The study protocol was approved by the Medical Ethics Committee of the Second People’s Hospital of Lianyungang City (Lianyungang, China; approval no: 2020005). The clinical strains were isolated from sputum and nasal swab samples collected from 2020 to 2022. All patients agreed the use of the samples in this study and completed a written consent form.

Two strains of H. influenzae (non-capsulated, American Type Culture Collection ATCC 49247; and capsulated, ATCC 9334) were purchased from Shanghai Covey Chemical Technology Co., Ltd. (Shanghai, China). The specificity of the RPA-LFS assay was investigated based on the detection of the omp6 gene of 20 H. influenzae isolates from sputum (ten non-capsulated and ten capsulated) along with 23 other common pathogenic bacteria provided by our laboratory (i.e., Acinetobacter calcoaceticus, Acinetobacter lwoffi, Acinetobacter haemolytius, Acinetobacter junii, Acinetobacter johnsonii, Candida albicans, Enterobacter cloacae, Enterococcus faecium, Escherichia coli O157, Mycobacterium tuberculosis H37Ra, Pseudomonas aeruginosa, Staphylococcus aureus, Staphylococcus capitis, Staphylococcus epidermidis, Staphylococcus haemolyticus, Staphylococcus hominis, Staphylococcus saprophyticus, Staphylococcus warneri, Stenotrophomonas maltophilia, Streptococcus pneumonia, Viridans streptococci, Klebsiella pneumoniae, and Acinetobacter baumannii; Table 1 ). In total, 209 samples (sputum or nasal swab) were collected from patients with suspected H. influenzae infection in four hospitals located in Lianyungang and other cities (i.e., Lianyungang Second People’s Hospital, Sichuan Provincial People’s Hospital, Xuzhou Central Hospital, The Second People’s Hospital of Huai’an). All bacterial samples were incubated for 10 min at 100°C and, if not otherwise specified, 1 µL of the heat-treated culture at 10⁵ CFU/mL was used as the template.

Two RPA primers were designed based on the sequences of the omp6 and bexA genes with Primer Premier 5.0 software (Premier Biosoft, Palo Alto, CA, USA). The minimum and maximum product sizes of the primers were set at 100 and 300 bp, while the minimum and maximum primer sizes were 30 and 35 bp, respectively. Primers with sequence pairing of more than three consecutive bases (and more than one base at the 3ʹ end) were abandoned. The sequences of the primers and probes were confirmed for species specificity using the Primer-Basic Local Alignment Search Tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/).

RPA reactions were performed using the TwistAmp^® Liquid DNA Amplification Kit (TwistDx Inc., Maidenhead, UK) in accordance with the manufacturer’s instructions. Each 50-µL reaction contained 25 μL of 2× reaction buffer, 5 μL of 10× Basic e-mix, 2.5 μL of 20× core mix, 2.4 μL of 10 μM forward primer, 2.4 μL of 10 μM reverse primer, and 9.2 μL of distilled water. In addition, 2.5 µL of 280 mM magnesium acetate and 1 μL of the template were added to the lid of the reaction tube. After brief centrifugation, the reaction mixture was incubated for 30 min at 37°C. The RPA amplification products were purified using a PCR Cleaning Kit (Shanghai Meiji Biotechnology Co., Ltd., Shanghai, China) and separated by electrophoresis on a 2% agarose gel.

Primer Premier 5 software was used to design specific probes between the forward and reverse primer targeting sequences of the omp6 and bexA genes. The formation of dimers, hairpin structures, and mismatches between the probe and reverse primer were theoretically avoided as much as possible by adhering to the following parameters: probe size, 46–51 bp; GC content, 20%–80%; melting temperature, 57–80°C; maximum hairpin score, 9; maximum primer-dimer score, 9; maximum poly-X, 5; and other parameters, default values. In addition, the 5’ end of the probe was labeled with FITC, the 3’ end was blocked with SpC3, the middle base of the probe was replaced with tetrahydrofuran (THF) with at least 30 bp before the THF site and 15 bp after, and the 5’ end of the reverse primer was labeled with biotin.

The RPA reactions were conducted using the TwistAmp^® DNA Amplification nfo Kit (TwistDx). Each reaction mixture consisted of 29.5 μL of rehydration buffer, 2.1 μL of 10 μM forward primer, 2.1 μL of 10 μM reverse primer, 0.6 μL of 10 μM probe, and 12.2 μL of distilled water. To initiate the reaction, 1 μL of the template and 2.5 μL of 280 mM magnesium acetate were added to the mixture. After brief centrifugation, the reaction mixture was incubated for 5–35 min at 30–45°C.

Owing to the remarkable sensitivity of colloidal gold test strips, only a small amount of product is needed for detection, thus appropriate dilutions might be required. Only 2 µL of the amplification products were used for LFS detection (Ustar Biotechnologies Ltd., Hangzhou, China). The amplification products were added to the sample pad of the LFS, while the stick of the LFS was inserted into 100 μL of the sample buffer (Ustar Biotech) for 2 min prior to visualizing the results.

Standard capsulated (ATCC 9334, type b) and non-capsulated H. influenzae were inactivated by 10-fold serial dilution with ddH₂O (10⁶–10⁰) for use as templates for the RPA-LFS assay. To determine whether contamination with other strains would interfere with detection sensitivity, 10⁵ CFU/μL of heat-treated S. pneumoniae were added to 10-fold dilutions of heat-treated capsulated and non-capsulated H. influenzae culture (10⁵–10⁰ CFU/μL), respectively.

Primers for the dual PCR assay based on the omp6 and bexA genes were used as the control group in this experiment ( Table 1 ). Each 50-μL reaction contained 2.5 M deoxynucleoside triphosphates, 250 nM upstream and downstream primers for the omp6 and bexA genes, 2.5 U of Ex-Taq DNA polymerase (TaKaRa Biotechnology Co., Ltd., Dalian, China), 2 μL of the template, and 1 × PCR buffer (pH = 8.3). The PCR cycling conditions consisted of an initial denaturation step at 94°C for 4 min, followed by 35 cycles at 94°C for 30 s, 54°C for 30 s, and 72°C for 45 s, and a final extension step at 72°C for 5 min. The amplification products were separated by electrophoresis on a 2% agarose gel that was prestained with 1 μg/mL of ethidium bromide.

The practical application of the RPA-LFS assay was verified by comparisons to the results of the dual PCR assay established by Zhao et al. with 209 clinical specimens.

In order to distinguish the capsulated from non-capsulated H. influenzae, two pairs of primers were designed based on the conserved regions of the omp6 and bexA genes ( Table 1 ). As shown in Figure 1A , each primer pair amplified the corresponding target strain, and except for the target band, there were no non-specific bands or primer dimers. However, as indicated by the gel map, the bands amplified with the primer pairs omp6-F1/R1 and bexA-F2/R2 were notably brighter, indicating greater amplification efficiency. Therefore, the primer pairs omp6-F1/R1 and bexA-F2/R2 were selected for further experimentation.

The introduction of probes into the RPA system not only improves the specificity and sensitivity of the reaction, but also reduces the generation of primer-dimers. In order to increase specificity and sensitivity, corresponding probes were designed based on the target regions of the primer pairs omp6-F1/R1 and bexA-F2/R2. Although the introduction of probes can reduce the generation of primer dimers, false-positive signals are unavoidable due to the formation of dimers between the probe and reverse primer with stable amplification ability. The formation of dimers formed between the probe and reverse primer was detected using Primer Premier 5 software. As shown in Figure 1B , each probe and the corresponding reverse primer have a certain complementary fragment and some cover the THF site or expose the reverse 3’-OH end of the reverse primer. These dimers can be stably amplified, resulting in false-positive signals. RPA can tolerate a certain extent of base mismatch with no effect on the amplification efficiency (Daher et al., 2016; Liu et al., 2019). Nonetheless, many previous studies successfully established RPA-LFS assays (Yang et al., 2020; Yang et al., 2021b). In order to avoid false-positive results, mismatched bases were introduced to the probe and forward and reverse primers to reduce the generation of dimers between the probe and reverse primer. The reason for introducing mismatches is to ensure that the probe and reverse primer have fewer than three consecutive complementary bases, the complementary region of the probe and primer does not cover the THF site, the 3’ end of the reverse primer and probe does not have complementary pairing of more than three bases, and A-G and T-C are preferentially used interchangeably. The introduction of mismatches resulted in suitable probes and reverse primers. The primer sequences are listed in Table 1 . As indicated by the experimental results presented in Figures 1C, D , the test line of the experimental group had an obvious red band, but not test line of the no template control group, indicating the absence of false-positive signals.

To confirm that the primer pair bexA-F3/R1B/P only amplifies capsulated H. influenzae and that the primer pair omp6-F3/R1B/P only amplifies non-capsulated H. influenzae, specific primers corresponding to the two genes were used to amplify 10 non-capsulated and 10 capsulated H. influenzae isolates (verified by the traditional culture method). As shown in Figures 2A, B , omp6-F3/R1B/P detected all non-capsulated H. influenzae, while bexA-F3/R1B/P detected only capsulated H. influenzae, which confirmed the specificity of the primer pairs. In addition, to confirm that omp6-F3/R1B/P was specific to H. influenzae, 23 pathogens were selected for interspecies specificity verification ( Table 2 ). The results showed that only H. influenzae produced red bands at both the test and control lines, while the other 23 pathogens produced red bands only at the control line ( Figure 3 ).

To verify the detection limit of the RPA-LFS assay, boiled cultures of the capsulated and non-capsulated H. influenzae were serially diluted. Briefly, 1 µL of diluted culture was added to each 50-mL reaction volume for a final concentration of ranging from 10⁶ to 10⁰ CFU/µL. As shown by the results presented in Figures 4A, C , the lowest detection limit for the primer pairs omp6-F3/R1B/P and bexA-F3/R1B/P was 10⁰ CFU/reaction, indicating that the RPA-LFS assay was more sensitive than the PCR assay. The addition of S. pneumoniae to different concentrations of H. influenzae had no effect on the sensitivity of the RPA-LFS assay, as the lowest detection limit remained at 10⁰ CFU ( Figures 4B, D ).

The practical application of the RPA-LFS assay was verified using 209 clinical H. influenzae samples. The results showed that 203 of the samples were correctly identified as H. influenzae, yielding a detection rate of 97.1%, which was consistent with the detection results of the dual PCR assay and traditional culture methods. In order to assess the specificity for capsulated H. influenzae, all samples were screened with the primer pair bexA-F3/R1B/P and the results were compared with those of the dual PCR assay. The results showed that 128 capsulated H. influenzae isolates were detected by both the RPA-LFS and dual PCR assays, yielding a detection rate of 63.2% ( Table 3 ), demonstrating that the specificity of the RPA-LFS assay was equal to those of the dual PCR assay and traditional culture method.