Rapid and sensitive diagnosis of live Mycobacterium tuberculosis using clustered regularly interspaced short palindromic repeat‐Cas13a point‐of‐care RNA testing

Mycobacterium tuberculosis (MTB) is the causal pathogen of tuberculosis (TB). Rapid and accurate detection of live MTB is important for transmission control and patient treatment. Here, we described a clustered regularly interspaced short palindromic repeat (CRISPR)‐Cas13a‐based molecular diagnosis approach for rapid and specific detection of live MTB. This detection method, which we termed CRISPR‐Live‐MTB, contained two consecutive reactions including nuclear acid sequence‐based amplification (NASBA) and CRISPR‐Cas13a collateral cleavage reaction. CRISPR‐Live‐MTB could efficiently detect MTB single‐stranded RNA (ssRNA) in 2 hours with high specificity over double‐stranded DNA (dsDNA). Importantly, CRISPR‐Live‐MTB exhibited a limit of detection of 2.4 copies for MTB ssRNA, which was 1000 times lower than that of the clinically used NASBA method. Moreover, lateral flow was integrated into the CRISPR‐Live‐MTB method to enable point‐of‐care testing application with a sensitivity of 95% and a specificity of 100%. Overall, our study demonstrated the feasibility of CRISPR‐Live‐MTB as a rapid, sensitive, and specific approach for live MTB detection.


INTRODUCTION
Tuberculosis (TB), caused by Mycobacterium tuberculosis (MTB), is a major public health challenge leading to 1.3 million deaths worldwide in 2020. 1 Clinical practice has proven that efficient diagnosis plays a critical role in the prevention and control of TB. 2 Conventional TB diagnostic methods include etiological, immunological and nucleic acid tests. 3Microbial cultivation is generally deemed the gold standard of etiological diagnosis for TB.Despite the convenience of detecting live MTB bacilli, cultivation methods can require up to 8 weeks 4 with multiple clinical visits, thus resulting in a delay of TB treatment. 5Tuberculin skin test and interferon γ release assay are important auxiliary diagnoses of TB, relying on MTB antigens or lymphocyte response to the antigens respectively. 3However, the results of these immunological tests can be complicated by the immunological status of patients such as Bacille Calmette-Guerin vaccination or co-infection of HIV. 6 Recently World Health Organization has highlighted nucleic acid amplification tests (NAATs) for TB diagnosis, including polymerase chain reaction (PCR)-based Xpert MTB/RIF technique. 7One drawback of Xpert MTB/RIF assay is that the capital cost of the companion components (GeneXpert System) impedes the widespread application in low-incoming countries or regions that in fact bear a higher risk of MTB infections. 1 Isothermal amplification-based NAATs such as loop-mediated isothermal amplification 8 and nuclear acid sequencebased amplification (NASBA) 9 have been developed to alleviate the demand for infrastructure facilities.However, these NAATs have limited ability to differentiate between live and dead MTB.
To overcome these disadvantages, rapid and affordable methods have been developed for specific detection of live MTB.These new diagnostic approaches include improved propidium monoazide (PMA) treatment that allows for selective PCR amplification of live MTB DNA, 10,11 fluorogenic chemical probes that can be activated in response to MTB proteins, [12][13][14] induced expression of specific MTB proteins 15 and bioaerosol capture techniques. 16Other emerging technologies are summarized in a recent review article 3 and are not discussed here.
Recent studies have highlighted clustered regularly interspaced short palindromic repeat (CRISPR)-associated proteins (Cas) (CRISPR-Cas) as a rapid nucleic acid detection technology. 17,18CRISPR diagnostics rely on the collateral cleavage activity of Cas nucleases which could induce the cleavage of single-stranded oligonucleotide reporters to generate fluorescent or immunogenic signals.
0][21][22] These diagnostic methods can be implemented with fluorescence signal and lateral flow, thus enabling point-of-care testing of MTB. 23Interestingly, CRISPR diagnostics can be used to detect cell-free circulating MTB DNA in human serum even in the case of HIV co-infection. 24Furthermore, CRISPR-Cas13a has been employed to identify drug-resistant mutations in MTB with single nucleotide precision. 25Despite the notable progress of CRISPR-based MTB diagnosis, it remains challenging to distinguish dead and live MTB bacilli.
In this study, we developed a CRISPR-Cas13a-based diagnostic platform that can specifically detect MTB ssRNA over genomic dsDNA, thus allowing for the diagnosis of live MTB.We analyzed the performance of this platform, referred to as CRISPR-Live-MTB, on the signalto-noise ratio for RNA over DNA, the limit of detection (LOD), and the specificity among mycobacteria species.Moreover, we investigated the feasibility of developing a CRISPR-Live-MTB-based point-of-care testing (POCT) diagnostic approach by integrating lateral flow.

Strains and clinical samples
Mycobacterium tuberculosis H37Rv (ATCC27294) was preserved in the Second Affiliated Hospital of Fujian Medical University.Clinical MTB was isolated from patients' sputum or bronchoalveolar lavage fluid or puncture fluid.The original sample was collected and pretreated as below.
Collected samples (5 mL) were transferred into a 50 mL centrifugation tube and mixed with an equal volume of 2% NALC-NaOH pretreatment solution, followed by vertexing for 20 s and incubation at room temperature for 15 min.Next sterile phosphate-buffered saline (PBS) was added to the tube to a total volume of 45 mL.The resuspended solution was centrifuged at 3000 g for 20 min and the supernatant was discarded.The precipitated bacteria were resuspended with 1 mL of sterile PBS.Finally, 0.5 mL of the treated specimens was inoculated into Middlebrook 7H9 liquid culture and incubated at 37 • C for 2 months.Collected clinical samples did not contain identifiable personal information and were used solely for molecular diagnosis purposes.

Specimen processing and RNA preparation
Specimen processing was performed using the N-acetyl-L-cysteine (NALC)-NaOH method. 26Sputum of approximately 5 mL was inactivated by thoroughly mixing with an equal volume of NALC-NaOH solution and placed at room temperature for 15 min.The sediment was washed once with a sterile 0.9% NaCl solution and resuspended in 1.5 mL sterile 0.9% NaCl solution.Treated sputum of 500 µL was transferred to a new sterile, nuclease-free 1.5 mL tube.After 5 min of centrifugation at 10,000 g, the pellet was resuspended in a 50 µL lysis buffer consisting of 10 mM sodium citrate, pH 8.0, and vortexed.Bacterial cells in the resuspended pellet were lyzed by a water bath sonicator (Shanghai Sheng-Yan Ultrasound Machines Co. Ltd.) with 300 W power at room temperature for 15 min, followed by 5 min of centrifugation at 10,000 g.
The supernatant was used as the template for subsequent experiments.

Expression and purification of Leptotrichia wadei protein
Purification of Cas13a protein from Leptotrichia wadei (LwCas13a) was performed as previously described 17 with minor modifications.The codon-optimized gene encoding LwCas13a protein was cloned into bacterial expression vector pET28a.A 6×His tag was added to the N-terminus of LwCas13a, the expression of which was under the control of the T7 promoter.LwCas13a expression vector was transformed into Escherichia coli BL21 (DE3).A fresh colony was picked and cultured in Luria-Bertani (LB) broth supplemented with 50 µg/mL kanamycin at 37 • C overnight with shaking.The overnight culture (10 mL) was inoculated into a 2-liter LB broth with 50 µg/mL kanamycin and incubated at 37 • C with shaking at 220 rpm.When the OD 600 reached 0.6 to 0.8, isopropyl-D-1-thiogalactopyranoside was added into the culture to a final concentration of 0.5 mM and protein expression was induced at 16 • C overnight with shaking.The cells were harvested by centrifugation at 5200 g for 15 min at 4 • C and cell pellets were stored at −80 • C.
For protein purification, cell pellets were re-suspended in lysis buffer containing 20 mM Tris-HCl, 500 mM NaCl, 1 mM dithiothreitol (DTT), and pH 8.0.The suspension was sonicated on ice with the following conditions: amplitude of 100 with cycles of 5 s on and 5 s off and a total sonication time of 10 min.The debris was removed by centrifugation for 1 h at 4 • C at 10,000 g and the supernatant was filtered using a Stericup 0.22 µm filter (EMD Millipore).The filtered supernatant was run through the Ni-NTA column (Qiagen) for affinity purification.Further purification was performed using fast protein liquid chromatography with a Superdex 200 filtration column (GE Healthcare Life Sciences).Purified proteins were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis and exchanged to storage buffer containing 50 mM Tris-HCl, pH 7.5, 600 mM NaCl, 5 % (v/v) glycerol, 2 mM DTT and frozen at −80 • C.

NASBA primer design and crRNA preparation
The 16S rRNA sequences of MTB H37Rv (ATCC27294) and 6 other reference strains were obtained from the National Center for Biotechnology Information database and aligned using MEGAX software to identify the conserved regions and variable regions.The NASBA primers were selected in the conserved nucleotide region of the 16S rRNA gene.The T7 promoter sequence was appended to the 5' end of the NASBA forward primers.The primer sequences are listed in Table S1 and synthesized by GENEWIZ.For crRNA preparation, the DNA templates of crRNA were synthesized using GENEWIZ primers with an appended T7 promoter sequence.crRNA template was annealed to a short T7 primer (T7-3G) and then transcribed to crRNA using HiScribe T7 Quick High Yield RNA Synthesis Kit (New England Biolabs) at 37 • C overnight.The crRNA was purified using RNAiso Plus (Takara, 9108), according to the manufacturer's instructions.The transcribed crRNA was stored at −80 • C.

Real time-PCR
RNA samples of 2 µL were reverse-transcribed to cDNA using the PrimeScript RT reagent Kit (Takara).PCR primers are listed in Table S1.For a 10 µL reaction, 0.5 µL of reverse-transcribed cDNA products, 5 µL of PowerUp SYBR Green Master Mix (Applied Biosystems), 0.5 µL of each forward and reverse primers with stock concentrations of 10 µM were mixed at 4 • C and incubated on QuantStudio 6 Flex System thermocycler (Applied Biosystems).The real-time (RT)-PCR reaction was performed using the following cycling conditions: 50

NASBA RNA-specific isothermal amplification
RNA-specific isothermal amplification was performed using a commercial NASBA kit according to the manufacturer's instructions.Briefly, a 20 µL reaction containing 6.7 µL of reaction buffer (Life Sciences, NECB-24), 3.3 µL of Nucleotide Mix (Life Sciences, NECN-24), 0.5 µL of nuclease-free water, 0.4 µL of 12.5 µM NASBA primers, 0.1 µL of Murine RNase inhibitor (Vazyme, R301-01) and 4 µL of RNA input (or water for the negative control) were mixed at 4 • C and incubated at 65 • C for 2 min and then 41 • C for 10 min.Afterward, 5 µL of enzyme mix (Life Sciences, NEC-1-24) was added to each reaction, and the reaction mixture was incubated at 41 • C for 1 to 2 h.Fluorescent NASBA was performed as described above by adding a 50 nM RNA probe (Table S1) into the 20 mL reaction.The RNA probe was labeled

LwCas13a reaction
LwCas13a detection assays were performed as previously described 17 with minor modifications.Reactions were performed in a 20 µL reaction containing 50 nM purified LwCas13a, 0.25 µM FAM-BQ1-labeled ssRNA probe (Genescript), 10 ng crRNA (Table S2), 6 mM MgCl 2 , 20 mM Tris-HCl, pH 7.4, 0.1 µL of Murine RNase inhibitor (Vazyme, R301-01) and 3 µL NASBA products.The reaction was kept at 37 • C for 1 h and the fluorescence was collected on SpectraMax iD3 Multi-Mode Microplate Reader with an excitation wavelength of 485 nm and an emission wavelength of 520 nm.The fluorescence kinetics was measured every 5 min.For lateral flow-based detection, the reaction was performed using commercially available lateral flow strips (Milenia HybriDetect 1, TwistDx) according to the manufacturer's instructions, where the FAM-BQ1-labeled ssRNA probe was replaced by FAM-biotin-labeled ssRNA probe with an adjusted final concentration of 1 µM in a 20 µL LwCas13a reaction.The LwCas13a reaction was performed at 37 • C for 1 h and then 20 µL products were transferred into 100 µL Hybridetect assay buffer, followed by loading of the products onto lateral flow strips.The band intensity of the strip was directly read with the naked eye or imaged by the camera for further densitometry analysis.

Design of NASBA-coupled CRISPR-Cas13a diagnostics for live MTB detection
As viable MTB can be determined by detecting RNA at the transcriptome level, 27 we designed an MTB 16S rRNAspecific NASBA reaction to amplify the RNA as substrates for downstream CRISPR-Cas13 recognition.In the NASBA reaction, MTB rRNA is first reverse-transcribed into complementary DNA (cDNA), and then the rRNA in the RNA-DNA duplex is degraded by RNase H.A doublestranded DNA (dsDNA) is synthesized using the cDNA as the template.RNA product will be generated from the newly synthesized dsDNA by T7 transcription reaction.Importantly, MTB genomic DNA will not serve as the template for RNA production due to the lack of T7 promoter.Therefore, this NASBA reaction allows for MTB RNAdependent amplification of RNA signal.The products from the NASBA reaction are subjected to Cas13a reaction which induces the collateral cleavage of ssRNA reporter for subsequent fluorescence or lateral flow analyses.The schematic presentation of this approach referred to as CRISPR-Live-MTB, is shown in Figure 1A in comparison with the traditional cultivation approach (Figure 1B).

Establishment and optimization of the CRISPR-Live-MTB platform
We expressed and purified the Cas13a protein from LwCas13a.Ni-NTA and size exclusion chromatography purification of LwCas13a gave a purity of more than 95% (Figure S1A,B).To design crRNAs for Cas13a reaction, we first aligned the sequences of 16S ribosomal RNA (rRNA) from MTB and six non-tuberculosis mycobacteria (NTM) and identified suitable primer binding sites for mycobacterium-wide NASBA amplification (Figures S1C and S2A and Table S3).Within the NASBA-amplified region, we designed seven Cas13 crRNAs including six MTB-specific crRNAs and one mycobacterium universal crRNA (Figure S2A).Comparative analysis showed that MTB-crRNA5 resulted in the strongest fluorescent signal (Figure S2B).All the following CRISPR-Live-MTB reaction was thus performed with MTB-crRNA5 unless noted otherwise.
In addition, it was noted that NASBA amplification has a remarkable impact on the output of the fluorescence signal.Two forward primers and four reverse primers were designed for NASBA amplification and eight primer combinations were screened for the amplification efficiency by quantifying the fluorescence signal from a coupled Cas13a fluorescence reaction with MTB-crRNA5.It was found that the combination of MTB-NASBA-FWD-1 and MTB-NASBA-REV-4 showed the highest efficiency (Figure S2C).

Comparison of NASBA and CRISPR-Live-MTB for specific detection of MTB ssRNA
To investigate the advantage of CRISPR-Live-MTB reaction over conventional NASBA, we analyzed their specificity for detecting MTB RNA over DNA and their sensitivity on dilute samples.The input MTB RNA was quantified by RT quantitative PCR (RT-qPCR) using the primer set of MTB-NASBA-FWD-1 and MTB-NASBA-REV-4 (Figure S3).We fixed the total reaction time to 2 h where the NASBA reaction contained a non-stop 2 h reaction and CRISPR-Live-MTB contained a 1 h NASBA reaction followed by a 1 h CRISPR reaction.It was found that both NASBA and CRISPR-Live-MTB exhibited markedly higher fluorescence signals on the ssRNA template over the DNA template (Figure 2A,B), suggesting that specific detection of MTB ssRNA is a feasible approach.The detection specificity of ssRNA over DNA exhibited time-dependent dynamics.NASBA and CRISPR-Live-MTB showed peak specificity of more than 5 at around 40 min and 30 min, respectively (Figure 2A,B).It was also noted that CRISPR-Live-MTB resulted in a much higher fluorescence signal.
Next, we sought to compare the LOD of fluorescent NASBA reaction and CRISPR-Live-MTB for MTB RNA detection.It was found that the NASBA reaction had a LOD of 2400 copies (corresponding to 0.2 fM template concentration) (Figure 2C and Figure S4A).Notably, CRISPR-Live-MTB had a largely reduced LOD of 2.4 copies (corresponding to 0.2 aM template concentration) (Figure 2D and Figure S4B).It was also found that the CRISPR-Cas13a reaction alone, without a coupled NASBA amplification reaction, had a very high LOD for MTB RNA detection (Figure S4C,D), demonstrating the essential role of coupled NASBA reactions for CRISPR-Live-MTB.Collectively, the largely improved LOD of CRISPR-Live-MTB in comparison with NASBA strongly suggested its potential advantages in clinical applications.

Species-specific and universal detection of MTB and NTMs with CRISPR-Live-MTB
To investigate whether CRISPR-Live-MTB could be expanded to other mycobacteria, we designed specific crRNAs targeting the 16S rRNA for each of the six selected mycobacterium species.We used transcribed 16S rRNA from MTB, NTM, and non-mycobacterium bacteria as the template and determined the cross-species reactivity of the crRNAs over different bacterial RNA (Figure 3A and Figure S5A).These species-specific crRNAs showed notable selectivity toward designated mycobacteria RNA (Figure S5B).
Importantly, MTB-crRNA5 (target site shown in Figure S2A) was found to be highly specific to MTB over NTM and non-mycobacterium bacteria (Figure 3B).Additionally, we designed mycobacterium universal crRNA according to the conserved genomic sequence in MTB and NTM (Figure S2A) and found that the universal crRNA exhibited broad activity toward all tested mycobacteria including MTB and NTM while retaining the specificity over unrelated bacterium (Figure 3B).The LOD of universal crRNA on the MTB RNA template was determined to be 2.4 copies (corresponding to 0.2 aM template concentration) (Figure 3C).The near single-copy LODs of MTB-crRNA5 and mycobacterium universal crRNA demonstrated that CRISPR-Live-MTB was a highly sensitive diagnostic platform.

Detection of MTB 16S rRNA with CRISPR-Live-MTB in a lateral flow mode
To explore the POCT application of CRISPR-Live-MTB, we designed a lateral flow-based signal readout approach.In this case, the fluorogenic reporter was replaced with a FAM-biotin-labeled ssRNA, allowing the display of cleavage signal on commercial lateral flow strips.As shown in Figure 4A, the control line was displayed which relied on the interaction between immobilized streptavidin and flowing biotin.In the presence of MTB RNA, the ssRNA reporter will be cleaved, releasing the free FAM molecule that can be captured by the gold nanoparticle-conjugated anti-FAM antibody in the test line (Figure 4A).
We first analyzed MTB-crRNA5 for the specific detection of MTB and found that this crRNA probe was highly selective to MTB over other bacteria (Figure 4B,C).Mycobacterium universal crRNA could also specifically detect all examined mycobacteria over other unrelated bacteria (Figure 4D,E).We next sought to determine the LOD of MTB-crRNA5 and mycobacterium universal crRNA under lateral flow conditions.Consistent with the fluorescence reaction, both MTB-specific and universal crRNAs exhibited LOD of 2.4 copies (corresponding to 0.2 aM template concentration) in lateral flow reaction (Figure 4F-I).These results demonstrated that lateral flow CRISPR-Live-MTB was a rapid, specific, and highly sensitive method for detecting live MTB.

Detection of live MTB in clinical samples using CRISPR-Live-MTB
To enable the rapid detection of MTB RNA, we intended to integrate an extraction-free method of RNA preparation into CRISPR-Live-MTB.Thus, the clinical samples were inactivated for 15 min using NALC-NaOH method 26 followed by a 15 min sonication process for RNA release.The RNA samples were then subjected to CRISPR-Live-MTB and the results were resolved on lateral flow strips with densitometry analyses (Figure 5A).
We collected 60 MTB clinical samples without identifiable patient information.Among these samples, 40 have been confirmed by cultivation method to have live MTB (S1-S40), and the other 20 specimens (S41-S60) bearing no live MTB.Using CRISPR-Live-MTB and a positive signal threshold of mean plus three-fold standard deviation (mean + 3SD) of the mock group, 38 out of the 40 cultivation-positive specimens (S1-S40) were identified to contain live MTB with the exception of S18 and S39 (Figure 5B,C), thus revealing a 95.0% sensitivity.The failed detection on S18 and S39 could be due to insufficient nucleic acid quantity or contamination on specimens that interfered with CRISPR-Live-MTB reactions.The 20 negative specimens (S41-S60) were determined to be all negative by CRISPR-Live-MTB, meaning a 100.0%specificity.Overall, CRISPR-Live-MTB resulted in a positive predictive agreement of 100.0% and a negative predictive agreement of 90.9% (Figure 5C).Importantly, the present CRISPR-Live-MTB platform may be further optimized to enhance the detection efficiency, reduce the detection time, or enable closed-lid reactions, like in other CRISPR diagnostics platforms. 28he factors to be considered for optimization include the integrated NASBA reaction, the volume ratio between NASBA and CRISPR reactions, fluorescent probes, and others.In addition, it must be noted that CRISPR-Live-MTB is now implemented with an extraction-free RNA preparation method.Because of the compact MTB cell wall and the high viscosity of clinical specimens, the efficiency of RNA preparation can be a key consideration for the performance of CRISPR-Live-MTB.Further efforts can be made to maximize the compatibility of the RNA preparation method with CRISPR-Live-MTB.
The present study included clinical specimens without considering disease progression or the stages of treatment.In future studies, more clinic-directed analysis should be performed for CRISPR-Live-MTB.Particularly, it would be interesting to compare CRISPR-Live-MTB with clinically used methods such as cultivation or NASBA for monitoring live MTB in patients under treatment.In addition, mutation-specific crRNA can be investigated to establish CRISPR-Live-MTB as a convenient platform for identifying common drug-resistant MTB strains.
CRISPR diagnostics kits have been already approved by the authority for SARS-CoV-2 diagnosis. 29However, unlike other NAAT kits that typically detect two target genes, approved CRISPR diagnostic products only detect one target gene with a single fluorescence channel.One important reason supporting the use of only one target gene is that the CRISPR diagnostics have coupled isothermal amplification reaction prior to CRISPR collateral cleavage reaction.The coupled reaction setting provides dual detection specificity.Nevertheless, approaches for multiplexed gene target detection have been developed 30 and these approaches may be combined with CRISPR-Live-MTB for further improvement.
Finally, the combination of lateral flow with CRISPR-Live-MTB makes this technology one of the few methods that support POCT diagnosis of live MTB.POCT diagnosis enabled real-time monitoring of TB disease progression, which is key to evaluating the effects of treatment and adopting appropriate therapeutic strategies.Although the present study has made the proof-of-concept demonstration for POCT-enabled CRISPR-Live-MTB, further efforts should be made to optimize the reaction components and diagnostic strategy to establish a user-friendly process.For example, the number of components can be reduced to enable a one-pot reaction, as shown in previous studies. 31articularly, recent studies have highlighted the importance of protein engineering for improving the detection efficiency of LwaCas13a. 32It will be interesting to explore whether these engineering efforts can help enhance the sensitivity and specificity of CRISPR-Live-MTB on clinical samples.In conclusion, we believe that CRISPR-Live-MTB can greatly expedite the clinical practice of live MTB diagnosis and facilitate the control and treatment of tuberculosis.

F I G U R E 1
Schematic of Clustered regularly interspaced short palindromic repeat-Live-Mycobacterium tuberculosis (CRISPR-Live-MTB) workflow.(A) CRISPR-Live-MTB approach.CRISPR RNAs (crRNAs) targeting the 16S rRNA gene were designed for MTB RNA detection.Conventional RNA extraction can be used as an input.The signal readout can be a fluorescent signal or colorimetric bands on lateral flow strips.RT, reverse transcription.The T7 promoter is highlighted in blue.(B) Traditional cultivation approach.

F I G U R E 2
Comparison of NASBA and NASBA-linked clustered regularly interspaced short palindromic repeat-Live-Mycobacterium tuberculosis (CRISPR-Live-MTB) methods for the specific detection of MTB RNA.(A, B) Specificity of detecting MTB RNA over DNA using NASBA (A) and CRISPR-Live-MTB (B).The input concentrations of RNA and DNA templates for the above two reactions are 10 7 to 10 8 copies.NASBA reaction is monitored over a course of 2 h, and CRISPR-Live-MTB only the CRISPR reaction following the coupled 1 h NASBA reaction is monitored for 1 h.(C) Determination of the limits of detection (LODs) of the NASBA method.(n = 2 biological replicates; data shown as mean ± SD). (D) Determination of the LODs of the CRISPR-Live-MTB method.(n = 3 technical replicates; data shown as mean ± SD).Both NASBA and CRISPR-Live-MTB methods are monitored over a total reaction time of 2 h where CRISPR-Live-MTB contains coupled NASBA (1 h) and CRISPR-Cas13a (1 h) reactions.MTB-crRNA5 is used for CRISPR-Live-MTB reactions.The cut-off for positive signals is set as values more than the mean + 3SD of the mock group.

F I G U R E 3
Cross-species reactivity of MTB-specific and universal crRNAs.(A) Schematic diagram showing detection of mycobacteria using species-specific or universal crRNAs.(B) Performance of MTB-specific crRNA (MTB-crRNA5) and mycobacterium universal crRNA with Cas13a reaction.Reactivity of universal crRNA toward different mycobacteria.(C) Determination of the limit of detection (LOD) of the mycobacterium universal crRNA on MTB RNA template.The data from three technical replicates are shown as mean ± SD.The cut-off for positive signals is set as mean + 3SD of the mock group.

F I G U R E 4
Lateral flow clustered regularly interspaced short palindromic repeat-Live-Mycobacterium tuberculosis (CRISPR-Live-MTB) for detection of Mycobacterium tuberculosis (MTB) and non-tuberculosis mycobacteria (NTM).(A) Schematic illustration of the design of lateral flow CRISPR-Live-MTB.(B, C) Specific detection of MTB using MTB-crRNA5.(B) Images of lateral flow strips.(C) Quantification of band density on the strips.(D, E) Detection of mycobacteria using universal crRNAs.(D) Images of lateral flow strips.(E) Quantification of band density on the strips.(F, G) Determination of the limit of detection (LOD) of MTB-crRNA5 toward MTB RNA.(F) Images of lateral flow strips.(G) Quantification of band density on the strips.(H, I) Determination of the LOD of universal crRNA toward MTB RNA.(H) Images of lateral flow strips.(I) Quantification of band density on the strips.ImageJ is used for the densitometry analyses.The data from three technical replicates are shown as mean ± SD.The cut-off for positive signals is set as mean + 3SD of the mock group.

4 DISCUSSION
Mycobacterium tuberculosis (MTB) is a public health threat worldwide.Rapid diagnosis of live MTB is not only important for the control of pathogen reservoirs but is also vital to the assessment of clinical practice.Unfortunately, existing culturing and molecular biology diagnostic technologies are largely limited by excessive detection time, the requirement of centralized infrastructure, or the lack of ability to distinguish genomic DNA and transcriptomic RNA.Recent studies have uncovered the potential of CRISPR diagnostics as a rapid and convenient option for clinical microbes.Here we have developed a CRISPR-Cas13-based diagnostic platform that allows for the detection of live MTB within 2 h.Compared with NASBA, which is used in clinical practice for MTB RNA detection, our platform CRISPR-Live-MTB has a 1000-fold improvement in detection sensitivity while maintaining the specificity toward MTB over mycobacteria or other unrelated bacteria.Importantly, we have demonstrated that CRISPR-Live-MTB can be implemented in a POCT setting, which is critical to patients in developing countries where inconvenient transportation and visits to clinics can disrupt clinical practice.

F I G U R E 5
Detection of live Mycobacterium tuberculosis (MTB) in clinical samples using lateral flow clustered regularly interspaced short palindromic repeat-Live-Mycobacterium tuberculosis (CRISPR-Live-MTB).(A) Schematic diagram showing an RNA extraction-free approach for CRISPR-Live-MTB.RT, room temperature.LFD, lateral flow detection.(B) Images showing the detection results of CRISPR-Live-MTB on lateral flow strips.The control band (C) is shown at the bottom, and the test band (T) is shown at the top.Images are taken at 3 min after sample loading.(C) Quantification of band density on the strips in (B) using ImageJ.The data from three technical replicates are shown as mean ± SD.The cut-off for positive signals is set as mean + 3SD of the mock group.PPA, positive predictive agreement.NPA, negative predictive agreement.