A novel method for visual microarray detection of antibiotic resistance genes

Abstract

Antibiotic resistance genes (ARGs) are emerging pollutants that pose significant threats to both the environment and human health. Rapid detection of ARGs is essential for monitoring their levels and controlling their spread. However, traditional detection methods are often time-consuming and require specialized equipment, leading to delays. To address this issue, this study developed a novel method for visual microarray detection of ARGs, including sul1, tetA, qepA, macB, and vanR. This method employs ARG-specific dual probes combined with silver staining signal cascade amplification technology. A centrifuge-free concentration device has been developed that can directly enrich nucleic acid fragments from environmental samples, with traditional nucleic acid extraction and PCR amplification being eliminated. Using a dual probe combined with silver staining enhancement dual probe amplification technology, rapid high-throughput detection of low-concentration ARG is achieved. No specialized equipment is required, and on-site visual detection of ARG was realized. The detection method can detect target genes (sul1, tetA, qepA, macB, and vanR) within a concentration range of 0.411 µg mL⁻¹ to 55.9 µg mL⁻¹, with the lowest detectable target gene concentration being 0.411 µg mL⁻¹. This study marks the first integration of a centrifuge-free concentration device, specific dual probes, and silver staining signal cascade amplification technology to establish a rapid visual detection method for ARGs using a microarray. The developed detection method holds significant potential not only for the detection and monitoring of ARGs but also as a prototype for devising future detection strategies.

---

Introduction

Antibiotic resistance genes (ARGs) are emerging pollutants that can transfer between bacteria and human pathogens.^1–3 This transfer poses significant threats to both public health and ecosystems.^4,5 Research indicates that antibiotic resistance is responsible for approximately 700 000 deaths worldwide each year.⁶ Without urgent intervention, this figure is projected to rise to around 10 million by 2050.⁷ The World Health Organization (WHO) has recognized it as one of the most significant threats to human health in the 21st century.^8,9 In recent years, the spread of ARGs has accelerated due to the increased discharge of antibiotics and other pollutants, such as heavy metals, into the environment.^10–12 As a result, the environmental impact of ARGs, including their role in antibiotic-resistant bacteria (ARB), has become a major focus in environmental science.^13,14 Monitoring and controlling ARGs is crucial. Among these efforts, the detection of ARGs plays a particularly critical role.

Nucleic acid detection methods, including ultraviolet spectrophotometry, fluorescent dye method, and polymerase chain reaction (PCR), each have their own advantages and significant drawbacks.^15–17 For instance, PCR technology can efficiently amplify low-abundance nucleic acids of specific sequences, but it may produce false positive signals, and its primer design is complex, with the equipment being expensive.¹⁸ The fluorescent dye method offers high sensitivity and repeatability; however, the nucleic acid fluorescent dyes are toxic, and the test results are influenced by the binding rate of the fluorescent dye to the nucleic acids.¹⁶

In comparison, traditional detection methods for ARGs in the environment primarily include fluorescent quantitative PCR (qPCR) and metagenomic sequencing.^19–21 Metagenomic sequencing allows for a comprehensive analysis of the diversity and abundance of ARGs in environmental samples through gene capture. In contrast, qPCR, particularly high-throughput quantitative PCR (HT-qPCR), offers better detection limits, reduced sample requirements, and the ability for absolute quantification. However, current detection methods still face some urgent challenges, especially in terms of detection time limits, simplicity, and cost. For instance, qPCR requires high sample purity, precise primer design, and appropriate sample handling.²² Covalently closed circular DNA, such as plasmids, may fail to fully denature during the pre-denaturation stage, leading to an underestimation of ARGs abundance. Furthermore, inter-laboratory methodological differences continue to challenge qPCR standardization.²³ While metagenomic sequencing does not require prior sequence information and enables comprehensive detection of both known and unknown ARGs, it has certain limitations, including sample quality requirements and high sequencing costs.²⁴

Despite the advances in traditional ARG detection techniques such as PCR and metagenomic sequencing, challenges remain in terms of time, cost, and the need for specialized equipment. The use of signal amplification technologies has proven effective in overcoming some of these limitations, but achieving a low detection limit without expensive equipment has remained difficult. To address these issues, this study introduces a centrifuge-free concentration device and a dual probe signal amplification method for detecting ARG. This new approach not only simplifies the process but also enhances the sensitivity of ARG detection, making it feasible for on-site, rapid, and high-throughput applications in environmental monitoring. The soil samples for ARG detection were collected from the breeding farm of Gansu Kangmei Modern Animal Husbandry and Agriculture Industry Group in Fucheng Town, Kangle County, Gansu Province, China. These samples were subjected to metagenomic analysis and quantitative comparison with the proposed method. Based on the results of the metagenomic analysis and the distribution of ARGs in previous studies, five ARGs were selected for method validation: sulfonamide sul1, tetracycline tetA, quinolone qepA, macrolide macB, and vancomycin vanR.^25,26

Materials and methods

Materials

The PCR amplification kit and DNA marker were purchased from Sangon Biotech Co., Ltd (Shanghai, China). The nucleic acid extraction kit (E.Z.N.A.™ Mag-Bind Soil DNA Kit) was purchased from Omega Bio-Tek (OMEGA, USA). The 0.22 µm microporous membrane, AgNO₃, and bovine serum albumin (BSA) were obtained from Takara Biotech Co., Ltd (Dalian, China). APTES was purchased from Macklin Biochemical Co., Ltd (Shanghai, China). Glass slides, with a product size of 25.4 × 76.2 mm, were purchased from Shanghai Machinery Import & Export Group Co., Ltd (Shanghai, China). Hydroquinone, glutaraldehyde, and chloroauric acid were purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd (Shanghai, China). HindIII and EcoRI were obtained from Bioneer Corporation (Shanghai, China). tris–EDTA (Tromethamine–Ethylenediaminetetraacetic Acid) was procured from Sangon Biotech Co., Ltd (Shanghai, China). Sodium chloride (NaCl) was sourced from Tianjin Damao Chemical Reagent Co., Ltd (Tianjin, China). The dual probes, target genes, and detection probes are to be diluted to a concentration of 100 mM in ultrapure water. The magnetic beads were homemade with a particle size of 200 nm. The TEM characterization of the magnetic beads is shown in SI Fig. S1.

Experimental methods

The probe and PCR primers were designed. ARG sequences of sul1, qepB, macB, tetA, and vanA were obtained from GenBank. The specific probes (carboxylated modified dual probes and thiolated modified detection probes) and PCR primers were designed using Primer Premier 5.0 software. Both the probes and primers were synthesized by Sangon Biotech Co., Ltd (Shanghai, China). The designed probes and PCR primers are shown in SI Table S1. The designed dual probes were connected to the aminated array substrate at both ends to obtain a microarray chip, while the other two ends of the dual probes were specifically complementary connected to the forward and reverse strands of the target genes. Ultimately, the dual probes, target genes, and detection probes were interconnected to form a sandwich structure, playing a role in signal amplification. Fig. 1 shows the schematic diagram of the designed detection probe.

Fig. 1 Principle diagram of the centrifuge-free concentration device and microarray detection.

Sample collection and processing. Soil samples were collected from three locations at the breeding farm in Fucheng Town, Kangle County, Gansu Province, China, which is under the Gansu Kangmei Modern Animal Husbandry and Agriculture Industry Group. The extracted soil genomic DNA was then pooled into a single sample, amplified by PCR, and gel-purified. The concentration of the recovered fragment was determined using a NanoDrop instrument (Eppendorf, Germany), and the purified gene was used as the standard for establishing the detection method. At the same time, the DNA was sent to Shanghai Sangon Biotechnology Co., Ltd for metagenomic analysis. The PCR was performed using an improved 20 µL reaction system as described in the literature: 10 µL was mixed, primers (1 µL each, 100 µM) were added, template DNA (2 µL, 100 µM) was included, and ddH₂O (6 µL) was added. The amplification conditions were: pre-denaturation at 94 °C for 5 min; denaturation at 94 °C for 30 s, annealing at 53–55 °C for 30 s, and extension at 72 °C for 30 s, repeated for 35 cycles; finally, extension was carried out at 72 °C for 10 min.²⁷

A centrifuge-free concentration device, based on ultrafiltration membranes and magnetic beads, was developed. This device allows for direct enrichment of ARGs from soil without the need for total DNA extraction, enabling detection in real samples. The procedure is as follows: 1.0 g of soil is placed into a 1.5 mL centrifuge tube, 500 µL of 0.85% NaCl is added, and the mixture is vortexed 30–50 times to suspend the particles and then allowed to settle.^28,29 The supernatant is transferred into the enrichment device, 500 µL of lysis solution (50 mM tris-HCl, pH 8.0, 100 mM EDTA, pH 8.0, 2% SDS) is added, the solution is mixed thoroughly and incubated at 65 °C for 15 min. Subsequently, 5 µL of EcoR I and 5 µL of Hind III are added sequentially, and digestion is performed for 5 min. Thirty microlitres of 10 mg per mL magnetic beads are added, the mixture is allowed to bind for 3 min, and impurities are removed. Elution is carried out with 100 µL of TE buffer (10 mM tris-HCl, pH 8.0, 1 mM EDTA, pH 8.0) for 10 min; after magnetic separation, the supernatant is collected, and purified nucleic acid fragments are obtained. The principle of the device is illustrated in the Fig. 1.

Preparation of colloidal gold nanoparticles (AuNPs). The synthesis of AuNPs was refined based on prior research.^30,31 A solution of 1 mmol per L HAuCl₄ and 38.8 mmol per L sodium citrate was prepared. HAuCl₄ was heated and stirred for 10 minutes, after which the sodium citrate solution was rapidly introduced and mixed for an additional 15 minutes. The AuNPs solution was subsequently procured through filtration. The prepared AuNPs were characterized by TEM and UV-vis spectroscopy.

Preparation of aldehyde slides and morphological characterization. The preparation of aminated glass slides refers to the literature and was improved. The slides underwent ultrasonication with acetone and ethanol, each for a duration of 30 minutes, followed by washing using triple-distilled water. Activation of the slides was achieved by immersing them in a concentrated solution of H₂SO₄/H₂O₂ at a ratio of 3 : 1 for one hour at a temperature of 80 °C, utilizing a water bath. Post-activation, the slides were rinsed with ethanol. Amination of the slides was accomplished by immersion in a mixture of APTES and ethanol at a ratio of 1 : 9 for two hours, followed by ethanol washes twice, each lasting five minutes, and subsequent baking at 100 °C for 45 minutes. The aminated glass slides were then immersed in 5% glutaraldehyde for 30 minutes to produce aldehyde-modified slides, which were washed with deionized water three times and stored submerged in water.^32,33 Images of the array substrate surface were captured using the Being-Nano Atomic Force Microscope, manufactured in Guangzhou, China. Photographs of both spotted and blank slides were obtained using a gel imaging system from Bio-Rad Laboratories, located in Hercules, California, USA.

Preparation of microarrays. The ARGs dual probes were spotted onto the surface of aminolated glass slides, with a spot spacing of 1 cm, a spot volume of 1 µL, and three replicates for each concentration. They were then fixed in a water bath at 37 °C for 12 h. Unreacted amino groups were blocked with BSA. After washed three times, a microarray was obtained. Due to the special properties of the array substrate, dual probes labeled with fluorescent Cy5 were simultaneously spotted onto the surface of aminated glass slides to verify the feasibility of coupling the dual probes to the array substrate. A 1.5 µL aliquot of the dual probe was applied to each microarray, and the dual probe concentrations were set to 100 mM, 10 mM, 1 mM, and 0 mM.

Mark of detection probes. The detection probe was labeled onto centrifugally resuspended AuNPs to form AuNPs probe. AuNPs probe labeling method: 1 mL of 10 nmol per L AuNPs solution was taken and centrifuged at 9000 rpm for 5 minutes, the supernatant was discarded, and 100 µL of deionized water was added to resuspend. Then, 5 µL of detection probe with a concentration of 100 µmol per L was added to achieve a final probe concentration of 5 µmol per L, and it was left at room temperature overnight. Subsequently, 1 mol per L NaCl and 0.1 mol per L PB⁻¹ (pH 7.2) were gradually added in 3 portions, up to final concentrations of 0.1 mol per L (optimal for NaCl) and 0.01 mol per L, respectively, thoroughly mixed, and left at room temperature for 48 h. The mixture was washed twice with 0.01 mol L⁻¹ PB⁻¹ (0.1 mol per L NaCl) solution, centrifuged at 9000 rpm for 5 minutes, then the supernatant was discarded, and 100 µL of 0.1 mol per L PB was added to resuspend. Afterward, unreacted aldehyde groups were blocked with 10 mg per mL BSA, and the mixture was placed in a 37 °C water bath for 1 h. It was then washed twice with 0.1 mol per L PBST (1 × PBS + 0.05% Tween) for 5 minutes each, and finally resuspended in 100 µL of 0.1 mol per L PB, stored at 4 °C for future use. The prepared gold nanoparticle probes were characterized using a Japan Electron (JEOL Ltd) JEM-2100 TEM and an Agilent Synergy H1M2 enzyme plate reader (California, USA).

Hybridization of oligonucleotides and silver staining signal amplification. The PCR products were diluted at concentrations of 10⁰, 10¹, 10², 10³, and 10⁴, and then hybridized with AuNPs probes for microarray detection. Three replicates for each concentration. After silver staining to amplify the signal, the results were observed. First, 20 µL of the chip hybridization solution was dropped onto the sample area and covered with a porous cover slip. The chip was placed in a special hybridization box with a small amount of ddH₂O for humidity control and kept in a 42 °C water bath for 3 h. Each slide included a negative control using non-target gene DNA. Next, in a preheated washing solution (0.1% SDS, 0.03 M NaCl, 0.3 mM SSC) at 42 °C, the chip was gently shaken in a dark room for 5 min to be cleaned, rinsed with ddH₂O, and then dried with nitrogen gas. Silver-staining color development was improved by eliminating light-sensitivity artifacts and background darkening. To achieve this, 300 µL of citrate buffer (pH 3.5) and 150 µL of hydroquinone (final concentration 20 mmol L⁻¹) were added and mixed thoroughly. Subsequently, 50 µL of 1% (w/v) AgNO₃ solution was added, and the mixture was allowed to develop at room temperature for 12 min. The reaction was terminated by addition of pure water, and the results were observed immediately.

The nucleic acid fragment samples obtained from the enrichment device were spotted onto the row of chips bearing immobilized probes, with three replicates for each immobilized probe, and a volume of 1.5 µL per spot. Hybridization and silver staining were subsequently performed according to the previously reported protocol, and the visually detectable silver-staining signals of the genes in the environmental samples were obtained.

Result and discussion

Preparation of AuNPs and aminated slide AuNPs

AuNPs are characterized by excellent stability, good biocompatibility, and non-toxicity, making them suitable for biological detection.^34–36 Moreover, the large surface area of AuNPs can increase the number of surface nucleic acid probe connections, thereby improving detection performance. This allows for the detection of low-concentration DNA target samples without the need for amplification.^37,38

AuNPs were prepared from HAuCl₄. The particles were found to be 14–20 nm in diameter, exhibited a single UV-absorption maximum at 520 nm, and were observed to be uniform, sharply defined, and well dispersed (Fig. 2b and d), which can be further used for connecting nucleic acid probes for signal transduction. Previous studies have also shown that the unique characteristics of Nanosurface Energy Transfer (NSET) of AuNPs make it possible to occur Surface Plasmon Resonance (SPR) in the visible-to-infrared wavelength range, which can be detected in the range of 515–560 nm.^39,40

Fig. 2 TEM graphs of AuNPs and AuNPs probes (b and d are AuNPs, a and c are AuNPs probes).

AuNPs can also be made into a variety of shapes, including spherical, rod-shaped, nanoshell, nanopyramid, ring-shaped, disc-shaped, and nanobowl-shaped, etc., with different optical properties.^41–44 Among them, spherical AuNPs are commonly used in biological detection due to their good biocompatibility, high specific surface area, and low toxicity.⁴⁵

Preparation of AuNPs probes

Due to the concentration of NaCl leading to the aggregation of AuNPs, it interferes with the detection background.⁴⁶ Therefore, before the connection of AuNPs with nucleic acid probes, the optimal settling concentration of AuNPs in NaCl solution was determined. The experimental results show that when the concentration of NaCl is 100 mM, the dispersion of AuNPs is better, and they can maintain a stable red color. When the NaCl solution is between 200−800 mM, the AuNPs solution begins to aggregate, and the color gradually deepens. Moreover, the absorbance at a wavelength of 520 nm also continuously decreases. Based on the UV-vis absorption spectra and the color change of AuNPs, 100 mM is determined as the optimal NaCl concentration (Fig. 3).

Fig. 3 Stability of AuNPs in NaCl Solution. (A) Absorption spectra of AuNPs in NaCl solutions at different concentrations. (B) Aggregation of AuNPs in NaCl solutions.

Labeling nucleic acid probes onto AuNPs results in the formation of gold nanoparticle probes used for connecting target gene detection to drug-resistant genes. TEM characterization of the prepared gold nanoparticle probes shows that, as seen from Fig. 2a and c, there is a layer of material around the labeled probes on the AuNPs, causing the particle size to increase. The AuNPs show aggregation within a certain range, as anticipated.⁴⁷ Labeling nucleic acid probes on the surface of AuNPs leads to an increase in the particle size, and the corresponding UV absorption peak exhibits a trend towards redshift in the long-wavelength direction. Fig. 4 results show that the maximum absorption peak of AuNPs is at 520 nm; after labeling with the AuNPs probes, the maximum absorption peak is at 530 nm. The wavelength of the maximum absorption peak of AuNPs labeled with nucleic acid probes shifted by about 10 nm. This is due to the labeling of nucleic acid probes onto the surface of AuNPs, altering the size and shape of the AuNPs, causing the surface plasmon resonance of individual AuNPs particles to become coupled and shift the absorption spectrum, thereby changing the spectral characteristics of the AuNPs.^48,49

Fig. 4 UV-vis spectrum of the as-prepared AuNPs probe.

Similarly, the use of adsorption peak shifts has been observed in other studies. For example, Dong et al. demonstrated a colorimetric method for Cd(II) detection using chloroform-acidified silver nanoparticles, where the absorption peak shifted from 396 to 522 nm upon particle aggregation.⁵⁰ Likewise, Wang et al. developed a sulfide sensor using triangular gold nanoplates, where the localized surface plasmon resonance (LSPR) peak redshifted in proportion to sulfide concentration, with a detection limit of 16 nM.⁵¹

Aminosilane slide preparation and carboxylic acid probe tethering

The surface of the silicon-based glass slide was aminated to enrich it with amino functional groups. An arrayed substrate was prepared by connecting the aminated and carboxylated substrate probes. Fig. S2 shows the atomic force microscope (AFM) characterization image of the aminated substrate. It can be seen that the untreated glass slide surface is smooth and flat without obvious protrusions, while the aminated glass slide surface has granular protrusions, and the particles are evenly distributed. The difference in the surface structure of the glass slides, before and after aldehyde treatment, is evident. The reason for this difference is that after the aldehyde treatment, a thin film forms on the surface of the glass slide, and the surface of this thin film has chemical bonds and physical protrusions (Fig. 4).⁵² The amination process functionalizes the surface of the glass slide, increasing its surface area, which is beneficial for connecting a larger number of substrate probes and for detecting lower concentration DNA target samples.

The designed substrate probe (5-carboxylic acid linkage) was connected to the aminated substrate, and due to the unique structure of the substrate, the fluorescent substrate probe (3-linkage Cy5) was simultaneously synthesized on the aminated glass slide to verify the feasibility of the connection. The feasibility results are shown in Fig. 5, where we connected different concentrations of fluorescent substrate probes to the aminated substrate array. Fig. 5a–d are fluorescent images under a 200 µm scale, while Fig. 5e–h are also fluorescent images under a 200 µm scale. Fig. 5 shows that the fluorescence intensity of different concentration probes varies, indicating good connection efficiency. It can also be observed that the dual probe concentration of 100 mM, with a drop volume of 1.5 µL, binds effectively with the microarray.

Fig. 5 Fluorescent probes connected to aminated slides. The fluorescent probe concentrations were set to 0 mM (a and e), 100 mM (b and f), 10 mM (c and g), and 1 mM (d and h) (a–d), 200 µm; (e–h), 500 µm; (a and e) are control (CK)).

Sample DNA extraction and PCR

To assess the viability of the detection method, we utilized the total DNA derived from environmental samples as a template for PCR amplification of the target genes. Subsequent to this, we subjected the amplified results to gel electrophoresis on a 1.5% agarose gel, conducted at 100 V for a duration of 30 minutes. We observed these results through the use of a gel imaging system. The findings of the PCR product gel electrophoresis are depicted in Fig. 6. The electrophoresis revealed distinct bands with the following band sizes: qepA at 216 bp, vanR at 328 bp, sul1 at 210 bp, macB at 247 bp, and tetA at 280 bp. Notably, these sizes align with the fragment lengths that were anticipated. Additionally, the PCR products were recovered from the gel and their gene concentrations were evaluated using a nucleic acid protein analyzer. The measured concentrations were as follows: sul1 at 55.9 µg mL⁻¹, tetA at 52.5 µg mL⁻¹, qepA at 41.1 µg mL⁻¹, macB at 42.7 µg mL⁻¹, and vanR at 52.5 µg mL⁻¹.

Fig. 6 Gel electrophoresis of antibiotic ARG PCR amplification.

Detection of ARGs by microarray.
Feasibility study. The target genes obtained from PCR were diluted in a gradient of 10⁰, 10¹, 10², 10³, and 10⁴, then hybridized with the prepared AuNPs probes on the surface of microarray chips. After silver staining, the results were observed with the naked eye, and the experimental outcomes are shown in Fig. 7a. The results demonstrate the effectiveness of the method in detecting ARGs by hybridizing target gene sequences with AuNPs probes on microarray chips followed by silver staining. This process is simple and fast, enabling visual detection of ARGs. The microarray detection method can identify target genes within a concentration range of 0.411 µg mL⁻¹ to 55.9 µg mL⁻¹, indicating high sensitivity. The lowest detectable target gene was diluted 10² times at 0.411 µg mL⁻¹. The hybridization results on the microarrays were measured using the Bio-Rad tri-color imaging system. Fig. 7b shows that the hybridization of five target gene concentrations with AuNPs probes and microarray silver staining all yielded positive results. As the concentration of the target genes decreases, the positive signals weaken, and the peak values decrease.

Fig. 7 Microarray detection of resistance gene sensitivity. (a) Shows the microarray silver staining results, and (b) shows the result from the Bio-Rad tri-color imaging system. Numbers 1–5 represent the 10⁴–10⁰ oligonucleotide gradient dilutions, and CK represents the 0 oligonucleotide gradient dilutions. Three replicates were performed for each concentration.

Anti-interference capability. The specificity of the method was tested by sul1 and tetA gene target sequences and microarray chip hybridization. In Fig. 8a, the array probe and AuNPs probe use the sul1 gene sequence, and the target sequence is a different gene sequence; in Fig. 8b, the array probe and AuNPs probe use the tetA gene sequence, and the target sequence is a different gene sequence. The results showed that only the microarrays with added sul1 and tetA gene target sequences displayed positivity, with no cross-reaction with non-target genes, indicating that the detection method has high specificity.

Fig. 8 Array slide detection method specificity (a) is the sul1 detection result, and (b) is the tetA detection result).

The colorimetric assay, as a simple, low-cost, and sensitive detection method, requires no expensive instruments and complex operations.¹⁸ The commonly used materials in colorimetric assays are AuNPs, G-quadruplex, magnetic particles (MPs) and graphene oxide (GO).^53,54 In comparison, AuNPs are the ideal materials for nucleic acid detection due to their versatile interactions with nucleic acids and more obvious color changes.^55,56 The results can be observed directly by the naked eye without the assistance of equipment.⁵⁷

Microarray applications. The ARGs glass slide probes of five types were connected on the same substrate through the microarray preparation method mentioned earlier, enabling the simultaneous detection of five ARGs. Subsequently, the DNA extracted from environmental samples was used as the target probe, and the five ARGs were detected using the detection method of this study. At the same time, metagenomics was employed to compare the concentration of ARGs in the soil. The microarray silver staining results showed that the method in this study could visually detect ARGs, and the results were visible to the naked eye.
Microarray detection of the actual sample. The use of microarrays to detect ARGs in soil samples from farms is shown in Fig. 9. Among them, simple probe represents the results of single probe detection (Fig. 9a), while dual probe represents the results of dual probe detection (Fig. 9b). It can be seen that the results of the dual probe are more pronounced than those of the single probe. The dual probe can enhance the binding efficiency of the target gene, thereby achieving a signal amplification effect. The microarray detection results are analyzed through peak detection of the gel imaging system and compared with the sensitivity experiment results in Section 3.5 to understand the pollution level of ARGs (Fig. 9c). Due to the stability of the microarray chip preparation process and the stability of probe binding with each other, there may be errors in the comparison results. This point has also been mentioned in previous studies.^58,59

Fig. 9 Microarray detection of ARGs in breeding farm (a and b) is the result of microarray, (c) is the result of metagenomic sequencing).

While the Bio-Rad tri-color imaging system effectively presents the results in the laboratory, we acknowledge its lack of portability for on-site applications. To address this limitation, future improvements may focus on developing portable, handheld devices capable of directly quantifying the silver-stained signals on-site, enabling faster and more convenient ARG monitoring. Recent advancements in AI and smartphone-assisted fast and quantitative readout technologies have shown significant potential for rapid and quantitative detection in diagnostic environments. For instance, Prasad et al. demonstrated the feasibility of using smartphone-based systems to detect chemical and biological agents, with high accuracy and portability.⁶⁰ Similarly, Du et al. emphasized the importance of using AI algorithms to enhance the sensitivity and speed of detection methods.⁶¹ Integrating such technologies into ARG detection platforms could significantly improve the speed and convenience of monitoring, making these methods more suitable for on-site applications.

Comparison of the array slide detection method. The DNA extracted from environmental samples was subjected to metagenomic detection and compared with the silver staining intensity values of ARGs on microarrays, verifying the results and accuracy of the detection method in this study. The metagenomic sequencing results obtained relative quantification results of sul1, tetA, qepA, macB, and vanR genes, as shown in Fig. 9B. The silver staining intensity values on the microarrays are directly proportional to the abundance of ARGs, showing that the abundance of ARGs in the soil presents a trend of macB > vanR > tetA > sul1 > qepA. Metagenomic detection results show that the abundance of ARGs in the soil presents a trend of macB > vanR > qepA > sul1 > tetA. The trends of ARGs abundance detected by the two methods are consistent, indicating that the microarray method has high accuracy and can be used for the detection of ARGs.

The silver staining amplification effect has been widely used in the detection of pathogenic bacteria, effectively improving the sensitivity of detection and playing an important role in trace detection of nucleic acids and proteins.^62,63 Ma et al. established a visual detection method for trace lead ions using aptamers and silver staining, and successfully applied it to the qualitative and semi-quantitative determination of Pb²⁺ in soil samples by adding S²⁻ to produce PbS.⁶⁴ Some studies use microfluidic paper-based analytical devices (µPAD) to colorimetrically detect aflatoxin M1 (Afl M1) in milk samples through AuNPs probes and 21-mer aptamers, with detection limits of 3 pM and 10 nM in standard buffer and spiked milk samples, respectively.⁶⁵

This study established a microarray detection method for simultaneously detecting 5 types of ARG through the joint amplification technology of dual-chip probes and silver staining signals. We prepared amino-microarray chips, fixed the designed ARGs-specific dual-chip probes on the microarray, and paired the dual-chip probes with the forward and reverse chains of the ARGs. Then, the AuNPs probes were connected to amplify the silver staining signals. This method amplifies signals by connecting the dual-chip probes to the forward and reverse chains of the ARGs, and the signals are then amplified a second time through silver staining, enabling the detection of the target gene at 0.411 µg mL⁻¹. In addition, the specificity and sensitivity of the microarray chip were also tested. The target genes were detectable within the concentration range of 0.411 µg µL⁻¹ to 55.9 µg µL⁻¹, and the results of silver staining were clearly visible to the naked eye.

Based on numerous studies on silver staining signal amplification, techniques such as strand displacement amplification and enzyme-assisted signal amplification are used to enhance the sensitivity of AuNPs colorimetric detection.^66,67 However, fewer studies design dual probes to connect target sequences for additional signal methods. By combining dual probes with silver staining amplified signals, a new strategy is introduced for rapid visual detection.

Conclusion

This study establishes a visual microarray signal amplification technique that enables high-throughput detection of ARG (sul1, tetA, qepA, macB, and vanR) with a detection range from 0.411 µg mL⁻¹ to 55.9 µg mL⁻¹. By incorporating a centrifuge-free concentration device and dual probe amplification technology, this method significantly reduces the detection time while maintaining high sensitivity and specificity. Previous studies on rapid detection were unable to guarantee detection at low concentrations, and the speed of detection was often compromised when attempting to detect at these lower concentrations. The signal amplification strategy, including the use of silver staining, is ensured to detect ARGs even at low concentrations by the centrifuge-free concentration device and dual probe amplification technology. Experimental results show that the detection method is capable of identifying target genes at concentrations as low as 0.411 µg mL⁻¹, which is a notable improvement over traditional methods. This approach represents a promising solution for on-site environmental monitoring and may serve as a prototype for future detection systems. Moving forward, further optimization of the enrichment and detection processes will be undertaken to enable real-time, high-sensitivity ARG detection in diverse environmental settings. Especially the research on smartphone-based and AI-based detection methods, which makes the detection methods more suitable for on-site applications.

Author contributions

Shenglong Ma: formal analysis, investigation, data curation, writing – original draft. Baowei Zhao: conceptualization, supervision, writing—review & editing. Rui Zhang: formal analysis, methodology, supervision. Hui Liu: conceptualization, supervision. Yunxia Li: data curation, formal analysis, software.

Conflicts of 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.

Data availability

The data presented in this study are available in the article and the supplementary information (SI). Data will be made available on request. Supplementary information is available. See DOI: https://doi.org/10.1039/d5ay01830g.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (22366025, 22166022), the Science and Technology Program of Gansu Province, China (25YFFA096). We sincerely thank the Institute of Sensing Technology, Gansu Academy of Science for providing the basic biological experimental and testing conditions.

References

1. J. Bengtsson-Palme and D. G. Larsson, Nat. Rev. Microbiol., 2015, 13, 396.
2. B. Li, Y. Yang, L. Ma, F. Ju, F. Guo, J. M. Tiedje and T. Zhang, ISME J., 2015, 9, 2490–2502.
3. T. P. Van Boeckel, C. Brower, M. Gilbert, B. T. Grenfell, S. A. Levin, T. P. Robinson, A. Teillant and R. Laxminarayan, Proc. Natl. Acad. Sci. U. S. A., 2015, 112, 5649–5654.
4. M.-C. Danner, A. Robertson, V. Behrends and J. Reiss, Sci. Total Environ., 2019, 664, 793–804.
5. W. Li, B. Wang, T. Wang, J. Li, J. Qi, J. Luo, T. Zhang, X. Xu and L. a. Hou, Environ. Pollut., 2025, 371, 125920.
6. C. J. L. Murray, K. S. Ikuta, F. Sharara, L. Swetschinski and G. Robles Aguilar, Lancet, 2022, 399, 629–655.
7. S. Li, C. Zhang, F. Li, T. Hua, Q. Zhou and S.-H. Ho, J. Hazard. Mater., 2021, 411, 125148.
8. F. Li, Z. Mai, C. Qiu, L. Long, A. Hu and S. Huang, Mar. Environ. Res., 2023, 188, 105978.
9. T. Zhang, L. Fan and Y.-N. Zhang, Aquat. Toxicol., 2025, 284, 107392.
10. J. Bengtsson-Palme, F. Boulund, J. Fick, E. Kristiansson and D. G. Larsson, Front. Microbiol., 2014, 5, 648.
11. N. Czekalski, E. Gascón Díez and H. Bürgmann, ISME J., 2014, 8, 1381–1390.
12. Y. Yang, C. Xu, X. Cao, H. Lin and J. Wang, Ecotoxicology, 2017, 26, 831–840.
13. Y. Yang, W. Song, H. Lin, W. Wang, L. Du and W. Xing, Environ. Int., 2018, 116, 60–73.
14. H. Gao, F. Zhao, R. Li, S. Jin, H. Zhang, K. Zhang, S. Li, Q. Shu and G. Na, J. Environ. Chem. Eng., 2022, 10, 108297.
15. P. R. Desjardins and D. S. Conklin, Curr. Protoc. Mol. Biol., 2011, 3, 3j.
16. V. L. Singer, L. J. Jones, S. T. Yue and R. P. Haugland, Anal. Biochem., 1997, 249, 228–238.
17. N. Aarskog and C. Vedeler, Hum. Genet., 2000, 107, 494–498.
18. W. Zhou, X. Gao, D. Liu and X. Chen, Chem. Rev., 2015, 115, 10575–10636.
19. D. Zheng, G. Yin, M. Liu, L. Hou, Y. Yang, T. P. Van Boeckel, Y. Zheng and Y. Li, Sci. Adv., 2022, 8, eabq8015.
20. I. M. Willms, J. Yuan, C. Penone, K. Goldmann, J. Vogt, T. Wubet, I. Schöning, M. Schrumpf, F. Buscot and H. Nacke, Genes , 2020, 11.
21. Z. Xu, S. Hu, D. Zhao, J. Xiong, C. Li, Y. Ma, S. Li, B. Huang and X. Pan, J. Environ. Manage., 2024, 358, 120827.
22. X. Mao, X. Yin, Y. Yang, Y. Che, X. Xu, Y. Deng, L. Li and T. Zhang, Crit. Rev. Environ. Sci. Technol., 2024, 54, 1633–1650.
23. X. Xia, Q. Gu, L. Chen, J. Zhang, W. Guo, Z. Liu, A. Li, X. Jiang, M. Deng, J. Zeng, X. Lin, F. Peng, W. Chen, Y. Ye and Q. Wu, J. Environ. Chem. Eng., 2025, 13, 115381.
24. V. A. C. de Abreu, J. Perdigão and S. Almeida, Front. Genet., 2020, 11, 575592.
25. C. Zhu, L. Wu, D. Ning, R. Tian and S. Gao, Nat. Commun., 2025, 16, 4006.
26. M. Zhuang, Y. Achmon, Y. Cao, X. Liang, L. Chen, H. Wang, B. A. Siame and K. Y. Leung, Environ. Pollut., 2021, 285, 117402.
27. J. Lu, Y. Zhang, J. Wu, J. Wang, C. Zhang and Y. Lin, Environ. Pollut., 2019, 252, 450–460.
28. T. B.-A. Nguyen, M. Bonkowski, K. Dumack, Q.-L. Chen, J.-Z. He and H.-W. Hu, ISME J., 2023, 17, 2182–2189.
29. C. Wang, L. Xue, Y. Dong, Y. Wei and R. Jiao, Forests, 2018, 9, 532.
30. S. Yokota, Methods Mol. Biol., 2016, 1474, 61–71.
31. C. Daruich De Souza, B. Ribeiro Nogueira and M. E. C. M. Rostelato, J. Alloys Compd., 2019, 798, 714–740.
32. V. Afanassiev, V. Hanemann and S. Wölfl, Nucleic Acids Res., 2000, 28, E66.
33. Z. L. Zou, S. Q. Wang and Z. Q. Wang, Chin. J. Biotechnol., 2001, 17, 498–502.
34. O. I. Guliy and L. A. Dykman, Biosens. Bioelectron. X, 2024, 17, 100457.
35. G. Suneetha, D. Ayodhya and P. Sunitha Manjari, Results Chem., 2023, 5, 100688.
36. R. Eivazzadeh-Keihan, Z. Saadatidizaji, M. Mahdavi, A. Maleki, M. Irani and I. Zare, Talanta, 2024, 275, 126099.
37. Q. Wang, H. Cui, C. Li, X. Song, Q. Lv and Z. Li, Sens. Acutators Rep., 2020, 2, 100021.
38. N. T. Thanh and Z. Rosenzweig, Anal. Chem., 2002, 74, 1624–1628.
39. A. Loiseau, L. Zhang, D. Hu, M. Salmain, Y. Mazouzi, R. Flack, B. Liedberg and S. Boujday, ACS Appl. Mater. Interfaces, 2019, 11, 46462–46471.
40. E. Ferrari, Biosensors, 2023, 13, 411.
41. A. Fernández-Lodeiro, J. Djafari, J. Fernández-Lodeiro, M. P. Duarte, E. Muchagato Mauricio, J. L. Capelo-Martínez and C. Lodeiro, Nanomaterials, 2021, 11, 1338.
42. A. D'Agostino, A. M. Giovannozzi, L. Mandrile, A. Sacco, A. M. Rossi and A. Taglietti, Talanta, 2020, 216, 120936.
43. S. Mussa Farkhani, P. Dehghankelishadi, A. Refaat, D. Veerasikku Gopal, A. Cifuentes-Rius and N. H. Voelcker, Prog. Mater. Sci., 2024, 142, 101229.
44. Z. Hua, T. Yu, D. Liu and Y. Xianyu, Biosens. Bioelectron., 2021, 179, 113076.
45. N. Elahi, M. Kamali and M. H. Baghersad, Talanta, 2018, 184, 537–556.
46. A. Z. Bunkas, K. Kalnins, V. V. Kim, A. N. K. Reddy, A. Bundulis, A. Atvars, A. Sarakovskis, A. Ubelis and R. A. Ganeev, Appl. Phys. B, 2024, 130, 21.
47. F. Millozzi, P. Milán-Rois, A. Sett, G. Delli Carpini, M. De Bardi, M. Gisbert-Garzarán, M. Sandonà, C. Rodríguez-Díaz, M. Martínez-Mingo, I. Pardo, F. Esposito, M. T. Viscomi, M. Bouché, O. Parolini, V. Saccone, J.-J. Toulmé, Á. Somoza and D. Palacios, Nat. Commun., 2025, 16, 577.
48. K. A. Willets and R. P. Van Duyne, Annu. Rev. Phys. Chem., 2007, 58, 267–297.
49. Y. Q. He, S. P. Liu, L. Kong and Z. F. Liu, Spectrochim. Acta, Part A, 2005, 61, 2861–2866.
50. Y. Dong, L. Ding, X. Jin and N. Zhu, Microchim. Acta, 2017, 184, 3357–3362.
51. Q. Wang, Y. Wang, M. Guan, S. Zhu, X. Yan, Y. Lei, X. Shen, L. Luo and H. He, Microchem. J., 2020, 159, 105429.
52. K. W. Shinato, F. Huang and Y. Jin, Corros. Rev., 2020, 38, 423–432.
53. H. Li, Z. Wu, L. Qiu, J. Liu, C. Wang, G. Shen and R. Yu, Biosens. Bioelectron., 2013, 50, 180–185.
54. K. S. Park, M. I. Kim, D. Y. Cho and H. G. Park, Small, 2011, 7, 1521–1525.
55. D. Vilela, M. C. González and A. Escarpa, Anal. Chim. Acta, 2012, 751, 24–43.
56. X. Su and R. Kanjanawarut, ACS Nano, 2009, 3, 2751–2759.
57. H. Wang, H. Rao, M. Luo, X. Xue, Z. Xue and X. Lu, Coord. Chem. Rev., 2019, 398, 113003.
58. C. Marquette, P. N. N. Marche, M.-B. Villiers, C. Brakha and A. Buhot, in Electropolymerization, ed, E. Schab-Balcerzak, IntechOpen, Rijeka, 2011,  DOI:10.5772/30309.
59. M. A. H. Capelle, R. Gurny and T. Arvinte, Eur. J. Pharm. Biopharm., 2007, 65, 131–148.
60. A. Prasad, S. M. A. Hasan, S. Grouchy and M. R. Gartia, Analyst, 2019, 144, 197–205.
61. J. Du, C. Cao, Z. Xue, W. Wang, X. Lu, Y. Wei, J. Huang, L. Zhao, L. Wang, F. Xu, C. Yao, T. Wen and M. You, Anal. Chem., 2025, 97, 24196–24208.
62. T. A. Taton, C. A. Mirkin and R. L. Letsinger, Science, 2000, 289, 1757–1760.
63. N. R. Jana and J. Y. Ying, Adv. Mater., 2008, 20, 430–434.
64. L.-H. Ma, H.-B. Wang, B.-Y. Fang, F. Tan, Y.-C. Cao and Y.-D. Zhao, Colloids Surf., B, 2018, 162, 415–419.
65. A. Kasoju, D. Shahdeo, A. A. Khan, N. S. Shrikrishna, S. Mahari, A. M. Alanazi, M. A. Bhat, J. Giri and S. Gandhi, Sci. Rep., 2020, 10, 4627.
66. Y. V. Gerasimova and D. M. Kolpashchikov, Chem. Soc. Rev., 2014, 43, 6405–6438.
67. B. Yurke, A. J. Turberfield, A. P. Mills Jr, F. C. Simmel and J. L. Neumann, Nature, 2000, 406, 605–608.

---