An engineered CRISPR-Cas12a variant and DNA-RNA hybrid guides enable robust and rapid COVID-19 testing

The study addresses the need for rapid, robust, and accessible COVID-19 diagnostics that remain accurate despite viral mutations and RNA editing. Conventional qRT-PCR, while sensitive, is slow, equipment-intensive, and not ideal for point-of-need use. Existing rapid tests (serology, antigen) have limitations in identifying infectious individuals or sensitivity. CRISPR-based diagnostics offer programmability and specificity but face challenges: susceptibility to target mutations/editing (e.g., ADAR/APOBEC), multi-temperature workflows due to Cas enzyme temperature constraints vs RT-LAMP conditions, frequent reliance on purified RNA, and lack of integrated human internal controls. The research aims to engineer and optimize a CRISPR-Cas12a-based assay (VaNGuard) that is mutation-tolerant, fast, single-temperature compatible with RT-LAMP, applicable directly to clinical specimens without extraction, and includes an internal control.

The paper situates its work within CRISPR-Dx developments (e.g., DETECTR, SHERLOCK, STOPCovid) and notes thousands of SARS-CoV-2 genomes with mutations affecting diagnostic targets, including qRT-PCR primer/probe sites. Prior CRISPR-Dx often operated around 37 °C, conflicting with RT-LAMP’s 60–65 °C, necessitating two heat blocks and additional handling. Some alternatives used RPA (supply limitations) or omitted amplification (reduced sensitivity). AapCas12b and TtCsm were previously used for higher-temperature workflows. LAMP-only diagnostics are susceptible to non-specific amplification; CRISPR provides a specificity gate. Reports on guide modifications (3′-extensions, chemical modifications) showed potential to enhance Cas12 collateral activity, but generalizability remained unclear. Guanidine chloride was reported to enhance LAMP sensitivity and speed.

Assay design and components: - Cas enzymes and guides: Evaluated five Cas12a variants (AsCas12a, enAsCas12a E174R/S542R/K548R, enRR, enRVR, LbCas12a) with multiple 20-nt guides targeting SARS-CoV-2 ORF1ab, S, and N (including N-Mam as DETECTR benchmark). Assessed mismatch tolerance via single-nucleotide mismatches across spacer positions. - Reaction temperatures/buffers: Characterized enAsCas12a collateral activity across 37–65 °C in buffers (CutSmart, Tango, NEB 2.1, 3.1) with/without DTT. - LAMP optimization: Systematically tested primer-set design (core F3/B3, FIP/BIP, loop primers LF/LB), primer concentration titrations, and augmented schemes (swarm primers; stem primers evaluated but deprioritized). Assessed mismatch effects at primer 3′ and 5′ ends and introduced truncated internal primers (tPM-3, tPM-5) and Q5 high-fidelity polymerase (0.15 U) to rescue mismatch-induced losses. Explored chemical additives (glycine, taurine, DMSO) and guanidine (40 mM). Selected final primer set targeting S gene with high specificity to SARS-CoV-2. - Guide engineering: Tested 5′ extensions (+4, +9 nt), chemically modified gRNAs (2′-O-methyl, 2′-fluoro, phosphorothioates), and DNA-RNA hybrid guides with 2 or 4 DNA substitutions at selected spacer positions for S2 and S6 loci. Selected hybrid guides (4 DNA bases) for best performance without background. - Quasi-one-pot workflow: Performed RT-LAMP (typically 65 °C; also tested 63 °C) followed by immediate addition of Cas12a RNP mix directly into the LAMP tube at 60 °C, enabling a single heat block operation. Assay total time ~30 min (22 min RT-LAMP, 5 min CRISPR cleavage, ~2 min readout). - Readouts: Fluorescence (microplate readers) and lateral flow dipsticks (FITC-biotin reporter; HybriDetect) with ratio-based interpretation. - Sample handling: Evaluated impact of collection media (UTM, QuickExtract, SAFER), human RNA/DNA backgrounds, and direct specimen use. For extraction-free workflows, applied proteinase K treatment and heat (95 °C, 5 min) to saliva, contrived UTM, and clinical NP swabs to inactivate RNases and enhance LAMP. - Internal control: Screened LAMP primer sets for human genes POP7, ACTB, GAPDH using heat-treated saliva. Selected ACTB Set2 as internal control. Implemented dual-colour readout: green DNA-binding dye for human amplification and red Cy5-quencher reporter for SARS-CoV-2 CRISPR signal. Optimized internal control conditions by halving human primer concentrations and adding thermostable pyrophosphatase (2 U) to the CRISPR step to mitigate pyrophosphate/Mg2+ effects. - Analytical sensitivity/specificity: Determined limits of detection with synthetic IVT RNA (wild-type and S254F and S254F/N234N mutant targets), contrived virus in saliva/UTM, and assessed cross-reactivity with other coronaviruses and respiratory viruses. - Clinical evaluation: Tested purified RNA from NP swabs (45 qRT-PCR positives, 30 negatives) and crude NP swabs without RNA extraction (21 positives, 21 negatives). RT-LAMP at 65 or 63 °C; CRISPR at 60 °C with enAsCas12a and S2+S6 hybrid guides. - Statistical analyses: Student’s t-tests (one-sided or two-sided as appropriate) were used to compare conditions; biological replicate counts specified throughout.

- Engineered enAsCas12a displayed superior mismatch tolerance and higher on-target collateral activity than LbCas12a and other tested variants across multiple guides. It remained active from 37 °C to 65 °C, enabling same-temperature workflows with RT-LAMP. - Two-guide strategy (S2 + S6) buffered against single-nucleotide variations at target sites; presence of S6 rescued activity when S2 site carried a disruptive SNV (e.g., MM10 or real S254F mutation). - LAMP robustness: Mismatches at FIP/BIP 3′ ends and FIP 5′ end significantly slowed amplification; combining truncated internal primers (tPM-3) with Q5 high-fidelity polymerase restored kinetics. Swarm primers markedly accelerated LAMP when combined with loop primers; doubling B3 primer slightly improved sensitivity. - Guide engineering: DNA-RNA hybrid guides (2–4 DNA substitutions) significantly accelerated Cas12a trans-cleavage and suppressed off-target/background, outperforming unmodified guides without template-triggered activation. 5′ 9-nt extensions improved activity at 37–60 °C for some guides but could induce background at 60 °C (S2), thus excluded from final design. Chemically modified gRNAs improved speed but were costlier and not superior to hybrids. - Buffer/additives: Buffer 2.1 with DTT or CutSmart with DTT improved kinetics; Tango buffer facilitated 60 °C operation. Guanidine chloride (40 mM) enhanced LAMP sensitivity and overall assay performance more than glycine. - Quasi-one-pot single-heat-block workflow achieved total assay time ≤30 min; detection possible within 5 min of CRISPR step. - Analytical LoD: Using optimized conditions and hybrid guides, LoD for synthetic RNA was 20 copies/reaction; lateral flow detected as low as 2 copies in some runs under quasi-one-pot at 60–63 °C. - Specificity: No cross-reactivity observed with SARS-CoV, MERS-CoV, common human coronaviruses (OC43, 229E, NL63, HKU1), or other respiratory viruses (influenza A/B, parainfluenza, mumps, measles, EV-D68, HRV-89, bocavirus) at 1E6 copies input. - Clinical purified RNA (NP swabs): On 45 positives and 30 negatives, specificity 100%; clear positives for Ct ≤ 33.32; effective LoD ~50 copies/reaction (2 copies/µL). Eight false negatives at higher Ct values; zero false positives. - Direct clinical NP swabs (no extraction; proteinase K + heat): On 21 positives and 21 negatives, specificity 100%; positive calls consistent for Ct ≤ 28.98 (~≥1000 copies/reaction). Overall LoD ~1000 copies/reaction (40 copies/µL). Four false negatives near Ct 29–33; zero false positives. - Internal control: Integrated ACTB internal control in the same tube with dual-colour readout functioned on crude clinical samples; reducing human primer load and adding pyrophosphatase improved SARS-CoV-2 detection without compromising internal control.

The VaNGuard assay directly addresses key limitations of prior CRISPR-Dx: it tolerates target sequence variation via an SNV-tolerant enzyme (enAsCas12a) and dual-guide design; achieves rapid turnaround by accelerating both LAMP (swarm primers, guanidine) and CRISPR (hybrid guides), and by enabling a single-temperature, quasi-one-pot workflow; operates on crude clinical matrices without extraction to reduce time and complexity; and incorporates a human internal control within the same reaction, improving result validity. The robust temperature range of enAsCas12a simplifies instrumentation needs and mitigates supply-chain constraints associated with alternative isothermal methods. Specificity remains high due to CRISPR’s sequence-resolved detection, eliminating false positives from non-specific LAMP products. The findings show that, within 30 minutes, clinically relevant viral loads are detected with high specificity, and that the system remains effective against certain real-world spike gene mutations (e.g., S254F) and across backgrounds of human RNA/DNA and common collection media. These attributes make VaNGuard suitable for point-of-need deployment and adaptable to evolving viral populations.

This work introduces VaNGuard, a rapid, robust CRISPR-Cas12a diagnostic for COVID-19 featuring: an engineered enAsCas12a tolerant to SNVs; dual-guide targeting to buffer mutations; hybrid DNA-RNA guides to accelerate CRISPR detection and suppress background; optimized LAMP (truncated primers, Q5 polymerase, swarm primers, guanidine) for speed/sensitivity; a quasi-one-pot single-temperature workflow; compatibility with direct NP swabs without RNA extraction; and an integrated human internal control. The assay achieves 100% specificity, with LoDs of ~20 copies/reaction (synthetic), ~50 copies/reaction on purified clinical RNA, and ~1000 copies/reaction on crude NP swabs, and completes within 30 minutes. Future work should expand validation on larger, diverse clinical cohorts, further benchmark across emerging variants, refine matrix tolerance for inconsistent crude samples, and translate to field-deployable platforms for broader infectious disease diagnostics.

- Sensitivity on crude NP swabs shows variability near Ct 29–32, likely due to matrix-dependent inhibition and sample-to-sample variability; LoD is higher (1000 copies/reaction) than for purified RNA. - Mutation tolerance was demonstrated for specific sites (e.g., S254F, N234N) and may not generalize to all possible target variations; ongoing variant surveillance and guide updates are required. - Some guide modifications (5′ extensions) risk template-independent activation at elevated temperatures and were excluded; careful guide engineering is necessary per locus. - LAMP non-specific amplification remains a risk without the CRISPR specificity gate; assay performance depends on precise primer optimization. - The study relies on specific buffers/additives and enzyme sources; performance may vary with reagent lots or alternative suppliers. - Clinical evaluations, while substantial, would benefit from larger, multi-center studies and broader sample types (e.g., saliva) to confirm generalizability.