The chemicals diazinon, fenthion, fenitrothion, and malathion (HPLC purification grade) were purchased from Sigma–Aldrich Inc. (St. Louis, MO, USA). All chemical insecticides were dissolved in acetone and then diluted in 1x SELEX buffer (20 mM HEPES, 1 M NaCl, 10 mM MgCl${}_2$, 5 mM KCl, pH 7.5) to formulate the desired working concentration and were stored at 4 ${}^{\circ }$C.

The size of ssDNAs in the library used was 72 bp, which included the following: N-30 region generated by randomly mixing the A-C-T-G bases and primer region (forward and reverse): (1) Library: 5${}^{\prime }$-GGA GGC TCT CGG GAC GAC GAC-(N30)-GTC GTC CCG ATG CTG CAA TCG TAA-3${}^{\prime }$; (2) Biotinylated column immobilizing capture strand: 5${}^{\prime }$-GTC GTC CCG AGA GCC ATA - BioTEG-3${}^{\prime }$; (3) Forward primer: 5${}^{\prime }$-GGA GGC TCT CGG GAC GAC GAC-3${}^{\prime }$; (4) Reverse primer: 5${}^{\prime }$-TTA CGA TTG CAG CAT CGG GAC G-3${}^{\prime }$; and (5) Biotinylated reverse primer: 5${}^{\prime }$-Biotin-TTA CGA TTG CAG CAT CGG GAC G-3${}^{\prime }$. All oligonucleotides were dissolved in nuclease-free water and stored at –20 ${}^{\circ }$C.

A 500 pmol ssDNA library was mixed with 2500 pmol capture strand (1:5 ratio) in 250 $\mu$L of 1x SELEX buffer. The mixture was incubated in a hot water bath (95 ${}^{\circ }$C) for denaturation. After 5 minutes, the mixture was removed and then allowed to cool at room temperature for 10 minutes. The oligonucleotide mixture was transferred to a micro biospin chromatography column (Cat. No. 732-6204, Bio–Rad, Hercules, CA, USA) containing 200 $\mu$L of streptavidin agarose resin (1–3 mg biotinylated BSA/mL resin, Thermo Scientific, Waltham, MA, USA). The oligonucleotide mixture was slowly passed through the column three times, and the eluent was collected. Then, the buffer was washed fifty times with 250 $\mu$L of 1x SELEX buffer each time. The eluent from the first wash (W1) and the final wash (W50) was collected and used to monitor the efficiency of washing. We then added three aliquots of 100 $\mu$M diazinon to the column, and each 250 $\mu$L aliquot was collected as the eluent. Small-scale PCR was run with 1 cycle of 95 ${}^{\circ }$C for 2 min, N cycles of 92 ${}^{\circ }$C for 15 s, 59 ${}^{\circ }$C for 30 s, and 72 ${}^{\circ }$C for 45 s and 1 cycle of 72 ${}^{\circ }$C for 2 min. Then, the PCR products were separated by 3% agarose gel electrophoresis (100 V, 20 min) to evaluate the efficiency and specificity of the selection procedure.

Three aliquots of eluent after diazinon addition were concentrated by a 3 KDa filter (EMD Millipore Amicon Ultra0.5 Centrifugal Filter Units, Millipore, cat. No. 732-6204) and were used as the templates for large-scale PCR. The reverse primer was replaced by a biotinylated reverse primer in this step. Four PCR tubes containing 30 $\mu$L samples were run similarly to a small-scale PCR program with 8, 10, 12, and 14 cycles. All the PCR products were loaded in the gel, and the optimal number of PCR cycles to amplify the template was estimated. Then, PCR was run with the rest of the master mix. Using a 10 KDa MWCO filter, the concentrated PCR products were passed through a mini gravity column containing 0.2 mL of streptavidin agarose resin. The eluent was collected after incubating with 400 $\mu$L of 0.2 M NaOH, and 0.2 M HCl was used to adjust pH 7.5. After adding 750 $\mu$L 2x SELEX buffer, the eluent was concentrated by a 3 KDa MWCO filter. This concentrated product was used as a new library for the next round of selection.

Purified DNA from the last round of library selection was used to clone the candidate aptamers. As mentioned earlier, PCR amplification of selected DNA was carried out, and cloning was performed using the T-Blunt™ PCR Cloning kit (SolGent, Daejeon, Korea) following the manufacturer’s instructions. A total of 96 individual colonies were randomly picked for plasmid extraction. Plasmid DNAs were extracted by using the HiGene™ Plasmid Mini Kit (Ver 2.0; BIOFACT, Daejeon, Korea) following the manufacturer’s instructions. All plasmids were sent for sequencing service by COSMO Tech Company (Seoul, Korea). The candidate aptamer sequences were analyzed and classified into groups using the Clustal Omega program [27] and aligned manually to remove spaces. Next, SELEX was performed with a few representative clones whose copy numbers were predominant in sequence data to confirm diazinon binding.

Sequences were analyzed to make a 2D structure using the Mfold program [28]. The aptamer sequences that form a stem–loop structure with eight base-pair stems were selected for the aptamer assay with fluorescein (FAM) conjugated to the 5${}^{\prime }$-end of the sensor. A 13-bp quencher sequence modified with dabcyl dye at the 3${}^{\prime }$-end (5${}^{\prime }$-GTCGTCCCGAGAG -Dab-3${}^{\prime }$) was used to quench the fluorescence signal.

The quenching efficiency was determined by optimization of the ratio of the sensor to the quencher. The 2x concentration of FAM-sensor 100 nM was mixed with a 13-bp quencher by twofold serial dilution of range from 0 nM to 500 nM with an equal volume (60 $\mu$L). The sensor and quencher mixture was heated to 96 ${}^{\circ }$C for 5 min and gradually cooled to room temperature for 30 min. One hundred microliters of each mixture was transferred into a 96-well black polystyrene assay plate (Cat. No. 3340, Corning, NY, USA). The fluorescence was measured by a 485 nm excitation wavelength and emission collection with a 515 nm to 535 nm cutoff on a SpectraMax GeminiTM XPS/EM Microplate Reader (San Jose, California 95134, United States). F/F0 was used to generate the binding curve, where F is the FL intensity in the presence of the quencher and F0 is the FL intensity without the quencher. Each binding curve was plotted through nonlinear least-squares progression using GraphPad Prism 5.0 software (GraphPad Software, 2365 Northside Dr., Suite 560, San Diego, CA 92108). All measurements were performed in nine replicates. The quenching efficiency was defined as q% = (1 – F/F0) %.

The K${}_d$ values of the sensor against the target were calculated. The final concentrations of both the sensor and quencher were set at a ratio that was determined by a previous quencher testing experiment, which has more than 80% quenching efficiency and is used at a sensor: quencher ratio of 1:5. To make a standard curve, the concentration of diazinon was made from 0 mM to 1 mM by twofold serial dilution. Briefly, 30 $\mu$L of each 4x sensor and a quencher were mixed and heated at 96 ${}^{\circ }$C for 5 min and cooled gradually to room temperature for 30 min. Then, 60 $\mu$L of various concentrations of 2x diazinon (0, 0.02, 0.04, 0.08, 0.16, 0.32, 0.64, 1.28, 0.78125, 1.5625, 3.125, 6.25, 12.5, 25, and 50 $\mu$M) was added and incubated at room temperature for 40 min. A total of 100 $\mu$L of each solution was transferred onto a 96-well microwell plate, and the FL intensity was measured. All measurements were performed in nine replicates.

The LOD for each sensor was calculated based on the formula CL = K*Sd/S [29], where Sd is the standard deviation of blank sample; K refers to a signal-to-noise ratio (S/N) (which was used in this study with 3.3); S is the slope of detecting linear between the fluorescence intensity enhancement versus low concentrations [29] of diazinon. The specificity of the sensors toward diazinon was estimated via cross-reactivity testing. Three organophosphate pesticides, fenthion, fenitrothion, and malathion, were used to compare the cross-reactivity of the selected aptamers.

The optimal conditions for detecting diazinon from vegetable and fruit tissue extracts were evaluated using different pH values, temperatures, and incubation times. The effect of pH was studied using all samples (sensor/quencher and diazinon) by preparing in 1x SELEX buffer with different pH values (5.0, 6.0, 7.0, 7.5, and 8.0) incubating at room temperature for 40 min before measurement. The effect of incubation time on all samples was studied by preparing samples in 1x SELEX buffer at pH 7.5 and measuring the signal at different time points from 0 min to 60 min. Finally, the effect of temperature was estimated by performing the reaction in 1x SELEX buffer (pH 7.5 and 40 min) at various temperatures, such as 20 ${}^{\circ }$C, 30 ${}^{\circ }$C, 35 ${}^{\circ }$C, 40 ${}^{\circ }$C, 45 ${}^{\circ }$C, 50 ${}^{\circ }$C, and 60 ${}^{\circ }$C.

To prepare plant tissue extracts, tomato, Chinese cabbage, and apples were selected as examples to develop a biosensor assay. Chinese cabbage (cultivar: Wa Wa Wa Sai), tomato (cultivar: Chal), and apple (cultivar: Fuji) were purchased from TopMart (Jinju, Gyeongsangnam-do, Korea). Fifty grams of sample was first cut into small pieces and homogenized with 50 mL of 20 mM Tris-HCl buffer (pH 7.5) by using a food blender and then centrifuged for 10 minutes at 4 ${}^{\circ }$C and 14000 rpm. The supernatant was collected and filtered using a Miracloth filter (Cat. No. 475855-1R, Sigma–Aldrich, St. Louis, MO, USA) and then used as an original plant extract. To obtain the proper dilution series of plant extract samples, we prepared five different plant extract solutions by 10 times serial dilution (0, 10X, 100X, 1000C, and 10000X), and each sample was spiked with 10 $\mu$M of diazinon. Additionally, we used different concentrations of diazinon to obtain optimal spiked concentrations and found that a concentration of 10 $\mu$M, which was within the linear range of detection, was most suitable.

The structure-switching SELEX process was employed [23] to identify ssDNA aptamers specific to diazinon. This method was based on the principle of structure switching and elution upon target binding [15, 18]. The random ssDNA library (of $>$10${}^{15}$ DNA molecules) was hybridized with a short capture strand with a biotinylated complementary sequence (Fig. 1A, Ref. [28]). The candidate aptamer oligonucleotides were immobilized on agarose beads containing streptavidin through a biotin-functionalized capture strand. First, diazinon was introduced as a specific target at a concentration of 100 $\mu$M from the 1st round to the 10th round of the SELEX process. Later, its concentration was reduced to 10 $\mu$M from the 11th round to the 15th round. Next, three rounds of negative selection were performed using fenthion, fenitrothion, and malathion (nonspecific target molecules). To generate the enriched library, the ssDNA oligonucleotides were eluted with diazinon, collected, and amplified via polymerase chain reaction (PCR). With an increase in the round of the SELEX process, the proportion of diazinon binding sequences in the regenerated ssDNA pool was increased. Subsequently, lower affinity bound or free nucleotide sequences were washed off. After washing 50 times with 1x SELEX buffer, unbound sequences in the ssDNA pool were removed from the column (Fig. 2C).

Fig. 1.

Design of oligonucleotide probes and monitoring of the SELEX process. (A) An overview of a typical SELEX cycle to identify and isolate specific ssDNA aptameric sequences. Random ssDNA libraries flanked by fixed primer annealing sites for PCR amplification are incubated with the target molecule, and aptamer-target complexes are separated from unbound ssDNA. PCR amplification with the selected ssDNA followed by cloning, sequencing, and strand separation for the next SELEX round. The SELEX cycle was repeated several times to enrich the high-affinity binding ssDNA. (B) The random ssDNA library was hybridized with a short capture strand via biotinylated complementary sequences, wherein ‘N’ means any nucleotide generated by random mixing of A-T-G-C bases. (C) Electrophoresis of eluent obtained after the 10th SELEX round (10 $\mu$M diazinon) and the 16th SELEX round (1 $\mu$M fenthion, 1 $\mu$M fenitrothion, 1 $\mu$M malathion and 10 $\mu$M diazinon); where NC, Negative control; Af. C, the mixture obtained after addition to the column; W1, the first wash; W50, the 50th wash; Ex, the eluent after incubation with target diazinon; N-f, negative selection of fenthion; N-fn, negative selection of fenitrothion; N-m, negative selection of malathion.

Fig. 2.

Confirmation and prediction of the secondary structures of candidate aptamers. (A) Stepwise illustration of aptamer selection—individual clones (plasmids), sequence amplification, strand separation, and confirmation of size and band intensity via electrophoresis, where NC, Negative control; Af. C, the mixture obtained after adding a column; W1, the first wash; W50, the 50th wash; Ex, the eluent after incubation with the target diazinon. (B) The secondary structure of three candidate aptamers was predicted using the Mfold program [28].

The selected PCR products were cloned, and numerous individual colonies were obtained that carry plasmids with a particular PCR fragment representing each candidate aptameric ssDNA. A total of 96 colonies were randomly picked for sequencing. The sequences were subjected to bioinformatics analysis; sequence alignment and similarity were measured. Each sequence contained an 18 bp forward primer and 22 bp reverse primer at the 5${}^{\prime }$ end and 3${}^{\prime }$-end, respectively, while the middle part consisted of a random variable region of 30 nucleotides (Fig. 1B). Of the 96 colonies sequenced, 64 yielded high sequence similarity and high abundance, with 18 unique sequences (Table 1). The remaining 32 clones showed high sequence divergence and low abundance; some clones failed to yield good sequence data (Supplementary Table 1). The sequence alignment and frequency helped select and classify aptamers into three classes: DIAZ-01, DIAZ-02, and DIAZ-03 (Table 1). Next, to confirm the binding affinity of the aptamers for diazinon, DIAZ-01, DIAZ-02, and DIAZ-03 were selected as representatives of each class (Fig. 2A). The results indicated that all three aptamers displayed a high binding affinity to diazinon.

The secondary structure of the newly discovered candidate aptameric sequences was predicted to examine the suitability and stability of the novel biosensor design (Fig. 2B). The sequences were analyzed by Mfold software [28]. The free energy of the sequences was compared, and all aptamers were found to have a short stem of 11 to 13 bp that could be used to form base pairing with a 13-bp dabcyl quencher.

The unique properties of ssDNA aptamers provide unparalleled opportunities to develop novel biosensors that can specifically bind to target molecules and produce a detectable signal. To obtain the fluorescence signal, the candidate ssDNA aptamer was modified with fluorescein amidite (FAM) at the 5${}^{\prime }$-end. Fluorescein is one of the most commonly used fluorophores for labeling biomolecules (DNA, RNA, and proteins) [15, 16, 17, 30]. It possesses a relatively high quantum yield, strong emission peak, specific excitation wavelength, and acceptable water solubility. Moreover, the fluorophore’s small size has minimal impact on ssDNA properties. However, dabcyl has been widely used as a universal quencher for many fluorophores, as it has strong potential to absorb emitted photons [30]. Hence, 6-FAM was successfully quenched, and it lacked any residual fluorescence. Additionally, 6-FAM or dabcyl molecules are relatively easy to introduce at any position in ssDNA or dsDNA sequences, including at the 5${}^{\prime }$- or 3${}^{\prime }$-end [30]. A 13-mer quencher carrying dabcyl dye was used to quench (block) the fluorescence signal. When the sensors were allowed to hybridize with a quencher, the dabcyl could successfully come in proximity to the FAM to form an F-Q complex, resulting in quenching of the FAM signal. Subsequently, the diazinon sensors undergo structural switching due to target-induced conformational change. This change in conformation leads to a unique 3D folding that causes dissociation of the quencher strand [15, 30], the quenching effect is eliminated, and a measurable fluorescence signal is regenerated (Fig. 3A). Achieving higher quenching efficiency ($>$80%) is essential for evaluating successful and effective aptameric sensors for detecting the target. To determine the optimal concentrations of the DIAZ-F sensors and quencher for the detection of the target molecule diazinon, we performed quencher testing in the absence of diazinon. We observed that the ssDNA aptameric sensors showed varied quencher efficiency; DIAZ-01-F, DIAZ-02-F, and DIAZ-03-F reached 89%, 86%, and 92% quenching when the concentration of quencher was applied at 125 nM, 250 nM, and 50 nM, respectively (Fig. 3B).

Fig. 3.

Detection strategy and characterization of fluorescent aptamer sensors against diazinon. (A) Graphic illustration of detection using a strand-displacement-based fluorescence assay and quenching with a dabcyl-modified quencher. Sensors are shown in blue line with 6-FAM (green circle) at the 5${}^{\prime }$-end, quencher-strand shown in gray with dabcyl (black circle), and diazinon molecule (red star). (B) Evaluation of the quenching efficiency of selected ssDNA aptamers. (C) Evaluation of the binding potential of 6-FAM-modified sensors to diazinon.

Next, the binding affinity of aptameric sensors with diazinon was determined by measuring the gain of fluorescence intensity (FL) by applying various concentrations of the target molecule. DIAZ-01-F, DIAZ-02-F, and DIAZ-03-F were incubated with quencher-strand in the optimal ratio that was determined by prior quenching analysis. A total of 8 different concentrations of diazinon (0 $\mu$M to 50 $\mu$M) were used, and a change in the FL intensity was observed. The FL intensity was found to increase due to diazinon concentration (Fig. 3C). The binding curve was further analyzed to measure the K${}_d$ value, and the binding analysis revealed that DIAZ-02-F had the lowest K${}_d$ value of 4.571 $\pm$ 0.714 $\mu$M among all aptamers tested. The K${}_d$ values were estimated following the method described by Hu and Easley [31]. Furthermore, by exploring the linearity of detection of the target molecule using each sensor, we calculated the limit of detection (LOD) value [29]. Linear detection was observed based on the change in FL intensity in the presence of a wide range of diazinon concentrations. The LOD values for each aptameric sensor were determined as follows: DIAZ-01-F, DIAZ-02-F, and DIAZ-03-F, 226 nM, 148 nM, and 341 nM, respectively. The analysis showed that the DIAZ-02-F aptamer had the highest affinity for diazinon among all aptamers used. As per US EPA guidelines, the maximum residue level for diazinon is 0.7 mg/Kg [32], which corresponds to 2.3076 $\mu$M for Brassica sp; whereas our assay could detect the pesticide with LOD of 44.89 $\mu$g/Kg (148 nM). Hence, the biosensor has high potential to recognize pesticide contamination of foods.

To determine the specificity of the novel ssDNA aptameric sensors developed in this study, we tested the cross-reactivity of the DIAZ-F biosensors against three other organophosphate insecticides, fenthion, fenitrothion, and malathion (Fig. 4). The selected insecticides possess structural and physicochemical properties similar to those of diazinon’s specific target. Detailed analysis of specificity showed that both DIAZ-01-F and DIAZ-02-F had a relatively high level of cross-reactivity to fenthion and fenitrothion. The organophosphorus insecticides shared an aromatic ring structure (Fig. 4B–E). Notably, all three aptamers successfully discriminated between malathion and diazinon’s specific target. Moreover, malathion did not bind to any aptasensors. Among all aptamers, DIAZ-03F was found to be the most specific, as it did not bind to fenitrothion or malathion and showed weaker binding to fenthion insecticide. Altogether, the ssDNA aptamers exhibited high affinity and specific binding to the target molecule diazinon. However, DIAZ-1-F and DIAZ-2-F showed weaker binding to the nonspecific targets fenthion and fenitrothion due to the presence of core ring structures and phosphorothioate bonds. Overall, DIAZ-03-F performed best in the specific recognition of the target molecule.

Fig. 4.

Measurement of the cross-reactivity of aptamer sensors. Cross-reactivity analysis using three other insecticides: (A) diazinon; (B) fenthion; (C) fenitrothion; (D) malathion; and (E) chemical structure of insecticides.

Vegetables and fruits form a significant ingredient of food and nutrition and are essential for a balanced human diet. This study assessed the applicability of newly designed aptameric sensors to detect insecticide contamination in food. The effect of pH, temperature, and incubation time was investigated as food samples present complexities. The DIAZ-02-F sensor and quencher mixture (1:5 ratios, i.e., 50 nM:250 nM) was incubated with 10 $\mu$M diazinon under different conditions and a control experiment sensor-quencher mixture without diazinon. We found that the DIAZ-02-F sensor worked well at pH values of 7.0, 7.5, and 8.0 (Fig. 5A). The FL of the sensor was similar to pH 7.5 and 8.0. Since the SELEX process was performed at pH 7.5, the food testing assay was conducted at pH 7.5. However, we noticed that under acidic conditions (at pH 5.0 or 6.0), the DIAZ-02-F sensor and quencher exhibited weaker quenching efficiency, and the aptamer did not bind to diazinon effectively. Hence, the observed FL was similar to that in the control experiments (Fig. 5A). Next, we probed the effect of incubation time on assay performance, for which we used 0 to 60 min; the sensor performed best at 40 min. Subsequently, we examined the effect of temperature on the FL and found that the highest signal was achieved at 30 ${}^{\circ }$C.

Fig. 5.

Analysis of diazinon concentration using the aptameric sensor in vegetable and fruit tissue extracts. (A) Determination of optimal conditions for diazinon detection using the DIAZ-02-F sensor. (B) Analysis of diazinon concentration using the DIAZ-02-F sensor in tissue extracts of vegetables/fruits—tomato, Chinese cabbage, and apple.

Altogether, the optimization experiments provided the best protocol to conduct the experiments using the following conditions: 1x SELEX buffer (for sensor: quencher mixture or food extract), pH 7.5, incubation at 30 ${}^{\circ }$C for 40 min. We eventually explored the potential application of the DIAZ-02-F sensor in real sample testing. Plant tissue extracts of Chinese cabbage, tomato, and apple were analyzed for diazinon contamination (Fig. 5B). Undiluted plant tissue extracts interfered with the FL signal. Hence, we diluted the extract to improve the assay’s performance. Therefore, we diluted the tissue extracts and found that 1000X dilution produced a higher signal than undiluted extracts, while for tomato extract, we needed to dilute 10000X. In the case of apple extract, the assay worked well even with 100X dilution. Notably, the sensor DIAZ-02-F performed well and could detect diazinon from pure Chinese cabbage extract. Overall, these results provide strong potential for sensing toxic insecticide contamination in food samples such as vegetable and fruit extracts. The sensor could be developed as a point-of-care tool to monitor the entry of toxic agri-waste contaminants into the human food chain and protect the environment, fulfilling several SDGs and ensuring a clean and safe planet for future generations.