Even large proteins can be studied using traditional NMR methods; no deuteration or TROSY necessary.

The pressures required for using ethane are those that induce structural transition in proteins.

For 30 sites in eGFP (54kDa) the average T2 in ethane (76 ms) is nearly two-fold longer than in water (39ms)!

Some advantages:

Following on an old result by Montal with rhodopsin, we have been trying to adapt the reverse micelle encapsulation strategy to solution NMR of integral membrane proteins. Our initial efforts have been with the ~50 kDa construct of the KcsA potassium channel, which is a relatively stable, easily producted homotetramer. Our view of this reverse micelle assembly is quite different from that envisaged for soluble proteins where the protein is completely encapsulated. We believe that the protein-reverse micelle complex is dumbbell in shape (see below figure).The dilemma that had to be overcome was how to transition an integral membrane protein from detergent micelles in water to reverse micelles in alkane. A graduate student (Joe Kielec) realized that one should use a hybrid detergent that could act as a aqueous detergent and then, in the presence of a cosurfactant, could form reverse micelles (see left figure below). After a comprehensive screen for encapsulation conditions (right figure below), a functionally active (i.e. binds K+ selectively over an excess background of Na+) preparation was obtained (figure below).

See Kielec et al. (2009) for more information.

Membrane Protein Encapsulation Strategy

• extraction into hybrid surfactant in water
• controlled dehydration
• addition of converting co-surfactant & alkane

Membrane Protein Encapsulation Strategy

Functional selectivity filter confirms proper KcsA structure

Much to our surprise it turns out that despite the fact that the several percent of the human proteome is lipidated at any given time, there are very few structures of lipidated proteins in the membrane anchored conformation i.e. with the lipid group exposed. Common lipid modifcations include myristolyaton and palmititation. The former is often involved in triggered structural transitions of the protein from the lipid sequestered state to a state where the lipid moeity is extruded from the protein and capable of anchoring to the membrane. Recoverin is the classic example where binding of the calcium second messenger results in the binding of the protein to the membrane. Another example is the HIV matrix protein, which assists in the assembly of the immature virion at the plasma membrane. We are studying these proteins with Professor Jim Ames (UC, Davis) and Professor Mike Summers (UMBC & HHMI), respectively. The paucity of structural information in the lipid extruded states is clearly due to the poor solution properties of the protein. The reverse micelle provides an excellent vehicle to support the extruded state by providing a membrane mimic for the lipid group to enter. The superior spectroscopic performance of preparations of encapsulated myristoylated proteins is impressive.

See Valentine et al. (2010) for more detail.

Remarkably, perhaps 30% of expressed proteins are predicted to be unfolded. It is clear that a significant fraction of these are truly "intrinsically" unfolded but are simply unstable suggesting that they have evolved to take advantage of the stabilization provided excluded volume effects due to the tight packing of macromolecules in the cellular millieu. Obviously this makes it difficult to study the structural properties of these proteins in vitro, particularly in the NMR tube. Using the defined internal volume of the reverse micelle, we have developed this as a means to impart additional stability to an otherwise only marginally stable protein. The internal volume of a reverse micelle can be largely controlled by the amount of water that is provided. The nice thing about this approach is that in principle the native folded state is not only the best packed but also, through the thermodynamic hypothesis, the most stable.

Thus a properly sized reverse micelle will leave the free energy and structure of the native state alone but will destabilize the more extensive partially unfolded states causing the population to shift towards the desired folded native state. We have demonstrated this idea with a purposefully destabilized three helix bundle where the internal core is left unperturbed (i.e. can fold) but the helical propensity is decreased by surface mutations. Decreasing the water content ("loading") and corresponding decreasing the internal volume available to the protein causes it to fold (below figure).

See Peterson et al. (2004) JACS for more detail.

High resolution solution NMR of the myristoylated proteins in the extruded state are enabled by reverse micelle encapsulation

Forced folding by confinement:

Assuming:

• Thermodynamic hypothesis for protein folding
• The (almost) perfect packing of native proteins

Forced folding by confinement:

Assuming:

• Thermodynamic hypothesis for protein folding
• The (almost) perfect packing of native proteins

Current Project Personnel

Sabrina Bedard
Igor Dodveski
Lia Athanasoula
Nimu Sidhu
Megan Tzakas
Nathaniel Nucci
John Gledhill
Kathy Valentine
Adam Seitz

On-Going & Future Projects

• Structural biology of lipidated proteins
• Structural biology of G-protein coupled receptors
• Structural biology of large soluble proteins
• Protein hydration & structural water
• The structural biology of unstable proteins by forced folding