A general onepot-method for nucleic acid detection with CRISPR-Cas12a

CRISPR/Cas12a system has been shown promising for nucleic acid diagnostics due to its rapid, portable and accurate features. In combination with isothermal amplication technology, single-copy sensitivity can be achieved. However, cleavage of the amplicons and primers by the cis- and trans-activity of Cas12a hinders the attempts to integrate the amplication and detection steps into a single reaction. Through phosphorothioate modication of primer and design of crRNA that allow for the cutting site locating at the modied site of the primer, we realized onepot detection of SARS-CoV-2 with single-copy sensitivity. We also identied the activated Cas12a has a much higher anity to C nucleotide-rich reporter than others. By applying such reporters, we signicantly reduced the reaction time required for the lateral-ow readout. Furthermore, to improve the specicity of the strip-based assay, we created a novel reporter and, when combined with a customized strip, the unspecic signal could be completely eliminated. This established system termed Targeting DNA by Cas12a-based Eye Sight Testing in Onepot Reaction (TESTOR) was validated using clinical cervical samples for human papillomaviruses (HPVs) detection. Our system represents a general approach to integrating the nucleic acid amplication and detection into a onepot reaction in CRISPR-Cas systems, highlighting its potential as a rapid, portable and accurate detection platform of nucleic acids.


Introduction
Recent outbreak of SARS-CoV-2 has highlighted the challenges of detecting viral infections, especially in areas where specialized equipment is not available [1][2][3] . Polymerase chain reaction (PCR) is the most commonly used method and has been considered as a "gold standard" for nucleic acid diagnosis due to high sensitivity and speci city [4][5][6] . However, the requirement of expensive equipment and well-trained personnel as well as long reaction times (normally more than 2 h) makes it unsuitable for point-of-care test (POCT) diagnostics. These limitations hinder its applications in many cases and whereby delay the prescription and administration of antiviral agents to patients. In contrast, serology tests are rapid and require minimal equipment but have lower sensitivity and speci city 7,8 . It may need several days to weeks following symptom onset for a patient to mount a detectable antibody response 9 .
CRISPR-Cas systems are adaptive immune systems in archaea and bacteria 10,11 . Some Cas nucleases display strong collateral activities after binding to their speci c cis targets, which has been fully evaluated for diagnostic use [12][13][14] . By combination with recombinase polymerase ampli cation (RPA), Cas12 and Cas13 has been shown to permit single molecule detection in reactions 13,15,16 . Although both of them are RNA-guided nucleases 16 , CRISPR/Cas13a may be less promising because it uses single stranded RNA (ssRNA) as reporters, which could be false-positive-prone, as RNases are common and highly stable in the environment.
In the addition to the commonly used RPA method, LAMP (loop-mediated isothermal ampli cation) and PCR approaches has also been reported to facilitate the Cas12a-based nucleic acid detection 17,18 . However, all of these methods have to separate the ampli cation step from the detection step owing to the fact that cleavage of the amplicons and primers by cis-and trans-activities of Cas12a could prevent the chain reactions. The requirement of nucleic acid pre-ampli cation makes this system more timeconsuming and manpower demanding as well as generation of aerosol that usually causes false positive.
We herein report an assay termed Targeting DNA by Cas12a-based Eye Sight Testing in Onepot Reaction ( TESTOR) for rapid, ultrasensitive and speci c detection of nucleic acids. The primers were modi ed with phosphorothioate at certain sites to prevent the degradation by activated Cas12a. The crRNAs were designed to allow the cleavage on amplicons occur at the modi ed sites, whereby the amplicon can be just nicked, enabling the chain reactions to continue. We also optimized the reporters and identi ed that activated Cas12a has the highest a nity to C nucleotide-rich reporters, which enables us to reduce the required time of reaction. Furthermore, we designed a novel reporter labeled with FAM, DIG and Biotin, and modi ed with phosphorothioate at the sites between DIG and Biotin. When combined with a strip that can capture DIG and Biotin at the control and test line, respectively, the unspeci c signal at the test line was completely eliminated. This Cas12-based onepot nucleic acid detection platform (TESTOR) holds the potential to address the key challenges for viral diagnostics and will undoubtedly have a better clinical potential.

Result
Development of Cas12a-based TESTOR detection system Studies of Cas12a-based diagnostics previously applied a separate pre-ampli cation step prior to Cas12a-mediated detection 14,[17][18][19] , which inevitably complicates the procedures and brings about contaminations. To simplify the operations, we attempted to develop an assay in which the components of RPA and Cas12a enzyme were added all together in a single reaction. As an application example, we rst designed primers and crRNA targeting N gene of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a novel coronavirus responsible for the COVID-19 global pandemic. After incubation of the RPA with Cas12a at 37 °C for 40 minutes, however, there was no signal detected (Fig. 1a).
Previous study showed that Cas12 family enzymes generate double-strand DNA breaks in a sequential dependent manner. The non-target DNA strand cleavage is known to precede target DNA strand cleavage. The cleavage of the NTS in dsDNA substrates activates the collateral activity of Cas12a and initiates cleavage of non-target substrates 20 . As the single-stranded primers could also be a substrate of the activated Cas12a, we hypothesized that both cleavage of the amplicons and degradation of the primers contribute to the failing detection of the target gene. To verify this, we performed phosphorothioate modi cations at multiple sites of primers (Fig. 1b). Such modi cations is typically resistant to the cleavage by divalent cation-dependent nucleases 20,21 . A crRNA overlapping multiple nucleotides with the primer was designed, which enables the cleavage by Cas12a-crRNA complex occur at the modi ed sites 20 . As shown in Fig. 1c, the non-target strand without modi cations would be cleaved while the target strand would keep intact after cleavage. Although the detailed mechanisms are unclear, this nicked structure might serve as a template for the DNA ampli cation, enabling continuously production of amplicons. When incubating the components of RPA and Cas12a in onepot at 37°C, RPA ampli cation is rst initiated and the ampli ed target sequence is recognized by Cas12a-crRNA complex. The Cas12a endonuclease is then activated and cleaves the nearby FQ reporters to generate uorescence (Fig. 1d).
In initial experiments with the modi ed primers, we were surprised to nd a strong signal accumulation over time when target was present in the reaction (Fig. 1e, S1a). Nevertheless, when we tried to repeat this nding with other primer pairs, the signals become extremely weak (Fig. S1b, S1c), suggesting instability of this onepot system. To investigate either failed ampli cation of the nucleic acid or unsuccessful cleavage of the amplicons by Cas12a contributed to the faint signal, we resolved the reactions by agarose electrophoresis. Consistent with the uorescent signals, three out of ve reactions showed obvious bands whereas the rest two had little products (Fig. S1d). To identify whether the observed bands were the intended products, 5 mL of the reaction was added into 45 mL of 1X Cas12a mixture. Similar trends were observed but the signals were much stronger compared to that of the onepot assay (Fig.  S1d). We wondered if the RPA reaction system might impede the collateral activity of activated Cas12a, because RPA reaction system is rather viscous. By adding additional 5 mL of water into the onepot system, we found a remarkable increase of signals generated by cleaved reporters (Fig. S1e). These observations drove us to think that if the inhibited cleavage of reporter was resulted from the evaporation of the reaction as an uncapped 96-well plate was used during the monitoring process. We therefore switched to a capped PCR detection system for signal monitoring. As expected, a strong uorescence increasing was observed after incubation at 37 o C for 20 min (Fig. 1f).
We next evaluated the limit of detection (LoD) of the TESTOR system. Whereas our onepot assay lacked the sensitivity to detect single-copy molecule, it unambiguously identi ed 10 copies of target in the reaction within a relative longer time (approximate 1 hour; Fig. 1g). To assess whether this primermodi ed method can serve as a general way for Cas12a-based onepot detection, we designed other primer sets targeting a different region of N (N2) and ORF1ab gene of SARS-CoV-2. In the presence of high number of targets, all reactions for both N2 and ORF1ab could readily distinguish their corresponding targets from NTC (non-target control) within a short incubation time (Fig. 1h, S2a). However, when copy numbers in the system were low, the uorescence-differential time was bit long, and all cases failed to detect single-copy molecule (Fig. 1i, S2b, S2c). Together, these results reveal the feasibility of Cas12a-based onepot detection by phosphorothioate modi cations on the backbone of primers.

Optimization of TESTOR assay
To achieve a better sensitivity in a shorter time, we rst evaluated the effect of primer concentration on the e ciency of our assay. As shown in Fig. 2a, we found the optimal concentration was 0.32 μM.
Cleavage e ciency of CRISPR/Cas9 systems seemed to be in uenced by GC content and purine residues in the gRNA end 22 . We were therefore interested in knowing whether the trans activity of Cas12a was dependent on the sequence of reporters. We found that C nucleotide-rich reporters exhibited the strongest uorescence signal, indicating the highest a nity of this reporter to crRNA-Cas12a complex (Fig. 2b, S3a, S3c). Interestingly, the G nucleotide reporter showed negligible increase of uorescence, implying there was almost no cleavage induced by crRNA-Cas12a complex (Fig. 2b). We further tested the speci city of these reporters, and none of them showed obvious signal increasing over time in NTC groups (Fig. S3b,  s3c). Besides, we found two sites of modi cation on the ends of forward and reverse primers is su cient to trigger the strongest detection. However, too many modi cations seemed to inhibit the reaction (Fig.  2c). We thus used a condition of two-site modi ed primers at the concentration of 0.32 μM for each and combined with C nucleotide-rich reporters in the reaction for the following experiments.
With these optimizations, we compared the canonical two-step method with TESTOR assay in the presence of 100 copies of target in the reaction. The former of which began to produce signal upon incubation at 37 o C, while the latter showed a lag around 10 minutes before reporter cleavage ( Fig. 2d-2f).
This phenomenon could be explained by the required time for amplicon accumulation before recognition and cleavage by crRNA-Cas12a complex. Moreover, we examined the analytic performance of this optimized system for detection of N gene and demonstrated it had a sensitivity of single-copy (Fig. 2g).

Development of a lateral ow TESTOR assay
For any on site detection assays, pairing with a visual readout is crucial especially in situations when instruments are not available. We herein developed a method enabling colorimetric readout with a lateral ow strip. We designed a reporter with a dye label (FAM) on the 5' end and a biotin on the 3' end. When coupled with a ready-to-use strip (Fig. 3a), the destruction of the reporter is visible by naked eye. As described previously 23 , the control line captures abundant reporter that binds to anti-FAM antibody-gold conjugates with streptavidin molecules, which prevents accumulation of the antibody-gold conjugates to secondary antibodies on the test line; the cleaved reporter-antibody-gold conjugates ow over the control line and will be xed there by species-speci c antibodies on the test line (Fig. 3a).
Whereas the positive samples displayed an easy-to-see band after 5 min of lateral ow in the diluted reaction solution (1:5), a faint signal, in the absence of target, was also present at the test line (Fig. 3b).
This unspeci c signal was even more obvious when the dilution ratio increased from 1:5 to 1:10 (Fig. 3c). Time point monitoring demonstrated the unspeci c band was produced as early as 1 min of incubation at room temperature (Fig. 3c).
In clinical applications, presence of an unspeci c faint band at the test line would make judgement of positivity rather di cult, as weak positive samples can also give a faint signal. The unspeci c band might be generated by the antibody-gold conjugates that did not bind to intact or destroyed reporter, making it a problem rather tough to be solved 17 . Here, we designed a novel reporter labeled with biotin, FAM and DIG (Fig. 3d), and modi ed with phosphorothioate at the sites between DIG and Biotin. Besides, we created a dipstick on which the gold particles were coupled with the biotin-ligand molecules in the sample application area, and anti-FAM antibodies and anti-DIG antibodies were immobilized on the control and test line, respectively. When reacted solution with the novel reporter was applied onto the strip, intact reporter will be trapped by anti-FAM antibodies at the rst line while the degraded reporter without FAM uorophore on its 3' end will be captured by anti-DIG antibodies at the second line. As a proof-ofconcept assay, we rst tested our novel system using N gene of SARS-CoV-2 and demonstrated it was highly speci c, and did not yield any signal at the test line in NTC group (Fig. 3e, 3f). Moreover, the signal was much stronger when the reporter was modi ed with phosphorothioate compared to that without such modi cation (Fig. 3e). The inhibited cleavage at the site between biotin and DIG, which leads to increased accumulation of intact ssDNA-antibody-gold conjugates at the test line, may contribute to this result. To con rm it, we applied a phsphorothioate modi ed reporter that was labeled with a FAM molecule on its 5' end and a BHQ1 quencher in the middle, and found the uorescence produced by the reporter cleavage was completely abolished (Fig. 3g, 3h). Consistent with the results obtained by uorescence, the C-nucleotide-rich reporter was able to shorten the reaction time in lateral ow assay as well (Fig. 3i). These results indicate that the TESTOR system for lateral ow assay is of high rapidity and speci city.

TESTOR validation with clinical samples for detection of HPV16 and HPV18
Human papillomaviruses (HPVs) are the major causative agents of cervical carcinomas with types 16 (HPV16) and 18 (HPV18) accounting for most precancerous lesion 24 . We designed sets of primers targeting L1 region of HPV16 or HPV18, and selected the best performance primer pairs for a rapid and sensitive detection (Fig. S4). We then used the selected primer pairs to test extracted DNA from 53 cervical scrape samples taken with a cytobrush. Of the 53 samples, 20 and 12 were positive for HPV16 and HPV18 infection, respectively, by qPCR testing, and 21 were negative for HPV16 or HPV18 infection but either positive for other type of HPVs or negative by all testing. All positive samples for HPV16 or HPV18 by qPCR were positive in our assay, con rming that the false negative rate for TESTOR is very low (Fig. 4a-4d, S5). However, 1 out of 20 the PCR-negative samples showed slight increase in uorescence when the reaction time was more than 40 min (Fig. S6). To explore whether the signal was generated by speci c activation of Cas12a, we performed PCR ampli cation using the product. The ampli ed solution was then subjected to Cas12a detection but, as shown in Fig.4e, no signal was detected, indicating the signal was generated by unspeci c reporter cleavage, which may be due to the long incubation time. Together, our onepot TESTOR assay had 100% positive and negative agreements relative to the qPCR assay, for detection of the HPV16 and HPV18 in a total of 53 clinical samples.
Next, we evaluated our established lateral ow system using 20 positive and 20 negative cervical scrape samples. The positive and negative agreements of the TESTOR assay relative to the qPCR assay were 100%, for detection of both HPV16 and HPV18. One samples with HPV18 infection gave a very faint band at the test line, which is consistent with its qPCR result, revealing the advantages of our established lateral ow system over the traditional one in detecting samples with low viral load, as there was no unspeci c signal at all at the test line in negative samples. These data together illustrate the high sensitivity and speci city of the TESTOR assay.

Discussion
In present study, we provide a general onepot method for Cas12a-based nucleic acid detection by modifying the primers with phosphorothioate and designing crRNA allowing the cleavage occur at the modi ed sites. With these innovations, we integrate the ampli cation and detection into a single reaction system, which is highly valuable as avoidance of uncapping is important for prevention of aerosol generation that easily causes false positivity. Previous studies adopted a physical separation method to combine nucleic acid ampli cation and Cas12a detection into a single tube 16,25 . Wang et. al. added Cas12a on the inner wall of the reaction tube, and a centrifugation step was followed to initiate the cisand trans-cleavage of Cas12a after RPA reaction 16 . However, this method is troublesome and also unreliable. The TESTOR system applies all components into a single common-used tube, which completely circumvents the preampli cation of target nucleic acids and the requirement of special tubes or devices 26 . We also report here, for the rst time, that the trans activity of the activated Cas12a prefers a C-nucleotide sequence. By leveraging this feature, the reaction time for a sensitive detection was signi cantly reduced when C-nucleotide rich reporters were used. To make it more eld deployable, we created a novel lateral ow assay, which surmounts the shortcoming of the routine lateral ow method, and provided a more speci c strip-based system.
Our TESTOR system needs minimal instrumentation and can be performed by lay users, making it suitable for diagnostic tests in resource-restricted areas. As proof-of-concept assays, TESTOR has been developed for SARS-CoV-2 detection. The infection cases of SARS-CoV-2 are increasing rapidly around the world and it seems now to be driven by community transmission 27,28 . As numerous infections are asymptomatic, nucleic acid testing is vital to differentiate infected from healthy individuals 29 . Therefore, a rapid diagnostic method for SARS-CoV-2 is urgently needed. Our Cas12a-based TESTOR technology can be recon gured within days to detect SARS-CoV-2 and has the promise to address the key challenges for this global pandemic. We also validated the TESTOR system using specimens for HPV detection. By combining with a rapid sample processing method, the sample-to-result can be achieved in 30 minutes with high sensitivity (1-10 copies per reaction, data not shown).
Our system reported here could be an alternative to qRT-PCR test as it is faster, simpler and highly speci c.
To facilitate routine surveillance of pathogens and other comprehensive applications, integration of the TESTOR with micro uidic system would be necessary because in many cases there is a need for diagnostic technologies to be able to test many samples while simultaneously testing for many targets. A recent study developed a massive multiplexing system for nucleic acid detection by combining a microwell array that harnesses solution-based uorescent color codes with CRISPR-Cas13 detection. This novel system is capable of testing >4,500 crRNA-target pairs on a single array 30 . Another appealing development in the future for TESTOR would be accommodation of a portable cartridge to streamline the work ow and to enable point-of-care testing in in diverse environments, such as airports, clinics, local communities and other locations. The cartridge could also reduce the risk of aerosol contamination as all processes including sample manipulation, onepot reaction and lateral ow strip visualization could be completed in a closed-evironment 31 .

Nucleic acid preparation
The synthetic DNA fragment of ORF1ab or N gene was ligated into a pUC19 vector (BIOLIGO, Shanghai, China) and ampli ed in E.coli. system. Plasmid DNA was extracted with a commercially available kit (Tiangen, Wuxi, China) and the concentration was quanti ed using a spectrophotometer (Thermo Fisher, NJ, USA). Copy number of the plasmid was calculated based on the concentration using the following equation: DNA copy number = (M × 6.022 × 10 23 )/(n ×1x10 9 x 650), in which M represents the amount of DNA in nanograms, n is the length of the plasmid in base pair, and the average weight of a base pair is assumed to be 650 Daltons.
Primer selection RPA primers for SARS-CoV-2 and HPVs detection were designed according to a protocol described by Twist-Dx (Maidenhead, UK). Primer pairs were screened using a single forward primer against all reverse primers, where the best reverse primer was selected and then used to screen all the forward primers. The best performance primer pairs were used in subsequent experiments.

Fluorescence TESTOR assay
Each lyophilized pellet from the Basic DNA RPA kit was resuspended in a solution comprising 29.4 mL of rehydration buffer, 4 mL of primer mix each at 5 mM, 10.1 mL of H 2 O. This mix was then divided into 2 parts, with each comprising 21.75 mL, and followed by addition of 2 mL of template DNA, 1.25 mL magnesium acetate (280mM) and 6.25 mL of 5x Cas12a mixture containing NEB 2.1 reaction buffer, 250nM Cas12a enzyme, 500nM crRNA, and 1.25mM ssDNA reporters. Reactions were incubated in a Real-Time PCR Detection System for up to 120 minutes at 37°C with uorescent signals collected every 30 s (ssDNA FQ reporter = lex: 485 nm; lem: 535 nm). For all endpoint analysis, uorescence was taken at 30 min after reaction.

Lateral ow TESTOR assay
The components of strip-based TESTOR reaction was similar to that for uorescence-based TESTOR assay, except replacing the FQ reporter with a ssDNA labelled with Biotin, DIG and FAM. The reaction was incubated at 37°C for 30 minutes, followed by dilution at the indicated ratio in Tris-HCL buffer (0.05M, PH=7.6), and then a strip (TwistDx, Cambridge, UK; Bioustar, Hangzhou, China) was inserted and incubated at room temperature. After 2 minutes of incubation, the strip were removed and photographed with a smartphone camera.
Specimen collection and DNA extraction for HPV detection A conventional cytological scrape was taken with a cytobrush from women visiting the gynecological outpatient clinic of the Shenzhen Luohu People's Hospital in China. The specimens were placed into tubes containing 3 mL of cell collection medium (Yaneng Bio, Shenzhen, China) and stored at -20 o C until use. Total DNA was extracted by using a lysis buffer containing 800 mM guanidine hydrochloride, 50 mM Tris-HCl (pH 8.0), 1% Triton X-100, and 0.5% Tween-20. The pellets from 1mL of the liquid-based cervical cytology samples after centrifugation were mixed with 200mL of the lysis buffer and heated for 10 min at 95 o C. The supernatant was used for HPV detection assays following a centrifugation step.
qPCR assay for HPV testing HPV DNA was detected using a HPV test kit (Yaneng Bio, Shenzhen, China) according to the instruction of manufacturer. The assay was performed using the Biorad CFX96 instrument, with the following program: pre-denaturation at 95 o C for 10 minutes, denaturation at 95 o C for 10 seconds, and annealing and extension as well as signal detection at 55 o C for 45 seconds.