For every CRISPR system bacteria evolved to fight phages, the phages evolved a counter-weapon. Anti-CRISPR proteins, or Acrs, are small phage-encoded proteins that block CRISPR-Cas activity in bacteria — and they have become indispensable tools for the gene editing field. Acrs let scientists turn CRISPR off on demand, which sounds like a niche capability but turns out to be one of the most useful safety levers ever discovered. Want to limit how long Cas9 stays active in a patient? Apply an Acr. Want tissue-specific editing? Drive an Acr in non-target cells. Want to prevent off-target editing after the on-target work is done? Acrs are the cleanest answer the field has.
What Are Anti-CRISPR Proteins?
Anti-CRISPR proteins were first reported by Joe Bondy-Denomy, Alan Davidson, Karen Maxwell, and colleagues in Nature in 2013 ("Bacteriophage genes that inactivate the CRISPR/Cas bacterial immune system"). The team studied phages of Pseudomonas aeruginosa that evaded a Type I-F CRISPR system and found small phage genes — initially anonymous open reading frames — that abolished CRISPR activity.
Within five years, the field exploded. Over 100 distinct anti-CRISPR families are now known, blocking all major CRISPR types: I, II, III, V, and VI. Each Acr family is named for the system it inhibits — for example, AcrIIA4 is the fourth family of Listeria-derived inhibitors of Type II-A systems, including the SpCas9 used in most therapeutic editing.
The story is a classic example of evolutionary arms-racing. CRISPR-Cas systems exist because phages threaten bacteria; anti-CRISPRs exist because CRISPR threatens phages; and engineered Cas variants and de novo Acrs are now being developed because researchers want finer control over both.
How Acrs Work at the Molecular Level
Different Acrs use very different strategies. The diversity is itself instructive — there is no single "off switch."
Blocking DNA binding. AcrIIA4 is the prototype. Crystal structures (Dong et al., Nature 2017; Shin et al., Science Advances 2017) show AcrIIA4 occupying the same surface on Cas9 that normally engages PAM-containing DNA. With AcrIIA4 bound, Cas9 cannot recognize its target.
Blocking cleavage after binding. Some Acrs, like AcrIIC1, allow Cas9 to bind DNA but stop the HNH nuclease domain from making the cut. This produces a dead-Cas9-like state without needing to engineer the protein.
Blocking guide loading. A handful of Acrs interfere with Cas-crRNA assembly, preventing the active complex from forming in the first place.
Targeting Cascade or other multi-subunit systems. Type I CRISPR uses a multi-protein Cascade complex; Acrs against Type I are often larger and bind specific Cascade subunits.
Acr enzymes. A few Acrs are enzymes — for instance, AcrVA1 is a guide-RNA cleaving nuclease that destroys Cas12a's crRNA. These are catalytic and thus more potent on a per-molecule basis.
The mechanistic diversity matters because it means Acrs can be combined and titrated. Not all are needed simultaneously, and different mechanisms produce different kinetics of CRISPR shutdown.
Key Papers and Milestones
- Bondy-Denomy et al., 2013 (Nature). Discovery of Type I-F anti-CRISPRs.
- Pawluk et al., 2014, 2016. Expansion to additional Type I systems and the recognition of Acr family diversity.
- Pawluk, Davidson, Maxwell et al., 2016 (Cell). First Type II (Cas9) anti-CRISPRs identified.
- Rauch, Bondy-Denomy et al., 2017 (Cell). AcrIIA4 — the first SpCas9 inhibitor used in human cells.
- Dong et al., 2017 (Nature). Crystal structure of AcrIIA4 bound to Cas9-sgRNA.
- Shin et al., 2017 (Science Advances). Independent structural and functional characterization of AcrIIA4.
- Marino et al., 2018 (Science). Discovery of Cas12a anti-CRISPRs.
- Watters et al., 2018 (Science). Cas13 anti-CRISPRs.
- Bondy-Denomy lab, 2019 onward (Nature and Cell papers). Discovery of dozens more families and mechanistic characterization.
- Hoffmann et al., 2019 (Nucleic Acids Research). Cell-type-specific Acr expression for tissue-restricted editing.
Applications and Use Cases
Temporal control of CRISPR activity in vivo. This is the killer application. When Cas9 is delivered systemically — via LNP or AAV — it can persist for days to weeks. The longer it persists, the more chance for off-target editing. By co-delivering or sequentially delivering an Acr, researchers can shut Cas9 down once the on-target work is complete, dramatically improving the safety profile.
Tissue-specific editing. Drive an Acr from a tissue-specific promoter so it is expressed in cells where editing is undesirable but absent from cells where editing should proceed. This converts a broadly-delivered editor into a tissue-restricted one.
Reducing off-target editing. Several studies (Shin 2017, Aschenbrenner 2020) showed that timed Acr delivery preserves on-target editing efficiency while substantially reducing off-target indels — because off-target events accumulate more slowly than on-target events.
Safer therapeutic editing. Acrs are being explored as built-in safety switches for in vivo CRISPR therapies. If clinical adverse events emerge, an Acr could be administered as a "rescue" intervention to neutralize residual Cas activity.
Allele-specific editing. Combining temporal Acr control with allele-discriminating guides has been used to improve specificity at heterozygous loci.
Biocontainment of gene drives. Anti-CRISPRs can serve as a brake on synthetic gene drives in case of unintended ecological spread.
Acrs vs Other CRISPR Safety Strategies
| Strategy | Mechanism | Reversible | Suitability for in vivo |
|---|---|---|---|
| Anti-CRISPR proteins | Direct Cas inhibition | Yes (with timed delivery) | Excellent |
| High-fidelity Cas variants | Engineered specificity | No | Good |
| Cas9 RNP delivery (vs DNA) | Short half-life | Naturally limited | Good |
| Dose limiting | Lower exposure | Naturally limited | Moderate |
| Self-deleting Cas9 | Cas9 cuts its own coding sequence | No | Good |
Acrs are unique in that they preserve normal Cas9 activity until the moment you want to shut it down. Other safety strategies trade activity for specificity from the start.
Connection to the Broader Gene Editing Ecosystem
Anti-CRISPRs are a safety layer that complements every other CRISPR modality discussed on Gene Editing 101 — including standard CRISPR-Cas9, base editing, and prime editing. They are most relevant to in vivo therapeutic editing where exposure cannot be controlled by simply withdrawing the editor — exactly the regime where delivery system choices matter most. Acrs are also a key part of the broader conversation about CRISPR off-target effects, and as more CRISPR therapies follow Casgevy into the clinic, Acr-based safety layers may become a standard component of trial designs.
Current Limitations and Challenges
- Delivery timing. The therapeutic value of an Acr depends on getting the right amount to the right place at the right time. Too early and editing fails; too late and off-targets accumulate.
- Immunogenicity. Acrs are foreign proteins. Pre-existing immunity is unlikely, but anti-Acr antibodies could develop with repeat dosing.
- Specificity. Each Acr targets a specific Cas family. There is no universal off-switch.
- Compact delivery. Co-packaging an editor and an Acr in a single AAV consumes scarce cargo capacity.
- Mechanistic gaps. Many Acrs remain incompletely characterized at the structural and kinetic level.
FAQ
Who discovered anti-CRISPR proteins?
Joseph Bondy-Denomy, Alan Davidson, and Karen Maxwell, reported in Nature in 2013 in Pseudomonas aeruginosa phages. Bondy-Denomy went on to lead the field at UCSF.
What is AcrIIA4?
AcrIIA4 is a 87-amino-acid protein from a Listeria prophage that inhibits SpCas9 by occupying the PAM-binding surface — preventing target DNA recognition.
Why would you ever want to turn CRISPR off?
Because the longer Cas9 is active, the more off-target edits accumulate. Once the desired on-target edit is made, additional activity is pure risk. Acrs provide a clean way to stop further editing.
Can Acrs be used as drugs?
In principle, yes. Several groups are exploring Acrs as safety adjuncts to in vivo CRISPR therapies, though no Acr-based therapeutic has yet entered clinical trials as of 2026.
How many anti-CRISPR families are known?
Over 100 distinct families spanning all major CRISPR types (I, II, III, V, VI), with new families discovered every year.
Do Acrs work against base editors and prime editors?
Yes, indirectly. Because base editors and prime editors are built on dCas9 or Cas9-nickase chassis, Acrs that block the underlying Cas9-DNA interaction will also inhibit them.
