Finding. MICOS does not assemble by piling subunits. Mic60 has to open. Daumke, van der Laan, Liu and colleagues solve a Chaetomium thermophilum Mic60-Mic19 crystal structure and then beat on it with modelling, mutagenesis, cross-links, and hydrogen-deuterium exchange. An extended helical region in the Mic60 mitofilin domain, plus a Mic19 helix, is the handshake. Dynamics say that helix is a switch. Mic60 starts as a closed dimer. Bound Mic19 takes it to a hetero-octamer. That transition is their activation model, and they treat it as the prerequisite for building a crista junction.
Why this paper matters
Crista junctions are how the inner membrane keeps respiratory-chain neighborhoods without sealing them off. MICOS is the conserved machine. Structures of pieces have piled up; an activation step has not. A named switch helix you can mutate is a biochemical object. If the closed dimer cannot bind Mic19, junctions should fail. That is a test, not a slogan.
What they actually measured
A crystal structure, an interface, HDX dynamics, cross-links, mutants. The octamer is a model that comes out of that pile, not a single-particle reconstruction of a junction in a mitochondrion.
How to read the score
Low 90s. Primary mitochondrial ultrastructure mechanism. Confidence is high for the interface, medium-high for the cellular activation story as stated.
Caveats
Fungal protein. No in-cell crista morphometry in the abstract. Do not draw a human disease Mic60 mutant onto this helix without the alignment.
What to do with it
If you mutate MICOS, start with the EHR and the Mic19 helix. If you model crista biogenesis, add a closed-dimer to octamer step before Mic10 arrives. Pair it with this week's complex I papers; junctions are the rooms those enzymes sit in.