Laszlo I. Horvath

Interaction of membrane-spanning proteins with peripheral and lipid-anchored membrane proteins: perspectives from protein± lipid interactions (Review)

Studies of lipid-protein interactions in double-reconstituted systems involving both integral and peripheral or lipid-anchored proteins are reviewed. Membranes of dimyristoyl phosphatidylglycerol containing either myelin proteolipid protein or cytochrome c oxidase were studied. The partner peripheral proteins bound to these membranes were myelin basic protein or cytochrome c, respectively. In addition, the interactions between the myelin proteolipid protein and avidin that was membrane-anchored by binding to N -biotinyl phosphatidylethanolamine were studied in dimyristoyl phosphatidylcholine membranes. Steric exclusion plays a significant role when sizes of the peripheral protein and transmembrane domain of the integral protein are comparable. Even so, the effects on avidin-linked lipids are different from those induced by myelin basic protein on freely diffusible lipids, both interacting with the myelin proteolipid protein. Both the former and the cytochrome c /cytochrome oxidase couple evidence a propagation of lipid perturbation out from the intramembrane protein interface that could be a basis for formation of microdomains.

Distance measurements using paramagnetic ion-induced relaxation in the saturation transfer electron spin resonance of spin-labeled biomolecules. Application to phospholipid bilayers and interdigitated gel phases

The saturation transfer electron spin resonance (STESR) spectra of spin-labeled phosphatidylcholines in gel phase lipid bilayers are shown to be sensitive to dipolar spin-spin interactions with paramagnetic ions in the aqueous phase. The reciprocal integrated intensity of the STESR spectrum is linearly dependent on aqueous Ni(2+) ion concentration, hence, confirming the expectation that the STESR intensity is directly proportional to the spin-lattice relaxation time of the spin label. The gradient of the relaxation rate with respect to Ni(2+) ion concentration decreases strongly with the position of the nitroxide group down the sn-2 chain of the spin-labeled lipid and is consistent with a 1/R(3) dependence on the distance, R, from the bilayer surface. The values derived for the dimensions of the bilayer and lipid molecules in the case of dipalmitoyl phosphatidylcholine (DPPC) are in good agreement with those available from x-ray diffraction studies. Allowance for the multibilayer nature of the DPPC dispersions gives an estimate of the water layer thickness that is also consistent with results from x-ray diffraction. The profile of the paramagnetic ion-induced relaxation is drastically changed with DPPC dispersions in glycerol for which the lipid chains are known to be interdigitated in the gel phase. The terminal methyl groups of the lipid chains are located approximately in register with the C-3 atoms of the sn-2 chain of the oppositely oriented lipid molecules in the interdigitated phase. The thickness of the lipid layer and the effective thickness of the lipid polar group are reduced by approximately 40% in the interdigitated phase as compared with the bilayer phase. The calibrations of the distance dependence established by use of spin labels at defined chain positions should be applicable to STESR measurements on other biological systems.

Saturation transfer electron spin resonance of Ca2(+)-ATPase covalently spin-labeled with beta-substituted vinyl ketone- and maleimide-nitroxide derivatives. Effects of segmental motion and labeling levels.

The Ca2(+)-ATPase in native sarcoplasmic reticulum membranes was selectively spin-labeled for saturation transfer electron spin resonance (ESR) studies by prelabeling with N-ethylmaleimide and by using low label/protein ratios. Results with the nitroxide derivative of the standard sulphydryl-modifying reagent, maleimide, were compared with a series of six novel nitroxide beta-substituted vinyl aryl ketone derivatives which differed (with two exceptions) in the substituent at the ketone position. The two exceptions had a different electron withdrawing group at the alpha-carbon, to enhance further the electrophilic character of the beta-carbon. Although differing in their reactivity, all the conjugated unsaturated ketone nitroxide derivatives displayed saturation transfer ESR spectra indicative of much slower motion than did the maleimide derivative. The saturation transfer ESR spectra of maleimide-labeled Ca2(+)-ATPase therefore most likely contain substantial contributions from segmental motion of the labeled group. The effects of the level of spin labeling were also investigated. With increasing degree of spin label incorporation, the linewidths of the conventional ESR spectrum progressively increased and the intensity of the saturation transfer spectrum dropped dramatically, as a result of increasing spin-spin interactions. The hyperfine splittings of the conventional spectrum and the outer lineheight ratios of the saturation transfer spectrum remained relatively unchanged. Extrapolation back to zero labeling level yielded comparable values for the effective rotational correlation times deduced from the saturation transfer spectrum intensities and from the lineheight ratios, for the vinyl ketone label. For the maleimide label the extrapolated values from the integral are significantly lower than those from the lineheight ratios, probably because of the segmental motion. Comparison is made of the effective rotational correlation time for the vinyl ketone label with the predictions of hydrodynamic models for the protein diffusion, in a discussion of the aggregation state of the Ca2(+)-ATPase in the native sarcoplasmic reticulum membrane. The implications for the study of protein rotational diffusion and segmental motion, and of the proximity relationships between labeled groups, using saturation transfer ESR spectroscopy are discussed.

Kinetics and dynamics of annealing during sub-gel phase formation in phospholipid bilayers: A saturation transfer electron spin resonance study.

The saturation transfer electron spin resonance (STESR) spectra of spin-labeled phosphatidylcholine have been used to follow the kinetics of conversion from the gel phase to the sub-gel phase in aqueous bilayers of dipalmitoyl phosphatidylcholine. This is a simple, well-defined model system for lipid domain formation in membranes. The integrated intensity of the STESR spectrum from the chain-labeled lipid first increases and then decreases with time of incubation in the gel phase at 0 degrees C. The first, more rapid phase of the kinetics is attributed to the conversion of germ nuclei to growth nuclei of the sub-gel phase. The increase in STESR intensity corresponds to the reduction in chain mobility of spin labels located in the gel phase at the boundaries of the growth nuclei and correlates with the increase in the diagnostic STESR line height ratios over this time range. The second, slower phase of the kinetics is attributed to growth of the domains of the sub-gel phase. The decrease in STESR intensity over this time regime corresponds to exclusion of the spin-labeled lipids from the tightly packed sub-gel phase and correlates quantitatively with calibrations of the spin label concentration dependence of the STESR intensity in the gel phase. The kinetics of formation of the sub-gel phase are consistent with the classical model for domain formation and growth. At 0 degrees C, the half-time for conversion of germ nuclei to growth nuclei is approximately 7.7 h and domain growth of the sub-gel phase is characterized by a rate constant of 0.025 h(-1). The temperature dependence of the STESR spectra from samples annealed at 0 degrees C suggests that the subtransition takes place via dissolution of sub-gel phase domains, possibly accompanied by domain fission.

Interactions between lipid-anchored and transmembrane proteins. Spin-label ESR studies on avidin-biotinyl phosphatidylethanolamine in membrane recombinants with myelin proteolipid proteins.

Interactions between lipid-anchored and transmembrane proteins are relevant to the intracellular membrane sorting of glycosyl phosphatidylinositol-linked proteins. We have studied the interaction of a spin-labeled biotinyl diacyl phospholipid, with and without specifically bound avidin, with the myelin proteolipid protein (or the DM-20 isoform) reconstituted in dimyristoylphosphatidylcholine. Tetrameric avidin bound to the N-biotinyl lipid headgroup is a surface-anchored protein, and the myelin proteolipid is an integral protein containing four transmembrane helices. The electron spin resonance (ESR) spectrum of N-biotinyl phosphatidylethanolamine spin-labeled at the C-14 position of the sn-2 chain consists of two components in fluid-phase membranes of dimyristoylphosphatidylcholine containing the proteolipid. In the absence of avidin, this is characteristic of lipid-protein interactions with integral transmembrane proteins. The more motionally restricted component represents the lipid population in direct contact with the intramembranous surface of the integral protein, and the more mobile component corresponds to the bulk fluid lipid environment of the bilayer. In the presence of avidin, the biotin-lipid chains have reduced mobility because of the binding to avidin, even in the absence of the proteolipid [Swamy, M. J., and Marsh, D. (1997) Biochemistry 36, 7403-7407]. In the presence of the proteolipid, the major fraction of the avidin-anchored chains is further restricted in its mobility by interaction with the transmembrane protein. At a biotin-lipid concentration of 1 mol %, approximately 80% of the avidin-linked chains are restricted in membranes with a phosphatidylcholine:proteolipid molar ratio of 37:1. This relatively high stoichiometry of interaction can be explained when allowance is made for the closest interaction distance between the lipid-anchored avidin tetramer and the transmembrane proteolipid hexamer, without any specific interaction between the two types of membrane-associated proteins. The interaction is essentially one of steric exclusion, but the lipid chains are rendered more sensitive to interaction with the integral protein by being linked to avidin, even though they are removed from the immediate intramembrane protein-lipid interface. This could have implications for the tendency of lipid-anchored chains to associate with membrane domains with reduced lipid mobility.