The oligonucleotides used for exoAgo were ordered from Sangon Biotechnology Co., Ltd., (Shanghai, China). The magnetic nanoparticle (MNP) was purchased from Nanjing XFNANO Materials Tech Co., Ltd., (Nanjing, China). The 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) were purchased from Sangon Biotechnology Co., Ltd., (Shanghai, China). TtAgo and Exo III were obtained from New England Biolabs Co., Ltd., (Beijing, China). RNase-free water (Takara Biotechnology Co., Ltd.) was used for all of the experiments.

The signal probe (SP) was constructed by covalently linking the FAM-probe to MNP. The sequence of the FAM-probe is NH₂-TTTTTTTTTTAGATCACAGATTTTGGGCGGGCCAAACTGCTGGGTG-CGGAAGAGAAAGAA-FAM. Briefly, the -COOH groups on carboxylated MNP (200 μL, 50 μg/mL) surface were activated by EDC and NHS, followed by adding an excess FAM probe (20 μM) and vibrating (600 rpm) the mixture under room temperature overnight. By this means, the FAM-probe was covalently linked with the MNP through the amido bond. The obtained SP was washed with 1× phosphate buffer three times and stored at 4°C for further use.

At first, the miRNA target was added into 1× phosphate buffer and hybridized with 250 nM of the L-gDNA through 85°C denaturation for 5 min and slowly annealed to room temperature (2°C/min) in 1× phosphate buffer. Then, 0.05 U/μL Exo III was added to hydrolyze the target-hybridized L-gDNA into the S-gDNA at 37°C for 30 min. After that, 50 nM TtAgo protein was added and the mixture was incubated at 75°C for 15 min in 1× Thermopol buffer (10 mM Na₂HPO₄, 1.75 mM KH₂PO₄, 137 mM NaCl, 2.65 mM KCl, pH 7.4) to make the S-gDNA assemble with TtAgo protein. Subsequently, 5 μL MNP-functionalized SP was added and incubated at 85°C for 40 min so that the assembled TtAgo/S-gDNA complex can cleave the phosphate diester bond of the SP at the 10-11 position of the S-gDNA complementary region. As a result, the fragments-bearing FAM fluorophore was separated from the MNP, and the fluorescence signal could be detected from the supernatant after the sample was placed on a magnetic rack.

The blood samples were first centrifuged at 1,600 g for 15 min. Then, the supernatant was further centrifuged at 16,000 g for another 15 min to obtain plasma. After that, 10 times continuously diluted let-7a (from 10 pM to 1 µM) were dispersed into 2% plasma and analyzed by exoAgo.

All obtained data are displayed as mean ± standard deviations (mean ± S.D.) of three independent experiments. Statistical analysis was performed with OriginPro 2023 10.0.

TtAgo can be served as a programmable endonuclease that is guided by a short DNA (∼20 nt) with a phosphate group at its 5′-end. Relatively, long DNA cannot be loaded into TtAgo protein even if it has a 5′-phosphorylated end (Swarts et al., 2014). On this basis, an Exo III-assisted Ago-mediated signal amplification system was designed for miRNA detection. The principle is shown in Figure 1. The L-gDNA with a 5′-phosphate group that was designed to have a complementary 3′-end with the target miRNA cannot directly assemble with the TtAgo protein. After being hybridized with the target miRNA, the miRNA-complementary region of the L-gDNA can be digested by Exo III, which can hydrolyze a single strand in a duplex with a blunt or a 5′-overhang but cannot digest a single-stranded DNA (ssDNA) and a double-strand DNA (dsDNA) with a 3′-overhang (Liu et al., 2020; Wang S. et al., 2021; Lee et al., 2022). In contrast, L-gDNA and Exo III could coexist in the same system without inference in the absence of the target miRNA. Only when the L-gDNA was recognized by the target miRNA and formed a duplex at its 3′-end could Exo III digest its 3′-end to produce TtAgo-available S-gDNA, which could guide TtAgo for further signal amplification. The signal probe was constructed by covalently functionalizing MNP with FAM probes that bearing an S-gDNA complementary region. As a result, activated TtAgo could efficiently cleave the FAM probe, and the signal of the FAM-labeled fragment could leave the MNP surface and be tested from the supernatant, and the fluorescence intensity could be proportional to the target miRNA concentration. Inversely, the sample without miRNA could not obtain cleaved L-gDNA and active TtAgo/S-gDNA complex. Accordingly, the FAM probes would still be linked with the MNP and adsorbed by the magnet, and the supernatant would have no fluorescence signal.

First, the effect of the length of gDNA on TtAgo cleavage activity was verified by polyacrylamide gel electrophoresis (PAGE). As shown in Figures 2A, B, various L-gDNAs with different lengths of 38, 39, or 40-nt, respectively, could not guide TtAgo to cleave substrate, while S-gDNA with lengths of 16, 17, or 18-nt, respectively, could guide TtAgo to efficiently hydrolyze the substrate, of which the sequence is the same as SP. Therefore, the S-gDNA with a length of 16 nt was selected to perform the following experiments. Accordingly, the L-gDNA was designed to have a 16-nt SP complementary region at its 5′-end and a target miRNA complementary region at its 3′-end. To further ensure that the S-gDNA could not be non-specifically cleaved by Exo III, a thiophosphate bond was modified between the 16 and 17 bases of the L-gDNA.

Considering that excess Exo III could cause non-specific DNA cleavage while deficient Exo III could result in a narrow dynamic range, the concentration of Exo III was optimized. Different concentrations of Exo III (from 0.5 U to 1.8 U) were incubated with the L-gDNA and L-gDNA+miRNA, respectively. The effect of Exo III on the L-gDNA was also analyzed by PAGE. The results in Figure 2C show that 1 U Exo III can cleave 1 μM L-gDNA into proper length in the presence of the target miRNA, while more Exo III induced non-specific cleavage of the L-gDNA in the absence of the target miRNA. Thus, 1 U Exo III was used to perform the following experiments.

Subsequently, the effect of the substrate’s length and the position of the gDNA recognition region on TtAgo cleavage activity were also investigated. For these purposes, we designed a series of substrates with various lengths and gDNA complementary region positions and incubated them with TtAgo/S-gDNA under 85°C temperature. The PAGE results are displayed in Figures 2D, E, F. It can be seen that the substrate without extra nucleotides at the left or right side of the gDNA complementary region could not be cleaved by TtAgo/S-gDNA no matter the length. Only the substrate with extra nucleotides at both the left and right side of the gDNA complementary region could be cleaved by TtAgo/S-gDNA. Therefore, it is necessary to consider the length and position of the gDNA complementary region when designing the TtAgo/gDNA substrate.

In addition, the TtAgo reaction conditions, including temperature and Mg²⁺ concentration, were also optimized. The exoAgo was performed under different temperatures (from 70°C to 90°C), and executed with various Mg²⁺ concentrations (from 2 to 10 mM), respectively. From the results in Figures 2G, H, 85°C temperature and 8 mM Mg²⁺ resulted in the highest cleavage efficiency of TtAgo. Therefore, exoAgo was carried out with 8 mM Mg²⁺ under 85°C.

Considering that the cleavage efficiency of TtAgo is significantly affected by the length of the substrate DNA, a typical fluorophore and quencher dual-labeled SP would not be suitable for exoAgo because the quenching efficiency is inversely proportional to the square of the distance between the fluorophore and quencher. The length of SP that should be beyond 31-nt (approximately 10.54 nm) out-of-work distance between fluorophore and quencher is much longer than the effective quenching distance (<10 nm). Also, the molecular beacon is not suitable for exoAgo because its metastable hairpin structure makes it hard to work under the high reaction temperature of TtAgo (85°C). Therefore, we designed a fluorophore (FAM)-functionalized MNP as the SP of exoAgo. In this way, the TtAgo cleavage can be quantified by MNP-mediated separation with the FAM-labeled cleavage fragment.

To verify the successful construction of the SP, the fluorescence signal, diameter, and ζ-potential of the SP were tested and compared to bare MNP. Before that, BSA was also added to block the site without a FAM-probe on the MNP surface, which may adsorb oligonucleotides non-specifically. The fluorescence detection results (Figure 3A) showed that bare MNP has no fluorescence, while the FAM-probe modified MNP has strong fluorescence signals. Also, it can be seen that the modification of the FAM probe increased the diameter of MNP from 169.9 to 267.2 nm (Figure 3B), and the ζ-potential changed from −4.4 to −17.9 mV (Figure 3C). From these results, it can be confirmed that the SP was successfully constructed.

In order to verify the cleavage ability of TtAgo on SP, the TtAgo/S-gDNA complex was directly assembled and incubated with SP under a temperature of 85°C, and the fluorescence signal from the supernatant was detected and compared with the pure SP solution. The fluorescence detection results of the supernatant in Figure 4A confirmed that the TtAgo/S-gDNA can efficiently cleave the SP. Under optimized conditions, we test the feasibility of exoAgo for let-7a. According to the PAGE results in Figure 4B, the exoAgo sample appears to have obvious bands of cleaved SP that are significantly lower than the bands of complete SP, while no cleaved SP bands appear for the blank control sample without let-7a. Therefore, the exoAgo is feasible for miRNA detection.

To analyze the sensitivity of exoAgo, a series of miRNA with different concentrations (from 10 pM to 1 μM) were added into the exoAgo system, respectively. The detection result is shown in Figure 4C. With the increase in the target miRNA concentration, the fluorescence signal also increased. The linear analysis results displayed that exoAgo can achieve a detection limit of 12.2 pM, which is defined as the concentration of the target that yields a net signal equivalent to three times the standard deviation of the blank control sample (n = 3) for let-7a detection with a linear equation of lg (FL) = 0.15 × lgC_let-7a + 3.4 (R ² = 0.995).

Furthermore, we tested the specificity of exoAgo by adding different miRNA (let-7a, let-7i, let-7f, and let-7g) from a homologous family into the exoAgo system, and tested the fluorescence signal of the supernatant. The sequences of let-7a, let-7i, let-7f, let-7g, and L-gDNA with thiophonsphate bond were listed in Figure 4D. The fluorescence detection results in Figure 4E indicate that only let-7a caused a significant fluorescence signal, while the signal of the other three miRNAs that have only 1∼2 base variation with let-7a was negligible. The corresponding PAGE results are shown in Figure 4F. The results confirmed that the exoAgo has a single-base resolution.

In order to test the detection performance of exoAgo for complex samples, we first investigated the specificity of exoAgo for plasma samples. A measurement of 1 µM of let-7a, let-7i, let-7f, and let-7g were added into the reaction system. Figure 5A shows that the fluorescence intensity gained from the sample with let-7a was extremely higher than that of other samples. Furthermore, the recovery (%) of exoAgo was evaluated. Different concentrations of let-7a (0.01, 0.1, 1, and 10 nM) were spiked into the diluted plasma sample. We collected the fluorescence intensity data from the supernatant after exoAgo and calculated the test concentration (C_t) according to the standard curves. The ratio of C_t and the actual added concentration of let-7a (C_s) was defined as the recovery. The results in Figure 5B shows that the recovery of exoAgo fluctuates around 100%, which indicates that the exoAgo possesses good reproducibility.