SHERLOCK & DETECTR — Cas13/Cas12 collateral cleavage for attomolar viral RNA/DNA detection
CRISPR diagnostics begin not with the Cas enzyme, but with an isothermal amplification step that boosts trace viral genetic material to detectable levels. Recombinase Polymerase Amplification (RPA) — or its LAMP cousin — runs at a single constant temperature, eliminating the bulky thermocyclers that PCR requires and making field-deployable, point-of-care testing possible.
CRISPR-Cas collateral detection is extraordinarily specific, but a single Cas-gRNA complex binding one target molecule does not by itself generate enough signal to see with the naked eye or a simple reader. Clinical samples — nasal swabs, blood, saliva — often contain only a handful of viral genome copies per microliter, especially early in infection.
To bridge this gap, SHERLOCK and DETECTR both pair the CRISPR detection module with an isothermal pre-amplification step. This raises the effective target concentration from single digits to billions of copies before the Cas enzyme ever sees the sample, so that even a faint starting signal becomes an unambiguous one.
Critically, the amplification step is not what confers specificity here — CRISPR guide RNA recognition is. Pre-amplification simply supplies enough raw material for the collateral-cleavage signal to be read out reliably.
RPA replaces PCR's heat-driven strand melting with protein machinery borrowed from bacterial DNA repair:
• Recombinase-primer filaments: a recombinase enzyme (uvsX from T4 phage-derived systems) coats short primers, forming nucleoprotein filaments that scan double-stranded DNA for a homologous sequence — no heat denaturation required. • Strand invasion: the filament invades the duplex and displaces one strand, forming a D-loop at the primer binding site. • Single-stranded binding proteins (SSB) stabilize the displaced strand, preventing re-annealing. • Strand-displacing polymerase (Bsu or Sau polymerase) extends from the primer, displacing the downstream strand as it synthesizes — enabling exponential amplification at a constant 37–42°C in about 20–30 minutes.
For RNA viruses (SARS-CoV-2, Zika, Ebola, Dengue), a reverse transcriptase is added to the same isothermal pot (RT-RPA), first converting viral RNA into cDNA before RPA amplifies it — all in a single tube, single temperature.
RPA is one of several isothermal amplification chemistries compatible with CRISPR detection:
• LAMP (Loop-mediated isothermal amplification): uses 4–6 primers and a strand-displacing polymerase (Bst) at ~65°C, producing highly branched concatemer products. Very robust to inhibitors but requires more primer design effort. • RPA/RT-RPA: runs cooler (37–42°C, near body temperature), completes in 20–30 min, and needs minimal equipment — a heat block or even body heat suffices, making it ideal for field and low-resource settings. • PCR: requires precise thermocycling (20–40 cycles of 94°C/55°C/72°C), specialized instrumentation, and 1–2 hours, but offers mature, extensively validated quantitative chemistry (qPCR).
Both RPA and LAMP trade some of PCR's quantitative precision for speed, simplicity, and equipment independence — exactly the properties needed for CRISPR-based point-of-care diagnostics.
RPA can amplify a target from single-digit copy numbers to over 10⁹ copies in about 20 minutes at a temperature no more precise than a gloved hand — the entire pre-amplification step can, in principle, run in a shirt pocket.
Once amplified nucleic acid floods the reaction, the detection module takes over. A Cas13 or Cas12 enzyme, pre-complexed with a synthetic guide RNA programmed to match a 20–28 nt stretch of the pathogen genome, searches the pool for its complementary sequence — ignoring everything else, including the patient's own genetic background.
CRISPR diagnostics repurpose two distinct Cas nuclease families depending on whether the pathogen genome is RNA or DNA:
• Cas13 (used in SHERLOCK): an RNA-guided RNA-targeting enzyme. It recognizes single-stranded RNA directly — ideal for RNA viruses like SARS-CoV-2, Zika, Dengue, and influenza, or for detecting cDNA-transcribed RPA products as RNA via a T7 transcription step. • Cas12a/Cpf1 (used in DETECTR): an RNA-guided DNA-targeting enzyme, originally characterized as a genome-editing tool. It recognizes double-stranded DNA target sites (requiring a PAM sequence, typically 5′-TTTV-3′) — suited to DNA viruses (HPV) or RT-RPA cDNA products.
Both enzymes are loaded with a CRISPR RNA (crRNA) — the "guide" — synthesized to be exactly complementary to a unique, conserved region of the target pathogen's genome, giving the system its programmability: swapping the guide sequence retargets the same enzyme to a completely different pathogen within days.
The Cas-gRNA ribonucleoprotein complex diffuses through the reaction and continuously samples nucleic acid strands by partial base-pairing with its spacer sequence:
• For Cas12a: the enzyme first locates a PAM (protospacer adjacent motif) on double-stranded DNA, then locally unwinds the duplex to test complementarity between guide and target strand. • For Cas13a: no PAM is needed on the target itself, but a protospacer flanking site (PFS) preference exists; the guide directly base-pairs with the single-stranded RNA target. • Seed-region sensitivity: the proximal ~6–8 nucleotides of the guide-target duplex ("seed region") must pair near-perfectly; mismatches here abort binding early, giving the system single-nucleotide discrimination that engineers exploit to distinguish viral variants (e.g., SARS-CoV-2 lineages) or drug-resistance SNPs. • Non-target strands — including human genomic background and unrelated microbial DNA/RNA — fail to satisfy the seed-pairing requirement and are released unbound, leaving the enzyme catalytically silent.
This recognition step is what gives CRISPR diagnostics their PCR-rivaling specificity: sequence discrimination is built into the guide-target hybridization itself, independent of amplification chemistry.
Because Cas13 orthologs (LwaCas13a, PsmCas13b, CcaCas13b) and Cas12 orthologs each have distinct reporter cleavage preferences (some prefer poly-U reporters, others poly-A or poly-C), a single reaction can house multiple orthologs, each with its own guide RNA and its own distinctly-colored fluorescent reporter.
This multiplexed architecture — demonstrated in SHERLOCKv2 (Gootenberg et al., Science 2018) — allows four pathogens (or four variants of the same pathogen) to be tested simultaneously in one tube, each reported through a different fluorescence channel, dramatically increasing the diagnostic throughput per sample without additional amplification reactions.
The defining discovery behind CRISPR diagnostics: when Cas13 or Cas12 correctly binds its programmed target, it does not stop at cutting that one molecule. Instead, the enzyme undergoes a conformational change that exposes a nuclease domain which indiscriminately shreds any single-stranded nucleic acid nearby — a phenomenon called collateral or "trans" cleavage.
In its resting state, Cas13's two HEPN (Higher Eukaryotes and Prokaryotes Nucleotide-binding) domains are held apart, catalytically inert. Cas12a's RuvC domain is similarly sequestered before target engagement.
When the crRNA-target duplex forms correctly across the full spacer length, the enzyme undergoes a large-scale conformational rearrangement: the two HEPN lobes of Cas13 swing together to form a single composite active site, while Cas12a's RuvC domain becomes exposed and catalytically competent. This structural transition, resolved by cryo-EM in landmark 2017–2018 structural papers, converts the enzyme from a sequence-specific "cis" nuclease acting only on its bound target into a non-specific "trans" nuclease.
Once activated, a single Cas13/Cas12 ribonucleoprotein complex remains bound to its original target (a stable, high-affinity interaction) while its exposed active site processively cleaves any single-stranded RNA (Cas13) or single-stranded DNA (Cas12) it randomly collides with in solution — including molecules with no sequence relationship to the target whatsoever.
This is a catalytic, multi-turnover reaction: one activated enzyme molecule can cleave on the order of 1,000 reporter molecules per second, providing enormous signal amplification from what may originally have been a single target-binding event. The reaction continues for as long as the enzyme remains bound and active reporters remain available — typically tens of minutes.
Collateral cleavage is sequence-agnostic for the reporter (it only requires single-strandedness and, for Cas13 orthologs, a preferred homopolymeric linker such as poly-U), which is precisely what allows a generic fluorescent reporter probe — unrelated to the pathogen sequence — to serve as a universal readout molecule.
A single activated Cas13 enzyme can cleave roughly 1,000 reporter molecules per second — turning one specific molecular recognition event into an enzymatically amplified signal without needing any further nucleic acid amplification.
Collateral cleavage was not an engineered feature — it is an intrinsic biological property of type VI (Cas13) and certain type V (Cas12) CRISPR-Cas systems, believed to serve as an abortive-infection immune strategy in bacteria: once a Cas13 enzyme confirms it has encountered invading phage RNA, indiscriminate RNA degradation can drive the infected bacterial cell into growth arrest or dormancy, sacrificing the individual cell to protect the surrounding bacterial population from a spreading phage epidemic.
The Zhang and Doudna laboratories independently recognized in 2016–2018 that this "downstream" collateral activity — originally an obscure biological curiosity — could be repurposed as a diagnostic signal amplifier, decoupled entirely from its ancestral immune function.
Collateral cleavage becomes visible through short synthetic reporter probes engineered specifically to convert nuclease activity into light or color. A quencher sits at one end of a single-stranded oligo and a fluorophore at the other; while intact, the quencher suppresses fluorescence. Cleavage separates the two, releasing a measurable optical signal.
The most common reporter design is a short single-stranded oligonucleotide (5–6 nucleotides, often a poly-U sequence for Cas13 or poly-T/mixed for Cas12) synthesized with a fluorophore (e.g., FAM, 6-carboxyfluorescein) covalently attached at the 5′ end and a quencher (e.g., Iowa Black FQ) at the 3′ end.
While the reporter is intact, the fluorophore and quencher are held in close spatial proximity, and Förster resonance energy transfer (FRET) suppresses virtually all fluorescence emission — the well stays dark. When an activated Cas13/Cas12 enzyme collaterally cleaves the reporter backbone, the fluorophore diffuses away from the quencher, FRET quenching is lost, and the fluorophore emits freely under excitation light — the well lights up.
Because every activated enzyme molecule cleaves many reporters per second, and because more target molecules mean more activated enzyme complexes, total fluorescence intensity rises in proportion to the amount of original target nucleic acid present — giving the assay a quantitative, not just binary, readout.
For field and point-of-care use, DETECTR and SHERLOCK-based assays can swap the fluorescence reporter for a reporter labeled with FAM at one end and biotin at the other. After the reaction, a paper lateral-flow strip (like a pregnancy test) is dipped into the tube:
• Intact reporters (biotin + FAM both present) are captured by a gold-nanoparticle-conjugated anti-FAM antibody at the sample pad, then flow up the strip and are captured entirely at the control line by anti-species antibody, producing no signal at the test line. • Cleaved reporters separate FAM from biotin; the small cleaved fragment carrying FAM binds the gold-anti-FAM conjugate and is captured at the test line by streptavidin (which binds the separately-migrating biotin fragment) — producing a visible colored test line.
The result: a positive sample produces a visible line on ordinary paper within minutes, needing no fluorescence reader, electricity, or laboratory — the same simplicity as a home antigen or pregnancy test, but built on CRISPR-level molecular specificity.
Reaction fluorescence is typically monitored in real time on a plate reader, with signal traced as relative fluorescence units (RFU) over 10–60 minutes:
• Positive samples show a rapid, roughly linear-then-saturating rise in RFU as reporters are progressively depleted (substrate exhaustion sets an eventual plateau). • Negative/no-target control wells show only a flat, low background trace, arising from rare non-specific reporter degradation or trace nuclease contamination. • The signal-to-noise ratio (positive RFU ÷ negative RFU) commonly exceeds 10–100× within 30 minutes for samples containing the target at clinically relevant concentrations, giving unambiguous separation between positive and negative calls without requiring a quantitative cycle-threshold interpretation as in qPCR.
The complete SHERLOCK or DETECTR workflow — sample lysis, isothermal pre-amplification, and CRISPR collateral detection — collapses what used to require a certified PCR laboratory into a single-tube or single-strip test deliverable at the point of care, with a limit of detection rivaling gold-standard PCR.
SHERLOCK (Specific High-sensitivity Enzymatic Reporter unLOCKing) was introduced by Jim Collins, Feng Zhang and colleagues (Gootenberg et al., Science 2017), combining RPA pre-amplification with Cas13a collateral cleavage to detect Zika and Dengue virus RNA at attomolar concentrations directly from patient samples.
DETECTR (DNA Endonuclease-Targeted CRISPR Trans Reporter) followed in 2018 from Janice Chen, Jennifer Doudna and colleagues, using Cas12a instead of Cas13a to detect human papillomavirus (HPV) DNA with comparable sensitivity and single-nucleotide strain discrimination.
Both platforms were commercialized rapidly: Sherlock Biosciences (spun out of the Broad Institute) and Mammoth Biosciences (co-founded by Doudna) each pursued clinical and point-of-care development, culminating in emergency deployment during the COVID-19 pandemic.
In May 2020, the FDA granted Emergency Use Authorization (EUA) to Sherlock Biosciences' SHERLOCK CRISPR SARS-CoV-2 assay — the first CRISPR-based diagnostic ever authorized for clinical use in the United States. Mammoth Biosciences similarly advanced a DETECTR-based SARS-CoV-2 assay, reporting results in under an hour with sensitivity and specificity comparable to RT-PCR in validation studies.
These authorizations validated CRISPR diagnostics as a legitimate clinical modality, not just a research curiosity — demonstrating that a test born from a bacterial antiviral defense system could, within roughly five years, become an FDA-authorized tool actively used to fight a global pandemic.
The 2020 FDA Emergency Use Authorization of SHERLOCK for SARS-CoV-2 marked the first-ever clinical clearance of a CRISPR-based diagnostic — a technology whose founding papers were published just three years earlier, in 2017.
CRISPR-Cas collateral detection is frequently compared to quantitative PCR (qPCR), the long-standing gold standard for nucleic acid diagnostics:
• Equipment: qPCR requires a precision thermocycler capable of rapid, accurate temperature cycling; SHERLOCK/DETECTR need only a simple heat block or incubator (RPA/LAMP step) plus either a basic fluorescence reader or, for lateral-flow formats, no instrument at all. • Time: qPCR typically takes 1–2 hours including cycling; CRISPR assays complete in 30–60 minutes end-to-end. • Specificity: both approaches offer high specificity, but CRISPR guide-RNA recognition adds an independent, programmable layer of sequence verification on top of amplification primers — useful for discriminating closely related viral variants. • Quantification: qPCR's cycle-threshold (Ct) value gives well-established quantitative viral load estimates; CRISPR fluorescence kinetics can be quantitative but are less standardized clinically. • Deployment: CRISPR diagnostics are explicitly designed for decentralized, point-of-care, and low-resource settings — the core motivation behind their development — while qPCR remains centralized in equipped laboratories.
In practice, the two technologies are complementary: qPCR remains the quantitative reference standard, while CRISPR diagnostics extend rapid, sensitive nucleic acid testing to settings where a full molecular laboratory is unavailable.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| SHERLOCK (Cas13a) | RNA viruses (SARS-CoV-2, Zika, Dengue) | RT-RPA pre-amplification + Cas13a collateral RNase activity on RNA reporters | Attomolar sensitivity; multiplexable; lateral-flow compatible |
| DETECTR (Cas12a) | DNA viruses / DNA targets (HPV, SARS-CoV-2 cDNA) | RPA pre-amplification + Cas12a collateral ssDNase activity on DNA reporters | Single-nucleotide strain discrimination; fast (~30–45 min) |
| qPCR (real-time PCR) | DNA or RNA (with RT step) | Thermocycled exponential amplification with fluorescent probe/intercalating dye detection | Quantitative gold standard; extensively validated; high multiplexing |
| LAMP / RT-LAMP | DNA or RNA (with RT step) | Isothermal strand-displacement amplification with 4–6 primers at ~65°C | Simple visual/colorimetric readout; very rapid; minimal equipment |