CV values measured for various concentrations of sulfamethazine (2 ng/mL, 4 ng/mL, 16 ng/mL) were 7.07%, 5.63%, and 5.77%, respectively. Overall, these results showed that the 96-well MSPRAN biosensor displayed favorable performance during sulfamethazine detection.

Sulfamethazine Analysis in Whole Eggs
The practical application of the 96-well MSPRAN biosensor was evaluated using sulfamethazine-spiked whole eggs. Recovery rates of 87.29-119.30% were obtained for sulfamethazine detection using the MSPRAN biosensor, which were similar to the recovery rates obtained with ELISA ( Table 2). The experimental steps of ELISA for sulfamethazine detection in whole eggs are shown in the Supplemental Materials. The consistency of recoveries between the 96-well MSPRAN biosensor and ELISA indicates that the biosensor could be used as a reliable alternative for small-molecule detection. In addition, the ELISA-based detection of sulfamethazine requires more than 1 h, while the detection with MSPRAN required only 10 min; thus, the MSRPAN approach facilitated fast measurements and high-throughput. Moreover, the operational steps of ELISA are complex, requiring five sample addition steps and multiple washing steps. However, our approach only required two sample addition steps and no washing steps, and can be performed following the manufacturer's instructions, without specific training. Therefore, the developed MSPRAN biosensor was easy to operate. Taken together, the developed MSPRAN biosensor shows multiple advantages for small-molecule detection.

Conclusions
In this study, an MSPRAN biosensor for small-molecule detection was developed using a competitive immunoassay. The biosensor was designed to have low cost and portability, and demonstrated fast reads, high sensitivity, and high accuracy. The detection of sulfamethazine using the MSPRAN biosensor could be completed within only 10 min.
Overall, our results demonstrated that the performance of the MSPRAN is comparable to that of the corresponding ELISA, a more widely used technique for small-molecule analysis. However, a limitation of the developed MSPRAN biosensor for sulfamethazine detection is that the antibody needs to be conjugated to the surface of AuNPs, which is a complicated and time-consuming operation. In the future, the sensitivity of the MetaSPR chip can be improved to detect small molecules directly by immobilizing antibodies on the chip surface. The novel MSPRAN biosensor for small-molecule detection developed here has not been reported previously and may pave the way for small-molecule detection based on a MetaSPR chip for use in food safety control and environmental monitoring. In addition, this approach holds potential for small/macro-molecule multiplexing analysis using a MetaSPR chip that is modified with multichannel plates with more than 96 wells. This could then be applied for large-scale food safety monitoring, environmental monitoring, and medical diagnoses.
Supplementary Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/bios13070681/s1, Figure S1: Sulfamethazine detection of blank eggs using LC-MS; Figure S2: The standard curve of varies concentration of sulfamethazine (0-72 ppb) spiked in eggs according to response unit detected by the MetaSPR Biosensor; Figure S3: The standard curve of different concentration of sulfamethazine (0-162 ppb) spiked in eggs corresponding to reaction signal detected by an ELISA kit.