Supercomplex assembly, disassembly and biogenesis

Sousa, Mills, Vonck and Kühlbrandt (2016) discussed a study of respirasome biogenesis by Moreno-Lastres and colleagues (2012). That study suggested that supercomplexes form by association of individual subunits or intermediate assemblies, rather than from fully assembled complexes². It presented incorporation of the NADH:dehydrogenase module of complex I as the final step³. It also reported a significant delay in integrating the 51 kDa subunit, a core subunit of the N-catalytic unit of complex I³. Sousa and colleagues described these observations as in agreement with their own cryo-EM data on the bovine respirasome³.

In the bovine respirasome work, about 30% of the initial dataset fell into classes with no density for the NADH:dehydrogenase module, or without the whole matrix arm. The authors allowed that this might be partial loss during purification or grid preparation. They did not rule out that these are assembly, disassembly or recycling intermediates¹. They gave a second reading. Several nuclear-encoded subunits of mature complex I can be exchanged for newly imported copies, alongside de novo assembly⁴. Most N-catalytic unit subunits are particularly susceptible to oxidative damage, so direct replacement could repair them⁴. The classes could therefore reflect complex I regeneration⁴.

Protasoni and Zeviani (2021) reviewed assembly of the other respiratory complexes and described it as modular. For ATP synthase, three sub-assemblies form individually and are then put together: the F1 module, the c-ring and the peripheral stalk⁸. For cytochrome c oxidase, the MTCO1 module forms in parallel with other parts. Its first binding factors are COX14 and COA3, which assist the subunit during and after membrane insertion⁷.

Caruana and Stroud (2020) noted a first structure of a complex I assembly intermediate. A homologue of the assembly factor NDUFAF2 was bound to an incomplete complex from cells lacking NDUFS6. The authors of that structure concluded that NDUFAF2 occupies the position later taken by NDUFA12 and blocks NDUFS6 binding⁶. Fedor and Hirst (2018) tested a possible function of the finished assemblies and found that quinone is not channeled within supercomplexes but acts as a common pool⁵. Rao and colleagues (2026) reported that cristae collapse impairs supercomplex assembly⁹.