A multiplexed Cas13-based assay with point-of-care attributes for simultaneous COVID-19 diagnosis and variant surveillance

Point-of-care (POC) nucleic acid detection technologies are poised to aid gold-standard technologies in controlling the COVID-19 pandemic, yet shortcomings in the capability to perform critically needed complex detection—such as multiplexed detection for viral variant surveillance—may limit their widespread adoption. Herein, we developed a robust multiplexed CRISPR-based detection using LwaCas13a and PsmCas13b to simultaneously diagnose SARS-CoV-2 infection and pinpoint the causative SARS-CoV-2 variant of concern (VOC)—including globally dominant VOCs Delta (B.1.617.2) and Omicron (B.1.1.529)—all while maintaining high levels of accuracy upon the detection of multiple SARS-CoV-2 gene targets. The platform has several attributes suitable for POC use: premixed, freeze-dried reagents for easy use and storage; convenient direct-to-eye or smartphone-based readouts; and a one-pot variant of the multiplexed detection. To reduce reliance on proprietary reagents and enable sustainable use of such a technology in low- and middle-income countries, we locally produced and formulated our own recombinase polymerase amplification reaction and demonstrated its equivalent efficiency to commercial counterparts. Our tool—CRISPR-based detection for simultaneous COVID-19 diagnosis and variant surveillance which can be locally manufactured—may enable sustainable use of CRISPR diagnostics technologies for COVID-19 and other diseases in POC settings.


Expression and purification of protein components of RPA
We used near-identical expression and cell lysis protocols for UvsX, UvsY, gp32, and Bsu LF, but the purification step and final storage buffer conditions were different for each enzyme, as given below.
An expression plasmid for UvsX, UvsY, gp32, or Bsu LF was transformed into E. coli BL21(DE3) cells. The cells were grown in LB medium at 37°C until OD600 reached 0.7-0.8. Protein expression was induced using the following concentrations of isopropyl-β-D-thiogalactopyranoside (IPTG): 1 mM for UvsX and UvsY; 0.2 mM for gp32; and 0.5 mM for Bsu LF. The cells were grown for additional 16 hours at 16°C and harvested by centrifugation. The cell pellet was resuspended in lysis buffer (50 mM sodium phosphate pH 8.0, 500 mM sodium chloride, 10 mM imidazole). Phenylmethylsulfonyl fluoride (PMSF) was added into resuspended solution at 1 mM final concentration followed by sonication for a total burst time of 3 minutes (3 sec on for short burst and 9 sec off for cooling). The cell lysate was clarified by centrifugation at 27,000×g for 30 minutes at 4 °C. Nickel-nitrilotriacetic acid (Ni-NTA) agarose beads (Qiagen) was washed with two column volumes of water and equilibrated with lysis buffer. The clarified cell lysate was incubated with the Ni-NTA agarose beads (Qiagen) at 4 °C for 40 minutes. The column was washed with ten column volumes (CVs) of washing buffer (50 mM sodium phosphate pH 8.0, 500 mM sodium chloride, and 20 mM imidazole), and the bound proteins were eluted with elution buffer (50 mM sodium phosphate pH 8.0, 500 mM sodium chloride, 250 mM imidazole).
For UvsX, the protein was further purified by Heparin Sepharose Fast Flow (GE Healthcare Life Sciences) with a linear gradient of 0.1-1 M NaCl in Buffer A (20 mM Tris-HCl pH 8, 5 mM βmercaptoethanol (βME)). Proteins were concentrated and stored in 20 mM Tris-HCl pH 8.0, 500 mM NaCl, and 40% (v/v) glycerol at −20 °C.
For UvsY, the protein was further purified by Heparin Sepharose Fast Flow (GE Healthcare Life Sciences) with a linear gradient of 0.1-1 M NaCl in Buffer A (20 mM Tris-HCl pH 8, 5 mM βME). Afterward, proteins were dialyzed against 20 mM Tris-HCl pH 7.5, 400 mM NaCl, 20% (v/v) glycerol and 5 mM βME, and the glycerol concentration adjusted to be 50% (v/v) with the addition of UvsY freezing buffer (20 mM Tris-HCl pH 7.5, 400 mM NaCl, and 80% (v/v) glycerol). Protein aliquots were stored in 20 mM Tris-HCl pH 7.5, 400 mM NaCl, and 50% (v/v) glycerol at −20 °C. µg/mL ampicillin, and the recombinant gene expression induced by the addition of 500 µM IPTG at 16 ºC for overnight. Cells were collected and lysed as for PsmCas13b. Following lysate clarification, the soluble fraction was filtered through 0.2 µm polyethersulfone membrane and purified by HisTrap FF column connected to FPLC system. The column was pre-equilibrated with binding buffer (50 mM Tris-HCl, 500 mM NaCl, 5 mM imidazole; pH 7.5). The soluble fraction was loaded at 2 mL/min, washed with 5 column volumes of binding buffer. The recombinant protein was eluted in linear gradient of elution buffer (50 mM Tris-HCl, 0.5 M NaCl, and 0.5 M imidazole; pH 7.5). The eluted fractions were pooled and dialyzed with 50 mM Tris-HCl, 0.5 mM EDTA, 1 mM DTT and 200 mM NaCl; pH 7.5 at 4 ºC for 3 hours, before proceeding to MBP tag removal.
The MBP tag was cleaved with ultraTEV protease using 20:1 substrate: protease ratio in 50 mM Tris-HCl, 0.5 mM EDTA, 1 mM DTT and 200 mM NaCl; pH 7.5. The reaction mixture was incubated at 4 ºC for overnight with gentle shaking before applied onto HisTrap column, which was pre-equilibrated with (50 mM Tris-HCl, 500 mM NaCl, 5 mM imidazole; pH 7.5). The flowthrough and unbound fractions containing RfxCas13d and RfxCas13d-dsRBD protein were then collected while retained MBP and other contaminants were later washed out with high concentration of imidazole. The purified protein was exchanged against 40 mM Tris, 400 mM NaCl; pH 7.5 in DEPC-treated water and concentrated using centricon 30k, and diluted to a concentration of 1.22 mg/mL (for RfxCas13d) and 1.89 mg/mL (for RfxCas13d-RBD).

RPA primers, crRNAs, and RNA reporters
The oligonucleotides used are listed in Table S2. The crRNAs used for variant detection were synthesized via in vitro transcription. The T7-3G oligonucleotide (at the final concentration of 0.5 µM) and the ssDNA crRNA template (at the final concentration of 0.5 µM) were subjected to 34 cycles of 50 μL Q5 ® High-Fidelity DNA Polymerase reaction (New England Biolabs). PCR products were purified using DNA Clean & Concentrator-5 kits (Zymo Research) and eluted in 20 μL of nuclease-free water. Four picomoles of the purified dsDNA crRNA template were used in in vitro transcription reactions using RiboMAX Large Scale RNA Production System-T7 (ProMega) or MEGAshortscript™ T7 Transcription Kit (Invitrogen). The reactions were performed at 37 ºC overnight, treated with DNase I, and purified using phenol-chloroform extraction and alcohol precipitation. 15% acrylamide/7.5 M urea PAGE with GelRed ® (Biotium) staining was used to assess the size and purity of the transcribed product. Gels were imaged on ImageQuant™ LAS 4000 (GE Healthcare); quantifications of produced crRNAs were performed by measuring band densitometries using Fiji ImageJ software 4 and comparing to RNA standards.

Design of RPA primers and crRNAs
General RPA primers and crRNAs design RPA primers and crRNA were designed according to a published protocol. 5

Design of RPA primers and crRNA targeting specific SARS-CoV-2 mutations
Representative genome sequences of Alpha (n = 1,100), Beta (n = 126), Gamma (n = 729), and Delta (n = 1,329) were retrieved from NCBI SARS-CoV-2 data packages using PANGO designations 6 as queries (dates of data retrieval, 27 April and 5 May 2021). Retrieved sequences were aligned using MAFFT 7 . The alignments were visualized and used to create consensus sequence for each variant using Jalview. 8 The four consensus sequences along with the SARS-CoV-2 isolate Wuhan-Hu-1 NCBI reference genome (Accession ID NC_045512.2) were realigned using MAFFT and visualized in SnapGene software (Insightful Science). Using visualized multiple sequence alignment, RPA primers and crRNAs were manually designed to cover regions spanning the target mutation while avoiding regions with genetic variations among SARS-CoV-2 variants. The design and chosen regions are shown in Figure S14 and Figure S15. , with the maximal number of target sequences set to 5000. The results were filtered to include complete genomes (>29,000 nt) with 90-100% query coverage and 90-100% identity (date of data retrieval 1 March 2022). The clade assignment of genomes was done using NextClade (https://clades.nextstrain.org) 9 . The genomes with bad and mediocre NextClade overall quality control status were excluded. The analysis was summarized in Figure S20.

Inclusivity evaluation of designed primers and crRNAs
GenBank accession lists of SARS-CoV-2 variant genomes sampling by the NextStrain COVID-19 global analysis open-data build were acquired by applying filter data by clade before downloading author metadata (date of data retrieval, 3 March 2022); Alpha (n = 110), Beta (n = 30), Gamma (n = 31), Delta (n = 1,044), and Omicron (n = 785)). FASTA files comprising listed genomes were downloaded from GenBank and aligned using MAFFT. The resulted alignments were visualized and inspected using JalView 9 . The prevalence of the sequences aligned with primers and crRNAs in different variants were calculated as a percentage of genomes with the complete matched sequence observed in the alignments and represented by the frequency of the sequence among the variant isolates (%F). Consensus sequences of each variant along with their %F values are provided in Figure S15 for primers and Figure S14 for crRNAs.
The NextStrain Clades or PANGO lineages were designated to each WHO classified variant of concern as described in Table S1 RPA Primer and primer combination screening with colorimetric lateral-flow Cas13a-based readout (for Figure S1) RT-RPA was set up as previously described 3 using TwistAmp Basic Kit (TwistDx) and EpiScript reverse transcriptase (Lucigen). Sequences of RPA primers used were given in Table S2. LwaCas13a-based detection reactions were also set up exactly as previously described, 3 then visualized with HybriDetect lateral-flow strips (Milenia Biotec). Sequences of LwaCas13a-crRNAs for each amplicon were given in Table S2.

One-pot, multiplexed RT-RPA/CRISPR-Cas13a detection
We prepare the one-pot, monophasic detection reactions by preparing the multiplex RT-RPA and CRISPR-Cas13a reaction mixes separately, before mixing them together, as follows.
To initiate amplification and detection, 5.3 μL of the RNA sample was added to each aliquot, followed by 0.7 μL magnesium acetate (280 mM). The reactions were incubated at 39 °C and generated FAM fluorescence was monitored using a real-time thermal cycler (CFX Connect Real-Time PCR System -Bio-Rad).

Optimizing RPA through titration of protein components
Concentrations of the main protein components of RPA (UvsX, UvsY, Gp32, and Bsu LF) were varied as indicated in Figure 4D. We also switched to using RPA primers for the n gene amplification (the F4/R1 pair) and pUC57-2019-nCoV-N plasmid (MolecularCloud cat no. #MC_0101085) at 10,000 copies/µL as the DNA template. The RPA reaction conditions were otherwise identical to the standard RPA reaction (see previous section), and were allowed to proceed at 42°C for 60 min. Thereafter, 2 µL of the RPA products were mixed with 18 µL of the Cas13a-based detection reaction, which contained 20 mM Tris-HCl pH 7.4, 60 mM NaCl, 6 mM MgCl2, 1 mM of each rNTPs, 1.5 U/μL NxGen T7 RNA polymerase, 6.3 µg/mL LwaCas13a, 0.5 ng/µL LwaCas13a crRNA, and 0.3 µM FAM-PolyU reporter. The generated FAM fluorescence was monitored at 37 °C over 90 min using a fluorescence microplate reader (Varioskan, Thermo Scientific or Infinite M Plex, Tecan).

RT-qPCR
The RT-qPCR was performed according to a published protocol. 11 In brief, a 5 µL of the RNA sample was added to a 20-µL reaction prepared with Luna® Universal Probe One-Step RT-qPCR Kit (New England Biolabs) and nCoV_N1 primer-probe set (Integrated DNA Technologies). The reactions were monitored using CFX Connect Real-Time PCR System (Bio-Rad).

Droplet digital RT-PCR (RT-ddPCR)
The RT-ddPCR was performed using One-Step RT-ddPCR Advanced Kit for Probes (Bio-Rad) according to the manufacturer's protocol and in line with RT-qPCR protocol described above. A 20-µL reaction consisted of 5 µL of RNA input, 5 µL of Supermix, 2 µL of reverse transcriptase, 1 µL of 300 mM DTT, 4 µL nuclease-free water, and each 1 µL of following component from nCoV_N1 primer-probe set (Integrated DNA Technologies): 10 µM forward primer, 10 µM reverse primer, and 10 µM probe. The reaction droplets were generated and read using a QX100™ Droplet Digital™ PCR system (Bio-Rad) connected to a T100™ Thermal Cycler (Bio-Rad). Thermocycling conditions were reverse transcription at 50 °C for 60 minutes, enzyme activation at 95 °C for 10 minutes, followed by 40 cycles of 95 °C for 15 seconds, 55 °C for 30 seconds, 58 °C for 10 seconds. The enzyme deactivation was done at 98 °C for 10 minutes. The quantification was carried out using QuantaSoft™ software (Bio-rad).

Standard curve of qPCR Ct versus ddPCR-derived copies/uL RNA input see Figure S27
Data analysis and visualization Statistical analyses were performed, and graphical representations created with GraphPad Prism 9, unless indicated otherwise. Schematic diagrams were created with Biorender.com and Adobe Illustrator CC 2017. RT-RPA using each primer pair was performed with SARS-CoV-2 RNA at two different concentrations as a template. Cas13a-based detection with lateralflow readout was then used to assess successful amplification. Experiments with R1 reverse primers (Qian et al.) and R2 reverse primers (bottom) were performed using different SARS-CoV-2 RNA dilutions, but with an identical control primer pair (F1, R1) included in both sets of experiments, allowing us to compare relative performance of all primer pairs. Based on this screening, we selected the (F4, R1) primer pair for subsequent n gene amplifications. D) Multiplexed RPA reactions with pairwise combinations of RPA primers (4 different primer pairs; 6 pairwise combinations). A multiplexed RPA is individually assessed with a LwaCas13a-based detection reaction, each programmed with a specific crRNA for the SARS-CoV-2 gene, and lateral-flow readout. Successful amplification was marked by production of a strong-colored band at a test band (T), which indicates target-activated Cas13a activity and resulting in efficient cleavage of the FAM-biotin reporter. Figure S2: Evaluation of the sensitivity and specificity of the Multiplexed RT-RPA assay with real-time fluorescence detection. (A) Optimizing total primer concentrations in multiplexed RPA with 1:1 malar ratio of n to s primer. Multiplexed RPA for the s and n gene of SARS-CoV-2 was performed using indicated primer concentrations and serially diluted SARS-CoV-2 genomic RNA template. In the second step, generated s and n amplicons were detected using a multiplexed CRISPR-Cas reaction containing LwaCas13a and PsmCas13b enzymes. LwaCas13a is programmed with a crRNA targeting the s amplicon and cleaves a FAM/IABkFQ -functionalized polyU reporter once targetactivated (left), while PsmCas13b is programmed with a crRNA targeting the n amplicon and cleaves a Cy5/IABkRQ-functionalized polyA reporter (right). (B) Fine-tuning primer concentrations in the multiplexed RPA reaction. Multiplexed RPA for the s and n gene of SARS-CoV-2 was performed using indicated primer concentrations, with serially diluted SARS-CoV-2 genomic RNA as a template. Amplicons were detected in a multiplexed CRISPR-Cas reaction, via FAM fluorescence generated from s-targeted LwaCas13a (left) and Cy5 fluorescence from N-targeted PsmCas13b (right). (C) Optimizing RNase H concentrations in the RT-RPA reaction. RT-RPA for s gene amplification from serially diluted SARS-CoV-2 RNA, followed by LwaCas13a-based detection, was used. (D) Orthogonality of LwaCas13a and PsmCas13b in a multiplexed CRISPR-Cas detection. RPA reactions were performed with only s primers (s), only n primers (n), or combined s and n primers (s+n). Endpoint fluorescence intensities of the multiplexed detection reactions (90 min) containing all components for LwaCas13-based detection of the s gene and PsmCas13b-based detection of the n gene were shown. Generated FAM (left) and Cy5 (right) fluorescence indicated cleavage of the reporters by the s-targeted LwaCas13a and the n-targeted PsmCas13b, respectively. For C-F, negative controls have no SARS-CoV-2 RNA template input. Data are mean ± s.d. from 3 replicates. (E) Analytical sensitivity of increasing of RPA pellet. Multiplexed RT-RPA for the s and n gene of SARS-CoV-2 was performed using indicated number of RPA pellets, with diluted SARS-CoV-2 genomic RNA at Ct 37 as a template. Subsequently, the amplicons were detected via FAM fluorescence generated from LwaCas13a-based reaction programmed with sand n-targeted crRNAs.  (A) Screening for additives that enhance multiplexed RT-RPA. Left: nine additives were assessed in a conventional two-step (RPA, then LwaCas13a-based CRISPR-Cas) SHERLOCK detection for the s gene of SARS-CoV-2. Middle: four active additives were evaluated further in triplicates, using less RNA input. Right: two best-performing additives-triglycine and taurine-were finally assessed in a multiplexed RPA reaction for the s and n gene of SARS-CoV-2. Two serial dilutions of SARS-CoV-2 RNA were used as a template; negative control has no SARS-CoV-2 RNA template input. Generated s and n amplicons were detected in separate LwaCas13a-based detection reactions, whose FAM fluorescence signal is shown. (B) Screening for additives that improve the multiplexed CRISPR reaction. Twenty-two additives were assessed in the multiplexed CRISPR-Cas detection, using an RPA product from a multiplexed. Endpoint FAM (Qian et al.) and Cy5 (bottom) fluorescence intensities were normalized against intensities obtained from the no-additive controls. Data are mean ± s.d. from 3 replicates. L-Arginine EE. 2HCl, L-Arginine ethyl ester dihydrochloride. TMAO, trimethylamine N-oxide. (C) Cumulative benefits of triglycine additive in the multiplexed RPA and betaine monohydrate additive in the multiplexed CRISPR-Cas reaction. Multiplexed RPA to amplify the s and n genes of SARS-CoV-2 was performed in the presence or absence of 40 mM triglycine, using serially diluted SARS-CoV-2 RNA as a template, and multiplexed RPA was allowed to proceed for 25 min at 42 ºC. Multiplexed CRISPR-Cas reactions to detect the s and n amplicons were then performed in the presence or absence of 500 mM betaine. FAM and Cy5 fluorescence signal (indicative of Cas13a-mediated s gene and Cas13b-mediated n gene detection respectively) is shown. RNase-free water was used as input of all negative control reactions. Data are mean ± s.d. from 3 replicates.     Thereafter, the multiplexed RPA products were used in a freshly prepared Cas13a-based detection with sand n-targeted crRNAs. FAM fluorescence generated over 120 min for each condition is shown. Middle vs right graphs: lyophilized LwaCas13a-based detection reaction has similar sensitivity as freshly prepared reactions.

Figure S11: Effects of cryoprotectants on lyophilized CRISPR-Cas13a reactions.
LwaCas13a-based detection reactions were prepared with or without 5% (w/v) PEG20000 and 6% (w/v) trehalose added as a cryoprotectant. MgCl2 was omitted from all reactions and was added at the rehydration step. The detection was performed at 37 °C using n gene RPA product at various dilution levels and monitored under FAM channel using a real-time thermal cycler. The input volume of the RPA dilutions was 2 μL except for the 1:100 dilution sample in which we used either 2-μL or 4-μL input volume. The end-point fluorescence in tubes was visualized using a BluPAD transilluminator.         Reactions were monitored on a microplate reader over 120 min, then imaged using blue/white LED transilluminator excitation, lighting gels for emission filtration, and smartphone-based image capture. The ROX signal for one Ct 37 sample is very weak due to the amount of RNA input being near the limit of detection.