Amino-terminal modification and tyrosine phosphorylation of [corrected] carboxy-terminal fragments of the amyloid precursor protein in Alzheimer's disease and Down's syndrome brain

The carboxy-terminal fragments (CTFs) of the amyloid precursor protein (APP) are considered beta-amyloid (Abeta) precursors as well as molecular species possibly amyloidogenic and neurotoxic by [corrected] in vitro or in animal models. The CTF's role in the pathogenesis of Alzheimer's disease (AD) is however relatively unexplored in human brain. In this study, we analyzed brain extracted CTFs in subjects with AD, non-AD control, and Down's syndrome (DS) cases. Our data indicate that: (i) In fetal DS subjects CTFs levels are increased in comparison to age-matched control, suggesting that the enhanced CTFs formation is important for the early occurrence of plaques deposition in DS. No significant difference in CTFs level [corrected] between AD and age-matched control cases. (ii) CTFs modified at their N-terminus are the direct precursors of similarly N-terminally modified Abeta peptides, which constitute the most abundant species in AD and DS plaques. This observation suggests that N-truncated Abeta peptides are formed directly at beta-secretase level and not through a progressive proteolysis of full-length Abeta1-40/42. (iii) Among the differently cleaved CTFs, only the 22- and 12.5-kDa CTF polypeptides are tyrosine phosphorylated in both AD and control brain while the full-length APP and the CTFs migrating below the 12.5-kDa marker are not phosphorylated, suggesting that APP and CTFs may be involved in different pathways depending on their length and sequences. This study provides evidence that CTFs constitute in human brain a molecular species directly involved in AD pathogenesis and in the development of the AD-like pathology in DS subjects.

Opposite roles of apolipoprotein E in normal brains and in Alzheimer's disease

We have characterized the interaction between apolipoprotein E (apoE) and amyloid beta peptide (Abeta) in the soluble fraction of the cerebral cortex of Alzheimer's disease (AD) and control subjects. Western blot analysis with specific antibodies identified in both groups a complex composed of the full-length apoE and Abeta peptides ending at residues 40 and 42. The apoE-Abeta soluble aggregate is less stable in AD brains than in controls, when treated with the anionic detergent SDS. The complex is present in significantly higher quantity in control than in AD brains, whereas in the insoluble fraction an inverse correlation has previously been reported. Moreover, in the AD subjects the Abeta bound to apoE is more sensitive to protease digestion than is the unbound Abeta. Taken together, our results indicate that in normal brains apoE efficiently binds and sequesters Abeta, preventing its aggregation. In AD, the impaired apoE-Abeta binding leads to the critical accumulation of Abeta, facilitating plaque formation.

Heterogeneity of water-soluble amyloid beta-peptide in Alzheimer's disease and Down's syndrome brains

Water-soluble amyloid beta-peptides (sA beta), ending at residue 42, precede amyloid plaques in Down's syndrome (DS). Here we report that sA beta consists of the full-length A beta(1-42) and peptides truncated and modified by cyclization of the N-terminal glutamates, A beta[3(pE)-42] and A beta[11(pE)-42]. The A beta[3(pE)-42] peptide is the most abundant form of sA beta in Alzheimer's disease (AD) brains. In DS, sA beta[3(pE)-42] concentration increases with age and the peptide becomes a dominant species in the presence of plaques. Both pyroglutamate-modified peptides and the full-length A beta form a stable aggregate that is water soluble. The findings point to a crucial role of the aggregated and modified sA beta in the plaque formation and pathogenesis of AD.

Presence of soluble amyloid beta-peptide precedes amyloid plaque formation in Down's syndrome

Abnormal and excessive accumulation of the amyloid beta-peptide (A beta) in the brain is a major and common characteristic of all Alzheimer's disease (AD) forms irrespective of their genetic background. Insoluble aggregates of A beta are identified as amyloid plaques. These deposits are thought to form when the amount of A beta is increased in the brain parenchyma as a result of either overexpression or altered processing of the amyloid precursor protein (APP). Soluble A beta ending at carboxyl-terminal residue 40 (A beta 40) and, in lesser amount, the form ending at residue 42 (A beta 42), are normal products of the APP metabolism in cell cultures. Increased secretion of soluble A beta 42 has been observed in cells transfected with constructs modeling APP gene mutations of familial forms of AD (refs 4, 5). On the basis of these in vitro data it has been hypothesized that the presence of soluble A beta 42 plays a role in the formation of amyloid plaques. Subjects affected by Down's syndrome (DS) have an increased APP gene dosage and overexpress APP. Apparently because of this overexpression, they almost invariably develop amyloid deposits after the age of 30 years, although they are free of them at earlier ages. Moreover, it has been observed that A beta 42 precedes A beta 40 in the course of amyloid deposition in DS brain. Thus, DS subjects provide the opportunity to investigate in the human brain the metabolic conditions that precede the formation of the amyloid deposits. Here we report that soluble A beta 42 is present in the brains of DS-affected subjects aged from 21 gestational weeks to 61 years but it is undetectable in age-matched controls. It is argued that overexpression of APP leads specifically to A beta 42 increase and that the presence of the soluble A beta 42 is causally related to plaque formation in DS and, likely, in AD brains.

Truncated forms of the human prion protein in normal brain and in prion diseases

The cellular form of the prion protein (PrPc) is a glycoprotein anchored to the cell membrane by a glycosylphosphatidylinositol moiety. An aberrant form of PrPc that is partially resistant to proteases, PrPres, is a hallmark of prion diseases, which in humans include Cruetzfeldt-Jakob disease (CJD), Gerstmann-Sträussler-Scheinker syndrome, and fatal familial insomnia. We have characterized the major forms of PrP in normal and pathological human brains. A COOH-terminal fragment of PrPc, designated C1, is abundant in normal and CJD brains as well as in human neuroblastoma cells. Sequence analysis revealed that C1 contains alternative NH2 termini starting at His-111 or Met-112. Like PrPc, C1 is glycosylated, anchored to the cell membrane, and is heat-stable. Consistent with the lack of the NH2-terminal region of PrPc, C1 is more acidic than PrPc and does not bind heparin. An additional fragment longer than C1, designated C2, is present in substantial amounts in CJD brains. Like PrPres, C2 is resistant to proteases and is detergent-insoluble. Our data indicate that C1 is a major product of normal PrPc metabolism, generated by a cleavage that disrupts the neurotoxic and amyloidogenic region of PrP comprising residues 106-126. This region remains intact in C2, suggesting a role for C2 in prion diseases.

Correlation of intron-exon organisation with the three-dimensional structure in glutamate dehydrogenase

The positions of the intron-exon boundaries in the genes for glutamate dehydrogenase from Chlorella sorokiniana rat, and human have been located on the three-dimensional structure of the highly homologous enzyme from Clostridium symbiosum and analysed for their position in the protein structure. This analysis shows no correlation between the positions of these boundaries in the mammalian and Chlorella glutamate dehydrogenase genes and no correlation with units of function in the enzyme and suggests that the present day exons do not represent the protein modules of an ancestral glutamate dehydrogenase. There appears to be no clear preference for the residues at the splice junctions to be either buried or exposed to solvent. However, the frequency with which the introns appear in the loops linking elements of secondary structure, rather than in either the alpha-helical or beta-sheet segments, is higher than predicted on the basis of the proportion of residues in the loops. This is consistent with but not proof of a role for exon modification/exchange in protein evolution since changes at these positions are less likely to disturb the structure and hence maintain function.

The catalytic role of aspartate in the active site of glutamate dehydrogenase

A putative catalytic aspartyl residue, Asp-165, in the active site of clostridial glutamate dehydrogenase has been replaced with serine by site-directed mutagenesis. The mutant enzyme is efficiently overexpressed in Escherichia coli as a soluble protein and can be successfully purified by the dye-ligand chromatographic procedure normally employed for the wild-type enzyme. By several criteria, including circular dichroism spectrum, sulphydryl reactivity with Ellman's reagent, crystallization and mobility in non-denaturing electrophoresis, the enzyme appears to be correctly folded. NAD+ protects the D165S mutant against modification by Ellman's reagent, suggesting unimpaired binding of coenzyme. In standard assays the specific activity is decreased 10(3)-fold in the reductive amination reaction and 10(5)-fold for oxidative deamination. Kinetic studies show that apparent Km values for NADH and 2-oxoglutarate are almost unchanged. The large reduction in the reaction rate coincides with a weakening of the affinity for ammonium ion (Km > 300 mM, compared with 60 mM for the wild-type). The data are entirely consistent with the direct involvement of D165 in catalysis rather than in the binding of coenzyme or 2-oxoglutarate.

Glutamate-malate metabolism in liver mitochondria. A model constructed on the basis of mitochondrial levels of enzymes, specificity, dissociation constants, and stoichiometry of hetero-enzyme complexes

The level of aspartate aminotransferase in liver mitochondria was found to be approximately 140 microM, or 2-3 orders of magnitude higher than its dissociation constant in complexes with the inner mitochondrial membrane and the high molecular weight enzymes (M(r) = 1.6 x 10(5) to 2.7 x 10(6)) carbamyl-phosphate synthase I, glutamate dehydrogenase, and the alpha-ketoglutarate dehydrogenase complex. The total concentration of aminotransferase-binding sites on these structures in liver mitochondria was more than sufficient to accommodate all of the aminotransferase. Therefore, in liver mitochondria, the aminotransferase could be associated with the inner mitochondrial membrane and/or these high molecular weight enzymes. The aminotransferase in these hetero-enzyme complexes could be supplied with oxalacetate because binding of aminotransferase to the high molecular weight enzymes can enhance binding of malate dehydrogenase, and binding of both malate dehydrogenase and the aminotransferase facilitated binding of fumarase. The level of malate dehydrogenase was found to be so high (140 microM) in liver mitochondria, compared with that of citrate synthase (25 microM) and the pyruvate dehydrogenase complex (0.3 microM), that there would also be a sufficient supply of oxalacetate to citrate synthase-pyruvate dehydrogenase.

Kinetics and regulation of hepatoma mitochondrial NAD(P) malic enzyme

Kinetic studies of Morris 7777 hepatoma mitochondrial NAD(P) malic enzyme were consistent with an ordered mechanism where NAD adds to the enzyme before malate and dissociation of NADH from the enzyme is rate-limiting. In addition to its active site, malate apparently also associates with a lower affinity with an activator site. The activator fumarate competes with malate at the activator site and facilitates dissociation of NADH from the enzyme. The ratio of NAD(P) malic enzyme to malate dehydrogenase activity in the hepatoma mitochondrial extract was found to be too low, even in the presence of known inhibitors of malate dehydrogenase, to account for the known ability of NAD(P) malic enzyme to intercept exogenous malate from malate dehydrogenase in intact tumor mitochondria (Moreadith, R.W., and Lehninger, A.L. (1984) J. Biol. Chem. 259, 6215-6221). However, NAD(P) malic enzyme may be able to intercept exogenous malate because according to the present results, it can associate with the pyruvate dehydrogenase complex, which could localize NAD(P) malic enzyme in the vicinity of the inner mitochondrial membrane. The activity levels of some key metabolic enzymes were found to be different in Morris 7777 mitochondria than in liver or mitochondria of other rapidly dividing tumors. These results are discussed in terms of differences among tumors in their ability to utilize malate, glutamate, and citrate as respiratory fuels.

The glutamate dehydrogenase gene of Clostridium symbiosum. Cloning by polymerase chain reaction, sequence analysis and over-expression in Escherichia coli

The gene encoding the NAD(+)-dependent glutamate dehydrogenase (GDH) of Clostridium symbiosum was cloned using the polymerase chain reaction (PCR) because it could not be recovered by standard techniques. The nucleotide sequence of the gdh gene was determined and it was overexpressed from the controllable tac promoter in Escherichia coli so that active clostridial GDH represented 20% of total cell protein. The recombinant plasmid complemented the nutritional lesion of an E. coli glutamate auxotroph. There was a marked difference between the nucleotide compositions of the coding region (G + C = 52%) and the flanking sequences (G + C = 30% and 37%). The structural gene encoded a polypeptide of 450 amino acid residues and relative molecular mass (M(r) 49,295 which corresponds to a single subunit of the hexameric enzyme. The DNA-derived amino acid sequence was consistent with a partial sequence from tryptic and cyanogen bromide peptides of the clostridial enzyme. The N-terminal amino acid sequence matched that of the purified protein, indicating that the initiating methionine is removed post-translationally, as in the natural host. The amino acid sequence is similar to those of other bacterial GDHs although it has a Gly-Xaa-Gly-Xaa-Xaa-Ala motif in the NAD(+)-binding domain, which is more typical of the NADP(+)-dependent enzymes. The sequence data now permit a detailed interpretation of the X-ray crystallographic structure of the enzyme and the cloning and expression of the clostridial gene will facilitate site-directed mutagenesis.

Interactions among mitochondrial aspartate aminotransferase, malate dehydrogenase, and the inner mitochondrial membrane from heart, hepatoma, and liver

The inner mitochondrial membranes from bovine heart, rat liver, and Morris hepatoma 7777 all bound the mitochondrial isozymes of aspartate aminotransferase and malate dehydrogenase with comparable affinities and binding ratios (mg of enzyme bound per mg of membrane protein). A low molecular weight fraction separated from a detergent extract of the heart membrane by chromatography on Sephacryl S-300 contained most of the binding activity of the extract for the aminotransferase and had a dissociation constant for the aminotransferase of 0.2 microM. The protein component of the membrane binding sites for the aminotransferase was apparently present in this fraction because binding activity was largely eliminated by proteolysis with trypsin. When this fraction was chromatographed on an aminotransferase affinity column, only the portion that was bound and eluted by 0.25 M KCl associated with added aminotransferase. Unlike the membrane, which was markedly inhibited by the non-ionic detergent Genapol but was inhibited only 20% by trypsin, the binding activity of this subfraction was completely inhibited by trypsin but not by Genapol. This suggests, on the membrane, that the aminotransferase binds to the binding protein and is then transferred to lipids specifically associated with the binding protein. These putative lipids are presumably removed on the affinity column. Although the yield of the binding protein was low, there is probably ample binding protein in mitochondria to accommodate the aminotransferase. In every case, binding of the aminotransferase to the membrane inactivated the malate dehydrogenase binding site whereas malate dehydrogenase had little effect on the binding of the aminotransferase and only associated with the higher molecular weight fractions from the Sephacryl column that contained Complex I activity. Inactivation of the malate dehydrogenase site by the aminotransferase, but not vice versa, could result from aminotransferase associating with the binding protein and malate dehydrogenase with Complex I followed by association of the enzymes with lipids located in the same region of the membrane. However, since aminotransferase is more cationic, it is not displaced readily from the lipids by malate dehydrogenase. The relevance of these interactions to the organization of the enzymes is discussed.

Purification of tumor mitochondrial malic enzyme by specific ligand affinity chromatography

A two-step chromatographic procedure, based on a specific ligand-binding approach, for the purification of tumor NAD(P)(+)-dependent malic enzyme is described. The enzyme was purified to near homogeneity by extraction from mitochondria, negative cellulose phosphate chromatography, ammonium sulfate precipitation, and application of specific elution from a malate-agarose column. The rationale for the use of the affinity column is also described.

Regulation of glutamate dehydrogenase by Mg2+ and magnification of leucine activation by Mg2+

In the presence of Mg2+, pure glutamate dehydrogenase is more reactive with NADPH than with NADH and is markedly activated by elevations in the ADP/ATP ratio or the addition of leucine. Because these are properties of glutamate dehydrogenase in mitochondria but not properties of the pure enzyme studied in the absence of Mg2+, Mg2+ could be a ligand that confers upon glutamate dehydrogenase the regulatory properties of this enzyme found in situ. In the absence of the allosteric activators ADP, leucine, or succinyl-CoA, Mg2+ is an inhibitor and increases product inhibition by alpha-ketoglutarate in the forward reaction and substrate inhibition by alpha-ketoglutarate in the reverse reaction. However, the allosteric activators convert Mg2+ from an inhibitor into an activator of the forward reaction. In the reverse reaction, ADP also converts Mg2+ from an inhibitor into an activator and leucine eliminates inhibition by Mg2+. Because Mg2+ is an inhibitor in the absence of activator that also increases inhibition by alpha-ketoglutarate, whereas in the presence of activator Mg2+ has no effect or is itself an activator, Mg2+ magnifies the effect of the activator, and magnification increases with increases in the concentration of alpha-ketoglutarate. Leucine and its analog 2-aminobicyclo (2.2.1) heptane 2-carboxylic acid (BCH) have almost identical effects on both human and bovine glutamate dehydrogenase in both the presence and absence of Mg2+. However, advantages of BCH over leucine as a potential pharmacological activator of glutamate dehydrogenase are that BCH is not metabolized and, unlike leucine, BCH does not inhibit ornithine transcarbamylase. Isoleucine and valine alone have little effect on human glutamate dehydrogenase, but isoleucine slightly inhibits the enzyme in the presence of leucine.

Kinetic advantages of hetero-enzyme complexes with glutamate dehydrogenase and the alpha-ketoglutarate dehydrogenase complex

We have found previously (Fahien, L.A., Kmiotek, E.H., MacDonald, M. J., Fibich, B., and Mandic, M. (1988) J. Biol. Chem. 263, 10687-10697) that glutamate-malate oxidation can be enhanced by cooperative binding of mitochondrial aspartate aminotransferase and malate dehydrogenase to the alpha-ketoglutarate dehydrogenase complex. The present results demonstrate that glutamate dehydrogenase, which forms binary complexes with these enzymes, adds to this ternary complex and thereby increases binding of the other enzymes. Kinetic evidence for direct transfer of alpha-ketoglutarate and NADH, within these complexes, has been obtained by measuring steady-state rates of E2 when most of the substrate or coenzyme is bound to the aminotransferase or glutamate dehydrogenase (E1). Rates significantly greater than those which can be accounted for by the concentration of free ligand, calculated from the measured values of the E1-ligand dissociation constants, require that the E1-ligand complex serve as a substrate for E2 (Srivastava, D. K., and Bernhard, S. A. (1986) Curr. Tops. Cell Regul. 28, 1-68). By this criterion, NADH is transferred directly from glutamate dehydrogenase to malate dehydrogenase and alpha-ketoglutarate is channeled from the aminotransferase to both glutamate dehydrogenase and the alpha-ketoglutarate dehydrogenase complex. Similar evidence indicates that GTP bound to an allosteric site on glutamate dehydrogenase functions as a substrate for succinic thiokinase. The potential physiological advantages to channeling of activators and inhibitors as well as substrates within multienzyme complexes organized around the alpha-ketoglutarate dehydrogenase complex are discussed.

The effect of nucleotides on glutamate dehydrogenase from the mealworm fat body

The effect of nucleotides: AMP, cAMP, ADP, ATP, GDP and GTP, on glutamate dehydrogenase (GDH) purified from the mealworm fat body was studied. Guanine nucleotides and ATP inhibited the enzyme strongly in both directions. GDH was partially protected from the inhibition by the addition of ADP to an assay medium. AMP and cAMP activated the enzyme slightly. The concerted effects of ADP and ATP indicate the importance of adenylate energy charge in the regulation of fat body GDH. It is suggested that GDH may play amphibolic role in the fat body and that the direction of GDH catalysed reaction is under strong influence of nucleotides. The enzyme may synthesize glutamate at high energy charge, but when the energy reserves are low, it oxidizes glutamate.