STRC Research

Prime Editing for STRC

8 min read

Prime Editing for STRC E1659A

Core claim

Instead of replacing the entire STRC gene (5,325 bp), correct just the mutated base. Misha’s maternal allele carries c.4976A>C (p.E1659A) — a missense substitution. STRC is on the minus strand of chr15. The coding-strand mutation is A→C (mutant C at chr15:43600551). Correcting it requires C→A on the coding strand, or G→T on the + strand. Only prime editing (PE3/PE3b) can make this specific edit among currently published tools. PAM survey completed April 2026 — one optimal PAM confirmed.

Why standard base editors don’t work

STRC is on the minus strand. At chr15:43600551:

    • strand (reference): T (WT), G (mutant)
  • Coding (minus) strand: A (WT), C (mutant)
  • Required correction: C→A on coding strand = G→T on + strand
EditorActionVerdict
CBE (C→T)Acts on C. Mutant coding strand has C. Gives C→T (coding = T, not A).Wrong base
ABE (A→G)Acts on A. No A at the mutant position on either strand.Cannot apply
ACBE (A→C, Chen et al. 2024)Acts on A and converts to C. No A at the mutant position.Cannot apply (also wrong direction — would CREATE this mutation)
CGBE (C→G)C on coding strand → G.Wrong base
PE3/PE3bInstalls any base via reverse transcription from pegRNA template. Corrects C→A on coding strand.Only viable precision editor

PAM survey — confirmed April 2026

Genomic sequence fetched from Ensembl REST API (chr15:43600500–43600620, GRCh38). SpCas9 NGG PAMs searched within ±30 bp of variant on both strands.

PAM positionStrandPAM seqDistance to variantWindow
chr15:43600540+TGG14 ntOptimal (3–16 nt)
chr15:43600570−NGG22 ntExtended
chr15:43600571−NGG23 ntExtended
chr15:43600572−NGG24 ntExtended
chr15:43600573−NGG25 ntExtended

Best PAM: TGG at chr15:43600540 (+ strand), 14 nt from the variant. This is the sweet spot. Nick falls at chr15:43600537. The pegRNA RT template extends downstream and delivers the A correction at position 4976 in a single step. PE3b strategy: nicking sgRNA targets the non-edited strand to improve efficiency and suppress indels.

4 additional PAMs on the minus strand in the 22–25 nt extended range are usable with reduced efficiency.

Genomic coordinates

  • GRCh38: chr15:43600551 (STRC exon 29)
  • GRCh37: chr15:43892749

What the literature actually shows

Prime editing in cochlear cells

Zero published papers. As of April 2026 no study has applied prime editing to OHCs or any inner ear cell type. This is a genuine gap.

Closest cochlear analogue: base editing

Wang et al. 2025 (Nature Communications, PMID 40968144) delivered ABE (SchABE8e) via Anc80L65 AAV to neonatal mice targeting a POU4F3 Q113* stop codon (DFNA15 model). Near-complete hearing recovery sustained 4+ months. [⚠ Cited as “Zhang 2025” in prior h07 notes — first author is Man Wang, not Zhang. The co-author list includes a Zhang (Ziyu Zhang), explaining the confusion. PMID 40968144 confirmed. Corrected 2026-04-25.] Demonstrates the full chain: AAV → OHCs → precise base correction → functional rescue. Wrong edit class for Misha’s mutation, but the delivery and cell-access proof is directly relevant.

LNP-CBE cochlear delivery — citation corrected 2026-04-23. The prior “Gao et al. 2020 (Sci Transl Med, PMID 32493791)” citation was PHANTOM: PMID 32493791 is “COVID-19 diagnostics in context” (STM 2020), unrelated to cochlear gene therapy. The 0.3-0.6% Organ-of-Corti editing figure is unverified by any single primary paper the author has located (2026-04-23 audit). Candidate real sources in the same claim space: Yeh et al. 2018 Nat Biomed Eng on disease-specific base-editor cochlear delivery, or Gao et al. 2018 Nature 553:217 (Beethoven Tmc1 lipid-Cas9 RNP, not LNP-mRNA). Until a specific paper is pinned, treat “LNP-CBE achieves sub-percent cochlear editing” as a qualitative direction, not a quantified number. LNP remains the wrong vector for PE regardless.

PE in post-mitotic cells (best proxy for OHCs)

OHCs are terminally differentiated and never divide. HDR (used by most CRISPR approaches) approaches zero in post-mitotic cells. PE does not require HDR — it uses nick + reverse transcription, which is mechanistically compatible with non-dividing cells.

Best in vivo data:

  • Chen et al. 2024 (Nature Biotechnology, PMID 37142705): Dual-AAV split-intein PE in adult mouse cortex. Up to 42% editing efficiency in post-mitotic cortical neurons. 46% in liver, 11% in heart. Closest tissue analogue available.
  • Chemla et al. 2025 (Mol Therapy Nucleic Acids, PMID 41210585): PE4 in iPSC-derived cardiomyocytes (post-mitotic). 34.8% average efficiency. Functional phenotypic rescue confirmed.
  • Anzalone et al. 2019 original paper: confirmed PE works in primary post-mitotic mouse cortical neurons (low frequency, no number given).

ACBE as alternative (single-AAV)

Chen et al. 2024 (Nature Biotechnology, PMID 37322276) introduced ACBE. [⚠ Previously cited as “Kim 2023” — no author named Kim in the paper. First author is Liang Chen (Li lab + Liu lab). Published 2024, epub June 2023, explaining the year confusion. Corrected 2026-04-25.] Lab-evolved mAAG-TadA8e-nCas9 achieves up to 73% A→C in cell culture, 44–56% in mouse embryos. Key advantage: smaller than PE3 (~130 kDa vs ~180 kDa), potentially fits in single AAV. No in vivo post-mitotic data published yet.

For Misha’s variant: ACBE goes the wrong direction (A→C creates the mutation). But for other DFNB16 patients with different variant classes, ACBE single-AAV may be the simpler path.

Dual-AAV split-intein delivery (PE-specific)

PE3 (SpCas9 + RT fusion) encodes ~6.3 kb — too large for a single AAV (~4.7 kb limit). Solution is established:

  • Zhi et al. 2022 (Mol Therapy 30:283, PMID 34298129): Split PE delivered by dual-AAV8, active editor achieved editing in adult mouse retina. [🚨 Previously cited as “Villiger 2021” — wrong author. The paper at PMID 34298129 is Zhi et al., epub July 2021, published January 2022. No “Villiger 2021” dual-AAV split PE retina paper found in PubMed. Corrected 2026-04-25.]
  • Chen et al. 2024: same split-intein architecture achieved 42% in cortical neurons.
  • Shubina-Oleinik et al. 2021 (Sci Adv, PMID 34910522): Dual-AAV9-PHP.B STRC delivery to OHCs; restored STRC expression, normalized hair bundle, cochlear amplification substantially enhanced. [⚠ Previously cited as “Fang et al. 2021” — first author is Olga Shubina-Oleinik (Holt lab, Harvard). No author named Fang in the paper. The “50% DPOAE recovery” figure is an interpretation from h07 prose; exact numbers in PMC8673757. Corrected 2026-04-25.] Co-delivery precedent for this exact tissue.

Realistic efficiency estimate for OHCs

No OHC data exists. Extrapolating from best analogues:

  • Cortical neurons (post-mitotic, AAV-accessible): 42%
  • Cardiomyocytes (post-mitotic, ex vivo): 35%
  • Cochlear ABE (wrong class, but same tissue/vector): near-complete functional recovery

Plausible range for OHC prime editing: 15–40%. Not verified. This is the key unknown experiment.

Prime editing vs gene replacement — parallel tracks

Shubina-Oleinik et al. 2021 and Iranfar 2026 both use dual-AAV STRC cDNA replacement. This works for loss-of-function alleles regardless of the causative mutation. For Misha:

  • Paternal allele: large STRC deletion — gene replacement is the only option
  • Maternal allele (c.4976A>C): gene replacement also covers this (provides functional STRC)

Gene replacement is simpler and further clinically validated. Prime/ACBE editing is theoretically more elegant (corrects endogenous allele, no ectopic integration) but at earlier TRL for OHCs. These are parallel tracks, not competitors.

Mutation visualization

The substitution creates a visible chemical difference at position 1659:

  • Glu (E): -CH₂-CH₂-COO⁻, negatively charged carboxylate, two H-bond acceptors
  • Ala (A): -CH₃, nonpolar methyl, no charge, no H-bond donors/acceptors
  • Volume change: 138.4 ų → 88.6 ų (49.8 ų cavity created)
  • Hydrophobicity: −3.5 (hydrophilic) → +1.8 (hydrophobic)

This is precisely why the mutation is pathogenic. See STRC Electrostatic Analysis E1659A.

Open questions

  1. PAM survey for ACBE/SpRY/SaCas9 — does NG (SpRY) or NNGRRT (SaCas9) extend the toolkit?
  2. Editing efficiency in OHCs — animal experiment required. Dual-AAV PE vs ACBE vs cochlear base editing.
  3. Which allele matters more — Misha’s paternal deletion requires gene replacement regardless. Does correcting the maternal VUS allele provide additive benefit?
  4. Age window — Wang et al. 2025 and Shubina-Oleinik et al. 2021 used neonatal injection. At Misha’s age (~5 years), OHC AAV transduction is substantially lower.

Status

AspectStatus
PAM site feasibility✅ Confirmed — TGG at 14 nt
pegRNA design✅ Phase 1–3 complete 2026-04-21. Phase 1 STRC PE Phase1 pegRNA E1659A: spacer GCCCAGCTCCCCACCTGCTA, RT 20 nt, PBS 13 nt, nick-to-edit 14 nt. Phase 2 STRC PE Phase2 PAM Expansion: PE3b filter bug fixed; 5 PE3b nickers for SpCas9 NGG, SpG NGN removes geometry penalty. Phase 3 STRC PE Phase3 Allele Discrimination: Phase 2 lead reclassified as WEAK (position-5 distal); true SpCas9 seed lead = CCTGAGATCTTCACTGAAAT (position 17, nick 0.5 nt), balanced SpG lead = TTCACTGAAATTGGCACCAT (position 8, nick 9.5 nt). Revised efficiency: 10–30% SpCas9+seed-PE3b, 20–40% SpG+mid-PE3b.
Active STRC PE programNone known as of 2026-04
Relationship to AAV gene therapyOrthogonal: corrects the mutation vs. delivering a functional copy
Timeline to clinicalUnknown; PE in Phase I for other conditions (e.g., BEAM-101 for sickle cell)

Connections

Graph View

Backlinks