Intact Red Cell Membranes Research Articles

Enhanced Identification of Zero-Length Chemical Cross-Links Using Label-Free Quantitation and High-Resolution Fragment Ion Spectra

Chemical cross-linking coupled to mass spectrometry provides structural information that is useful for probing protein conformations and providing experimental support for molecular models. "Zero-length" cross-links have greater value for these applications than longer cross-links because they provide more stringent distance constraints. However, this method is less commonly utilized because it cannot take advantage of isotopic labels, MS-labile bonds, or enrichment tags to facilitate identification. In this study, we combined label-free precursor ion quantitation and targeted tandem mass spectrometry with a new software tool, Zero-length Cross-link Miner (ZXMiner), to form a multitiered analysis strategy. A major, critical objective was to simultaneously achieve very high accuracy with essentially no false-positive cross-link identifications while maintaining a good depth of analysis. Our strategy was optimized on several proteins with known crystal structures. Comparison of ZXMiner to several existing cross-link analysis software showed that other algorithms detected less true positive cross-links and were far less accurate. Although prior use of zero-length cross-linking was typically restricted to small proteins, ZXMiner and the associated strategy enable facile analysis of very large protein complexes. This was demonstrated by identification of zero-length cross-links using purified 526 kDa spectrin heterodimers and intact red cell membranes and membrane skeletons.

New Insights into the Function of N-Terminal 11 Amino Acids of Band 3 from Structural and Functional Study of a Naturally Occuring Band 3 Variant.

The 911 amino acid human erythroid AE1 (eAE1) Cl-/HCO3- exchanger SLC4A1 (usually called band 3) is the major intrinsic membrane protein of red cells. The N-terminal cytoplasmic domain of AE1 represents the anchoring site for membrane-associated proteins such as ankyrin, protein 4.2, protein 4.1, glycolytic enzymes (including aldolase and glyceraldeyde-3-phosphate dehydrogenase (GAPDH) and hemoglobin. We identified marked band 3 deficiency in the second son of a consanguineous marriage with a life-threatening nonimmune hemolytic anemia. The patient was transfusion-dependent prior to splenectomy. SDS-PAGE and immunoblotting analysis of the proband red cell membrane proteins showed approximately 12±4% of band 3 and protein 4.2 compared to controls. Direct nucleotide sequence of SLC4A1 gene showed a single base substitution (T->C) at position +2 in the donor splice site of intron 2 (Band 3 Neapolis). Functionally, the mutation causes an altered splicing with the consequent formation of two different mature mRNAs, one including intron 2 and one skipping exon 2. While intron 2 retention leads to premature translation termination, exon 2 skipping causes the loss of the normal start site of eAE1 protein translation. The purification of mutant band 3 and its characterization by MALDI mass spectrometry demonstrated the lack of the first 11 amino acids due to the usage of second in frame start site. Real-time RT-PCR analyses of reticulocyte mRNA showed a marked decrement in band 3 transcription accounting for protein deficiency. The lack of the 11 N-terminal amino acids resulted in complete absence of membrane bound aldolase while other glycolitic enzymes (for example GAPDH) were membrane bound. Syk tyrosine kinase recognized the truncated band 3 as a substrate in vitro. In spite of this ability to be phosphorylated by Syk and to recruit Lyn tyrosine kinase in vitro, we were unable to demonstrate Tyr-phosphorylation of mutant band 3 in intact erythrocytes following stimulation by oxidative stress. This finding implies a requirement for the 11 N-terminal amino acids for the sequential Tyr-phosphorylation of band 3 in intact red cell membranes. The mutant band 3 was largely present in the high molecular weight aggregate fraction (about 5.2 fold higher than control), indicating its increased tendency to cluster in the membrane. The spontaneous clustering of truncated band 3 strongly suggests that the negatively charged N-terminal domain may regulate oligomeric state of band 3 in the membrane. Biophysical characterization showed that band 3 deficiency resulted in decreased cohesion between lipid bilyer and spectrin based membrane skeleton accounting for membrane loss. The structural and functional characterization of the naturally occuring mutant band 3 has enabled us to identify a significant role for the 11 N-terminal amino acids in band 3 function and in red cell membrane physiology.

Initial steps of ?and ?-d-glucose binding to intact red cell membrane

The kinetics of the initial phases of D-glucose binding to the glucose transport protein (GLUT1) of the human red cell can be followed by stopped-flow measurements of the time course of tryptophan (trp) fluorescence enhancement. A number of control experiments have shown that the trp fluorescence kinetics are the result of conformational changes in GLUT1. One shows that nontransportable L-glucose has no kinetic response, in contrast to D-glucose kinetics. Other controls show that D-glucose binding is inhibited by cytochalasin B and by extracellular D-maltose. A typical time course for a transportable sugar, such as D-glucose, consists of a zero-time displacement, too fast for us to measure, followed by three rapid reactions whose exponential time courses have rate constants of 0.5-100 sec-1 at 20 degrees C. It is suggested that the zero-time displacement represents the initial bimolecular ligand/GLUT1 association. Exponential 1 appears to be located at, or near, the external membrane face where it is involved in discriminating among the sugars. Exponential 3 is apparently controlled by events at the cytosolic face. Trp kinetics distinguish the Kd of the epimer, D-galactose, from the Kd for D-glucose, with results in agreement with determinations by other methods. Trp kinetics distinguish between the binding of the alpha- and beta-D-glucose anomers. The exponential 1 activation energy of the beta-anomer, 13.6 +/- 1.4 kcal mol-1, is less than that of alpha-D-glucose, 18.4 +/- 0.8 kcal mol-1, and the two Arrhenius lines cross at approximately 23.5 degrees C. The temperature dependence of the kinetic response following alpha-D-glucose binding illustrates the interplay among the exponentials and the increasing dominance of exponential 2 as the temperature increases from 22.3 to 36.6 degrees C. The existence of these interrelations means that previously acceptable approximations in simplified reaction schemes for sugar transport will now have to be justified on a point-to-point basis.

The calmodulin-binding domain as an endogenous inhibitor of the p-nitrophenylphosphatase activity of the Ca 2+ pump from human red cells

Digestion of red cell membranes with chymotrypsin elicited p-nitrophenylphosphatase activity. During digestion, the p-nitrophenylphosphatase appeared in parallel with the activation of the Ca 2+-ATPase (in the absence of calmodulin). The chymotrypsin-activated p-nitrophenylphosphatase was inhibited by C20W, a 20 amino acid peptide modelled after the sequence of the calmodulin-binding site of the red cell Ca 2+ pump (Vorherr et al. (1990) Biochemistry 29, 355–365). On the contrary, the (ATP + Ca 2+)-dependent p-nitrophenylphosphatase activity of intact red cell membranes was not affected by C20W. Ca 2+ inhibited the chymotrypsin-induced p-nitrophenylphosphatase ( K ifor Ca 2+ = 2 μM). In the absence of ATP, C20W and Ca 2+ did not interact in apparent affinity as inhibitors of this activity. On the other hand, in the presence of 2 mM ATP, Ca 2+ antagonized the inhibition produced by C20W. The results are consistent with the idea that the calmodulin-binding site is an ‘autoinhibitory domain’ of the Ca 2+ pump, and that removal of this domain by proteolysis, or its modification by calmodulin binding is the reason for the activation of both the ATPase and the p-nitrophenylphosphatase activity of the pump. The results presented in this paper give new information about the mechanism of the two kinds of p-nitrophenylphosphatase and about the nature of the apparent competition between C20W and Ca 2+.

Effect of magnesium ions on red cell membrane properties

Spectrin forms aggregates in solution when incubated at relatively high concentrations (several millimolar) of divalent cations. According to the evidence of electron microscopy, aggregates of globular appearance and rather uniform size are cooperatively formed from spectrin dimers, no intermediate structures being seen. Inter-dimer chemical cross-linking of spectrin in intact red cell membranes is enhanced if magnesium ions at a concentration of 0.5 mM or more are present. On the other hand, the elimination of magnesium from the interior of intact cells causes no significant change in shear elastic modulus, measured by micropipette assays, nor is there any dependence of membrane filtration rate on intracellular free magnesium concentration in the range 0-1 mM. Magnesium-depleted cells are, however, converted into echinocytes within a short period, in which control cells, exposed to ionophore and external magnesium ions, remain completely discoid. Magnesium-depleted cells also undergo structural changes on heating below the temperature at which vesiculation sets in. These reveal themselves by the transformation of the cells to a unique characteristic shape, by grossly reduced filtrability, and by extensive agglutination of the cells when treated with a bifunctional reagent. Magnesium ions thus regulate the stability, but not to any measurable extent the gross elasticity, of the red cell membrane.

Partial covalent labeling with pyridoxal 5′-phosphate induces bis(sulfosuccinimidyl)suberate crosslinking of band 3 protein tetramers in intact human red blood cells

Partial covalent labeling of band 3 protein lysines with pyridoxal 5'-phosphate (a substrate and affinity probe) changes the bis(sulfosuccinimidyl)suberate crosslinking pattern of band 3 in intact red cells from a mixture of dimers and tetramers to all tetramers as the exclusive crosslinked product. This is the first demonstration of band 3 crosslinkage to the tetrameric level within membranes of intact red cells. The possible implications of the ligand-induced change in the band 3 crosslinking pattern are discussed.

Pre-steady-state phosphorylation of the human red cell Ca2+-ATPase.

The pre-steady-state kinetics of phosphorylation of the Ca2+-ATPase by ATP was studied at 37 degrees C and in intact red cell membranes to approach physiological conditions. ATP and Ca2+ activate with K0.5 of 4.9 and 26.4 microM, respectively. Preincubation with Ca2+ did not change the K0.5 for ATP. Preincubation with ATP did not alter the initial velocity of phosphorylation suggesting that binding of ATP was not rate-limiting. Mg2+ added at the start of the reaction increased the initial rate of phosphorylation from 4 to 8 pmol/mg/s. With 30 microM Ca2+, the K0.5 for Mg2+ was 60 microM. Mg2+ and Ca2+ added together beforehand accelerated phosphorylation to 70 pmol/mg/s. Phosphorylation of calmodulin-bound membranes was the fastest (280 pmol/mg/s), and its time course showed a neat overshoot before steady state. The results suggest that either preincubation with Ca2+ plus Mg2+ or calmodulin accelerated phosphorylation shifting toward E1 the equilibrium between the E1 and E2 conformers of the enzyme. K+ had no effect on the initial rate of phosphorylation and lowered by 40% the steady-state level of phosphoenzyme in the absence of Mg2+. Phosphorylation is not rate-limiting for the overall reaction since its initial rate was always higher than ATPase activity. In the absence of K+, the turnover of the phosphoenzyme was 2000 min-1, which is close to the values for other transport ATPases.

Fluidity of intact erythrocyte membranes. Correction for fluorescence energy transfer from diphenylhexatriene to hemoglobin

Membranes of intact erythrocyte were labeled by the fluorescent probe 1,6-diphenyl-1,3,5-hexatriene (DPH) using an improved labeling procedure described previously (Plášek J. and Jarolím, P. (1987) Gen. Physiol. Biophys. 6, 425–437). The relationship between the steady-state DPH fluorescence anistoropy r and the mean corpuscular hemoglobin concentration (MCHC) was studied. Fluorescence anisotropy increased with increasing MCHC. A linear dependence of r = 0.0026 ( MCHC) + 0.113 was obtained which enabled us to measure the fluidity of intact red cell membranes. Without this correction for fluorescence quenching by hemoglobin, incorrect conclusions about membrane fluidity could be made. This fact is demonstrated in a group of pyruvate kinase deficient patients compared with a group of healthy blood donors.

Quantitative Immunoferritin Microscopy of Fya Fyb, Jka U, and Dib Antigen Site Numbers on Human Red Cells

The Fya, Fyb, Jka, U, and Dib antigen site numbers and ultrastructural distribution patterns on the human erythrocyte membrane were determined using quantitative immunoferritin microscopy. For homozygous antigen-positive red cells, the average number of determinants per red cell was about 14,000 for Jka, 17,000 for Fya and Fyb, 19,000 for Dib, and 23,000 for the U antigen, assuming that the equilibrium binding observed represented 80% saturation of the accessible antigen sites. The site numbers for this group of antigens were less than that for the Rh antigens, but considerably more than the Kell and Cellano antigens. The technique used was capable of demonstrating a twofold difference in antigen density between heterozygous and homozygous Fy (a+) red cells. More than 85% of the Fya and Fyb antigen sites were lost following pretreatment of the red cells with papain, consistent with the serologic lability of the Fy antigens following proteolysis. The ferritin distribution observed following conjugate staining of antibody-sensitized ghost membranes was similar for all five antigens studied and showed a random, clustered ferritin pattern. Although the quantitative estimates are valid, the remarkable similarity in antigen distribution pattern for this diverse group of antigens, as well as other considerations, suggest that the findings with ghose membranes probably do not reflect faithfully the antigen arrangement on the intact red cell membrane.

Quantitative immunoferritin microscopy of Fya, Fyb, Jka, U, and Dib antigen site numbers on human red cells

The Fya, Fyb, Jka, U, and Dib antigen site numbers and ultrastructural distribution patterns on the human erythrocyte membrane were determined using quantitative immunoferritin microscopy. For homozygous antigen- positive red cells, the average number of determinants per red cell was about 14,000 for Jka, 17,000 for Fya and Fyb, 19,000 for Dib, and 23,000 for the U antigen, assuming that the equilibrium binding observed represented 80% saturation of the accessible antigen sites. The site numbers for this group of antigens were less than that for the Rh antigens, but considerably more than the Kell and Cellano antigens. The technique used was capable of demonstrating a twofold difference in antigen density between heterozygous and homozygous Fy (a+) red cells. More than 85% of the Fya and Fyb antigen sites were lost following pretreatment of the red cells with papain, consistent with the serologic lability of the Fy antigens following proteolysis. The ferritin distribution observed following conjugate staining of antibody- sensitized ghost membranes was similar for all five antigens studied and showed a random, clustered ferritin pattern. Although the quantitative estimates are valid, the remarkable similarity in antigen distribution pattern for this diverse group of antigens, as well as other considerations, suggest that the findings with ghose membranes probably do not reflect faithfully the antigen arrangement on the intact red cell membrane.

The effect of cholesterol and other intercalated amphipaths on the contour and stability of the isolated red cell membrane.

Three membrane properties were strikingly affected when the cholesterol of human erythrocytes, normally approximately 0.8 mol/mol of phospholipid (i.e. C/P approximately 0.8), was altered by equilibration with phospholipid dispersions of an appropriate cholesterol content. 1) While the sterol in intact red cell membranes of C/P less than or equal to 0.8 was resistant to cholesterol oxidase digestion, enrichment to C/P greater than or equal to 0.9 rendered the entire cholesterol pool sensitive to enzyme attack. Susceptibility to oxidation was reversed by removal of the excess cholesterol. Treatment of cells with 1 X 10(-4) M chlorpromazine also rendered the entire cholesterol pool susceptible to cholesterol oxidase. 2) Whereas ghosts with C/P less than or equal to 0.8 underwent invagination in low ionic strength, alkaline buffers to form inside-out vesicles, enrichment to C/P greater than or equal to 1.0 promoted right-side-out vesicle formation instead. This effect was mimicked by treatment with 2,4-dinitrophenol, an amphipath known to cause eversion of the membrane of red cells. In contrast, the presence of chlorpromazine, which promotes invagination of the intact red cell membrane, favored the formation of inside-out vesicles in ghosts. 3) The breakdown of ghosts into endocytic vesicles in dilute, alkaline media and the concomitant release and dissolution of the submembrane reticulum of spectrin and actin were retarded by excess cholesterol and promoted by its removal. This cholesterol effect was mimicked by exposing ghosts to chlorpromazine, but not dinitrophenol. Our data suggest that cholesterol may act physiologically both to stabilize the red cell membrane and to constrain its contour against invagination, and that red cell membrane cholesterol is maintained in vivo just below a critical level at which important organizational changes can occur.

Variable Ca sensitivity of a K-selective channel in intact red-cell membranes

THE membranes of many cells exhibit a K-selective permeability mechanism with a gating process controlled by the concentration of Ca2+ in the cytosol1. Originally described in human red cells2, where its function still remains unknown, this or similar mechanisms3 seem to mediate important physiological actions in other cells4–8.

Mechanism of hemolysis induced by nematocyst venom: Roles of phospholipase A and direct lytic factor

Pure venom from the acontial nematocysts of the sea anemone Aiptasia pallida exhibited phospholipase A (phosphatide acyl-hydrolase; EC 3.1.1.4) activity on a mixture of free phospholipids. Diethyl aminoethyl cellulose fractionation of the venom gave four distinct protein peaks with the phospholipase A activity being restricted to fractions III and IV. These two fractions tested separately also were able to lyse red blood cells weakly. Fractions I and II enhanced the hemolytic activity of fractions III and IV, with fractions I and III giving as much as ninefold enhancement over that of III alone. Fraction I appears analogous to the direct lytic factor of some snake and bee venoms. Fraction III, which could not appreciably hydrolyze the phospholipids of the intact red cell membrane, was able to do so in the presence of fraction I. The sequential interactions of these two nematocyst venom proteins with the red blood cell membrane to produce hemolysis is discussed.

Dual effect of 2,3-diphosphoglycerate on the Bohr effects of human blood.

The influence of the red cell concentration of 2,3-diphosphoglycerate (2,3-DPG, 0.5-26 mumoles/g erythrocytes) on the "CO2-Bohr effect" (pH varied by CO2 at constant base excess) and the "fixed acid-Bohr effect" (pH varied by fixed acid or base at constant PCO2) was studied in human blood at plasma pH values ranging between pH 7.2 and pH 7.6. Elevation of red cell 2,3-DPG concentration leads to a numerical decrease of the "CO2-Bohr coefficient" referring to plasma pH. The "fixed acid-Bohr coefficients" are numerically smaller than the corresponding "CO2-Bohr coefficients" and exhibit a maximum at normal red cell 2,3-DPG concentrations. The Bohr coefficients referring to red cell pH are distinctly higher than those referring to plasma pH, especially at high 2,3-DPG levels. This is due on the one hand to the physico-chemical properties of the intact red cell membrane, and on the other hand to a 2,3-DPG-induced decrease in the ratio deltapHcell/deltapHplasma. From the results it is concluded that 2,3-DPG exerts a dual effect on the Bohr coefficients of whole blood which is mediated 1. by the direct effect of 2,3-DPG on the allosteric properties of hemoglobin (as reflected by changes of the Bohr coefficients referring to red cell pH), and 2. by the effect of 2,3-DPG on deltapHcell/deltapHplasma.

Permeability of bimolecular membranes made from lipid extracts of human red cell ghosts to sugars

Spherical lipid bimolecular membranes of a large surface area separating two aqueous solutions were formed from the total lipid extracts of human red cell ghosts and from their individual lipid components. The isotopic permeabilities of these membranes to biologically important sugars and to a related polyol were measured. The observedD-glucose permeabilities of the bimolecular membranes of the total lipid, phosphatidyl choline, phosphatidyl ethanolamine, sphingomyelin, and cholesterol were 2.35, 2.51, 2.23, 1.35, and 0.62×10(-10) cm/sec, respectively. These permeabilities are about four to five orders of magnitude lower than that of the intact red cell membrane. The permeabilities of the bimolecular membrane made from an identical extract of the total lipid to different sugars varied: the values forD-glucose,D-mannose,D-ribose,D-fructose, 2-deoxy-D-glucose, 3-0-methyl-D-glucose, andD-mannitol were 2.3, 2.6, 8.9, 0.38, 16.1, 11.2, and 0.44×10(-10) cm/sec, respectively. The pattern of the difference is neither parallel with nor as extensive as that observed with the intact red cell membrane. The observed permeabilities of the lipid membranes, however, agree qualitatively with what is predicted by an analysis of non-specific movements of nonelectrolytes across the cell membranes. The failure of the membrane lipids to reproduce the carrier function in a structure most closely approximating that of living membranes strongly suggests that some membrane components other than lipids are required for this function.