Cas13a was purified as referred to in previous studies (East-Seletsky et al., 2016, 2017; Su et al., 2020). Briefly, the Cas13a gene in plasmid pC019-LwCas13a from Leptotrichia wadei (Addgene, United States) was subcloned into the psmarti vector (xhoI restriction site), and the correct plasmid was transformed into BL21(DE3) (General Biosystem, China). The expression of the target protein was induced at different temperatures (15 and 37°C) and different concentrations of IPTG (0.2 and 1.0 mM) (Amresco, United States). The Cas13a protein was purified by Ni-NTA (Smart-Lifesciences, China) using the 6 × His Tag antibody and horseradish peroxidase conjugate (Invitrogen, United States). After the addition of SUMO Protease (General Biosystem, China) to remove the fusion SUMO label, and the purified target protein was dialyzed into protein buffer [50 mM Tris, pH 7.5, 600 mM NaCl, 5% (vol/vol) glycerol, and 2 mM DTT].

Salmonella and other diarrheal pathogenic bacteria (Staphylococcus aureus, Listeria monocytogenes, Enterococcus faecalis, Shigella flexneri, Vibrio parahemolyticus, Escherichia coli O157, Yersinia enterocolitica, Campylobacter jejuni, Vibrio vulnificus) were provided by Jiangsu Provincial Center for Disease Control and Prevention. A total of 84 clinical stool samples that were selected for clinical validation of the method were collected from different hospitals in Jiangsu Province in 2019. The nucleic acid was extracted by EX-DNA Bacterial Genomic Extraction Kit (Tianlong, China) on a fully automatic nucleic acid extractor (Tianlong, China).

Salmonella-specific invA gene was retrieved from GenBank, and the conservative sequence region was selected to design RPA amplification primers using Primer 5 software, and Primer-BLAST software was used to verify the specificity of the primer sequence. The designed primers, probes, and oligonucleotides were synthesized by Sangon Biotech (Shanghai, China). The oligonucleotides containing T7 promoter, repeat, and spacer sequences were annealed with a T7 primer. Then, the crRNA was synthesized by incubating at 42°C for 2 h with T7 RNA polymerase (TaKaRa, China). The synthesized crRNA was digested with DNase I (TaKaRa, China) at 37°C for 1 h followed by purification with RNA rapid concentration and purification kit (Sangon Biotech, China) according to the manufacturer’s protocol. The concentration of crRNA was quantified using Qubit 2.0 (Invitrogen, United States). All nucleic acid sequences used in this study are shown in Table 1.

The amplicon of the RPA reaction was used as the template for transcription in vitro with T7 RNA polymerase, and the transcription product was purified to obtain the Salmonella target RNA. Then, the target RNAs were diluted gradiently (10⁰–10⁵ pM). The synthesized oligonucleotide Salmonella-RPA-R was used as a non-target control (500 ng/μL). Verification of CRISPR-Cas13a cutting activity was completed as follows: 5 × reaction buffer 5 μL (final concentration 20 mM HEPES, 60 mM NaCl, 6 mM MgCl₂, pH 6.8), 25 mM NTP mix 1 μL (ATP, GTP, CTP, UTP, Bio Lab, China), Recombinant RNase Inhibitor (TaKaRa, China) 0.5 μL, 1.2 μM crRNA 0.6 μL, 10 μM RNA-probe 0.5 μL, 1 μM Cas13a 1 μL, or 10³ pM target RNA or non-target control 2 μL, adding RNase-free ddH₂O to 25 μL. The reaction was conducted at 39°C for 30 min, and RNase A was used as the positive control.

In the two-step RPA-Cas13a assay, basic RPA reactions were first conducted according to the instructions of the TwistAmp Basic Kit (Twist Dx, Cambridge, United Kingdom). Each RPA reaction was carried out in a 50 μL reaction volume containing Primer Free Rehydration buffer 29.5 μL, 0.24–0.96 μM forward and reverse primers each 0.5 μL, target DNA template 2 μL, and RNase-free ddH₂O 15 μL. The reaction mixes were vortexed and spun briefly, and then 280 mM magnesium acetate (MgOAc) 2.5 μL was added and mixed well to start a reaction in fluorescence detector F1620 (Qitian, China) at 37, 39, and 41°C for 20 min. Then, a 25 μL CRISPR-Cas13a reaction system was performed with 5 × reaction buffer 5 μL, 25 mM NTP mix 1 μL, Recombinant RNase Inhibitor 0.5 μL, 1.2 μM Salmonella-crRNA 0.6 μL, 10 μM RNA-probe 0.5 μL, T7 RNA Polymerase 0.3 μL, 1 μM Cas13a 1 μL, and added RNase-free ddH₂O to 23 μL, and RPA products 2 μL. The fluorescent signals were collected at 39°C for 30 min in fluorescence detector F1620.

The one-tube RPA-Cas13a assay combined RPA with CRISPR-Cas13a in a one-tube reaction system. Briefly, the 50 μL one-tube reaction system consisted of Primer Free Rehydration buffer 29.5 μL, 25 mM NTP mix 4 μL, Recombinant RNase Inhibitor 4 μL, T7 RNA Polymerase 1 μL, 0.24–0.96 μM primers each 0.5 μL, 1.2 μM crRNA 1 μL, 10 μM RNA-probe 1 μL, 1 μM Cas13a 2 μL, target DNA template 2 μL, and RNase-free ddH₂O 1 μL. The reaction mixes were vortexed and spun briefly, and then 280 mM MgOAc 2.5 μL was added to the tube cap and was centrifuged into the reaction solution. Fluorescent signals were collected at 37, 39, and 41°C for 2 h in fluorescence detector F1620.

To evaluate the minimum detection limit, serial dsDNA standards were prepared as follows: using the extracted nucleic acid of Salmonella as a template, a 50 μL LA Taq PCR (TaKaRa, China) containing LA Taq 0.5 μL, 2.5 μM dNTP 4 μL, 10 × buffer 5 μL, 10 μM forward/reverse primer each 1 μL, and RNase-free ddH₂O 37.5 μL was performed. The PCR amplification product was purified by High Pure PCR Product Purification Kit (Roche, CH). The dsDNA quantification was performed in a Qubit digital fluorimeter using the Qubit dsDNA BR Assay (Invitrogen, Waltham, United States). The dsDNA copy number was then determined using the following formula: {[6.02 × 10¹⁴ × dsDNA concentration (ng/μL) × 10^–9]}/[DNA in length × 660]. The gradiently diluted dsDNA standards (10⁷∼10^–1 copies/μL) were detected by the one-tube and two-step RPA-Cas13a. We also chose real-time PCR (Hein et al., 2006) (TaKaRa, China) as the reference test. The specificity of the one-tube and two-step RPA-Cas13a was tested by using nucleic acid extracted from diarrheal pathogenic bacteria strains: Salmonella, S. aureus, L. monocytogenes, E. faecalis, S. flexneri, V. parahemolyticus, E. coli O157, Y. enterocolitica, C. jejuni, and V. vulnificus.

Real-time PCR was performed on the LightCycler 2.0 instrument in which premix Ex Taq^TM (Takara, China) was used (Hein et al., 2006). The reaction was performed as follows: 95°C for 30 s, followed by 40 cycles of 95°C for 5 s and then 60°C for 20 s.

Genomic DNAs from 84 clinical stool samples were selected for a clinical comparison experiment. All samples were tested with the two-step, one-tube RPA-Cas13a, and real-time PCR assay (Hein et al., 2006). The consistency of the three methods above was compared using SPSS 24.0 software. κ-tests were used for consistency analysis.

A significant fluorescent signal in the reactions of the positive control within was verified with RNase A. Compared with the detection of Cas13a for target RNA and non-target RNA control, the results showed that Cas13a could achieve specific detection for target RNA of Salmonella in the presence of its corresponding crRNA. Furthermore, there was no detection signal generated for Cas12a compared to Cas13a. The detection signal of CRISPR-Cas13a for 10³ pM was almost identical to that of the positive control (Figure 1). The results of different dilutions for RNA standards showed that the minimum test limit was 10 pM (Figure 2).

We first tested the best RPA amplification temperature using 0.48 μM primer concentration. The results showed the highest fluorescent signals were obtained at 39°C in all three Cas13a detecting time points (Figure 3A). Therefore, 39°C was selected as the best RPA amplification temperature. Then, the best primer concentrations were explored at 39°C. The optimal primer concentration of 0.48 μM was found (Figure 3B). Therefore, we chose 0.48 μM and 39°C as the best RPA primer concentration and amplification temperature.

We first tested the different temperature (37, 39, 41°C) in a one-tube RPA-Cas13a detection system and found that 39°C had the highest fluorescent signals. The detection signal grew stronger from 10 to 90 min and reached its peak at 90 min, whereas the negative control had no visible detection signal (Figure 4A). We then explored the optimal primer concentration under 39°C and found the highest fluorescent signals were obtained at 0.48 μM primer concentrations. Therefore, the optimized conditions of one-tube RPA-Cas13a were 39°C with 0.48 μM primer concentration (Figure 4B).

As shown in Figure 5A, the limit of detection of one-tube RPA-Cas13a for Salmonella was 10² copies/μL, which showed the same sensitivity to real-time PCR (10² copies/μL, Figure 6) and slightly lower than that of two-step RPA-Cas13a (10⁰ copies/μL, Figure 5B). The whole detection time of one-tube RPA-Cas13a for 10² copies/μL was only 20 min (Figure 5A), which was shorter than those for two-step RPA-Cas13a (RPA: 20 min, transfer time: 5 min, and CRISPR-Cas13a: 20 min, Figure 5B) and real-time PCR (60 min or much longer, Figure 6). For specificity assessment, whether one-tube RPA-Cas13a at 60 min (Figure 7A) or two-step RPA-Cas13a at 30 min (Figure 7B), only the Salmonella showed detection signals among the 10 common diarrheal bacteria, and no cross-reactions were found.

For evaluating the applicability of developed one-tube and two-step RPA-Cas13a for Salmonella in clinical stool samples, a total of 84 clinical stool samples from different hospitals were selected and tested by real-time PCR, two-step RPA-Cas13a, and one-tube RPA-Cas13a. Sixteen samples were detected positive for Salmonella by real-time PCR. Besides those 16 samples, three additional samples were judged as positive by two-step RPA-Cas13a in which one was also detected positive by one-tube RPA-Cas13a. However, the signal intensity was significantly lower than that of the other 16 positive samples but significantly higher than that of the negative control sample. At the same time, those three samples were also identified by Sanger sequencing technology and determined to be positive for Salmonella. The two-step RPA-Cas13a detection results are in good consistency with real-time PCR (κ = 0.892), the one-tube RPA-Cas13a detection results are also in good consistency with real-time PCR (κ = 0.962), and the statistical results are shown in Table 2.