Detergent solubilisation and purification of respiratory complexes

Kühlbrandt (2015) records that respiratory chain supercomplexes were first postulated from blue-native gels of yeast and bovine heart mitochondria solubilised in digitonin¹. Sousa and colleagues (2016) state that digitonin has been the detergent of choice since the supercomplexes were discovered by Schägger and Pfeiffer in 2000³. They add that density gradient centrifugation was the preferred purification method. Separating the two digitonin-solubilised supercomplexes on a preparative scale was hard, because their densities differ little³. In their own preparation, bovine heart mitochondria were solubilised with 1% digitonin at a detergent-to-protein ratio of 28:1, or with a different detergent, PCC-a-M, at 1:1⁴.

Kühlbrandt (2015) notes that respirasomes were long thought to be artifacts of solubilisation, despite their well defined structure². He cites genetic evidence for their existence in vivo and recent cryo-electron tomography showing them in cristae². Brzezinski and colleagues (2021) argue that III2IV1/2 supercomplexes in S. cerevisiae are held by weak interactions. They point to the need for weak detergents such as digitonin or glyco-diosgenin, and to dissociation on adding dodecyl-maltoside⁸. Protasoni and Zeviani (2021) report that the share of free complex I differs by method. Tomography of bovine heart found only 56% associated with other complexes⁹. They also cite blue-native gel estimates of 14-16% free complex I in the same tissue. They say such differences might be due to various reasons, including artefacts introduced during sample preparation or actual differences between organisms⁹.

Fedor and Hirst (2018) tested whether supercomplexes enhance catalysis by quinone channelling. Their added enzyme, AOX, needs dodecyl-maltoside to stay in solution. Adding more AOX to membrane vesicles therefore also adds detergent, which could disrupt the supercomplex or the membrane⁵. Reverse electron transfer is a test of membrane integrity. Its rate fell sharply above 0.25 mg AOX or 0.17 mg DDM per mg, well above the 0.1 mg AOX needed for maximal catalysis⁶. Individual enzyme activities stayed high, which they read as failure of the membrane. They concluded that the detergent carried in with the AOX caused the failure⁶.

Letts and colleagues (2019) reported a practical difference between detergents when purifying complex I. The yield from the first anion exchange step was lower after extraction with the branched chain detergent LMNG than with single-chain dodecyl-maltoside⁷. In LMNG, much complex I did not elute until the high salt wash. A modified gradient gave two peaks of complex I activity, and the second also absorbed at 420 nm, indicating heme⁷. Rerunning that peak on anion exchange or size exclusion separated it into complex I and a complex III dimer, confirmed by mass spectrometry⁷. Sousa and colleagues exchanged the detergent for amphipol A8-35 during purification⁴.