A Lateral Flow Strip Based on a Truncated Aptamer-Complementary Strand for Detection of Type-B Aflatoxins in Nuts and Dried Figs

Type-B aflatoxins (AFB1 and AFB2) frequently contaminate food, especially nuts and fried figs, and seriously threaten human health; hence, it is necessary for the newly rapid and sensitive detection methods to prevent the consumption of potentially contaminated food. Here, a lateral flow aptasensor for the detection of type-B aflatoxins was developed. It is based on the use of fluorescent dye Cy5 as a label for the aptamer, and on the competition between type-B aflatoxins and the complementary DNA of the aptamer. This is the first time that the complementary strand of the aptamer has been used as the test line (T-line) to detect type-B aflatoxins. In addition, the truncated aptamer was used to improve the affinity with type-B aflatoxins in our study. Therefore, the lengths of aptamer and cDNA probe were optimized as key parameters for higher sensitivity. In addition, binding buffer and organic solvent were investigated. The results showed that the best pair for achieving improved sensitivity and accuracy in detecting AFB1 was formed by a shorter aptamer (32 bases) coupled with the probe complementary to the AFB1 binding region of the aptamer. Under the optimal experimental conditions, the test strip showed an excellent linear relationship in the range from 0.2 to 20 ng/mL with a limit of detection of 0.16 ng/mL. This aptamer-based strip was successfully applied to the determination of type-B aflatoxins in spiked and commercial peanuts, almonds, and dried figs, and the recoveries of the spiked samples were from 93.3%−112.0%. The aptamer-complementary strand-based lateral flow test strip is a potential alternative tool for the rapid and sensitive detection of type-B aflatoxins in nuts and dried figs. It is of help for monitoring aflatoxins to avoid the consumption of unsafe food.


Introduction
Type-B aflatoxins, including aflatoxin B 1 and B 2 (AFB 1 and AFB 2 ), produced by Aspergillus species, have the highest toxicity among mycotoxins that seriously threaten the health of humans and animals [1].

Principles for the Aptamer-Complementary Strand-Based Lateral Flow Test Strip
The principle for the aptamer-based lateral flow test strip (ALFs) was based on the competition for Cy5-Apt between the complementary strand on the test-line and AFB 1 in the samples (Figure 1). In the absence of AFB 1 , the solution was chromatographed from the sample pad to the nitrocellulose (NC) membrane, and the Cy5-labeled aptamer was first combined with a partially complementary strand at the test-line, as shown in Figure 1a-1. The remaining aptamer was then combined with the poly T (Table 1) of the control line, as shown in Figure 1a-2, which resulted in a higher fluorescence value of the test-line and a lower fluorescence value of the control line. Conversely, in the presence of Toxins 2020, 12, 136 3 of 12 AFB 1 , the Cy5-labeled aptamer preferred to bind with AFB 1 (Figure 1b-3), rather than couple with the partially complementary strand at the test-line (Figure 1b). In other words, the more aptamer-AFB 1 complexes there were in the solution, the less free Cy5-labeled aptamer was captured on the test-line. As a result, the fluorescence intensity of the test-line was decreased. Moreover, aptamer probes would hybridize with poly T on the control line in any case, ensuring the detection validity. Quantitative analysis was performed by measuring the fluorescent intensity of the two lines with analysis software, and the intensity ratio of the test-line to control line (T/C) was inversely proportional to the AFB 1 concentration in the samples.
Toxins 2020, 12, x FOR PEER REVIEW  3 of 12 other words, the more aptamer-AFB1 complexes there were in the solution, the less free Cy5-labeled aptamer was captured on the test-line. As a result, the fluorescence intensity of the test-line was decreased. Moreover, aptamer probes would hybridize with poly T on the control line in any case, ensuring the detection validity. Quantitative analysis was performed by measuring the fluorescent intensity of the two lines with analysis software, and the intensity ratio of the test-line to control line (T/C) was inversely proportional to the AFB1 concentration in the samples. In the presence of AFB1, the Cy5-labeled aptamer preferred to bind with AFB1 (b-3), and the fluorescence intensity of the T-line decreased. The aptamer bound to AFB1 was still bound to poly T of the C-line (b-4), while the fluorescence intensity of the C-line increased, and the value of test-line to control line (T/C) was used to quantify AFB1.

Optimization of the Aptamer Length and Concentration
The length of AFB1 aptamer is a key factor for consideration when rationally designing a structure-switching aptamer [22,23]. A previous study showed that the short segment close to the 3' end of the 50-mer AFB1 aptamer was not essential for target binding [21], and the use of the shorter aptamer could save costs without compromising the AFB1 recognition capability. Based on this result, the length of AFB1 aptamer was optimized with ALFs analysis. The DNA sequences of the aptamers are presented in Table 1. The apt 32 and apt 26 for AFB1 binding in our study were truncated from the reference aptamer (apt 50), which was first reported by Le et al. [24], and their predicted secondary structures are shown in Figure 2a. The three Cy5-labeled aptamers were mixed with the AFB1 standard of 100 ng/mL at a ratio of 1:9. Fluorescence intensities were measured with and without the target. As shown in Figure 2b, there was no significant difference in the T/C values among three aptamers. However, for the T-line, the fluorescence intensity of apt 50 and apt 32 added to the target was lower than that of apt 26, indicating weaker binding between AFB1 and apt 26. All three aptamers could bind with the AFB1, which further confirmed that the sequence of the loop region is important for aptamer binding affinity. Moreover, compared with apt 26, apt 32 contains three more G-C base pairs, which may further stabilize the secondary structure and strengthen the binding towards AFB1. The previous study also showed that the proper numbers of base pairs in the stem region is required for the aptamer to have strong binding affinity to AFB1, which is helpful to Figure 1. Schematic illustration of adapter side-flow tomography sensor for detecting aflatoxin B 1 (AFB 1 ). (a) In the absence of AFB 1 , the Cy5-labeled aptamer was first combined with a partially complementary strand at the test-line (T-line) (a-1); the remaining aptamer was then combined with the poly T of the control line (C-line) (a-2). (b) In the presence of AFB 1 , the Cy5-labeled aptamer preferred to bind with AFB 1 (b-3), and the fluorescence intensity of the T-line decreased. The aptamer bound to AFB 1 was still bound to poly T of the C-line (b-4), while the fluorescence intensity of the C-line increased, and the value of test-line to control line (T/C) was used to quantify AFB 1 .

Optimization of the Aptamer Length and Concentration
The length of AFB 1 aptamer is a key factor for consideration when rationally designing a structure-switching aptamer [22,23]. A previous study showed that the short segment close to the 3' end of the 50-mer AFB 1 aptamer was not essential for target binding [21], and the use of the shorter aptamer could save costs without compromising the AFB 1 recognition capability. Based on this result, Toxins 2020, 12, 136 4 of 12 the length of AFB 1 aptamer was optimized with ALFs analysis. The DNA sequences of the aptamers are presented in Table 1. The apt 32 and apt 26 for AFB 1 binding in our study were truncated from the reference aptamer (apt 50), which was first reported by Le et al. [24], and their predicted secondary structures are shown in Figure 2a. The three Cy5-labeled aptamers were mixed with the AFB 1 standard of 100 ng/mL at a ratio of 1:9. Fluorescence intensities were measured with and without the target. As shown in Figure 2b, there was no significant difference in the T/C values among three aptamers. However, for the T-line, the fluorescence intensity of apt 50 and apt 32 added to the target was lower than that of apt 26, indicating weaker binding between AFB 1 and apt 26. All three aptamers could bind with the AFB 1 , which further confirmed that the sequence of the loop region is important for aptamer binding affinity. Moreover, compared with apt 26, apt 32 contains three more G-C base pairs, which may further stabilize the secondary structure and strengthen the binding towards AFB 1 . The previous study also showed that the proper numbers of base pairs in the stem region is required for the aptamer to have strong binding affinity to AFB 1 , which is helpful to form a stable stem loop structure [21]. Judging by the performance of these three aptamers, apt 32 was selected as the AFB 1 binding aptamer in this study.  The concentration of the aptamer has a great influence on the sensitivity and detection limit of the test strip. In order to determine the optimal aptamer concentration, the target was mixed with different concentrations of apt 32 (0.005, 0.01, 0.02, 0.03, 0.04, 0.05 µM) in a ratio of 9:1, and 60 µL mixture was loaded onto the test strip. The T/C inhibition rate was used to evaluate the analytical performance. The fluorescence intensities (T and C-line) were gradually enhanced with the increase of aptamer levels, but the T/C inhibition rate sharply decreased from 0.02 to 0.05 µM apt 32 ( Figure 2c). Moreover, there was no significant difference in the T/C inhibition rate from 0.005 to 0.02 µM apt 32. Since the T-line fluorescence intensity with 0.005 and 0.01 µM of aptamer concentration was low, this may lead to large systematic errors. Hence, 0.02 µM apt 32 was the best choice for the optimal concentration for further studies.

Optimization of the Complementary Strand on the T-Line
Apart from forming a complex with its target, the aptamer can also hybridize with a complementary DNA strand. Therefore, the sequence of the complementary strand could also have a significant impact on the interaction between aptamer and target [22,23]. Ideally, the complementary strand on the T-line should be able to capture free aptamer in the absence of target. On the other hand, it should not compete with the target for aptamer binding. In this study, we synthesized four types of complementary DNA modified with biotin. The sequences of the complementary DNA are listed in Table 1, which include one full complement sequence (probe 1) with apt 32, and three partial complementary sequences (probe 2, probe 3, and probe 4) as DNA probes. The length of each partial complementary ( Figure 3a) strand was 15 bases. All DNA probes were tested to identify the optimal DNA probe as the complementary strand for the ALFs assay. As shown in Figure 3b, in the presence of 100 ng/mL AFB 1 standard, the T-line fluorescence intensity of all four DNA probes decreased. However, the inhibition rates of probe 3 and 4 were significantly lower than that of probe 1 and 2, suggesting that neither of them is suitable for serving as the complementary strand. In the case of probe 1 and 2, the inhibition rates were calculated as 87.4% (probe 1) and 95.9% (probe 2), respectively. The relatively lower inhibition rate of probe 1 might be caused by the competitive binding between the complementary strand and AFB 1 target. A previous study showed that the aptamer tends to hybridize more strongly with complementary DNA than a target if the length of complementary DNA is the same as that of an aptamer [25]. Based on this result, probe 2, with the highest inhibition rate, was chosen as the complementary strand for the T-line.

Optimization of the Complementary Strand on the T-Line
Apart from forming a complex with its target, the aptamer can also hybridize with a complementary DNA strand. Therefore, the sequence of the complementary strand could also have a significant impact on the interaction between aptamer and target [22,23]. Ideally, the complementary strand on the T-line should be able to capture free aptamer in the absence of target. On the other hand, it should not compete with the target for aptamer binding. In this study, we synthesized four types of complementary DNA modified with biotin. The sequences of the complementary DNA are listed in Table 1, which include one full complement sequence (probe 1) with apt 32, and three partial complementary sequences (probe 2, probe 3, and probe 4) as DNA probes. The length of each partial complementary ( Figure 3a) strand was 15 bases. All DNA probes were tested to identify the optimal DNA probe as the complementary strand for the ALFs assay. As shown in Figure 3b, in the presence of 100 ng/mL AFB1 standard, the T-line fluorescence intensity of all four DNA probes decreased. However, the inhibition rates of probe 3 and 4 were significantly lower than that of probe 1 and 2, suggesting that neither of them is suitable for serving as the complementary strand. In the case of probe 1 and 2, the inhibition rates were calculated as 87.4% (probe 1) and 95.9% (probe 2), respectively. The relatively lower inhibition rate of probe 1 might be caused by the competitive binding between the complementary strand and AFB1 target. A previous study showed that the aptamer tends to hybridize more strongly with complementary DNA than a target if the length of complementary DNA is the same as that of an aptamer [25]. Based on this result, probe 2, with the highest inhibition rate, was chosen as the complementary strand for the T-line.

Optimization of Reaction Conditions between Aptamer and AFB1
The composition of the binding buffer has a crucial effect on spatial conformation of the aptamer, which directly affects the binding strength between aptamer and target. Therefore, the composition of the binding buffer was also optimized in this study. Firstly, three commonly used binding buffers (Tris-HCl, HEPES, and PBS) were compared, then the concentration of Mg 2+ was investigated. Finally, the type and maximum concentration of organic solvent in the binding buffer

Optimization of Reaction Conditions between Aptamer and AFB 1
The composition of the binding buffer has a crucial effect on spatial conformation of the aptamer, which directly affects the binding strength between aptamer and target. Therefore, the composition of the binding buffer was also optimized in this study. Firstly, three commonly used binding buffers (Tris-HCl, HEPES, and PBS) were compared, then the concentration of Mg 2+ was investigated. Finally, the type and maximum concentration of organic solvent in the binding buffer was evaluated.
The ideal buffer can improve the specificity and stability of the test strips, and thus the fluorescence intensity of C-line was measured in order to determine the binding strength between the aptamer and poly-T in different buffer systems. Among the three tested buffers, Tris-HCl and HEPES provided similar inhibition rates (89.2% and 91.1%, respectively), while the inhibition rate was significantly lower in the PBS buffer system (Figure 4a). On the other hand, the T-line fluorescence intensity of Tris-HCl buffer was higher than that of HEPES buffer. According to this result, Tris-HCl buffer was identified as the binding buffer for the subsequent experiments.
result, Tris-HCl buffer was identified as the binding buffer for the subsequent experiments.
The concentration of Mg 2+ has a significant effect on the hybridization of DNA strands, it also affects the binding affinity of targets and aptamers by influencing DNA structure [26,27]. For the purpose of optimizing Mg 2+ concentration in our system, binding buffers with different concentrations of MgCl2 (10, 20, 50, 100, 200, 400 mM) were prepared. As the concentration of Mg 2+ increased, not only did the fluorescence intensity of T-line increase significantly, but also the inhibition rate increased from 50.3% to 89.1% (Figure 4b). As a result, 200 mM of Mg 2+ in Tris-HCl buffer was identified to be the optimal concentration.
Acetonitrile and methanol are widely introduced to address the poor solubility of AFB1 in aqueous medium. However, an excess amount of organic solvent might lead to incomplete binding between DNA strands [28] and could even brake down the structure of the NC membrane. Therefore, it is necessary to determine which type of organic solvent is more suitable for this sensor. Moreover, the maximum ratio of organic solvent should also be determined. In this experiment, different volumes of acetonitrile or methanol were added to prepare the binding buffer with 0, 5, 10, 20, 30, and 40% of organic solvent, respectively. The results in Figure 4c showed that the addition of acetonitrile greatly affected the binding between the aptamer and target, as the inhibition rate obviously decreased from 84.0% to 16.3% with the increase of acetonitrile concentration (0 to 20%). In contrast, the inhibition rate was basically unchanged upon addition of methanol (Figure 4d). Further comparing of T-line fluorescence intensity revealed that the concentration of methanol should not exceed 10% to guarantee the minimum interference induced by the solvent.  The concentration of Mg 2+ has a significant effect on the hybridization of DNA strands, it also affects the binding affinity of targets and aptamers by influencing DNA structure [26,27]. For the purpose of optimizing Mg 2+ concentration in our system, binding buffers with different concentrations of MgCl 2 (10, 20, 50, 100, 200, 400 mM) were prepared. As the concentration of Mg 2+ increased, not only did the fluorescence intensity of T-line increase significantly, but also the inhibition rate increased from 50.3% to 89.1% (Figure 4b). As a result, 200 mM of Mg 2+ in Tris-HCl buffer was identified to be the optimal concentration. Acetonitrile and methanol are widely introduced to address the poor solubility of AFB 1 in aqueous medium. However, an excess amount of organic solvent might lead to incomplete binding between DNA strands [28] and could even brake down the structure of the NC membrane. Therefore, it is necessary to determine which type of organic solvent is more suitable for this sensor. Moreover, the maximum ratio of organic solvent should also be determined. In this experiment, different volumes of acetonitrile or methanol were added to prepare the binding buffer with 0, 5, 10, 20, 30, and 40% of organic solvent, respectively. The results in Figure 4c showed that the addition of acetonitrile greatly affected the binding between the aptamer and target, as the inhibition rate obviously decreased from 84.0% to 16.3% with the increase of acetonitrile concentration (0 to 20%). In contrast, the inhibition rate was basically unchanged upon addition of methanol (Figure 4d). Further comparing of T-line fluorescence intensity revealed that the concentration of methanol should not exceed 10% to guarantee the minimum interference induced by the solvent.

Specificity Test
The specificity towards the target AFB 1 by using our developed ALFs was studied. Several relevant mycotoxins including AFB 2 , aflatoxin G 1 (AFG 1 ), AFG 2 , ochratoxin A (OTA), zearalenone (ZEN), fumonisin B 1 (FB 1 ), and deoxynivalenol (DON) were measured at the same concentration of 100 ng/mL. As shown in Figure 5a, the aptamer showed similar high binding affinity to AFB 1 and AFB 2 . No change of T-line fluorescence intensity was observed compared with the control for the other types of mycotoxins, which indicated that the ALFs had high specificity for type-B aflatoxins. Furthermore, it showed much decreased binding strength to AFG 1 and AFG 2 . The T-line fluorescence intensity of AFG 1 and AFG 2 was 2.3-and 2.2-fold higher than that of AFB 1 , respectively. The distinction of AFB 1 and AFB 2 proved to be challenging because of the high similarity between their chemical structures ( Figure 5b). Presumably, the C-C double bond on one of the furan rings of AFB 1 does not participate in the aptamer binding. However, it is clear that the tested sensor showed high binding affinity to these two groups of aflatoxins than other types of mycotoxins. More importantly, the ALFs can simultaneously detect both AFB 1 and AFB 2 , which is an advantage for food security. The specificity towards the target AFB1 by using our developed ALFs was studied. Several relevant mycotoxins including AFB2, aflatoxin G1 (AFG1), AFG2, ochratoxin A (OTA), zearalenone (ZEN), fumonisin B1 (FB1), and deoxynivalenol (DON) were measured at the same concentration of 100 ng/mL. As shown in Figure 5a, the aptamer showed similar high binding affinity to AFB1 and AFB2. No change of T-line fluorescence intensity was observed compared with the control for the other types of mycotoxins, which indicated that the ALFs had high specificity for type-B aflatoxins. Furthermore, it showed much decreased binding strength to AFG1 and AFG2. The T-line fluorescence intensity of AFG1 and AFG2 was 2.3-and 2.2-fold higher than that of AFB1, respectively. The distinction of AFB1 and AFB2 proved to be challenging because of the high similarity between their chemical structures ( Figure 5b). Presumably, the C-C double bond on one of the furan rings of AFB1 does not participate in the aptamer binding. However, it is clear that the tested sensor showed high binding affinity to these two groups of aflatoxins than other types of mycotoxins. More importantly, the ALFs can simultaneously detect both AFB1 and AFB2, which is an advantage for food security.

Quantitative Analysis of Type-B Aflatoxins
Under the optimized conditions, the sensitivity of the test strip was examined. The AFB1 and AFB2 stock solution were diluted into the concentrations of 0.2, 0.5, 1, 2, 5, 10, 20 ng/mL. The standard solutions were mixed with 0.02 μM aptamer solution with the ratio of 9:1. After 10 min, the mixed solution was loaded onto the test strip and subsequently scanned with a card reader. It was shown that, as the concentration of type-B aflatoxins increased, the fluorescence intensity of the T-line gradually decreased, while that the fluorescence intensity of the C-line increased, which resulted in the value of T/C decreasing (Figure 6a). The calculated F (T/C) value had a good linear relationship with the logarithm of the type-B aflatoxins concentration. A linear range between 0.2 and 20 ng/mL was determined for AFB1 and AFB2 with a correlation coefficient (r) of 0.9869 and 0.9906, respectively (Figure 6b). The limit of detection (LOD) was determined to be as low as 0.16

Quantitative Analysis of Type-B Aflatoxins
Under the optimized conditions, the sensitivity of the test strip was examined. The AFB 1 and AFB 2 stock solution were diluted into the concentrations of 0.2, 0.5, 1, 2, 5, 10, 20 ng/mL. The standard solutions were mixed with 0.02 µM aptamer solution with the ratio of 9:1. After 10 min, the mixed solution was loaded onto the test strip and subsequently scanned with a card reader. It was shown that, as the concentration of type-B aflatoxins increased, the fluorescence intensity of the T-line gradually decreased, while that the fluorescence intensity of the C-line increased, which resulted in the value of T/C decreasing (Figure 6a). The calculated F (T/C) value had a good linear relationship with the logarithm of the type-B aflatoxins concentration. A linear range between 0.2 and 20 ng/mL was determined for AFB 1 and AFB 2 with a correlation coefficient (r) of 0.9869 and 0.9906, respectively (Figure 6b). The limit of detection (LOD) was determined to be as low as 0.16 ng/mL. This result demonstrated that our test strip is capable of quantitative analysis of type-B aflatoxins at an extremely low level. Although the sensitivity of our developed ALFs is similar to that of a previous study (Table 2), this is the first time that ALFs have been applied for the simultaneous detection of AFB 1 and AFB 2 . Moreover, the other benefit of these ALFs is that the complementary strand instead of hapten-BSA as test line can save costs, and facilitate synthesis and reproducibility.

Analysis of Type-B Aflatoxins in Foods
To evaluate the reliability and accuracy of the prepared ALFs biosensors, peanuts, almonds, and dried figs samples were spiked with type-B aflatoxins standard solution at a concentration of 3 and 10 μg/kg. The spiked samples were processed by simply extracting with acetonitrile and filtering through an solid phase extraction (SPE) cartridge. The extracts were re-dissolved in binding buffer and analyzed by our ALFs. The recovery rates were summarized in Table 3. The recovery range of the different matrix for AFB1 was 96.5%-109.4% with the relative standard deviation (RSD) of 2.0% to 6.5%; for AFB2 it was 93.3%-112.0%, and the RSD ranged from 2.7% to 7.9%. The limit of quantitation (LOQ) was determined to be 0.5 μg/kg, which is far below the European Union (EU) limit. This result matches well with the standard HPLC-MS/MS method. It is proven in this study that the proposed ALFs are capable of detecting type-B aflatoxins in food samples. Seventeen samples of peanuts, dried figs, and almonds were selected and extracted. The extracts were measured using ALFs and HPLC-MS/MS, respectively. The comparison results showed that AFB1 and AFB2 were not detected from the collected samples, and the results of the two detection methods were consistent.

Analysis of Type-B Aflatoxins in Foods
To evaluate the reliability and accuracy of the prepared ALFs biosensors, peanuts, almonds, and dried figs samples were spiked with type-B aflatoxins standard solution at a concentration of 3 and 10 µg/kg. The spiked samples were processed by simply extracting with acetonitrile and filtering through an solid phase extraction (SPE) cartridge. The extracts were re-dissolved in binding buffer and analyzed by our ALFs. The recovery rates were summarized in Table 3. The recovery range of the different matrix for AFB 1 was 96.5%-109.4% with the relative standard deviation (RSD) of 2.0% to 6.5%; for AFB 2 it was 93.3%-112.0%, and the RSD ranged from 2.7% to 7.9%. The limit of quantitation (LOQ) was determined to be 0.5 µg/kg, which is far below the European Union (EU) limit. This result matches well with the standard HPLC-MS/MS method. It is proven in this study that the proposed ALFs are capable of detecting type-B aflatoxins in food samples. Seventeen samples of peanuts, dried figs, and almonds were selected and extracted. The extracts were measured using ALFs and HPLC-MS/MS, respectively. The comparison results showed that AFB 1 and AFB 2 were not detected from the collected samples, and the results of the two detection methods were consistent.

Conclusions
In summary, we introduced an ultrafast, sensitive, specific, and easy-to-operate method for type-B aflatoxins analysis by designing and preparing a lateral flow strip biosensor based on the specific recognition of type-B aflatoxins by an aptamer labeled with a fluorescent tag. This sensor improved the sensitivity and accuracy towards AFB 1 by employing a shortened aptamer sequence, which made it more suitable for practical applications. The fluorescence intensity on the strip could be recorded by a portable strip reader. For quantitative detection, the ratio of fluorescence intensities between T and C-line provided a good linear relationship with the logarithm of the type-B aflatoxins concentration (0.2-20 ng/mL). By using this ALFs sensor, rapid detection of type-B aflatoxins could be achieved within 20 minutes. Finally, peanuts, almonds, and dried figs samples were spiked with type-B aflatoxins and analyzed using this method. The result verified that our ALFs could satisfy the requirement of on-site detection of spiked samples. Compared with the traditional HPLC-MS/MS technique, the reported sensing device has the advantages of being low cost, fast to analyze and easy to operate, which make it more suitable for on-site and rapid monitoring of type-B aflatoxins contamination in food samples. Moreover, this strategy is also applicable to construct test strips for other mycotoxins.
The absorbent pad (H5076), sample pad (GL-b04), nitrocellulose (NC) membrane (Sartorius CN140), and PVC plastic adhesive backing used in the experiments were purchased from Shanghai Jieyi Biotechnology Co., Ltd, China. Each lateral flow chromatography strip was cut using a 3050 dispensing platform and KM-3100 Cutter (BioDot, Irvine, CA, USA). The fluorescence signals of the test strip were scanned by an ESEQuant LR3 FL Premium reader (Qiagen, Frankfurt, Germany).
The AFB 1 aptamer sequence [23,24] and complementary probe sequences used in this project were listed in Table 1. All DNA probes were synthesized by Sangon Biotechnology Co., Ltd. (Shanghai, China).

Preparation of ALFs
The ALFs are composed of four parts: a sample pad, a NC membrane, an absorbent pad, and a PVC plastic adhesive backing. The sample pad, NC membrane, and absorbent pad is sequentially attached to the PVC plastic adhesive backing from left to right according to the direction of migration. The sample pad and absorbent pad sequentially laminate 3 mm with the NC membrane to guarantee smooth migration of solutions. The NC membrane was used to prepare test zones.
The NC membrane can only immobilize proteins on its surface, so the detection probe and the control probe labeled with biotin are reacted with streptavidin to form a biotin-streptavidin conjugate, which was anchored on the NC membrane in advance. The 3 µM biotin labeled partial complementary strand (biotin-probe 1) and 3 µM poly-T (biotin-probe 2) were mixed separately with streptavidin (1 mg mL −1 ) by 7: 1 (v/v) and incubated for 2 hours at room temperature. The streptavidin-biotin labeled DNA probe 1 (SA-biotin probe 1) and probe 2 (SA-biotin probe 2) were dispensed on the NC membrane of test (T) line and control (C) line using the 3050-dispensing platform, respectively. The distance between the two lines is 5 mm. It was dried at 37 • C for 1 h and the whole assembled sheet was cut into 4 mm strips by KM-3100 Cutter, and then stored in a desiccator before use.

Type-B Aflatoxins Analysis
The sample was mixed with Cy5-labeled aptamer in an Eppendorf tube at a ratio of 9:1 (v/v) and incubated at room temperature for 10 min. After that, each 60 µL portion of the mixture was loaded on the sample pad of an ALFs. All liquid was allowed to migrate freely along the strip until absorbed by the absorbent pad. After 10 min, the fluorescence intensity on the T and C-lines was recorded by scanning the test strip with ESEQuant LR3 FL Premium Reader.

Optimization of Test Conditions
In order to achieve the desirable performance of the ALFs, a series of impact factors that potentially affect the T and C-lines in the test strip were optimized, including the lengths and concentrations of the aptamer, the sequence of the complementary strands and the compositions of binding buffer solution (including which type of buffer, the concentration of Mg 2+ , and the ratio of organic solvent). The inhibition rate was introduced for evaluating the performance of the sensors under different conditions. It was calculated following the formula: T 0 and C 0 correspond to the fluorescence intensity of T and C-line in the absence of AFB 1 , while T 1 and C 1 represent the fluorescence intensity of T and C-line in the presence of 100 ng/mL AFB 1 . In the case of conditions with similar inhibition rates, the value of T 1 was compared to determine the optimal condition.
For quantitative analysis, the ratio of T and C fluorescence intensities was calculated as the concentration evaluation index of type-B aflatoxins, which was used to draw the standard curve.

Analysis of Food Samples
In this study, peanut, almond, and dried figs were chosen as food samples to further confirm the outcome of practical application by spiked recovery. A total of 17 samples consisting of peanuts (9), almonds (3), and dried figs (5) were randomly collected from different supermarkets and local markets in Beijing, China. The extraction of the specimen was performed following previous work [32]. In detail, 5.0 g of sample spiked with known amount of AFB 1 was placed into a 50 mL polypropylene centrifugation tube and diluted with 25 mL 80% acetonitrile aqueous solution containing 100 mM citric acid. After that, the tube was transferred to a shaker and shaken at room temperature with a speed of 150 rpm for 30 minutes. Subsequently, the sample was centrifuged at 10,000 rpm for 10 min at 4 • C using a centrifuge. A volume of 5.0 mL supernatant layer was aspirated and filtered through a SPE cartridge column. Finally, the collected solution was evaporated to dryness at 50 • C under nitrogen atmosphere, and re-dissolved in 1 mL binding buffer as the extract solution. For each measurement, a mixture containing 54 µL of the extract solution and 6 µL of the Cy5 labeled aptamer solution was added on the aptamer-based flow chromatography test strip. The average recovery was calculated by comparing with the actual concentration of type-B aflatoxins.