Complement activation by neurofibrillary tangles in Alzheimer's disease.

Brain inflammation is widely documented to occur in Alzheimer's disease (AD), but its sources are still incompletely understood. Here, we present in vitro and in situ evidence that, like amyloid beta peptide (Abeta), tau, the major protein constituent of the neurofibrillary tangle, is a potent, antibody-independent activator of the classical complement pathway. Complement activation, in turn, is known to drive numerous inflammatory responses, including scavenger cell activation and cytokine production. Because Abeta deposits and extracellular tangles are present from early preclinical to terminal stages of AD, their ability to activate complement provides a ready mechanism for initiating and sustaining chronic, low-level inflammatory responses that may cumulate over the disease course.

Characterization of the glycolipid associated with Alzheimer paired helical filaments.

In the present study, analytical techniques including gas chromatography/mass spectrometry (GC/MS)-assisted carbohydrate linkage-analysis, one- and two-dimensional NMR, and matrix-assisted laser desorption/ionization time of flight mass spectroscopy (MALDI-MS) have been used to characterize the structure of the glycolipid associated with the paired helical filaments (PHF) isolated from the neurofibrillary tangles of Alzheimer's diseased brain. The 1H NMR spectrum of acid-hydrolyzed protein-resistant core PHF (prcPHF) displays resonances that can be assigned to fatty acid and glucose. There are no resonances present that would indicate the presence of protein, amino acids, or a sphingosine base. Using two-dimensional homonuclear correlated spectroscopy, homonuclear Hartmann-Hahn, and heteronuclear multiple quantum coherence experiments, resonances in the 1H and 13C NMR spectrum of native PHF were assigned to a nonreducing terminal alpha-1,6-glycosidically linked glucose, an internal alpha-1,6-linked glucose, and an alpha-1,2,6-linked glucose. The narrow line-widths observed for these residues suggest that they arise from glucose residues undergoing rapid segmental motion. The carbohydrate portion of the PHF-associated glycolipid was analyzed using GC/MS linkage analysis and confirmed the presence of terminal and internal alpha-1,6-linked glucose and alpha-1,2,6-linked glucose in a molar ratio of 2:1:1. Three components of the PHF-associated glycolipid fraction having masses 2,416, 2,325, and 2,237 Da were observed using MALDI-MS. The least abundant, heavier mass component (2,416 Da) was best fit to a structure with a tridecamer of glucose having a single esterified C20 fatty acid (Glc13 + C20 or Glc13 + C20:1), whereas the more abundant, lower mass components were best fit to noncovalently associated glycolipid dimers, each with a glucose pentamer or hexamer having two C14, C16, or C18 esterified fatty acids {D[(Glc5 + C18) + (Glc6 + C16)] or D[(Glc5 + C14) + (Glc6 + C14)]}. The ratio of glucose to fatty acid calculated from these best-fit structures of the more abundant mass components (5.5 +/- 1.1:1.0) is in reasonable agreement with the same ratio calculated from peak integrations in the NMR spectra of acid-hydrolyzed prcPHF (6.2 +/- 1.6). Structural similarities between PHF-associated glycolipid and other glycolipid amphiphiles known to form PHF-like filaments indirectly suggest that this unique glycolipid may be an integral component of the PHF suprastructure.

Analysis of the core components of Alzheimer paired helical filaments. A gas chromatography/mass spectrometry characterization of fatty acids, carbohydrates and long-chain bases.

We have carried out a fatty acid and carbohydrate compositional analysis of the protease-resistant core of paired helical filaments (prcPHF) isolated from six Alzheimer's diseased brains. Fatty acids, long-chain bases and monosaccharides were characterized by gas chromatography/mass spectrometry (GC/MS) of fatty acid methyl esters, trimethylsilylated long-chain bases, peracetylated alditol acetates and trimethylsilyl methyl glycosides. Glucose and mannose were found to be the only carbohydrate components. Four of the six prcPHF samples contained only glucose while the remaining two samples contained between 30-40% mannose in addition to glucose. None of the samples were found to contain either hydroxylated fatty acids or long-chain bases. The average fatty acid profile of prcPHF was highest in stearic (C18:0) and palmitic acids (C16:0) with less than 10% unsaturated fatty acids. By comparing the carbohydrate and lipid composition of prcPHF to similar data for other brain glycolipids, it was determined that prcPHF is a unique glycolipid, distinct from cerebrosides, gangliosides or brain phospholipids. The fatty acid and carbohydrate composition of a glycolipid isolated from a population of normal brains according to the prcPHF protocol was found to be identical to that of prcPHF glycolipid. It is possible that subtle differences in structure or indigenous factors are responsible for the initiation of PHF formation in vivo.

Assembly of Alzheimer-like, insoluble filaments from brain cerebrosides.

Amphiphiles are molecules which contain a polar head and a hydrophobic tail. When the head contains a chiral center, amphiphiles, incubated in the presence of some di- and trivalent metal ions, have been shown to form large fibrous molecular aggregates. In this study, naturally occurring brain cerebroside was tested to determine if it had sufficient amphiphilic properties to form similar supramolecular structures. When galactocerebroside was heated in the presence of magnesium, it was able to form tubules, vesicles and filaments. The filaments included long fibrils that aggregated into dense bundles and short fibrils that were associated to form smaller bundles. These fibrils were shown to be resistant to solubilization in boiling SDS in the presence of reducing agents. This is the first report of a naturally occurring glycolipid being able to form filaments. Since their structural and physical properties are similar to the paired helical filaments of Alzheimer's disease, they may serve as an experimental model for their assembly.

Utilization of aspartate as a nitrogen source in Escherichia coli. Analysis of nitrogen flow and characterization of the products of aspartate catabolism.

15N NMR, reverse-phase high performance liquid chromatography and gas chromatography/mass spectrometry have been used to follow nitrogen metabolism in Escherichia coli labeled on medium containing L-[15N]aspartate. The flow of 15N through various nitrogen-containing metabolites was followed over the course of the labeling period. For wild-type E. coli labeled on L-[15N]aspartate as sole source of nitrogen, significant 15N labeling was detected only in the intracellular L-glutamate, L-alanine, L-aspartate, and putrescine pools. Intracellular concentrations of L-aspartate and L-glutamate differed significantly in extracts of an arginine auxotroph (argG-), which is deficient in a potential aspartate-assimilating reaction. When the L-[15N]aspartate containing labeling medium was supplemented with unlabeled arginine, extracts of wild-type E. coli were shown to contain significant amounts of unlabeled ammonia and putrescine. There was substantial dilution of 15N in the glutamate pool. The observation that glutamate, aspartate, and alanine are 15N-labeled, but ammonia is not, suggests that these amino acids are not the immediate source of ammonia. Our results suggest that arginine may be an intermediate for the degradation of some of the aspartate and that arginine may be an intermediate for ammonia production during nitrogen-limited growth. Our results also strongly suggest the presence of a previously uncharacterized pathway of arginine degradation.

The application of NMR-pattern-recognition methods to the classification of reduced, peracetylated oligosaccharide residues.

In the present paper homo- and hetero-nuclear correlation spectroscopies have been used to assign proton and carbonyl carbon resonances of a number of reduced, peracetylated mono- and oligo-saccharide derivatives. Each of the native structures for which assignments were made represent residues or substructures typically found in N- or O-linked glycans. Using the assigned NMR parameters as a basis, residues contained in parent structures were classified according to their residue type and glycosidic substitution sites using a relatively simple K-Nearest Neighbor pattern recognition approach. The method was able to correctly assign 99% of 77 "test residues" to their correct structural class using the full set of 19 assigned parameters as a basis. Similar correlations made between data and structure were less successful when reduced variable sets selected on the basis of SIMCA optimization were used.

A convenient method for the preparation of a variety of 13C-substituted D-fructose phosphates using readily available enzymes of the glycolytic pathway.

Methods are presented for the preparation of a variety of D-fructose phosphates, 13C-substituted at any single carbon site or at any two symmetrically disposed carbon sites, from either 13C-substituted pyruvate or L-alanine. It is demonstrated that millimole quantities of product can be obtained in good yield following a "one-pot" incubation of 13C-substituted precursors with commercially available enzymes and cofactors of the glycolytic pathway. Since it has previously been shown that a wide variety of aldehydes serve as acceptable substrates for the final rabbit muscle aldolase-catalyzed condensation step, the method can potentially be applied to prepare a wide variety of 13C-substituted sugars and sugar phosphates.

Alzheimer disease paired helical filament core structures contain glycolipid.

The core structures of sodium dodecyl sulfate extracted, pronase digested paired helical filaments of Alzheimer disease were solubilized by heating in dimethyl sulfoxide. Electron microscopy revealed that after heating in dimethyl sulfoxide, intact paired helical filaments were no longer present in the dimethyl sulfoxide soluble fractions or in the insoluble lipofuscin-containing fractions. Enzyme-linked immunosorbent assays of the various fractions with the monospecific antibody A128 to paired helical filaments demonstrated 96% of the immunoreactivity to be in the dimethyl sulfoxide soluble fraction, and only 4% in the dimethyl sulfoxide insoluble fractions. Lyophilization of the dimethyl sulfoxide soluble supernatant and resuspension in water failed to reassociate the paired helical filaments, but did result in an insoluble precipitate. Analysis of the dimethyl sulfoxide solubilized paired helical filament fraction by nuclear magnetic resonance revealed it to be composed of glycolipid in a form that was distinct from similar fractions isolated from normal aged control brains. The aggregation of an altered glycolipid to form paired helical filaments in Alzheimer disease could explain their insolubility.

Phosphorus-31 nuclear magnetic resonance spectroscopy reveals two conformational forms of chloroacetol phosphate-bound triosephosphate isomerase.

Chloroacetol phosphate covalently reacts with Glu-165 in the catalytic center of triosephosphate isomerase. Reaction of the enzyme with the substrate analogue results in two 31P resonances at 6.8 and 5.5 ppm. Dissociation with guanidinium chloride results in a single resonance at 4.5 ppm. Reassociation and redimerization of the triosephosphate isomerase-chloroacetol phosphate complex restores only the resonance at 5.5 ppm. The two 31P resonances appear to represent different conformations of the enzyme which are trapped upon reaction with the affinity label.

The determination of complex carbohydrate structure by using carbonyl carbon resonances of peracetylated derivatives.

Carbonyl carbon resonances have been assigned to specific acetyl substituents in peracetylated derivatives of a variety of di- and tri-saccharides that occur as substructures of N- and O-linked glycoprotein glycans. Assignments were made by correlating shifts of these resonances to previously assigned pyranoid-ring proton and acetyl methyl proton shifts by means of 2D 13C-1H shifts correlation spectra. It was found that, when the shift assignment data for carbonyl carbon atoms, acetyl methyl protons, and pyranoid-ring protons are plotted in three dimensions, patterns appear that are unique to the different types of residues occurring in a parent structure. It is suggested that these shift data can complement existing 1H- and 13C-n.m.r. methods for determining primary structures of complex carbohydrates.

The assignment of carbonyl resonances in 13C-n.m.r. spectra of peracetylated mono- and oligo-saccharides containing D-glucose and D-mannose: an alternative method for structural determination of complex carbohydrates.

In the present study, proton homonuclear (COSY) and 13C-1H heteronuclear shift-correlation, n.m.r. spectroscopies have been used to assign the carbonyl carbon resonances of peracetylated D-gluco- and D-mannopyranose monosaccharides and oligosaccharides containing residues of parent D-glucopyranose monomers. Chemical shifts of these assigned resonances, particularly those arising from acetyl groups near to aglycon substitution sites, were found to be sensitive to the position and configuration of glycosidic linkages present. In addition, evidence is presented that indicates that the shifts of these carbonyl carbon resonances depend on long-range interactions with other peracetylated pyranose monomers resulting from folding of the oligosaccharide chain. These results suggest that carbonyl carbon resonances of peracetylated carbohydrates may be useful probes of oligosaccharide structure.

Metal ion binding properties of hen ovalbumin and S-ovalbumin: characterization of the metal ion binding site by 31P NMR and water proton relaxation rate enhancements.

In this study, water proton relaxation rate (PRR) enhancements have been used to characterize the binding of metal ions to native ovalbumin, ovalbumin in which phosphate has been enzymatically cleaved from one or both of the two protein phosphoserines, and a heat-stabilized form of the protein (S-ovalbumin). With Scatchard plots constructed from water PRR enhancements, it was found that native ovalbumin and S-ovalbumin had one strong binding site for Mn2+ ion (KD approximately equal to 6.0 X 10(-4) M). Alkaline phosphatase treated ovalbumin, a protein having a single phosphoserine, had one Mn2+ binding site of slightly weaker affinity (KD approximately equal to 8.3 X 10(-4) M), while acid phosphatase treated ovalbumin, a dephosphorylated protein, had two much weaker Mn2+ ion binding sites (KD approximately equal to 1.3 X 10(-3) M). Competitive binding studies on the native protein suggested that Zn2+ ion competes with Mn2+ for the single strong-affinity site (KD approximately equal to 6.1 X 10(-3) M) while Mg2+ and Ca2+ do not. In a second set of experiments, the paramagnetic contribution to the 31P spin-lattice (T1P) and spin-spin (T2P) relaxation times at three separate magnetic field strengths was measured. Correlation times tau c characterizing Mn2+-31P dipolar relaxation were estimated from the ratios of T1P/T2P at a single field and from the ratios of spin-lattice relaxation rates at three different field strengths. The correlation times so obtained, ranging from about 0.7 to 7.7 ns at the three field strengths, were used in calculating distances from the bound Mn2+ ion to the phosphoserines of native ovalbumin, S-ovalbumin, and alkaline phosphatase treated ovalbumins. It was determined that the phosphate of phosphoserine-68 was 5.95 +/- 0.26 and 6.29 +/- 0.18 A from the Mn2+ in the native and alkaline phosphatase treated protein, respectively, and 6.99 +/- 0.30 A away from the Mn2+ in S-ovalbumin. The phosphate of phosphoserine-344 was determined to be 5.31 +/- 0.20 and 5.75 +/- 0.10 A from the Mn2+ ion in native ovalbumin and S-ovalbumin, respectively. The 13C nucleus of [1-13C]galactose enzymatically transferred to the nonreducing end of the ovalbumin oligosaccharide chain was not found to be significantly relaxed by Mn2+ bound to the protein, even at 1:1 stoichiometric ratio of metal:protein. Using this, we estimate the nonreducing terminal of the ovalbumin oligosaccharide to be at least 39 A from the metal ion binding site on the protein.

Amine inversion in proteins. A 13C-NMR study of proton exchange and nitrogen inversion rates in N epsilon,N epsilon,N alpha,N alpha-[13C]tetramethyllysine,N epsilon,N epsilon,N alpha,N alpha-[13C]tetramethyllysine methyl ester, and reductively methylated concanavalin A.

Exchange rates were calculated as a function of pH from line widths of methylamine resonances in 13C-NMR spectra of N epsilon,N epsilon,N alpha,N alpha-[13C]tetramethyllysine (TML) and N epsilon,N epsilon,N alpha,N alpha-tetramethyllysine methyl ester (TMLME). The pH dependence of the dimethyl alpha-amine exchange rate could be adequately described by assuming base-catalyzed chemical exchange between two diastereotopic methyl populations related by nitrogen inversion. Deprotonation of the alpha-amine was assumed to occur by proton transfer to (1) OH-, (2) water, (3) a deprotonated amine or (4) RCO2-. Microscopic rate constants characterizing each of these transfer processes (k1, k2, k3 and k4, respectively) were determined by fitting the rates calculated from line width analysis to a steady-state kinetic model. Using this procedure it was determined that for both TML and TMLME k2 approximately equal to 1-10 M-1 s-1, k3 approximately equal to 10(6) M-1 s-1 and ki, the rate constant for nitrogen inversion was about 10(8)-10(9) s-1. Upper limits of 10(12) and 10(3) M-1 s-1 could be determined for k1 and k4, respectively. A similar kinetic analysis was used to explain pH-dependent line-broadening effects observed for the N-terminal dimethylalanyl resonance in 13C-NMR spectra of concanavalin A, reductively methylated using 90% [13C]formaldehyde. From exchange data below pH 4 it could be determined that amine inversion was limited by the proton transfer rate to the solvent, with a rate constant estimated at 20 M-1 s-1. Above pH 4, exchange was limited by proton transfer to other titrating groups in the protein structure. Based upon their proximity, the carboxylate side chains of Asp-2 and Asp-218 appear to be likely candidates. The apparent first-order microscopic rate constant characterizing proton transfer to these groups was estimated to be about 1 X 10(4) s-1. Rate constants characterizing nitrogen inversion (ki), proton transfer to OH- (k1) and proton transfer to the solvent (k2) were estimated to be of the same order of magnitude as those determined for the model compounds. On the basis of our results, it is proposed that chemical exchange processes associated with base-catalyzed nitrogen inversion may contribute to 15N or 13C spin-lattice relaxation times in reductively methylated peptides or proteins.

An NMR study of 13C-enriched galactose attached to the single carbohydrate chain of hen ovalbumin. Motions of the carbohydrates of glycoproteins.

In order to facilitate the study of motions of the carbohydrate moiety of glycoproteins, a method of enzymatically attaching uniformly 13C-enriched galactopyranose ([U-13C]galactose) onto the nonreducing end of the carbohydrate chain has been developed. Specifically, [U-13C]glucose (20% 13C) was converted through a sequence of four enzymatic steps to [U-13C]UDP-galactose which was then attached, via a beta-1,4 glycosidic bond, to a nonreducing terminal N-acetylglucosamine of hen ovalbumin using the enzyme bovine galactosyltransferase. The overall reaction was followed using 31P and 13C NMR. 13C NMR spectra of the modified ovalbumin show six well resolved resonances arising from the six carbons of the attached [U-13C]galactose. Values of the spin-lattice relaxation time, the line width, the 13C (1H) nuclear Overhauser effect, and the intensity ratio, a measure of the off-resonance rotating frame spin-lattice relaxation time, were obtained for the individual carbons of the [U13-C]galactose as well as other carbon resonances of the protein. Relaxation data for the protein alpha-carbon and aromatic methine carbon resonances could be fit to theoretical values by assuming a correlation time of 25 ns for overall isotropic motion of the glycoprotein. Relaxation data for the anomeric carbon of the attached [U-13C]galactose could only be fit by assuming an additional correlation time between 40 and 80 ps for free internal motion, subject to the constraint that its carbon-proton bond vector makes an angle of about 30 degrees with an effective axis of rotation.

Studies of chemically modified histidine residues of proteins by carbon 13 nuclear magnetic resonance spectroscopy. Reaction of hen egg white lysozyme with iodoacetate.

It is shown that natural abundance 13C NMR spectroscopy can be used to determine the structures and relative amounts of chemically modified forms of a histidine residue of a peptide or protein. The unfractionated product of the reaction of N alpha-acetyl-L-histidine with bromoacetate yields four resonances of nonprotonated aromatic carbons. These resonances are assigned (on a one-to-one basis) to C gamma of the intact amino acid, the two monocarboxymethylated derivatives (at N delta1 and N epsilon2), and the dicarboxymethylated derivative. The effect of pH on the chemical shift of C gamma is characteristic for each of the four species. This property is used to study the carboxymethylation of His-15 of hen egg white lysozyme upon treatment with iodoacetate. With the use of various reaction conditions, His 15 is carboxymethylated in detectable quantities only at N epsilon2. The spectra of the various reaction mixtures indicate which conditions are best for maximizing the yield of this derivative. A comparison of the spectrum of chromatographically pure [N epsilon2-carboxymethylhistidine-15]lysozyme with that of the intact protein indicates that the chemical modification does not significantly affect the conformation of the protein (at least in the regions of all aromatic amino acid residues).