Characterization of the EACHA process
The EACHA reaction was verified using polyacrylamide gel electrophoresis (PAGE) analysis (Fig. S1). Upon addition of the primer to the mixture of H1 and H2, CHA products (H complexes) were formed. However, with the addition of RNase H, the RNA sequence within the complex was broken down, leading to a significant decrease in the amount of the H complex. The operation of EACHA and CHA was confirmed by real-time fluorescence analysis. I and II in Fig. 2A illustrate the CHA and EACHA reaction processes, respectively. As shown in Fig. 2B, in the absence of the primer, a small increase in fluorescence was observed for both EACHA and CHA. Nevertheless, EACHA exhibited an enhanced rapid fluorescence signal in response to the same concentration of primer P as the conventional CHA. The initial rate of the EACHA was significantly higher than that of the CHA. Moreover, the increase in the fluorescence signal of EACHA was significantly higher than that of CHA when additional aliquots of H2 were added to both systems (Fig. 2C). These results indicated a significant improvement in the reaction rate of EACHA. This may be due to the fact that in addition to the primer recycle, RNase H breaks up the RNA sequence of the H product into the original H1 state, resulting in high recyclability of the H1.
To verify the recycling of the primer and H1, the stepwise reaction process of EACHA was monitored using time-course fluorescence experiments. The primer labeled with FAM at the 3' end (FAM-P) and H1 labeled with BHQ1 at the 5' end (BHQ1-H1) were prepared. As shown in Fig. 2D, as the FAM-P/BHQ1-H1 complex formed, the fluorescence intensity decreased rapidly (Process I, black curve). In the CHA system, the hybridization of H2 and the FAM-P/BHQ1-H1 complex led to the release of FAM-P and the formation of the BHQ1-H1/H2 complex. Therefore, the decreased fluorescence signal gradually recovered and then reached a plateau (process II, red curve), indicating that the primer was recycled. In the EACHA system, the fluorescence intensity decreased (process I), increased (process II), and finally decreased again (process III, blue curve), indicating that the formed BHQ1-H1/H2 complex disassembled into BHQ1-H1, which further reacted with FAM-P to form FAM-P/BHQ1-H1, thus revealing that H1 was effectively recycled. To clearly show the structural changes in H1 during the recycling process, FAM/BHQ1-labeled H1 was monitored in real-time (Fig. 2E). The interaction between P and H1 (process I) and the CHA reaction (process II) could open the hairpin structure of H1, leading to an increase in the fluorescence intensity (black and red curves). However, the fluorescence signal followed a pattern of growth, decline, and then growth again in the EACHA system (process III, blue curve), which proved that H1 experienced a process of opening (formation of the P-H1 complex), closing (reverting the H1/H2 complex into the original state of H1), and opening (formation of the P-H1 complex again when H2 was exhausted). These results further suggest that H1 can be recycled well in the EACHA system. Meanwhile, PAGE experiments showed that the cleavage products of the H complex gradually increased with the increase in H2 concentration ranging from 1 to 4 µM (Fig. 2F), indicating that H1 could consume excessive amounts of H2.
EACHA exhibits superior kinetics and sensitivity
To highlight the rapid reaction rate of EACHA, the rate constants of conventional CHA and the innovative EACHA were evaluated by performing time-dependent fluorescence assays using primers at different concentrations. With an increase in primer concentration from 0 to 16 nM, the fluorescence variation in CHA and EACHA gradually increased (Fig. 3A and 3B). The Kobs values (observed rates) were measured during the linear growth phases at various primer concentrations and fitted to a linear equation with an intercept. The equation \({K}_{obs}={V}_{0}+\frac{({V}_{\text{m}\text{a}\text{x} }-{ V}_{0})\left[\text{P}\right]}{{K}_{\text{M}} + \left[\text{P}\right]}\) is used to calculate Vmax (the maximum reaction rate) and KM.[39]Here, [P] is the primer concentration, and V0 is the Kobs at zero primer concentration. As shown in Fig. 3C, the Kobs of CHA and EACHA increased linearly with increasing primer concentrations, with rate constants of 0.012 nM-1 s-1 and 0.451 nM-1 s-1, respectively. The rate constant of EACHA was approximately 37.6-fold larger than that of CHA (Fig. S2).
Next, the sensitivity of EACHA was evaluated and compared with that of CHA and primer hybridization reactions without signal amplification by detecting primer P at different concentrations under optimal reaction conditions, such as 0.5 U of RNase H and a H1 to H2 ratio of 3:4 (Fig. S3). The primer hybridization reaction, in which primer P hybridizes and opens BHQ1/FAM-labeled H1 to produce a fluorescence signal, could only detect as low as 5 nM primer P, with a detection range from 5 to 200 nM (Fig. 3D and S4). The lowest detectable concentration of CHA was 0.1 nM for primer P, with a detection range from 0.1 to 40 nM (Fig. 3E and S5), which was similar to that previously reported.[39–41] The developed EACHA responded to primer P in a broad concentration range from 1 pM to 20 nM, with the lowest detectable concentration of 1 pM (Fig. 3F and S6), which was 100-fold lower than that of CHA and 5000-fold lower than that of the primer hybridization reaction. Researchers have devised a variety of DNA circuits based on the classic CHA, including self-replicating CHA,[42] dual-layer CHA,[43] and localized CHA.[44] Compared with CHA circuits that require complicated sequence designs or long reaction times to obtain high sensitivity, EACHA exhibited comparable sensitivity, simple design, and short assay time. Moreover, our work presents, for the first time, the hairpin reactant recycling of CHA with the assistance of an enzyme without the need for additional DNA reactants.
Development of an EACHA-based method for rapid and one-pot detection of Salmonella enterica Enteritidis
Motivated by the remarkable improvements in the sensitivity and reaction rate of EACHA, we used it to develop a method for S. Enteritidis detection. Figure 4A illustrates the developed allosteric probe (AP)-triggered EACHA (AP-EACHA) method for the one-pot, rapid detection of S. Enteritidis. This method uses AP and EACHA substrates. Specifically, the AP contains two functional regions: an aptamer domain for identifying S. Enteritidis and a primer (P) domain for triggering the subsequent EACHA reaction. The annealed AP had a hairpin structure, and its P domain was blocked by the aptamer sequence. In the presence of S. Enteritidis, the aptamer domain of AP specifically recognizes S. Enteritidis, resulting in its exposure to P (Fig. 4B). EACHA was then triggered, producing an amplified fluorescence signal for S. Enteritidis detection.
To prove the capacity of AP to bind to S. Enteritidis and subsequently expose the P sequence, the ends of AP were linked to FAM and BHQ1. As shown in Fig. 4B, laser scanning confocal microscopy (LSCM) revealed that S. Enteritidis incubated with AP appeared green (right) compared to those without any treatment (left). The generated fluorescence signals gradually improved with increasing S. Enteritidis concentrations from 102 to 108 CFU mL− 1 (Fig. S7A). Moreover, flow cytometry results showed that a significantly enhanced signal was observed in the experimental group (S. Enteritidis treated with AP) compared to the negative control (S. Enteritidis treated with FAM-labeled random strands) and the blank control (S. Enteritidis incubated with 1× PBS buffer; Fig. S7B). These results demonstrated that AP could successfully identify S. Enteritidis and expose the P domain.
To evaluate the feasibility of the AP-EACHA method for S. Enteritidis detection, the fluorescence signals for different reaction and substrate combinations were compared. As depicted in Fig. 4C, in comparison to other control methods in the absence of AP or S. Enteritidis, this method generated a stronger fluorescence signal, even when the concentration of S. Enteritidis was decreased from 108 CFU mL− 1 to 2 × 101 CFU mL− 1 (P < 0.001 or P < 0.01). These results prove the feasibility of this method for the one-pot detection of S. Enteritidis. It is worth noting that the fluorescence intensity increased quickly and peaked within 10 min in the presence of a high concentration of S. Enteritidis (108 CFU mL− 1); however, at low concentrations (2 × 101 CFU mL− 1), the fluorescence intensity peaked at approximately 20 min (Fig. S8). To achieve excellent sensitivity for S. Enteritidis detection, the reaction time for the AP-EACHA method was set to 20 min. Furthermore, to guarantee that AP could not only adequately interact with different concentrations of bacteria but also trigger a minimized background signal, the concentration ratio of AP to the hairpins was optimized in the presence of 108 CFU mL− 1 of S. Enteritidis. As shown in Fig. 4D, the ratio of the fluorescence signal to background noise (S/BG) peaked when the ratio of AP, H1, and H2 was 2:3:4, while the S/BG decreased with a further increase in AP concentration. Therefore, the optimal AP concentration was determined to be 200 nM. Subsequently, we investigated the most favorable temperature for AP-EACHA. Compared to other temperatures, the strongest fluorescence signal and the highest S/BG were obtained at 37 ℃ (Fig. 4E). This may be attributed to the negative impact of lower or higher temperatures on the activity of RNase H. Therefore, the reaction temperature of this method for S. Enteritidis detection was set at 37 ℃.
Simple, sensitive, and specific detection of S. Enteritidis using AP-EACHA
To evaluate the sensitivity of AP-EACHA, a series of samples containing different concentrations of S. Enteritidis were prepared by spiking with known amounts of S. Enteritidis. S. Enteritidis was first counted by a plate colony counting method (Fig. S9). As shown in Fig. 5A, the fluorescence intensity gradually increased as the concentration of S. Enteritidis increased from 2 × 101 to 2 × 107 CFU mL− 1. A fine linear calibration curve was observed between the fluorescence signal and the concentration of S. Enteritidis ranging from 2 × 101 to 2 × 105 CFU mL− 1, and the corresponding regression equation was Y = 39.8 X + 317 (R2 = 0.9988), where Y represents the fluorescence signal and X represents S. Enteritidis concentration (Fig. 5B). The limit of detection (LOD) was determined to be 15 CFU mL− 1 based on three standard deviations from the blank control (black dashed line; Fig. 5C). Furthermore, the genomic DNA of S. Enteritidis extracted from the prepared samples was analyzed using a clinical RT-PCR method. Although the RT-PCR method showed a wide linear range from 2 × 102 to 2 × 107 CFU mL− 1, it was unable to detect 2 × 101 CFU mL− 1 of S. Enteritidis, given that a cycle threshold (Ct) value greater than 35 is clinically interpreted as a negative test result (Fig. S10). Therefore, the sensitivity of the AP-EACHA method was higher than that of RT-PCR. In contrast to other previously reported methods (Table S1), the AP-EACHA strategy showed a shorter detection time and a lower LOD.
Subsequently, the specificity of the AP-EACHA was evaluated. As shown in Fig. 5D, a markedly improved fluorescence signal was observed in the presence of S. Enteritidis compared to other bacterial species, including Shigella flexneri, Listeria monocytogenes, Klebsiella pneumoniae, Staphylococcus aureus, Acinetobacter baumannii, Vibrio parahaemolyticus, Escherichia coli, and Pseudomonas aeruginosa. Even though ten-fold concentration of these bacteria was mixed with S. Enteritidis, minimal variations in fluorescence intensity were observed between samples of S. Enteritidis and these mixed bacteria (Fig. 5E). These results indicated the high specificity of AP-EACHA for S. Enteritidis. Furthermore, we demonstrated that AP-EACHA could identify other S. Enteritidis strains derived from humans and mice, such as CMCC 50040, CMCC 50041, and CMCC 50335 (Fig. 5F).
Detection of S. Enteritidis in complex matrices
To validate the ability of AP-EACHA to detect S. Enteritidis in complex matrices, we performed a recovery experiment by detecting the bacteria at concentrations of 20, 103, and 105 CFU mL− 1 in undiluted serum, drinking water, and 10-fold diluted milk, respectively. The signal corresponding to 105 CFU mL− 1 of S. Enteritidis was recovered at 100.38%, 100.14%, and 100.10% in undiluted serum, drinking water, and 10-fold diluted milk, respectively. Even when incubated with 20 CFU mL− 1 of S. Enteritidis, the recovery rates of AP-EACHA were 98.58%, 99.62%, and 97.50% in undiluted serum, drinking water, and 10-fold diluted milk, respectively (Fig. S11 and Table S2). These preliminary results indicate that this method was capable of detecting S. Enteritidis in various complex samples.
To demonstrate the application of AP-EACHA in discerning contaminated foods, 60 milk samples, half of which were contaminated with S. Enteritidis, were prepared. 10 µL of the prepared samples were directly analyzed by AP-EACHA (Fig. 6A). Fluorescence signal analysis showed that although the three contaminated milk samples could not be identified based on the cut-off line (Fig. 6B), AP-EACHA remained effective in distinguishing between contaminated and pasteurized milk (P < 0.001; Fig. 6C). This is probably because the concentration of S. Enteritidis in these three samples was lower than 15 CFU mL− 1. Furthermore, RT-PCR results showed that these two groups had differential genomic DNA abundances of S. Enteritidis (P < 0.001; Fig. 6D), and the amplification curves of several contaminated samples still overlapped with those of the pasteurized milk samples (Fig. S12). Receiver operating characteristic (ROC) analysis indicated that AP-EACHA showed an area under the curve (AUC) of 0.997 in discriminating contaminated milk from pasteurized milk samples, with a sensitivity and specificity of 96.7% (Fig. 6E), which was better than that of RT-PCR, which had an AUC of 0.982, a sensitivity of 96.7%, and a specificity of 90.0% (Fig. 6F and Table S3). These results indicate the potential application of AP-EACHA for discriminating S. Enteritidis-infected foods.
Rapid diagnosis of S. Enteritidis infection in mice
To evaluate the potential application of AP-EACHA in clinical diagnosis, fecal and serum samples from mice were collected and analyzed (Fig. 7A). Specifically, six mice were infected by injecting S. Enteritidis solution orally, and six healthy mice were injected with the same amount of saline and used as the control group. Subsequently, their feces and blood were collected, and S. Enteritidis was detected every 1 h for the first 6 h and then every 12 h for 5 days. We observed that the infected mice showed clinical symptoms after 30 min, and one infected mouse died at 24 h.
The average fluorescence signal from the fecal samples of the infected mice increased gradually over the first 3 h and then rapidly decreased between 3 and 60 h, after which the average signal improved slightly and decreased again (Fig. S13A), indicating that the amount of S. Enteritidis in the feces remained unchanged over time. This may be the result of a strong interaction between S. Enteritidis and gut microbiota. Similarly, we concluded that the number of bacteria in the serum samples of infected mice peaked within three hours and then decreased gradually (Fig. S13B), which may have resulted from the immune system clearing pathogenic bacteria. At all stages of testing, the signals from the feces and serum samples of the infected mice were significantly higher than those from the feces and serum samples of the healthy mice (P<0.05; Fig. 7B and 7C), and ROC curves of the signals from serum and feces samples showed AUC values of 0.966 and 0.966, respectively, with a specificity of 92.6% 89.7% and a sensitivity of 97.1% and 94.1% for the serum fecal samples, respectively (Fig. S14 and Table S3).
Furthermore, we evaluated the diagnostic performance of AP-EACHA at different S. Enteritidis infection stages in mice, including the early (2–6 h) and late stages (12 h − 5 d). As shown in Fig. 7D, for the feces samples, AP-EACHA showed effective diagnostic performance in the identification of early-stage (AUC = 1.00; sensitivity = 100%; specificity = 100%) and late-stage infections (AUC = 0.959; sensitivity = 87.5%; specificity = 90%). Similarly, for the serum samples, this method showed good performance in identifying early-stage (AUC = 0.986; sensitivity = 97.2%; specificity = 100%) and late-stage infections (AUC = 0.980; sensitivity = 96.2%; specificity = 91.7%) (Fig. 7E and Table S3). The diagnostic performance of AP-EACHA in early-stage infection was superior to that during late-stage infection because S. Enteritidis was gradually cleared by the immune system. By comparing the AUC values, we suggest detecting S. Enteritidis in fecal samples in the earlier stages of infection and serum samples in the later stages of infection. Collectively, our results indicated that AP-EACHA has the potential to diagnose bacterial infections.