#bicelles

 Chemists at the University of Oxford, UK, have developed a way to look at the cell's membrane lipids without using detergents — charged molecules which can interfere with exactly those structures they are being used to visualise. 

In the illustration above: Lipid heads at the top and bottom of the bilayers (purple dots); the protein belt used to protect lipid tails from coming into contact with water (cyan); and the membrane protein encapsulated in the centre (red). The disc encased in a charged water droplet (right) undergoes evaporation to form the naked disc (left) which can then be studied in a mass spectrometer.

Membrane proteins are key to many biological processes and comprise half of all current drug targets. They are notoriously difficult to study in their natural environment (see my previous post on reverse micelles), but this new technique has allowed just that — combining the use of sophisticated nanodiscs with mass spectrometry.

some detergents can promote unfolding, and they do not mimic the lateral forces and curvature of the cellular membrane that can be important for maintaining protein structure... nanodiscs and bicelles that employ small, discoidal arrangements of phospholipid bilayers have demonstrated great potential for X-ray crystallography, nuclear magnetic resonance (NMR) spectroscopy and electron microscopy.

The study confirmed that "oligomeric complexes or proteins requiring defined lipid environments are stabilized to a greater extent in the absence of detergent". The team set themselves a challenge in selecting proteins known to be tricky analytes:

To test the preservation of oligomeric state, we analyzed E. coli diacylglycerol kinase (DgkA), a trimeric cytoplasmic membrane protein9 whose activity and oligomeric state varies based on its preparation10 and the presence of lipids11 . The third protein we studied, sensory rhodopsin II (pSRII) from Natronomonas pharaonis, is a seven-transmembrane (7-TM) receptor of negative phototaxis12,13 ; this protein represents a ubiquitous yet notoriously unstable class of membrane protein...

DgkA, pSRII and LacYGFP were released from micelles possessing average charge states of 7+, 9+ and 18+, respectively, indicative of compact structures. Unfolded proteins, with larger surface areas and/or more exposed basic residues, result in a wider charge-state distribution centered around higher charge. For DgkA, we observed the monomeric protein to be dominant, with a smaller population corresponding to dimers. However, the expected stoichiometry of DgkA is trimeric, indicating that detergent micelles are inadequate for preserving this complex in the gas phasep>s.

Using a high-tech nano flow system, molecules are transmitted into the instrument in charged water droplets, which then undergo evaporation releasing molecules into the gas phase of the mass spectrometer.

Hydrophobic membrane proteins pose a problem to this method, as by their very nature they won't dissolve in water, leading to the common use of detergents to force them to do so, however this can damage the protein's structure making it an unreliable model of the in vivo reality.

First author John Hopper explains on his faculty website that nanodiscs are tiny disc-like structures made of lipids...

...the same material that membrane proteins occupy in the body. It's essentially as if you took a round cookie cutter to remove a section of the natural bilayer, so the conditions are just like they would be in the body. The discs are stabilised by wrapping a belt of proteins around them to keep the exposed lipid tails from the water.

They consist of a segment of bilayer encapsulated by an amphipathic protein coat, rather than a lipid or detergent layer. Nanodiscs are more stable than bicelles and micelles at low concentrations, and have a well-defined size (10-20 nm, depending on the type of protein coat).

As an alternative to using bicelles, we next assembled membrane scaffold protein 1D1 (MSP1D1) nanodiscs (following established protocols16 ), to solubilize the target protein. MSP1D1 nanodiscs also provide a lipid bilayer environment, in this case DMPC. The hydrophobic perimeter of a nanodisc is stabilized by two copies of a long α-helical membrane scaffold protein (MSP). Measurements of dynamic light scattering revealed homogeneous size distributions (Supplementary Fig. 9), and transmission electron microscopy (TEM) showed uniformly sized disc structures, some in face-to-face stacked arrangements, of the correct geometry. LacY-GFP allowed us to monitor its incorporation into membrane scaffold protein 1E3D1 (MSP1E3D1) nanodiscs through size exclusion chromatography and affinity purification (SupplementaryFig. 1). We observed a series of charge states for LacY-GFP (3+ to 5+, Supplementary Fig. 11) from nanodiscs (compare to 14+ to 21+ from DDM micelles, Supplementary Fig. 5), implying that in nanodiscs LacY is folded and undergoes extensive lipid binding... Mass spectra (CE of 400 V) clearly showed the presence of well-defined trimers as well as monomers and dimers of DgkA.

...The low charge states observed for proteins ejected from nanodiscs or bicelles, compared with those from micelles, minimize the potential for coulomb-induced unfolding in the gas phase. Both nanodiscs and bicelles maintained the native trimeric stoichiometry of DgkA, though a larger population of trimeric assemblies was preserved after dissociation from bicelles compared with nanodiscs (Figs. 1c and 2c). Larger bicelles, presumably with greater lateral forces that effectively ‘compress’ the protein subunits along the plane of lipid bilayer, could maintain dimeric forms of pSRII.

Hopper explains further on the UoOx site,

Aside from the nanodiscs, we actually got great results from bicelles, which are made in a similar way. The main difference is that instead of putting a belt of proteins around the edge, we plug the gap with short-chain lipids instead. This actually gives us much more control over the size and structure of the disc.

The paper itself however is more critical of bicelles, highlighting that "both nanodiscs and bicelles have the ability to preserve protein-lipid interactions, although the range of lipids that can be explored is limited for bicelles.".

Figures below show some of the interesting conformations the molecules involved take on during these experiments (amphipols = amphipathic polymers)