Finding. Respiration falls in a straight line as mtDNA-deletion heteroplasmy rises. Redox imbalance does not. The integrated stress response and translation failure wait until heteroplasmy is past about 70%, and they reverse if you give asparagine or raise pyruvate carboxylase. Cells that cannot respire and cannot express much PC become asparagine auxotrophs. In a high-heteroplasmy, low-PC thyroid tumor, L-asparaginase slows growth.
Why this paper matters
Mitochondrial genetics is a dose problem. The same deletion can be silent or lethal depending on mutant-genome fraction. Cui, Ryu, Fu, Sfeir, Thompson and colleagues use isogenic deletion lines to pull that dose apart into separate physiological layers.
Linear respiration loss is the expected titration. The interesting result is the delay: redox stays relatively stable until heteroplasmy crosses ~50%, and ISR/translation break later still. That means “more mutant mtDNA” is not a single stress. It is a sequence. Models that bolt ISR directly to any OXPHOS dip will mis-time the biology.
The metabolic rescue
High-heteroplasmy defects reverse with asparagine or with PC overexpression. That points at anaplerosis and non-essential amino-acid synthesis, not at a mysterious respiratory subunit. When the ETC cannot reoxidize NADH or supply aspartate/oxaloacetate efficiently, pyruvate carboxylase becomes the workaround that keeps asparagine production — and therefore translation — alive.
The authors then invert the logic. Across other respiration-deficient cells, exogenous asparagine dependence tracks inversely with PC level. Patient-derived thyroid tumor cells with a high-heteroplasmy Complex I mutation and little PC are asparagine-auxotrophic; asparaginase suppresses tumor growth. That is a rare case where a mitochondrial genotype plus a matrix enzyme predicts a licensed metabolic drug.
Caveats
Thresholds belong to this deletion series until someone repeats them in other genotypes. The abstract does not prove that mitochondrial PC, rather than a cytosolic pool, is the required enzyme, nor that asparagine is the only missing product. Asparaginase has a brutal systemic history; one tumor model is a lead, not an oncology program.
What to do with it
If you score mtDNA-disease papers, this one is a conceptual upgrade: heteroplasmy thresholds, ISR timing, and anaplerotic suppression of auxotrophy. If you score cancer metabolism, add PC-low / OXPHOS-low / high-heteroplasmy as an asparaginase hypothesis. If you build agents that mine preprints for drug–biomarker pairs, the pair is L-asparaginase × PC expression in respiration-deficient cells.