A CRISPR-based ultrasensitive assay detects attomolar concentrations of SARS-CoV-2 antibodies in clinical samples

Coronavirus disease 2019 (COVID-19) caused by SARS-CoV-2 created an urgent need for sensitive and specific assays to detect infection and associated immune responses at both individual and population levels. CRISPR-based point-of-care tests (POCTs), including Cas13 SHERLOCK and Cas12 DETECTR, have received EUA from the U.S. FDA for nucleic acid testing. However, despite their success as nucleic acid tests (NATs), CRISPR systems have not been exploited for sensitive detection of anti-SARS-CoV-2 antibodies, which are critical for diagnosis, seroprevalence estimation, pandemic monitoring, and assessment of humoral immunity after infection or vaccination. Conventional immunoassays such as ELISA and CLIA are widely used but are not optimal as POCT and often lack the sensitivity to detect early seroconversion or low antibody levels in immunocompromised populations. The authors introduce an ultrasensitive CRISPR-based nucleic acid detection (UCAD) assay that is 10,000 times more sensitive than commercial ELISA for detecting anti-RBD IgG and IgM directly in undiluted human serum, aiming to enable early diagnosis and improved monitoring of immune responses, including in immunocompromised individuals.

Prior work established CRISPR-based diagnostics (e.g., SHERLOCK, DETECTR) as rapid, sensitive NATs for SARS-CoV-2 RNA detection under resource-limited conditions. These platforms gained EUA and are effective for viral RNA detection. Conventional serological assays (ELISA/CLIA) are standard for antibody detection but can lack sensitivity for early infection and in immunocompromised patients. Previous CRISPR-linked protein assays and ultrasensitive protein detection methods exist, but there remains a need for a simple, homogeneous, highly sensitive approach adaptable to POCT. The present work extends CRISPR diagnostics from nucleic acids to protein (antibody) detection by converting antibody binding events into a DNA barcode that triggers Cas12a activity.

Assay design and principle: UCAD converts antibody recognition into a CRISPR-Cas12a-detectable double-stranded DNA (dsDNA) barcode. Two DNA strands (target strand, TS; non-target strand, NTS) each contain half of the crRNA binding sequence. In the absence of target antibody, the truncated duplex has a low melting temperature (~10 °C) and does not stably form or activate Cas12a. TS is conjugated to SARS-CoV-2 spike receptor binding domain (RBD), and NTS is conjugated to anti-human IgG (or IgM). When anti-RBD antibodies bridge RBD and anti-human IgG/IgM, the two DNA strands are brought into proximity, forming a stable duplex (Tm ~46 °C), enabling primer extension to synthesize a full dsDNA barcode that can be amplified and detected by Cas12a. Workflow: A homogeneous, isothermal, single-tube, three-step protocol at 37 °C: (1) Antibody-specific primer extension produces the full dsDNA barcode from the proximity-assembled TS/NTS probes. (2) Recombinase polymerase amplification (RPA) amplifies the barcode. (3) CRISPR-Cas12a collateral cleavage of a fluorophore-quencher ssDNA reporter generates a fluorescence signal. An alternative lateral flow readout is achieved by using a FAM/Dig-labeled reporter to enable visual detection on a strip. Optimization and analytical performance: Assay conditions were optimized for endpoint fluorescence at 40 min for best dynamic range. Sensitivity reached at least attomolar levels, with detection of extremely low antibody quantities (reported as detection as low as 10 anti-RBD human monoclonal antibodies). UCAD detected anti-RBD IgG and IgM in certified positive human sera, with minimal signal in negatives. Compared to ELISA, UCAD detected signals in sera diluted up to 1:100,000, whereas ELISA worked reliably only up to 1:10, indicating ~10,000-fold greater sensitivity. Specificity was demonstrated by low cross-reactivity with anti-MERS-CoV RBD, anti-SARS-CoV-2 nucleocapsid protein antibodies, and rabbit polyclonal anti-RBD across 1 pM to 1 nM, and by background-level signals in sera from SARS-CoV (2003) patients. Modularity: By switching the antigen on the TS probe, UCAD can target antibodies to WT, Delta, or Omicron RBDs, or to SARS-CoV-2 nucleocapsid protein. Variant-specific TS probes yielded high signals only for matching monoclonal antibodies (100 nM to 1 pM), with moderate cross-reactivity consistent with shared conserved RBD epitopes and the lowest cross-reactivity observed between WT and Omicron. Clinical validation cohorts and procedures: Sera were collected under an ethics-approved protocol. Clinical validation included 65 CLIA-confirmed positive post-vaccination sera and 132 negatives (77 pre-pandemic and 55 in-pandemic CLIA-negative) totaling n = 197 for performance analysis. UCAD IgG/IgM signals were acquired as time-resolved fluorescence and analyzed at 40 min for ROC and cutoff determination. A confusion matrix compared UCAD results with standard CLIA. Additional testing involved immunocompromised kidney transplant recipients (KTRs) after 2 doses and after a third vaccine dose to assess UCAD’s ability to quantify low antibody levels. Experimental details (selected): DNA probes were synthesized and HPLC-purified; biotin-streptavidin conjugation linked antigens (WT/Delta/Omicron RBD or N protein) to DNA probes. A typical 50 µL UCAD reaction contained serum (5 µL), antigen-modified TS probe (10 µM), anti-human IgG/IgM NTS probe (10 nM), dNTPs (40 µM), T4 polymerase (2 U) in buffer, incubated at 37 °C for 20 min. RPA was then performed at 37 °C for 20 min. Cas12a detection used 40 nM Cas12a, 10 nM crRNA, and fluorogenic ssDNA reporter with fluorescence read at 495/520 nm. For lateral flow, RPA amplicons were incubated with Cas12a/crRNA and FAM-labeled reporter before loading onto strips developed in SSC buffer. Data analysis used Excel and GraphPad; melting temperatures were estimated with OligoAnalyzer; statistics included unpaired two-tailed t-tests and one-way ANOVA with multiple-comparisons corrections.

- UCAD achieves ultrasensitive antibody detection, approximately 10,000-fold more sensitive than standard ELISA, detecting signals in sera diluted up to 1:100,000 versus ~1:10 for ELISA. - Analytical detection of extremely low amounts of anti-RBD human monoclonal antibodies; endpoint fluorescence optimized at 40 min for quantitative analysis. - High specificity with low cross-reactivity to anti-MERS-CoV RBD, anti-SARS-CoV-2 nucleocapsid antibodies, and rabbit anti-RBD across 1 pM to 1 nM; background-level signals in SARS-CoV (2003) sera confirm specificity. - Modularity: By swapping the antigen on the TS probe, UCAD specifically detects antibodies to WT, Delta, and Omicron RBDs, and to nucleocapsid; highest signals occur only with matched antigen–antibody pairs over 100 nM to 1 pM. - Clinical validation (n = 197): Sensitivity 100% and specificity 98.5% relative to standard CLIA. Significant differences between CLIA-positive post-vaccination and pre-pandemic controls (p < 0.0001). Only two discordant positives among 132 negatives suggest low-level antibodies below CLIA LOD. - Immunocompromised KTRs: UCAD quantified low anti-RBD IgG/IgM levels in KTRs that were “undetectable” by standard assays. After a third vaccine dose, 84.8% (28/33) of KTRs showed increased anti-RBD IgG levels, indicating effective booster responses detectable by UCAD. - A lateral flow format was engineered, enabling visual POCT-compatible readouts for both IgG and IgM.

The study addresses the gap in CRISPR diagnostics for protein detection by converting antibody binding events into a DNA barcode recognizable by Cas12a, thereby extending CRISPR testing from nucleic acids to proteins. UCAD retains the simplicity and isothermal operation of CRISPR assays and is compatible with existing SHERLOCK/DETECTR platforms. Its ultrahigh sensitivity (10,000-fold over ELISA) enables earlier detection of seroconversion and accurate quantification of low antibody levels in immunocompromised patients, such as kidney transplant recipients. The assay’s modularity allows rapid retargeting to different SARS-CoV-2 variants and to other antigens (e.g., nucleocapsid) by swapping the antigen recognition motif. The clinical validation demonstrates excellent sensitivity and specificity versus CLIA, and the lateral flow adaptation supports point-of-care use. These findings suggest broad applicability for early diagnosis, seromonitoring post-infection or vaccination, and potential extension to other infectious diseases and clinically relevant protein biomarkers.

UCAD is a simple, homogeneous, isothermal CRISPR-based assay that translates antibody detection into nucleic acid readout, achieving attomolar-level sensitivity and approximately 10,000-fold improved sensitivity over ELISA. It demonstrates excellent clinical performance (100% sensitivity, 98.5% specificity in n = 197), variant-specific modularity, and effective detection of low antibody levels in immunocompromised KTRs, including booster-induced increases undetected by standard tests. UCAD can be implemented with fluorescence or lateral flow readouts, supporting deployment in centralized labs and POCT settings. Future work will validate the assay in larger cohorts, track temporal antibody dynamics in diverse populations, and generalize the platform to other infectious agents and clinically important protein biomarkers.

- Cohort size for positives and negatives, while adequate for initial validation, remains limited; the authors note a need for larger-scale clinical validation across broader cohorts. - Two discordant positives among CLIA-negative samples may reflect low-level antibodies below CLIA LOD; resolving such discrepancies requires longitudinal follow-up and orthogonal assays. - Moderate cross-reactivity between some variant RBDs due to conserved epitopes may affect strict variant discrimination without further probe optimization. - Experimental design notes include no predetermined sample size, non-randomization, and lack of blinding, which may introduce bias. - Single-center sample sourcing may limit generalizability; external, multi-center validations are warranted.