Compact CRISPR editors are rewriting the rules of in vivo gene therapy. For a decade, the field has wrestled with a stubborn delivery ceiling: adeno-associated virus (AAV), the dominant vehicle for delivering genome editors into human tissue, can only carry about 4.7 kilobases of cargo. The workhorse editor, Streptococcus pyogenes Cas9 (SpCas9), burns through roughly 4.2 kb of that budget on its own, leaving almost no room for a guide RNA, a tissue-specific promoter, regulatory elements, or a repair template. Compact CRISPR editors — CasX, CasMINI, CasΦ, Cas12f variants, and Cas-CLOVER — were engineered or discovered to break through that ceiling, and they are now powering a new wave of therapeutic programs from Scribe Therapeutics, Poseida, and academic labs worldwide.
This deep dive walks through the biology, the key discoveries, the trade-offs, and the companies betting on miniature editors.
What Is a Compact CRISPR Editor?
A compact CRISPR editor is any RNA-guided nuclease (or nickase) small enough to fit comfortably inside a single AAV vector together with its guide RNA and the regulatory machinery needed to drive expression in target tissue. In practice, "compact" usually means under 1,000 amino acids, or roughly 3 kb of coding sequence — about 25 to 60 percent the size of SpCas9.
Size matters for three reasons:
- AAV cargo limit. Beyond ~4.7 kb, AAV packaging efficiency collapses. Dual-AAV strategies exist, but they require two infections of the same cell and reconstitution of a split editor, which cuts efficiency and complicates manufacturing.
- Tissue specificity. Strong tissue-specific promoters (liver, muscle, retina, CNS) can each consume 1–2 kb. Smaller editors leave budget for specificity.
- Immunogenicity. Smaller proteins present fewer epitopes and, in some cases, come from organisms humans rarely encounter, lowering the risk of pre-existing immunity.
For context, see our primer on what CRISPR is and how it works and the deeper look at gene-editing delivery systems, which explains why AAV is both indispensable and constraining.
Mechanism: How Small Editors Still Cut DNA
All compact Class 2 CRISPR effectors share the same basic biochemistry as SpCas9: an RNA guide directs the protein to a complementary DNA sequence flanked by a short protospacer-adjacent motif (PAM), the protein interrogates the DNA, and — if the match is good — it introduces a double-strand break or a nick.
What miniaturization sacrifices is usually one or more of the following:
- PAM flexibility. Smaller proteins often recognize stricter or longer PAMs, reducing the fraction of the genome they can target.
- On-target activity. Fewer protein–DNA contacts can mean lower cutting efficiency, especially in chromatinized cellular DNA.
- Thermal stability. Some ultracompact editors evolved in cold environments and lose activity at 37 °C until engineered.
Engineers compensate with directed evolution, guide-RNA scaffold redesign, and fusion to accessory domains that restore activity.
The Key Compact Editors
CasX (Cas12e) — Jennifer Doudna, 2019
CasX was discovered through metagenomic sequencing of groundwater samples from Deltaproteobacteria and Planctomycetes. Liu, Knott, and colleagues (Doudna lab, Nature 2019) showed it was an RNA-guided nuclease of roughly 980 amino acids — about 30 percent smaller than SpCas9 — with a distinct structural fold and minimal homology to Cas9 or Cas12a. Crucially, it retained robust genome editing in human cells while targeting a TTCN PAM.
CasX became the foundation of Scribe Therapeutics, co-founded by Doudna, which has built an entire therapeutic platform on an engineered CasX chassis called CasX Molecules. Scribe partnered with Biogen on ALS targets and with Sanofi on hematologic indications, with the compact size enabling single-AAV delivery to the CNS and other hard-to-reach tissues.
CasMINI — Lei Stanley Qi lab, 2021
If CasX was small, CasMINI was tiny. Qi's lab at Stanford (Molecular Cell, 2021) started with Cas12f2, a 529-amino-acid natural protein that barely worked in human cells, and used structure-guided engineering plus guide-RNA redesign to boost its activity to levels comparable with Cas12a. The resulting editor is, at roughly 529 amino acids, the smallest programmable DNA editor demonstrated to work efficiently in human cells.
CasMINI's DNA fits in well under half of AAV's cargo budget, leaving room for base-editor or prime-editor fusions, dual guides, and strong tissue-specific promoters. Several startups — and Qi's own academic spinouts — are pursuing CasMINI-based base editors.
CasΦ / CasPhi — Doudna lab, 2020
CasΦ (Cas12j) was discovered in huge bacteriophages — viruses that infect bacteria. Pausch and colleagues (Science, 2020) showed that CasΦ is only about 700 amino acids, uses a single RNA guide, and processes its own CRISPR array. Because phages are not typically encountered by the human immune system in the same way as Streptococcus or Staphylococcus, CasΦ offers a potential immunogenicity advantage.
Cas12f Variants
The broader Cas12f (formerly Cas14) family includes numerous natural hypercompact editors in the 400–700 amino acid range. Many have been engineered by groups in Korea, China, and the US, with variants like enAsCas12f and Un1Cas12f now approaching SpCas9-level efficiency on several targets. For a broader look at this family, see our article on Cas12 and Cas12a editors.
Cas-CLOVER — Demeetra / Poseida
Cas-CLOVER is the odd one out: not a single-protein nuclease, but a dual-guide chimeric nuclease that combines a catalytically inactive Cas9 (dCas9) with the FokI nuclease domain — similar in logic to TALEN dimers. Because cutting requires two guides to bind adjacent sites and the two FokI domains to dimerize, Cas-CLOVER has exceptionally low off-target activity. It was licensed and commercialized by Demeetra AgBio and, most prominently, by Poseida Therapeutics, which uses it as the gene-editing backbone for its allogeneic CAR-T and liver gene therapy programs. Roche acquired Poseida in 2024 in a deal valued at up to $1.5 billion, validating Cas-CLOVER as a differentiated editor.
Comparison Table
| Editor | Source | Size (aa) | PAM | Notable developer |
|---|---|---|---|---|
| SpCas9 | S. pyogenes | ~1,368 | NGG | Editas, Intellia, CRISPR Tx |
| SaCas9 | S. aureus | ~1,053 | NNGRRT | Editas (EDIT-101) |
| Cas12a (Cpf1) | Acidaminococcus, etc. | ~1,200 | TTTV | Mammoth Biosciences |
| CasX (Cas12e) | Deltaproteobacteria | ~980 | TTCN | Scribe Therapeutics |
| CasΦ (Cas12j) | Bacteriophages | ~700 | TBN | Academic (Doudna lab) |
| Cas12f / CasMINI | Uncultured bacteria | ~400–700 | TTN / TTR | Academic, emerging startups |
| Cas-CLOVER | Engineered (dCas9+FokI, dual-guide) | ~1,600 (dual) | Flexible | Poseida, Demeetra |
Note that Cas-CLOVER is larger in total protein mass but delivers a different value proposition (fidelity), not compactness.
Evidence and Clinical Trials
Most compact editors are still in preclinical development, but the translational signal is building fast:
- Scribe Therapeutics / Biogen (2022–2026): Single-AAV CasX programs targeting SOD1 for ALS, currently IND-enabling.
- Poseida (Phase 1, 2023–present): Cas-CLOVER–engineered allogeneic CAR-T (P-BCMA-ALLO1) in multiple myeloma, with durable responses reported at ASH 2023 and 2024.
- Cas12f academic work: Multiple Nature Biotechnology and Cell papers (2022–2024) showing enhanced Cas12f variants achieving 40–80 percent editing in mouse liver and muscle via single AAV.
- CasMINI base editors: Qi lab and collaborators have published cytosine and adenine base editors built on CasMINI with competitive editing rates in human cells (2023–2024).
The field is approaching a tipping point where compact editors move from "interesting biology" to "preferred clinical modality" for any AAV-delivered program.
Applications
Compact editors unlock therapies that dual-AAV or lipid nanoparticle (LNP) strategies struggle with:
- Central nervous system. AAV9 crosses the blood-brain barrier, but the tight cargo budget has blocked CNS gene editing programs. Compact editors change that math.
- Retina. AAV is the standard for ocular delivery, and single-AAV editors enable tissue-specific promoters critical for limiting off-target expression.
- Skeletal muscle. Duchenne muscular dystrophy and other muscle indications benefit from muscle-specific promoters, which only fit alongside compact editors.
- Allogeneic cell therapy. Cas-CLOVER's high fidelity matters when editing multiple loci in donor T cells or iPSCs without introducing translocations.
- Autoimmune disease. Precise in vivo editing of immune cell populations is a major frontier, and single-AAV compact editors may be the enabler.
Connection to the Broader CRISPR Ecosystem
Compact editors are not a replacement for SpCas9 — they are a complementary layer in a maturing editor toolbox. SpCas9 remains the workhorse for ex vivo editing, where cargo size is irrelevant and LNPs handle delivery. Compact editors dominate where delivery is constrained: in vivo, AAV, and tissues outside the liver. Base editors and prime editors built on compact chassis (a particularly active research area in 2024–2025) promise the precision of base editing with the deliverability of AAV.
On the delivery side, compact editors pair naturally with cell-penetrating peptides and engineered virus-like particles for non-AAV strategies.
Limitations and Lessons
- Efficiency trade-offs. Natural miniature editors are often weak; engineered versions require years of directed evolution to reach SpCas9-like activity.
- PAM constraints. Stricter PAMs reduce targetable genome space, sometimes by 4–10×.
- Guide-RNA design rules are new. The software tools trained on SpCas9 do not transfer directly to Cas12f or CasX, slowing adoption.
- IP complexity. Every compact editor carries its own patent estate, adding commercial risk.
- Manufacturing. Novel proteins require fresh CMC work, including new assays, potency tests, and stability studies.
The lesson: size is necessary but not sufficient. The winners will be editors that combine compactness with competitive activity, permissive PAMs, and a clean regulatory path.
FAQ
Why does AAV have a 4.7 kb cargo limit?
AAV genomes are naturally about 4.7 kilobases, and packaging efficiency drops sharply beyond that size. Larger cargos either fail to package or produce truncated vectors.
Is CasMINI better than CasX?
They serve different niches. CasMINI is smaller (easier to fit accessory elements) but has fewer years of optimization. CasX has more clinical momentum via Scribe Therapeutics.
Can compact editors do base editing and prime editing?
Yes. Cytosine and adenine base editors built on CasX and CasMINI have been published. Prime editors are harder because they are inherently larger (reverse transcriptase fusion), but compact prime editor variants are an active research area.
Is Cas-CLOVER really compact?
No — Cas-CLOVER is notable for fidelity, not size. It is included here because it is a leading non-Cas9 editor in active clinical development for allogeneic CAR-T.
Are compact editors safer than SpCas9?
Potentially. Smaller proteins present fewer T-cell epitopes, and non-pathogen-derived editors may face less pre-existing immunity. But safety depends mostly on delivery, target, and off-target profile.
When will the first compact-editor therapy reach patients?
Poseida's Cas-CLOVER CAR-T programs are already in Phase 1. A single-AAV in vivo CasX or CasMINI therapy is likely 2–4 years from first-in-human trials.
