Principle of Hybridization Assemble Nicked Displacement Amplification
The principle of HANDA was shown in Fig. 1. The “head” of the hairpin L6S0H2 was designed with the recognition area of the Nt.BbvCI. Target H0 (TH0) bound to two hairpin probes, L6S0H1 and L6S0H2, and triggered HCR to form a long dsDNA with gaps. Then Klenow Fragment (3’-5’exo) polymerase (KF) were added to combined with the gap of the HCR product, and used one of the chains as a template for polymerization. When polymerized to the template chain gap, the long HCR fragment disintegrated, forming a double-stranded structure including H1 and TH0 sequence hybridization and H1 and H2 staggered hybridization, which contained a large number of Nt.BbvCI recognition sites. In the subsequent amplification, Nt.BbvCI recognized and cleaved a single strand in the double strand to form a gap. KF combined with the gap and extended the DNA according to the template, and then replaced another ssDNA, which acted as a round of strand displacement amplification cycle. After several rounds of cycles, a large amount of ssDNA would be produced, which could hybridize with the fluorescent reporter probe to dissociate the DNA modified with the quenching group [32], resulting in the recovery of fluorescence intensity. If TH0 did not exist, H1 and H2 did not perform HCR. Moreover, the hairpin loop had a single-stranded structure, and Nt.BbvCI could not recognize the cleavage, that is, non-specific amplification would not be triggered, ensuring a low background value. The base sequence of the non-cutting site can be arbitrarily designed according to the target, so this method can be used for the detection of various types of targets.
Agarose gel electrophoresis to verify the feasibility of HANDA
In the electrophoresis experiment, primers L6S0H1 and L6S0H2 were used to amplify hairpin probes, and the results were shown in Fig. 2A. Comparing lane 3 and lane 4, we could find that the band of lane 4 was significantly brighter, indicating that KF would polymerize along 5'-3' to complement the sticky end of L6S0H1 [33]. Moreover, no other miscellaneous strips appeared in the two lanes, which proved that the card issuing would not assemble spontaneously. Lane 5 showed bands of different lengths indicating that HCR occurred. Lane 6 was the system where KF was added after the HCR reaction, showing three bright bands: the bottom was the unreacted hairpin band; the middle band (about 70 bp) was the dsDNA formed by KF polymerization after the primer sequence hybridized with L6S0H1; the upper band was a complete double-stranded structure formed by KF polymerization after dislocation hybridization of two L6S0H1 and L6S0H2. The length of it was about 90 bp, which was 1.5 times the hairpin after it was fully opened. Lane 7 showed a new short band. The added Nt.BbvCI would specifically recognize the site of the hairpin assembly product to cut the single strand, KF would bind to the gap and polymerize along the 5'-3' direction to replace the single strand. Lane 8 was the product formed by adding TH0, L6S0H1 and KF.
Amplification kinetics of HANDA
The two ends of the hairpin L6S0H2F were respectively labeled with fluorophore (6-FAM) and quencher (BHQ-1). When the hairpin was closed, 6-FAM and BHQ-1 were close to each other and the fluorescence was quenched. If the hairpin was opened, the fluorophore moved away from the quencher to restore the fluorescence. The dsDNA reporter probes were labeled with 6-FAM /BHQ-1. If TH0 was present, the dsDNA probe hybridized with the final amplification product, and 6-FAM would far away from the BHQ-1 to restore fluorescence. The amplification kinetics experiment measured the change of fluorescence intensity as the reaction progresses to infer the kinetics of amplification.
The result was shown in Fig. 2B. Section “a” was the HCR process in the presence of TH0. The sticky end of hairpin L6S0H1F hybridized with TH0 so that the hairpin was opened to expose the recognition fragment of L6S0H1F, and then L6S0H2 recognized and hybridized alternately. The progress of HCR made the fluorophore/quencher labeled at both ends away from each other to restore the fluorescence, which was specifically manifested as an increase in the fluorescence value of the “a”. As the amplification progressing, the energy of HCR gradually entered a steady state, that is, the amplification basically stagnated (the fluorescence curve entered a plateau). The “b” segment showed that the fluorescence value had further recovered. Because with the assistance of KF, the distance of some fluorophore/quencher pairs became farther (Fig. 2C). After adding Nt.BbvCI, the system would cut and polymerize the formed template. Hybridization of the amplified single strand with the reporter probe was achieved to continuously repeat the amplification process. More amplified fragments hybridized with the reporter probe to further increase the fluorescence value until the substrate was exhausted.
Exploration of the sequence design rules of hairpin probe
HANDA designed the nicking enzyme recognition sequence in the hairpin loop. Therefore, the number of bases in the hairpin loop and the size of the gap after the formation of dsDNA would affect the efficiency of HANDA. Gel electrophoresis and real-time fluorescence were used to explore the optimal number of bases in the hairpin loop. In this experiment, a total of 7 pairs of hairpins were designed: L6S0H1-L6S0H2, L10S0H1-L10S0H2, L14S0H1-L14S0H2, L18S0H1-L18S0H2, L10S4H1-L10S4H2 L14S8H1-L14S8H2, L18S12H1-L18S12H2. Among them, "L10" in the number "L10S4H1-L10S0H2" meant that the number of loop bases was 10; "S4" meant that the nicking enzyme recognition site was 4 bases away from the 5'end.
Figure 3A showed the electrophoresis diagram of the assembled hairpins: L6S0H1-L6S0H2, L10S0H1-L10S0H2, L14S0H1-L14S0H2, L18S0H1-L18S0H2. Among them, lanes 1–4 were hairpins mixed systems without TH0. It could be seen that there was no amplified band except for the shallow band around 60 bp in lane 3, indicating that the specificity of the hairpin was better. Lanes 5–8 were the HCR assembly system for hairpin after adding TH0. Lanes 5 and 6 had amplified bands, indicating that L6S0H1-L6S0H2, L10S0H1-L10S0H2 had stronger hybridization and amplification ability. Lanes 9–12 were added KF based on the previous period. It could be seen that two bands were produced at 70 bp and 100 bp, and the band amplified by the original HCR basically disappeared. Comparing lanes 13–16 with lanes 9–12, there was a band at 20 bp, indicating that Nt.BbvCI could assist in amplifying a large amount of ssDNA. Moreover, the band in lane 13 was the brightest, proving that L6S0H1-L6S0H2 had the strongest hairpin amplification ability.
Real-time fluorescence quantitative PCR was used to monitor the fluorescence changes during the amplification process, and the ratio of each card signal value to the background value (signal-to-noise ratio) was determined when the amplification time was 120 min (Fig. 3B). The fluorescence values of L6S0H1-L6S0H2 reached a plateau first before 120 min, and the signal-to-noise ratio was 12.0 (Fig. 3C). Compared with the signal-to-noise ratios of others, L6S0H1-L6S0H2 had the shortest amplification time, the highest signal-to-noise ratio and the strongest amplification ability. It might be that the two sticky ends of the hairpin were completely paired with the loop, making it easy to achieve a more stable state [34]. The results of this experiment were consistent with the conclusions of the electrophoresis experiment, so L6S0H1-L6S0H2 was used to for subsequent experiments.
Signal amplification capability of HANDA
The experiment used the same hairpin and TH0 concentration to compare the amplification ability (Fig. 4). The concentration of TH0 was quantified using conventional HCR amplification and the HANDA established in this experiment. The standard curve of conventional HCR was y = 1671.1x-727.16 (R2 = 0.9924), the detection linear range was between 7.8125 nM-1 µM, and the detection limit (S/N = 3) was 7.1 nM. The detection standard curve of HANDA was y = 1553.6x + 2720.6 (R2 = 0.9965), the detection linear range was 0.0305 nM-250 nM, and the detection limit (S/N = 3) was 26.31 pM. The linear range of HANDA was one order of magnitude wider than that of the conventional HCR amplification detection method, and the sensitivity was also increased by 270 times, which proved that HANDA had super signal amplification capability, which could significantly improve detection sensitivity.
The principle and feasibility of ricin detection
The principle of ricin detection was shown in Fig. 5. The magnetic beads coupled with the ricin monoclonal antibody were sealed and combined with the ricin and the colloidal gold probe to form a sandwich structure. After removing the unreacted gold probe by magnetic separation, DTT was added. DTT destroyed the gold-sulfur bond formed by the sulfhydryl-labeled barcode DNA and the gold nanoparticles to free the barcode DNA bound to the gold probe. The supernatant, collected by magnetically separated, was mixed with L6S0H1 and L6S0H2. The barcoded DNA in the supernatant would trigger HCR to form a long chain structure with gaps that hybridized with H1 and H2. After adding KF and dNTP, KF could insert into the gap and then polymerize along the 5'-3' end with complementary single-stranded as a template, disintegrating the originally formed HCR long chain to form a double-stranded fragment containing barcode and L6S0H1 hybridization, and the double-stranded fragments hybridized between L6S0H1 and L6S0H2, both of which were template of strand displacement amplification. Then the addition of Nt.BbvCI would recognize the sites on the templates to cut the single strands. KF inserted and then polymerized along the 5'-3' ends to replace the single strands, so that a large amount of ssDNA would be generated after continuous cycle amplification. They hybridized with the reporter probe to restore fluorescence. The fluorescence value was directly proportional to the concentration of the barcode DNA and ricin, which could detect the concentration of ricin with high sensitivity.
Three concentrations of ricin at 50 ng/mL, 5 ng/mL, and 0 ng/mL were detected, and their fluorescence values showed a gradient downward trend, indicating that the presence of ricin could trigger subsequent HANDA (Fig. 6A). At the same time, the fluorescence intensity of the reporter probe and the negative value (0 ng/mL) basically overlap, indicating that when there was no target, the supernatant collected by magnetic separation did not contain barcode DNA, so the hairpin probe was stably present in the system.
Analytical performance of DNA barcodes-HANDA
After the detection system was optimized, the detection range and sensitivity of the method needed to be further investigated. The detection result was shown in Fig. 6B/C. As the concentration of ricin increased, the fluorescence value at 520 nm also gradually increased. Different concentrations of ricin samples were used as the abscissa, and the fluorescence value of the corresponding target minus the negative fluorescence value (△F) was used as the ordinate to establish a standard curve. In the range of 0.2441 ng/mL-1000 ng/mL, the detected fluorescence value had a good linear relationship with the concentration of ricin. The linear equation was y = 242.25x + 213.21 (R2 = 0.9926), and the detection limit (S/N = 3) was 85.11 pg/mL.
Acacia toxin (Arbin), BSA, SEA, and IgG at a concentration of 20 ng/mL were selected as structural analogs for specificity experiments (Fig. 6D). Only ricin had an obvious response fluorescence value, and the fluorescence values of other targets were similar to the blank value. This was due to the good recognition specificity of monoclonal antibodies. In addition, this method had also been successfully applied to actual sample testing (Table S2).