Mitochondrial Genome Editing – Comprehensive Study Notes

Introduction & Background

• Mitochria = organelles responsible for oxidative phosphorylation (OXPHOS) and stress response.
• They harbour their own genome – mitochondrial DNA (mtDNA).
• Human mtDNA: $16,569$ bp, double-stranded, circular, encodes $37$ genes:
  • $13$ protein subunits of the OXPHOS complexes.
  • $2$ rRNAs.
  • $22$ tRNAs.
• Copy number is cell-type dependent (up to $100,000$/cell).
• Energy-demanding tissues (brain, heart, muscle, etc.) are most vulnerable to mtDNA defects.

mtDNA Mutations & Disease Spectrum

• ~$100$ pathogenic mtDNA mutations identified:
  • $95$ point mutations.
  • $2$ deletions, $2$ insertions, $1$ inversion.
• Prevalence ≈ $1:5,000$ adults.
• Representative disorders: Leigh Syndrome, MELAS, LHON, MERRF, etc.
• Mutation states:
  • Heteroplasmy – mixture of mutant & wild-type genomes.
  • Homoplasmy – uniform genome.
• Threshold effect: clinical symptoms appear only when mutant load exceeds tissue-specific threshold.

Heteroplasmy Dynamics

• During mitosis/meiosis mutant:wild-type ratios drift.
• Lowering mutant load below threshold can rescue phenotype (therapeutic goal).

mtDNA Repair Pathways

• DNA repair proteins are nuclear-encoded, imported with mitochondrial-targeting sequence (MTS).
• Key pathways:
  • Base Excision Repair (BER) – primary; damaged base → glycosylase → AP site → short/long patch fill-in.
  • Mismatch Repair (MMR) (less characterised).
• Double-strand breaks (DSBs): inefficiently repaired → damaged genomes degraded → remaining genomes replicate to restore copy number.

Rationale for Genome Editing

• Conventional CRISPR requires guide RNA – mitochondrial import inefficient, so protein-only editors are used.
• Categories:
  • Nucleases (create DSB → selective elimination).
  • Base editors (C→T or A→G changes without DSB).

Protein Targeting Strategy

• N-terminal MTS fused to editing protein ensures import into matrix.
• Often combined with nuclear export signal (NES) to minimise nuclear residence.

mtDNA Nucleases (Elimination Approach)

• Structure: DNA-binding domain + cleavage domain (usually FokI).
• Tools & notes:
  • Restriction endonucleases (mito‐REs) – fixed sequence sites.
  • ZFNs – zinc finger arrays + FokI; dimeric or single-chain.
  • TALENs – TALE arrays + FokI; dimeric or single-chain.
  • mitoTev-TALE – TALE + homing endonuclease I-TevI.
  • mitoARCUS – meganuclease platform.
• Mechanism: mutant-specific cleavage → degraded → wild-type genomes repopulate (Fig.1 left concept).
• Limitations:
  • Requires sequence divergence; less useful for homoplasmic mutations.
  • Cannot create novel mutations.

Cytosine Base Editing – DdCBE

• Discovery: DddAtox (Burkholderia cenocepacia toxin) deaminates C in dsDNA.
• Safety engineering: split at G$1333$ or G$1397$ → halves are fused to TALE arrays + UGI + MTS.
• Upon adjacent binding, heterodimer reconstitutes enzyme → deaminates C in $\text{TC}$ motif → U (retained by UGI) → T after replication (Fig.1 right).
• Editing window: spacer between TALEs; typically ~$4$–$6$ nts.

Expanding Sequence Scope

• Directed evolution (PANCE/PACE) → variants:
  • DddA6 (better TC).
  • DddA11 (HC preference, wider window).
• Homolog/ortholog mining:
  • DddSs (FZY2) → DdCBESs (DC motif).
  • Q2L7-DdCBE (GC preference).
  • RsDdCBE, mitoCBE2.0 (Ri) → NC compatibility.
• AI-guided discovery: Ddd1, Ddd7, Ddd8, Ddd9 (varied motif biases).

Size Reduction

• Zinc Finger Deaminase (ZFD): ZF array + split DddA + UGI → smaller.
• ZF-DdCBE: architecture-optimized; fits single AAV9.
• GSVG full-length variant → monomeric DdCBE (mDdCBE) – further size and delivery advantage.

Off-Target Mitigation

• Nuclear genome hits observed (GOTI, Detect-seq).
• Strategies:
  • C-terminal NES fusion (UGI-NES-DdCBE).
  • Co-express NLS-DddIA inhibitor (binds leaking enzyme).
  • Interface mutations (K$1389$A, T$1391$A, V$1411$A) → HiFi-DdCBE (reduces spontaneous assembly).
  • ZFQQ variant (R(-5)Q) and HS-ZF-DdCBE (charged truncations, catalytically dead N half).

Adenine Base Editing – TALED

• Uses TadA8e (ssDNA A-deaminase) + split DddAtox to transiently reveal ssDNA.
• Architectures:
  • split-TALED (sTALED, dimer).
  • dual-TALED (dTALED).
  • monomeric-TALED (mTALED).
• Converts A→I→G; editing window similar to DdCBE.

Strand-Selective Editors

• Limitation of double-strand editors ⇒ bystanders on both strands.
• mitoBEs: nickase (MutH / MutH* for $\text{GATC}/\text{GAT}$; Nt.BspD6I(C) no motif) + TadA8e or rAPOBEC1-UGI + TALE.
  • A-to-G (mitoABEMutH / mitoABENt.BspD6I).
  • C-to-T (mitoCBEMutH / mitoCBENt.BspD6I).
  • Can be monomeric or dimeric (AAV-friendly).
• CyDENT: TALE + FokI nickase + exonuclease + ssDNA cytidine deaminase → strand-preferred C editing.

Editing Window & Bystander Control

• Window shaped by:
  • Distance between DNA-binding sites.
  • Catalytic domain mutations (e.g., DddA11 widens; V$1411$A narrows).
• Strategically placing TALE or ZF binding sites refines window to avoid bystanders.

Delivery Platforms

• Non-viral (in vitro): lipofectamine, PEI, lipid nanoparticles, polyplexes.
  • Forms delivered: plasmid DNA, mRNA, circRNA.
• Physical:
  • Electroporation – common for cultured cells & zygotes.
  • Microinjection – embryos (mouse, rat, zebrafish, human 3PN/2PN) → germline editing.
• Viral: AAV (≈$4.8$ kb payload).
  • Dual-AAV for large constructs (TALE pairs).
  • Single-AAV for compact ZF-based or mDdCBE editors.
  • Tissue tropism via serotypes/promoters.
  • Safety: generally safe but high-dose toxicity reported.

Research & Therapeutic Applications

• Nucleases:
  • Shift heteroplasmy in patient cells (MELAS, MERRF, etc.).
  • In vivo rescue in NZB/BALB and m.$5024$C>T mice via mitoTALEN or mitoARCUS (AAV delivery).
• Base editors:
  • Generate disease models (mouse, rat, zebrafish, plant).
  • Correct point mutations in cell lines & embryos.
  • Library approaches allow systematic knockout of all $13$ mt-encoded OXPHOS genes.
• Example in vivo successes:
  • AAV-NES-TALED treated mice – efficient A→G edits in heart/liver/muscle.
  • Single-AAV9 ZF-DdCBE delivered to post-natal mice – modelling tRNA mutations.

Current Limitations & Future Directions

• Editing confined to transition mutations (C↔T, A↔G).
• RNA delivery barrier prevents CRISPR, prime editing, transversions.
• Off-target surveillance in nuclear genome remains critical.
• Need for improved delivery vectors with larger cargo or split-intein reconstitution.
• Potential of CRISPR import or engineered RNA chaperones; if solved, would unlock full suite of CRISPR technologies in mitochondria.
• Ethical considerations: germline editing, embryo experiments, off-target mutagenesis.

Practical & Philosophical Implications

• Therapeutic promise for untreatable mitochondrial disorders once safety & specificity are optimised.
• Insight into aging and metabolic diseases through precise modelling.
• Highlights interplay of nuclear-mitochondrial co-evolution: editing mtDNA may require concurrent nuclear adaptation.
• Raises questions about heritable editing, consent across generations, and equitable access.