The design of H. pylori-specific primers for MCDA reaction was performed based on the nucleotide sequence of ureB gene (accession no. CP003904) using the Primer Premier Version 5.0 software according to the principle of MCDA (Wang et al., 2015). The MCDA primers were composed of 10 primers targeting 10 distinct regions of target sequence, including a pair of cross primers (CP1 and CP2), a pair of displacement primers (F1 and F2), and three pairs of amplification primers (C1 and C2, D1 and D2, and R1 and R2). For biosensor detection, one amplification primer D1 was additionally labeled a fluorescein isothiocyanate (FAM) at the 5′ end (termed D1^#). Sequences of all the primers are shown in Table 1 . Details of primer sequences, locations, and modifications are shown in Figure 1C , and all the primers were synthesized and purified at HPLC purification grade by Tianyi‐Huiyuan Biotech. Co, Ltd. (Tianjin, China).

The universal isothermal amplification kits, biotin‐14‐dCTP, visual detection reagent (VDR), and LFB strips were obtained from HuiDeXin Biotechnology (Tianjin, China), and the article numbers are HC-300-L, HT10005, and HT10004. The DNA extraction kits were purchased from TransGen Biotech (Beijing, China, HT0600). Real-time turbidity (LA-320C) was provided by Eiken Chemical Co., Ltd (Japan).

A total of 40 gastric antrum mucosa samples collected from patients admitted at the Department of Gastroenterology, Capital Institute of Pediatrics from September 2021 to March 2022 were employed in this study. A total of 17 of the 40 samples were diagnosed as H. pylori infection by the positive results of urea breath test or stool antigen combining with rapid urease test (RUT), while the remainder were considered as functional gastrointestinal disorder with an H. pylori-negative result. Clinical samples were taken from the minor curvature of the gastric antrum and one piece from the greater curvature of the gastric body (the biopsy location of the gastric antrum was within 5 cm from the pylorus opening, and the sampling location of the stomach body was within 8 cm from the cardia). Genomic DNA of the 40 clinical samples were extracted and purified using the commercially available kit (Beijing Transgen Biotech Co., Ltd.) as per the manufacturer’s instruction. The extract nucleic acid was stored at −20°C for subsequent use. In addition, 29 non-Helicobacter pathogens obtained from the Chinese Center for Disease Control and Prevention (CDC) were processed as the clinical samples, and the extracted genomic DNA samples were employed for the specificity analysis of the HP-MCDA-LFB ( Table 2 ).

As per the previous studies (Wang et al., 2015, 2018), the HP-MCDA reaction was conducted in a 25-µl amplification mixture, containing 0.4 µM each of displacement primer (F1 and F2), 0.8 µM each of amplification primer (C1, C2, R1, R2, D1^#, and D2), 1.2 µM each of cross primer (CP1and CP2), 12.5 µl 2× reaction mixture, 1.5 µl VDR, 1 µl Bst 2.0 DNA polymerase, 1 µl DNA template from pure culture or 5 µl from clinical specimens, and moderate distilled water (DW) to a final volume of 25 µl. The MCDA reaction mixtures were incubated at 66°C for 40 min followed by 105°C for 5 min to stop the amplification. Genomic DNA of H. pylori was used as positive control, that of Listeria innocua (a resident of food, which may transiently exist in the stomach) as negative controls, and DW as blank control. The MCDA reaction results were determined according to the dynamic curve of real-time turbidity and the color change of the final reaction tube. Typically, turbidity value higher than 0.1 or a light blue color of reaction mixture indicated a positive result, while that lower than 0.1 or colorless implied a negative one.

For biosensor detection of the products of HP-MCDA reaction, LFB strip was utilized. In detail, an aliquot of 1 μl MCDA product was added to the sample pad region of LFB ( Figure 1B , II, step 2), followed by addition of an aliquot of 100 μl running buffer (10 mM PBS, pH 7.4 with 1% Tween 20) to the same region ( Figure 1B , II, step 2). After 5 min, the result was indicated by the presence or absence of red band in the NC membrane region ( Figure 1B , III), with red bands occurring in both the testing line (TL) and control line (CL) as positive result and only a red band in CL as negative.

The performance of the HP-MCDA-LFB assay may be influenced by the MCDA reaction temperature, reaction time, and the volume of MCDA products for LFB detection. The optimal reaction temperature of HP-MCDA reaction was determined by performing the reactions at temperatures of 60°C–67°C at 1°C intervals for 40 min. The optimal reaction time was examined by implementing MCDA reactions for 10–40 min (with 10 min interval), respectively. The optimal volume for LFB test was evaluated by adding varied volumes (0.25 µl, 0.5 µl, 1 µl, 1.5 µl, and 2.0 µl) of HP-MCDA products to the sample region of LFB strip and reporting the result within the same period of time.

Tenfold serial dilution of the genomic DNA of H. pylori (600 pg μl⁻¹ to 60 fg μl⁻¹) was prepared to determine the analytical sensitivity and limit of detection (LoD) of the MCDA-LFB assay for H. pylori detection. In parallel, MCDA products of the serial dilutions of genomic DNA of H. pylori was visually reported with VDR. Each test was implemented with three replicates.

The specificity of the HP-MCDA-LFB assay was evaluated using genomic DNA templates from 29 bacterial pathogens. After amplified under the conditions mentioned above, the obtained products were interpreted by both VDR and LFB detection formats. Each sample was tested independently at least three times.

In order to further validate the feasibility of the HP-MCDA-LFB technology in clinical H. pylori detection, gastric antrum mucosa samples from patients suspected of H. pylori infection at Children’s Hospital Capital Institute of Pediatrics were collected. These samples were detected by RUT method previously, a common rapid diagnostic method of H. pylori via detecting the activity of bacterial enzyme urease. Detail information of all the samples was listed in Table 3. Genomic DNA extracted from the rest of these samples was deployed to HP-MCDA-LFB test. Results obtained through both methods were compared to confirm the clinical application potential of the HP-MCDA-LFB assay.

The principle of the MCDA-LFB method for H. pylori detection is shown in Figure 1A . Briefly, the MCDA reaction was carried out as previously described, which used only a DNA polymerase with strand-replacement activity (Bst 2.0) and a set of 10 primers specifically spanning 10 different regions of the target sequence to achieve exponential amplification of the analytes. The MCDA assay was initiated by the cross primer CP1 annealing the target sequence, then replaced by the newly synthesized strand primed with F1, leading to release of the strand primed with CP1, which was employed as templates of subsequent amplification. Within a short time, plenty of strands complementary to the primer D1^# were produced, resulting in substantial double strands primed with D1^# generated. Particularly, biotin-labeled cytosine deoxynucleotide (biotin-14-dCTP) was additional added in the reaction mixture, leading to biotin labeled in the amplification products. As a result, the D1^# primed products contained both FAM and biotin labels and were considered as the detectable targets, which could be captured and visually inspected with LFB strips.

Details of biosensor detection of HP-MCDA products with LFB are shown in Figure 1B . The LFB strip was adapted for use by using anti-fluorophore antibodies (anti-FAM) to capture FAM-containing amplicons and using streptavidin-coated dyed (crimson red) polymer nanoparticles (SA-GNPs) to bind biotin-containing amplicons for readout visualization. As previously described (Wang et al., 2018; Zhou et al., 2023), SA-GNPs were deposited on the conjugate pad of the LFB strip; anti-FAM and biotinylated bovine serum albumin (biotin-BSA) were on the TL and CL regions of NC membrane, respectively. If the target analyte was present, the double-labeled (FAM and biotin-containing) products will flow down the strip and eventually captured by anti-FAM and visualized via SA-GNPs with red bands occurring in both TL and CL regions, while the absence of target amplicons only appeared as red band only in CL region due to the interaction between biotin-BSA and SA-GNPs.

Primer set feasibility for the HP-MCDA-LFB assay was determined by performing H. pylori MCDA reaction and reporting the results with LFB strip along with real-time turbidimeter and VDR. As shown in Figure 2A , by real-time turbidimeter, it was observed that only the positive control reaction reached a turbidity value higher than 0.1 within a short time, while the negative and blank controls all displayed blunted curves within the whole incubation time; with VDR, the tube of positive control mixture showed light green in color while that of negative and blank controls were colorless ( Figure 2B ); on the LFB strips, positive control reaction products induced two red bands in the TL and CL region, simultaneously, while the others only exhibited a red band in the CL region ( Figure 2C ). Data from the three detection formats demonstrated the effectiveness of the selected primer and its feasibility in the HP-MCDA-LFB assay for H. pylori detection.

Efficiency of the HP-MCDA assay at temperatures 60°C–67°C were exhibited in Figure 3 . According to the kinetic curves, it was observed that reactions at 66°C allowed for the most abundant products and the fastest speed to reach a turbidity of 0.1. Thus, 66°C was considered as the optimal reaction temperature of HP-MCDA reaction. Additionally, after amplifying LoD level of genomic DNA of H. pylori (60 fg) at 66°C for 10 min, 20 min, 30 min, and 40 min, respectively, the products were detected in formats of VDR and LFB. As shown in Figure 4 , the shortest time required to detect LoD levels of genomic DNA of H. pylori was 40 min. As a result, the whole detection procedure for HP diagnosis, including rapid template extraction (15 min), MCDA reaction (40 min), and result reporting using LFB (5 min), could be finished within 1 h ( Figure 1D ). Then, the optimal volume of products for LFB detection was determined by loading 0.25, 0.5, 1, 1.5, and 2 µl of amplification products to the sample pad region of LFB strip, respectively. As displayed in Figure 5 , the LFB strips loading 0.5–2 µl of product all showed two red bands at CL and TL regions within 5 min, while the one with 0.25 μl products showed only one red band at CL. Considering the potential of aerosol pollution, 0.5 μl of HP-MCDA product was accepted for LBF detection. In summary, the optimal reaction conditions, including a reaction temperature of 66°C, a reaction time of 40 min, and a product loading volume of 0.5 μl, were employed for the subsequent analysis.

Analytical sensitivity of the HP-MCDA-LFB was determined using the serial dilutions of genomic DNA templates of H. pylori. After testing each titer with three replicates, it was revealed that the detection limit of HP-MCDA-LFB assay was 60 fg μl⁻¹, which was demonstrated by LFB strip exhibiting two red bands in the TL and CL regions when loaded with products from templates not less than 60 fg ( Figure 6C ). Similarly, only the reaction with at least 60 fg templates displayed a sharp curve by real-time turbidimeter ( Figure 6A ) and showed light green in color ( Figure 6B ), revealing an identical LoD level to HP-MCDA-LFB assay. For the purpose of specificity analysis, 29 non-Helicobacter pathogens were detected. As envisaged, all the LFB strips loaded with MCDA products from non-Helicobacter pathogens showed red band only in the CL regions, while the positive control displayed two red bands in the TL and CL regions of LFB strip ( Figure 7 ), indicating a 100% specificity of HP-MCDA-LFB assay in H. pylori detection.

To further validate the feasibility of HP-MCDA-LFB as a diagnostic tool for H. pylori in clinical settings, we examined 20 endoscopically suspected of H. pylori infected gastric sinus tissues and 20 gastric sinuses with normal endoscopic mucosa using both HP-MCDA-LFB and RUT. As a result, 17 H. pylori-positive samples diagnosed by RUT were tested positive for H. pylori by HP-MCDA-LFB, and the 23 URT H. pylori-negative samples were negative in the HP-MCDA-LFB assay as well ( Figure 8 ), demonstrating high degree of agreement between the two methods. The results implied that the HP-MCDA-LFB method was an accurate and sensitive screening tool for H. pylori infection and indicated great potential to be widely used in clinical laboratories and basic research.