Initially, ssDNA library, aptamer sequences, primers, and mycotoxins used in this study were obtained from Sigma-Aldrich (India), except mentioned specifically. All the solutions were prepared with ultra high purity water from Millipore water purification system. The chemicals and solvents used in present study are listed in the Supporting Information 1.

The single stranded DNA (ssDNA) library used in the present study was designed with a randomized region of 40 nucleotides flanked by two known sequences of 20 nucleotides on either side. The two known sequences served as binding regions for forward and reverse primers in PCR amplification. The nucleotide sequences of the library and the primers used in the study are listed in Supporting Table 1.

In order to generate ssDNA, asymmetric PCR amplification (Mondal et al., 2015) of aptamer library was carried out with each reaction consisting of 4 μl of 1X PCR buffer, 3.2 μl of 1.6 mM MgCl₂, 2 μl of 0.2 mM dNTPs, 0.4 μl of 2U/μl of Taq DNA polymerase, 1 μl of template library, 27.4 μl of distilled water and different ratios of forward to reverse primers (1.1:0.9, 1.2:0.8, 1.3:0.7, 1.4:0.6, 1.5:0.5, 1.6:0.4, 1.7:0.3, 1.8:0.2, 1.9:0.1, and 2.0:0). The PCR was performed in thermocycler (Bio-Rad, India) with following amplification conditions: initial denaturation at 94°C for 5 min and 30 cycles of 94°C for 45 s, 56°C for 45 s, 72°C for 45 s and final extension at 72°C for 8 min. PCR amplicons were separated by 2% agarose gel electrophoresis and desired product size was purified using Qiagen MiniElute gel extraction kit following manufacturer’s protocol.

Aptamers exhibiting affinity to AFB1 were selected using modified immunoaffinity based SELEX (Figure 1). Detailed methodology is provided in the Supporting Informations 2 and 3.

Initially, pre-SELEX was performed with 2 nmol denatured aptamer library dissolved in 500 μl of 1X binding buffer (1X BB; 5 mM MgCl₂ in Dulbecco’s phosphate buffered saline, pH 7.4 ± 0.2) and mixture was incubated for 30 min in empty preconditioned IA column to partition aptamers having affinity to column components. The collected flowthrough was pooled and concentrated by PCR clean up kit following manufacturer’s protocol (Supporting Information 3A). The eluted DNA was used as template library for selection of aptamers against AFB1 in further rounds of SELEX. In each round of SELEX, 100 ng/μl of pre-prepared, denatured ssDNA pool was incubated with AFB1 (200 ng/ml) bound to IA column and washed with 500 μl of 1X BB (Supporting Information 3B). Bound aptamer pool was eluted with methanol. DNA from pooled elutes was recovered and dissolved in 50 μl of distilled water and concentration of obtained DNA was measured by Nanodrop-2000 (Thermo Scientific, India). The eluted DNA was amplified by earlier standardized PCR conditions and used in subsequent rounds of SELEX. The concentration and specificity of ssDNA pool toward AFB1 from each round was monitored by Nanodrop-2000 (Thermo Scientific, India) and ELONA (Mondal et al., 2015) using biotin-labeled aptamer pool (Supporting Information 4). To enhance the specificity of the aptamers, counter SELEX was introduced after 3rd, 5th, and 7th rounds. Aflatoxins (AFB2, AFG1, and AFG2) were conjugated to BSA molecule according to Cervino et al. (2008) with certain modifications and commercially procured BSA conjugated Ochratoxin A (OTA-BSA) were employed in the counter SELEX (Supporting Information 3B).

High affinity aptamer pool from 10th round was cloned using TOPO TA cloning kit following manufacturer’s instructions. Positive clones were selected and sequenced utilizing the in-house sequencing facility (ABI 3500 Genetic Analyzer, Applied Biosystems, India) (Supporting Information 5). Obtained sequences were analyzed by using online multiple sequence alignment tool, Multalin¹ (Corpet, 1988). Possible secondary structures and Gibbs free energies (dG) of selected aptamers were predicted using the M-fold software² at 26°C in 150 mM (Na+) and 1 mM (Mg++; SantaLucia, 1998; Zuker, 2003). Putative Quadruplex forming G Rich Sequences (QGRS) of high affinity aptamers using QGRS Mapper software³ were predicted (Kikin et al., 2006).

Binding affinity of aptamer pool and individual aptamers were evaluated by ELONA (Supporting Figure S4; Supporting Information 4). Microtiter plates were coated with BSA conjugated AFB1 (AFB1-BSA) at a concentration of 200 ng/ml (carbonate-bicarbonate buffer, pH 9.6 ± 0.2). After blocking with BSA (3%), biotin labeled ssDNA aptamers (250 ng/μl) were added and incubated for 1 h at 37°C. The plates were washed with 1X PBST prior to incubation with Streptavidin-HRP (1:3000 dilutions) at 37°C for 45 min. After incubation, plates were washed thoroughly and color was developed using chromogenic substrate (OPD/H₂O₂). The reaction was stopped using H₂SO₄ (2N) and absorbance was measured at 490 nm using Infinite M1000 spectrophotometer (TECAN, India). For evaluation of cross reactivity, the mentioned assay was repeated with other BSA conjugated mycotoxins (AFB2, AFG1, AFG2, and OTA) that were prepared according to Cervino et al. (2008).

Similar ELONA was performed for determination of apparent dissociation constants (Martín et al., 2013), where constant amounts of AFB1-BSA conjugate (200 ng/ml) was titrated with increasing concentrations (0–350 nM) of biotinylated aptamers. The resultant absorbance values were fitted in y = B_max^∗x/(K_d+x) via non-linear regression analysis, using Graph Pad Prism 6 software. Additionally, the detection potential of high affinity binders was tested with a various concentrations of AFB1-BSA (10–250 ng/ml) against constant amounts of biotinylated aptamers (Supporting Figure S4; Supporting Information 6).

Purchased corn samples were surface sterilized by rinsing with water followed by rapid shaking in 1% Sodium hypochlorite + 0.1% Tween 20 for 10–15 min and final washing 3–5 times for 5 min in sterile water and were grounded mechanically. For evaluating the purifying ability of selected aptamers, AFB1 ranging from 10 to 250 ng/ml was resuspended in methanol and incubated with 5 gm of powdered corn samples for 3 h. Toxin extraction from spiked samples was carried according to Stroka et al. (2000), with certain modifications (Supporting Information 7). Obtained toxin extracts were air dried in dark to remove remnants of organic solvents and resuspended in 1X BB (500 μl) and allowed to react with biotinylated ssDNA bound onto the streptavidin coated microtiter plates and eluted with methanol: water (80:20) (Figure 7). The recovery of eluted AFB1 was confirmed and quantified by HPLC (JASCO, UK) analysis according to Priyanka et al. (2014) with certain modifications. The chromatographic analysis of AFB1 from naive and spiked samples was carried on RP-C18 column (3 μm, 250 mm × 46 mm) using mobile phase consisting of water: acetonitrile: acetic acid in the ratio of 52:47:1, at a flow rate of 0.8 ml/min and fluorescence detection (emission-365 nm, excitation-455 nm). Purification efficiency of the aptamers was assessed by plotting measured relative peak area (μV) against various toxin concentrations employing linear regression equation (y = mx+c).

All experiments were repeated three times with similar conditions to ensure the stability and reliability of the data. Results were presented as the mean value ± standard deviation (SD). Statistical differences between treatments were analyzed by univariate (ANOVA) and Tukey’s test assuming p value (P < 0.05), (P < 0.01), and (P < 0.001) followed by Dunnett’s test using Graph Pad Prism6.

In the present study, amplification of aptamer pool using asymmetric PCR was performed. PCR parameters such as number of cycles, annealing temperature, primer concentration were optimized to evade mis-amplification of DNA (Supporting Figures S1A–C). Asymmetric PCR preferentially increased the target ssDNA and decreased the primer dimer as well as non-specific amplification. Optimized reactions inferred a sharp band of expected base pair size (agarose gel analysis) at a combination of 56°C annealing temperature with primer ratio of 1.6:0.4 (Forward: Reverse) after 30 cycles (Supporting Figure S2).

Immuno affinity column based SELEX procedure was performed to identify aptamers against AFB1 as described previously (Supporting Informations 2 and 3). Ten rounds of SELEX were performed, alternating with three rounds of counter SELEX using closely related and other mycotoxins after round 3rd, 5th, and 7th to isolate aptamers that could favorably bind to target toxin with high specificity. After each round of selection, bound aptamer pool was recovered using methanol and concentrated DNA pool was amplified by standardized PCR conditions (see “Optimization of PCR conditions”) and used in subsequent rounds of selection.

Prior to selection, binding saturation of toxin to IA column was analyzed by thin layer chromatography (TLC). Absence of toxin in flow through at concentration of 200 ng indicated optimal binding ability of IA column. During selection, pre SELEX was performed to remove aptamer sequences having affinity to IA column components. PCR amplified unbound flow through was used as aptamer pool in successive rounds of SELEX. After each round of selection, bound and unbound DNA concentrations (Figure 2A) were measured and ELONA (Figure 2B) was performed to monitor the binding affinity of enriched aptamer pool. As the number of selection rounds progressed, bound DNA concentration and absorbance values gradually increased and remained in steady state during 9th and 10th round of SELEX indicating binding saturation of enriched high affinity aptamer pool. Incorporation of counter SELEX aided in the enrichment of AFB1 specific aptamer pool and elimination of cross reacting aptamers (Figure 2B). The stringency of the selection process was controlled by adjusting incubation time and washes. After final round of selection, the purified PCR products were cloned into E. coli TOP10 cells.

A total of 105 positive transformants were sequenced and their structural homology was analyzed using Multalin web based tool. Observations of resultant phylogenetic tree (Figure 3A) revealed 12 groups of enriched sequences with 60–99% homology and each group contained certain common sequences: group 1 (3/9 sequences), group 2 (3/7 sequences), group 3 (4/7 sequences), group 4 (2/9 sequences), group 5 (3/9 sequences), group 6 (9/14 sequences), group 7 (7/10 sequences), group 8 (5/8 sequences), group 9 (4/10 sequences), group 10 (7/10 sequences), and group 12 (5/8 sequences), respectively. Jalview analysis (Figure 3B) indicated that each group having few base pair length conserved regions which might play an important role in target binding. Possible Secondary structures analyzed using M-fold software showed maximum number of aptamers having typical stem loop structures (Supporting Figure S3). Three aptamers (AFLA5, AFLA53, and AFLA71) with most stable predicted secondary confirmations (lowest Gibbs free energy) are shown in Figure 3C. Total 16 aptamers from 12 groups were chosen for further characterization with one to three sequences from each group (Supporting Table 2). Putative QGRS obtained from QGRS Mapper for AFLA5 and AFLA71 showed higher G-scores indicating their G-quadruplex forming ability (Table 1; Supporting Table 3).

Relative binding affinity and selectivity of selected aptamers was measured by ELONA. Each aptamer was biotin labeled at the 5′ end and incubated at higher concentration (250 ng/μl) with constant concentration (200 ng) of BSA-AFB1. Unbound aptamer was removed by washing and reaction was developed after incubation with streptavidin conjugated horseradish peroxidase. The same analysis was performed with each of the closely associated and other mycotoxins used in the negative selection to check the specificity of obtained aptamers. As results shown in Figure 4, each sequence bound significantly better to AFB1 than to the control species. The DNA sequence library pool showed no significant binding toward the targeted mycotoxin, ruling out possibility of binding by non-specific polyanionic effects.

However, aptamers (AFLA12 – 46% to AFB2, AFLA41 – 54% to AFB2, 42% to AFG1, 33% to AFG2 and AFLA68 – 48% to AFB2, 40% to AFG1 and 30% to AFG2) exhibited affinity toward other aflatoxins when absorbance values of target toxin and aptamer complex was considered as 100 percent. This inference facilitate the possibility of utilizing mentioned aptamers for simultaneous detection two or more aflatoxins.

Apparent dissociation constants (K_d) of selected aptamers with high binding affinity and specificity were measured by ELONA. In this assay constant amount of BSA conjugated AFB1 (200 ng) was titrated with various concentrations (0–350 nM) of biotinylated aptamers (Figure 5). Tested eight aptamers showed affinities (K_d = 40–120 nM) (Table 2). Three highest affinity binders were chosen for detection of AFB1. As shown in Figure 6, aptamer candidates (AFLA5, AFLA53, and AFLA71) bound significantly to increasing concentrations of AFB1-BSA (10–250 ng/ml) and the detection limits (LOD) were approximately (20, 40, and 40 ng/ml), respectively (Supporting Figure S5). Aptamer AFLA5 and AFLA71 were further investigated for affinity purification of AFB1 by spiking studies.

To demonstrate the potential use of high affinity aptamers, spiking study was conducted using corn sample. Extracted toxin from spiked sample was purified applying biotinylated aptamers. The toxin was then confirmed and quantified by HPLC. A standard curve was obtained by plotting measured relative peak area (μV) against various toxin concentrations (10–250 ng/ml). A linear correlation (R² = 0.9963, 0.9942) was obtained for AFLA5 and AFLA71, respectively (Figures 8A,B). Considering relative peak area (μV) of standard toxin as 100, estimates for percentage of recovery of toxin from spiked samples were drawn. Aptamer, AFLA5 (LOQ – 53.74 ng) showed recovery of 82.2–96.21% and 78.3–94.22% that of AFLA71 (LOQ – 66.75 ng) with a retention time of 8.88 min (Figures 9A,B; Table 3). This demonstrates that, selected aptamers from present study can be successfully employed for the detection and purification of Aflatoxin B1 from food and agro commodities.