Alternative Methods

However, the limited number of membrane protein structures available in the protein data bank also reflects the difficulties in obtaining well-ordered crystalline arrangements. Indeed, the lack of homogeneity in the starting solution is very often the major drawback for several reasons. First the detergent has to provide a homogeneous solubilization without denaturing the protein. The difference in dynamical behaviour of solvent and lipidic membranes certainly influences the dynamical properties of the protein. Extracting the protein from its native environment releases the lateral constraint applied by the membrane and might induce partial denaturing. This should be particularly critical for proteins that undergo large conformational changes for their functions, such as transporters. As a result, the protein solubilized in detergent micelles might be present with a large variety of conformations, including non-functional ones, a very unfavourable situation for crystallization. Individual endogenous lipids can interact strongly with the protein and help in stabilizing native conformations. The difficulty of controlling a constant (and appropriate) amount of well-defined lipids in the protein solutions also contributes to heterogeneity (Lemieux et al. 2003). More recent approaches tend to exploit the lipid phases and to reincorporate the proteins in a lipidic environment in order to get rid of the detergent. The idea was already exploited in the last decade with attempts at stacking two-dimensional crystals of bacteriorhodopsin (naturally present in the purple membrane) but no order in the third dimension could be obtained. More recently, in order to overcome the limitations due to two-dimensional space and to explore the three dimensions, lipidic cubic phases were introduced by Landau and Rosenbusch (1996). They demonstrated the feasibility of the method and obtained for the first time highly diffracting crystals of bacteriorhodopsin leading to the structure (Pebay-Peyroual et al. 1997; Belrahli et al. 1999; Luecke et al. 1999). Other bacterial rho-dopsins and reaction centres were further crystallized with this method (Chiu et al. 2000; Kolbe et al. 2000; Gordelyi et al. 2002; Katona et al. 2003; Royant et al. 2001; Vogeley et al. 2004). Interestingly, all these crystals consist of stacked layers of proteins forming type I crystals. The crystallization mechanism proposed by Nollert et al. (2001) is based on the lateral diffusion of the protein within the curved bilayer that constitutes the cubic phase. Increasing the curvature (by adding a precipitant) destabilizes the proteins by creating a hydrophobic mismatch and favours lateral protein-protein interactions. Locally, around the nucleation centre, lipids undergo a phase transition from cubic to lamellar. These various steps are illustrated in Fig. 3.3. The cubic phase can therefore by considered as a milieu that controls the relative orientations of the proteins during their diffusion in three-dimensional space.

The lipidic cubic phase approach is attractive because it can be generalized to other proteins in a rational way. There are still developments to take place either for a deeper mechanistic understanding, or for automating and miniaturizing the process (Qutub et al. 2004; Cherezov et al. 2004). In parallel, other individual cases of inducing membrane protein crystallization in a lipidic environment were also reported. The sarcoplasmic Ca-ATPase was crystallized by dialyz-

Lipid Cubic Phase

Fig. 3.3. Proposed crystallization mechanism in lipidic cubic phases, (a) Proteins solubilized in detergent micelles (2) or in lipid patches (1) are incorporated in the lipidic cubic phase (3). A modification of the cubic cell parameter (4) induces the stress that results in local protein reorganization possibly leading to 3D crystals (5). (b) Schematic representation of the various protein-membrane interactions encountered during the crystallization process (Nollert et al. 2001)

Fig. 3.3. Proposed crystallization mechanism in lipidic cubic phases, (a) Proteins solubilized in detergent micelles (2) or in lipid patches (1) are incorporated in the lipidic cubic phase (3). A modification of the cubic cell parameter (4) induces the stress that results in local protein reorganization possibly leading to 3D crystals (5). (b) Schematic representation of the various protein-membrane interactions encountered during the crystallization process (Nollert et al. 2001)

ing the detergent away in the presence of lipids leading to well-ordered stacks of 2D crystals. This technique was adapted from a two-dimensional crystallization protocol (Toyoshima et al. 2000). Similarly, the mitochondrial ADP/ATP carrier was crystallized from a protein-detergent solution in which excess detergent was adsorbed on polystyrene beads (Dahout-Gonzalez et al. 2003). The crystal is of type I; moreover each layer in the crystal packing contains only proteins, lipids and possibly monomeric detergent molecules tightly associated to the protein (Pebay-Peyroula et al. 2003). The success of this crystallization was due to the high amount of endogenous lipids still present after solubilization and purification. Adsorbing the detergent, probably forced the protein in lipid patches, which might nucleate the crystallization process. This example and others demonstrate the need for well-characterized systems prior to crystallization. In particular proteins that are overexpressed in heterologous systems might need the addition of a few lipids present in their native membrane and relevant for structural and/or functional reasons.

In conclusion, a new membrane protein after purification and solubilization in detergent should be characterized as far as possible in term of homogeneity, protein/detergent/lipid composition, importance of endogenous lipids, and lig-ands that stabilize a conformation; all of them are tools that will pave the way to crystallization. For some proteins a rapid analysis by negative stain electron microscopy could reveal the presence of several oligomeric states. Crystallization should be approached in parallel by several methods. The classical vapour diffusion method can now be efficiently explored with robots that use very small amounts of proteins (100 nL per drop compared to 1 ^L manually). A ten-fold reduction in the total amount of protein is noticeable for membrane proteins, which are so difficult to produce in large amounts. The number of parameters to explore can also be reduced by restricting the précipitants to polyethylenegly-cols of various molecular weights and alcohols that have proven to be favourable précipitants. Alternative methods such as the lipidic cubic phases should also be exploited.

Was this article helpful?

0 0
Healthy Chemistry For Optimal Health

Healthy Chemistry For Optimal Health

Thousands Have Used Chemicals To Improve Their Medical Condition. This Book Is one Of The Most Valuable Resources In The World When It Comes To Chemicals. Not All Chemicals Are Harmful For Your Body – Find Out Those That Helps To Maintain Your Health.

Get My Free Ebook


Post a comment