The engineered CRISPR‐Mb2Cas12a variant enables sensitive and fast nucleic acid‐based pathogens diagnostics in the field

Summary

Existing CRISPR/Cas12a‐based diagnostic platforms offer accurate and vigorous monitoring of nucleic acid targets, but have the potential to be further optimized for more efficient detection. Here, we profiled 16 Cas12a orthologs, focusing on their trans‐cleavage activity and their potential as diagnostic enzymes. We observed the Mb2Cas12a has more robust trans‐cleavage activity than other orthologs, especially at lower temperatures. An engineered Mb2Cas12a‐RRVRR variant presented robust trans‐cleavage activity and looser PAM constraints. Moreover, we found the existing one‐pot assay, which simultaneously performed Recombinase Polymerase Amplification (RPA) and Cas12a reaction in one system, resulted in the loss of single‐base discrimination during diagnosis. Therefore, we designed a reaction vessel that physically separated the RPA and Cas12a steps while maintaining a closed system. This isolated but closed system made diagnostics more sensitive and specific and effectively prevented contamination. This shelved Mb2Cas12a‐RRVRR variant‐mediated assay detected various targets in less than 15 min and exhibited equal or greater sensitivity than qPCR when detecting bacterial pathogens, plant RNA viruses and genetically modified crops. Overall, our findings further improved the efficiency of the current CRISPR‐based diagnostic system and undoubtedly have great potential for highly sensitive and specific detection of multiple sample types.

Introduction

Advances in genetic and pathogen detection promote crop health because they facilitate diagnosis and monitoring of pathogens and reduce unnecessary management such as chemical treatments. Currently, there are many methods for detecting pathogens in crops, classified broadly into conventional (e.g., culture‐based method) and modern molecular techniques (Nnachi et al., 2022), as well as integrated, automated diagnostic biosensors developed for disease monitoring in the field (Dincer et al., 2019). Traditional methods to identify pathogens require culture and determination of morphological, physiological, chemical and biochemical characteristics, which are time‐consuming and labor‐intensive (Buszewski et al., 2021). Molecular tests, such as PCR/qPCR‐based and next‐generation sequencing‐based diagnostics, often require a specialized laboratory and trained technicians, or expensive, precision instruments (Tian et al., 2022). Biosensors or microfluidic sensors have greater advantages since they are fast, sensitive and require less labour to detect bacterial pathogens, but such sophisticated sensors are often costly to design and their sensitivity is easily limited by the nature of the sample matrices (Nnachi et al., 2022). To overcome these shortcomings, the next‐generation diagnosis systems should be sensitive, convenient, easy to use, low‐cost and not require special equipment or training. These features also meet the needs of quarantine authorities for timely detection.

Diagnostic techniques based on isothermal amplification‐based assays, such as recombinase polymerase amplification (RPA) or loop‐mediated isothermal amplification (LAMP), have become mainstream for the rapid detection of pathogens due to features of simple, user‐friendly and rapid (Ghosh et al., 2018; Shin et al., 2018). These isothermal amplification techniques only require a simple, constant‐temperature device and can even rapidly amplify target sequences at room temperature (Mayboroda et al., 2018). However, isothermal amplification may still yield false positives through non‐specific amplification (Zhao et al., 2015). Encouragingly, the integration of isothermal amplification with a CRISPR‐Cas (Clustered Regularly Interspaced Short Palindromic Repeats‐CRISPR‐associated proteins) reaction not only diminishes false positives, but also further strengthens the sensitivity and specificity of the diagnostic assay (Wang et al., 2020b). CRISPR‐based detection relies primarily on the indiscriminate cleavage of neighbouring nucleic acids upon recognition of the desired target by a Cas protein, such as Cas12, Cas13 and Cas14 (now named as Cas12f) (Chen et al., 2018; Gootenberg et al., 2017; Harrington et al., 2018). An integrated CRISPR platform detects a target in four steps, i.e., target pre‐amplification, target recognition by Cas‐crRNA complex, Cas protein activation and reporter cleavage (Yin et al., 2021). The Cas protein is selected according to the detection target. For example, Cas12a can be activated by dsDNA/ssDNA targets and showed scramble activity on ssDNA substrates (Chen et al., 2018; Li et al., 2018a), while Cas13 can be activated by ssRNA targets and cleave ssRNA substrates (Myhrvold et al., 2018). By programming the Cas‐crRNA to recognize desired sequences, various diagnostic systems such as SHERLOCK (Kellner et al., 2019) and DETECTR (Broughton et al., 2020) have been developed and widely used in the detection of viruses (Chaijarasphong et al., 2019), pathogens (Zhang et al., 2020) and other small molecules (Liang et al., 2019). Cas12a protein has been widely used and engineered for developing diagnostic systems because its reaction system is relatively simple and stable compared to those of Cas13.

Although the existing CRISPR/Cas12a‐based detection systems are sensitive (attoM level) and show high specificity for SNP differentiation when combined with isothermal amplification (Chen et al., 2018; Gootenberg et al., 2018; Li et al., 2018b), the methods can still be further improved. The early CRISPR diagnostic platforms took typically 30 min or more (Tian et al., 2022; Wang et al., 2019; Zhou et al., 2020). This waiting time needs to be shortened for on‐site application. Additionally, the Cas protein requires thermostatic heating to effectively amplify the target and maintain its activity, which may be inconvenient for field application. The other limitation is that current Cas12a orthologs prefer the TTTV (V = A, C, or G) protospacer adjacent motif (PAM) and showed very low activity in recognizing other non‐canonical PAMs, which limits the targets of the crRNAs (Kleinstiver et al., 2019). To date, only a limited number of Cas12a proteins have been evaluated and utilized in the development of detection systems. However, Cas proteins derived from different microbial strains possess distinct recognition efficiency and cleavage activity (Li et al., 2018b). Therefore, screening novel Cas12a orthologs is an important step in establishing more effective CRISPR/Cas assays that can target relaxed PAMs and that can further speed up the detection process, especially at lower temperatures.

To break through the above limitations, 16 Cas12a orthologs were first evaluated in the present study, especially for their trans‐cleavage activity at lower temperatures. We found that the Mb2Cas12a protein, derived from Moraxella bovoculi strain 57 922, exhibited stronger activity than other Cas12a‐family proteins currently employed. Furthermore, we introduced five mutations into the Mb2Cas12a protein to create an improved Mb2Cas12a variant (termed Mb2Cas12a‐RRVRR) that has a broader PAMs range and great activity. Next, we systematically optimized and designed a diagnostics system with the Mb2Cas12a‐RRVRR variant for the detection of multiple sample types. This engineered Mb2Cas12a‐RRVRR variant shortened the CRISPR‐based detection time to 15 min, and the sensitivity was equal to or greater than that of qPCR assay, when detecting the bacterial pathogens, RNA virus in plant and genetically modified crops. Taken together, the novel detection platform based on CRISPR/Mb2Cas12a‐RRVRR variant enhanced the diagnostic capabilities of CRISPR‐based technology and has the potential to replace the commonly used Cas12a protein in the existing detection platform.

Results

Survey of trans‐cleavage activity for 16 Cas12a orthologs

To improve Cas12a‐based nucleic acid detection, we selected 16 proteins from the Cas12a‐family, including three well‐characterized Cas12a nucleases (FnCas12a, AsCas12a and LbCas12a), and 13 Cas12a orthologs that have not yet been evaluated for trans‐cleavage activity (Figure 1a and Table S1). These Cas12a proteins from different bacteria have low homology with each other at the amino acid level. However, as reported (Zetsche et al., 2020), the direct repeat (DR) sequences of mature crRNAs associated with these Cas12a orthologs display a high level of homology (Figure S1A) and form nearly identical secondary structures (Figure S1B). Based on the differential sequences between the left and right stems of the mature crRNAs, we divided the crRNA into 9 groups (crRNA1‐9). To better compare the properties of Cas12a orthologs, we purified these proteins with the same prokaryotic expression system after codon optimization.

Figure 1.

Comparison of protein domains, crRNA sequences and trans‐cleavage activities of 16 orthologs in the Cas12a‐family. (a) Phylogenetic analysis of 16 Cas12a orthologs. The protein sequences of each Cas12a ortholog were retrieved from previous studies and presented in Table S1. The multiple alignments were generated with protein sequences in the software Geneious Prime with MUSCLE. (b) Cas12a orthologs evaluated for in vitro trans‐cleavage activity with four dsDNA targets. The left diagram presents the dsDNA targets containing either the canonical PAM site (TTTA/G) or no PAMs. The sequences of the four dsDNA targets are shown in Table S2. Fluorescence values were measured continuously over 30 min, and given in Figure S2. The assays were performed three times at 37 °C with 300 nM Cas12a, 360 nM crRNA, 400 nM ssDNA FQ‐reporter and 500 nM dsDNA targets. As shown on the right, the fluorescence values at 15 min of reaction were taken out separately to create a heat map for comparing the trans‐cleavage activity between Cas12a proteins. Fluorescence intensities are shown as means from three independent reactions.

All 16 Cas12a proteins had been identified to possess a T‐rich PAM preference (Chen et al., 2020; Jacobsen et al., 2020; Wierson et al., 2019; Zetsche et al., 2015). We next evaluated their trans‐cleavage activity in vitro against four dsDNA targets. Among them, two were designed with canonical PAM sites (TTTA/G), while others lacked the canonical PAM. By monitoring the fluorescence intensity during the reaction, the 16 Cas12a proteins were found to cleave non‐targeted ssDNA, although there were large differences in activities (Figure 1b). For all Cas12a proteins, the presence of TTTA/G PAM in the dsDNA targets induced their trans‐cleavage activity, and almost no trans‐cleavage was observed in the presence of dsDNA targets without canonical PAMs. Nine Cas12a orthologs (As, Er, Lb, Mb2, MI, Fn, Ts, Px and Lp) completely degraded FQ‐labelled ssDNA reporters within 30 min and exhibited sufficient cleavage activity in vitro (Figure S2). Interestingly, the Mb2Cas12a derived from Moraxella bovoculi strain 57 922 showed the highest trans‐cleavage activity among the Cas12a orthologs, and was able to recognize the GCTC PAM variant (Figure 1b).

The effect of temperature on the in vitro cleavage efficiency of the 16 Cas12a proteins was evaluated. The proteins exhibited various levels of cleavage efficiency under different temperatures, ranging from 24 to 48 °C, and the optimal reaction temperature for these proteins was found to be around 40–44 °C (Figure S3). This temperature range is slightly higher than 37 °C commonly used, which is consistent with reports showing higher trans‐cleavage activity of LbCas12a protein at reaction temperatures around 42 °C (Jiao et al., 2022; Ning et al., 2020). A reaction temperature below 30 °C significantly inhibited cleavage activity, but Mb2Cas12a exhibited temperature‐insensitive trans‐cleavage activity and functioned rather robustly even at 24 and 28 °C in vitro (Figure 2a,b). When targeting the NTTN PAMs, Mb2Cas12a showed more enzymatic activity than LbCas12a at a relatively low reaction temperature (Figure 2c). Therefore, we speculated that Mb2Cas12a might be a superior alternative nuclease for Cas12a‐based detection.

Figure 2.

Mb2Cas12a exhibited temperature‐insensitive trans‐cleavage activity compared to other Cas12a orthologs. (a) Heat map depicting the effect of reaction temperature on the in vitro cleavage efficiency of 16 Cas12a proteins. Assays were performed at 37 °C with 300 nM of each Cas12a ortholog, 360 nM of the corresponding crRNA, 400 nM ssDNA FQ‐reporter, and 500 nM dsDNA T3 targets (shown in Figure 1b and Table S2). The results are also presented as a histogram in Figure S3. Fluorescence values were taken 15 min after reaction initiation. (b) Fluorescence time courses obtained from Mb2Cas12a and LbCas12a reactions at low temperature. The reaction conditions were the same as above, except that the reaction temperature was changed. Fluorescence measurements were taken every minute. NTC, non‐targets control reaction performed at 32 °C. (c) Mb2Cas12a showed higher enzymatic activity than LbCas12a at low reaction temperatures when targeting the NTTN PAM sequence (Table S2). The reaction conditions were the same as above, and fluorescence values were taken after 15 min of reaction time. All fluorescence intensities are shown as means ± SD from three independent reactions.

An engineered Mb2Cas12a variant enhanced trans‐cleavage activity and broadened PAM sequence preference

In LbCas12a, mutation of five sites (D156R, G532R, K538V, Y542R and K595R) significantly enhanced its enzymatic activity and extended the PAM range (Jiao et al., 2022). Four of the residues (G532R, K538V, Y542R and K595R) are present in the WED or PI structural domain of Cas12a and might interact with the PAM (Yamano et al., 2017). The D156R mutation was shown to improve the efficiency of gene editing in plants and animals (Kleinstiver et al., 2019; Tóth et al., 2020). Inspired by this, we next attempted to introduce these five mutations into Mb2Cas12a protein to further enhance the protein's ability. The corresponding five sites were identified in the Mb2Cas12a protein by sequence comparison (Figure 3a). After mutation, prokaryotic expression, and purification, the Mb2Cas12a‐RRVRR variant (D172R, N563R, K569V, N573R and K625R) was obtained.

Figure 3.

The Mb2Cas12a‐RRVRR variant possesses enhanced trans‐cleavage activity. (a) The sequence alignment (left) and mutation sites (right) in two domains of LbCas12a and Mb2Cas12a. The original and mutated amino acid sites are indicated in yellow and red, respectively. WED, wedge; REC, recognition; PI, protospacer adjacent motif (PAM)‐interacting; BH, bridge helix; Nuc, nuclease. Both the WED and RuvC structural domains are composed of three discontinuous protein sequences. (b) Fluorescence intensities of trans‐cleavage reactions at temperatures ranging from 24 to 32 °C. All assays were performed 6 times with 300 nM of each Cas12a protein, 360 nM of the corresponding crRNA, 400 nM ssDNA FQ‐reporter, and 500 nm dsDNA T3 target (Table S2). For comparison, the endpoint fluorescence values were taken 15 min after the reaction was started. Data are shown as means ± SD from six independent reactions. Asterisks indicate a significant difference in trans‐cleavage activity between the two proteins with two‐tailed Student's t‐test (**P < 0.01, *** P < 0.001). Numbers above the asterisk represent the ratio of two fluorescence values. NTC, non‐targets control reaction performed at 32 °C. (c) Enzyme activities of Cas12a‐catalysed trans‐cleavage reaction. The solid lines are the fit to the Michaelis–Menten equation. Calculated values of activity parameters are given in the table below the figure. Data are presented as the mean ± SD (n = 3).

Within the temperature range of 24–32 °C, both Mb2Cas12a‐RRVRR and LbCas2a‐RRVRR variants showed stronger activity than their wild‐type, but to varying degrees. Mb2Cas12a‐RRVRR exhibited stronger activity after 15 min of reaction, an approximately 1.27‐ to 2.23‐fold increase in activity compared to the LbCas2a‐RRVRR (Figure 3b). A Michaelis–Menten kinetics analysis was also performed to evaluate the difference in enzymatic activity between the different proteins at 28 °C. The catalytic efficiency of Mb2Cas12a‐RRVRR, as expressed by the K _cat/K _m value, was 1.18 × 10⁹ M⁻¹ s⁻¹, which was a 1.34‐fold increase compared to wild‐type Mb2Cas12a and a 1.40‐fold increase compared to LbCas2a‐RRVRR (Figure 3c). These results indicated that mutation of these amino acids in Mb2Cas12a could significantly enhance the trans‐cleavage activity in vitro. While this result is similar to the results observed in LbCas12a (Jiao et al., 2022), the overall activity does seem to be higher for Mb2Cas12a‐RRVRR.

To assess whether the mutated Mb2Cas12a‐RRVRR variant can recognize dsDNA targets with relaxed PAMs, we measured the trans‐cleavage activity with an in vitro PAM identification assay. We synthesized 256 pairs of base‐complementary oligonucleotides (47 bp), and obtained 256 dsDNA targets by annealing. These dsDNA substrates shared the same target spacer but different PAMs, constituted by random combinations of four base (Figure 4a and Table S4). By monitoring the fluorescence intensity in the reaction system, we found that the wild‐type Mb2Cas12a presented a high tolerance for variation at the 4th base, and can reliably target sites with NTTV PAMs. Moreover, Mb2Cas12a tolerated a C base at the 2nd or 3rd position of the PAMs and displayed relatively high activity at NTCV and NCTV PAMs (Figure 4b). Compared to wild‐type Mb2Cas12a protein, the RRVRR variant of Mb2Cas12a showed enhanced trans‐cleavage activity at the NTTV, NTCV and NCTV PAMs sites (Figure 4c and Figure S4). Furthermore, Mb2Cas12a‐RRVRR showed less restrictive PAM recognition, showing activity, albeit a lower overall activity, against NCCV.

Figure 4.

The in vitro trans‐cleavage activity of Mb2Cas12a and its RRVRR variant when targeting the dsDNA substrates with all 4‐base‐pair variants of the PAM. (a) Schematic design of in vitro PAM identification assay. The PAM library includes 256 dsDNA fragments (47 bp) consisting of 4 randomized nucleotides and was synthesized from complementary‐ssDNA by annealing. Only one of the dsDNA substrates was used as a target for each analysis. Heat maps depicting fluorescence intensity of wild‐type Mb2Cas12a (b) and its RRVRR variant (c) detecting dsDNA substrates with all variations at the 4‐base pair PAM using the fluorescence‐based reporter assay. The colour intensity in the heat maps represents the trans‐cleavage activity. All assays were performed 6 times at 37 °C with 300 nM Mb2Cas12a or Mb2Cas12a‐RRVRR, 360 nm crRNA, 400 nM ssDNA FQ‐reporter and 500 nM dsDNA targets (also seen in Table S2). For comparison, the endpoint fluorescence values were taken 15 min after the reaction was started.

Several CRISPR/Cas12 systems are highly sensitive to mismatches between the crRNA and DNA targets (Chen et al., 2018; Teng et al., 2019). To investigate whether the introduction of five mutations into the Mb2Cas12a protein would impact its specificity, we introduced mismatch into the dsDNA substrates sequentially (Figure S5A). Results demonstrated that a single‐base mismatch at the crRNA‐substrate interface, particularly in the PAM‐proximal region, significantly inhibited the trans‐cleavage activity of LbCas12a, Mb2Cas12a and Mb2Cas12a‐RRVRR. Nevertheless, the mutant Mb2Cas12a‐RRVRR exhibited a stronger fluorescence signal than the wild‐type protein and was more tolerant to a single‐base mismatch(Figure S5B). Furthermore, two continuous or random mismatches in the crRNA‐target pairing region resulted in a further reduction of the trans‐cleavage activity of both proteins. Nevertheless, the Mb2Cas12a‐RRVRR maintained high specificity even in the presence of more than two mutant sites in the target (Figure S5C). While the specificity of the mutant Mb2Cas12a‐RRVRR was slightly reduced, its ability to tolerate mismatches would increase its compatibility for detecting targets with mutant loci.

Compatibility of Mb2Cas12a‐RRVRR variant with RPA reaction system

To improve assay sensitivity, most Cas12a‐based detection systems require isothermal amplification of the target prior to trans‐cleavage reaction (Broughton et al., 2020; Chen et al., 2018; Patchsung et al., 2020). The process can be one‐step (amplification and Cas12a detection performed simultaneously; Figure S6A) (Ding et al., 2020) or two‐step (isothermal amplification followed by Cas12a detection; Figure S6B). The two‐step process has proven to be more sensitive, but one‐step method simplifies the operation and reduces the risk of contamination (Lu et al., 2022). To test Mb2Cas12a‐RRVRR compatibility with the two approaches, we synthesized a dsDNA sequence as the detection target, including a pair of RPA primers and four targeting sites for crRNA (Figure 5a). In the same amplification region, the four crRNAs are different and flanked by distinct PAM sequences, including canonical TTTC (crRNA2^TTTC) and ATTG (crRNA3^ATTG), as well as non‐canonical TGGT (crRNA1^TGGT) and CTAG (crRNA4^CTAG).

Figure 5.

Evaluating the compatibility of Mb2Cas12a‐RRVRR variant with RPA pre‐amplification. (a) Schematic diagram of RPA primers and crRNA‐targeting regions in a fragment of dsDNA sequences. The crRNA1^TGGT, 3^ATTG were designed to bind the anti‐sense strand and limited by the TGGT and ATTG PAMs, while crRNA 2^TTTC, 4^CTAG were designed to bind the positive‐sense strand and limited by the TTTC and CTAG PAMs. (b) Time courses of the fluorescence intensity in two‐step (top) and one‐step (bottom) RPA‐Cas12a assays. The primers and crRNAs are given in Table S2. (c) Mutated spacer sequences in crRNAs. The single‐base mismatch in each crRNA is highlighted in red. The targeting sites for each crRNA are given in (a). The numbers in the left column indicate the mismatch locations. Trans‐cleavage activity of Mb2Cas12a‐RRVRR/crRNA complexes with synthetic mismatches in two‐step (d) and one‐step (e) assays. The reactions were performed based on the process in (b). For comparison, the endpoint fluorescence values were taken 30 min after the reaction was started. All fluorescence intensities are shown as means ± SD from six independent reactions. The black asterisks indicate a significant difference using two‐tailed Student's t‐test (**** P < 0.0001; n. s., not significant). The detection fluorescence values were calculated by subtracting the background fluorescence value of non‐targets control reaction.

As expected, the crRNA2^TTTC and crRNA3^ATTG with canonical PAMs showed robust fluorescence responses, while the crRNA1^TGGT and crRNA4^CTAG with non‐canonical PAMs did not trigger the nonspecific cleavage of Mb2Cas12a‐RRVRR in the two‐step assay (the top in Figure 5b). In the one‐step assay, all four crRNAs activated the collateral cleavage of Mb2Cas12a‐RRVRR, but they showed lower fluorescence responses than the two‐step assay. Moreover, canonical PAMs of TTTC and ATTG showed slower kinetics of nonspecific cleavage with reduced fluorescence signals in the one‐pot reaction (the bottom in Figure 5b). The result is consistent with a previous report showing that a one‐step reaction could overcome the limitations of PAM for designing the crRNA (Ding et al., 2020). Because canonical PAMs could be better at activating Mb2Cas12a‐RRVRR activity, the lower activity of crRNA2^TTTC and crRNA3^ATTG in the one‐step assay is likely attributable to the fact that the dsDNA templates derived from RPA amplification are rapidly cleaved by highly activated Mb2Cas12a‐RRVRR, thereby decreasing the amplification and trans‐cleavage efficiency (Lu et al., 2022).

Next, we examined the mismatch tolerance of the Mb2Cas12a‐RRVRR variant in both two‐step and one‐step RPA/Cas12a assays. Synthetic mismatches were introduced into the spacer sequences of each crRNA, which enables Cas12a to discriminate between targets that differ by a single‐base mismatch (Figure 5c). The same dsDNA fragment shown in Figure 5b was used as target substrates. In the two‐step assay, the Mb2Cas12a‐RRVRR variant was able to detect single‐base differences. When single‐base mutations were introduced into the 1st, 3rd, 5th positions of the spacer sequences in crRNA2^TTTC and crRNA3^ATTG, the enzyme activity of Cas12a was significantly declined (Figure 5d). However, in the one‐step assay, all of the mutated crRNAs activated strong trans‐cleavage activity of the Mb2Cas12a‐RRVRR variant showing similar fluorescence values in different reactions (Figure 5e). We initially speculated that the RRVRR variant might affect Mb2Cas12a specificity. However, when Mb2Cas12a‐RRVRR was replaced with LbCas12a, AsCas12a, FnCas12a and Mb2Cas12a in the one‐step system, we observed that these wild‐type Cas12a orthologs (Lb, As, Fn and Mb2) were also unable to distinguish single‐base mutations (Figure S7). Furthermore, continuous mismatches were also introduced into the spacer sequences of crRNA1^TGGT and crRNA2^TTTC, and we found the sequential mutations of more than 6 bases were required to substantially reduce the trans‐activity of Mb2Cas12a‐RRVRR and LbCas12a in one‐step RPA/Cas12a assays (Figures S8 and S9). In one‐step CRISPR assay (Figure S6A), the simultaneous RPA reaction could help to open up the dsDNA strands and expose the binding sites of the Cas12a‐crRNA complexes through strand displacement. This can lead to the formation of single‐stranded DNA at the crRNA‐targeting sites, which facilitates the binding of the crRNA to the target DNA and activates Cas12a. However, it should be noted that MbCas12a, like LbCas12a (Chen et al., 2018), is unable to distinguish single‐base differences in the recognition of ssDNA targets (as shown in Figure S10). As a result, when the RPA amplification and Cas12a reaction are mixed and initiated simultaneously in the same tube, Cas12a‐based assays may lose their ability to discriminate single‐base variations. Overall, RPA amplification allows simultaneous Cas12a reaction to break through the restriction of the PAM, but it also reduced the detection specificity.

Strategies to enhance the sensitivity of RPA/Mb2Cas12a‐RRVRR reaction

Since the Mb2Cas12a‐RRVRR protein displayed more activity than other Cas12a orthologs, we then wondered if it has a better performance in terms of assay sensitivity. In the absence of RPA amplification, we found that the Mb2cas12a‐RRVRR showed higher sensitivity than other Cas12a orthologs (Figure S11A). Furthermore, the PAM sequence of the dsDNA target significantly affected the sensitivity of the Mb2Cas12a‐RRVRR detection. When the target harboured NTTV, NTCV, NCTV or NCCV PAMs, the sensitivity reached 2.4 × 10⁸ to 2.4 × 10⁹ copies after 10 min of reaction (Figure S11B). Overall, the sensitivity of the detection system was higher when the dsDNA could better activate the trans‐cleavage activity of the Mb2Cas12a‐RRVRR system.

When there are fewer targets, RPA pre‐amplification could amplify the number of targets, thus significantly improving the detection sensitivity of the subsequent Cas12a reaction. We evaluated the reaction temperature and found that different temperatures affected the amplification efficiency of RPA. RPA amplified the dsDNA target efficiently only when the reaction temperature was at 30–50 °C (Figure 6a). Therefore, the RPA reaction needs to be performed with the aid of a heating device, as the target cannot be amplified at room temperatures of around 24–26 °C. To make the assay process simpler and less expensive, we explored the feasibility of performing the RPA reaction through body heat instead of an additional heating device. By monitoring the temperature of a tube incubated in the axilla outside the clothing (Figure 6b), we found the reaction solution in the tube reached about 35 °C within 3 min (Figure 6c), which is the temperature required for RPA reactions to amplify 24 copies of DNA targets to detectable levels (Figure 6d). Lower reaction temperatures also severely affected the sensitivity of the subsequent Cas12a assay, but the Mb2Cas12a‐RRVRR performed significantly better than its wild‐type or LbCas12a proteins (Figure S12A,B). The dsDNA target was efficiently amplified at 35 °C, and with the powerful activity of Mb2Cas12a‐RRVRR, the detection sensitivity of RPA/ Mb2Cas12a‐RRVRR assay at this temperature reached 2.4 copies per reaction (Figure S12B). These datapoints suggest that the use of the axilla as an incubation site allows for high sensitivity of the RPA/Mb2Cas12a‐RRVRR assay.

Figure 6.

Optimizing equipment‐free incubation for Mb2Cas12a‐RRVRR reaction. (a) Temperature range for effective RPA reaction. The RPA reaction was performed with TwistAmp Basic kit (TwistDx) and 2 μL synthetic DNA fragments (1 ng/μL) according to the manufacturer's instructions. After incubation for 10 min at different temperatures, 10 μL of RPA products were run on 2% agarose gel electrophoresis. The synthetic DNA fragments and RPA primers were consistent with those used in Figure 5a and Table S2. (b) The reactions were secured in the axilla using an elastic sweatband. The arrow indicates the approximate position of the tubes, which are covered with the sweatband. (c) The temperature curve of reaction mixture incubated in the axilla. Each plot shows the temperature traces of reactions incubated by 5 volunteers. The reaction was simulated by a 0.5 mL microcentrifuge tube filled with 50 μL water. The temperature inside the tube was monitored by a temperature detector with a puncture probe (Yusheng DT1311, China). (d) The sensitivity of RPA reactions incubated in the axilla. Each of serially diluted dsDNA targets were subjected to RPA amplification. After a 10‐min incubation in the axilla, 10 μL of RPA products were run on 2% agarose gel electrophoresis. The synthetic DNA fragments and RPA primers were consistent with those used in Figure 5a and Table S2. (e) Workflow of a one‐tube assay, which improved the sensitivity of Mb2Cas12a‐RRVRR detection by separating the RPA and Cas12a reactions, with detection by a lateral flow assay. The tube has an extra shelf or ledge manufactured into its internal surface. The parameters of the tube for 3D printing are provided in supplementary data. The entire system has been named the Shelved Mb2^RRVRR assay. (f) The detection sensitivity reached a single‐molecule level within 15 min when detecting targets harbouring TTTC PAMs. The reactions were performed as described in Figure 6e. One strip from each set of three replicates was shown. NTC, non‐targets control reaction.

Next, we screened other parameters (buffer, mixing ratio, incubation time and concentration) to improve the Mb2Cas12a‐RRVRR protein trans‐cleavage activity and shorten the reaction time at 35 °C. We found that commercial rCutSmart buffer obtained from New England Biolabs (NEB) was more suitable for Mb2Cas12a‐RRVRR and allowed complete degradation of the fluorescent probe within 5 min (Figure S13A). The fluorescence signal could reach the peak at 3 min when the concentrations of Mb2Cas12a‐RRVRR and crRNA were 600 nm and 720 nm, respectively, for detecting the targets harbouring TTTC PAMs (Figure S13B). In the above assay procedure, it was necessary to open the tube's cap after the RPA reaction to transfer a small amount of RPA products into another tube to minimize changes in the Cas12a system. However, opening the lid to transfer the RPA product not only increases the handling steps, but also does not take full advantage of the RPA product. We optimized the mixing ratio of RPA product and Cas12a system and found that the subsequent fluorescence signal was further enhanced when 20 μL of RPA product was added to 40 μL of Cas12a system (Figure S13C). The RPA product, pre‐amplified for 10 min, could activate the Mb2Cas12a‐RRVRR protein to cleave more reporter molecules within 5 min (Figure S13D).

To separate the RPA and Mb2Cas12a reactions, we designed a new reaction vessel with a small shelf or ledge that can be 3D printed. This one‐tube detection using the ledge to separate the 20‐μL RPA system from the 40‐μL Cas12a solution into two separate locations within the tube (Figure 6e). After completion of the first reaction (the RPA), the two systems are subsequently mixed by inversion of the tube. As shown in Figure 6e, the ledge tube is first incubated in the axilla for 10 min to initiate the RPA reaction. The tube is turned upside down and shaken to mix the RPA and Cas12a systems and then placed back into the axilla to initiate the Cas12a reaction. After incubation for another 3–5 min, the detection of the target is evaluated by measuring the fluorescence value or by lateral flow assay (Figure S14). We aimed to visualize the results using lateral flow test strips, because it is difficult to distinguish the nature and intensity of fluorescence using UV light to excite FQ‐reporter molecules, especially under different ambient light conditions in the field. The entire assay took <15 min and possessed a detection sensitivity at the single‐molecule level when detecting targets harbouring NTTV, NCCV, NCTV and NTCV PAMs (Figure 6f, Figures S13E and S15). Moreover, we lyophilized the RPA and Cas components in the tube and found that the re‐activated reagents performed well and were stabile for up to 20 days when stored at ambient temperature (Figure S13F). We call this prepared, layered reaction with the Mb2Cas12a‐RRVRR variant and a flow assay read‐out a “Shelved Mb2^RRVRR” assay.

Diagnostic applications for bacterial and viral pathogens and genetically modified organisms

Screening Cas12a orthologs and their variants showed that the Mb2Cas12a‐RRVRR variant showed robust trans‐cleavage activity and less restrictive PAM requirements, which make it a suitable candidate for nucleic acid detection compared to other Cas12a proteins. We further developed the one‐tube Shelved Mb2^RRVRR assay with enhanced speed, sensitivity and reproducibility. Next, we wondered if these approaches would be universally effective in detecting viral and bacterial pathogens as well as genetically modified organisms.

The spread of diseases caused by pathogens is one of the primary issues in crop production worldwide (Nelson and Bone, 2015). The bacterial pathogens Erwinia amylovora and Acidovorax citrulli, due to their devastating reductions of crops in the Rosaceae and Cucurbitaceae families, respectively, have been prioritized for quarantine around the world to prevent cross‐border transmission through seeds, seedlings and other plant tissues (Liebhold et al., 2012). Apple necrotic mosaic virus (ApNMV) is a single‐stranded, positive‐sense RNA virus that was identified in apple trees with mosaic disease (Noda et al., 2017). With the help of rapid nucleic acid preparation techniques for field samples, the one‐tube Shelved Mb2^RRVRR assay had a limit of detection (LOD) for bacterial pathogens E. amylovora and A. citrulli of 10 CFU/reaction (Figure 7a). For comparison, qPCR assay showed a LOD of about 2.4–2.7 × 10² CFU/reaction for the two bacteria (Figure S16A,B), suggesting that the sensitivity of qPCR is approximately 20 times lower than that of Shelved Mb2^RRVRR assay. For diluted in vitro RNA transcript, the LOD of Shelved Mb2^RRVRR assay for ApNMV was 2.2 × 10² viral copies per reaction, probably due to the efficiency of reverse transcription when detecting the RNA target. The LOD of RT‐qPCR for ApNMV was the same as the Shelved Mb2^RRVRR assay (Figure S16C).

Figure 7.

Sensitivity and validity of one‐tube Shelved Mb2^RRVRR assays when detecting bacterial and viral pathogens and genetically modified organisms. (a) Sensitivity of Shelved Mb2^RRVRR assays. Three types of targets, including bacterial pathogens, RNA viruses, and transgenic events, were used as detection substrates for sensitivity evaluation. All reactions were incubated in the axilla and performed as described in Figure 6e using the Shelved Mb2^RRVRR assays. NTC, non‐targets control reaction. (b) Identification of A. citrulli, ApNMV and transgenic events from field samples. In the top panel, the fluorescent signals were shown as means ± SD from six independent reactions, with reaction times of 10 min for RPA and 3 min for the Mb2Cas12a‐RRVRR reaction. The lateral flow strip in the middle indicates the visualized readout after the Shelved Mb2^RRVRR assay. One strip from each set of three replicates is shown. The qPCR or RT‐qPCR results are shown in the bottom. The Cq values represent average values from three independent experiments.

To investigate the sensitivity for detecting the CaMV‐35S promoter in genetically modified soybean, genomic DNA obtained from the GM soybean was serially diluted with that of non‐GM soybean. For the Shelved Mb2RRVRR assay, the LOD was estimated by identifying the lowest dilution ratio that produced a visible band on the lateral flow strip. Both the Shelved Mb2^RRVRR and qPCR displayed LODs of 0.01% (w/w) (Figure 7A and Figure S16D), but the Shelved Mb2^RRVRR assay does not require fluorescence monitoring under laboratory conditions, and the entire process took only 15 min.

In the final verification, we gathered some commercial watermelon seeds, apple leaf tissues and GM materials to determine the effectiveness of the Shelved Mb2^RRVRR assay for detecting pathogens, viruses and transgenic events, respectively (Figure 7b). The Shelved Mb2^RRVRR detected 4 A. citrulli‐positive seed samples, and 13 ApNMV‐ positive leaf samples, which were consistent with qPCR or RT‐qPCR assay. The GM maize samples MON810, MON863, Bt‐176 and soybean samples GTS40‐3‐2 and A2704‐12 showed positive results with the Shelved Mb2^RRVRR assay, while the other samples displayed negative results. These results were also consistent with the qPCR assay, indicating an excellent diagnostic agreement between the two methods. Overall, the Shelved Mb2^RRVRR system containing the Mb2Cas12a‐RRVRR variant could detect multi‐type nucleic acid targets in a very short time with high specificity and sensitivity compared to the existing detection platform.

Discussion

Nucleic acid testing is essential for pathogen detection and is extensively used for disease control, field diagnosis, biosafety and environmental monitoring. The current PCR‐based technology is less suitable for rapid diagnosis, due to the long detection time and requirements for a laboratory environment, technical personnel and specialized hardware equipment (Tian et al., 2022). A CRISPR/Cas nuclease‐based diagnostic system can be easy to use, fast, highly sensitive and specific and exhibits great promise for the next generation of point‐of‐care (POC) diagnostics. The robust detection system developed in this study has the following advantages over existing Cas12a‐based methods (Chen et al., 2018; Ding et al., 2020; Joung et al., 2020; Ooi et al., 2021): (1) the Mb2Cas12a‐RRVRR variant maintain sufficient exonuclease activity at low temperatures (Figure 3), allowing diagnostics to be completed within 15 min without additional heating devices (Figure 6); (2) this variant also has a broader PAM sequence preference (Figure 4), facilitating the design of crRNA targeting sites; (3) a reaction vessel was designed to physically separate the RPA and Cas assays in a closed system, making the diagnosis more sensitive and specific as well as effectively preventing contamination (Figure 6); and (4) the Shelved Mb2^RRVRR assay was highly sensitive and specific detection for multiple sample types (Figure 7). These advantages make it more suitable for on‐site detection, providing a better alternative to existing diagnostics platforms (Table S10).

The discovery of novel Cas protein variants with special properties could enrich the CRISPR toolbox and further improve the CRISPR/Cas‐based diagnostic systems. The LbCas12a^D156R variant was proven to have improved gene editing efficiency (Schindele and Puchta, 2020) as well as trans‐cleavage capability that further increased the detection sensitivity (Nguyen et al., 2022). As for AsCas12a, variants carrying the mutations S542R/K607R, S542R/K548V/N552R and E174R/S542R/K548R, were successively demonstrated to possess higher gene editing efficiency (Gao et al., 2017), looser PAM constraints (Kleinstiver et al., 2019) and the enhanced trans‐cleavage activity that was used for detecting SARS‐CoV‐2 (Ooi et al., 2021). Subsequently, these mutations were introduced into the corresponding sites in LbCas12a, and the resulting variant extended the PAM range to TNTN (Tóth et al., 2020) and exhibited superior properties over the wild‐type in pathogen detection (Jiao et al., 2022). In screening the Cas12a‐family protein, Mb2Cas12a and its engineered Mb2Cas12a‐RVRR variant (N563R, K569V, N573R and K625R), had also been shown to exhibit higher cis‐cleavage efficiency and a relaxed PAM requirement (Zhang et al., 2021). However, the use of the trans‐cleavage activity for nucleic acid detection has not been fully elucidated for Mb2Cas12a. Considering the correlation between cis‐ and trans‐cleavage abilities of the Cas12a protein, it is reasonable to speculate that Mb2Cas12a has excellent properties for development as a CRISPR/Cas‐based diagnostic tool. On this basis, the present study evaluated the trans‐cleavage activity of 16 Cas12a orthologs and observed that Mb2Cas12a showed robust enzymatic activity (Figure 1), which confirmed the speculation above. Moreover, analogous mutations of all five sites (D172R, N563R, K569V, N573R and K625R) were introduced into Mb2Cas12a as well, generating the Mb2Cas12a‐RRVRR variant, which further improved the testing capability. We believe that it is feasible to replace the dominant LbCas12a protein with the Mb2Cas12a‐RRVRR variant, thereby improving the detection efficiency of the CRISPR/Cas‐based diagnostic system.

Previous Cas12/Cas13‐based diagnostic systems used a discrete pre‐amplification step to amplify the target prior to the Cas‐mediated assay (Broughton et al., 2020; Chen et al., 2018; Patchsung et al., 2020), which unavoidably complicated the process and introduced the risk of cross‐contamination. To overcome this drawback, different studies have integrated the components of RPA or LAMP into the Cas system, allowing both systems to react in one tube (Ding et al., 2020; Joung et al., 2020). Although the development of one‐step CRISPR detection has become mainstream, there were still a few reports showing both cleavage of the amplicons and degradation of the primers in this blended system, caused by cis‐ and trans‐cleavage of Cas12a, respectively, that could hinder the continuous production of amplicons, resulting in detection failure (Gong et al., 2021). A recent study showed that if crRNA is designed to target a sequence with suboptimal PAMs, the cis‐cleavage efficiency of LbCas12a on dsDNA substrate was inhibited, thus allowing isothermal amplification to rapidly enrich enough substrates for activating the subsequent trans‐cleavage of Cas12a on ssDNA reporter, which greatly improved detection efficiency (Lu et al., 2022). However, the thus far reported one‐step assays did not evaluate their ability to differentiate from the single base pair. We found that Cas12a/crRNA complexes lost their single‐base discrimination in the currently prevalent one‐step system (Ding et al., 2020; Lu et al., 2022; Wang et al., 2019) (Figure 5 and Figure S7). Since the RPA reaction opens up and amplifies the dsDNA targets, single‐stranded DNA can be formed at the crRNA‐targeting sites, thus facilitating crRNA‐guided target DNA binding and activation of Cas12a. Previous study confirmed that the non‐specific cleavage of ssDNA reporter by the Cas12a protein is related to crRNA/target binding, but not to the cleavage of dsDNA target (Jeon et al., 2018). Moreover, MbCas12a was unable to distinguish single‐base differences in recognition of ssDNA targets (Figure S10), as reported for LbCas12a (Chen et al., 2018). That might explain the low specificity of the one‐step CRISPR assay. In view of our findings, we prefer to physically separate the components of Cas12a and RPA in one tube, and then mix the two systems after the RPA reaction is completed.

In previous studies, CRISPR/Cas12a‐based methods allowed for quantitative measurement of the target load by plotting standard curves and monitoring fluorescence signal intensity (Ding et al., 2021; Ning et al., 2020). We believe the Mb2Cas12a‐RRVRR variant could replace wild‐type proteins and also enable quantitative analysis. However, the reaction time, the concentrations of RPA primer and other components must be re‐optimized to obtain the best linear relationship between fluorescence values and targets concentration. In addition, fluorescence monitoring equipment is also necessary, but this undoubtedly brings inconvenience to field detection. When we tried to predict virus/ bacteria load in the sample, the crude DNA extracts seemed to cause interference, creating poor linearity. Furthermore, we observed that due to the high efficiency of CRISPR/Cas systems, it is difficult to quantify high concentrations of targets, which can easily saturate the detection signal. Therefore, most current diagnostic systems suitable for field applications can only perform qualitative analysis based on endpoint values.

The major challenges in the field of CRISPR‐based diagnosis are the time for sample preparation, the portability of reagents, the complexity of the procedure, and the read‐out time, as well as the sensitivity and specificity when used for multiple sample types. In this study, we optimized a CRISPR‐based nucleic acid detection platform with a newly engineered Mb2Cas12a variant. This variant, to our knowledge, is probably the most active compared to other Cas12a orthologs available. The Shelved Mb2^RRVRR assay system developed based on this variant could detect several types of samples with high sensitivity and specificity within 15 min, which is significantly better than other CRISPR‐based platforms. This is an on‐site testing method that does not require any heating equipment, and the one‐tube design reduces the risk of cross‐contamination. Although some functions still need to be improved, such as quantitative analysis, this robust and field‐based Shelved Mb2^RRVRR diagnostic method provides a powerful tool for enhancing the reliability and efficiency of pathogen detection.

Methods

Cas12a‐family protein expression and purification

The amino acid sequences of each Cas12a ortholog (Table S1) and the corresponding mature crRNA were obtained from previous reports (Chen et al., 2020; Jacobsen et al., 2020; Teng et al., 2019; Wierson et al., 2019; Zetsche et al., 2015; Zhang et al., 2021). The bacterial expression plasmids for AsCas12a and FnCas12a were constructed by PCR‐amplification of the plasmids from the previous construct (Addgene, Cat: 113430 and 113 432) and inserting them into the BamHI and XhoI sites of the pET‐N‐His‐TEV vector (Beyotime Biotechnology, China). The other Cas12a proteins and the Mb2Cas12a‐RRVRR variant (D172R, N563R, K569V, N573R and K625R) were codon‐optimized for E. coli using the codon optimization tool (GenSmart™ Codon Optimization of GenScript) and then were synthesized and also cloned into pET‐N‐His‐TEV vector. The vector for Mb2Cas12a‐RRVRR (D172R, N563R, K569V, N573R and K625R) was constructed by introducing five mutations into pET‐Mb2Cas12a. The bacterial expression vector of LbCas12a and its RRVRR variant was derived from our previous report (Jiao et al., 2022). Taken together, a total of 18 Cas12a proteins, including two variants, were expressed and purified by the preparation process as described in Supplementary experimental procedures (Data S1).

The evaluation of in vitro trans‐cleavage activity for Cas12a orthologs

The in vitro trans‐cleavage by the Cas12a orthologs was evaluated by detecting the fluorescence signal in 50 μL of reaction system, including 300 nM Cas12a, 360 nM of the corresponding crRNAs, 400 nM ssDNA FQ‐reporter (5’‐FAM‐TTATT‐Quencher‐ 3′, synthesized by Sangon Biotech), 20 U of RNase inhibitor, 500 nM dsDNA activator and 1 × NEBuffer 3.1. To ensure uniformity, all crRNAs used in this study were synthesized directly by Sangon Biotech (Shanghai, China) and diluted to 10 μM. Prior to the cleavage reaction, Cas12a and crRNA were mixed and incubated at 25 °C for 10 min, thus promoting the formation of crRNA/Cas12a complexes. The 47‐nt dsDNA activators were generated by annealing two complementary DNA oligonucleotides (10 μM for each, Table S2) in 1× annealing buffer (Solarbio, Beijing, China) according to the manufacturers' instructions. The Bio‐Tek FLx800 microplate fluorescence reader was used for fluorescence detection with excitation/emission at 485/520 nm, and different incubation temperatures and detection frequencies were set according to experimental needs.

When Michaelis–Menten analysis was used to compare the trans‐activity of Cas12a orthologs, the final concentrations were 10 nM Cas12a, 12 nM crRNA and 0.1 nM dsDNA activator in the reaction system. The various concentrations of ssDNA fluorescent reporter, ranging from 0, 10 nM, 100 nM, 200 nM, 400 nM, 600 nM, 800 nM, 1000 nM and 1500 nM, were subsequently added to the reactions as cleavage substrates. All reactions were incubated at 37 °C for 10 min and the fluorescence signal intensity was detected every 30 s. Finally, enzyme activity parameters such as K _m and K _cat were calculated using GraphPad Prism 8 software.

In vitro PAM identification assay

To evaluate the PAMs preference of the Cas12a proteins, we synthesized a PAMs library containing a total of 256 dsDNA sequences (Table S4). These sequences share the same protospacer, but the upstream PAM sequences consisted of 4 random bases. Each dsDNA target was obtained from two complementary oligonucleotides by annealing, then purified by Small Molecular DNA Gel Mini Purification Kit (Zoman Biotech, Beijing, China) and diluted to the same concentration. Each in vitro cleavage reaction was performed at 37 °C with 300 nM Cas12a, 360 nM of the corresponding crRNAs, 400 nM ssDNA FQ‐reporter, 20 U of RNase inhibitor, 500 nM of each dsDNA target and 1 × NEBuffer 3.1. For comparison, the endpoint fluorescence values were taken 15 min after the reaction was started.

Isolated one‐tube assays

The isolated reactions, one‐tube system was prepared separately, as reagents A and B, in a newly designed reaction vessel that can be easily made by 3D printing. The reaction vessel was modified from a common 0.5 mL tube by adding a shelf or ledge inside the tube and changing the lid to a spherical shape (Figure 6f; Supplementary data for 3D printing parameters, Data S2). Reagent A is the RPA component prepared using TwistAmp Liquid Basic Kit (TwistDx) contained 1 × reaction buffer, 1 × basic E‐mix, 1 × core reaction mix, 480 nM of each RPA primers, 1.8 mM dNTPs and 14 mM MgOAc. Reagent B is the Cas12a component prepared using ddH₂O, 900 nM Cas12a, 1080 nM crRNA, 1.5 × rCutSmart buffer, 1.5 U/μL RNase inhibitor and 600 nM ssDNA FQ‐reporter (5’‐FAM‐TTATT‐Quencher‐3′). For detecting the RNA targets, most components were the same as those in the two system above, except supplementing 10 U/μL HiScript II Reverse Transcriptase (Vazyme Biotech, China) in Reagent A of RPA component. To pre‐store the reagents in the reaction vessel, 20 μL of reagent A and 40 μL of reagent B were quickly deposited on the shelf and bottom of the vessel, respectively, and then lyophilized for 6 h using a benchtop freeze‐dry system (−50 °C and 0.1 mbar).

In a shelved‐one‐tube reaction, 2 μL of DNA mixed with 18 μ L ddH₂O was dripped onto the shelf of the tube to dissolve the dry powder of reagent A and initiate the RPA reaction. At the same time, 40 μL ddH₂O was dropped to the bottom of the tube to dissolve the lyophilized reagent B. After being incubated for 10 min in the armpit, the tube was manually turned upside down to allow the two parts to mix, thus initiating the Cas12a reaction. The endpoint fluorescence was the raw fluorescence determined by the Bio‐Tek FLx800 microplate fluorescence reader (λex: 485 nm; λem: 535 nm). For lateral flow visual detection, the FQ‐reporter was replaced with 6 μM of lateral flow cleavage reporter (5’‐FAM‐TTATTATT‐Biotin‐ 3′, synthesized from Sangon Biotech), and other components were the same as those in the shelved‐one‐tube system above. At the end of the reaction, 100 μL of assay buffer was added, and a lateral flow strip (Zoonbio Biotechnology Co., Ltd.; Nanjing, China) was inserted directly into the solution for 2 min. The presence/absence of both bands was scored.

Sample preparation for sensitivity assay and field detection

The strains of A. citrulli GDMCC 803619 and E. amylovora GDMCC 1.1352 were obtained from Guangdong Microbial Culture Collection Center in China (GDMCC, http://www.gdmcc.net). After activation, all strains were grown in LB medium and incubated at 37 °C for 16–24 h until OD = 1.0. The concentrations were determined by gradient dilution and plate counting. Subsequently, the bacterial solution was 10‐fold serially diluted to a range of 10⁷–10⁰ CFU/mL with sterile water. An NaOH‐based crude extraction was employed for rapid DNA extraction (Wang et al., 2020a). Briefly, the different concentrations of bacterial suspension (100 μL) were mixed with 200 μL of 0.5 m NaOH solution and shaken vigorously, followed by standing at room temperature for 1 min. The cell lysis was 10‐fold diluted with Tris–HCl buffer (20 mM, pH = 7.5), and the resulting solution was directly used as a template for qPCR or Shelved Mb2^RRVRR analysis. For the detection of seed‐transmitted pathogen A. citrulli, 50 seeds of each product were crushed with a mortar, immersed in 20 mL of sterile water (supplemented with 0.1% Tween 80), and vortexed gently for 10 min to promote the bacterial release. Then, 100 μL of the bacterial suspension was used for rapid DNA extraction as described above. The detailed information of RPA primers and crRNA‐targeting sequences were given in Table S8.

The in vitro RNA transcription of ApNMV and sensitivity assay, as well as the sample preparation for field detection of ApNMV in apple leaf, were performed as previous report (Jiao et al., 2021). A total of 16 leaf samples from commercial orchards located in different areas (Henan province, China) were collected for the qPCR or Shelved Mb2^RRVRR analysis.

The powders of genetically modified (GM) maize (event MON810, MON863, Bt‐176, GA21) and GM soybean (event GTS40‐3‐2, MON87705, A2704‐12) were purchased from Monsanto company (St. Louis, MO). Other non‐GM materials of primary cultivars were collected from local farms. For the detection of the CaMV‐35S promoter in GM crops, a NaOH‐based extraction‐free method was used for DNA preparation of plant tissues (Wang et al., 2020c). Briefly, 20–30 mg of leaf tissue or powder was homogenized by a hand‐held tissue homogenizer in a 2‐mL tube with 150 μL of lysis buffer (0.5 M NaOH, 10 mM Na₂EDTA, pH 8.0), then strongly shaken for 10 s. After incubation at room temperature for 1 min, the crude extract was diluted with 1.5 mL sterile water, and the supernatants were directly added into the qPCR or Shelved Mb2^RRVRR system as the DNA template. Moreover, DNA standard of GM soybean A2704‐12 was mixed with non‐GM maize DNA to different final concentrations of 10%, 1%, 0.1%, 0.01% and 0.001%, and the resulting mixtures were used for sensitivity assay.

Detection of targets by gold‐standard qPCR or qRT‐PCR.

The qPCR assay for detecting A. citrulli and E. amylovora were performed as previously reported (Jiao et al., 2022). For detecting the CaMV35 promoter in genetically modified materials with qPCR, the assay was performed with a Taq Pro HS Probe Master Mix (Vazyme, China) and cycling conditions of 30 s at 95 °C, followed by 45 cycles of 10 s at 95°C and 30 s at 60 °C, and a final 5 min at 40°C. Fluorescence information was collected during the 60 °C annealing step each cycle. The RT‐qPCR assays for detecting RNA virus of ApNMV were carried out using a HiScript II One‐Step qRT‐PCR Probe Kit (Vazyme, China) according to the manufacturer's protocol. All assays were performed using a LightCycler480 (Roche) and the primer sets were listed in Table S9. Sensitivity and linearity of the RT‐qPCR assays were estimated by constructing standard curves. In this study, results with Cq values greater than 35 were considered negative. The LOD was estimated by identifying the lowest concentration that produced a positive result.

For the detection of ApNMV in apple leaf, a modified alkaline polyethylene glycol (PEG)‐based extraction method was used for sample preparation (Silva et al., 2018). Briefly, a small amount of leaf tissue (approximately 20 mg) was taken into a centrifuge tube, and 300 μL of extraction buffer (6% PEG 200 and 20 mM NaOH) was added followed by crushing the tissue with a handheld electric tissue homogenizer. After the homogenate was incubated for 3 min at room temperature, the resulting crude extract was used directly for qRT‐PCR or Shelved Mb2^RRVRR analysis.

Author contributions

J. C. F. and X. B. Z. conceived the project and provided overall supervision. J. J. designed and performed the most experiments with the help of Y. Q. L., M. L. Y. and J. C. Z. Some conceptual advice was provided by W. X. Y. and C. H. L., and they also revised the draft manuscript. S. S. W. and T. H. B. analysed the most data. S. C. H. and M. M. W. contributed the testing materials and reviewed the manuscript. S. J. L. and R. W. performed the Shelved Mb2^RRVRR analysis for detecting A. citrulli. K. X. Z. and P. B. H. performed the Shelved Mb2^RRVRR analysis for detecting ApNMV and GMO. J.J. wrote the manuscript that was read and approved by all authors.

Conflict of interest

The authors declare no competing interests.

Supporting information

Figure S1 The conservation of direct repeat sequences within crRNAs.

Figure S2 Fluorescence curves of different Cas12a/crRNA complexes trans‐cleavage reactions with four dsDNA substrates.

Figure S3 Effect of different temperatures on trans‐cleavage activity of different Cas12a orthologs.

Figure S4 Mb2Cas12a and Mb2Cas12a‐RRVRR variant was tested for trans‐cleavage activity when recognizing dsDNA targets harbouring NTTN, NCCN, NCTN and NTCN.

Figure S5 The tolerance to single‐base and double‐base mismatches in the target sequence in trans‐cleavage reactions with different Cas protein.

Figure S6 Schematic of the RPA‐Cas12a detection systems, namely a one‐step process (A) and a two‐step process (B).

Figure S7 Trans‐cleavage activity of different wild‐type Cas12a proteins when used in two‐step (A) and one‐step (B) RPA/Cas12a assays.

Figure S8 Tolerance of Cas12a protein to continuous target mismatch when crRNA was designed close to the binding region of the RPA primer.

Figure S9 Tolerance of Cas12a protein to continuous target mismatch when crRNA was designed away from the binding region of the RPA primer.

Figure S10 The mismatch tolerance of Mb2Cas12a‐RRVRR‐based detection when recognizing the dsDNA and ssDNA targets.

Figure S11 The fluorescence intensity of Cas12a reactions when detecting serially diluted dsDNA targets.

Figure S12 Sensitivity of the RPA/Cas12a assay with different Cas12a proteins under the incubation at 30 °C (A) and 35 °C (B).

Figure S13 The detection sensitivity of the Cas12a system was improved by optimizing the reaction system at 35 °C.

Figure S14 Mb2Cas12a reaction was visualized with lateral‐flow test strip.

Figure S15 Comparison of the sensitivity of the one‐tube Shelved Mb2^RRVRR assay for targets with different PAMs.

Figure S16 Standard curve for TaqMan qPCR or RT‐qPCR detection of E. amylovora (A), A. citrulli (B), ApNMV (C) and CaMV‐35S promoter (D).

Table S1 Detailed information of Cas12a‐family proteins used in this study.

Table S2 Oligonucleotides and crRNAs used in Figure 1 and Figure S2.

Table S3 The oligonucleotides and crRNAs used in Figure 2, Figure 3 and Figure S3 .

Table S4 The oligonucleotides and crRNA used in Figure 4, Figure S4 and Figure S11.

Table S5 The dsDNA targets, primers and crRNAs used in Figure 5A‐B, Figure 6A, Figure 6D‐F, Figure S12, Figure S13.

Table S6 The dsDNA targets, primers and crRNAs used in Figure 5C‐D, Figure S5 and Figure S7.

Table S7 The dsDNA targets, primers and crRNAs used in Figure S15.

Table S8 Sequences of RPA primers, crRNAs and targeting domain for E. amylovora, A. citrulli, ApNMV and CaMV35S promoter, respectively.

Table S9 Primers and TaqMan‐MGB probes used in the one‐step RT‐qPCR and qPCR assay.

Table S10 Characteristics of the developed CRISPR/Cas12a‐based detection systems.

Data S1 Supplementary experimental procedures.

Data S2 Parameters for 3D printing.