It has been suggested that binding of extrinsic membrane proteins to lipid bilayer surfaces could be affected by curvature stress in the lipid bilayer, the level of curvature stress being increased by the preference of a phospholipid such as phosphatidylethanolamine for a curved, hexagonal Hn phase (Gruner 1985). However, in two cases of extrinsic membrane proteins where crystal structures are available, it is clear that binding specificity is determined by specific molecular interactions rather than by curvature stress. The first case concerns the channel-forming toxins a-hemolysin and LukF of Staphylococcus aureus (Olson et al. 1999; Galdiero and Gouaux 2004). a-Hemolysin permeabilizes liposomes of phosphatidylcholine or sphingomyelin, but not those of phosphatidylethanolamine, phosphatidylserine, phosphatidylglycerol, or phosphatidylinositol (Watanabe et al. 1987). The crystal structures make clear the basis of this selectivity (Olson et al. 1999; Galdiero and Gouaux 2004). The crystal structure for LukF crystallized in the presence of dipropanoylphosphatidylcholine is shown in Fig. 6.6. Only the phosphorylcholine headgroup is resolved, this being largely buried. The phosphate group interacts with the side-chain of Arg-198 and its main chain nitrogen, the quaternary ammonium group is in contact with Glu-192 and Tyr-180, and the CH2-CH2 group is near Trp-177 (Fig. 6.6). The specificity of the toxins for phosphatidylcholine or sphingomyelin can, therefore, be understood in terms of the specific molecular interactions between the lipid headgroup and its binding site on the toxin.
Fig. 6.7. Binding of glycerophosphoserine (GPS) and glycerophosphoethanolamine (GPE) to annexin V.The binding sites for GPS (a) and GPE (b) are shown. The filled sphere is Ca2+. Some of the residues Important for binding are shown In ball-and-stlck mode (PDB files 1A8Aand 1A8B)
Fig.6.8. Structure ofthe phosphatldylethanolamlne molecule bound to the photosynthetlc reaction centre from T. tepidum. The residues from subunlts HandM interacting with the lipid headgroup are shown (PDB file 1EYS)
The second case concerns the annexins, which show specificity for phosphatidylethanolamines and phosphatidylserines. Annexin V hardly binds to bilayers of phosphatidylcholine or sphingomyelin but binds to bilayers of phosphatidyleth-anolamine, and more strongly, to bilayers containing anionic phospholipids such as phosphatidylserine (Schlaepfer et al. 1987; Blackwood and Ernst 1990; Andree et al. 1990; Raynal and Pollard 1994; Campos et al. 1998). The crystal structure of rat annexin V in the presence of glycerophosphoserine (GPS) and glycerophos-phoethanolamine (GPE) (Swairjo et al. 1995) shows the phosphoryl oxygen coordinated to a bound Ca2+ ion, with the GPE headgroup extending along the molecular surface in the opposite direction to GPS, in a shallower binding site (Fig. 6.7); the more extensive interactions observed in the crystal structure with GPS than with GPE explains why binding to phosphatidylserine is stronger than to phosphatidylethanolamine. In this case the binding site is too small to accommodate the bulky phosphorylcholine group.
In some cases it is clear that interactions between intrinsic membrane proteins and the surrounding lipid molecules also have to be understood in terms of specific molecular level interactions. For example, Fig. 6.8 shows the structure of a phosphatidylethanolamine molecule resolved in the crystal structure of pho-tosynthetic reaction centre from Thermochromatium tepidum (Nogi et al. 2000). The conformation adopted by the phosphatidylethanolamine is very different to that shown in Fig. 6.2; the glycerol backbone is not oriented parallel to the fatty acyl chains, and the headgroup is not oriented perpendicular to the chain axis. Clearly, the adopted structure is determined by specific interactions, particularly between the headgroup and the protein, the headgroup being bent down towards the chain, for example, to allow a hydrogen-bonded interaction between the quaternary ammonium nitrogen and the backbone of Gly-256 in subunit M (Fig. 6.8). Other examples of specific interactions of this type have been presented in Lee (2003) and an analysis of the conformations of bound lipids shows that they are often different from the structures adopted in simple lipid bilayers (Marsh and Pali 2004).
It is likely that the lipid molecules resolved in crystal structures of intrinsic membrane proteins are rather unrepresentative of the bulk of the lipid molecules in contact with a membrane protein, the resolved lipid molecules often being located between transmembrane a-helices, often at protein-protein interfaces in multi-subunit proteins (Lee 2003, 2004). Most of the lipid molecules interacting with the hydrophobic surface of a membrane protein (the boundary or annular lipids) will interact with the protein rather non-specifically, although there is evidence for binding hot-spots on the surface of a protein close, for example, to clusters of positively charged residues where anionic phospholipids will bind relatively strongly (Powl et al. 2005). Hydrogen bonding and charge interactions with the lipid headgroup must also be important for non-specific binding of lipid molecules to a membrane protein. The binding constants for phospholipids to annular sites on membrane proteins will depend on the strengths of the lipid-protein interactions relative to those of lipid-lipid interactions. ESR studies show that the on and off rate constants for annular binding are fast, suggesting that the strength of the lipid-protein interaction is comparable to that of a lipid-lipid interaction (East et al. 1985; Lee 2003) consistent with the general lack of selectivity in binding phospholipids to the annular sites on a membrane protein (Lee 2003). Given the importance of hydrogen bonding and charge interactions between lipid molecules in the headgroup region of a lipid bilayer, this suggests that hydrogen bonding and charge interactions with membrane proteins must also be important. In agreement with this expectation, uncharged molecules such as long-chain alcohols bind less weakly to membrane proteins than charged molecules such as long-chain amines or acids (Froud et al. 1986; Lee 2003).
The importance ofhydrogen bonding has been shown in a molecular dynamics simulation of the mechanosensitive channel of large conductance MscL from T. tuberculosis in bilayers of (C16:0,C18:1)PE. The simulation showed a large number of hydrogen bonds between the lipid molecules and MscL, about half involving the NH3+ group of the phosphatidylethanolamine headgroup (Elmore and Dougherty 2001). Other lipid headgroups such as phosphatidylcholine and phosphatidylglycerol do not have a hydrogen bond donating group analogous to the NH3+ of phosphatidylethanolamine and so show a very different pattern of hydrogen bonding. A molecular dynamics simulation of MscL in bilayers of (C16:0,C18:1)PC suggests that the loss ofhydrogen bonding observed on replacement of the phosphatidylethanolamine headgroup by the phosphatidylcholine headgroup is compensated for by a conformational change in the C-terminal region of the protein, bringing the C-terminal region closer to the membrane, leading to stronger interactions with the membrane (Elmore and Dougherty 2003). A molecular dynamics simulation of bacteriorhodopsin also shows the importance of headgroup interactions with the protein, different interactions with the different conformational states of the protein having an important effect on the energetics of the changes between these conformational states (Jang et al. 2004).
It is well established that the conformational equilibrium between the two major intermediates of the rhodopsin photocycle, metarhodopsin I (MI) and me-tarhodopsin II (Mil), depend on lipid structure, and that the presence of phos-phatidylethanolamines increases the ratio MII/MI (Litman and Mitchell 1996; Brown 1997). These effects are usually described in terms of the tendency of
Fig. 6.9. The structure of rhodopsin.The hydrophobic thickness of rhodopsin, defined largely by the positions ofTyr and Trp residues, is shown bythe horizontal lines. Residue Glu-134 is shown in space fill format (PDB file 1F88)
phosphatidylethanolamines to adopt a curved, hexagonal Hn phase. However, as described in Lee (2003), the results can also be interpreted in terms of head-group interactions with rhodopsin. The effect of phosphatidylethanolamines on the MII/MI ratio for rhodopsin follows from a shift in the pK describing the pH dependence of this equilibrium (Brown 1994). Mutation of Glu-134 leads to a loss of the pH dependence of the MII/MI ratio, suggesting that the pH dependence of the equilibrium follows from protonation/deprotonation of Glu-134 (Amis et al. 1994). As shown in Fig. 6.9, Glu-134 is probably located very close to the glycerol backbone region of the surrounding lipid bilayer, so that the pK of this residue is very likely to change with changing lipid structure (Lee 2003). Indeed, calculations of the pK of this residue show that it is environmentally sensitive (Periole et al. 2004).
Further examples oflipid headgroup-protein interactions and their effects on protein function are given in Lee (2003, 2004).
Was this article helpful?
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.