Search

Transient lateral microdomains or lipid rafts play important roles in many physiological membrane-mediated cell processes. Detergent-resistant membranes (DRMs) are good models for the study of lipid rafts. Here we report that DRMs can be obtained by treating human erythrocytes with the nonionic detergents Triton X-100 or octaethylene glycol monododecyl ether (C(12)E(8)) at 37 degrees C, and by treatment at 4 degrees C of cholesterol-depleted erythrocytes. Electron paramagnetic resonance with spin labels inserted at different membrane depths (5- and 16-doxyl stearic acids, 5-SASL and 16-SASL) were used to measure the order parameter (S) of the cell membranes and DRMs. We previously reported significantly higher S values in DRMs with respect to intact erythrocyte membranes. Here we show that higher S values were still measurable in DRMs prepared from intact erythrocytes at 37 degrees C, or from cholesterol-depleted cells at 4 degrees C, for both detergents. For 5-SASL only, increased S values were measured in 4 degrees C DRMs obtained from cholesterol-depleted versus intact erythrocytes. Flotillin-2, a protein marker of lipid rafts, was found in DRMs from intact cells in trace amounts but it was sensitively increased in C(12)E(8) DRMs prepared at 4 degrees C from cholesterol-depleted erythrocytes, while the membrane-skeletal proteins spectrin and actin were excluded from both Triton X-100 and C(12)E(8) DRMs. However, contrary to the 4 degrees C treatment results, flotillin-2 and stomatin were not resistant to Triton X-100 and C(12)E(8) treatment at physiological temperature. The role of cholesterol in DRMs formation is discussed and the results presented provide further support for the use of C(12)E(8) to the study of DRMs.

The dimerizations of membrane proteins, Outer Membrane Phospholipase A (OMPLA) and glycophorin A (GPA), have been simulated by an adapted Brownian Dynamics program. To mimic the membrane protein environment, we introduced a hybrid electrostatic potential map of membrane and water for electrostatic interaction calculations. We added a van der Waals potential term to the force field of the current version of the BD program to simulate the short-range interactions of the two monomers. We reduced the BD sampling space from three dimensions to two dimensions to improve the efficiency of BD simulations for membrane proteins. The OMPLA and GPA dimers predicted by our 2D-BD simulation and structural refinement is in good agreement with the experimental structures. The adapted 2D-BD method could be used for prediction of dimerization of other membrane proteins, such as G protein-coupled receptors, to help better understanding of the structures and functions of membrane proteins.

Self-association of two hydrophobic alpha-helices is studied via unrestrained Monte Carlo (MC) simulations in a hydrophobic slab described by an effective potential. The system under study represents two transmembrane (TM) segments of human glycophorin A (GpA), which form homo-dimers in membranes. The influence of TM electrostatic potential, thickness and hydrophobicity degree of lipid bilayer is investigated. It is shown that the membrane environment stabilizes alpha-helical conformation of GpA monomers, induces their TM insertion and facilitates inter-helical contacts. Head-to-head orientation of the helices is promoted by the voltage difference across the membrane. Subsequent "fine-tuned" assembling of the dimer is mediated by van der Waals interactions. Only the models of dimer, calculated in a hydrophobic slab with applied voltage agree with experimental data, while simulations in vacuo or without TM voltage fail to give reasonable results. The moderate structural heterogeneity of GpA dimers (existence of several groups of states with close energies) is proposed to reflect their equilibrium dynamics in membrane-mimics. The calculations performed for GpA mutants G83A and G86L permit rationalization of mutagenesis data for them. The results of Monte Carlo simulations critically depend on the parameters of the membrane model: adequate description of helix association is achieved in the water-cyclohexane-water system with the membrane thickness 30-34 A, while in membranes with different hydrophobicities and thickness unrealistic conformations of the dimer are found. The computational approach permits efficient prediction of TM helical oligomers based solely on the sequences of interacting peptides.

The three-dimensional structure of a glycopeptide, His-Thr*-Ser*-Thr*-Ser*-Ser*-Ser*-Val-Thr-Lys, with 2-acetamido-2-deoxy-alpha-D-galactose (GalNAc) residues linked to six adjacent amino acids from Thr-10 to Ser-15, was studied by NMR spectroscopy and molecular dynamics (MD) simulations. The hexaglycosylated decapeptide is part of the extracellular domain of human glycophorin A and shows an extended structure of the peptide backbone due to O-glycosylation. Furthermore, each GalNAc residue exhibits one and only one NOE contact from the NHAc proton to the backbone amide proton of the amino acid that the sugar is directly bound to. This indicates a strong preference for the orientation of all GalNAc residues towards the N-terminus. NOE build-up curves were used to determine 42 inter-proton distances that, in connection with phi angles of the peptide backbone obtained from 3J-coupling constants, resulted in constraints for a MD simulation in water. The NMR data and the MD simulations show a preference for an extended backbone structure. The GalNAc residues are located alternatingly on opposite sides of the backbone and reduce the flexibility of the peptide backbone. The conformation of the molecule is relatively rigid and shows a 'wave-type' 3D structure of the peptide backbone within the glycosylation cluster. This new structural element is also supported by the unusual CD spectrum of the glycopeptide.

We have used a spin-echo difference NMR pulse sequence to measure three-bond J couplings between lambda- and alpha-carbons of the leucine residues in a micelle-associated helical peptide dimer that corresponds to residues 62-101 of the transmembrane erythrocyte protein glycophorin A. The observed 3J couplings correlate strongly with the 13C chemical shift of the lambda-methyl groups, and within experimental error both the shift distribution of the methyl carbons and the variations in 3J can be accounted for by variations in side-chain rotamer populations. We infer that all leucine side chains in this peptide dimer are in fast exchange among chi 2 rotamers and sample two of the three possible rotameric states, even when the side chain forms part of the dimer interface. The observed correlation of chemical shift with couplings can be traced to a gamma-gauche interaction of methyl and alpha-carbons. This correlation may provide an alternate route to rotamer analysis in some protein systems.

Anti-Pr agglutinins (CAs) with the subspecificities anti-Pr1h, -Pr1d, -Pr2, -Pr3h, -Pr3d, -PrM and anti-Sa CAs recognize immunodominant N-acetylneuraminic acid (NeuN Ac) groups of tetra and/or trisaccharides (O-glycans) of glycophorin. These O-glycans are sialylated in alpha 2,3- and/or alpha 2,6-linkages. Sa and most Pr antigens have been inactivated by alpha 2,3-specific sialidases. Antigenicity was reconstituted on desialylated glycophorin by alpha 2,3-specific Gal beta 1,3GalN Ac-sialyltransferase indicating that alpha 2,3-linked NeuN Ac groups are the immunodominant components of Sa and most Pr antigens. Some Pr antigens were resistant to alpha 2,3-specific sialidase and were not reconstituted by alpha 2,3-specific Gal beta 1,3GalN Ac-sialyltransferase, which indicates that alpha 2,6-linked NeuN Ac group represents an immunodominant component of some Pr antigens.

Alterations in N- and O-linked glycosylation affect cell surface expression and antigenicity of recombinant glycophorin A expressed in transfected Chinese hamster ovary (CHO) cells. To understand these effects further, glycophorin A was purified by immunoaffinity chromatography from transfected wild type and glycosylation deficient CHO cells. The O-glycans were characterized both biochemically, using gel filtration and high performance anion exchange chromatography, and immunologically, using carbohydrate specific monoclonal antibodies to probe Western blots. The O-glycans of human erythrocyte glycophorin A consist mainly of short oligosaccharides with one, two, or three sialic acid residues linked to a common disaccharide core, Gal beta 1-3GalNAc alpha 1-Ser/Thr, with the disialylated structure being the most abundant. With the exception of the trisialylated derivative, the same structures were found on recombinant glycophorin A expressed by wild type CHO cells. However, in contrast to human erythrocyte glycophorin A, the monosialylated oligosaccharide was the most abundant structure on the recombinant protein. Furthermore, recombinant glycophorin A was shown to express a small amount of the Tn antigen (GalNAc alpha 1-Ser/Thr). Recombinant glycophorin A had the same O-glycan composition, whether purified from clones expressing high or moderate levels of the recombinant glycoprotein. This indicates that the level of expression of the transfected glycoprotein did not affect its O-glycan composition. Deletion of the N-linked glycosylation site at Asn26, by introducing the Mi.I mutation (Thr28-->Met) by site-directed mutagenesis, did not markedly affect the O-glycan composition of the resulting recombinant glycoprotein expressed in wild type CHO cells.(ABSTRACT TRUNCATED AT 250 WORDS)

Human erythrocyte glycophorin was desialylated by mild acid hydrolysis and degalactosylated by Smith degradation. Two monoclonal antibodies (Tn5 and Tn56) obtained by immunization of mice with this 'artificial' Tn antigen were characterized and compared in some experiments with two antibodies (BRIC111 and LM225) obtained in other laboratories by immunization with Tn erythrocytes. The specific binding of the antibodies to glycophorins desialylated and degalactosylated on the nitrocellulose blot and to asialo-agalactoglycophorin-coated ELISA plates, and reactions with authentic Tn antigen served for identification of their anti-Tn specificity. The antibodies were further characterized in inhibition assay with various glycoproteins. The antibody Tn5 (similar to BRIC111) was shown to be specific for human erythrocyte Tn antigen, whereas Tn56 reacted strongly with different glycoproteins carrying O-linked GalNAc alpha- residues, and was strongly bound to the murine adenocarcinoma cell line Ta3-Ha. The antibodies Tn5, Tn56 and BRIC111 were similarly inhibited by ovine submaxillary mucin (OSM) and asialoOSM, but the antibody LM225 showed a distinct preference in reaction with OSM (sialosyl-Tn antigen). The results show that Tn antigen, obtained by chemical modifications of human glycophorin, enables the preparation and characterization of anti-Tn monoclonal antibodies, without using rare Tn erythrocytes.

Sialic acids and the major O-glycosidic oligosaccharide of glycophorin MK from monkey (Japanese monkey, Macaca fuscata) erythrocyte membranes were characterized. N-Glycolylneuraminic acid (Neu5Gc) was found as the major sialic acid, which was confirmed by gas-liquid chromatography-mass spectrometry as the trimethylsilyl methyl ester. Three O-glycosidic oligosaccharide units were obtained from a tryptic glycopeptide that contained all of the carbohydrate units in glycophorin MK by mild alkaline borohydride/borotritide treatment. Carbohydrate analyses of the oligosaccharides revealed that they were composed of Neu5Gc, galactose and N-acetylgalactosaminitol in the molar ratios of 1:1:1 (trisaccharide), 2:1:1 (tetrasaccharide) and 3:1:1 (pentasaccharide). The content of oligosaccharide units was estimated to be 1:12:5 for penta-, tetra- and trisaccharide, respectively, based on the yields, the molecular weight, and the number of oligosaccharide attachment sites in the amino-acid sequence. The tetrasaccharide was the major oligosaccharide and its structure was proposed to be Neu5Gc alpha 2-3Gal beta 1-3[Neu5Gc alpha 2-6]GalNAcol.