The sample pad, conjugate pad, nitrocellulose membrane (NC), absorbent pad and backing card were purchased from the Jie Yi Biotechnology Co., Ltd. (Shanghai, China). The streptavidin-immobilized gold nanoparticles (SA-G), rabbit anti-fluorescein antibody (anti-FITC) and biotinylated bovine serum albumin (biotin-BSA) were purchased from the Resenbio Co., Ltd. (XiAn, China). The QIAamp DNA Stool Mini Kit and QIAamp DNA Mini Kit (QIAamp DNA minikits; Qiagen, Hilden, Germany) were purchased from Qiagen (Beijing, China). Loopamp^TM Fluorescent Detection Reagent (FD) and the Loopamp kits were purchased from Eiken Chemical (Beijing, China).

The dry-reagent dipstick (5 mm × 70 mm), illustrated in Figure 1, consisted of an absorbent pad, a NC membrane, a conjugate pad and an immersion pad assembled on a plastic adhesive backing card. The capture reagents, including anti-FITC (0.15 mg/ml) and biotin-BSA (4 mg/ml) in 0.01 M phosphate-buffered saline (PBS, PH 7.4), were dispensed onto the reaction regions. On the NC membrane, there are two zones as the test zone (conjugated with anti-FITC) and control zone (conjugated with biotin-BSA), with each line separated by 5 mm. SA-G in 0.01M PBS (PH 7.4) was deposited on the conjugate pad of the biosensor. Then, the assembled cards were cut at 5 mm widths, and the biosensors were dryly stored at the room temperature until use.

A 0.5 μl aliquot of MCDA amplicons was deposited to the sample application area of the biosensor. Then, the strip was directly immersed into 120 μl of running buffer (10 mM PBS, PH 7.4 with 1% Tween 20) and the biosensor allowed absorbing the whole running buffer. After 2 min, the MCDA product detection was visualized in the form of red lines on the NC membrane.

In order to design Shigella spp. specific MCDA primers, the nucleotide sequence of the specific ipaH gene (GenBank accession no. M32063) was downloaded from the NCBI Genbank database, and a set of MCDA primers was designed by PrimerExplorer V4 (Eiken Chemical, Japan) and primer software PRIMER PREMIER 5.0 (Thiem et al., 2004). Blast analysis demonstrated that the MCDA primer set was specific for Shigella spp. strains. The details of primer design, primers sequences, locations and modifications of MCDA primers were displayed in Figure 2 and Table 1. All of the oligomers were synthesized and purified by TsingKe Biological Technology (Beijing, China) at HPLC purification grade.

A total of 60 bacterial strains were used in this study (Table 2). Twenty-three Shigella strains and 37 non-Shigella strains were included to test the specificity of the MCDA-LFB assay. All bacterial strains were stored in 10% (w/v) glycerol broth at -70°C and then were refreshed three times on nutrient agar plate at 37°C. The genomic DNA templates were extracted from all culture strains using DNA extraction kits according to the manufacturer’s instructions. The extracted templates were examined with ultraviolet spectrophotometer (Nano drop ND-1000, Calibre, Beijing, China) at A260/280 and stored under at -20°C before the templates were used. The strains of S. flexneri serovar 1d (ICDC-NPS001) were applied for confirmation performance, optimal temperature, sensitivity analysis and practical application conducted in the report. Moreover, the genomic templates of S. flexneri serovar 1d (ICDC-NPS001) were serially diluted (10 ng, 10 pg, 10 fg, 1 fg, and 0.1 fg) for sensitivity evaluation of MCDA-LFB detection.

In order to assess the feasibility of ipaH-MCDA primers, the MCDA reaction was conducted as the standard MCDA condition, which has been reported in previous report (Wang et al., 2015c). In brief, the MCDA assay was carried out in 25-μl amplification mixtures containing the following components: 0.4 μM each of displacement primers F1 and F2, 0.8 μM each of amplification primers C1 and C2, 1.2 μM each of amplification primers R1, R2, D1^∗ and D2, 1.2 μM each of cross primers CP1^∗ and CP1, 2.4 μM cross primer CP2, 12.5 μl 2× reaction mix (Loopamp kits), 1.25 μl of Bst DNA polymerase (10 U) and 1 μl DNA template. Four monitoring techniques, including colorimetric indicator (FD), gel electrophoresis, turbidimeters (LA-320C) and LFB detection, were employed to analyze the MCDA products. Furthermore, the endonuclease restriction-mediated real-time multiple cross displacement amplification (ET-MCDA), which was reported in a recent study, was employed to achieve real time fluorescence measurement of MCDA reaction (Wang et al., 2015c, 2016b).

Then, we tested the optimal reaction temperature of ipaH-MCDA primers. The MCDA reaction mixtures were performed at a constant temperature ranging from 60°C to 67°C for 1 h and then incubated at 85°C for 5 min to stop the amplification. Mixtures with 1 μl genomic template of Listeria monocytogens strain (L. monocytogenes, ATCC19114) and Salmonella strain (ICDC-NPsa001) were used as negative controls, and mixtures with 1 μl double distilled water (DW) were used as a blank control.

The templates of S. flexneri serovar 1d (ICDC-NPS001) were serially diluted to confirm the limit of detection (LoD), which was defined by genomic DNA amount of the template. The analytical sensitivity of MCDA by colorimetric indicator (FD reagent), real time turbidity, 2% agarose gel electrophoresis, real time fluorescence and LFB detection was determined as described above. At least three replicates of each dilution were examined to test the analytical sensitivity.

In order to assess the analytical specificity of MCDA-LFB methodology, the MCDA reactions were carried out under the conditions described above with purely genomic templates from 23 Shigella strains and 37 non-Shigella strains (Table 2). The MCDA products were tested using 2.5% agarose gel electrophoresis and LFB detection. Analysis of each sample was examined in at least two independent experiments.

Human fecal samples were acquired from a healthy donor with the written informed consent. Our study was reviewed and approved by the ethics committee of the National Institute for Communicable Disease Control and Prevention, China CDC, according to the medical research regulations of the Ministry of Health China (Approval No. ICDC2014003).

In order to evaluate the suitability of MCDA-LFB technique as a surveillance tool for Shigella, the MCDA-LFB assay was applied to rapidly diagnose the target pathogens in human fecal samples. Firstly, the human fecal samples were confirmed as being Shigella-negative by culture-based methods and PCR detection. Then, to test the minimal detectable colony forming units (CFUs), the cultures with S. flexneri strains were serially diluted (10^-1 to 10^-9), and the aliquots of 100 μl appropriate dilution (10^-6) was spread in triplicate onto brain heart infusion (BHI) agar. The CFUs were counted after 24 h at 37°C. Simultaneously, the aliquots of 100 μl appropriate dilution (10^-3 to 10^-8) with S. flexneri strains were inoculated into the fecal samples (0.2 g), and the number of Shigella was adjusted to approximate 1.42 × 10⁶, 1.42 × 10⁵, 1.42 × 10⁴, 1.42 × 10³, 1.42 × 10² and 1.41 × 10¹ CFU/g. Then, the artificially contaminated stool samples were applied to extract the genomic DNA templates, and the supernatants (2 μl) were used for MCDA detections. Non-contaminated fecal sample was used as negative control and this analysis was independently conducted in triplicate. The MCDA products were also analyzed by colorimetric indicator (FD reagent), real time turbidity, 2% agarose gel electrophoresis, real time fluorescence and LFB detection as described above.

A schematic of MCDA-LFB technique was shown in Figure 1. In the MCDA-LFB system, the cross primer (CP1 or CP2) involved in MCDA reaction were labeled at the 5′ end with FITC, and the amplification primers (D1 or D2) were modified at the 5′ end with biotin (Figure 1A). The new CP1, CP2, D1, and D2 primers were named as CP1^∗, CP2^∗, D1^∗, and D2^∗, respectively. For clarity, the CP2^∗ and D2^∗ primers were not displayed in outline of MCDA reaction during the reaction stage (Figure 1B). The CP1^∗ primer initiated MCDA reaction at the P1s site of the target sequence, and the newly synthesized strand was displaced by upstream synthesis from the primer F1 (Step 1). Five primers (D1^∗, C1, R1, CP2, and F2) annealed to the newly generated strand, and then the Bst polymerase extended in tandem producing four different products (Step 2). The D1^∗ product was used as the template by C1 and CP1^∗ primers, enter a cyclic process (Step 3, Cycle 1). In the cycle, a larger amounts of double-labled detectable amplicons, which contained biotin-labeled D1^∗ primer and a FITC-labeled CP1^∗ primer, were successfully yielded. The details of the reaction process for C1, R1, and CP2 products (Step 4, 5, 6) has been reported in previous study (Wang et al., 2015c). In addition, a double-labeled detectable product (CP2^∗/D2^∗ product), which was similar to the detectable CP1^∗/D1^∗ product, could be formed when the CP2 primer was modified with a FITC at the 5′ end and D2 primer for biotin.

The principle of LFB for visualization of MCDA amplicons was exhibited in Figure 1C. The LFB detected MCDA amplicons through specific recognition of the FITC labels at the end of products, which were formed by using FITC labeled primers (CP1^∗ primer). The other end, the amplicons labeled with biotin binded streptavidin-conjugated gold nanoparticles for visualization. The MCDA products were deposited onto on the sample application region of the biosensor, and then the biosensor was directly immersed in the running buffer. The running buffer moved along the biosensor by capillary action, which rehydrated the immobilized detector reagents (SA-G). The target amplicons was specifically captured by the immobilized anti-FITC at the first test zone and detector reagents rapidly accumulate in the reaction zone of the strip through biotion/streptavidin interaction, resulting in a visual red colored line on the test region. The proper function of the strip is demonstrated by the control line formation which contained biotinylated bovine serum albumin that captured excess detector reagent.

In order to verify the feasibility of Shigell-MCDA primers, the MCDA reactions were carried out in the presence or absence of genomic DNA templates within 60 min at a constant temperature (65°C). Three monitoring techniques, including colorimetric indicator (FD reagent), gel electrophoresis analysis and LFB detection, were employed to confirm the Shigella-MCDA products. A color shift of positive amplification in Shigella-MCDA tubes was directly observed from light gray to green (Figure 3A). The positive MCDA products were seen many bands of different sizes in a typical ladder-like pattern on ethidium bromide-stained 2% agarose gel electrophoresis, but not in the negative and blank control (Figure 3B). It was also observed that two visible red bands (Test line, TL; Control line, CL) were seen in positive amplifications, and only the CL were seen in negative and blank controls (Figure 3C). Therefore, the MCDA primer set was a good candidate for establishment of the MCDA-LFB method for Shigella detection.

In order to examine the optimal assay temperature during the amplification stage, the Shigella-MCDA reactions were conducted at eight distinct temperatures (60°C–67°C) with 1°C intervals. The strain S. flexneri serovar 1d (ICDC-NPS001) was employed as the positive control to evaluate the optimal amplification temperature at the level of 10 pg genomic templates per reaction. The reactions were analyzed by means of real time turbidity detection and the typical kinetics graphs corresponding to eight temperatures were obtained (Figure 4). Eight reaction temperatures provided a robust signal, with the faster amplifications generated from assay temperature of 63°C–67°C, which were recommended as the standard temperature for Shigella-MCDA-LFB assay during the amplification stage. The assay temperature of 65°C was used for the rest of MCDA-LFB tests conducted in this study.

The analytical sensitivity of MCDA-LFB technique on Shigella was determined by analyzing the products yielded from the serial dilutions (10 ng, 10 pg, 10 fg, 1 fg, and 0.1 fg per microliter) of Shigella genomic DNA in triplicate (Figure 5). The Shigell-MCDA reactions were monitored by real time measurement of turbidity and the LoD of MCDA-LFB assay for Shigella detection was 10 fg of genomic templates per reaction (Figure 5A). By the FD reagent, a color shift of positive amplification in Shigella-MCDA tubes was directly observed from light gray to green (Figure 5B). Then, the Shigella-MCDA products were analyzed by 2% agarose gel electrophoresis and positive products were observed as the ladder-like patterns but not in negative reactions, negative control and blank control (Figure 5C). The biosensor was also subjected to detect the Shigella-MCDA products (Figure 5E). As expected, the biosensor exhibited clear visible red bands for both TL and CL when the products came from positive MCDA amplifications, and only the CL were generated from for negative MCDA amplifications, negative control and blank control. The LoD of MCDA-LFB assay for detecting ipaH gene was also 10 fg of genomic templates per reaction. Moreover, the LoD of ET-MCDA assay for Shigella detection was also 10 fg of genomic DNA in pure culture (Figure 5D). These results indicated that the analytical sensitivity by FD reagent, real time turbidity, real time fluorescence and agarose gel electrophoresis detection for Shigella-MCDA amplifications was conformity with biosensor analysis.

Then, we assessed the optimal duration of time require for the MCDA-LFB assay during the amplification stage, and four different reaction times (10, 15, 20, and 25 min) were compared at 65°C according to the standard MCDA conditions. The lowest genomic DNA level (10 fg of Shigella templates per tube) showed two red bands when the reaction only lasted for 20 min at 65°C (Figure 6). A reaction time of 20 min was used as the optimal time for the MCDA-LFB assay during the reaction stage. Hence, the whole procedure, including specimen (such as fecal sample) processing (35 min), isothermal reaction (20 min), and result reporting (5 min), could be completed within 60 min.

The analytical specificity of the MCDA-LFB technique was evaluated by MCDA-LFB amplification of genomic DNA extracted from 23 Shigella strains and 37 non-Shigella strains (roughly 10 ng of genomic templates for each pathogen). The detection was positive only for the four Shigella species, and was negative for non-Shigella species and blank control (Figure 7). As shown in Figure 7, two red bands, including TL and CL, appeared on the biosensor from the positive test, and only a red band at the control line appeared, indicating negative results for non-Shigella strains and blank control. The results demonstrated that the MCDA-LFB assay has a 100% analytical specificity for Shigella detection.

In order to determine the suitability of the MCDA-LFB assay as a nucleic acid detection tool, the MCDA-LFB approach was examined by the artificially inoculating Shigella strains into human fecal samples. As shown in Figure 8A, the MCDA-LFB assay could generate positive results when the contaminated numbers of Shigella were more than 1.42 × 10³ CFU/g (∼5.68 CFU/reaction). The MCDA-LFB approach produced the negative results when the contaminated numbers of Shigella were lower than 1.42 × 10² CFU/g (∼0.568 CFU/reaction). Only a red band at the control line appeared, indicating negative results for negative control and blank control. Thus, the LoD of MCDA-LFB method was 5.68 CFU per tube, which was consistent with MCDA-FD, MCDA-turbidity and MCDA-gel electrophoresis assays (Figures 8B,D,E). In contrast, the analytical sensitivity of ET-MCDA assay for detection of Shigella in fecal samples was also 5.68 CFU per reaction, which was as sensitive as MCDA-LFB detection (Figures 8A,C). The results indicated that the analytical sensitivity of MCDA-LFB assay was in complete accordance with MCDA-FD, MCDA-turbidity, MCDA-gel electrophoresis and ET-MCDA assays.