Precision Mitochondria through a Layered Onion Nanocarrier for Matrix Delivery and Expression of Engineered Human mtDNA

Mitochondrial genome engineering remains constrained by a fundamental unsolved problem: the inability to deliver intact exogenous DNA across both mitochondrial membranes into the matrix and support autonomous expression within the organelle. Here we present a computational proof of concept for a staged six-layer “onion” nanocarrier designed to address this challenge through sequential barrier resolution. The architecture combines an outer ionizable lipid nanoparticle for endosomal escape, a triphenylphosphonium targeting layer for mitochondrial accumulation, a fusogenic outermembrane lipid layer, a protective metal-organic framework for controlled uncoating, an artificial cardiolipin-enriched inner-membrane vesicle with OPA1-inspired fusion logic, and a TFAM-condensed engineered mitochondrial DNA cargo that mimics a synthetic nucleoid. We simulated 5,000 candidate architectures spanning delivery, fusion, matrix localisation, reporter expression, persistence, and viability variables, and analysed the resulting design space using milestone gating, supervised machine learning, regression, and unsupervised clustering. End-to-end success emerged in 31.66% of architectures after cumulative attrition across six milestones. Nuclease resistance, metal-organic framework stability, nucleoid size, and inner-membrane content mixing were the strongest determinants of overall success, whereas matrix delivery showed a monotonic association with reporter output across orthogonal luminescent and fluorescent reporters. A high-performing cluster of architectures combined strong matrix delivery with preserved viability, supporting the feasibility of down-selecting balanced designs rather than toxicity-driven extremes. Although experimental validation remains necessary, these results define a quantitatively structured framework for mitochondrial DNA delivery and expression, identify innermembrane traversal as the dominant bottleneck, and suggest that mitochondrial genome engineering may become tractable when approached as an ordered systems-delivery problem rather than as an extension of nuclear gene therapy or editing alone.