PINK1 Is Selectively Stabilized on Impaired Mitochondria to Activate Parkin

Parkinson's disease has no disease-modifying treatment. That's the clinical reality behind this story. For decades, one of the most compelling clues about why neurons die in Parkinson's has pointed to mitochondria — the cell's energy-generating organelles. Mitochondria in the substantia nigra, the brain region most devastated in Parkinson's, accumulate DNA mutations at higher rates than elsewhere. Toxins that cripple mitochondrial energy production can produce a Parkinson's-like syndrome. The evidence kept pointing in the same direction. Then genetics handed researchers two proteins: PINK1, a kinase that is located on mitochondria, and Parkin, an E3 ubiquitin ligase that normally floats in the cytosol. Loss-of-function mutations in either gene cause early-onset inherited parkinsonism. Work in Drosophila showed something even more specific: overexpressing Parkin can partially rescue the effects of losing PINK1, but the reverse is not true. This asymmetry suggested an order: PINK1 works upstream of Parkin. But what does "upstream" actually mean at the molecular level? That was the unresolved question. Earlier work by Narendra and colleagues established that Parkin does not go to mitochondria randomly. It specifically targets damaged, depolarized mitochondria and promotes their destruction through a process called mitophagy — selective autophagy of mitochondria. In other words, the cell has a garbage collection system that can identify individual broken organelles and eliminate them. The puzzle was in the recognition step. How does Parkin know which mitochondria are damaged? The answer, as Narendra and colleagues showed, is PINK1. The mechanism is elegant in its simplicity. On a healthy mitochondrion, PINK1 is essentially invisible. The protein is continuously synthesized, imported into the mitochondrion, and immediately cleaved. Full-length PINK1 sits at around sixty-three kilodaltons, but on polarized mitochondria, it gets proteolytically processed down to a shorter fifty-two kilodalton fragment, which is then rapidly degraded by the proteasome. The net result is that PINK1 barely accumulates. It is produced and destroyed in a continuous cycle, maintaining near-zero steady-state levels on any healthy mitochondrion. The critical variable is membrane potential — the electrochemical charge that healthy mitochondria maintain across their inner membrane. That charge drives the cleavage. When a mitochondrion loses its membrane potential, the cleavage stops. Full-length PINK1 can no longer be processed, and it begins to accumulate rapidly on that specific organelle. Narendra and colleagues demonstrated this using CCCP, a chemical that collapses mitochondrial membrane potential. Within thirty minutes of CCCP treatment in HeLa cells, endogenous full-length PINK1 was already detectable, and it continued rising for at least three hours. Live imaging of PINK1 tagged with a fluorescent protein showed the same kinetics, with the signal appearing between one and five minutes after depolarization. Critically, this is a post-translational effect, not a transcriptional one. Quantitative PCR showed no significant change in PINK1 mRNA after an hour of CCCP treatment. The cell is not making more PINK1 message; it just stopped destroying the protein it was already making. The selectivity for damaged mitochondria is striking. In cells with heterogeneous mitochondrial populations — some healthy and some not — PINK1 accumulated specifically on the depolarized ones. The average correlation coefficient between PINK1 and a dye that marks bioenergetically active mitochondria was only 0.26, while the correlation between PINK1 and cytochrome c, which marks all mitochondria regardless of health, was 0.58. PINK1 was reading the damage signal and responding to it organelle by organelle. So, full-length PINK1 builds up on damaged mitochondria. The next question is what that buildup actually does. The answer is that it recruits Parkin. Narendra and colleagues showed this is both necessary and sufficient. On the necessity side: in primary mouse embryonic fibroblasts lacking PINK1, depolarization with CCCP recruited Parkin to mitochondria in forty-three percent of wild-type cells but zero percent of PINK1-null cells. Restoring PINK1 expression brought recruitment back to seventy-three percent. In human M17 neuroblastoma cells, stable PINK1 knockdown dropped CCCP-induced Parkin translocation from sixty-seven percent in controls to under five percent. No PINK1 means no Parkin recruitment. The signal chain is broken at the first link. On the sufficiency side: artificially stabilizing PINK1 on the outer mitochondrial membrane was enough to recruit Parkin even without any depolarization. When Narendra and colleagues fused a truncated PINK1 — one that bypasses the normal voltage-sensitive import and cleavage — to a protein that anchors it to the outer membrane, Parkin translocated to mitochondria in ninety-eight percent of cells, with no CCCP required. A completely orthogonal approach using chemically induced protein dimerization, acutely tethering PINK1 to the outer membrane via a rapalog compound, produced the same result: ninety-seven percent of cells showed Parkin on mitochondria. The depolarization itself isn't what matters. What matters is PINK1 accumulating on the outer membrane. That's the flag. Two features of PINK1 are required for it to function as that flag. First, it must be properly targeted to mitochondria — a truncated PINK1 missing its mitochondrial targeting sequence failed to rescue Parkin recruitment in null cells. Second, its kinase activity must be intact — a kinase-dead version of PINK1 also failed to rescue recruitment, even though the protein was processed normally. PINK1 needs to be in the right place and catalytically active to signal Parkin. With the core mechanism established, the disease-causing mutations become a natural experiment. They reveal something important: PINK1 and Parkin mutations don't break the pathway in the same way. They disrupt it at different steps. PINK1 mutations act at the upstream recruitment step. A kinase-deficient PINK1 cannot signal Parkin, so Parkin never arrives at the mitochondrion. The pathway stops at the first handoff. Parkin mutations are more varied. Some affect recruitment itself — deletion of Parkin's ubiquitin-like domain caused a moderate reduction in how efficiently Parkin was recruited after depolarization, though it didn't abolish recruitment entirely. The most revealing mutations are those that allow Parkin to arrive at the mitochondrion normally, then fail downstream. The R275W mutation in Parkin's RING1 domain caused only a minor recruitment deficit but severely disrupted mitophagy. Parkin got to the damaged mitochondrion but couldn't do the job once it arrived. That disconnection — between docking and execution — proves these are genuinely separate steps in the pathway, not just two descriptions of the same event. This maps cleanly onto the fly genetics. The reason Parkin overexpression can compensate for PINK1 loss in Drosophila is that flooding the cytosol with Parkin can partially bypass the defective recruitment signal. The reason PINK1 overexpression cannot compensate for Parkin loss is that no amount of flagging a mitochondrion matters if the executor is broken. Narendra and colleagues provide the biochemical explanation for a genetic relationship that had been observed but not understood. Stepping back to the larger picture: what this work describes is a quality control circuit. PINK1 is the sensor, reading membrane potential continuously across the mitochondrial network. When an organelle's potential collapses, PINK1 accumulates on its surface. That accumulation recruits Parkin from the cytosol. Parkin then marks the damaged mitochondrion for autophagic destruction. The cell removes the broken part. In neurons that never divide — like the dopaminergic cells of the substantia nigra — this circuit runs for a lifetime. There's no dilution of damaged mitochondria through cell division. The quality control pathway is the only mechanism of negative selection available. When PINK1 or Parkin is broken, damaged mitochondria are no longer recognized or cleared. They accumulate, generating oxidative stress and depressing local metabolism. Neurons with exceptionally high energy demands and calcium buffering burdens — like substantia nigral dopaminergic neurons — would be especially vulnerable to that accumulation. The tissue-specific pattern of Parkinson's disease becomes, at least in part, a story about which cells can least afford a broken garbage collection system. Two clear questions remain. The protease responsible for PINK1 cleavage in mammalian cells is still unidentified. PARL, the mammalian rhomboid protease related to the Drosophila enzyme Rhomboid-7, was tested and found dispensable — PINK1 levels were similar in PARL-null and wild-type cells. It remains mechanistically open exactly how PINK1's kinase activity drives Parkin recruitment — whether through direct phosphorylation of Parkin, modification of an adaptor, or some other substrate. Those are precise, tractable gaps. But the central architecture is clear: PINK1 reads the damage signal, Parkin reads PINK1, and together they ensure that broken mitochondria do not outlive their usefulness in cells that cannot afford to keep them. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.